Redis/redis-For-Windows
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 3 Wiki Activity

Compare commits

base: Redis/redis-For-Windows:Win2.6.8
Branches Tags
Redis/redis-For-Windows:develop Redis/redis-For-Windows:win-5.0 Redis/redis-For-Windows:win-6.2
Redis/redis-For-Windows:5.0.14.1 Redis/redis-For-Windows:v5.0.14.1 Redis/redis-For-Windows:v5.0.14 Redis/redis-For-Windows:v5.0.10 Redis/redis-For-Windows:v5.0.9 Redis/redis-For-Windows:v4.0.14.2 Redis/redis-For-Windows:v4.0.14.1 Redis/redis-For-Windows:v4.0.14 Redis/redis-For-Windows:v4.0.2.3-alpha Redis/redis-For-Windows:v4.0.2.2-alpha Redis/redis-For-Windows:v4.0.2.1-alpha Redis/redis-For-Windows:v4.0.2-alpha Redis/redis-For-Windows:win-3.2.100 Redis/redis-For-Windows:win-3.0.504 Redis/redis-For-Windows:win-3.0.503 Redis/redis-For-Windows:win-2.8.2402 Redis/redis-For-Windows:win-3.0.502 Redis/redis-For-Windows:win-2.8.2401 Redis/redis-For-Windows:3.2.1 Redis/redis-For-Windows:win-3.2.0000-preview Redis/redis-For-Windows:3.2.0 Redis/redis-For-Windows:3.2.0-rc3 Redis/redis-For-Windows:3.2.0-rc2 Redis/redis-For-Windows:win-2.8.2400 Redis/redis-For-Windows:win-3.0.501 Redis/redis-For-Windows:3.2-rc1 Redis/redis-For-Windows:3.0.6 Redis/redis-For-Windows:2.8.24 Redis/redis-For-Windows:win-3.0.500 Redis/redis-For-Windows:win-3.0.500-rc2 Redis/redis-For-Windows:win-3.0.500-rc1 Redis/redis-For-Windows:win-2.8.2104 Redis/redis-For-Windows:3.0.5 Redis/redis-For-Windows:2.8.23 Redis/redis-For-Windows:win-3.0.300-beta1 Redis/redis-For-Windows:win-2.8.2103 Redis/redis-For-Windows:2.8.22 Redis/redis-For-Windows:3.0.4 Redis/redis-For-Windows:win-3.0.300-alpha3 Redis/redis-For-Windows:win-2.8.2102 Redis/redis-For-Windows:win-3.0.300-alpha2 Redis/redis-For-Windows:win-3.0.100-alpha1 Redis/redis-For-Windows:3.0.3 Redis/redis-For-Windows:win-2.8.2101 Redis/redis-For-Windows:win-2.8.21 Redis/redis-For-Windows:3.0.2 Redis/redis-For-Windows:2.8.21 Redis/redis-For-Windows:2.8.20 Redis/redis-For-Windows:3.0.1 Redis/redis-For-Windows:win-2.8.19.1 Redis/redis-For-Windows:3.0.0 Redis/redis-For-Windows:3.0.0-rc6 Redis/redis-For-Windows:3.0.0-rc5 Redis/redis-For-Windows:win-2.8.19-rc1 Redis/redis-For-Windows:win-2.8.19 Redis/redis-For-Windows:3.0.0-rc4 Redis/redis-For-Windows:win-2.8.19-pre2 Redis/redis-For-Windows:win-2.8.17.4 Redis/redis-For-Windows:3.0.0-rc3 Redis/redis-For-Windows:3.0.0-rc2 Redis/redis-For-Windows:win-2.8.17.3 Redis/redis-For-Windows:win-2.8.19-pre1 Redis/redis-For-Windows:win-2.8.17.2 Redis/redis-For-Windows:win-2.8.17.1 Redis/redis-For-Windows:2.8.19 Redis/redis-For-Windows:2.8.18 Redis/redis-For-Windows:3.0.0-rc1 Redis/redis-For-Windows:2.8.17 Redis/redis-For-Windows:2.8.16 Redis/redis-For-Windows:2.8.15 Redis/redis-For-Windows:win2.8.12 Redis/redis-For-Windows:2.8.14 Redis/redis-For-Windows:3.0.0-beta8 Redis/redis-For-Windows:2.8.13 Redis/redis-For-Windows:3.0.0-beta7 Redis/redis-For-Windows:win_2.8.9 Redis/redis-For-Windows:2.8.12 Redis/redis-For-Windows:2.8.11 Redis/redis-For-Windows:3.0.0-beta6 Redis/redis-For-Windows:2.8.10 Redis/redis-For-Windows:3.0.0-beta5 Redis/redis-For-Windows:3.0.0-beta4 Redis/redis-For-Windows:2.6.14_msopentech Redis/redis-For-Windows:2.8.4_msopentech Redis/redis-For-Windows:msopentech_2.8.4 Redis/redis-For-Windows:3.0.0-beta3 Redis/redis-For-Windows:2.8.9 Redis/redis-For-Windows:2.8.8 Redis/redis-For-Windows:3.0.0-beta2 Redis/redis-For-Windows:2.8.7 Redis/redis-For-Windows:2.8.6 Redis/redis-For-Windows:3.0.0-beta1 Redis/redis-For-Windows:2.8.5 Redis/redis-For-Windows:2.8.4 Redis/redis-For-Windows:2.8.3 Redis/redis-For-Windows:2.8.2 Redis/redis-For-Windows:2.8.1 Redis/redis-For-Windows:2.8.0 Redis/redis-For-Windows:2.8.0-rc6 Redis/redis-For-Windows:2.8.0-rc5 Redis/redis-For-Windows:2.8.0-rc4 Redis/redis-For-Windows:2.8.0-rc3 Redis/redis-For-Windows:2.8.0-rc2 Redis/redis-For-Windows:2.8.0-rc1 Redis/redis-For-Windows:2.6.14 Redis/redis-For-Windows:Win2.6.8 Redis/redis-For-Windows:2.6.13 Redis/redis-For-Windows:2.6.12 Redis/redis-For-Windows:2.6.11 Redis/redis-For-Windows:2.6.10 Redis/redis-For-Windows:2.6.9 Redis/redis-For-Windows:2.6.8 Redis/redis-For-Windows:2.6.7 Redis/redis-For-Windows:2.6.6 Redis/redis-For-Windows:2.6.5 Redis/redis-For-Windows:2.6.4 Redis/redis-For-Windows:2.6.3 Redis/redis-For-Windows:2.6.2 Redis/redis-For-Windows:2.6.1 Redis/redis-For-Windows:2.6.0 Redis/redis-For-Windows:2.6.0-rc8 Redis/redis-For-Windows:2.6.0-rc7 Redis/redis-For-Windows:2.6.0-rc6 Redis/redis-For-Windows:2.6.0-rc5 Redis/redis-For-Windows:2.6.0-rc4 Redis/redis-For-Windows:2.6.0-rc3 Redis/redis-For-Windows:2.6.0-rc2 Redis/redis-For-Windows:2.6.0-rc1 Redis/redis-For-Windows:2.4.11 Redis/redis-For-Windows:2.4.10 Redis/redis-For-Windows:2.4.9 Redis/redis-For-Windows:2.4.8 Redis/redis-For-Windows:2.4.6_win32 Redis/redis-For-Windows:2.4.7 Redis/redis-For-Windows:2.4.6 Redis/redis-For-Windows:2.4.5 Redis/redis-For-Windows:2.4.4 Redis/redis-For-Windows:2.4.3 Redis/redis-For-Windows:2.4.2 Redis/redis-For-Windows:2.4.1 Redis/redis-For-Windows:2.4.0 Redis/redis-For-Windows:2.4.0-rc8 Redis/redis-For-Windows:2.4.0-rc7 Redis/redis-For-Windows:2.4.0-rc6 Redis/redis-For-Windows:2.4.0-rc5 Redis/redis-For-Windows:2.4.0-rc4 Redis/redis-For-Windows:2.4.0-rc3 Redis/redis-For-Windows:with-deprecated-diskstore Redis/redis-For-Windows:2.4.0-rc2 Redis/redis-For-Windows:2.4.0-rc1 Redis/redis-For-Windows:2.2.4 Redis/redis-For-Windows:2.2.3 Redis/redis-For-Windows:3.0-alpha0 Redis/redis-For-Windows:2.2.2 Redis/redis-For-Windows:2.2.1 Redis/redis-For-Windows:2.2.0 Redis/redis-For-Windows:2.2.0-rc4 Redis/redis-For-Windows:2.3-alpha0 Redis/redis-For-Windows:2.2.0-rc3 Redis/redis-For-Windows:2.2.0-rc2 Redis/redis-For-Windows:2.2.0-rc1 Redis/redis-For-Windows:2.2-alpha6 Redis/redis-For-Windows:2.2-alpha5 Redis/redis-For-Windows:2.2-alpha4 Redis/redis-For-Windows:2.2-alpha3 Redis/redis-For-Windows:2.2-alpha2 Redis/redis-For-Windows:2.2-alpha1 Redis/redis-For-Windows:2.2-alpha0 Redis/redis-For-Windows:v2.1.1-watch Redis/redis-For-Windows:v2.0.0-rc1 Redis/redis-For-Windows:v1.3.12 Redis/redis-For-Windows:v1.3.11 Redis/redis-For-Windows:v1.3.10 Redis/redis-For-Windows:v1.3.9 Redis/redis-For-Windows:v1.3.8 Redis/redis-For-Windows:v1.3.7 Redis/redis-For-Windows:1.3.6 Redis/redis-For-Windows:vm-playpen
..
compare: Redis/redis-For-Windows:2.8.23
Branches Tags
Redis/redis-For-Windows:win-5.0 Redis/redis-For-Windows:develop Redis/redis-For-Windows:win-6.2
Redis/redis-For-Windows:5.0.14.1 Redis/redis-For-Windows:v5.0.14.1 Redis/redis-For-Windows:v5.0.14 Redis/redis-For-Windows:v5.0.10 Redis/redis-For-Windows:v5.0.9 Redis/redis-For-Windows:v4.0.14.2 Redis/redis-For-Windows:v4.0.14.1 Redis/redis-For-Windows:v4.0.14 Redis/redis-For-Windows:v4.0.2.3-alpha Redis/redis-For-Windows:v4.0.2.2-alpha Redis/redis-For-Windows:v4.0.2.1-alpha Redis/redis-For-Windows:v4.0.2-alpha Redis/redis-For-Windows:win-3.2.100 Redis/redis-For-Windows:win-3.0.504 Redis/redis-For-Windows:win-3.0.503 Redis/redis-For-Windows:win-2.8.2402 Redis/redis-For-Windows:win-3.0.502 Redis/redis-For-Windows:win-2.8.2401 Redis/redis-For-Windows:3.2.1 Redis/redis-For-Windows:win-3.2.0000-preview Redis/redis-For-Windows:3.2.0 Redis/redis-For-Windows:3.2.0-rc3 Redis/redis-For-Windows:3.2.0-rc2 Redis/redis-For-Windows:win-2.8.2400 Redis/redis-For-Windows:win-3.0.501 Redis/redis-For-Windows:3.2-rc1 Redis/redis-For-Windows:3.0.6 Redis/redis-For-Windows:2.8.24 Redis/redis-For-Windows:win-3.0.500 Redis/redis-For-Windows:win-3.0.500-rc2 Redis/redis-For-Windows:win-3.0.500-rc1 Redis/redis-For-Windows:win-2.8.2104 Redis/redis-For-Windows:3.0.5 Redis/redis-For-Windows:2.8.23 Redis/redis-For-Windows:win-3.0.300-beta1 Redis/redis-For-Windows:win-2.8.2103 Redis/redis-For-Windows:2.8.22 Redis/redis-For-Windows:3.0.4 Redis/redis-For-Windows:win-3.0.300-alpha3 Redis/redis-For-Windows:win-2.8.2102 Redis/redis-For-Windows:win-3.0.300-alpha2 Redis/redis-For-Windows:win-3.0.100-alpha1 Redis/redis-For-Windows:3.0.3 Redis/redis-For-Windows:win-2.8.2101 Redis/redis-For-Windows:win-2.8.21 Redis/redis-For-Windows:3.0.2 Redis/redis-For-Windows:2.8.21 Redis/redis-For-Windows:2.8.20 Redis/redis-For-Windows:3.0.1 Redis/redis-For-Windows:win-2.8.19.1 Redis/redis-For-Windows:3.0.0 Redis/redis-For-Windows:3.0.0-rc6 Redis/redis-For-Windows:3.0.0-rc5 Redis/redis-For-Windows:win-2.8.19-rc1 Redis/redis-For-Windows:win-2.8.19 Redis/redis-For-Windows:3.0.0-rc4 Redis/redis-For-Windows:win-2.8.19-pre2 Redis/redis-For-Windows:win-2.8.17.4 Redis/redis-For-Windows:3.0.0-rc3 Redis/redis-For-Windows:3.0.0-rc2 Redis/redis-For-Windows:win-2.8.17.3 Redis/redis-For-Windows:win-2.8.19-pre1 Redis/redis-For-Windows:win-2.8.17.2 Redis/redis-For-Windows:win-2.8.17.1 Redis/redis-For-Windows:2.8.19 Redis/redis-For-Windows:2.8.18 Redis/redis-For-Windows:3.0.0-rc1 Redis/redis-For-Windows:2.8.17 Redis/redis-For-Windows:2.8.16 Redis/redis-For-Windows:2.8.15 Redis/redis-For-Windows:win2.8.12 Redis/redis-For-Windows:2.8.14 Redis/redis-For-Windows:3.0.0-beta8 Redis/redis-For-Windows:2.8.13 Redis/redis-For-Windows:3.0.0-beta7 Redis/redis-For-Windows:win_2.8.9 Redis/redis-For-Windows:2.8.12 Redis/redis-For-Windows:2.8.11 Redis/redis-For-Windows:3.0.0-beta6 Redis/redis-For-Windows:2.8.10 Redis/redis-For-Windows:3.0.0-beta5 Redis/redis-For-Windows:3.0.0-beta4 Redis/redis-For-Windows:2.6.14_msopentech Redis/redis-For-Windows:2.8.4_msopentech Redis/redis-For-Windows:msopentech_2.8.4 Redis/redis-For-Windows:3.0.0-beta3 Redis/redis-For-Windows:2.8.9 Redis/redis-For-Windows:2.8.8 Redis/redis-For-Windows:3.0.0-beta2 Redis/redis-For-Windows:2.8.7 Redis/redis-For-Windows:2.8.6 Redis/redis-For-Windows:3.0.0-beta1 Redis/redis-For-Windows:2.8.5 Redis/redis-For-Windows:2.8.4 Redis/redis-For-Windows:2.8.3 Redis/redis-For-Windows:2.8.2 Redis/redis-For-Windows:2.8.1 Redis/redis-For-Windows:2.8.0 Redis/redis-For-Windows:2.8.0-rc6 Redis/redis-For-Windows:2.8.0-rc5 Redis/redis-For-Windows:2.8.0-rc4 Redis/redis-For-Windows:2.8.0-rc3 Redis/redis-For-Windows:2.8.0-rc2 Redis/redis-For-Windows:2.8.0-rc1 Redis/redis-For-Windows:2.6.14 Redis/redis-For-Windows:Win2.6.8 Redis/redis-For-Windows:2.6.13 Redis/redis-For-Windows:2.6.12 Redis/redis-For-Windows:2.6.11 Redis/redis-For-Windows:2.6.10 Redis/redis-For-Windows:2.6.9 Redis/redis-For-Windows:2.6.8 Redis/redis-For-Windows:2.6.7 Redis/redis-For-Windows:2.6.6 Redis/redis-For-Windows:2.6.5 Redis/redis-For-Windows:2.6.4 Redis/redis-For-Windows:2.6.3 Redis/redis-For-Windows:2.6.2 Redis/redis-For-Windows:2.6.1 Redis/redis-For-Windows:2.6.0 Redis/redis-For-Windows:2.6.0-rc8 Redis/redis-For-Windows:2.6.0-rc7 Redis/redis-For-Windows:2.6.0-rc6 Redis/redis-For-Windows:2.6.0-rc5 Redis/redis-For-Windows:2.6.0-rc4 Redis/redis-For-Windows:2.6.0-rc3 Redis/redis-For-Windows:2.6.0-rc2 Redis/redis-For-Windows:2.6.0-rc1 Redis/redis-For-Windows:2.4.11 Redis/redis-For-Windows:2.4.10 Redis/redis-For-Windows:2.4.9 Redis/redis-For-Windows:2.4.8 Redis/redis-For-Windows:2.4.6_win32 Redis/redis-For-Windows:2.4.7 Redis/redis-For-Windows:2.4.6 Redis/redis-For-Windows:2.4.5 Redis/redis-For-Windows:2.4.4 Redis/redis-For-Windows:2.4.3 Redis/redis-For-Windows:2.4.2 Redis/redis-For-Windows:2.4.1 Redis/redis-For-Windows:2.4.0 Redis/redis-For-Windows:2.4.0-rc8 Redis/redis-For-Windows:2.4.0-rc7 Redis/redis-For-Windows:2.4.0-rc6 Redis/redis-For-Windows:2.4.0-rc5 Redis/redis-For-Windows:2.4.0-rc4 Redis/redis-For-Windows:2.4.0-rc3 Redis/redis-For-Windows:with-deprecated-diskstore Redis/redis-For-Windows:2.4.0-rc2 Redis/redis-For-Windows:2.4.0-rc1 Redis/redis-For-Windows:2.2.4 Redis/redis-For-Windows:2.2.3 Redis/redis-For-Windows:3.0-alpha0 Redis/redis-For-Windows:2.2.2 Redis/redis-For-Windows:2.2.1 Redis/redis-For-Windows:2.2.0 Redis/redis-For-Windows:2.2.0-rc4 Redis/redis-For-Windows:2.3-alpha0 Redis/redis-For-Windows:2.2.0-rc3 Redis/redis-For-Windows:2.2.0-rc2 Redis/redis-For-Windows:2.2.0-rc1 Redis/redis-For-Windows:2.2-alpha6 Redis/redis-For-Windows:2.2-alpha5 Redis/redis-For-Windows:2.2-alpha4 Redis/redis-For-Windows:2.2-alpha3 Redis/redis-For-Windows:2.2-alpha2 Redis/redis-For-Windows:2.2-alpha1 Redis/redis-For-Windows:2.2-alpha0 Redis/redis-For-Windows:v2.1.1-watch Redis/redis-For-Windows:v2.0.0-rc1 Redis/redis-For-Windows:v1.3.12 Redis/redis-For-Windows:v1.3.11 Redis/redis-For-Windows:v1.3.10 Redis/redis-For-Windows:v1.3.9 Redis/redis-For-Windows:v1.3.8 Redis/redis-For-Windows:v1.3.7 Redis/redis-For-Windows:1.3.6 Redis/redis-For-Windows:vm-playpen
1384 Commits
Win2.6.8 .. 2.8.23
Author SHA1 Message Date
antirez 973ce1b5bd Redis 2.8.23. 2015-10-15 15:40:57 +02:00
antirez d85eed2400 Redis.conf example: make clear user must pass its path as argument. 2015-10-15 12:46:19 +02:00
antirez c3a7fe4d32 Regression test for issue #2813. 2015-10-15 11:25:22 +02:00
antirez 34b3a045f8 Move end-comment of handshake states. 
For an error I missed the last handshake state.
Related to issue #2813.
 2015-10-15 10:04:35 +02:00
antirez 38b1867fe2 Make clear that slave handshake states must be ordered. 
Make sure that people from the future will not break this rule.
Related to issue #2813.
 2015-10-15 10:03:33 +02:00
antirez 816dda3971 Minor changes to PR #2813. 
* Function to test for slave handshake renamed slaveIsInHandshakeState.
* Function no longer accepts arguments since it always tests the
  same global state.
* Test for state translated to a range test since defines are guaranteed
  to stay in order in the future.
* Use the new function in the ROLE command implementation as well.
 2015-10-15 09:59:07 +02:00
Salvatore Sanfilippo db98e54982 Merge pull request #2813 from kevinmcgehee/2.8 
Fix master timeout during handshake
 2015-10-15 09:51:28 +02:00
Kevin McGehee 3e0b34cf6d Fix master timeout during handshake 
This change allows a slave to properly time out a dead master during
the extended asynchronous synchronization state machine.  Now, slaves
will record their last interaction with the master and apply the
replication timeout before a response to the PSYNC request is received.
 2015-10-14 12:03:47 -07:00
antirez 560142e6a7 redis-cli pipe mode: don't stay in the write loop forever. 
The code was broken and resulted in redis-cli --pipe to, most of the
times, writing everything received in the standard input to the Redis
connection socket without ever reading back the replies, until all the
content to write was written.

This means that Redis had to accumulate all the output in the output
buffers of the client, consuming a lot of memory.

Fixed thanks to the original report of anomalies in the behavior
provided by Twitter user @fsaintjacques.
 2015-09-30 16:27:26 +02:00
antirez 18bdc2790d Test: fix false positive in HSTRLEN test. 
HINCRBY* tests later used the value "tmp" that was sometimes generated
by the random key generation function. The result was ovewriting what
Tcl expected to be inside Redis with another value, causing the next
HSTRLEN test to fail.
 2015-09-15 09:38:30 +02:00
antirez 8819800310 Test: MOVE expire test improved. 
Related to #2765.
 2015-09-14 12:37:49 +02:00
antirez 5fd2dc3186 MOVE re-add TTL check fixed. 
getExpire() returns -1 when no expire exists.

Related to #2765.
 2015-09-14 12:37:49 +02:00
antirez 37843be92f MOVE now can move TTL metadata as well. 
MOVE was not able to move the TTL: when a key was moved into a different
database number, it became persistent like if PERSIST was used.

In some incredible way (I guess almost nobody uses Redis MOVE) this bug
remained unnoticed inside Redis internals for many years.
Finally Andy Grunwald discovered it and opened an issue.

This commit fixes the bug and adds a regression test.

Close #2765.
 2015-09-14 12:37:49 +02:00
antirez 54b368bb81 Redis 2.8.22. 2015-09-08 10:03:48 +02:00
antirez 35507e7871 Sentinel: command arity check added where missing. 2015-09-08 09:33:48 +02:00
Rogerio Goncalves fa81cc2624 Check args before run ckquorum. Fix issue #2635 2015-09-08 09:28:30 +02:00
MOON_CLJ 06894b8fe4 pfcount support multi keys 2015-09-08 09:22:11 +02:00
antirez 7bac7d37cb Fix merge issues in 490847c. 2015-09-07 17:29:43 +02:00
antirez 9155cdc3d6 Undo slaves state change on failed rdbSaveToSlavesSockets(). 
As Oran Agra suggested, in startBgsaveForReplication() when the BGSAVE
attempt returns an error, we scan the list of slaves in order to remove
them since there is no way to serve them currently.

However we check for the replication state BGSAVE_START, which was
modified by rdbSaveToSlaveSockets() before forking(). So when fork fails
the state of slaves remain BGSAVE_END and no cleanup is performed.

This commit fixes the problem by making rdbSaveToSlavesSockets() able to
undo the state change on fork failure.
 2015-09-07 16:21:33 +02:00
antirez 8c8a7cded9 Sentinel: fix bug in config rewriting during failover 
We have a check to rewrite the config properly when a failover is in
progress, in order to add the current (already failed over) master as
slave, and don't include in the slave list the promoted slave itself.

However there was an issue, the variable with the right address was
computed but never used when the code was modified, and no tests are
available for this feature for two reasons:

1. The Sentinel unit test currently does not test Sentinel ability to
persist its state at all.
2. It is a very hard to trigger state since it lasts for little time in
the context of the testing framework.

However this feature should be covered in the test in some way.

The bug was found by @badboy using the clang static analyzer.

Effects of the bug on safety of Sentinel
===

This bug results in severe issues in the following case:

1. A Sentinel is elected leader.
2. During the failover, it persists a wrong config with a known-slave
entry listing the master address.
3. The Sentinel crashes and restarts, reading invalid configuration from
disk.
4. It sees that the slave now does not obey the logical configuration
(should replicate from the current master), so it sends a SLAVEOF
command to the master (since the slave master is the same) creating a
replication loop (attempt to replicate from itself) which Redis is
currently unable to detect.
5. This means that the master is no longer available because of the bug.

However the lack of availability should be only transient (at least
in my tests, but other states could be possible where the problem
is not recovered automatically) because:

6. Sentinels treat masters reporting to be slaves as failing.
7. A new failover is triggered, and a slave is promoted to master.

Bug lifetime
===

The bug is there forever. Commit 16237d78 actually tried to fix the bug
but in the wrong way (the computed variable was never used! My fault).
So this bug is there basically since the start of Sentinel.

Since the bug is hard to trigger, I remember little reports matching
this condition, but I remember at least a few. Also in automated tests
where instances were stopped and restarted multiple times automatically
I remember hitting this issue, however I was not able to reproduce nor
to determine with the information I had at the time what was causing the
issue.
 2015-09-07 15:53:08 +02:00
antirez c47854615d Sentinel: clarify effect of resetting failover_start_time. 2015-09-07 15:53:08 +02:00
ubuntu 3ff2d65ffa SCAN iter parsing changed from atoi to chartoull 2015-09-07 13:25:38 +02:00
antirez 194b7e2186 Force slaves to resync after unsuccessful PSYNC. 
Using chained replication where C is slave of B which is in turn slave of
A, if B reconnects the replication link with A but discovers it is no
longer possible to PSYNC, slaves of B must be disconnected and PSYNC
not allowed, since the new B dataset may be completely different after
the synchronization with the master.

Note that there are varius semantical differences in the way this is
handled now compared to the past. In the past the semantics was:

1. When a slave lost connection with its master, disconnected the chained
slaves ASAP. Which is not needed since after a successful PSYNC with the
master, the slaves can continue and don't need to resync in turn.

2. However after a failed PSYNC the replication backlog was not reset, so a
slave was able to PSYNC successfully even if the instance did a full
sync with its master, containing now an entirely different data set.

Now instead chained slaves are not disconnected when the slave lose the
connection with its master, but only when it is forced to full SYNC with
its master. This means that if the slave having chained slaves does a
successful PSYNC all its slaves can continue without troubles.

See issue #2694 for more details.
 2015-08-21 15:32:04 +02:00
antirez 730f7c5f5e replicationHandleMasterDisconnection() belongs to replication.c. 2015-08-21 15:32:01 +02:00
antirez 12d2a89410 flushSlavesOutputBuffers(): details clarified via comments. 
Talking with @oranagra we had to reason a little bit to understand if
this function could ever flush the output buffers of the wrong slaves,
having online state but actually not being ready to receive writes
before the first ACK is received from them (this happens with diskless
replication).

Next time we'll just read this comment.
 2015-08-21 15:31:59 +02:00
antirez 7a02677097 Log client details on SLAVEOF command having an effect. 2015-08-21 15:31:11 +02:00
antirez 5630eeb12b startBgsaveForReplication(): handle waiting slaves state change. 
Before this commit, after triggering a BGSAVE it was up to the caller of
startBgsavForReplication() to handle slaves in WAIT_BGSAVE_START in
order to update them accordingly. However when the replication target is
the socket, this is not possible since the process of updating the
slaves and sending the FULLRESYNC reply must be coupled with the process
of starting an RDB save (the reason is, we need to send the FULLSYNC
command and spawn a child that will start to send RDB data to the slaves
ASAP).

This commit moves the responsibility of handling slaves in
WAIT_BGSAVE_START to startBgsavForReplication() so that for both
diskless and disk-based replication we have the same chain of
responsiblity. In order accomodate such change, the syncCommand() also
needs to put the client in the slave list ASAP (just after the initial
checks) and not at the end, so that startBgsavForReplication() can find
the new slave alrady in the list.

Another related change is what happens if the BGSAVE fails because of
fork() or other errors: we now remove the slave from the list of slaves
and send an error, scheduling the slave connection to be terminated.

As a side effect of this change the following errors found by
Oran Agra are fixed (thanks!):

1. rdbSaveToSlavesSockets() on failed fork will get the slaves cleaned
up, otherwise they remain in a wrong state forever since we setup them
for full resync before actually trying to fork.

2. updateSlavesWaitingBgsave() with replication target set as "socket"
was broken since the function changed the slaves state from
WAIT_BGSAVE_START to WAIT_BGSAVE_END via
replicationSetupSlaveForFullResync(), so later rdbSaveToSlavesSockets()
will not find any slave in the right state (WAIT_BGSAVE_START) to feed.
 2015-08-21 11:55:30 +02:00
antirez a7e9b38a48 slaveTryPartialResynchronization and syncWithMaster: better synergy. 
It is simpler if removing the read event handler from the FD is up to
slaveTryPartialResynchronization, after all it is only called in the
context of syncWithMaster.

This commit also makes sure that on error all the event handlers are
removed from the socket before closing it.
 2015-08-07 12:19:21 +02:00
antirez d4e4bd039a syncWithMaster(): non blocking state machine. 2015-08-07 12:18:29 +02:00
antirez c9df63c103 Fixed issues introduced during last merge. 2015-08-06 17:33:04 +02:00
antirez ce3a2d085b startBgsaveForReplication(): log what you really do. 2015-08-06 12:36:13 +02:00
antirez 6974e69f35 Replication: add REPLCONF CAPA EOF support. 
Add the concept of slaves capabilities to Redis, the slave now presents
to the Redis master with a set of capabilities in the form:

    REPLCONF capa SOMECAPA capa OTHERCAPA ...

This has the effect of setting slave->slave_capa with the corresponding
SLAVE_CAPA macros that the master can test later to understand if it
the slave will understand certain formats and protocols of the
replication process. This makes it much simpler to introduce new
replication capabilities in the future in a way that don't break old
slaves or masters.

This patch was designed and implemented together with Oran Agra
(@oranagra).
 2015-08-06 12:36:13 +02:00
antirez be56e4cf33 Fix synchronous readline "\n" handling. 
Our function to read a line with a timeout handles newlines as requests
to refresh the timeout, however the code kept subtracting the buffer
size left every time a newline was received, for a bug in the loop
logic. Fixed by this commit.
 2015-08-05 17:00:04 +02:00
antirez a67d67b561 Fix replication slave pings period. 
For PINGs we use the period configured by the user, but for the newlines
of slaves waiting for an RDB to be created (including slaves waiting for
the FULLRESYNC reply) we need to ping with frequency of 1 second, since
the timeout is fixed and needs to be refreshed.
 2015-08-05 17:00:04 +02:00
antirez 9a5560f4c3 Fix RDB encoding test for new csvdump format. 2015-08-05 14:05:34 +02:00
antirez 6da198cdf1 Remove slave state change handled by replicationSetupSlaveForFullResync(). 2015-08-05 14:00:25 +02:00
antirez 39994c2493 Make sure we re-emit SELECT after each new slave full sync setup. 
In previous commits we moved the FULLRESYNC to the moment we start the
BGSAVE, so that the offset we provide is the right one. However this
also means that we need to re-emit the SELECT statement every time a new
slave starts to accumulate the changes.

To obtian this effect in a more clean way, the function that sends the
FULLRESYNC reply was overloaded with a more important role of also doing
this and chanigng the slave state. So it was renamed to
replicationSetupSlaveForFullResync() to better reflect what it does now.
 2015-08-05 14:00:22 +02:00
antirez a89326f0f6 Test: csvdump now scans all DBs. 2015-08-05 14:00:18 +02:00
antirez b2ff48ef19 Don't send SELECT to slaves in WAIT_BGSAVE_START state. 2015-08-05 14:00:15 +02:00
antirez 7967f1bca6 PSYNC test: also test the vanilla SYNC. 2015-08-05 14:00:11 +02:00
antirez e684e7266c syncCommand() comments improved. 2015-08-05 14:00:08 +02:00
antirez 4b010572cd PSYNC initial offset fix. 
This commit attempts to fix a bug involving PSYNC and diskless
replication (currently experimental) found by Yuval Inbar from Redis Labs
and that was later found to have even more far reaching effects (the bug also
exists when diskstore is off).

The gist of the bug is that, a Redis master replies with +FULLRESYNC to
a PSYNC attempt that fails and requires a full resynchronization.
However, the baseline offset sent along with FULLRESYNC was always the
current master replication offset. This is not ok, because there are
many reasosn that may delay the RDB file creation. And... guess what,
the master offset we communicate must be the one of the time the RDB
was created. So for example:

1) When the BGSAVE for replication is delayed since there is one
   already but is not good for replication.
2) When the BGSAVE is not needed as we attach one currently ongoing.
3) When because of diskless replication the BGSAVE is delayed.

In all the above cases the PSYNC reply is wrong and the slave may
reconnect later claiming to need a wrong offset: this may cause
data curruption later.
 2015-08-05 14:00:04 +02:00
antirez dc4d24440f Test PSYNC with diskless replication. 
Thanks to Oran Agra from Redis Labs for providing this patch.
 2015-08-05 14:00:00 +02:00
Salvatore Sanfilippo d3688e8b68 Merge pull request #2666 from Abioy/2.8 
bugfix: errno might change before logging
 2015-07-17 10:44:18 +02:00
Salvatore Sanfilippo 88a38fe3d3 Merge pull request #2678 from Kiemes/fix-config-resetstat 
Fix: aof_delayed_fsync is not reset, fixes #2677
 2015-07-17 10:39:02 +02:00
Tom Kiemes cb98ae12f6 Fix: aof_delayed_fsync is not reset 
aof_delayed_fsync was not set to 0 when calling CONFIG RESETSTAT
 2015-07-16 11:57:07 +02:00
Yongyue Sun 9c7f98521e bugfix: errno might change before logging 
Signed-off-by: Yongyue Sun <abioy.sun@gmail.com>
 2015-07-13 14:59:05 +08:00
antirez fbb9d619f7 Fix 2.8.21 release notes to give full credits. 2015-06-04 12:00:21 +02:00
antirez 7f8b865a65 Redis 2.8.21 2015-06-04 11:32:24 +02:00
Ben Murphy 700b863f13 hide access to debug table 2015-06-03 13:36:02 +02:00
Ben Murphy 5a1b22ad7f disable loading lua bytecode 2015-06-03 13:36:02 +02:00
antirez 1eeb9bd714 Scripting: Lua cmsgpack lib updated to include str8 support 2015-06-03 08:45:56 +02:00
therealbill 22ee2f9cd5 adding a sentinel command: "flushconfig" 
This new command triggers a config flush to save the in-memory config to
disk. This is useful for cases of a configuration management system or a
package manager wiping out your sentinel config while the process is
still running - and has not yet been restarted. It can also be useful
for scripting a backup and migrate or clone of a running sentinel.
 2015-05-25 12:06:57 +02:00
antirez cdedad2322 Sentinel: CKQUORUM tests 2015-05-19 12:27:26 +02:00
antirez 76412ec9c0 Sentinel: SENTINEL CKQUORUM command 
A way for monitoring systems to check that Sentinel is technically able
to reach the quorum and failover, using the currently visible Sentinels.
 2015-05-19 12:27:26 +02:00
antirez f42fcff6d2 Rewrite smoveCommand test with ternary operator 2015-05-15 17:39:39 +02:00
Glenn Nethercutt 19382c8be6 uphold the smove contract to return 0 when the element is not a member of the source set, even if source=dest 2015-05-15 17:39:39 +02:00
Jungtaek Lim bea224308b protocol error log should be seen debug/verbose level 2015-05-15 17:06:45 +02:00
antirez 22979836e5 Fix 2.8.20 changelog typo 2015-05-05 11:50:30 +02:00
antirez 11934a9f32 Redis 2.8.20 2015-05-05 11:14:33 +02:00
FuGangqiang 8db3969971 sdsfree x and y 2015-05-04 13:03:06 +02:00
FuGangqiang 98756d4c0d fix doc example 2015-05-04 13:03:06 +02:00
FuGangqiang 7316fda33a fix typo 2015-05-04 13:03:06 +02:00
Itamar Haber a5bada19be update copyright year 2015-05-04 12:56:52 +02:00
therealbill f0ab4fd6fd Making sentinel flush config on +slave 
Originally, only the +slave event which occurs when a slave is
reconfigured during sentinelResetMasterAndChangeAddress triggers a flush
of the config to disk.  However, newly discovered slaves don't
apparently trigger this flush but do trigger the +slave event issuance.

So if you start up a sentinel, add a master, then add a slave to the
master (as a way to reproduce it) you'll see the +slave event issued,
but the sentinel config won't be updated with the known-slave entry.

This change makes sentinel do the flush of the config if a new slave is
deteted in sentinelRefreshInstanceInfo.
 2015-05-04 12:55:30 +02:00
antirez e4c54498db Sentinel: remove useless sentinelFlushConfig() call 
To rewrite the config in the loop that adds slaves back after a master
reset, in order to handle switching to another master, is useless: it
just adds latency since there is an fsync call in the inner loop,
without providing any additional guarantee, but the contrary, since if
after the first loop iteration the server crashes we end with just a
single slave entry losing all the other informations.

It is wiser to rewrite the config at the end when the full new
state is configured.
 2015-05-04 12:55:30 +02:00
Yossi Gottlieb 81a8ebd7f0 Fix Redis server crash when Lua command exceeds client output buffer 
limit.
 2015-05-04 12:20:28 +02:00
clark.kang c4d4c1ed12 fix sentinel memory leak 2015-05-04 12:18:13 +02:00
antirez 081a0c943f Fix Sentinel memory leak (hiredis bug) 
This fixes issue #2535, that was actually an hiredis library bug (I
submitted an issue and fix to the redis/hiredis repo as well).

When an asynchronous hiredis connection subscribes to a Pub/Sub channel
and gets an error, and in other related conditions, the function
redisProcessCallbacks() enters a code path where the link is
disconnected, however the function returns before freeing the allocated
reply object. This causes a memory leak. The memory leak was trivial to
trigger in Redis Sentinel, which uses hiredis, every time we tried to
subscribe to an instance that required a password, in case the Sentinel
was configured either with the wrong password or without password at
all. In this case, the -AUTH error caused the leaking code path to be
executed.

It was verified with Valgrind that after this change the leak no longer
happens in Sentinel with a misconfigured authentication password.
 2015-04-28 22:15:12 +02:00
antirez d593ae8ab6 sha1.c: use standard uint32_t. 2015-04-27 12:08:03 +02:00
Oran Agra c72253ec54 fixes to diskless replication. 
master was closing the connection if the RDB transfer took long time.
and also sent PINGs to the slave before it got the initial ACK, in which case the slave wouldn't be able to find the EOF marker.
 2015-04-01 15:21:08 +02:00
antirez 83eb5e7f54 Test: regression for issue #2473. 2015-03-27 12:11:46 +01:00
antirez 96aa610666 Replication: disconnect blocked clients when switching to slave role. 
Bug as old as Redis and blocking operations. It's hard to trigger since
only happens on instance role switch, but the results are quite bad
since an inconsistency between master and slave is created.

How to trigger the bug is a good description of the bug itself.

1. Client does "BLPOP mylist 0" in master.
2. Master is turned into slave, that replicates from New-Master.
3. Client does "LPUSH mylist foo" in New-Master.
4. New-Master propagates write to slave.
5. Slave receives the LPUSH, the blocked client get served.

Now Master "mylist" key has "foo", Slave "mylist" key is empty.

Highlights:

* At step "2" above, the client remains attached, basically escaping any
  check performed during command dispatch: read only slave, in that case.
* At step "5" the slave (that was the master), serves the blocked client
  consuming a list element, which is not consumed on the master side.

This scenario is technically likely to happen during failovers, however
since Redis Sentinel already disconnects clients using the CLIENT
command when changing the role of the instance, the bug is avoided in
Sentinel deployments.

Closes #2473.
 2015-03-24 16:14:49 +01:00
antirez 1f5e7b8863 Config: activerehashing option support in CONFIG SET. 2015-03-08 15:34:00 +01:00
antirez f8f40afd5d Fix iterator for issue #2438. 
Itereator misuse due to analyzeLatencyForEvent() accessing the
dictionary during the iteration, without the iterator being
reclared as safe.
 2015-03-04 11:49:53 -08:00
Charles Hooper 1f4074e957 override histfile from env - fixes #831 and copies #833 2015-02-10 14:55:17 +01:00
antirez 9e83d2d26f redis-cli --lru-test implemented (cache workload simulator). 2015-02-10 14:55:17 +01:00
antirez ccbec34fad redis-cli: interactive reconnection for latency modes. 
--stat mode already used to reconnect automatically if the server is no
longer available. This is useful since this is an interactive mode used
for debugging, however the same applies to --latency and --latency-dist
modes, so now both use the reconnecting command execution as well.

The reconnection code was modified to use basic VT100 escape sequences
in order to play better with different kinds of output on the screen
when the reconnection happens, and to hide the reconnection attempt
output when finally the reconnection happens.
 2015-02-10 14:55:17 +01:00
antirez d29eb68f20 redis-cli --latecy-dist reverted to gray scale. 
So far not able to find a color palette within the 256 colors which is
not confusing. However I believe it is a possible task, so will try
better later.
 2015-02-10 14:55:17 +01:00
antirez 657f719760 redis-cli --latency-dist now uses a color palette. 
Still not happy with the result but low grays are hard to see in certain
monitors with a non perfect gamma.
 2015-02-10 14:55:17 +01:00
antirez ff13fa787d redis-cli latency dist: add new top HL. 2015-02-10 14:55:17 +01:00
antirez f6f3cb9458 Add missing latency-dest legend symbol. 2015-02-10 14:55:17 +01:00
antirez 9d9be0e09d Initial implementation of redis-cli --latency-dist. 2015-02-10 14:55:17 +01:00
antirez d4047f7261 Faster memory efficiency test. 
This test on Linux was extremely slow, since in Tcl we can't enable
easily tcp-nodelay, so the busy loop used to take *a lot* with bigger
writes. Fixed using pipelining.
 2015-02-10 14:48:22 +01:00
mattcollier 61290d819d Update redis-cli.c 
Code was adding '\n'  (line 521) to the end of NIL values exlusively making csv output inconsistent.  Removed '\n'
 2015-02-02 22:50:53 +01:00
antirez 18dc769104 luaRedisGenericCommand(): log error at WARNING level when re-entered. 
Rationale is that when re-entering, it is likely due to Lua debugging
hooks. Returning an error will be ignored in most cases, going totally
unnoticed. With the log at least we leave a trace.

Related to issue #2302.
 2015-01-22 10:36:23 +01:00
antirez 1dea8b5a29 luaRedisGenericCommand() recursion: just return an error. 
Instead of calling redisPanic() to abort the server.

Related to issue #2302.
 2015-01-22 10:36:19 +01:00
antirez a0ba7b3f00 Panic on recursive calls to luaRedisGenericCommand(). 
Related to issue #2302.
 2015-01-22 10:36:15 +01:00
antirez e7aa5fb5ef AOF rewrite: set iterator var to NULL when freed. 
The cleanup code expects that if 'di' is not NULL, it is a valid
iterator that should be freed.

The result of this bug was a crash of the AOF rewriting process if an
error occurred after the DBs data are written and the iterator is no
longer valid.
 2015-01-21 16:42:56 +01:00
antirez 8a41dae49f Add "-lrt" in Makefile for Solaris. 
This fix is from @NanXiao, however I was not able to retain authorship
because the Pull Request original repository was removed.
 2015-01-09 11:54:52 +01:00
antirez 50cdedd127 Check for __sun macro in solarisfixes.h, not in includers. 2015-01-09 11:22:35 +01:00
antirez 77608ae5ac README section about make distclean reworded / extended. 2015-01-08 16:35:53 +01:00
Jungtaek Lim 5165fb98af Explain make distclean which seems not well known 2015-01-08 16:35:49 +01:00
antirez 22a0fe8de6 INFO loading stats: three fixes. 
1. Server unxtime may remain not updated while loading AOF, so ETA is
not updated correctly.

2. Number of processed byte was not initialized.

3. Possible division by zero condition (likely cause of issue #1932).
 2014-12-23 14:52:57 +01:00
Alon Diamant 585884da45 Fixed memory leaks in rdbSaveToSlavesSockets() 2014-12-22 11:01:21 +01:00
antirez ca3dd5e5ed Fix adjustOpenFilesLimit() logging to match real state. 
Fixes issue #2225.
 2014-12-19 10:05:09 +01:00
Rhommel Lamas 6f0a7353a6 Add redis-sentinel symlink to a relative path 2014-12-17 12:27:44 +01:00
antirez 902b877dc0 Redis 2.8.19. 2014-12-16 09:18:20 +01:00
antirez 02d465c8fe Don't log admin commands in MONITOR. 
Otherwise there are security risks, especially when providing Redis as a
service, the user may "sniff" for admin commands renamed to an
unguessable string via rename-command in redis.conf.
 2014-12-13 09:00:03 +01:00
antirez 4d8f4262a4 List of commands flagged as admin commands modified. 
The old list did not made much sense... and the flag is currently not
used at all, so no side effects.
 2014-12-13 08:59:43 +01:00
Salvatore Sanfilippo d4fc65309d Merge pull request #2092 from rhoml/rhoml-install-sentinel 
Adds a symlink for redis-sentinel when Make install
 2014-12-12 22:58:56 +01:00
antirez e47e460f1f Lua cmsgpack lib updated to latest version. 
It fixes a bad bug that crashes the server in certain conditions
as shown in issue #2210.
 2014-12-12 22:57:09 +01:00
Rhommel Lamas 5509c141f8 Add symlink to redis-sentinel during make install 2014-12-12 20:00:08 +01:00
antirez 7de1ef773e SORT: Don't sort Set elements if not needed. 
Related to #2094.
 2014-12-11 16:01:21 +01:00
Matt Stancliff e945a546af Fix zero-ordering SORT when called against lists 
People mostly use SORT against lists, but our prior
behavior was pretending lists were an unordered bag
requiring a forced-sort when no sort was requested.

We can just use the native list ordering to ensure
consistency across replicaion and scripting calls.

Closes #2079
Closes #545 (again)
 2014-12-11 16:01:09 +01:00
Ben Dowling d81c38316b Update redis_init_script.tpl 
status command currently reports success when redis has crashed and the pid file still exists. Changing to check the actual process is running.
 2014-12-11 15:21:53 +01:00
zhanghailei dba57ea910 FIXED redis-benchmark's idle mode.With idle mode shouldn't create write event 2014-12-11 15:21:53 +01:00
Sun He 888ea17548 zipmap.c: update comments above 2014-12-09 16:01:13 +01:00
Deepak Verma 86ebc139c2 replaced // comments #2150 2014-12-09 15:26:26 +01:00
azure provisioned user 3d73f088e1 redis-benchmark AUTH command to be discarded after the first send #2150 2014-12-09 15:26:26 +01:00
Sun He 76d53a6770 sds.c: Correct two spelling mistakes in comments 2014-12-09 15:14:14 +01:00
Sun He 4848cf9335 sds.c/sdscatvprintf: set va_end to finish va_list cpy 2014-12-09 15:14:14 +01:00
Sun He d2f584fa53 sds.c: Correct some comments 2014-12-09 15:14:13 +01:00
Serghei Iakovlev 2ed3f09cc3 Update whatisdoing.sh 
Improved getting pid
 2014-12-09 13:01:31 +01:00
antirez 77b997d429 Include stropts only if __sun is defined. 2014-12-09 12:59:43 +01:00
Jan-Erik Rediger d409371193 Fix implicit declaration of ioctl on Solaris 2014-12-09 12:57:59 +01:00
Johan Bergström 23b96c02a5 Silence _BSD_SOURCE warnings in glibc 2.20 and forward 
See https://sourceware.org/glibc/wiki/Release/2.20#Packaging_Changes
 2014-12-09 12:24:30 +01:00
antirez a47a042ea6 Mark whatisdoing.sh as deprecated in top-comment. 2014-12-09 12:18:47 +01:00
Serghei Iakovlev b5737d2d19 getting pid fixes 
```sh
$ ~ pidof redis-server
# nothing

$ ~ ps aux | grep [r]edis
redis      593  0.0  0.0  36900  5564 ?        Ssl  Dec02   1:37 /usr/bin/redis-server 127.0.0.1:6379
klay     15927  0.0  0.0  16772  6068 pts/6    S+   13:58   0:00 redis-cli

$ ~ uname -a
Linux edge 3.17.4-1-ARCH #1 SMP PREEMPT Fri Nov 21 21:14:42 CET 2014 x86_64 GNU/Linux
```
 2014-12-09 12:09:31 +01:00
Sun He a598e0894b sparkline.c: AddSample skip Empty label 2014-12-09 11:28:53 +01:00
Sun He 7d480abab4 sparkline.c: mov label-ini into the AddSample Function 2014-12-09 11:28:53 +01:00
Jan-Erik Rediger 2f3c860979 Only ignore sigpipe in interactive mode 
This allows shell pipes to correctly end redis-cli.

Ref #2066
 2014-12-09 09:01:10 +01:00
antirez 0c211a1953 Simplify lua_cmsgpack macro and fix build on old Linux distros. 
Thanks to @badboy for the help in checking the build after the fix.
 2014-12-05 10:52:56 +01:00
antirez d6ada4d1a8 Redis 2.8.18. 2014-12-04 11:36:52 +01:00
antirez 7c671c09a1 Check that tcp-backlog is matched by /proc/sys/net/core/somaxconn. 2014-12-04 11:07:55 +01:00
antirez d56ef62983 Network bandwidth tracking + refactoring. 
Track bandwidth used by clients and replication (but diskless
replication is not tracked since the actual transfer happens in the
child process).

This includes a refactoring that makes tracking new instantaneous
metrics simpler.
 2014-12-03 12:17:18 +01:00
antirez f60c8c0999 Don't show the ASCII logo if syslog is enabled. 
Closes issue #1935.
 2014-12-03 10:52:09 +01:00
antirez 04231cb7d7 Regression test for issue #2175. 2014-12-03 10:41:06 +01:00
antirez 6e0277769d Handle infinite explicitly in createStringObjectFromLongLong(). 2014-12-03 10:41:06 +01:00
antirez 42695fdfe6 Use exp format and more precision output for ZSCAN. 
Ref: issue #2175
 2014-12-03 10:41:06 +01:00
antirez 164ad2ead1 Over 80 chars comment trimmed in pfcountCommand(). 2014-12-02 17:03:15 +01:00
antirez 0e40124d0f Mark PFCOUNT as read-only, even if not true. 
PFCOUNT is technically speaking a write command, since the cached value
of the HLL is exposed in the data structure (design error, mea culpa), and
can be modified by PFCOUNT.

However if we flag PFCOUNT as "w", read only slaves can't execute the
command, which is a problem since there are environments where slaves
are used to scale PFCOUNT reads.

Nor it is possible to just prevent PFCOUNT to modify the data structure
in slaves, since without the cache we lose too much efficiency.

So while this commit allows slaves to create a temporary inconsistency
(the strings representing the HLLs in the master and slave can be
different in certain moments) it is actually harmless.

In the long run this should be probably fixed by turning the HLL into a
more opaque representation, for example by storing the cached value in
the part of the string which is not exposed (this should be possible
with SDS strings).
 2014-12-02 17:01:55 +01:00
antirez d0cc96f053 Mark diskless replication as experimental in redis.conf. 2014-12-02 14:13:44 +01:00
antirez a81043959a Test: wait for actual startup in start_server. 
start_server now uses return value from Tcl exec to get the server pid,
however this introduces errors that depend from timing: a lot of the
testing code base assumed the server to be actually up and running when
server_start returns.

So the old code that waits to see the pid in the log file was restored.
 2014-11-28 12:57:40 +01:00
antirez f0c00a14ed Test: try to cleanup still running Redis instances on exit. 
It's hard to run the Redis test continuously if it leaks processes on
exceptions / errors.
 2014-11-28 12:57:30 +01:00
antirez 3ac62dff23 Test framework: exit on timeout with report. 
There was no sane way to detect tests that may never end because of
Redis bugs or tests bugs.
 2014-11-28 12:54:17 +01:00
antirez 60c09dcf09 redis-benchmark: default num of requests is now 100000. 
10000 completes in a too short time and may easily provide unreliable
figures because of tiny duration.
 2014-11-28 10:38:58 +01:00
Matthias Petschick e689e93663 fix benchmark memleak in loop mode 2014-11-28 10:38:54 +01:00
antirez 01854f87e8 Document redis-cli --stat in --help output. 2014-11-25 18:25:58 +01:00
Matt Stancliff e305df099a Fix lua-cmsgpack 64 bit integer on 32 bit platform 
This syncs lua-cmsgpack with the mattsta/lua-cmsgpack upstream.

Fixes #2161
 2014-11-25 15:56:57 +01:00
antirez 9854d03fad Avoid valgrind memory leak false positive in processInlineBuffer(). 
zmalloc(0) cauesd to actually trigger a non-zero allocation since with
standard libc malloc we have our own zmalloc header for memory tracking,
but at the same time the returned pointer is at the end of the block and
not in the middle. This triggers a false positive when testing with
valgrind.

When the inline protocol args count is 0, we now avoid reallocating
c->argv, preventing the issue to happen.
 2014-11-25 14:51:04 +01:00
antirez 5ddf56f9b1 Attempt to prevent false positives in replication test. 2014-11-24 11:55:16 +01:00
antirez b19e745780 scripting.tcl tests order fixed to match unstable branch. 2014-11-14 17:28:58 +01:00
antirez ce96069676 lua_cjson.c Lua includes: angled -> quoted. 2014-11-14 17:28:54 +01:00
antirez ead8e31b28 Fix non-linux builds error introduced with THP checks. 2014-11-14 17:28:49 +01:00
Matt Stancliff 5d8c88fdc1 Lua: add cmsgpack scripting tests 
Basically: test to make sure we can load cmsgpack
and do some sanity checks to make sure pack/unpack works
properly.  We also have a bonus test for circular encoding
and decoding because I was curious how it worked.
 2014-11-14 17:28:45 +01:00
Matt Stancliff 381863be5c Lua: upgrade cmsgpack to 0.4.0 
Main reasons for upgrade:
  - Remove a warning when building Redis
  - Add multi pack/unpack
  - Improve memory usage and use Lua allocator properly
  - Fix some edge case encoding/decoding bugs
 2014-11-14 17:28:40 +01:00
Matt Stancliff a7b58f1569 Lua: remove new warning added by cjson header 
clang doesn't like "extern inline" when no definition
is given right away.
 2014-11-14 17:28:35 +01:00
Matt Stancliff 56e1861398 Lua: Use Redis solaris compatability for cjson too 
cjson calls isinf, but some Solaris versions don't have isinf
even with the attempted fix we have in deps/Makefile.

We can harmlessly include the Redis solarisfixes.h header to
give cjson proper isinf.

Note: cjson has a compile-time setting for using their own defined
isinf, but the Redis definition in solarisfixes.h is more complete.

Fixes antirez#1620
 2014-11-14 17:28:32 +01:00
Matt Stancliff e2983dd03f Lua: Upgrade cjson to 2.1.0 (2012-03-01) 
The new cjson has some improvements over our current version including
increased platform compatability, a new resource limit to restrict
decode depth, and better invalid number handling.

One minor change was required to deps/Makefile because this version
of cjson doesn't export itself globally, so we added a quick little
define of -DENABLE_CJSON_GLOBAL.

cjson now has an optional higher performing float parsing interface,
but we are not including it (g_fmt.c, dtoa.c) because it requires
endianness declaration during compile time.

This commit is exactly lua_cjson.c from 2.1.0 with one minor
change of altering the two Lua includes for local search
instead of system-wide importing.
 2014-11-14 17:28:27 +01:00
Matt Stancliff 2828bbd511 Lua: add cjson scripting test 
Two simple decode tests added mainly to check that
the 'cjson' global gets registered and is usable.
 2014-11-14 17:28:23 +01:00
Matt Stancliff 1e501d9fd6 Lua: Add bitop 
A few people have written custom C commands because bit
manipulation isn't exposed through Lua.  Let's give
them Mike Pall's bitop.

This adds bitop 1.0.2 (2012-05-08) from http://bitop.luajit.org/

bitop is imported as "bit" into the global namespace.

New Lua commands: bit.tobit, bit.tohex, bit.bnot, bit.band, bit.bor, bit.bxor,
bit.lshift, bit.rshift, bit.arshift, bit.rol, bit.ror, bit.bswap

Verification of working (the asserts would abort on error, so (nil) is correct):
127.0.0.1:6379> eval "assert(bit.tobit(1) == 1); assert(bit.band(1) == 1); assert(bit.bxor(1,2) == 3); assert(bit.bor(1,2,4,8,16,32,64,128) == 255)" 0
(nil)
127.0.0.1:6379> eval 'assert(0x7fffffff == 2147483647, "broken hex literals"); assert(0xffffffff == -1 or 0xffffffff == 2^32-1, "broken hex literals"); assert(tostring(-1) == "-1", "broken tostring()"); assert(tostring(0xffffffff) == "-1" or tostring(0xffffffff) == "4294967295", "broken tostring()")' 0
(nil)

Tests also integrated into the scripting tests and can be run with:
./runtest --single unit/scripting

Tests are excerpted from `bittest.lua` included in the bitop distribution.
 2014-11-14 17:28:19 +01:00
antirez d9326cb83d THP detection for LATENCY DOCTOR. 2014-11-12 11:17:31 +01:00
antirez 011cfdfb6b Check THP support at startup and warn about it. 2014-11-12 11:17:27 +01:00
antirez c0266d87ae THP detection / reporting functions added. 2014-11-12 11:17:24 +01:00
antirez a5fcf44fcc Diskless SYNC: fix RDB EOF detection. 
RDB EOF detection was relying on the final part of the RDB transfer to
be a magic 40 bytes EOF marker. However as the slave is put online
immediately, and because of sockets timeouts, the replication stream is
actually contiguous with the RDB file.

This means that to detect the EOF correctly we should either:

1) Scan all the stream searching for the mark. Sucks CPU-wise.
2) Start to send the replication stream only after an acknowledge.
3) Implement a proper chunked encoding.

For now solution "2" was picked, so the master does not start to send
ASAP the stream of commands in the case of diskless replication. We wait
for the first REPLCONF ACK command from the slave, that certifies us
that the slave correctly loaded the RDB file and is ready to get more
data.
 2014-11-11 17:27:22 +01:00
antirez ec1ab0f18b Disconnect timedout slave: regression introduced with diskless repl. 2014-11-11 17:25:58 +01:00
Matt Stancliff eb67ca0f00 redis-cli: ignore SIGPIPE network errors 
Closes #2066
 2014-10-30 12:02:32 +01:00
antirez bb9ad32cb0 Fix DEBUG POPULATE warning for lack of casting. 2014-10-29 14:36:55 +01:00
antirez a37f8ec8b4 DEBUG POPULATE two args form implemented. 
The old DEBUG POPULATE form for automatic creation of test keys is:

    DEBUG POPULATE <count>

Now an additional form is available:

    DEBUG POPULATE <count> <prefix>

When prefix is not specified, it defaults to "key", so the keys are
named incrementally from key:0 to key:<count-1>. Otherwise the specified
prefix is used instead of "key".

The command is useful in order to populate different Redis instances
with key names guaranteed to don't collide. There are other debugging
uses, for example it is possible to add additional N keys using a count
of N and a random prefix at every call.
 2014-10-29 14:36:55 +01:00
antirez b2509d7759 redis-cli: add missing newline in error message. 2014-10-29 14:35:48 +01:00
antirez 9b7a190c82 Diskless replication: missing listRewind() added. 
This caused BGSAVE to be triggered a second time without any need when
we switch from socket to disk target via the command

    CONFIG SET repl-diskless-sync no

and there is already a slave waiting for the BGSAVE to start.
Also comments clarified about what is happening.
 2014-10-29 14:34:58 +01:00
antirez 49629a2e34 Log slave ip:port in more log messages. 2014-10-29 14:34:58 +01:00
antirez 081eeb9f6b Use new slave name function for diskless repl reporting. 2014-10-29 14:34:58 +01:00
antirez 72ea77af0b Added a function to get slave name for logs. 2014-10-29 14:34:51 +01:00
antirez 16319c69b1 Diskless replication: log BGSAVE delay only when it is non-zero. 2014-10-29 14:34:25 +01:00
antirez bafd3f646f Document repl-diskless-sync-delay in redis.conf. 2014-10-29 14:34:25 +01:00
antirez f71c4d54d1 Diskless sync delay is now configurable. 2014-10-29 14:34:18 +01:00
antirez 4b520177b3 Remove duplicated log message about starting BGSAVE. 2014-10-29 14:33:50 +01:00
antirez 7e8a471baf Diskless replication documented inside example redis.conf. 2014-10-29 14:33:50 +01:00
antirez d57ed95222 Diskless replication tested with the multiple slaves consistency test. 2014-10-29 14:33:50 +01:00
antirez 7d32dfc69f Diskless replication: child -> parent communication improved. 
Child now reports full info to the parent including IDs of slaves in
failure state and exit code.
 2014-10-29 14:33:50 +01:00
antirez 7728ee809d Translate rio fdset target EWOULDBLOCK error into ETIMEDOUT. 
EWOULDBLOCK with the fdset rio target is returned when we try to write
but the send timeout socket option triggered an error. Better to
translate the error in something the user can actually recognize as a
timeout.
 2014-10-29 14:33:50 +01:00
antirez 7cffd74b38 Diskless replication: set / reset socket send timeout. 
We need to avoid that a child -> slaves transfer can continue forever.
We use the same timeout used as global replication timeout, which is
documented to also affect I/O operations during bulk transfers.
 2014-10-29 14:33:50 +01:00
antirez eff8fb6f39 anet.c: new API anetSendTimeout(). 2014-10-29 14:33:50 +01:00
antirez c38bb5b2ea Diskless replication: less debugging printfs around. 2014-10-29 14:33:50 +01:00
antirez 348c4f6ce3 rio.c fdset write() method fixed: wrong type for return value. 2014-10-29 14:33:50 +01:00
antirez a7d0137075 rio fdset target: handle short writes. 
While the socket is set in blocking mode, we still can get short writes
writing to a socket.
 2014-10-29 14:33:50 +01:00
antirez 23bf435f54 anet.c: API to set sockets back to blocking mode. 2014-10-29 14:33:50 +01:00
antirez 9a521e6c74 Diskless replication: rio fdset target new supports buffering. 
To perform a socket write() for each RDB rio API write call was
extremely unefficient, so now rio has minimal buffering capabilities.
Writes are accumulated into a buffer and only when a given limit is
reacehd are actually wrote to the N slaves FDs.

Trivia: rio lacked support for buffering since our targets were:

1) Memory buffers.
2) C standard I/O.

Both were buffered already.
 2014-10-29 14:33:50 +01:00
antirez 9588973fda Diskless replication: Various fixes to backgroundSaveDoneHandlerSocket() 2014-10-29 14:33:50 +01:00
antirez 8a618971bd Diskless replication: don't send "\n" pings to slaves. 
This is useful for normal replication in order to refresh the slave
when we are persisting on disk, but for diskless replication the
child is already receiving data while in WAIT_BGSAVE_END state.
 2014-10-29 14:33:50 +01:00
antirez b794b94797 Diskless replication: remove 40 bytes EOF mark from end of RDB file. 2014-10-29 14:33:50 +01:00
antirez 3f1f29fdf3 Diskless replication: swap inverted branches to compute read len. 2014-10-29 14:33:50 +01:00
antirez dbc6e9a1ef Diskless replication: don't enter the read-payload branch forever. 2014-10-29 14:33:49 +01:00
antirez adf55b2337 syncReadLine(): actually enforce buffer size limits. 2014-10-29 14:33:49 +01:00
antirez 89f82a25e1 Diskless replication: EOF:<mark> streaming support slave side. 2014-10-29 14:33:49 +01:00
antirez 583a762cd9 Diskless replication: redis.conf and CONFIG SET/GET support. 2014-10-29 14:33:49 +01:00
antirez 31b0e13268 Diskless replication: trigger a BGSAVE after a config change. 
If we turn from diskless to disk-based replication via CONFIG SET, we
need a way to start a BGSAVE if there are slaves alerady waiting for a
BGSAVE to start. Normally with disk-based replication we do it as soon
as the previous child exits, but when there is a configuration change
via CONFIG SET, we may have slaves in WAIT_BGSAVE_START state without
an RDB background process currently active.
 2014-10-29 14:33:49 +01:00
antirez d403e4ae63 Diskless replication flag renamed repl_diskless -> repl_diskless_sync. 2014-10-29 14:33:49 +01:00
antirez 2f0b58e4ed Diskless replication: trigger diskless RDB transfer if needed. 2014-10-29 14:33:49 +01:00
antirez 82bfae5b70 Diskless replication: handle putting the slave online. 2014-10-29 14:33:49 +01:00
antirez 132c108f17 Diskless replication: read report from child. 2014-10-29 14:33:49 +01:00
antirez e82ec2c5d8 Diskless replication: child writes report to parent. 2014-10-29 14:33:49 +01:00
antirez 2ed67e4a9e rio.c fdset target: tolerate (and report) a subset of FDs in error. 
Fdset target is used when we want to write an RDB file directly to
slave's sockets. In this setup as long as there is a single slave that
is still receiving our payload, we want to continue sennding instead of
aborting. However rio calls should abort of no FD is ok.

Also we want the errors reported so that we can signal the parent who is
ok and who is broken, so there is a new set integers with the state of
each fd. Zero is ok, non-zero is the errno of the failure, if avaialble,
or a generic EIO.
 2014-10-29 14:33:49 +01:00
antirez 51864fcdd7 Diskless replication: parent-child pipe and a few TODOs. 2014-10-29 14:33:49 +01:00
antirez 1b4cadb664 Diskless replication: RDB -> slaves transfer draft implementation. 2014-10-29 14:33:49 +01:00
antirez 38dee9dc51 rio.c: draft implementation of fdset target implemented. 2014-10-29 14:33:49 +01:00
antirez 5d295fea12 rio.c refactoring before adding a new target. 2014-10-29 14:33:49 +01:00
antirez aef4c60c78 Add some comments in syncCommand() to clarify RDB target. 2014-10-29 14:33:49 +01:00
antirez 81abd1e178 Define different types of RDB childs. 
We need to remember what is the saving strategy of the current RDB child
process, since the configuration may be modified at runtime via CONFIG
SET and still we'll need to understand, when the child exists, what to
do and for what goal the process was initiated: to create an RDB file
on disk or to write stuff directly to slave's sockets.
 2014-10-29 14:33:49 +01:00
antirez 1e633f21c5 RDB file creation refactored to target non-disk target. 2014-10-29 14:33:49 +01:00
antirez ff8a3baaf7 Replication: better way to send a preamble before RDB payload. 
During the replication full resynchronization process, the RDB file is
transfered from the master to the slave. However there is a short
preamble to send, that is currently just the bulk payload length of the
file in the usual Redis form $..length..<CR><LF>.

This preamble used to be sent with a direct write call, assuming that
there was alway room in the socket output buffer to hold the few bytes
needed, however this does not scale in case we'll need to send more
stuff, and is not very robust code in general.

This commit introduces a more general mechanism to send a preamble up to
2GB in size (the max length of an sds string) in a non blocking way.
 2014-10-29 14:33:30 +01:00
antirez b6203138d3 Linenoise README updated to match source code. 2014-10-06 10:02:08 +02:00
Miguel Parramon 85b15acb44 redis.conf small grammar fix 
😄

Closes #2034
 2014-10-06 10:02:08 +02:00
Ezequiel Lovelle 09fc0fbbe4 Fix typo 
Closes #2029
 2014-10-06 10:02:08 +02:00
Gregory Petrosyan 03b36eef72 Fix typos in comments 
Closes #2002
 2014-10-06 10:01:36 +02:00
Matt Stancliff e1ccd7243f Add missing 'by' 2014-10-06 10:01:36 +02:00
Juarez Bochi 364aeed96d Fix typo in scripting.c comment 
Closes #1960
 2014-10-06 10:01:36 +02:00
T.J. Schuck a7a2f69bdb Fix typo radis -> redis 
Closes #1938
 2014-10-06 10:01:36 +02:00
Matt Stancliff 1adc4a7888 Cleanup wording of dictScan() comment 
Some language in the comment was difficult
to understand, so this commit: clarifies wording, removes
unnecessary words, and relocates some dependent clauses
closer to what they actually describe.

I also tried to break up longer chains of thought
(if X, then Y, and Q, and also F, so obviously M)
into more manageable chunks for ease of understanding.
 2014-10-06 10:01:36 +02:00
Matt Stancliff c3375fbdff Remove trailing spaces from tests 2014-10-06 10:01:27 +02:00
Matt Stancliff c4d32c03d0 Fix spelling in some test cases 2014-10-06 10:01:05 +02:00
Jan-Erik Rediger 919a4c3126 Fix sentinel.conf typo 
Closes #1739
 2014-10-06 10:00:48 +02:00
Xiaojie Zhang 8a1d16fe5c Fix comment inconsistencies in ziplist.c 
Closes #1523
 2014-10-06 10:00:48 +02:00
Matt Stancliff e0ffc863c6 Clean up text throughout project 
- Remove trailing newlines from redis.conf
  - Fix comment misspelling
  - Clarifies zipEncodeLength usage and a C API mention (#1243, #1242)
  - Fix cluster typos (inspired by @papanikge #1507)
  - Fix rewite -> rewrite in a few places (inspired by #682)

Closes #1243, #1242, #1507
 2014-10-06 10:00:43 +02:00
Matt Stancliff 22e6f434f8 Spell software correctly 2014-10-06 09:59:52 +02:00
Aaron Rutkovsky 3101760937 Fix typos 
Closes #1513
 2014-10-06 09:59:02 +02:00
Ben d3005539c5 Grammar and typo fixes in redis.conf 
Closes #1441
 2014-10-06 09:58:56 +02:00
xuxiang 3c99e429cf Fix typo in redis.h 
Closes #1386
 2014-10-06 09:58:18 +02:00
Ted Nyman 55597bbd67 Fix on-place -> in-place 
Closes #1373
 2014-10-06 09:58:18 +02:00
Michael Parker a5044c254d Fix hash table size in comment for dictScan 
Closes #1351
 2014-10-06 09:58:18 +02:00
Jan-Erik Rediger 33daf8e9a9 Fix many small typos 
Closes #1871
 2014-10-06 09:58:11 +02:00
zionwu b72422e100 Fix incorrect comments 
error != success; and 0 != number of bytes written

Closes #1806
 2014-10-06 09:57:01 +02:00
Jan-Erik Rediger 6cdd38f947 Fix typo in redis.conf 
Closes #1713
 2014-10-06 09:57:01 +02:00
Juarez Bochi 6b64fa49f8 Fix typo 
Closes #1682
 2014-10-06 09:57:01 +02:00
Agis Anastasopoulos fc9bce45c5 Fix typo 
Closes #1544
 2014-10-06 09:57:01 +02:00
Jan-Erik Rediger 575e293ddd Fix typo: ad -> and 
Closes #1537
 2014-10-06 09:57:01 +02:00
Matt Stancliff d05fcaffa0 Update linenoise accessory files 
The core linenoise code was being backported, but not
the README or example.  It's less confusing for users
if everything matches across directories.

Fix inspired by @thrig

Closes #1872
 2014-10-06 09:57:01 +02:00
Manuel Meurer a99d187bc4 Improve disable save comment in redis.conf 
Closes #1897
 2014-10-06 09:57:01 +02:00
Lynn 904bd0400c Remove surplus double quotes 
Closes #1877
 2014-10-06 09:57:01 +02:00
antirez 04b6945ebc Linenoise lib updated. 2014-09-22 14:56:01 +02:00
antirez ade6c5e782 Redis 2.8.17 2014-09-19 17:49:16 +02:00
Matt Stancliff cf143b571e Fix hiredis getaddrinfo leak 
Fixed in Redis by 1a5e5b6, but since that part of code
is largely copy/paste from Redis, the fix needs to be
ported over too.

Closes #2012
 2014-09-19 14:21:36 +02:00
antirez a69534967b Redis 2.8.16. 2014-09-16 15:15:56 +02:00
antirez 5e38bc389c Better truncated AOF loading tests. 
Now there are tests to write more data after loading a truncated AOF,
testing that the loaded data is correct, appending more, and testing
again.
 2014-09-16 11:05:50 +02:00
antirez 01f7db608d Seek at the end of AOF after truncate call. 
It is not clear if files open in append only mode will automatically fix
their offset after a truncate(2) operation. This commit makes sure that
we reposition the AOF file descriptor offset at the end of the file
after a truncated AOF is loaded and trimmed to the last valid command.
 2014-09-16 11:05:50 +02:00
antirez 9faeff0215 On AOF end of file, truncate the AOF to last valid command. 
Recently we introduced the ability to load truncated AOFs, but
unfortuantely the support was broken since the server, after loading the
truncated AOF, continues appending to the file that is corrupted at the
end. The problem is fixed only in the next AOF rewrite.

This commit fixes the issue by truncating the AOF to the last valid
opcode, and aborting if it is not possible to truncate the file
correctly.
 2014-09-16 10:41:15 +02:00
antirez fbd9dc6039 Don't propagate SAVE. 
This is a general fix (check that dirty delta is positive) but actually
should have as the only effect fixing the SAVE propagation to
AOF and slaves.
 2014-09-16 10:41:15 +02:00
antirez 551ffc486c version.h updated to 2.8.15. 2014-09-12 16:16:58 +02:00
antirez 78dca8977f Redis 2.8.15. 2014-09-12 16:15:48 +02:00
antirez a6695e341b Sentinel sentinelGetLeader() top comment improved. 2014-09-11 19:28:10 +02:00
antirez 9fe4429a09 Sentinel: fix computation of total number of votes. 
The code to check the number of voters was never updated to follow the new
Sentinel specification, so the number of voters was computed using only
the set of Sentinels that provided a vote.

This means that there is a changing majority on partitions, even if
usually the issue is not triggered because of the configured quorum
check (what was broken was the other implicit check that requires anyway
half of the known sentinels to agree in order to start a failover).
 2014-09-11 19:28:09 +02:00
antirez 6fbcb19275 luaRedisGenericCommand() cached argv handling simplified. 
As discussed in issue #1945.
 2014-09-10 15:47:02 +02:00
xiaost 8375a6d0aa Limit the *SCAN command dictScan iterations 
*SCAN will cause redis server to hang for seconds
after millions of keys was deleted by SCAN/DEL pairs
 2014-09-10 09:58:20 +02:00
antirez 55024e9a9a Sentinel: don't set announce-ip if is empty. 2014-09-10 09:26:29 +02:00
antirez 3cc31da349 Sentinel: clarify announce-ip/port options in sentinel.conf. 2014-09-10 09:26:29 +02:00
antirez 8c3507aa8f Sentinel: announce ip/port changes + rewrite. 
The original implementation was modified in order to allow to
selectively announce a different IP or port, and to rewrite the two
options in the config file after a rewrite.
 2014-09-10 09:26:29 +02:00
Dara Kong 9f3abf436b sentinel: Decouple bind address from address sent to other sentinels 
There are instances such as EC2 where the bind address is private
(behind a NAT) and cannot be accessible from WAN.

https://groups.google.com/d/msg/redis-db/PVVvjO4nMd0/P3oWC036v3cJ
 2014-09-10 09:26:29 +02:00
antirez ff47d5ebf6 Tests for aof-load-truncated = yes. 2014-09-08 10:58:00 +02:00
antirez b3fdddfd1f AOF tests fixed turning aof-load-truncated to no. 
When aof-load-truncated option was introduced, with a default of "yes",
the past behavior of the server to abort with trunncated AOF changed, so
we need to explicitly configure the tests to abort with truncated AOF
by setting the option to no.
 2014-09-08 10:58:00 +02:00
antirez 670084b724 Make aof-load-truncated option actually configurable. 2014-09-08 10:58:00 +02:00
antirez 03a5c0ecb5 Correct cleanup when aborting AOF loading. 
Because of the new ability to start with a truncated AOF, we need
to correctly release all the memory on EOF error. Otherwise there is a
small leak, that is not really a problem, but causes a false positive in
the tests that detect memory leaks.
 2014-09-08 10:58:00 +02:00
antirez 683cf5b872 AOF ability to load truncated files. 2014-09-08 10:57:57 +02:00
antirez 94f7979414 Test AOF format error detection. 2014-09-08 10:57:11 +02:00
antirez 537f3053fa AOF loading: split handling of format errors from unexpected EOF. 2014-09-08 10:57:10 +02:00
Matt Stancliff 667e2991dc Return empty string if GETRANGE of empty string 
Previously, GETRANGE of a key containing nothing ("")
would allocate a large (size_t)-1 return value causing
crashes on 32bit builds when it tried to allocate the
4 GB return string.
 2014-09-04 11:47:21 +02:00
Matt Stancliff 4284b07220 Increase size of range request in getrange 
32 bit builds don't have a big enough long to capture
the same range as a 64 bit build.  If we use "long long"
we get proper size limits everywhere.

Also updates size of unsigned comparison to fit new size of `end`.

Fixes #1981
 2014-09-04 11:46:57 +02:00
antirez db6e874a25 Redis 2.8.14. 2014-09-01 17:00:38 +02:00
Matt Stancliff dd36f18c76 Sentinel: Abort Hello quicker if not connected 
We can save a little work by aborting when we enter the function
if we're disconnected.
 2014-09-01 16:35:04 +02:00
Paddy Byers d1f89bdc2b Add regression test for issue #1939 2014-09-01 10:41:15 +02:00
Paddy Byers 59ad68ca4f Store the length of the static argv when first allocated. 2014-09-01 10:40:04 +02:00
antirez 8fb8a474fd Fix dictRehash assert casting type. 
Also related to #1929.
 2014-08-27 10:31:52 +02:00
antirez 85b3db4065 Cast to right type in dictNext(). 
This closes issue #1929, the other part was fixed in the context of issue
 2014-08-27 10:31:48 +02:00
antirez 3a706ba792 Use long for rehash and iterator index in dict.h. 
This allows to support datasets with more than 2 billion of keys
(possible in very large memory instances, this bug was actually
reported).

Closes issue #1814.
 2014-08-27 10:31:44 +02:00
antirez 97fdcd8746 Missing va_end also added in hiredis copy of sds.c. 2014-08-27 10:31:41 +02:00
Mariano Pérez Rodríguez 45fac9481a Missing assert removal from sdsIncrLen() 
Companion for 8eeb1802ec, but dealing with hiredis.
 2014-08-27 10:31:35 +02:00
antirez c3eac3ddff No longer useful assert removed from sdsIncrLen(). 2014-08-27 10:31:32 +02:00
Matt Stancliff fdcadd16c6 Fix redis-benchmark abort output 2014-08-27 10:31:29 +02:00
Matt Stancliff ac48aada30 Deny CLIENT command in scripts 
We don't want scripts doing CLIENT SETNAME
or CLIENT KILL or CLIENT LIST or CLIENT PAUSE.

Originally reported by Chris Wj then proper
action inspired by Itamar Haber.

Reference: https://groups.google.com/forum/#!topic/redis-db/09B2EYwyVgk
 2014-08-27 10:31:20 +02:00
Matt Stancliff abe29c83e6 redis-benchmark: Remove unused buffer 
This just deletes old code that didn't get removed when
logic changed.  We were setting offsets that never
got read anywhere.

Since clients are now just cloned, we don't need to track
per-client buffer offsets anywhere because they are all
the same from the original client.
 2014-08-27 10:30:49 +02:00
CharSyam 8fcce90b25 redis-benchmark: add auth option 
Closes #1097
 2014-08-27 10:30:49 +02:00
Matt Stancliff 4ac8472cac Rename two 'buf' vars to 'ip' for better clarity 
Clearly ip[32] is wrong, but it's less clear that buf[32] was wrong
without further reading.
 2014-08-27 10:30:42 +02:00
Eiichi Sato f7b5e2d11e Sentinel: fix bufsize to support IPv6 address 
Closes #1914
 2014-08-27 10:30:37 +02:00
Jan-Erik Rediger 32a94beca4 Use correct github url to find commands.json 
Once this is merged:
  - merge the latest changes to commands.json in antirez/redis-doc
  - re-run: utils/generate-command-help.rb > src/help.h

Then we'll have nice and easy tab-completed help in redis-cli again.

Closes #1909
 2014-08-27 10:30:34 +02:00
Mariano Pérez Rodríguez f862e5ddfa Stop tests from leaving a black background 
Uses ANSI "default background" color code after closing tests
so any non-black terminals don't remain polluted.

Fixes #1649
Closes #1912
 2014-08-27 10:30:29 +02:00
Matt Stancliff 32168a5e8a Sentinel: Reject config from STDIN 
Sentinel needs to die with a more accurate error message
when attempted to open a config from STDIN.

See: https://groups.google.com/forum/#!topic/redis-db/sYx7VNMWaNM
 2014-08-27 10:30:26 +02:00
Matt Stancliff 4bf7c5d0f8 Remove unused global variable 
It has an important name, but nothing uses it...
 2014-08-27 10:30:22 +02:00
antirez 33b37b1400 Fix invalid expire error for SET family commands. 2014-08-27 10:30:19 +02:00
Cong Ding ae2cd2e16f Remove unused function 
Closes #878
 2014-08-27 10:30:11 +02:00
David Palm 8c0743d80e Extend range of bytesToHuman to TB and PB 
Also adds a fallthrough case for when given
large values (like overflow numbers of 2^64 by mistake).

Closes #858
 2014-08-27 10:30:08 +02:00
yoav ef12b4c3eb Add error check for writing RDB checksum 
Closes #857
 2014-08-27 10:30:04 +02:00
Dov Murik 1580a378b8 redis-cli: fix prompt after shutdown command 
Fix redis-cli prompt to state "not connected" after a SHUTDOWN command
is sent.
 2014-08-27 10:30:01 +02:00
xiaoyu 467848460a Clarify argument to dict macro 
d is more clear because the type of argument is dict not dictht

Closes #513
 2014-08-27 10:29:57 +02:00
antirez 5e707c66dd Fix undefined behavior in ll2string(). 
The bug was found by @CAFxX, thanks!
See issue #1940.
 2014-08-27 10:29:53 +02:00
antirez 65d47452f8 Remove warnings and improve integer sign correctness. 2014-08-27 10:29:26 +02:00
antirez 09757420a4 Add -W to compilation flags. 2014-08-27 10:25:57 +02:00
antirez 134345e332 Use unsigned integers in SDS header. 
This raises the max string to 4GB without any downside.
 2014-08-27 10:25:54 +02:00
Pieter Noordhuis f89a7bbe11 va_copy must be matched by va_end 
Hat tip to @rfuchs. See: https://github.com/redis/hiredis/pull/178.

Fixes #1187
 2014-08-27 10:25:44 +02:00
charsyam dbcf381099 Remove duplicate prototypes in redis.h 
Also moves acceptHandler() to be near the other accept...() functions.

Closes #1105
 2014-08-27 10:25:40 +02:00
NanXiao 6bc74b4393 Use resolv library in Solaris 
For some Solaris flavours, the inet_aton in in resolv library.
Not linking this library will introduce link error.

Improves compatability with older Solaris and still
works on new Solaris.

Closes #1092
 2014-08-27 10:25:37 +02:00
Daniel Price 1aff711d90 Define AR to help with lua cross-compilation 
Closes #997
 2014-08-27 10:25:34 +02:00
Matt Stancliff 866f6bcc80 Cleanup double semicolons 
Closes #1161
 2014-08-27 10:25:31 +02:00
Chris Lamb 29ed617ac4 redis-check-dump: Prevent segfault if can't malloc 
Found by The Mayhem Team (Alexandre Rebert, Thanassis Avgerinos,
Sang Kil Cha, David Brumley, Manuel Egele) Cylab, Carnegie Mellon
University. See http://bugs.debian.org/716259 for more.

Signed-off-by: Chris Lamb <lamby@debian.org>

Fixes #1191
 2014-08-27 10:25:28 +02:00
charsyam a43e45e391 Remove unused LINE_BUFLEN definition 
Closes #1129
 2014-08-27 10:25:25 +02:00
Matt Stancliff 25a4699a9e Reject MOVE to non-integer DBs 
Previously, "MOVE key somestring" would move the key to
DB 0 which is just unexpected and wrong.
String as DB == error.

Test added too.

Modified by @antirez in order to use the getLongLongFromObject() API
instead of strtol().

Fixes #1428
 2014-08-27 10:25:22 +02:00
Matt Stancliff 09a3976478 pubsub: Return integers for NUMSUB, not strings 
Also adds test for numsub — due to tcl being tcl,
it doesn't capture the "numberness" of the fix,
but now we at least have one test case for numsub.

Closes #1561
 2014-08-27 10:25:15 +02:00
Xiaojie Zhang 60c748628b Avoid unnecessary decoding in ziplist.c 
Closes #1519
 2014-08-27 10:25:12 +02:00
Samuel Chou 96614f5866 Improve accuracy of HAVE_ATOMIC dependency check 
[I had to split out the clang check due to
 clang *really* not liking the __GLIBC_PREREQ macro; -matt]

Closes #1456
 2014-08-27 10:25:07 +02:00
Jim Howard d1df50ad59 Remove redundant event loop fd processing 
Reason:
https://github.com/antirez/redis/pull/1450#issuecomment-30432925

Closes #1450
 2014-08-27 10:25:04 +02:00
Xian Li 282c6106ce memtest: Add missing free() 
Closes #1614
 2014-08-27 10:25:00 +02:00
Joshua Anderson 25b5abe8cf src/Makefile: Reword "to run make test" message 
Closes #1610
 2014-08-27 10:24:56 +02:00
cubicdaiya a7c46df612 Use 'void' for zero-argument functions 
According to the C standard,
it is desirable to give the type 'void'
to functions have no argument.

Closes #1631
 2014-08-27 10:24:52 +02:00
Matt Stancliff 8dbd0134e5 Fix assert technical correctness 
dictAdd returns DICT_OK, not REDIS_OK. They both
have the same underlying values, so it works even though
the code is technically wrong.

Fixes #1512
 2014-08-27 10:24:46 +02:00
Jan-Erik Rediger 0c8999dfeb install_server.sh: add missing bang 
This was discovered by _bodya and reported in the IRC channel.
Everything worked fine as these scripts are always executed as shell
scripts.

Closes #1728
 2014-08-27 10:24:42 +02:00
Edgars Irmejs 0f9ab49d58 Change unixsocketperm comment to 700 from 755 
According to unix manuals, "Connecting to the socket object requires
read/write permission." -- mode 755 is useless for anybody
other than the owner.

Fixes #1696
 2014-08-27 10:24:39 +02:00
ripcurld00d 4be50f7fbb redis-check-dump: use names instead of magic nums 
Use constants to avoid magic numbers in `types`, which is an array
that stores the names of `REDIS` types.

Closes #1681
 2014-08-27 10:24:36 +02:00
Kashif Rasul 8fc2247c9b Fix issues raised by clang analyzer 
Modified by @antirez since the original fix to genInfoString() looked
weak. Probably the clang analyzer complained about `section` being
possibly NULL, and strcasecmp() called with a NULL pointer. In the
practice this can never happen, still for the sake of correctness
the right fix is not to modify only the first call, but to set `section`
to the value of "default" if it happens to be NULL.

Closes #1660
 2014-08-27 10:24:33 +02:00
antirez ebd25710dd Force quit when receiving a second SIGINT. 
Also quit ASAP when we are still loading a DB, since care is not needed
in this special condition, especially for a SIGINT.
 2014-08-27 10:24:29 +02:00
Matt Stancliff 524b2626f7 Add graceful exit when Ctrl-C is received 2014-08-27 10:24:26 +02:00
Matt Stancliff 78bde472c0 Fix intset midpoint selection 
The classic (min+max)/2 is provably unsafe.  Fixed
as recommended in research:
http://googleresearch.blogspot.com/2006/06/extra-extra-read-all-about-it-nearly.html

Fix inspired by @wjin, but I used a different approach.
(later, I found @kuebler fixed the same issue too).

Fixes #1741, #1602
 2014-08-27 10:24:22 +02:00
Jan-Erik Rediger 07b2779bff Handle large getrange requests 
Previously the end was casted to a smaller type
which resulted in a wrong check and failed
with values larger than handled by unsigned.

Closes #1847, #1844
 2014-08-27 10:24:19 +02:00
Matt Stancliff 4024220fe6 scripting: no eval with negative key count 
Negative key count causes segfault in Lua functions.

Fixes #1842
Closes #1843
 2014-08-27 10:24:15 +02:00
Jan-Erik Rediger 50669f7b66 redis-cli: fix latency result output 
(Cleaned up a little by @mattsta)

Closes #1774
 2014-08-27 10:24:11 +02:00
Wei Jin 9b53f79e00 Remove redundant else/return block 
Fixes #1741
 2014-08-27 10:24:08 +02:00
antirez 9840df7bad Clarify AIX "#undef hz". 2014-08-27 10:22:59 +02:00
siahl b5dc2c5b91 Add support for compiling on AIX 
Closes #1900
 2014-08-27 10:22:28 +02:00
Matt Stancliff cbd1607c6a redis-cli: Add --no-raw option 
Some people need formatted output even when they have no
interactive tty.

Fixes #760
 2014-08-27 10:21:19 +02:00
Matt Stancliff 9a039f76c2 redis-cli: stop showing incorrectly selected DB 
Previously redis-cli would happily show "-1" or "99999"
as valid DB choices.

Now, if the SELECT call returned an error, we don't update
the DB number in the CLI.

Inspired by @anupshendkar in #1313

Fixes #566, #1313
 2014-08-27 10:21:19 +02:00
Matt Stancliff 28b54b5d94 redis-cli: Re-attach selected DB after auth 
Previously, if you did SELECT then AUTH, redis-cli
would show your SELECT'd db even though it didn't
happen.

Note: running into this situation is a (hopefully) very limited
used case of people using multiple DBs and AUTH all at the same
time.

Fixes antirez#1639
 2014-08-27 10:21:19 +02:00
Yossi Gottlieb ff58d5a380 Fail SYNC if background save child aborted due to a signal. 2014-07-28 14:46:03 +02:00
Salvatore Sanfilippo eb7b3427e7 Merge pull request #1894 from veselov/2.8 
* fixed doc URL for keyspace events
 2014-07-28 09:38:33 +02:00
vps 4d7b3815d5 * fixed doc URL for keyspace events 2014-07-27 20:14:39 -07:00
antirez 9fa51dd4d1 Example redis.conf: improve slaveof description. 2014-07-25 14:20:31 +02:00
antirez d558af8e6f Fix util.c compilation by including stdint.h. 2014-07-23 18:01:51 +02:00
antirez e426e77896 Faster ll2string() implementation. 
Based on ideas documented in this blog post:

https://www.facebook.com/notes/facebook-engineering/three-optimization-tips-for-c/10151361643253920

The original code was modified to handle signed integers, reformetted to
fit inside the Redis code base, and was stress-tested with a program
in order to validate the implementation against snprintf().

Redis was measured to be measurably faster from the point of view of
clients in real-world operations because of this change, since sometimes
number to string conversion is used extensively (for example every time
a GET results into an integer encoded object to be returned to the
user).
 2014-07-23 14:58:26 +02:00
antirez 8a44e10848 PFSELFTEST: less false positives. 
This is just a quickfix, for the nature of the test the right way to fix
it is to average the error of N runs, since otherwise it is always
possible to get a false positive with a bad run, or to minimize too much
this possibility we may end testing with too much "large" error ranges.
 2014-07-23 11:45:03 +02:00
antirez 0bfdd4bb49 Test: check sorted set elements order after union. 
This is not a regression but issue #1786 showed the need for this test.
 2014-07-22 17:58:41 +02:00
antirez 098bb97703 Remove useless var and check in zunionInterGenericCommand(). 2014-07-22 17:58:26 +02:00
antirez 316e1db970 ZUNIONSTORE reimplemented for speed. 
The user @kjmph provided excellent ideas to improve speed of ZUNIONSTORE
(in certain cases by many order of magnitude), together with an
implementation of the ideas.

While the ideas were sounding, the implementation could be improved both
in terms of speed and clearness, so that's my attempt at reimplementing
the speedup proposed, trying to improve by directly using just a
dictionary with an embedded score inside, and reusing the single-pass
aggregate + order-later approach.

Note that you can't apply this commit without applying the previous
commit in this branch that adds a double in the dictEntry value union.

Issue #1786.
 2014-07-22 17:58:05 +02:00
antirez c8a6cd94b3 Add double field in dict.c entry value union. 2014-07-22 17:53:06 +02:00
antirez 7b11dfe05e Test: Pub/Sub PING. 2014-07-18 12:03:38 +02:00
antirez 156029b336 Test: small integer sharing depends on maxmemory policy. 2014-07-18 10:55:19 +02:00
antirez 06e9b3cad8 PING: backward compatible error for wrong number of args. 2014-07-18 10:46:21 +02:00
antirez ab5f20d19c tryObjectEncoding(): use shared objects with maxmemory and non-LRU policy. 
In order to make sure every object has its own private LRU counter, when
maxmemory is enabled tryObjectEncoding() does not use the pool of shared
integers. However when the policy is not LRU-based, it does not make
sense to do so, and it is much better to save memory using shared
integers.
 2014-07-18 10:46:05 +02:00
antirez 70e39481ab Variadic PING with support for Pub/Sub. 
PING can now be called with an additional arugment, behaving exactly
like the ECHO command. PING can now also be called in Pub/Sub mode (with
one more more subscriptions to channels / patterns) in order to trigger
the delivery of an asynchronous pong message with the optional payload.

This fixes issue #420.
 2014-07-18 10:41:47 +02:00
antirez 294bcfc4d2 PubSub clients refactoring and new PUBSUB flag. 
The code tested many times if a client had active Pub/Sub subscriptions
by checking the length of a list and dictionary where the patterns and
channels are stored. This was substituted with a client flag called
REDIS_PUBSUB that is simpler to test for. Moreover in order to manage
this flag some code was refactored.

This commit is believed to have no effects in the behavior of the
server.
 2014-07-18 10:41:22 +02:00
antirez 0c29cadd46 Redis 2.8.13. 2014-07-14 17:48:14 +02:00
michael-grunder 3df2ab673c Fix OBJECT arity 
Previously, the command definition for the OBJECT command specified
a minimum of two args (and that it was variadic), which meant that
if you sent this:

OBJECT foo

When cluster was enabled, it would result in an assertion/SEGFAULT
when Redis was attempting to extract keys.

It appears that OBJECT is not variadic, and only ever takes 3 args.

https://gist.github.com/michael-grunder/25960ce1508396d0d36a
 2014-07-14 15:38:32 +02:00
antirez c970507478 Fix typo in LATENCY DOCTOR output. 2014-07-11 10:57:34 +02:00
antirez bb10ad9c13 Test: more reliable AOF rewrite test under write load. 2014-07-10 16:42:52 +02:00
antirez c785d4cccc LATENCY command / monitor basic tests. 2014-07-10 16:24:24 +02:00
antirez e1fd73f634 LATENCY HISTORY returns empty array if no data is available. 
Previously we returned an error.
 2014-07-10 16:24:24 +02:00
antirez 315536d012 Test: enable latency monitor in the default config. 
This way as a side effect of running the test we also stress the latency
monitor data collection.
 2014-07-10 15:04:59 +02:00
antirez f0e67afd94 Test: AOF rewrite during write load. 2014-07-10 11:24:59 +02:00
antirez 1f3c097b1c Fixed a few missing newline in createLatencyReport(). 2014-07-09 20:02:01 +02:00
antirez e4db34e97c createLatencyReport(): compile before commit avoids commits. 2014-07-09 20:01:57 +02:00
antirez f9765990b2 createLatencyReport(), fix mount option name. 2014-07-09 20:01:53 +02:00
antirez 54ca682887 createLatencyReport(): initialize all advices to zero. 2014-07-09 20:01:48 +02:00
antirez 95d880a7bf LATENCY DOCTOR first implementation complete. 2014-07-09 20:01:44 +02:00
antirez 2e581217d9 Fix analyzeLatencyForEvent() MAD computation. 2014-07-09 20:01:41 +02:00
antirez 121a9a3579 LATENCY DOCTOR: initial draft and events summary output. 2014-07-09 20:01:37 +02:00
antirez 3dd113bf33 Latency: low level time series analysis implemented. 2014-07-09 20:01:33 +02:00
antirez 3e8f5113fa LATENCY SAMPLES renamed LATENCY HISTORY. 2014-07-09 20:01:30 +02:00
antirez c145854bac LATENCY RESET implemented. 2014-07-09 20:01:27 +02:00
antirez 65dbb3546a redis-benchmark: abort when all clients are disconnected. 2014-07-09 20:01:18 +02:00
antirez 4c68319370 Use fsync instead of aof_fsync in final AOF sync. 
This happens in the child process so we don't care about latency:
better to sync metadata as well.
 2014-07-09 20:01:01 +02:00
antirez f3eab7a9b4 Fixed conditional for aof-write-pending-fsync latency event selection. 2014-07-09 20:00:57 +02:00
antirez f3c4111edd Dependencies updated. 2014-07-09 20:00:50 +02:00
antirez 07b4d33734 Cast void* to char* to avoid waring in latencyCommand(). 2014-07-09 19:58:06 +02:00
antirez 58fb90fdac Properly initialize min/max in latency.c. 2014-07-09 19:57:59 +02:00
antirez c005315a64 latencyStartMonitor() modified to avoid warnings. 2014-07-09 19:57:53 +02:00
antirez e81b51ad8c Latency monitor: specialize delayed aof writes events. 2014-07-09 19:57:48 +02:00
antirez 8cc912db43 LATENCY GRAPH: filling under the curve is more readable. 2014-07-09 19:57:42 +02:00
antirez e905588062 LATENCY GRAPH implemented. 2014-07-09 19:57:37 +02:00
antirez b38996aab1 latencyTimeSeries structure max field type fixed. 2014-07-09 19:57:34 +02:00
antirez 062564c261 Free labels in freeSparklineSequence(). 2014-07-09 19:57:28 +02:00
antirez d20d89e630 LATENCY LATEST: add the max field. 2014-07-09 19:57:22 +02:00
antirez dc1c91a8bf Latency monitor trheshold value is now configurable. 
This commit adds both support for redis.conf and CONFIG SET/GET.
 2014-07-09 19:57:17 +02:00
antirez 350792bd72 ASCII sparklines generation API. 2014-07-09 19:57:08 +02:00
antirez ed95915d33 License added to latency.h. 2014-07-09 19:56:47 +02:00
antirez 34afc91030 Latency monitor turned off by default. 
It is not a good idea to bloat the code with gettimeofday() calls if the
instance is working well, and turning monitoring on at runtime is a
joke.
 2014-07-09 19:56:43 +02:00
antirez 35a62f7408 Latency monitor: more hooks around the code. 2014-07-09 19:56:43 +02:00
antirez 855c97ce1b Latency monitor: don't add new samples in the same second. 
Instead we update the old sample with the new latency if it is greater.
 2014-07-09 19:56:35 +02:00
antirez 3cb03b68a4 LATENCY LATEST implemented. 2014-07-09 19:56:30 +02:00
antirez e400ae60f5 Latency monitor: command duration is in useconds. Convert. 2014-07-09 19:04:02 +02:00
antirez a57eb3272d LATENCY SAMPLES implemented. 2014-07-09 19:03:59 +02:00
antirez 72c84acc99 Latency monitor: collect slow commands. 
We introduce the distinction between slow and fast commands since those
are two different sources of latency. An O(1) or O(log N) command without
side effects (can't trigger deletion of large objects as a side effect of
its execution) if delayed is a symptom of inherent latency of the system.

A non-fast command (commands that may run large O(N) computations) if
delayed may just mean that the user is executing slow operations.

The advices LATENCY should provide in this two different cases are
different, so we log the two classes of commands in a separated way.
 2014-07-09 19:03:26 +02:00
antirez 58ecc0bd68 Latency monitor: basic samples collection. 2014-07-09 18:57:57 +02:00
antirez 0baab66d16 Fixed 2.8 build breakage introduced back porting COMMAND. 2014-07-02 22:42:00 +02:00
antirez 670cf9598e Fix Solaris compilation due to ctime_r() call. 
Introduced in Redis 2.8.10 because of a change in Sentinel.
This closes issue #1837.
 2014-06-30 16:29:14 +02:00
antirez 41d7c90417 Test: fix instances.tcl restart_instance abort error. 2014-06-30 12:06:24 +02:00
antirez 63feb93a9f COMMAND COUNT subcommand added. 2014-06-27 18:38:08 +02:00
antirez 0a6649a7c0 COMMAND: fix argument parsing. 
This fixes detection of wrong subcommand (that resulted in the default
all-commands output instead) and allows COMMAND INFO to be called
without arguments (resulting into an empty array) which is useful in
programmtically generated calls like the following (in Ruby):

    redis.commands("command","info",*mycommands)

Note: mycommands may be empty.
 2014-06-27 18:38:08 +02:00
antirez 53377f8c6d COMMANDS command renamed COMMAND. 2014-06-27 18:38:03 +02:00
antirez f3efd529df COMMANDS command: remove static + aesthetic changes. 
Static was removed since it is needed in order to get symbols in stack
traces. Minor changes in the source code were operated to make it more
similar to the existing Redis code base.
 2014-06-27 18:37:34 +02:00
Matt Stancliff fdc5dbd5b1 Cluster: Add COMMANDS command 
COMMANDS returns a nested multibulk reply for each
command in the command table.  The reply for each
command contains:
  - command name
  - arity
  - array of command flags
  - start key position
  - end key position
  - key offset step
  - optional: if the keys are not deterministic and
    Redis uses an internal key evaluation function,
    the 6th field appears and is defined as a status
    reply of: REQUIRES ARGUMENT PARSING

Cluster clients need to know where the keys are in each
command to implement proper routing to cluster nodes.

Redis commands can have multiple keys, keys at offset steps, or other
issues where you can't always assume the first element after
the command name is the cluster routing key.

Using the information exposed by COMMANDS, client implementations
can have live, accurate key extraction details for all commands.

Also implements COMMANDS INFO [commands...] to return only a
specific set of commands instead of all 160+ commands live in Redis.
 2014-06-27 18:37:29 +02:00
antirez afe949efcb Remove infinite loop from PSYNC test. 
Added for debugging and forgot there.
 2014-06-27 18:36:59 +02:00
antirez 1066fed769 Test: hopefully more robust PSYNC test. 
This is supposed to fix issue #1417, but we'll know if this is enough
only after a couple of runs of the CI test without false positives.
 2014-06-27 18:36:59 +02:00
Matt Stancliff cd4f0dc775 Allow __powerpc__ to define HAVE_ATOMIC too 
From mailing list post https://groups.google.com/forum/#!topic/redis-db/D3k7KmJmYgM

In the file “config.h”, the definition HAVE_ATOMIC is used to indicate
if an architecture on which redis is implemented supports atomic
synchronization primitives.  Powerpc  supports atomic synchronization
primitives, however, it is not listed as one of the architectures
supported in config.h. This patch  adds the __powerpc__ to the list of
architectures supporting these primitives. The improvement of redis
due to the atomic synchronization on powerpc is significant,
around 30% to 40%, over the default implementation using pthreads.

This proposal adds __powerpc__ to the list of architectures designated
to support atomic builtins.
 2014-06-27 18:36:59 +02:00
Matt Stancliff 58fea46976 Allow atomic memory count update with C11 builtins 
From mailing list post https://groups.google.com/forum/#!topic/redis-db/QLjiQe4D7LA

In zmalloc.c the following primitives are currently used
to synchronize access to single global variable:
__sync_add_and_fetch
__sync_sub_and_fetch

In some architectures such as powerpc these primitives are overhead
intensive. More efficient C11 __atomic builtins are available with
newer GCC versions, see
http://gcc.gnu.org/onlinedocs/gcc-4.8.2/gcc/_005f_005fatomic-Builtins.html#_005f_005fatomic-Builtins

By substituting the following  __atomic… builtins:
__atomic_add_fetch
__atomic_sub_fetch

the performance improvement on certain architectures such as powerpc can be significant,
around 10% to 15%, over the implementation using __sync builtins while there is only slight uptick on
Intel architectures because it was already enforcing Intel Strongly ordered memory semantics.

The selection of __atomic built-ins can be predicated on the definition of ATOMIC_RELAXED
which Is available on in gcc 4.8.2 and later versions.
 2014-06-27 18:36:59 +02:00
Matt Stancliff ff3ca17bac Use predefined macro for used_memory() update 2014-06-27 18:36:59 +02:00
antirez 04749ee015 Fixed assert conditional in ROLE command test. 2014-06-26 22:13:57 +02:00
antirez 60ff8095d6 No more trailing spaces in Redis source code. 2014-06-26 18:52:16 +02:00
antirez e9cd75dcd1 CLIENT KILL: don't kill the master as a normal client. 
Technically the problem is due to the client type API that does not
return a special value for the master, however fixing it locally in the
CLIENT KILL command is better currently because otherwise we would
introduce a new output buffer limit class as a side effect.
 2014-06-26 18:43:42 +02:00
antirez ac6b656fd6 Old form of CLIENT KILL should still allow suicide. 2014-06-24 12:49:18 +02:00
antirez 32ebb3e651 Redis 2.8.12. 2014-06-23 16:56:18 +02:00
antirez 5d16a838db Sentinel test: more correct sentinels config reset. 
In the initialization test for each instance we used to unregister the
old master and register it again to clear the config.
However there is a race condition doing this: as soon as we unregister
and re-register "mymaster", another Sentinel can update the new
configuration with the old state because of gossip "hello" messages.

So the correct procedure is instead, unregister "mymaster" from all the
sentinel instances, and re-register it everywhere again.
 2014-06-23 14:04:09 +02:00
antirez d3c2918434 Sentinel implementation of ROLE. 2014-06-23 12:07:55 +02:00
antirez 95f492ee57 Silence different signs comparison warning in sds.c. 2014-06-23 11:53:50 +02:00
antirez 364464e790 Use Redis updated sds.c for deps/hiredis. 2014-06-23 11:53:45 +02:00
Matt Stancliff 8066334986 Sentinel: bind source address 
Some deployments need traffic sent from a specific address.  This
change uses the same policy as Cluster where the first listed bindaddr
becomes the source address for outgoing Sentinel communication.

Fixes #1667
 2014-06-23 11:53:40 +02:00
Matt Stancliff e59eef246b Add REDIS_BIND_ADDR access macro 
We need to access (bindaddr[0] || NULL) in a few places, so centralize
access with a nice macro.
 2014-06-23 11:53:25 +02:00
antirez a8ecd3f040 Dummy zmalloc.h restored into deps/hiredis. 2014-06-23 11:53:15 +02:00
Matt Stancliff 6ee25dc439 hiredis: Update to latest version 
This is hiredis f225c276be7fd0646019b51023e3f41566633dfe

This update includes all changes that diverged inside of Redis since
the last update.  This version also allows optional source address
binding for connections which we need for some Sentinel deployments.
 2014-06-23 11:53:12 +02:00
Matt Stancliff ed705f4789 Cancel SHUTDOWN if initial AOF is being written 
Fixes #1826 (and many other reports of the same problem)
 2014-06-23 10:09:25 +02:00
antirez 9a5408f23e Basic tests for the ROLE command. 2014-06-23 09:09:02 +02:00
antirez 061fd99767 Test: AOF test false positive when running in slow hosts. 
The bug was triggered by running the test with Valgrind (which is a lot
slower and more sensible to timing issues) after the recent changes
that made Redis more promptly able to reply with the -LOADING error.
 2014-06-21 16:04:20 +02:00
antirez 2b805ce119 Allow to call ROLE in LOADING state. 2014-06-21 15:29:12 +02:00
antirez 8c460a2872 ROLE command: array len fixed for slave output. 2014-06-21 15:29:11 +02:00
antirez 7d0992dad6 Sentinel: send SLAVEOF with MULTI, CLIENT KILL, CONFIG REWRITE. 
This implements the new Sentinel-Client protocol for the Sentinel part:
now instances are reconfigured using a transaction that ensures that the
config is rewritten in the target instance, and that clients lose the
connection with the instance, in order to be forced to: ask Sentinel,
reconnect to the instance, and verify the instance role with the new
ROLE command.
 2014-06-21 15:27:57 +02:00
antirez 674194ad42 CLIENT KILL API modified. 
Added a new SKIPME option that is true by default, that prevents the
client sending the command to be killed, unless SKIPME NO is sent.
 2014-06-21 15:27:57 +02:00
antirez 61d9a73d87 CLIENT KILL: fix closing link of the current client. 2014-06-21 15:27:57 +02:00
antirez 09dc6dadc1 New features for CLIENT KILL. 2014-06-21 15:27:57 +02:00
antirez cad13223f0 Assign an unique non-repeating ID to each new client. 
This will be used by CLIENT KILL and is also a good way to ensure a
given client is still the same across CLIENT LIST calls.

The output of CLIENT LIST was modified to include the new ID, but this
change is considered to be backward compatible as the API does not imply
you can do positional parsing, since each filed as a different name.
 2014-06-21 15:27:53 +02:00
antirez b6a26b52bf Client types generalized. 
Because of output buffer limits Redis internals had this idea of type of
clients: normal, pubsub, slave. It is possible to set different output
buffer limits for the three kinds of clients.

However all the macros and API were named after output buffer limit
classes, while the idea of a client type is a generic one that can be
reused.

This commit does two things:

1) Rename the API and defines with more general names.
2) Change the class of clients executing the MONITOR command from "slave"
   to "normal".

"2" is a good idea because you want to have very special settings for
slaves, that are not a good idea for MONITOR clients that are instead
normal clients even if they are conceptually slave-alike (since it is a
push protocol).

The backward-compatibility breakage resulting from "2" is considered to
be minimal to care, since MONITOR is a debugging command, and because
anyway this change is not going to break the format or the behavior, but
just when a connection is closed on big output buffer issues.
 2014-06-21 15:27:12 +02:00
antirez 8060de98f8 ROLE output improved for slaves. 
Info about the replication state with the master added.
 2014-06-21 15:27:12 +02:00
antirez 41a15205b6 ROLE command added. 
The new ROLE command is designed in order to provide a client with
informations about the replication in a fast and easy to use way
compared to the INFO command where the same information is also
available.
 2014-06-21 15:27:12 +02:00
antirez d8d415e717 CLIENT LIST speedup via peerid caching + smart allocation. 
This commit adds peer ID caching in the client structure plus an API
change and the use of sdsMakeRoomFor() in order to improve the
reallocation pattern to generate the CLIENT LIST output.

Both the changes account for a very significant speedup.
 2014-06-21 15:27:08 +02:00
antirez 52189cb9c3 Use sdscatfmt() in getClientInfoString() to make it faster. 2014-06-21 15:27:02 +02:00
antirez 4acc3daa92 Added new sdscatfmt() %u and %U format specifiers. 
This commit also fixes a bug in the implementation of sdscatfmt()
resulting from stale references to the SDS string header after
sdsMakeRoomFor() calls.
 2014-06-21 15:26:58 +02:00
antirez 3a915ace44 sdscatfmt() added to SDS library. 
sdscatprintf() relies on printf() family libc functions and is sometimes
too slow in critical code paths. sdscatfmt() is an alternative which is:

1) Far less capable.
2) Format specifier uncompatible.
3) Faster.

It is suitable to be used in those speed critical code paths such as
CLIENT LIST output generation.
 2014-06-21 15:26:54 +02:00
antirez ebc62a6f35 Jemalloc updated to 3.6.0. 
Not a single bug in about 3 months, and our previous version was
too old (3.2.0).
 2014-06-20 15:01:36 +02:00
antirez a0e0fba7fa Jemalloc update script added. 2014-06-20 14:53:47 +02:00
antirez 2355ae9bc6 Sentinel test: tolerate larger delays in init tests. 2014-06-19 16:10:20 +02:00
antirez fd2faee5de Sentinel test: unit 02, avoid some time related false positives. 2014-06-19 16:10:17 +02:00
antirez 93ee0f2666 Sentinel: send hello messages ASAP after config change. 
Eventual configuration convergence is guaranteed by our periodic hello
messages to all the instances, however when there are important notices
to share, better make a phone call. With this commit we force an hello
message to other Sentinal and Redis instances within the next 100
milliseconds of a config update, which is practically better than
waiting a few seconds.
 2014-06-19 15:25:57 +02:00
antirez 9b883974af Sentinel: handle SRI_PROMOTED flag correctly. 
Lack of check of the SRI_PROMOTED flag caused Sentienl to act with the
promoted slave turned into a master during failover like if it was a
normal instance.

Normally this problem was not apparent because during real failovers the
old master is down so the bugged code path was not entered, however with
manual failovers via the SENTINEL FAILOVER command, the problem was
easily triggered.

This commit prevents promoted slaves from getting reconfigured, moreover
we now explicitly check that during a failover the slave turning into a
master is the one we selected for promotion and not a different one.
 2014-06-19 10:35:03 +02:00
antirez 7665cc67b3 Sentinel test: add manual failover test. 2014-06-19 10:32:38 +02:00
antirez 1267e28edc Test: use higher level redis.tcl proc to read replies. 2014-06-18 16:25:45 +02:00
antirez 2cff0c041f Test: Tcl client initial support for automatic reconnection. 2014-06-18 16:19:22 +02:00
antirez 9a39db3bb6 Test: tcl client, unset deferred var on close. 2014-06-18 16:19:19 +02:00
antirez b0964b7dd9 redis.tcl: return I/O error message when peer closes connection. 2014-06-18 16:18:45 +02:00
antirez eef369b66a Sentinel framework new layout back ported 2.8. 
Functionally the same, but makes cherry picking simpler.
 2014-06-18 16:14:33 +02:00
Matt Stancliff f392c7737c Add correct exit value to failed tests 
(Note from @antirez: modified to apply to 2.8)
 2014-06-18 15:03:06 +02:00
Alex Suraci d178cd03ed add missing signal.h include 2014-06-18 09:34:23 +02:00
Matt Stancliff 1b044af402 Add SIGINT handler to cli --intrinsic-latency 
If we run a long latency session and want to Ctrl-C out of it,
it's nice to still get the summary output at the end.
 2014-06-17 23:40:05 +02:00
antirez 258f3cd360 Scripting: regression test for issue #1811. 2014-06-16 10:30:42 +02:00
antirez ca6b95dfdd Fix semantics of Lua calls to SELECT. 
Lua scripts are executed in the context of the currently selected
database (as selected by the caller of the script).

However Lua scripts are also free to use the SELECT command in order to
affect other DBs. When SELECT is called frm Lua, the old behavior, before
this commit, was to automatically set the Lua caller selected DB to the
last DB selected by Lua. See for example the following sequence of
commands:

    SELECT 0
    SET x 10
    EVAL "redis.call('select','1')" 0
    SET x 20

Before this commit after the execution of this sequence of commands,
we'll have x=10 in DB 0, and x=20 in DB 1.

Because of the problem above, there was a bug affecting replication of
Lua scripts, because of the actual implementation of replication. It was
possible to fix the implementation of Lua scripts in order to fix the
issue, but looking closely, the bug is the consequence of the behavior
of Lua ability to set the caller's DB.

Under the old semantics, a script selecting a different DB, has no simple
ways to restore the state and select back the previously selected DB.
Moreover the script auhtor must remember that the restore is needed,
otherwise the new commands executed by the caller, will be executed in
the context of a different DB.

So this commit fixes both the replication issue, and this hard-to-use
semantics, by removing the ability of Lua, after the script execution,
to force the caller to switch to the DB selected by the Lua script.

The new behavior of the previous sequence of commadns is to just set
X=20 in DB 0. However Lua scripts are still capable of writing / reading
from different DBs if needed.

WARNING: This is a semantical change that will break programs that are
conceived to select the client selected DB via Lua scripts.

This fixes issue #1811.
 2014-06-16 10:30:37 +02:00
antirez b0f3922383 Redis 2.8.11. 2014-06-11 10:30:05 +02:00
antirez a3b0dbccf5 Scripting: Fix for a #1118 regression simplified. 
It is more straightforward to just test for a numerical type avoiding
Lua's automatic conversion. The code is technically more correct now,
however Lua should automatically convert to number only if the original
type is a string that "looks like a number", and not from other types,
so practically speaking the fix is identical AFAIK.
 2014-06-11 10:14:26 +02:00
Matt Stancliff ba76daa42e Scripting: Fix regression from #1118 
The new check-for-number behavior of Lua arguments broke
users who use large strings of just integers.

The Lua number check would convert the string to a number, but
that breaks user data because
Lua numbers have limited precision compared to an arbitrarily
precise number wrapped in a string.

Regression fixed and new test added.

Fixes #1118 again.
 2014-06-11 10:14:26 +02:00
Matt Stancliff 8f774ee7ca Fix lack of strtold under Cygwin 
Renaming strtold to strtod then casting
the result is the standard way of dealing with
no strtold in Cygwin.
 2014-06-09 11:43:19 +02:00
Matt Stancliff 2af2414344 Fix lack of SA_ONSTACK under Cygwin 
Fixes #232
 2014-06-09 11:43:19 +02:00
Matt Stancliff b4f9761d85 Fix blocking operations from missing new lists 
Behrad Zari discovered [1] and Josiah reported [2]: if you block
and wait for a list to exist, but the list creates from
a non-push command, the blocked client never gets notified.

This commit adds notification of blocked clients into
the DB layer and away from individual commands.

Lists can be created by [LR]PUSH, SORT..STORE, RENAME, MOVE,
and RESTORE.  Previously, blocked client notifications were
only triggered by [LR]PUSH.  Your client would never get
notified if a list were created by SORT..STORE or RENAME or
a RESTORE, etc.

Blocked client notification now happens in one unified place:
  - dbAdd() triggers notification when adding a list to the DB

Two new tests are added that fail prior to this commit.

All test pass.

Fixes #1668

[1]: https://groups.google.com/forum/#!topic/redis-db/k4oWfMkN1NU
[2]: #1668
 2014-06-09 11:39:44 +02:00
Andy Grunwald 6ab4775a2a Fixed typo in word avarege in result message of --intrinsic-latency analyzer 2014-06-06 11:20:04 +02:00
Jan-Erik Rediger b98ee12f61 Small typo fixed 2014-06-06 10:45:46 +02:00
yoav 3783d90e53 Fix eval usage in tests to conform with eval semantics 2014-06-06 10:44:08 +02:00
zionwu e3112bed8d fix issue 1787 2014-06-06 10:38:04 +02:00
antirez 987a47b823 Redis 2.8.10. 2014-06-05 11:02:23 +02:00
antirez 430bec3461 Don't process min-slaves-to-write for slaves. 
Replication is totally broken when a slave has this option, since it
stops accepting updates from masters.

This fixes issue #1434.
 2014-06-05 10:50:31 +02:00
antirez 3a30be51aa Tests for min-slaves-* feature. 2014-06-05 10:50:31 +02:00
antirez 42504169c7 Fixed dbuf variable scope in luaRedisGenericCommand(). 
I'm not sure if while the visibility is the inner block, the fact we
point to 'dbuf' is a problem or not, probably the stack var isx
guaranteed to live until the function returns. However obvious code is
better anyway.
 2014-06-04 18:58:58 +02:00
antirez 26359d570b Regression test for issue #1118. 2014-06-04 18:51:31 +02:00
antirez 768994b683 Scripting: better Lua number -> string conversion in luaRedisGenericCommand(). 
The lua_to*string() family of functions use a non optimal format
specifier when converting integers to strings. This has both the problem
of the number being converted in exponential notation, which we don't
use as a Redis return value when floating point numbers are involed,
and, moreover, there is a loss of precision since the default format
specifier is not able to represent numbers that must be represented
exactly in the IEEE 754 number mantissa.

The new code handles it as a special case using a saner conversion.

This fixes issue #1118.
 2014-06-04 18:45:49 +02:00
antirez 1c0c42e668 More trailing spaces in sentinel.c removed. 2014-05-28 15:46:14 +02:00
Matt Stancliff 75266018ee Disable recursive watchdog signal handler 
If we are in the signal handler, we don't want to handle
the signal again.  In extreme cases, this can cause a stack overflow
and segfault Redis.

Fixes #1771
 2014-05-26 17:53:40 +02:00
antirez edc41baca5 Sentinel example config: explain you don't need to specify slaves. 2014-05-26 10:18:07 +02:00
antirez 1a14eba2d5 Test: fixed scripting.tcl test false positive. 2014-05-22 15:46:06 +02:00
antirez f482349707 Process events with processEventsWhileBlocked() when blocked. 
When we are blocked and a few events a processed from time to time, it
is smarter to call the event handler a few times in order to handle the
accept, read, write, close cycle of a client in a single pass, otherwise
there is too much latency added for clients to receive a reply while the
server is busy in some way (for example during the DB loading).
 2014-05-22 15:46:02 +02:00
antirez f3d3c606f1 Accept multiple clients per iteration. 
When the listening sockets readable event is fired, we have the chance
to accept multiple clients instead of accepting a single one. This makes
Redis more responsive when there is a mass-connect event (for example
after the server startup), and in workloads where a connect-disconnect
pattern is used often, so that multiple clients are waiting to be
accepted continuously.

As a side effect, this commit makes the LOADING, BUSY, and similar
errors much faster to deliver to the client, making Redis more
responsive when there is to return errors to inform the clients that the
server is blocked in an not interruptible operation.
 2014-05-22 15:45:58 +02:00
antirez 4113473997 While ANET_ERR is -1, check syscall retval for -1 itself. 2014-05-22 15:45:54 +02:00
antirez e8669ca8a7 Regression test for issue #1764. 2014-05-20 16:20:45 +02:00
michael-grunder d491a479bc Fix LUA_OBJCACHE segfault. 
When scanning the argument list inside of a redis.call() invocation
for pre-cached values, there was no check being done that the
argument we were on was in fact within the bounds of the cache size.

So if a redis.call() command was ever executed with more than 32
arguments (current cache size #define setting) redis-server could
segfault.
 2014-05-20 16:20:38 +02:00
antirez d575f7a147 Remove trailing spaces from scripting.c 2014-05-20 16:11:28 +02:00
antirez 6d2fddd2a3 Remove trailing spaces from sentinel.c. 2014-05-20 14:22:34 +02:00
antirez ad935769ae HyperLogLog regression test for issue #1762. 2014-05-19 15:45:24 +02:00
Mike Trinkala 43278af63e Correct the HyperLogLog stale cache flag to prevent unnecessary computations. 
Set the MSB as documented.
 2014-05-19 15:45:24 +02:00
antirez 3203e2dcdd Fix 2.8 backport of fastscript branch. 
No REDIS_ENCODING_EMBSTR in 2.8 internals.
 2014-05-14 18:39:42 +02:00
Akos Vandra 6e478ec890 Fixed possible buffer overflow bug if RDB file is corrupted. 
(Note: commit message modified by @antirez for clarity).
 2014-05-12 11:49:16 +02:00
Akos Vandra c0eaac9572 fixed possible buffer overflow error 2014-05-12 11:19:36 +02:00
antirez 23011dea1f Sentinel: Add "dir /tmp" directive in example sentinel.conf. 2014-05-12 10:46:56 +02:00
antirez 6ae4c1fcbc DEBUG POPULATE: call dictExpand() to avoid useless rehashing. 2014-05-09 15:05:31 +02:00
antirez 13d8b2b096 Sentinel: log when a failover will be attempted again. 
When a Sentinel performs a failover (successful or not), or when a
Sentinel votes for a different Sentinel trying to start a failover, it
sets a min delay before it will try to get elected for a failover.

While not strictly needed, because if multiple Sentinels will try
to failover the same master at the same time, only one configuration
will eventually win, this serialization is practically very useful.
Normal failovers are cleaner: one Sentinel starts to failover, the
others update their config when the Sentinel performing the failover
is able to get the selected slave to move from the role of slave to the
one of master.

However currently this timeout was implicit, so users could see
Sentinels not reacting, after a failed failover, for some time, without
giving any feedback in the logs to the poor sysadmin waiting for clues.

This commit makes Sentinels more verbose about the delay: when a master
is down and a failover attempt is not performed because the delay has
still not elaped, something like that will be logged:

    Next failover delay: I will not start a failover
    before Thu May  8 16:48:59 2014
 2014-05-08 16:49:45 +02:00
antirez 909d1883e7 Sentinel: generate +config-update-from event when a new config is received. 
This event makes clear, before the switch-master event is generated,
that a Sentinel received a configuration update from another Sentinel.
 2014-05-08 16:49:45 +02:00
antirez 80b96bc4a7 Scripting test: check that Lua can call commands rewirting argv. 
SPOP, tested in the new test, is among the commands rewritng the
client->argv argument vector (it gets rewritten as SREM) for command
replication purposes.

Because of recent optimizations to client->argv caching in the context
of the Lua internal Redis client, it is important to test for SPOP to be
callable from Lua without bad effects to the other commands.
 2014-05-07 16:17:11 +02:00
antirez 093eab56de Test: handle new osx 'leaks' error. 
Sometimes the process is still there but no longer in a state that can
be checked (after being killed). This used to happen after a call to
SHUTDOWN NOSAVE in the scripting unit, causing a false positive.
 2014-05-07 16:17:11 +02:00
antirez 32971f6251 Scripting: objects caching for Lua c->argv creation. 
Reusing small objects when possible is a major speedup under certain
conditions, since it is able to avoid the malloc/free pattern that
otherwise is performed for every argument in the client command vector.
 2014-05-07 16:17:11 +02:00
antirez 87516f1a4d Scripting: Use faster API for Lua client c->argv creation. 
Replace the three calls to Lua API lua_tostring, lua_lua_strlen,
and lua_isstring, with a single call to lua_tolstring.

~ 5% consistent speed gain measured.
 2014-05-07 16:17:11 +02:00
antirez 1b5524f697 Scripting: don't call lua_gc() after Lua script run. 
Calling lua_gc() after every script execution is too expensive, and
apparently does not make the execution smoother: the same peak latency
was measured before and after the commit.

This change accounts for scripts execution speedup in the order of 10%.
 2014-05-07 16:17:11 +02:00
antirez 8237d88f48 Scripting: cache argv in luaRedisGenericCommand(). 
~ 4% consistently measured speed improvement.
 2014-05-07 16:17:11 +02:00
antirez 6938a3d58f Fixed missing c->bufpos reset in luaRedisGenericCommand(). 
Bug introduced when adding a fast path to avoid copying the reply buffer
for small replies that fit into the client static buffer.
 2014-05-07 16:17:11 +02:00
antirez 08ea98c9e9 Scripting: replace tolower() with faster code in evalGenericCommand(). 
The function showed up consuming a non trivial amount of time in the
profiler output. After this change benchmarking gives a 6% speed
improvement that can be consistently measured.
 2014-05-07 16:17:11 +02:00
antirez 3a1a0c3240 Scripting: luaRedisGenericCommand() fast path for buffer-only replies. 
When the reply is only contained in the client static output buffer, use
a fast path avoiding the dynamic allocation of an SDS string to
concatenate the client reply objects.
 2014-05-07 16:17:11 +02:00
antirez d70281575c Define HAVE_ATOMIC for clang. 2014-05-07 16:17:11 +02:00
antirez 67e350367d Scripting: simpler reply buffer creation in luaRedisGenericCommand(). 
It if faster to just create the string with a single sdsnewlen() call.
If c->bufpos is zero, the call will simply be like sdsemtpy().
 2014-05-07 16:17:11 +02:00
Matt Stancliff b06de30dd7 Add test for deleting an expired key 
Verify proper expire-before-delete behavior.

This test passes with the expire-before-delete commit and fails
without it.
 2014-04-23 16:14:40 +02:00
Matt Stancliff e9d0d86dbf Check key expiration before deleting 
Deleting an expired key should return 0, not success.

Fixes #1648
 2014-04-23 16:14:40 +02:00
Glauber Costa 96b5dc0b98 fix null pointer access with no file pointer 
I happen to be working on a system that lacks urandom. While the code does try
to handle this case and artificially create some bytes if the file pointer is
empty, it does try to close it unconditionally, leading to a segfault.
 2014-04-23 15:35:12 +02:00
Salvatore Sanfilippo 3d2af57974 Merge pull request #1700 from nirvdrum/patch-1 
Fixed typos.
 2014-04-23 11:55:50 +02:00
antirez 35252037fd Missing return REDIS_ERR added to processMultibulkBuffer(). 
When we set a protocol error we should return with REDIS_ERR to let the
caller know it should stop processing the client.

Bug found in a code auditing related to issue #1699.
 2014-04-23 10:21:26 +02:00
Kevin Menard 662f505727 Fixed typos. 2014-04-22 10:51:19 -04:00
antirez de7ebf83e1 redis-cli help.h updated. 2014-04-22 16:15:01 +02:00
antirez 91dcc5b12f generate-command-help.rb updated with new hyperloglog group. 2014-04-22 16:15:01 +02:00
antirez 26375709c5 Redis 2.8.9. 2014-04-22 10:14:57 +02:00
antirez c5ee48a9d9 Fuzzy test for ZREMRANGEBYLEX added. 2014-04-18 16:16:20 +02:00
antirez be145453c6 ZREMRANGEBYLEX memory leak removed calling zslFreeLexRange(). 2014-04-18 16:16:20 +02:00
antirez 115782b1d7 PFCOUNT multi-key test added. 2014-04-18 16:16:20 +02:00
antirez 2cd20d3c4c Speedup hllRawSum() processing 8 bytes per iteration. 
The internal HLL raw encoding used by PFCOUNT when merging multiple keys
is aligned to 8 bits (1 byte per register) so we can exploit this to
improve performances by processing multiple bytes per iteration.

In benchmarks the new code was several times faster with HLLs with many
registers set to zero, while no slowdown was observed with populated
HLLs.
 2014-04-18 16:16:20 +02:00
antirez 3f5cd0c660 Speedup SUM(2^-reg[m]) in HyperLogLog computation. 
When the register is set to zero, we need to add 2^-0 to E, which is 1,
but it is faster to just add 'ez' at the end, which is the number of
registers set to zero, a value we need to compute anyway.
 2014-04-18 16:16:20 +02:00
antirez f0eca148ff PFCOUNT support for multi-key union. 2014-04-18 16:16:20 +02:00
antirez d25fd8be64 HyperLogLog low level merge extracted from PFMERGE. 2014-04-18 16:16:20 +02:00
antirez 4a26648a22 ZREMRANGEBYLEX implemented. 2014-04-18 16:16:20 +02:00
antirez 3fc22561e7 Always pass sorted set range objects by reference. 2014-04-18 16:16:04 +02:00
antirez ea0af9ace3 ZREMRANGE* commands refactored into a single generic function. 2014-04-18 16:14:38 +02:00
antirez 019e6f1889 Sorted set lex ranges stress tester. 2014-04-18 16:14:38 +02:00
antirez 9c9aef24a2 Basic ZRANGEBYLEX / ZLEXCOUNT tests. 2014-04-17 00:08:38 +02:00
antirez 11f89497b4 Pass by pointer and release of lex ranges. 
Given that the code was written with a 2 years pause... something
strange happened in the middle. So there was no function to free a
lex range min/max objects, and in some places the range was passed by
value.
 2014-04-17 00:08:37 +02:00
antirez d975bb587f ZLEXCOUNT implemented. 
Like ZCOUNT for lexicographical ranges.
 2014-04-16 15:26:28 +02:00
antirez 71dfb87f76 More HyperLogLog tests. 2014-04-16 15:26:28 +02:00
antirez f2e59eae44 HyperLogLog invalid representation error code set to INVALIDOBJ. 2014-04-16 15:26:28 +02:00
antirez 469b7c5168 PFDEBUG TODENSE added. 
Converts HyperLogLogs from sparse to dense. Used for testing.
 2014-04-16 15:26:28 +02:00
antirez 614fcd491e User-defined switch point between sparse-dense HLL encodings. 2014-04-16 15:26:28 +02:00
antirez 08da2b79fc PFSELFTEST improved with sparse encoding checks. 2014-04-16 15:26:27 +02:00
antirez 5a0bab2a8f PFDEBUG ENCODING added. 2014-04-16 15:26:27 +02:00
antirez 05563310f4 Set HLL_SPARSE_MAX to 3000. 
After running a few benchmarks, 3000 looks like a reasonable value to
keep HLLs with a few thousand elements small while the CPU cost is
still not huge.

This covers all the cases where the dense representation would use N
orders of magnitude more space, like in the case of many HLLs with
carinality of a few tens or hundreds.

It is not impossible that in the future this gets user configurable,
however it is easy to pick an unreasoable value just looking at savings
in the space dimension without checking what happens in the time
dimension.
 2014-04-16 15:26:27 +02:00
antirez e77a3e40f2 Error message for invalid HLL objects unified. 2014-04-16 15:26:27 +02:00
antirez c4b2a7a7f4 PFMERGE fixed to work with sparse encoding. 2014-04-16 15:26:27 +02:00
antirez ee682c4c48 Mark PFDEBUG as write command in the commands table. 
It is safer since it is able to have side effects.
 2014-04-16 15:26:27 +02:00
antirez 9ef8ef410b Correctly replicate PFDEBUG GETREG. 
Even if it is a debugging command, make sure that when it forces a
change in encoding, the command is propagated.
 2014-04-16 15:26:27 +02:00
antirez 91e05e1618 Added assertion in hllSparseAdd() when promotion to dense occurs. 
If we converted to dense, a register must be updated in the dense
representation.
 2014-04-16 15:26:27 +02:00
antirez 68a4bf70e0 hllSparseAdd(): speed optimization. 
Mostly by reordering opcodes check conditional by frequency of opcodes
in larger sparse-encoded HLLs.
 2014-04-16 15:26:27 +02:00
antirez 21e328a6d7 Detect corrupted sparse HLLs in hllSparseSum(). 2014-04-16 15:26:27 +02:00
antirez 4244aa29f9 hllSparseAdd(): faster code removing conditional. 
Bottleneck found profiling. Big run time improvement found when testing
after the change.
 2014-04-16 15:26:27 +02:00
antirez 4cd11077d2 Comment typo in hllSparseAdd(). first -> fits. 2014-04-16 15:26:27 +02:00
antirez 08183cfef5 Merge adjacent VAL opcodes in hllSparseAdd(). 
As more values are added splitting ZERO or XZERO opcodes, try to merge
adjacent VAL opcodes if they have the same value.
 2014-04-16 15:26:27 +02:00
antirez d428569d13 More robust HLL_SPARSE macros protecting 'p' with parens. 
Now the macros will work with arguments such as "ptr+1".
 2014-04-16 15:26:27 +02:00
antirez 2d1e35a58a hllSparseAdd() opcode seek stop condition fixed. 2014-04-16 15:26:27 +02:00
antirez 29c7d1974b Fixed error message generation in PFDEBUG GETREG. 
Bulk length for registers was emitted too early, so if there was a bug
the reply looked like a long array with just one element, blocking the
client as result.
 2014-04-16 15:26:27 +02:00
antirez 023f2f7a10 Fixed memmove() count in hllSparseAdd(). 2014-04-16 15:26:27 +02:00
antirez 91391760ca hllSparseAdd(): more correct dense conversion conditional. 
We want to promote if the total string size exceeds the resulting size
after the upgrade.
 2014-04-16 15:26:27 +02:00
antirez 5d87160f8a hllSparseToDense(): sanity check added. 
The function checks if all the HLL_REGISTERS were processed during the
convertion from sparse to dense encoding, returning REDIS_OK or
REDIS_ERR to signal a corruption problem.

A bug in PFDEBUG GETREG was fixed: when the object is converted to the
dense representation we need to reassign the new pointer to the header
structure pointer.
 2014-04-16 15:26:27 +02:00
antirez a18e7afbe8 PFDEBUG DECODE added. 
Provides a human readable description of the opcodes composing a
run-length encoded HLL (sparse encoding).
The command is only useful for debugging / development tasks.
 2014-04-16 15:26:27 +02:00
antirez 5c86d9e7b0 PFDEBUG added, PFGETREG removed. 
PFDEBUG will be the interface to do debugging tasks with a key
containing an HLL object.
 2014-04-16 15:26:27 +02:00
antirez 0e0ddcc4fa hllSparseToDense API changed to take ref to object. 
The new API takes directly the object doing everything needed to
turn it into a dense representation, including setting the new
representation as object->ptr.
 2014-04-16 15:26:27 +02:00
antirez 533235f7e3 hllSparseAdd() sanity check for span != 0 added. 2014-04-16 15:26:27 +02:00
antirez f12fb49bb2 Fix hllSparseAdd() new sequence replacement when next is NULL. 
sdsIncrLen() must be called anyway even if we are replacing the last
oppcode of the sparse representation.
 2014-04-16 15:26:27 +02:00
antirez a3e2adc0ae Fix seqlen computation in hllSparseAdd(). 2014-04-16 15:26:27 +02:00
antirez eaa73ef73a Abstract hllSparseAdd() / hllDenseAdd() via hllAdd(). 2014-04-16 15:26:27 +02:00
antirez 95de3eaea0 hllSparseSum(): multiply 1 * runlen for zero entries. 2014-04-16 15:26:27 +02:00
antirez c277169ed7 Macro HLL_SPARSE_XZERO_LEN fixed. 2014-04-16 15:26:27 +02:00
antirez b6ca71a40a Fix HLL sparse object creation #2. 
Two vars initialized to wrong values in createHLLObject().
 2014-04-16 15:26:27 +02:00
antirez 8b39fee4c8 Increment pointer while iterating sparse HLL object. 2014-04-16 15:26:27 +02:00
antirez c1ddead779 Fix HLL sparse object creation. 
The function didn't considered the fact that each XZERO opcode is
two bytes.
 2014-04-16 15:26:27 +02:00
antirez 0a7ba6faef Create HyperLogLog objects with sparse encoding. 2014-04-16 15:26:27 +02:00
antirez 8e889728bf HyperLogLog sparse to dense conversion function. 2014-04-16 15:26:27 +02:00
antirez aadd513160 HyperLogLog sparse representation initial implementation. 
Code never tested, but the basic layout is shaped in this commit.
Also missing:

1) Sparse -> Dense conversion function.
2) New HLL object creation using the sparse representation.
3) Implementation of PFMERGE for the sparse representation.
 2014-04-16 15:26:27 +02:00
antirez e5e5facba1 hllCount() refactored to support multiple representations. 2014-04-16 15:26:27 +02:00
antirez 88847a216f hllAdd() refactored into two functions. 
Also dense representation access macro renamed accordingly.
 2014-04-16 15:26:27 +02:00
antirez 21c97e71cd HyperLogLog refactoring to support different encodings. 
Metadata are now placed at the start of the representation as an header.
There is a proper structure to access the representation.
Still work to do in order to truly abstract the implementation from the
representation, commands still work assuming dense representation.
 2014-04-16 15:26:27 +02:00
antirez ed96e115d0 HyperLogLog sparse representation slightly modified. 
After running a few simulations with different alternative encodings,
it was found that the VAL opcode performs better using 5 bits for the
value and 2 bits for the run length, at least for cardinalities in the
range of interest.
 2014-04-16 15:26:27 +02:00
antirez 93c75f1bf4 HyperLogLog sparse representation description and macros. 2014-04-16 15:26:26 +02:00
antirez 2b1385b9ec Add casting to match printf format. 
adjustOpenFilesLimit() and clusterUpdateSlotsWithConfig() that were
assuming uint64_t is the same as unsigned long long, which is true
probably for all the systems out there that we target, but still GCC
emitted a warning since technically they are two different types.
 2014-04-16 15:26:22 +02:00
antirez 12d1d18b59 ZRANGEBYLEX and ZREVRANGEBYLEX implementation. 2014-04-16 15:26:09 +02:00
antirez a10bfded15 PFCOUNT: always unshare/decode the object. 
This will be a non-op most of the times since the object will be
unshared / decoded, however it is more technically correct to start this
way since the object may be decoded even in the read-only code path.
 2014-04-16 15:26:09 +02:00
antirez ee764b0f8d Changed HyperLogLog hash seed to a non-zero value. 
Using a seed of zero has the side effect of having the empty string
hashing to what is a very special case in the context of HyperLogLog: a
very long run of zeroes.

This did not influenced the correctness of the result with 16k registers
because of the harmonic mean, but still it is inconvenient that a so
obvious value maps to a so special hash.

The seed 0xadc83b19 is used instead, which is the first 64 bits of the
SHA1 of the empty string.

Reference: issue #1657.
 2014-04-16 15:26:09 +02:00
antirez 091f5677a0 Initial HyperLogLog tests. 2014-04-16 15:26:09 +02:00
antirez a97675923d Return "WRONGTYPE" error on PF* type mismatch. 2014-04-16 15:26:09 +02:00
antirez faa7f259cc Fix PFADD infinite loop. 
We need to guarantee that the last bit is 1, otherwise an element may
hash to just zeroes with probability 1/(2^64) and trigger an infinite
loop.

See issue #1657.
 2014-04-16 15:26:09 +02:00
antirez 8ae6887091 Make hll-gnuplot-graph.rb callable from cli. 2014-04-16 15:26:09 +02:00
antirez bdd8701c60 Remove HyperLogLog type checking duplicated code. 2014-04-16 15:26:09 +02:00
antirez b3baa51403 PFGETREG added for testing purposes. 
The new command allows to get a dump of the registers stored
into an HyperLogLog data structure for testing / debugging purposes.
 2014-04-16 15:26:09 +02:00
antirez 348d5ea246 PFCOUNT: unshare the object when cached cardinality is modified. 2014-04-16 15:26:09 +02:00
antirez ba0dc0be2c PFSELFTEST improved to test the approximation error. 2014-04-16 15:26:09 +02:00
antirez c2cd48718d hll-gnuplot-graph.rb improved with new filter. 
The function to generate graphs is also more flexible as now includes
step and max value. The step of the samples generation function is no
longer limited to min step of 1000.
 2014-04-16 15:26:09 +02:00
antirez ff551fad24 HyperLogLog: added magic / version. 
This will allow future changes like compressed representations.
Currently the magic is not checked for performance reasons but this may
change in the future, for example if we add new types encoded in strings
that may have the same size of HyperLogLogs.
 2014-04-16 15:26:09 +02:00
Raymond Myers e08100a9ba Fixed pfadd/pfcount commands emitting hll* events instead of pf* events 2014-04-16 15:26:09 +02:00
Raymond Myers 23acae6e1c Change HLL* to PF* in error messages 2014-04-16 15:26:09 +02:00
antirez e3773b0a0e Include redis.h before other stuff in hyperloglog.c. 
Otherwise fmacros.h is included later and this may break compilation on
different systems.
 2014-04-16 15:26:09 +02:00
antirez badf23f57b HyperLogLog API prefix modified from "P" to "PF". 
Using both the initials of Philippe Flajolet instead of just "P".
 2014-04-16 15:26:03 +02:00
antirez 6e71459c73 Makefile.dep updated with hyperloglog.o deps. 2014-04-16 15:25:28 +02:00
antirez 2ee2bec2d6 HyperLogLog: make API use the P prefix in honor of Philippe Flajolet. 2014-04-16 15:24:15 +02:00
antirez 1e90c0066c HLLMERGE fixed by adding a... missing loop! 2014-04-16 15:23:21 +02:00
antirez be7eef911b hll-gnuplot-graph.rb: added new filter "all". 2014-04-16 15:23:21 +02:00
antirez f2e707b679 HyperLogLog apply bias correction using a polynomial. 
Better results can be achieved by compensating for the bias of the raw
approximation just after 2.5m (when LINEARCOUNTING is no longer used) by
using a polynomial that approximates the bias at a given cardinality.

The curve used was found using this web page:

    http://www.xuru.org/rt/PR.asp

That performs polynomial regression given a set of values.
 2014-04-16 15:23:21 +02:00
antirez be7fe2b92b HLLMERGE implemented. 
Merge N HLL data structures by selecting the max value for every
M[i] register among the set of HLLs.
 2014-04-16 15:23:21 +02:00
antirez 39dd1f648d HLLCOUNT is technically a write command 
When we update the cached value, we need to propagate the command and
signal the key as modified for WATCH.
 2014-04-16 15:23:21 +02:00
antirez f07b1514bb HLLADD: propagate write when only variable name is given. 
The following form is given:

    HLLADD myhll

No element is provided in the above case so if 'myhll' var does not
exist the result is to just create an empty HLL structure, and no update
will be performed on the registers.

In this case, the DB should still be set dirty and the command
propagated.
 2014-04-16 15:23:21 +02:00
antirez 32e9996846 hll-gnuplot-graph.rb: Use |error| when filter is :max 2014-04-16 15:23:21 +02:00
antirez 9faa80ee8c Ignore txt files inside utils/hyperloglog. 
Those are generated to trace graphs using gnuplot.
 2014-04-16 15:23:21 +02:00
antirez aad5959ddc HyperLogLog: use LINEARCOUNTING up to 3m. 
The HyperLogLog original paper suggests using LINEARCOUNTING for
cardinalities < 2.5m, however for P=14 the median / max error
curves show that a value of '3' is the best pick for m = 16384.
 2014-04-16 15:23:21 +02:00
antirez d6ce659ffd hll-gnuplot-graph.rb added to plot HyperLogLog error graphs. 2014-04-16 15:23:21 +02:00
antirez c3df5965f3 HyperLogLog approximated cardinality caching. 
The more we add elements to an HyperLogLog counter, the smaller is
the probability that we actually update some register.

From this observation it is easy to see how it is possible to use
caching of a previously computed cardinality and reuse it to serve
HLLCOUNT queries as long as no register was updated in the data
structure.

This commit does exactly this by using just additional 8 bytes for the
data structure to store a 64 bit unsigned integer value cached
cardinality. When the most significant bit of the 64 bit integer is set,
it means that the value computed is no longer usable since at least a
single register was modified and we need to recompute it at the next
call of HLLCOUNT.

The value is always stored in little endian format regardless of the
actual CPU endianess.
 2014-04-16 15:23:21 +02:00
antirez 9e178afae9 String value unsharing refactored into proper function. 
All the Redis functions that need to modify the string value of a key in
a destructive way (APPEND, SETBIT, SETRANGE, ...) require to make the
object unshared (if refcount > 1) and encoded in raw format (if encoding
is not already REDIS_ENCODING_RAW).

This was cut & pasted many times in multiple places of the code. This
commit puts the small logic needed into a function called
dbUnshareStringValue().
 2014-04-16 15:22:56 +02:00
antirez 1c795db9d0 Use endian neutral hash function for HyperLogLog. 
We need to be sure that you can save a dataset in a Redis instance,
reload it in a different architecture, and continue to count in the same
HyperLogLog structure.

So 32 and 64 bit, little or bit endian, must all guarantee to output the
same hash for the same element.
 2014-04-16 15:19:40 +02:00
antirez 7f30998432 HyperLogLog internal representation modified. 
The new representation is more obvious, starting from the LSB of the
first byte and using bits going to MSB, and passing to next byte as
needed.

There was also a subtle error: first two bits were unused, everything
was carried over on the right of two bits, even if it worked because of
the code requirement of always having a byte more at the end.

During the rewrite the code was made safer trying to avoid undefined
behavior due to shifting an uint8_t for more than 8 bits.
 2014-04-16 15:19:40 +02:00
antirez 430d0ade75 Remove a few useless operations from hllCount() fast path. 2014-04-16 15:19:40 +02:00
antirez 83650a72a1 HLLCOUNT 3x faster taking fast path for default params. 2014-04-16 15:19:40 +02:00
antirez 3afd25ecc7 Use processor base types in HLL_(GET|SET)_REGISTER. 
This speedups the macros by a noticeable factor.
 2014-04-16 15:19:40 +02:00
antirez 0ffd5e4a30 HyperLogLog: use precomputed table for 2^(-M[i]). 2014-04-16 15:19:40 +02:00
antirez ca6973d3b6 hll-err.rb: speedup using pipelining. 2014-04-16 15:19:40 +02:00
antirez 2500f2a334 hll-err.rb added to test error rate of Redis HyperLogLog. 2014-04-16 15:19:40 +02:00
antirez 42592a7faa HyperLogLog algorithm fixed in two ways. 
There was an error in the computation of 2^register, and the sequence of
zeroes computed after the hashing did not included the "1".
 2014-04-16 15:19:40 +02:00
antirez 3035075b2c HLLCOUNT implemented. 2014-04-16 15:19:40 +02:00
antirez 959d0f012a HLLADD implemented. 2014-04-16 15:19:33 +02:00
antirez ac29bb2ae4 hllAdd() low level HyperLogLog "add" implemented. 2014-04-16 15:19:05 +02:00
antirez 2a91f548c4 HyperLogLog: redefine constants using "P". 2014-04-16 15:19:05 +02:00
antirez 3f159bbd44 HLL_SET_REGISTER fixed. 
There was an error in the first version of the macro.
Now the HLLSELFTEST test reports success.
 2014-04-16 15:19:05 +02:00
antirez a18e880bf1 Use REDIS_HLL_REGISTER_MAX when possible. 2014-04-16 15:19:05 +02:00
antirez 00409d8742 HLL_(SET|GET)_REGISTER types fixed. 2014-04-16 15:19:05 +02:00
antirez de3f821af6 HLLSELFTEST command implemented. 
To test the bitfield array of counters set/get macros from the Redis Tcl
suite is hard, so a specialized command that is able to test the
internals was developed.
 2014-04-16 15:18:57 +02:00
antirez 02d88fb201 HyperLogLog: initial sketch of registers access. 2014-04-16 15:17:26 +02:00
antirez ddbdd3b0df Document TTL changes in the 2.6 -> 2.8 migration doc. 2014-03-27 15:53:14 +01:00
antirez d2e59c2715 Redis 2.8.8. 2014-03-25 11:30:42 +01:00
antirez 99fc582f8c Test: do not complain when "leaks" can't run because process died. 2014-03-25 09:34:11 +01:00
antirez 3580bb485a adjustOpenFilesLimit() refactoring. 
In this commit:
* Decrement steps are semantically differentiated from the reserved FDs.
  Previously both values were 32 but the meaning was different.
* Make it clear that we save setrlimit errno.
* Don't explicitly handle wrapping of 'f', but prevent it from
  happening.
* Add comments to make the function flow more readable.

This integrates PR #1630
 2014-03-25 09:07:21 +01:00
Matt Stancliff c3510af1c0 Fix potentially incorrect errno usage 
errno may be reset by the previous call to redisLog, so capture
the original value for proper error reporting.
 2014-03-25 09:07:21 +01:00
Matt Stancliff 771f8ad0e7 Add REDIS_MIN_RESERVED_FDS define for open fds 
Also update the original REDIS_EVENTLOOP_FDSET_INCR to
include REDIS_MIN_RESERVED_FDS. REDIS_EVENTLOOP_FDSET_INCR
exists to make sure more than (maxclients+RESERVED) entries
are allocated, but we can only guarantee that if we include
the current value of REDIS_MIN_RESERVED_FDS as a minimum
for the INCR size.
 2014-03-25 09:07:21 +01:00
Matt Stancliff 3ce742d1d5 Fix infinite loop on startup if ulimit too low 
Fun fact: rlim_t is an unsigned long long on all platforms.

Continually subtracting from a rlim_t makes it get smaller
and smaller until it wraps, then you're up to 2^64-1.

This was causing an infinite loop on Redis startup if
your ulimit was extremely (almost comically) low.

The case of (f > oldlimit) would never be met in a case like:

    f = 150
    while (f > 20) f -= 128

Since f is unsigned, it can't go negative and would
take on values of:

    Iteration 1: 150 - 128 => 22
    Iteration 2:  22 - 128 => 18446744073709551510
    Iterations 3-∞: ...

To catch the wraparound, we use the previous value of f
stored in limit.rlimit_cur.  If we subtract from f and
get a larger number than the value it had previously,
we print an error and exit since we don't have enough
file descriptors to help the user at this point.

Thanks to @bs3g for the inspiration to fix this problem.
Patches existed from @bs3g at antirez#1227, but I needed to repair a few other
parts of Redis simultaneously, so I didn't get a chance to use them.
 2014-03-25 09:07:21 +01:00
Matt Stancliff b59585396b Improve error handling around setting ulimits 
The log messages about open file limits have always
been slightly opaque and confusing.  Here's an attempt to
fix their wording, detail, and meaning.  Users will have a
better understanding of how to fix very common problems
with these reworded messages.

Also, we handle a new error case when maxclients becomes less
than one, essentially rendering the server unusable.  We
now exit on startup instead of leaving the user with a server
unable to handle any connections.

This fixes antirez#356 as well.
 2014-03-25 09:07:21 +01:00
Matt Stancliff 6826af1b50 Replace magic 32 with REDIS_EVENTLOOP_FDSET_INCR 
32 was the additional number of file descriptors Redis
would reserve when managing a too-low ulimit.  The
number 32 was in too many places statically, so now
we use a macro instead that looks more appropriate.

When Redis sets up the server event loop, it uses:
    server.maxclients+REDIS_EVENTLOOP_FDSET_INCR

So, when reserving file descriptors, it makes sense to
reserve at least REDIS_EVENTLOOP_FDSET_INCR FDs instead
of only 32.  Currently, REDIS_EVENTLOOP_FDSET_INCR is
set to 128 in redis.h.

Also, I replaced the static 128 in the while f < old loop
with REDIS_EVENTLOOP_FDSET_INCR as well, which results
in no change since it was already 128.

Impact: Users now need at least maxclients+128 as
their open file limit instead of maxclients+32 to obtain
actual "maxclients" number of clients.  Redis will carve
the extra REDIS_EVENTLOOP_FDSET_INCR file descriptors it
needs out of the "maxclients" range instead of failing
to start (unless the local ulimit -n is too low to accomidate
the request).
 2014-03-25 09:07:21 +01:00
Matt Stancliff 611372fa97 Fix maxclients error handling 
Everywhere in the Redis code base, maxclients is treated
as an int with (int)maxclients or `maxclients = atoi(source)`,
so let's make maxclients an int.

This fixes a bug where someone could specify a negative maxclients
on startup and it would work (as well as set maxclients very high)
because:

    unsigned int maxclients;
    char *update = "-300";
    maxclients = atoi(update);
    if (maxclients < 1) goto fail;

But, (maxclients < 1) can only catch the case when maxclients
is exactly 0.  maxclients happily sets itself to -300, which isn't
-300, but rather 4294966996, which isn't < 1, so... everything
"worked."

maxclients config parsing checks for the case of < 1, but maxclients
CONFIG SET parsing was checking for case of < 0 (allowing
maxclients to be set to 0).  CONFIG SET parsing is now updated to
match config parsing of < 1.

It's tempting to add a MINIMUM_CLIENTS define, but... I didn't.

These changes were inspired by antirez#356, but this doesn't
fix that issue.
 2014-03-25 09:07:21 +01:00
Matt Stancliff 3bd3240660 Sentinel: remove variable causing warning 
GCC-4.9 warned about this, but clang didn't.

This commit fixes warning:
sentinel.c: In function 'sentinelReceiveHelloMessages':
sentinel.c:2156:43: warning: variable 'master' set but not used [-Wunused-but-set-variable]
     sentinelRedisInstance *ri = c->data, *master;
 2014-03-25 08:06:04 +01:00
antirez 79349affd5 Fixed undefined variable value with certain code paths. 
In sentinelFlushConfig() fd could be undefined when the following if
statement was true:

        if (rewrite_status == -1) goto werr;

This could cause random file descriptors to get closed.
 2014-03-25 08:03:11 +01:00
Matt Stancliff 80dec5e4df Sentinel: Notify user when config can't be saved 2014-03-25 08:03:07 +01:00
Jan-Erik Rediger a2ec9a9068 Small typo fixed 2014-03-25 08:02:56 +01:00
Matt Stancliff 88c6c669c3 Fix data loss when save AOF/RDB with no free space 
Previously, the (!fp) would only catch lack of free space
under OS X.  Linux waits to discover it can't write until
it actually writes contents to disk.

(fwrite() returns success even if the underlying file
has no free space to write into.  All the errors
only show up at flush/sync/close time.)

Fixes antirez/redis#1604
 2014-03-24 21:56:22 +01:00
Jan-Erik Rediger 6182dad37c Finally fix the install_server.sh script. 
Includes changes from a dozen bug reports and pull requests.
Was tested on Ubuntu, Debian and CentOS.
 2014-03-24 18:12:58 +01:00
antirez 4ebc7e3738 Sample and cache RSS in serverCron(). 
Obtaining the RSS (Resident Set Size) info is slow in Linux and OSX.
This slowed down the generation of the INFO 'memory' section.

Since the RSS does not require to be a real-time measurement, we
now sample it with server.hz frequency (10 times per second by default)
and use this value both to show the INFO rss field and to compute the
fragmentation ratio.

Practically this does not make any difference for memory profiling of
Redis but speeds up the INFO call significantly.
 2014-03-24 12:03:44 +01:00
antirez 72257f4b87 sdscatvprintf(): Try to use a static buffer. 
For small content the function now tries to use a static buffer to avoid
a malloc/free cycle that is too costly when the function is used in the
context of performance critical code path such as INFO output generation.

This change was verified to have positive effects in the execution speed
of the INFO command.
 2014-03-24 10:23:58 +01:00
antirez 7054f2a029 Cache uname() output across INFO calls. 
Uname was profiled to be a slow syscall. It produces always the same
output in the context of a single execution of Redis, so calling it at
every INFO output generation does not make too much sense.

The uname utsname structure was modified as a static variable. At the
same time a static integer was added to check if we need to call uname
the first time.
 2014-03-24 10:02:58 +01:00
antirez 7a20f09647 sdscatvprintf(): guess buflen using format length. 
sdscatvprintf() uses a loop where it tries to output the formatted
string in a buffer of the initial length, if there was not enough room,
a buffer of doubled size is tried and so forth.

The initial guess for the buffer length was very poor, an hardcoded
"16". This caused the printf to be processed multiple times without a
good reason. Given that printf functions are already not fast, the
overhead was significant.

The new heuristic is to use a buffer 4 times the length of the format
buffer, and 32 as minimal size. This appears to be a good balance for
typical uses of the function inside the Redis code base.

This change improved INFO command performances 3 times.
 2014-03-24 09:45:18 +01:00
antirez 001775f5fb Use 24 bits for the lru object field and improve resolution. 
There were 2 spare bits inside the Redis object structure that are now
used in order to enlarge 4x the range of the LRU field.

At the same time the resolution was improved from 10 to 1 second: this
still provides 194 days before the LRU counter overflows (restarting from
zero).

This is not a problem since it only causes lack of eviction precision for
objects not touched for a very long time, and the lack of precision is
only temporary.
 2014-03-21 11:19:28 +01:00
antirez c68189a19f Specify lruclock in redisServer structure via REDIS_LRU_BITS. 
The padding field was totally useless: removed.
 2014-03-21 11:17:52 +01:00
antirez ff8c81873a Set LRU parameters via REDIS_LRU_BITS define. 2014-03-21 11:17:36 +01:00
antirez e3b71a1c02 Unify stats reset for CONFIG RESETSTAT / initServer(). 
Now CONFIG RESETSTAT makes sure to reset all the fields, and in the
future it will be simpler to avoid missing new fields.
 2014-03-21 11:16:12 +01:00
antirez 0937377aa2 Sentinel: sentinelRefreshInstanceInfo() minor refactoring. 
Test sentinel.tilt condition on top and return if it is true.
This allows to remove the check for the tilt condition in the remaining
code paths of the function.
 2014-03-21 11:16:12 +01:00
antirez 686839b477 Sentinel test: 02 unit better coverage + refactoring. 2014-03-21 11:16:12 +01:00
antirez 6d0e408a27 Sentinel test: foreach_instance_id implements 'break'. 2014-03-21 11:16:12 +01:00
antirez ba2edc4191 Sentinel: instance_is_killed proc added to sentinel.tcl. 2014-03-21 11:16:12 +01:00
antirez 9c2063fb8f Sentinel: propagate down-after-ms changes to slaves and sentinels. 2014-03-21 11:16:12 +01:00
antirez ffa8f479a5 Sentinel: down-after-milliseconds is not master-specific. 
addReplySentinelRedisInstance() modified so that this field is displayed
for all the kind of instances: Sentinels, Masters, Slaves.
 2014-03-21 11:16:12 +01:00
antirez 42091a79bb Sentinel failure detection implementation improved. 
Failure detection in Sentinel is ping-pong based. It used to work by
remembering the last time a valid PONG reply was received, and checking
if the reception time was too old compared to the current current time.

PINGs were sent at a fixed interval of 1 second.

This works in a decent way, but does not scale well when we want to set
very small values of "down-after-milliseconds" (this is the node
timeout basically).

This commit reiplements the failure detection making a number of
changes. Some changes are inspired to Redis Cluster failure detection
code:

* A new last_ping_time field is added in representation of instances.
  If non zero, we have an active ping that was sent at the specified
  time. When a valid reply to ping is received, the field is zeroed
  again.
* last_ping_time is not reset when we reconnect the link or send a new
  ping, so from our point of view it represents the time we started
  waiting for the instance to reply to our pings without receiving a
  reply.
* last_ping_time is now used in order to check if the instance is
  timed out. This means that we can have a node timeout of 100
  milliseconds and yet the system will work well since the new check is
  not bound to the period used to send pings.
* Pings are now sent every second, or often if the value of
  down-after-milliseconds is less than one second. With a lower limit of
  10 HZ ping frequency.
* Link reconnection code was improved. This is used in order to try to
  reconnect the link when we are at 50% of the node timeout without a
  valid reply received yet. However the old code triggered unnecessary
  reconnections when the node timeout was very small. Now that should be
  ok.

The new code passes the tests but more testing is needed and more unit
tests stressing the failure detector, so currently this is merged only
in the unstable branch.
 2014-03-21 11:16:11 +01:00
antirez 38241c4b1e Sentinel: use CLIENT SETNAME when connecting to Redis. 
This makes debugging / monitoring of Sentinels simpler since you can
identify sentinels in CLIENT LIST output of Redis instances.
 2014-03-21 11:16:11 +01:00
Matt Stancliff 9de0755869 Fix segfault from accessing array out of bounds 
argc == 2; argv[2] == crash
 2014-03-14 22:57:10 +01:00
antirez a31a0b4336 Sentinel: be safe under crash-recovery assumptions. 
Sentinel's main safety argument is that there are no two configurations
for the same master with the same version (configuration epoch).

For this to be true Sentinels require to be authorized by a majority.
Additionally Sentinels require to do two important things:

* Never vote again for the same epoch.
* Never exchange an old vote for a fresh one.

The first prerequisite, in a crash-recovery system model, requires to
persist the master->leader_epoch on durable storage before to reply to
messages. This was not the case.

We also make sure to persist the current epoch in order to never reply
to stale votes requests from other Sentinels, after a recovery.

The configuration is persisted by making use of fsync(), this is
considered in the context of this code a good enough guarantee that
after a restart our durable state is restored, however this may not
always be the case depending on the kind of hardware and operating
system used.
 2014-03-14 22:57:06 +01:00
antirez 6b0e36ffc6 Sentinel: fake PUBLISH command to receive HELLO messages. 
Now the way HELLO messages are received is unified.
Now it is no longer needed for Sentinels to converge to the higher
configuration for a master to be able to chat via some Redis instance,
the are able to directly exchanges configurations.

Note that this commit does not include the (trivial) change needed to
send HELLO messages to Sentinel instances as well, since for an error I
committed the change in the previous commit that refactored hello
messages processing into a separated function.
 2014-03-14 11:04:54 +01:00
antirez bd48ff69b0 Sentinel: HELLO processing refactored into sentinelProcessHelloMessage(). 2014-03-14 10:56:56 +01:00
antirez 3703112671 Linenoise updated, multiline mode enabled in redis-cli. 2014-03-13 15:12:04 +01:00
zhanghailei ccd2c18c1c According to context,the size should be 16 rather than 64 2014-03-11 10:12:19 +01:00
zhanghailei daa5d2a6c9 FIXED a typo more thank should be more than 2014-03-11 10:12:13 +01:00
zhanghailei ca0720b694 refer to updateLRUClock's comment REDIS_LRU_CLOCK_MAX is 22 bits,but #define REDIS_LRU_CLOCK_MAX ((1<<21)-1) only 21 bits 2014-03-11 10:12:00 +01:00
Matt Stancliff ffe742f92a Reset op_sec_last_sample_ops when reset requested 
This value needs to be set to zero (in addition to
stat_numcommands) or else people may see
a negative operations per second count after they
run CONFIG RESETSTAT.

Fixes antirez/redis#1577
 2014-03-11 10:10:54 +01:00
antirez 6fb9e2d717 Typo in sentinel.conf, exists -> exits. 2014-03-11 10:10:30 +01:00
antirez eb9e1526e6 DEBUG ERROR implemented. 
The new "error" subcommand of the DEBUG command can reply with an user
selected error, specified as its sole argument:

    DEBUG ERROR "LOADING please wait..."

The error is generated just prefixing the command argument with a "-"
character, and replacing newlines with spaces (since error replies can't
include newlines).

The goal of the command is to help in Client libraries unit tests by
making simple to simulate a command call triggering a given error.
 2014-03-10 23:04:37 +01:00
Matt Stancliff a6970570e3 Fix return value check for anetTcpAccept 
anetTcpAccept returns ANET_ERR, not AE_ERR.

This isn't a physical error since both ANET_ERR
and AE_ERR are -1, but better to be consistent.
 2014-03-10 15:47:20 +01:00
antirez fe0ab7d234 Fixed memory leak in SORT LIMIT option argument parsing on error. 2014-03-10 15:45:29 +01:00
antirez 464fef9bf8 Redis 2.8.7. 2014-03-05 14:42:50 +01:00
antirez aab16ead92 Cast saveparams[].seconds to long for %ld format specifier. 2014-03-05 11:26:50 +01:00
antirez 86afc0706d Merge branch '2.8' of github.com:/antirez/redis into 2.8 2014-03-05 10:22:56 +01:00
antirez 55b8f6ec1c Document why we update peak memory in INFO. 2014-03-05 10:16:20 +01:00
Matt Stancliff 647a261465 Force INFO used_memory_peak to match peak memory 
used_memory_peak only updates in serverCron every server.hz,
but Redis can use more memory and a user can request memory
INFO before used_memory_peak gets updated in the next
cron run.

This patch updates used_memory_peak to the current
memory usage if the current memory usage is higher
than the recorded used_memory_peak value.

(And it only calls zmalloc_used_memory() once instead of
twice as it was doing before.)
 2014-03-05 10:16:16 +01:00
antirez 9c66cd91be Sentinel test: Makefile target added. 2014-03-05 10:16:12 +01:00
michael-grunder 6991792943 Improved bigkeys with progress, pipelining and summary 
This commit reworks the redis-cli --bigkeys command to provide more
information about our progress as well as output summary information
when we're done.

 - We now show an approximate percentage completion as we go
 - Hiredis pipelining is used for TYPE and SIZE retreival
 - A summary of keyspace distribution and overall breakout at the end
 2014-03-05 10:16:06 +01:00
antirez 7d65b7199a BITPOS fuzzy testing. 2014-03-05 10:16:02 +01:00
antirez c19cfde65d Basic BITPOS tests. 2014-03-05 10:15:55 +01:00
antirez 55f4b20f31 Sentinel test: set less time sensitive defaults. 
This commit sets the failover timeout to 30 seconds instead of the 180
seconds default, and allows to reconfigure multiple slaves at the same
time.

This makes tests less sensible to timing, with the result that there are
less false positives due to normal behaviors that require time to
succeed or to be retried.

However the long term solution is probably some way in order to detect
when a test failed because of timing issues (for example split brain
during leader election) and retry it.
 2014-03-05 10:15:32 +01:00
antirez 1606978a0b Sentinel: more aggressive failover start desynchronization. 
Sentinel needs to avoid split brain conditions due to multiple sentinels
trying to get voted at the exact same time.

So far some desynchronization was provided by fluctuating server.hz,
that is the frequency of the timer function call. However the
desynchonization provided in this way was not enough when using many
Sentinel instances, especially when a large quorum value is used in
order to force a greater degree of agreement (more than N/2+1).

It was verified that it was likely to trigger a split brain
condition, forcing the system to try again after a timeout.
Usually the system will succeed after a few retries, but this is not
optimal.

This commit desynchronizes instances in a more effective way to make it
likely that the first attempt will be successful.
 2014-03-05 10:15:32 +01:00
antirez 6200191943 CONFIG REWRITE should be logged at WARNING level. 2014-03-05 10:15:32 +01:00
antirez 2af2173a60 Sentinel test: debugging console improved. 2014-03-05 10:15:32 +01:00
antirez 1dc1e31c2b Sentinel test: initial debugging console. 2014-03-05 10:15:32 +01:00
antirez b132121a71 Sentinel test: be more patient in create_redis_master_slave_cluster. 2014-03-05 10:15:32 +01:00
antirez d1d706c923 Sentiel test: add test start time in output. 2014-03-05 10:15:32 +01:00
antirez c99cd2fd40 Sentinel test: use 1000 as retry in initial 00 unit test. 2014-03-05 10:15:32 +01:00
antirez f33d4a6c24 Sentinel test: initial tests in 03 unit. 2014-03-05 10:15:32 +01:00
antirez c2d99d49b5 Sentinel test: foreach_instance_id now supports 'continue'. 2014-03-05 10:15:32 +01:00
antirez e40f8378bd Sentienl test: fixed typo in unit 03 top comment. 2014-03-05 10:15:32 +01:00
antirez a0335e18ba Sentinel test: Makefile target added. 2014-02-28 16:00:17 +01:00
antirez 950cb76e7b BITPOS fuzzy testing. 2014-02-27 15:58:34 +01:00
antirez 42e3630d77 Basic BITPOS tests. 2014-02-27 15:58:30 +01:00
antirez 1892b56224 warnigns -> warnings in redisBitpos(). 2014-02-27 15:58:27 +01:00
antirez 2c8036f7b2 More consistent BITPOS behavior with bit=0 and ranges. 
With the new behavior it is possible to specify just the start in the
range (the end will be assumed to be the first byte), or it is possible
to specify both start and end.

This is useful to change the behavior of the command when looking for
zeros inside a string.

1) If the user specifies both start and end, and no 0 is found inside
   the range, the command returns -1.

2) If instead no range is specified, or just the start is given, even
   if in the actual string no 0 bit is found, the command returns the
   first bit on the right after the end of the string.

So for example if the string stored at key foo is "\xff\xff":

    BITPOS foo (returns 16)
    BITPOS foo 0 -1 (returns -1)
    BITPOS foo 0 (returns 16)

The idea is that when no end is given the user is just looking for the
first bit that is zero and can be set to 1 with SETBIT, as it is
"available". Instead when a specific range is given, we just look for a
zero within the boundaries of the range.
 2014-02-27 15:58:21 +01:00
antirez 3294f74fef Initial implementation of BITPOS. 
It appears to work but more stress testing, and both unit tests and
fuzzy testing, is needed in order to ensure the implementation is sane.
 2014-02-27 15:58:14 +01:00
antirez 82d2e295b8 Added two more BITCOUNT tests stressing misaligned access. 2014-02-27 15:57:56 +01:00
antirez a3eb3f9c3b BITCOUNT fuzzy test with random start/end added. 
It was verified in practice that this test is able to stress much more
the implementation by introducing errors that were only trivially to
detect with different offsets but impossible to detect starting always
at zero and counting bits the full length of the string.
 2014-02-27 15:57:53 +01:00
antirez 30a92b6c76 Fix misaligned word access in redisPopcount(). 2014-02-27 15:57:49 +01:00
antirez c1cc28f230 warnigns -> warnings in redisBitpos(). 2014-02-27 15:55:56 +01:00
antirez d79f9ebdb5 More consistent BITPOS behavior with bit=0 and ranges. 
With the new behavior it is possible to specify just the start in the
range (the end will be assumed to be the first byte), or it is possible
to specify both start and end.

This is useful to change the behavior of the command when looking for
zeros inside a string.

1) If the user specifies both start and end, and no 0 is found inside
   the range, the command returns -1.

2) If instead no range is specified, or just the start is given, even
   if in the actual string no 0 bit is found, the command returns the
   first bit on the right after the end of the string.

So for example if the string stored at key foo is "\xff\xff":

    BITPOS foo (returns 16)
    BITPOS foo 0 -1 (returns -1)
    BITPOS foo 0 (returns 16)

The idea is that when no end is given the user is just looking for the
first bit that is zero and can be set to 1 with SETBIT, as it is
"available". Instead when a specific range is given, we just look for a
zero within the boundaries of the range.
 2014-02-27 15:55:47 +01:00
antirez 25e2791ec3 Initial implementation of BITPOS. 
It appears to work but more stress testing, and both unit tests and
fuzzy testing, is needed in order to ensure the implementation is sane.
 2014-02-27 15:54:31 +01:00
antirez c955b47d62 Added two more BITCOUNT tests stressing misaligned access. 2014-02-27 15:53:40 +01:00
antirez bfe4c560e6 BITCOUNT fuzzy test with random start/end added. 
It was verified in practice that this test is able to stress much more
the implementation by introducing errors that were only trivially to
detect with different offsets but impossible to detect starting always
at zero and counting bits the full length of the string.
 2014-02-27 15:53:40 +01:00
antirez 580ed0768f Fix misaligned word access in redisPopcount(). 2014-02-27 15:53:40 +01:00
antirez d90e3e3a70 Sentinel test: add stub for unit 04. 2014-02-25 15:55:37 +01:00
antirez 89d95ec2e7 Sentinel test: added TODO items in 02 unit. 2014-02-25 15:55:37 +01:00
antirez 65194d11c3 redis-cli: also remove useless uint8_t. 2014-02-25 15:09:15 +01:00
antirez 16c2189d4e redis-cli: don't use uint64_t where actually not needed. 
The computation is just something to take the CPU busy, no need to use a
specific type. Since stdint.h was not included this prevented
compilation on certain systems.
 2014-02-25 15:09:15 +01:00
antirez 1b2bcd4210 redis-cli: check argument existence for --pattern. 2014-02-25 15:09:15 +01:00
antirez 6ec3dc2c42 redis-cli: --intrinsic-latency run mode added. 2014-02-25 15:09:15 +01:00
antirez 02ae6cb01a redis-cli: added comments to split program in parts. 2014-02-25 15:09:15 +01:00
antirez fe05112029 Sentinel test: check role at end of unit 01. 2014-02-25 15:08:41 +01:00
michael-grunder cf76fdd04f Update --bigkeys to use SCAN 
This commit changes the findBigKeys() function in redis-cli.c to use the new
SCAN command for iterating the keyspace, rather than RANDOMKEY.  Because we
can know when we're done using SCAN, it will exit after exhausting the keyspace.
 2014-02-25 15:08:38 +01:00
antirez 9e34e296cf Sentinel test: kill masters instead of using DEBUG SLEEP in all tests. 2014-02-25 15:08:34 +01:00
antirez dfacfafc2a Sentinel test: restart instances left killed by previous unit. 
An unit can abort in the middle for an error. The next unit should not
assume that the instances are in a clean state, and must restart what
was left killed.
 2014-02-25 10:30:35 +01:00
antirez 2fab983c69 Sentinel test: jump to next unit on test failure. 
Sentinel tests are designed to be dependent on the previous tests in the
same unit, so usually we can't continue with the next test in the same
unit if a previous test failed.
 2014-02-25 10:30:35 +01:00
antirez a6d560db64 Sentinel test: test majority crashing Sentinels. 
The test was previously performed by removing the master from the
Sentinel monitored masters. The test with the Sentinels crashed is
more similar to real-world partitions / failures.
 2014-02-25 10:30:35 +01:00
antirez 3f16fee39a Sentinel test: restart_instance should refresh pid attrib. 
Also kill_instance was modified to warn when a test will try to kill the
same instance multiple times for error.
 2014-02-25 10:30:35 +01:00
antirez e8e2ac072c Sentinel test: more stuff mored from 00-base to init. 
The area a number of mandatory tests to craete a stable setup for
testing that is not too sensitive to timing issues. All those tests
moved to includes/init-tests, and marked as (init).
 2014-02-25 10:30:35 +01:00
antirez 85fa77e02e Sentinel: log quorum with +monitor event. 2014-02-25 10:30:35 +01:00
antirez 2d48d7c2fc Sentinel test: removed useless code to set SDOWN timeout. 
The new common initialization code used to start a new unit already set
the timeout to 2000 milliseconds.
 2014-02-25 10:30:35 +01:00
antirez 6e6106797e Sentinel: generate +monitor events at startup. 2014-02-25 10:30:35 +01:00
antirez 96162c0c58 Sentinel: log +monitor and +set events. 
Now that we have a runtime configuration system, it is very important to
be able to log how the Sentinel configuration changes over time because
of API calls.
 2014-02-25 10:30:35 +01:00
antirez 39eacde136 Sentinel: added missing exit(1) after checking for config file. 2014-02-25 10:30:35 +01:00
antirez 28aa15a6ea Sentinel test: tmp dir and gitignore added. 2014-02-25 10:30:35 +01:00
antirez 3917224f7c Sentinel test: minor fixes to --pause-on-error. 2014-02-25 10:30:35 +01:00
antirez 001ab29915 Sentinel test: --pause-on-error option added. 
Pause the test with running instances available for state inspection on
error.
 2014-02-25 10:30:35 +01:00
antirez d31a394148 Sentinel test: added empty units to fill later. 2014-02-25 10:30:35 +01:00
antirez d83ab8f9d4 Sentinel: IDONTKNOW error removed. 
This error was conceived for the older version of Sentinel that worked
via master redirection and that was not able to get configuration
updates from other Sentinels via the Pub/Sub channel of masters or
slaves.

This reply does not make sense today, every Sentinel should reply with
the best information it has currently. The error will make even more
sense in the future since the plan is to allow Sentinels to update the
configuration of other Sentinels via gossip with a direct chat without
the prerequisite that they have at least a monitored instance in common.
 2014-02-25 10:30:35 +01:00
antirez 1dfab3d107 Sentinel test: framework improved and conf-update unit added. 
It is now possible to kill and restart sentinel or redis instances for
more real-world testing.

The 01 unit tests the capability of Sentinel to update the configuration
of Sentinels rejoining the cluster, however the test is pretty trivial
and more tests should be added.
 2014-02-25 10:30:35 +01:00
antirez aee165e084 Sentinel test: move init tests as includes. 
Most units will start with these two basic tests to create an
environment where the real tests are ran.
 2014-02-25 10:30:35 +01:00
antirez 73b290cc5b Sentinel test: ability to run just a subset of test files. 2014-02-25 10:30:35 +01:00
Matt Stancliff 3ced78dc6c Auto-enter slaveMode when SYNC from redis-cli 
If someone asks for SYNC or PSYNC from redis-cli,
automatically enter slaveMode (as if they ran
redis-cli --slave) and continue printing the replication
stream until either they Ctrl-C or the master gets disconnected.
 2014-02-25 10:30:35 +01:00
antirez 6441a41f94 Sentinel: report instances role switch events. 
This is useful mostly for debugging of issues.
 2014-02-20 12:28:03 +01:00
antirez 05dde71919 Sentinel test: some reliability fixes to 00-base tests. 2014-02-20 12:28:03 +01:00
antirez c499d82f28 Sentinel test: check that role matches at end of 00-base. 2014-02-20 12:28:03 +01:00
antirez c01aff1234 Sentinel test: ODOWN and agreement. 2014-02-20 12:28:03 +01:00
antirez 56cb037a86 Sentinel test: check reconfig of slaves and old master. 2014-02-20 12:28:03 +01:00
antirez d4edf63bc4 Sentinel test: basic failover tested. Framework improvements. 2014-02-20 12:28:02 +01:00
antirez 722d4f0c41 Sentinel test: basic tests for MONITOR and auto-discovery. 2014-02-20 12:28:02 +01:00
antirez b26660e130 Sentinel test: info fields, master-slave setup, fixes. 2014-02-20 12:28:02 +01:00
antirez ac5fd5a519 Prefix test file names with numbers to force exec order. 2014-02-20 12:28:02 +01:00
antirez 890559ea1c Sentinel test: provide basic commands to access instances. 2014-02-20 12:28:02 +01:00
antirez 905c55d5ae Sentinel: SENTINEL_SLAVE_RECONF_RETRY_PERIOD -> RECONF_TIMEOUT 
Rename define to match the new meaning.
 2014-02-18 10:30:29 +01:00
antirez 1b345ec3f0 Sentinel: fix slave promotion timeout. 
If we can't reconfigure a slave in time during failover, go forward as
anyway the slave will be fixed by Sentinels in the future, once they
detect it is misconfigured.

Otherwise a failover in progress may never terminate if for some reason
the slave is uncapable to sync with the master while at the same time
it is not disconnected.
 2014-02-18 10:30:28 +01:00
antirez 351e790032 Sentinel: initial testing framework. 
Nothing tested at all so far... Just the infrastructure spawning N
Sentinels and N Redis instances that the test will use again and again.
 2014-02-17 17:38:58 +01:00
antirez f71c7de88e Test: colorstr moved to util.tcl. 2014-02-17 17:38:54 +01:00
antirez 0886c98fef Test: code to test server availability refactored. 
Some inline test moved into server_is_up procedure.
Also find_available_port was moved into util since it is going
to be used for the Sentinel test as well.
 2014-02-17 12:29:54 +01:00
antirez 4237f14a8a Get absoulte config file path before processig 'dir'. 
The code tried to obtain the configuration file absolute path after
processing the configuration file. However if config file was a relative
path and a "dir" statement was processed reading the config, the absolute
path obtained was wrong.

With this fix the absolute path is obtained before processing the
configuration while the server is still in the original directory where
it was executed.
 2014-02-17 12:14:19 +01:00
antirez 5efee4f0bb Sentinel: better specify startup errors due to config file. 
Now it logs the file name if it is not accessible. Also there is a
different error for the missing config file case, and for the non
writable file case.
 2014-02-17 12:10:12 +01:00
antirez 1d4d9e7b14 Redis 2.8.6. 2014-02-13 15:43:40 +01:00
antirez 85492dcfef Update cached time in rdbLoad() callback. 
server.unixtime and server.mstime are cached less precise timestamps
that we use every time we don't need an accurate time representation and
a syscall would be too slow for the number of calls we require.

Such an example is the initialization and update process of the last
interaction time with the client, that is used for timeouts.

However rdbLoad() can take some time to load the DB, but at the same
time it did not updated the time during DB loading. This resulted in the
bug described in issue #1535, where in the replication process the slave
loads the DB, creates the redisClient representation of its master, but
the timestamp is so old that the master, under certain conditions, is
sensed as already "timed out".

Thanks to @yoav-steinberg and Redis Labs Inc for the bug report and
analysis.
 2014-02-13 15:13:35 +01:00
antirez eb10bb31ac Log when CONFIG REWRITE goes bad. 2014-02-13 14:33:50 +01:00
antirez ebdb37cea7 Test: regression for issue #1549. 
It was verified that reverting the commit that fixes the bug, the test
no longer passes.
 2014-02-13 12:27:08 +01:00
antirez 9b73a274ac Fix script cache bug in the scripting engine. 
This commit fixes a serious Lua scripting replication issue, described
by Github issue #1549. The root cause of the problem is that scripts
were put inside the script cache, assuming that slaves and AOF already
contained it, even if the scripts sometimes produced no changes in the
data set, and were not actaully propagated to AOF/slaves.

Example:

    eval "if tonumber(KEYS[1]) > 0 then redis.call('incr', 'x') end" 1 0

Then:

    evalsha <sha1 step 1 script> 1 0

At this step sha1 of the script is added to the replication script cache
(the script is marked as known to the slaves) and EVALSHA command is
transformed to EVAL. However it is not dirty (there is no changes to db),
so it is not propagated to the slaves. Then the script is called again:

    evalsha <sha1 step 1 script> 1 1

At this step master checks that the script already exists in the
replication script cache and doesn't transform it to EVAL command. It is
dirty and propagated to the slaves, but they fail to evaluate the script
as they don't have it in the script cache.

The fix is trivial and just uses the new API to force the propagation of
the executed command regardless of the dirty state of the data set.

Thank you to @minus-infinity on Github for finding the issue,
understanding the root cause, and fixing it.
 2014-02-13 12:16:34 +01:00
antirez 96973a7c33 AOF write error: retry with a frequency of 1 hz. 2014-02-12 16:57:17 +01:00
antirez fadbbdd3f4 AOF: don't abort on write errors unless fsync is 'always'. 
A system similar to the RDB write error handling is used, in which when
we can't write to the AOF file, writes are no longer accepted until we
are able to write again.

For fsync == always we still abort on errors since there is currently no
easy way to avoid replying with success to the user otherwise, and this
would violate the contract with the user of only acknowledging data
already secured on disk.
 2014-02-12 16:57:13 +01:00
antirez 688d32e16b Don't count time to feed MONITORs in SLOWLOG. 2014-02-07 18:29:26 +01:00
antirez 3e4968339b Sentinel: allow SHUTDOWN command in Sentinel mode. 2014-02-07 11:22:30 +01:00
antirez 301a0cfc69 Check for EAGAIN in sendBulkToSlave(). 
Sometime an osx master with a Linux server over a slow link caused
a strange error where osx called the writable function for
the socket but actually apparently there was no room in the socket
buffer to accept the write: write(2) call returned an EAGAIN error,
that was not checked, so we considered write(2) == 0 always as a connection
reset, which was unfortunate since the bulk transfer has to start again.

Also more errors are logged with the WARNING level in the same code path
now.
 2014-02-05 16:41:04 +01:00
antirez 4e809a9a19 Redis 2.8.5. 2014-02-04 11:17:21 +01:00
antirez ddcf160309 Move mstime_t define outside sentinel.c. 
The define is now used in other parts of Redis 2.8 tree instead of long
long.

A nice side effect is that now 2.8 and unstable sentinel.c files are
identical as it should be.
 2014-02-03 16:34:46 +01:00
antirez c5bc592650 Scripting: expire keys in scripts only at first access. 
Keys expiring in the middle of the execution of Lua scripts are to
create inconsistencies in masters and / or AOF files. See the following
example:

    if redis.call("exists",KEYS[1]) == 1
    then
        redis.call("incr","mycounter")
    end

    if redis.call("exists",KEYS[1]) == 1
    then
        return redis.call("incr","mycounter")
    end

The script executes two times the same *if key exists then incrementcounter*
logic. However the two executions will work differently in the master and
the slaves, provided some unlucky timing happens.

In the master the first time the key may still exist, while the second time
the key may no longer exist. This will result in the key incremented just one
time. However as a side effect the master will generate a synthetic
`DEL` command in the replication channel in order to force the slaves to
expire the key (given that key expiration is master-driven).

When the same script will run in the slave, the key will no longer be
there, so the script will not increment the key.

The key idea used to implement the expire-at-first-lookup semantics was
provided by Marc Gravell.
 2014-02-03 16:29:25 +01:00
antirez 5201ca0ca1 Allow CONFIG and SHUTDOWN while in stale-slave state. 2014-02-03 15:51:07 +01:00
antirez 3da5cbe5bb Scripting: use mstime() and mstime_t for lua_time_start. 
server.lua_time_start is expressed in milliseconds. Use mstime_t instead
of long long, and populate it with mstime() instead of ustime()/1000.

Functionally identical but more natural.
 2014-02-03 15:46:47 +01:00
PatrickJS 0be31e2d22 update copyright year 2014-02-03 11:19:25 +01:00
antirez 66304fb122 Test: fixed osx msg passing issue in testing framework. 
The Redis test uses a server-clients model in order to parallelize the
execution of different tests. However in recent versions of osx not
setting the channel to a binary encoding caused issues even if AFAIK no
binary data is really sent via this channel. However now the channels
are deliberately set to a binary encoding and this solves the issue.

The exact issue was the test not terminating and giving the impression
of running forever, since test clients or servers were unable to
exchange the messages to continue.
 2014-01-31 16:25:13 +01:00
antirez 1406112ec4 Redis.conf comment about tcp-backlog option improved. 2014-01-31 15:03:25 +01:00
antirez 917b851491 Option "backlog" renamed "tcp-backlog". 
This is especially important since we already have a concept of backlog
(the replication backlog).
 2014-01-31 15:03:22 +01:00
Nenad Merdanovic 8dda9dbef0 Add support for listen(2) backlog definition 
In high RPS environments, the default listen backlog is not sufficient, so
giving users the power to configure it is the right approach, especially
since it requires only minor modifications to the code.
 2014-01-31 15:03:19 +01:00
antirez f0652c37a5 Sentinel: check arity for SENTINEL MASTER command. 
This fixes issue #1530.
 2014-01-31 10:14:07 +01:00
antirez a2c9d38a9f SENTINEL SET master quorum implemented. 2014-01-28 11:23:51 +01:00
antirez 89a731b3aa Fixed inverted if condition in MISCONF error code path. 2014-01-28 10:10:56 +01:00
antirez ecfefde760 Don't log MONITOR clients as disconnecting slaves. 2014-01-25 11:54:02 +01:00
antirez 43503ae5d6 redis-cli --help output improved with --scan and periods. 2014-01-22 12:08:08 +01:00
antirez b9c84d4aef redis-cli: support for --scan option. 2014-01-22 12:08:04 +01:00
antirez 248e916550 Use fflush() before fsync() in rio.c. 
Incremental flushing in rio.c is only used to avoid huge kernel buffers
synched to slow disks creating big latency spikes, so this fix has no
durability implications, however it is certainly more correct to make
sure that the FILE buffers are flushed to the kernel before calling
fsync on the file descriptor.

Thanks to Li Shao Kai for reporting this issue in the Redis mailing
list.
 2014-01-22 09:56:35 +01:00
antirez d401022d14 Fix typo in aofRewriteBufferAppend() comment. 2014-01-14 15:38:26 +01:00
antirez 4ea91aadfb Set REDIS_AOF_REWRITE_MIN_SIZE to 64mb. 
64mb is the default value in redis.conf. For some reason instead the
hard-coded default was 1mb that is too small.
 2014-01-14 11:28:18 +01:00
antirez 71f7eb2590 Redis 2.8.4. 2014-01-13 17:09:58 +01:00
antirez df00b6ac05 SENTINEL SET: error on bad option name + flush config on error. 2014-01-13 16:39:56 +01:00
antirez b8db8a0c48 SENTINEL SET implemented. 
The new command allows to change master-specific configurations
at runtime. All the settable parameters can be retrivied via the
SENTINEL MASTER command, so there is no equivalent "GET" command.
 2014-01-13 16:39:51 +01:00
antirez 8ed0b49525 Sentinel: fix wrong arity error message. 2014-01-13 16:39:47 +01:00
antirez 560e548dc6 Sentinel: SENTINEL REMOVE command added. 
The command totally removes a monitored master.
 2014-01-13 16:39:44 +01:00
antirez 0ca750d94a Sentinel: releaseSentinelRedisInstance() top comment fixed. 
The claim about unlinking the instance from the connected hash tables
was the opposite of the reality. Also the current actual behavior is
safer in most cases, so it is better to manually unlink when needed.
 2014-01-13 16:39:40 +01:00
antirez 1f9580c125 Sentinel: flush config on disk when new master is added. 2014-01-13 16:39:36 +01:00
antirez bb207007ca anetResolveIP() prototype added to anet.h. 2014-01-13 16:39:31 +01:00
antirez b7dc3204a6 Sentinel: SENTINEL MONITOR command implemented. 
It allows to add new masters to monitor at runtime.
 2014-01-13 16:39:27 +01:00
antirez 2e0ba7bb5b anetResolveIP() added to anet.c. 
The new function is used when we want to normalize an IP address without
performing a DNS lookup if the string to resolve is not a valid IP.

This is useful every time only IPs are valid inputs or when we want to
skip DNS resolution that is slow during runtime operations if we are
required to block.
 2014-01-13 16:39:24 +01:00
antirez 52cf0975c7 Sentinel: added SENTINEL MASTER <name> command. 
With SENTINEL MASTERS it was already possible to list all the configured
masters, but not a specific one.
 2014-01-13 16:39:18 +01:00
antirez 6bcc370c9d Add all the configurable fields to addReplySentinelRedisInstance(). 
Note: the auth password with the master is voluntarily not exposed.
 2014-01-13 16:39:15 +01:00
antirez ea2bffa030 Trip comment to 80 cols in SentinelCommand(). 2014-01-13 16:39:11 +01:00
antirez bca65866ff Test: regression for issues #1483. 2014-01-09 11:19:11 +01:00
antirez e4c51759f5 Fix RESTORE ttl handling in 32 bit archs. 
long was used instead of long long in order to handle a 64 bit
resolution millisecond timestamp.

This fixes issue #1483.
 2014-01-09 11:12:53 +01:00
antirez 51f71e2d8e Fix keyspace events flags-to-string conversion. 
Fixes issue #1491 on Github.
 2014-01-08 17:18:00 +01:00
antirez 05d219a6fa Test: stress events flags to/from string conversion. 2014-01-08 17:16:09 +01:00
antirez 2a1a31ca9d Don't send REPLCONF ACK to old masters. 
Masters not understanding REPLCONF ACK will reply with errors to our
requests causing a number of possible issues.

This commit detects a global replication offest set to -1 at the end of
the replication, and marks the client representing the master with the
REDIS_PRE_PSYNC flag.

Note that this flag was called REDIS_PRE_PSYNC_SLAVE but now it is just
REDIS_PRE_PSYNC as it is used for both slaves and masters starting with
this commit.

This commit fixes issue #1488.
 2014-01-08 14:27:49 +01:00
antirez 418d3d358a Clarify a comment in slaveTryPartialResynchronization(). 2014-01-08 14:11:02 +01:00
antirez 0a1a236e3e Log disconnection with slave only when ip:port is available. 2013-12-25 18:41:10 +01:00
antirez 27d06111db anetPeerToString / SockName: port can be NULL on errors too. 2013-12-25 18:39:49 +01:00
antirez 5b7c16137d anetTcpGenericConnect() bug introduced in 9d19977 fixed. 
Durign a refactoring I mispelled _port for port.
This is one of the reasons I never used _varname myself.
 2013-12-25 18:38:33 +01:00
antirez d07d4a876c Remove useless goto from anetTcpGenericConnect(). 2013-12-25 18:24:04 +01:00
antirez 9d1997706c anetTcpGenericConnect() code improved + 1 bug fix. 
Now the socket is closed if anetNonBlock() fails, and in general the
code structure makes it harder to introduce this kind of bugs in the
future.

Reference: pull request #1059.
 2013-12-25 18:16:46 +01:00
antirez 4ad219adc8 Fix CONFIG REWRITE handling of unknown options. 
There were two problems with the implementation.

1) "save" was not correctly processed when no save point was configured,
   as reported in issue #1416.
2) The way the code checked if an option existed in the "processed"
   dictionary was wrong, as we add the element with as a key associated
   with a NULL value, so dictFetchValue() can't be used to check for
   existance, but dictFind() must be used, that returns NULL only if the
   entry does not exist at all.
 2013-12-23 12:50:52 +01:00
antirez c6db326d1d Configuring port to 0 disables IP socket as specified. 
This was no longer the case with 2.8 becuase of a bug introduced with
the IPv6 support. Now it is fixed.

This fixes issue #1287 and #1477.
 2013-12-23 11:34:15 +01:00
antirez 4456ee1173 Make new masters inherit replication offsets. 
Currently replication offsets could be used into a limited way in order
to understand, out of a set of slaves, what is the one with the most
updated data. For example this comparison is possible of N slaves
were replicating all with the same master.

However the replication offset was not transferred from master to slaves
(that are later promoted as masters) in any way, so for instance if
there were three instances A, B, C, with A master and B and C
replication from A, the following could happen:

C disconnects from A.
B is turned into master.
A is switched to master of B.
B receives some write.

In this context there was no way to compare the offset of A and C,
because B would use its own local master replication offset as
replication offset to initialize the replication with A.

With this commit what happens is that when B is turned into master it
inherits the replication offset from A, making A and C comparable.
In the above case assuming no inconsistencies are created during the
disconnection and failover process, A will show to have a replication
offset greater than C.

Note that this does not mean offsets are always comparable to understand
what is, in a set of instances, since in more complex examples the
replica with the higher replication offset could be partitioned away
when picking the instance to elect as new master. However this in
general improves the ability of a system to try to pick a good replica
to promote to master.
 2013-12-22 11:54:10 +01:00
antirez 5fa937d956 Slave disconnection is an event worth logging. 2013-12-22 10:16:02 +01:00
antirez 9ba9d78ada Log when a slave lose the connection with its master. 2013-12-21 00:24:29 +01:00
antirez d33f3689d7 Clarify include directive behavior in example redis.conf. 2013-12-19 16:03:52 +01:00
antirez fb8a480f54 CONFIG REWRITE: no special handling or include and rename-command. 
CONFIG REWRITE is now wiser and does not touch what it does not
understand inside redis.conf.
 2013-12-19 16:03:49 +01:00
Yubao Liu d62bbfda16 CONFIG REWRITE: don't throw some options on config rewrite 
Those options will be thrown without this patch:
  include, rename-command, min-slaves-to-write, min-slaves-max-lag,
appendfilename.
 2013-12-19 16:03:37 +01:00
antirez 090bcc946f CONFIG REWRITE: old development comments removed. 2013-12-19 16:02:50 +01:00
antirez ddd529bc3b CONFIG REWRITE: don't wipe unknown options. 
With this commit options not explicitly rewritten by CONFIG REWRITE are
not touched at all. These include new options that may not have support
for REWRITE, and other special cases like rename-command and include.
 2013-12-19 16:02:46 +01:00
antirez 673c42bb4e Example redis.conf formatted to better show appendfilename option. 2013-12-19 10:19:13 +01:00
antirez 33f6f35f34 Makefile.dep updated. 2013-12-13 13:14:47 +01:00
antirez 8eb1cb3b52 SDIFF iterator misuse bug regression test added. 
See commit c00453d for more info about the bug.
 2013-12-13 11:37:35 +01:00
antirez 993e0ede76 SDIFF iterator misuse fixed in diff algorithm #1. 
The bug could be easily triggered by:

    SADD foo a b c 1 2 3 4 5 6
    SDIFF foo foo

When the key was the same in two sets, an unsafe iterator was used to
check existence of elements in the same set we were iterating.
Usually this would just result into a wrong output, however with the
dict.c API misuse protection we have in place, the result was actually
an assertion failed that was triggered by the CI test, while creating
random datasets for the "MASTER and SLAVE consistency" test.
 2013-12-13 11:29:59 +01:00
antirez 2507e366b2 Sentinel: dead code removed. 2013-12-13 11:01:20 +01:00
antirez f9e9448a5b Makefile: remove odd syntax not compatible with some make versions. 
See issue #1448.
 2013-12-12 15:19:29 +01:00
Yubao Liu 45c60a903a fix typo in redis.conf and sentinel.conf 2013-12-12 11:32:05 +01:00
codeeply 5e75e681fe comment mistake fixed 2013-12-12 11:25:48 +01:00
antirez 8995634014 Redis 2.8.3. 2013-12-11 15:31:57 +01:00
antirez 9a8ae5a553 Replication: publish the slave_repl_offset when disconnected from master. 
When a slave was disconnected from its master the replication offset was
reported as -1. Now it is reported as the replication offset of the
previous master, so that failover can be performed using this value in
order to try to select a slave with more processed data from a set of
slaves of the old master.
 2013-12-11 15:24:50 +01:00
Yossi Gottlieb 0ff078d8d0 Return proper error on requests with an unbalanced number of quotes. 2013-12-11 13:21:55 +01:00
Yossi Gottlieb 25ba2e9607 Fix wrong repldboff type which causes dropped replication in rare cases. 2013-12-11 11:37:58 +01:00
antirez 563d6b3f98 Slaves heartbeats during sync improved. 
The previous fix for false positive timeout detected by master was not
complete. There is another blocking stage while loading data for the
first synchronization with the master, that is, flushing away the
current data from the DB memory.

This commit uses the newly introduced dict.c callback in order to make
some incremental work (to send "\n" heartbeats to the master) while
flushing the old data from memory.

It is hard to write a regression test for this issue unfortunately. More
support for debugging in the Redis core would be needed in terms of
functionalities to simulate a slow DB loading / deletion.
 2013-12-10 18:42:22 +01:00
antirez b6610a569d dict.c: added optional callback to dictEmpty(). 
Redis hash table implementation has many non-blocking features like
incremental rehashing, however while deleting a large hash table there
was no way to have a callback called to do some incremental work.

This commit adds this support, as an optiona callback argument to
dictEmpty() that is currently called at a fixed interval (one time every
65k deletions).
 2013-12-10 18:18:36 +01:00
antirez 26cf5c8ac6 Log empty DB + Loading data into two separated messages. 2013-12-10 17:51:16 +01:00
antirez 303cc97ff9 Don't send more than 1 newline/sec while loading RDB. 2013-12-10 17:50:26 +01:00
antirez 75bf5a4a4a Slaves heartbeat while loading RDB files. 
Starting with Redis 2.8 masters are able to detect timed out slaves,
while before 2.8 only slaves were able to detect a timed out master.

Now that timeout detection is bi-directional the following problem
happens as described "in the field" by issue #1449:

1) Master and slave setup with big dataset.
2) Slave performs the first synchronization, or a full sync
   after a failed partial resync.
3) Master sends the RDB payload to the slave.
4) Slave loads this payload.
5) Master detects the slave as timed out since does not receive back the
   REPLCONF ACK acknowledges.

Here the problem is that the master has no way to know how much the
slave will take to load the RDB file in memory. The obvious solution is
to use a greater replication timeout setting, but this is a shame since
for the 0.1% of operation time we are forced to use a timeout that is
not what is suited for 99.9% of operation time.

This commit tries to fix this problem with a solution that is a bit of
an hack, but that modifies little of the replication internals, in order
to be back ported to 2.8 safely.

During the RDB loading time, we send the master newlines to avoid
being sensed as timed out. This is the same that the master already does
while saving the RDB file to still signal its presence to the slave.

The single newline is used because:

1) It can't desync the protocol, as it is only transmitted all or
nothing.
2) It can be safely sent while we don't have a client structure for the
master or in similar situations just with write(2).
 2013-12-10 15:40:36 +01:00
antirez 8d0083ba25 Handle inline requested terminated with just \n. 2013-12-10 15:40:33 +01:00
antirez fba0b23e72 Sentinel: fix reported role info sampling. 
The way the role change was recoded was not sane and too much
convoluted, causing the role information to be not always updated.

This commit fixes issue #1445.
 2013-12-06 12:49:27 +01:00
antirez dceaca1f69 Sentinel: fix reported role fields when master is reset. 
When there is a master address switch, the reported role must be set to
master so that we have a chance to re-sample the INFO output to check if
the new address is reporting the right role.

Otherwise if the role was wrong, it will be sensed as wrong even after
the address switch, and for enough time according to the role change
time, for Sentinel consider the master SDOWN.

This fixes isue #1446, that describes the effects of this bug in
practice.
 2013-12-06 12:49:23 +01:00
antirez bf307cfb4d Fixed typo in redis.conf. 2013-12-06 10:48:40 +01:00
Anurag Ramdasan 4f9d30b33b Grammar fix. 2013-12-05 18:54:01 +01:00
Anurag Ramdasan b67f39da09 fixed typo 2013-12-05 17:18:37 +01:00
Anurag Ramdasan 4c53178c6c Fixed grammar: 'usually' to 'usual' 2013-12-05 16:42:28 +01:00
antirez 7f6743a581 Fixed grammar: before H the article is a, not an. 2013-12-05 16:37:21 +01:00
antirez edae78999c Fixed typos in redis.conf file. 2013-12-05 16:30:15 +01:00
antirez 3d7263aa41 Removed old comments and dead code from freeClient(). 2013-12-03 13:54:15 +01:00
antirez 333453646c Grammar fix in freeClient(). 2013-12-03 13:40:51 +01:00
antirez 4d650a3baa Redis 2.8.2. 2013-12-02 16:07:46 +01:00
antirez dd0ac4ac72 Sentinel: don't write HZ when flushing config. 
See issue #1419.
 2013-12-02 15:56:42 +01:00
antirez 75347ada7f Sentinel: better time desynchronization. 
Sentinels are now desynchronized in a better way changing the time
handler frequency between 10 and 20 HZ. This way on average a
desynchronization of 25 milliesconds is produced that should be larger
enough compared to network latency, avoiding most split-brain condition
during the vote.

Now that the clocks are desynchronized, to have larger random delays when
performing operations can be easily achieved in the following way.
Take as example the function that starts the failover, that is
called with a frequency between 10 and 20 HZ and will start the
failover every time there are the conditions. By just adding as an
additional condition something like rand()%4 == 0, we can amplify the
desynchronization between Sentinel instances easily.

See issue #1419.
 2013-12-02 15:56:39 +01:00
antirez 83333b08d0 Stop writes on MISCONF only if instance is a master. 
From the point of view of the slave not accepting writes from the master
can only create a bigger consistency issue.
 2013-11-28 16:25:49 +01:00
antirez 50d140e90b SLAVEOF command refactored into a proper API. 
We now have replicationSetMaster() and replicationUnsetMaster() that can
be called in other contexts (for instance Redis Cluster).
 2013-11-28 16:19:16 +01:00
antirez 2eb8a46061 Reply to PING with error when there is a MISCONF state. 2013-11-28 16:16:58 +01:00
antirez a46f841df3 Sentinel: log vote received from other Sentinels. 2013-11-28 15:23:51 +01:00
huangz1990 e5c577e679 fix a bug in sentinel.c about pub/sub link 2013-11-26 15:11:58 +01:00
antirez dd239c37db Redis 2.8.1. 2013-11-25 11:21:49 +01:00
Salvatore Sanfilippo c61cc761d5 Merge pull request #1370 from yaauie/deprecate-redis-copy-script-2.8 
Deprecate redis copy script 2.8
 2013-11-25 02:10:32 -08:00
antirez d13635b2a9 Sentinel: fixes inverted strcmp() test preventing config updates. 
The result of this one-char bug was pretty serious, if the new master
had the same port of the previous master, but just a different IP
address, non-leader Sentinels would not be able to recognize the
configuration change.

This commit fixes issue #1394.

Many thanks to @shanemadden that reported the bug and helped
investigating it.
 2013-11-25 10:57:20 +01:00
antirez d240202261 Sentinel: fix type specifier for Hello msg generation. 
This fixes issue #1395.
 2013-11-25 10:25:10 +01:00
antirez e83741746c Fix false positive in memory efficiency test. 
Fixes issue #1298.
 2013-11-25 10:21:18 +01:00
antirez 4e6d34bd2c Redis 2.8.0. 2013-11-22 10:25:29 +01:00
antirez f75fccafe7 Sentinel: example sentinel.conf updated. 2013-11-21 17:07:10 +01:00
antirez 312ca4dacc Sentinel: different comments updated to new implementation. 2013-11-21 16:23:11 +01:00
antirez c84275ee60 Sentinel: cleanup around SENTINEL_INFO_VALIDITY_TIME. 2013-11-21 16:06:07 +01:00
antirez 0b2639123b Sentinel: removed mem leak and useless code. 2013-11-21 15:44:10 +01:00
antirez 750b007d7c Sentinel: manual failover works again. 2013-11-21 15:24:27 +01:00
antirez 4bd1dd1c53 Sentinel: test for writable config file. 
This commit introduces a funciton called when Sentinel is ready for
normal operations to avoid putting Sentinel specific stuff in redis.c.
 2013-11-21 15:24:22 +01:00
antirez 46e2f3468c Sentinel: check for disconnected links in sentinelSendHello(). 
Does not fix any bug as the test is performed by the caller, but better
to have the check.
 2013-11-21 15:24:15 +01:00
antirez 87098a39f0 Sentinel: Hello message sending code refactored. 2013-11-21 15:24:11 +01:00
antirez 9ecd819ecb Sentinel: select slave with best (greater) replication offset. 2013-11-21 15:24:08 +01:00
antirez eeb8cb3305 Sentinel: take the replication offset in slaves state. 2013-11-21 15:24:03 +01:00
antirez c517f72343 CONFIG REWRITE: don't add the signature if it already exists. 
At the end of the file, CONFIG REWRITE adds a comment line that:

    # Generated by CONFIG REWRITE

Followed by the additional config options required. However this was
added again and again at every rewrite in praticular conditions (when a
given set of options change in a given time during the time).

Now if it was alrady encountered, it is not added a second time.

This is especially important for Sentinel that rewrites the config at
every state change.
 2013-11-21 15:23:47 +01:00
antirez 812f76a850 Sentinel: distinguish between is-master-down-by-addr requests. 
Some are just to know if the master is down, and in this case the runid
in the request is set to "*", others are actually in order to seek for a
vote and get elected. In the latter case the runid is set to the runid
of the instance seeking for the vote.
 2013-11-21 15:22:55 +01:00
antirez fc93198ff9 Sentinel: various fixes to leader election implementation. 2013-11-21 15:22:52 +01:00
antirez 44b3684633 Sentinel: failover script execution fixed. 2013-11-21 15:22:48 +01:00
antirez 6d0400f569 Sentinel: no longer used defines removed. 2013-11-21 15:22:44 +01:00
antirez 1bbce5bf01 Sentinel: when writing config on disk, remember sentinels runid. 2013-11-21 15:22:27 +01:00
antirez 679dc8b09d Sentinel: arity of known-sentinel/slave is 4 not 3. 2013-11-21 15:22:21 +01:00
antirez 3e1fc6278d Sentinel: rewriteConfigSentinelOption() sub-iterators var typo fixed. 2013-11-21 15:22:18 +01:00
antirez 9a9c0cfaa6 Sentinel: call sentinelFlushConfig() to persist state when needed. 
Also the sentinel configuration rewriting was modified in order to
account for failover in progress, where we need to provide the promoted
slave address as master address, and the old master address as one of
the slaves address.
 2013-11-21 15:22:14 +01:00
antirez 93d924ff1c Sentinel: sentinelFlushConfig() to CONFIG REWRITE + fsync. 2013-11-21 15:22:11 +01:00
antirez a52909c5f2 Sentinel: CONFIG REWRITE support for Sentinel config. 2013-11-21 15:22:07 +01:00
antirez 8c3e197040 Sentinel: can-failover option removed, many comments fixed. 2013-11-21 15:22:02 +01:00
antirez 04b1fb0b1a Fix typo 'configuraiton' in rewriteConfigRewriteLine() comment. 2013-11-21 15:21:58 +01:00
antirez 0b9853ecdc Sentinel: added config options useful to take state on config rewrite. 
We'll use CONFIG REWRITE (internally) in order to store the new
configuration of a Sentinel after the internal state changes. In order
to do so, we need configuration options (that usually the user will not
touch at all) about config epoch of the master, Sentinels and Slaves
known for this master, and so forth.
 2013-11-21 15:21:55 +01:00
antirez 737062745d Sentinel: failover abort function simplified. 2013-11-21 15:21:50 +01:00
antirez 66b03c1a40 Sentinel: slaves reconfig delay modified. 
The time Sentinel waits since the slave is detected to be configured to
the wrong master, before reconfiguring it, is now the failover_timeout
time as this makes more sense in order to give the Sentinel performing
the failover enoung time to reconfigure the slaves slowly (if required
by the configuration).

Also we now PUBLISH more frequently the new configuraiton as this allows
to switch the reapprearing master back to slave faster.
 2013-11-21 15:21:46 +01:00
antirez 8ba31c218b Sentinel: failover restart time is now multiple of failover timeout. 
Also defaulf failover timeout changed to 3 minutes as the failover is a
fairly fast procedure most of the times, unless there are a very big
number of slaves and the user picked to configure them sequentially (in
that case the user should change the failover timeout accordingly).
 2013-11-21 15:21:41 +01:00
antirez ccaba966bc Sentinel: state machine and timeouts simplified. 2013-11-21 15:21:37 +01:00
antirez 3c4497e83c Sentinel: election timeout define. 2013-11-21 15:21:34 +01:00
antirez e15ba6a697 Sentinel: fix address of master in Hello messages. 
Once we switched configuration during a failover, we should advertise
the new address.

This was a serious race condition as the Sentinel performing the
failover for a moment advertised the old address with the new
configuration epoch: once trasmitted to the other Sentinels the broken
configuration would remain there forever, until the next failover
(because a greater configuration epoch is required to overwrite an older
one).
 2013-11-21 15:21:30 +01:00
antirez 1a6abe7d79 Sentinel: master address selection in get-master-address refactored. 2013-11-21 15:21:26 +01:00
antirez 0eeb0a0782 Sentinel: fix conditional to only affect slaves with wrong master. 2013-11-21 15:21:21 +01:00
antirez 64c8de8657 Sentinel: simplify and refactor slave reconfig code. 2013-11-21 15:20:56 +01:00
antirez 782f9cacaf Sentinel: reconfigure slaves to right master. 2013-11-21 15:20:44 +01:00
antirez 7dbc0a63f5 Sentinel: remember last time slave changed master. 2013-11-21 15:20:40 +01:00
antirez 612dbb2a91 Sentinel: redirect-to-master is not ok with new algorithm. 
Now Sentinel believe the current configuration is always the winner and
should be applied by Sentinels instead of trying to adapt our view of
the cluster based on what we observe.

So the only way to modify what a Sentinel believe to be the truth is to
win an election and advertise the new configuration via Pub / Sub with a
greater configuration epoch.
 2013-11-21 15:20:36 +01:00
antirez 4ccf807abc Sentinel: safer slave reconfig, master reported role should match. 2013-11-21 15:20:31 +01:00
antirez e98d82c639 Sentinel: role reporting fixed and added in SENTINEL output. 2013-11-21 15:20:25 +01:00
antirez be19e5450c Sentinel: being a master and reporting as slave is considered SDOWN. 2013-11-21 15:20:20 +01:00
antirez 8c551d65a1 Sentinel: make sure role_reported is always updated. 2013-11-21 15:20:15 +01:00
antirez 9577fed8a3 Sentinel: track role change time. Wait before reconfigurations. 2013-11-21 15:20:11 +01:00
antirez fc10fb17da Sentinel: fix no-down check in master->slave conversion code. 2013-11-21 15:20:07 +01:00
antirez b02ef3d59a Sentinel: readd slaves back after a master reset. 2013-11-21 15:19:56 +01:00
antirez 1a1dc3de38 Sentinel: sentinelResetMaster() new flag to avoid removing set of sentinels. 
This commit also removes some dead code and cleanup generic flags.
 2013-11-21 15:19:52 +01:00
antirez 9f0e52a13d Sentinel: receive Pub/Sub messages from slaves. 2013-11-21 15:19:49 +01:00
antirez f7604b4c07 Sentinel: change event name when converting master to slave. 2013-11-21 15:19:45 +01:00
antirez 1569326277 Sentinel: added config-epoch to SENTINEL masters output. 2013-11-21 15:19:41 +01:00
antirez 9f20780de6 Sentinel: new failover algo, desync slaves and update config epoch. 2013-11-21 15:19:34 +01:00
antirez 2488257ad8 Sentinel: when starting failover seek for votes ASAP. 2013-11-21 15:19:30 +01:00
antirez 6593f33222 Sentinel: +new-epoch events. 2013-11-21 15:19:26 +01:00
antirez 48acc675dd Sentinel: wait some time between failover attempts. 2013-11-21 15:19:22 +01:00
antirez 447c2787a0 Sentinel: allow to vote for myself. 2013-11-21 15:19:16 +01:00
antirez eba4775b5d Sentinel: fix PUBLISH to masters and slaves. 2013-11-21 15:19:11 +01:00
antirez b95c6ed7b7 Sentinel: epoch introduced in leader vote. 2013-11-21 15:18:52 +01:00
antirez 663d79c0d5 Sentinel: leadership handling changes WIP. 
Changes to leadership handling.

Now the leader gets selected by every Sentinel, for a specified epoch,
when the SENTINEL is-master-down-by-addr is sent.

This command now includes the runid and the currentEpoch of the instance
seeking for a vote. The Sentinel only votes a single time in a given
epoch.

Still a work in progress, does not even compile at this stage.
 2013-11-21 15:18:45 +01:00
antirez b72985d7f7 Sentinel: handle Hello messages received via slaves correctly. 
Even when messages are received via the slave, we should perform
operations (like adding a new Sentinel) in the context of the master.
 2013-11-21 15:18:41 +01:00
antirez fe7f96f18c Sentinel: remove code not useful in the new design. 2013-11-21 15:18:36 +01:00
antirez be2ef1b59f Sentinel: epoch introduced. 
Sentinel state now includes the idea of current epoch and config epoch.
In the Hello message, that is now published both on masters and slaves,
a Sentinel no longer just advertises itself but also broadcasts its
current view of the configuration: the master name / ip / port and its
current epoch.

Sentinels receiving such information switch to the new master if the
configuration epoch received is newer and the ip / port of the master
are indeed different compared to the previos ones.
 2013-11-21 15:18:31 +01:00
antirez 98682cd178 Log to what master a slave is going to connect to. 2013-11-11 09:25:41 +01:00
antirez 68b911478f Fixed typo in release candidate notes. 2013-11-10 10:25:09 +01:00
Ryan Biesemeyer 7d30b3121d Deprecate utils/redis-copy.rb in favor of redis-copy gem 2013-11-08 06:33:04 +00:00
antirez 875bc51909 Fix broken rdbWriteRaw() return value check in rdb.c. 
Thanks to @PhoneLi for reporting.
 2013-11-07 23:54:46 +01:00
antirez c874e39c45 Sentinel: sentinelSendSlaveOf() was missing a var and the prototype. 2013-11-06 11:29:57 +01:00
antirez 1767998751 Sentinel: increment pending_commands counter in two more places. 
AUTH and SCRIPT KILL were sent without incrementing the pending commands
counter. Clearly this needs some kind of wrapper doing it for the caller
in order to be less bug prone.
 2013-11-06 11:29:53 +01:00
antirez 97810c45e8 Sentinel: always send CONFIG REWRITE when changing instance role. 
This change makes Sentinel less fragile about a number of failure modes.

This commit also fixes a different bug as a side effect, SLAVEOF command
was sent multiple times without incrementing the pending commands count.
 2013-11-06 11:29:49 +01:00
antirez f7f97bf730 Redis 2.7.106 (2.8 Release Candidate 6). 2013-11-06 09:24:26 +01:00
antirez 3a66e0c157 Fixed typo in parseScanCursorOrReply(): ad -> at. 
Thanks to @badboy for reporting it.
 2013-11-05 17:25:34 +01:00
antirez 060d56e7eb SCAN code refactored to parse cursor first. 
The previous implementation of SCAN parsed the cursor in the generic
function implementing SCAN, SSCAN, HSCAN and ZSCAN.

The actual higher-level command implementation only checked for empty
keys and return ASAP in that case. The result was that inverting the
arguments of, for instance, SSCAN for example and write:

    SSCAN 0 key

Instead of

    SSCAN key 0

Resulted into no error, since 0 is a non-existing key name very likely.
Just the iterator returned no elements at all.

In order to fix this issue the code was refactored to extract the
function to parse the cursor and return the error. Every higher level
command implementation now parses the cursor and later checks if the key
exist or not.
 2013-11-05 17:25:29 +01:00
antirez 162acd8ac2 SCAN: when iterating ziplists or intsets always return cursor of 0. 
The previous implementation assumed that the first call always happens
with cursor set to 0, this may not be the case, and we want to return 0
anyway otherwise the (broken) client code will loop forever.
 2013-11-05 17:24:33 +01:00
antirez 406be0ea28 Use strtoul() instead of sscanf() in SCAN implementation. 2013-11-05 17:24:30 +01:00
antirez 4b615eea8f Added tests for [SHZ]SCAN with MATCH. 2013-11-05 17:24:26 +01:00
antirez 54a5a7dff8 HSCAN/ZSCAN: skip value when matching. 
This fixes issue #1360 and #1362.
 2013-11-05 17:24:22 +01:00
antirez bebbc7f9f7 Pass int64_t to intsetGet() instead of long long. 2013-11-05 17:24:17 +01:00
antirez 736e343509 removed not used vars in dictScan(). 2013-11-05 17:24:12 +01:00
antirez 59b30b327d SCAN: no sdsEncodedObject() API in Redis 2.8. 2013-10-31 18:22:29 +01:00
antirez 3d3e350b12 SSCAN with integer encoded object test improved. 2013-10-31 10:38:56 +01:00
antirez 3cc7d486bc Inverted variable boolean value and name after scanGenericCommand() refactoring. 2013-10-31 10:38:52 +01:00
antirez 06455014aa scanGenericCommand() refactoring and handling of integer encoded elements. 
This commit fixes issue #1354.
 2013-10-31 10:38:48 +01:00
antirez 141f30f421 Regression test added for [SHZ]SCAN issue #1354. 2013-10-31 10:38:45 +01:00
antirez c53cab5dde Test: added a SCAN test trying to trigger HT resize. 2013-10-30 16:51:48 +01:00
antirez ab6f4195dd Test: added ZSCAN test. 2013-10-30 16:27:43 +01:00
antirez 13e879c9cf Test: added HSCAN test. 2013-10-30 16:27:39 +01:00
antirez 9ea69a58ac Test: added SSCAN test. 2013-10-30 16:27:36 +01:00
antirez bd1962acc4 SCAN test keys sorting turned into more idiomatic Tcl. 2013-10-30 16:27:31 +01:00
antirez 02617b6e92 SCAN: tests moved to unit/scan.tcl. 2013-10-30 16:27:28 +01:00
antirez bb2df795e0 Aesthetic fix (missing space) into HSCAN and ZSCAN implementations. 
Thanks to @badboy for reporting.
 2013-10-29 16:21:21 +01:00
antirez 35816a2e4d ZSCAN implemented. 2013-10-29 16:21:18 +01:00
antirez 1cc7e646af HSCAN implemented. 2013-10-29 16:21:13 +01:00
antirez 24ecabc995 SSCAN implemented. 2013-10-29 16:21:01 +01:00
antirez 17785f40b5 dictScan(): empty hash table requires special handling. 2013-10-29 16:20:13 +01:00
antirez 7f063b1c8c SCAN is a random command and does not require output sorting. 
Sorting the output helps when we want to turn a non-deterministic into a
deterministic command, in that case this is not possible.
 2013-10-29 16:20:10 +01:00
antirez 9e62d9e79c SCAN: refactored into scanGenericCommand. 
The new implementation is capable of iterating the keyspace but also
sets, hashes, and sorted sets, and can be used to implement SSCAN, ZSCAN
and HSCAN.
 2013-10-29 16:20:06 +01:00
antirez 1a06bbe430 Fixed typos in dictScan() comment. 2013-10-29 16:20:02 +01:00
antirez 0880bec4e2 dictScan() algorithm documented. 2013-10-29 16:19:57 +01:00
antirez 1d659784f4 SCAN: stay inside 80 cols. 2013-10-29 16:19:54 +01:00
antirez 854b8be616 Revert "Fixed typo in SCAN comment. iff -> if." 
Probably here Pieter means "if and only if".

This reverts commit 43fdf3b404.
 2013-10-29 16:19:49 +01:00
antirez 5fa5153bf5 SCAN: simplify keys list cleanup using listSetFreeMethod(). 2013-10-29 16:19:45 +01:00
antirez 108c6e6734 SCAN: Fix test after option renamed from PATTERN to MATCH. 2013-10-29 16:19:41 +01:00
antirez 21da9b06a4 SCAN: improve variable names for readability. 2013-10-29 16:19:37 +01:00
antirez 907e06cd61 SCAN: remove additional newlines to conform to Redis code base. 2013-10-29 16:19:33 +01:00
antirez b44f1589a4 SCAN: remove useless assertion, already enforced by command table. 2013-10-29 16:19:28 +01:00
antirez d1c1f62ed5 SCAN: use define REDIS_LONGSTR_SIZE instead of fixed len. 2013-10-29 16:19:25 +01:00
antirez 8cdaf604ce Fixed typo in SCAN comment. iff -> if. 2013-10-29 16:19:21 +01:00
antirez 33409b4c6c SCAN option name changed: pattern -> match. 2013-10-29 16:19:17 +01:00
Pieter Noordhuis fd22106a98 Fix error in scan algorithm 
The irrelevant bits shouldn't be masked to 1. This can result in slots being
skipped when the hash table is resized between calls to the iterator.
 2013-10-29 16:19:12 +01:00
Pieter Noordhuis 8c4ff679fe SCAN requires at least 1 argument 2013-10-29 16:19:08 +01:00
Pieter Noordhuis 0dd95e23ab Add SCAN command 2013-10-29 16:18:24 +01:00
antirez 9546f7846b redis-benchmark: update help for new __rand_int__ form. 2013-10-28 18:14:11 +01:00
antirez 508d3c1bb7 Redis 2.7.105 (2.8 Release Candidate 5). 2013-10-09 13:08:42 +02:00
antirez df0c96002d Replication: install the write handler when reusing a cached master. 
Sometimes when we resurrect a cached master after a successful partial
resynchronization attempt, there is pending data in the output buffers
of the client structure representing the master (likely REPLCONF ACK
commands).

If we don't reinstall the write handler, it will never be installed
again by addReply*() family functions as they'll assume that if there is
already data pending, the write handler is already installed.

This bug caused some slaves after a successful partial sync to never
send REPLCONF ACK, and continuously being detected as timing out by the
master, with a disconnection / reconnection loop.
 2013-10-04 16:14:57 +02:00
antirez d7fa6d9aba Replication: fix master timeout. 
Since we started sending REPLCONF ACK from slaves to masters, the
lastinteraction field of the client structure is always refreshed as
soon as there is room in the socket output buffer, so masters in timeout
are detected with too much delay (the socket buffer takes a lot of time
to be filled by small REPLCONF ACK <number> entries).

This commit only counts data received as interactions with a master,
solving the issue.
 2013-10-04 13:01:49 +02:00
antirez c8c1006cf4 PSYNC: safer handling of PSYNC requests. 
There was a bug that over-esteemed the amount of backlog available,
however this could only happen when a slave was asking for an offset
that was in the "future" compared to the master replication backlog.

Now this case is handled well and logged as an incident in the master
log file.
 2013-10-04 12:27:40 +02:00
antirez 81f614ef6f Add REWRITE to CONFIG subcommands help message. 2013-10-04 12:27:36 +02:00
Michel Martens 20c506f432 Document the redis-cli --csv option. 2013-09-26 10:12:51 +02:00
antirez af967ff9cb Allow AUTH / PING when disconnected from slave and serve-stale-data is no. 2013-09-17 09:46:40 +02:00
antirez 2d55f1ca7f redis-cli: fix big keys search when the key no longer exist. 
The code freed a reply object that was never created, resulting in a
segfault every time randomkey returned a key that was deleted before we
queried it for size.
 2013-09-04 10:37:55 +02:00
Maxim Zakharov 9a0f27a5d0 mistype fixed 2013-09-03 15:15:42 +02:00
Maxim Zakharov c713b1ebdd A mistype fixed 2013-09-03 15:15:34 +02:00
antirez 56ce068664 Redis 2.7.104 (2.8 Release Candidate 4). 2013-08-30 09:52:18 +02:00
antirez 23acf93f97 Test: Lua stack leak regression test added. 2013-08-30 08:59:16 +02:00
antirez 0ea9a20d47 Test: added a memory efficiency test. 2013-08-30 08:48:07 +02:00
antirez 752b1fca90 Fixed critical memory leak from EVAL. 
Multiple missing calls to lua_pop prevented the error handler function
pushed on the stack for lua_pcall() to be popped before returning,
causing a memory leak in almost all the code paths of EVAL (both
successful calls and calls returning errors).

This caused two issues: Lua leaking memory (and this was very visible
from INFO memory output, as the 'used_memory_lua' field reported an
always increasing amount of memory used), and as a result slower and
slower GC cycles resulting in all the CPU being used.

Thanks to Tanguy Le Barzic for noticing something was wrong with his 2.8
slave, and for creating a testing EC2 environment where I was able to
investigate the issue.
 2013-08-29 11:49:28 +02:00
Salvatore Sanfilippo 3e62d7c308 Merge pull request #1244 from yihuang/2.8 
fix lua_cmsgpack pack map as array
 2013-08-27 06:15:32 -07:00
antirez a57294b339 Fix an hypothetical issue in processMultibulkBuffer(). 2013-08-27 13:00:14 +02:00
antirez facc5c0661 tryObjectEncoding(): optimize sds strings if possible. 
When no encoding is possible, at least try to reallocate the sds string
with one that does not waste memory (with free space at the end of the
buffer) when the string is large enough.
 2013-08-27 12:36:09 +02:00
antirez 79a1d335c8 tryObjectEncoding(): don't call stringl2() for too big strings. 
We are sure that a string that is longer than 21 chars cannot be
represented by a 64 bit signed integer, as -(2^64) is 21 chars:

strlen(-18446744073709551616) => 21
 2013-08-27 12:30:42 +02:00
antirez 0f32c37fc9 Fix DEBUG SDSLEN after 2.8 back port. 2013-08-27 12:26:02 +02:00
antirez df21771c69 Don't over-allocate the sds string for large bulk requests. 
The call to sdsMakeRoomFor() did not accounted for the amount of data
already present in the query buffer, resulting into over-allocation.
 2013-08-27 12:22:12 +02:00
antirez f05b1f6889 DEBUG SDSLEN added. 
This command is only useful for low-level debugging of memory issues due
to sds wasting memory as empty buffer at the end of the string.
 2013-08-27 12:22:07 +02:00
antirez 9c3e6c3aa0 Update server.lastbgsave_status when fork() fails. 2013-08-27 10:16:25 +02:00
antirez 9df63fcabf Only run the fast active expire cycle if master & enabled. 2013-08-27 09:31:43 +02:00
antirez f5903846bc Opening TCP listening ports refactored into a function. 2013-08-22 14:06:54 +02:00
antirez bd6327855b Print error message when can't bind * on any address. 2013-08-22 13:03:07 +02:00
antirez 2c94d80f58 Fix for issue #1214 simplified. 2013-08-21 11:41:35 +02:00
Allan 778f753deb fixed initServer fail while having no IPv6 nor IPv4 2013-08-21 11:41:30 +02:00
Allan 6c2b34a465 fixed initServer failed if no IPV4 or no IPV6 2013-08-21 11:41:25 +02:00
Allan 0a4656d63f fixed bug issue of #1213 2013-08-21 11:41:21 +02:00
antirez 7c861340b1 Use printf %zu specifier to print private_dirty. 2013-08-20 12:09:36 +02:00
antirez 2c018b14fb dictFingerprint(): cast pointers to integer of same size. 2013-08-20 12:09:31 +02:00
antirez 6a88a84ec7 Redis 2.7.103 (2.8 Release Candidate 3). 2013-08-19 18:04:32 +02:00
antirez 5621272575 Revert "Fixed type in dict.c comment: 265 -> 256." 
This reverts commit 009515f914.
 2013-08-19 17:25:57 +02:00
antirez 009515f914 Fixed type in dict.c comment: 265 -> 256. 2013-08-19 15:11:07 +02:00
antirez 837817b2b7 assert.h replaced with redisassert.h when appropriate. 
Also a warning was suppressed by including unistd.h in redisassert.h
(needed for _exit()).
 2013-08-19 15:02:16 +02:00
antirez bd353d9a3c Added redisassert.h as drop in replacement for assert.h. 
By using redisassert.h version of assert() you get stack traces in the
log instead of a process disappearing on assertions.
 2013-08-19 15:02:13 +02:00
antirez c7da5fc6c1 dictFingerprint() fingerprinting made more robust. 
The previous hashing used the trivial algorithm of xoring the integers
together. This is not optimal as it is very likely that different
hash table setups will hash the same, for instance an hash table at the
start of the rehashing process, and at the end, will have the same
fingerprint.

Now we hash N integers in a smarter way, by summing every integer to the
previous hash, and taking the integer hashing again (see the code for
further details). This way it is a lot less likely that we get a
collision. Moreover this way of hashing explicitly protects from the
same set of integers in a different order to hash to the same number.

This commit is related to issue #1240.
 2013-08-19 15:02:10 +02:00
antirez 8368550733 Fix comments for correctness in zunionInterGenericCommand(). 
Related to issue #1240.
 2013-08-19 15:02:06 +02:00
antirez 3ad87c652b Properly init/release iterators in zunionInterGenericCommand(). 
This commit does mainly two things:

1) It fixes zunionInterGenericCommand() by removing mass-initialization
of all the iterators used, so that we don't violate the unsafe iterator
API of dictionaries. This fixes issue #1240.

2) Since the zui* APIs required the allocator to be initialized in the
zsetopsrc structure in order to use non-iterator related APIs, this
commit fixes this strict requirement by accessing objects directly via
the op->subject->ptr pointer we have to the object.
 2013-08-19 15:02:02 +02:00
antirez 05379f49ee dict.c iterator API misuse protection. 
dict.c allows the user to create unsafe iterators, that are iterators
that will not touch the dictionary data structure in any way, preventing
copy on write, but at the same time are limited in their usage.

The limitation is that when itearting with an unsafe iterator, no call
to other dictionary functions must be done inside the iteration loop,
otherwise the dictionary may be incrementally rehashed resulting into
missing elements in the set of the elements returned by the iterator.

However after introducing this kind of iterators a number of bugs were
found due to misuses of the API, and we are still finding
bugs about this issue. The bugs are not trivial to track because the
effect is just missing elements during the iteartion.

This commit introduces auto-detection of the API misuse. The idea is
that an unsafe iterator has a contract: from initialization to the
release of the iterator the dictionary should not change.

So we take a fingerprint of the dictionary state, xoring a few important
dict properties when the unsafe iteartor is initialized. We later check
when the iterator is released if the fingerprint is still the same. If it
is not, we found a misuse of the iterator, as not allowed API calls
changed the internal state of the dictionary.

This code was checked against a real bug, issue #1240.

This is what Redis prints (aborting) when a misuse is detected:

Assertion failed: (iter->fingerprint == dictFingerprint(iter->d)),
function dictReleaseIterator, file dict.c, line 587.
 2013-08-19 15:01:58 +02:00
yihuang 706056ca16 fix lua_cmsgpack pack map as array 2013-08-13 17:48:40 +08:00
antirez 14a1cba3e3 Use precomptued objects for bulk and mbulk prefixes. 2013-08-12 12:43:26 +02:00
antirez 004f00bfa4 replicationFeedSlaves() func name typo: feedReplicationBacklogWithObject -> feedReplicationBacklog. 2013-08-12 12:38:52 +02:00
antirez 45d4e06e93 replicationFeedSlave() reworked for correctness and speed. 
The previous code using a static buffer as an optimization was lame:

1) Premature optimization, actually it was *slower* than naive code
   because resulted into the creation / destruction of the object
   encapsulating the output buffer.
2) The code was very hard to test, since it was needed to have specific
   tests for command lines exceeding the size of the static buffer.
3) As a result of "2" the code was bugged as the current tests were not
   able to stress specific corner cases.

It was replaced with easy to understand code that is safer and faster.
 2013-08-12 12:31:20 +02:00
antirez 3eab283b65 Fix a PSYNC bug caused by a variable name typo. 2013-08-12 11:51:03 +02:00
antirez 0e72cc601f Fix sdsempty() prototype in sds.h. 2013-08-12 11:38:28 +02:00
antirez 689829f2e1 redis-benchmark: changes to random arguments substitution. 
Before this commit redis-benchmark supported random argumetns in the
form of :rand:000000000000. In every string of that form, the zeros were
replaced with a random number of 12 digits at every command invocation.

However this was far from perfect as did not allowed to generate simply
random numbers as arguments, there was always the :rand: prefix.

Now instead every argument in the form __rand_int__ is replaced with a
12 digits number. Note that "__rand_int__" is 12 characters itself.

In order to implement the new semantic, it was needed to change a few
thigns in the internals of redis-benchmark, as new clients are created
cloning old clients, so without a stable prefix such as ":rand:" the old
way of cloning the client was no longer able to understand, from the old
command line, what was the position of the random strings to substitute.

Now instead a client structure is passed as a reference for cloning, so
that we can directly clone the offsets inside the command line.
 2013-08-08 16:47:18 +02:00
antirez 4fe67cc103 redis-benchmark: replace snprintf()+memcpy with faster code. 
This change was profiler-driven, but the actual effect is hard to
measure in real-world redis benchmark runs.
 2013-08-08 14:31:54 +02:00
Jan-Erik Rediger a256b83448 Little typo 2013-08-07 16:09:33 +02:00
antirez a79862de2a redis-benchmark: fix memory leak introduced by 346256f 2013-08-07 16:09:13 +02:00
antirez bcc7965931 redis-benchmark: max pipeline length hardcoded limit removed. 2013-08-07 16:09:09 +02:00
antirez 46c1bafeab redis-benchmark: fix db selection when :rand: feature is used. 2013-08-07 11:22:33 +02:00
antirez 151ad68540 redis-benchmark: ability to SELECT a specifid db number. 2013-08-07 11:22:28 +02:00
antirez fa48b1fa32 Add per-db average TTL information in INFO output. 
Example:

db0:keys=221913,expires=221913,avg_ttl=655

The algorithm uses a running average with only two samples (current and
previous). Keys found to be expired are considered at TTL zero even if
the actual TTL can be negative.

The TTL is reported in milliseconds.
 2013-08-06 15:36:43 +02:00
antirez 31d0f34100 activeExpireCycle(): fix about fast cycle early start. 
We don't want to repeat a fast cycle too soon, the previous code was
broken, we need to wait two times the period *since* the start of the
previous cycle in order to avoid there is an even space between cycles:

.-> start                   .-> second start
|                           |
+-------------+-------------+--------------+
| first cycle |    pause    | second cycle |
+-------------+-------------+--------------+

The second and first start must be PERIOD*2 useconds apart hence the *2
in the new code.
 2013-08-06 15:36:39 +02:00
antirez 00c8cfef74 Some activeExpireCycle() refactoring. 2013-08-06 15:36:35 +02:00
antirez 8686a70303 Remove dead code and fix comments for new expire code. 2013-08-06 15:36:31 +02:00
antirez d54e373b84 Darft #2 for key collection algo: more improvements. 
This commit makes the fast collection cycle time configurable, at
the same time it does not allow to run a new fast collection cycle
for the same amount of time as the max duration of the fast
collection cycle.
 2013-08-06 15:36:27 +02:00
antirez 500155b91b Draft #1 of a new expired keys collection algorithm. 
The main idea here is that when we are no longer to expire keys at the
rate the are created, we can't block more in the normal expire cycle as
this would result in too big latency spikes.

For this reason the commit introduces a "fast" expire cycle that does
not run for more than 1 millisecond but is called in the beforeSleep()
hook of the event loop, so much more often, and with a frequency bound
to the frequency of executed commnads.

The fast expire cycle is only called when the standard expiration
algorithm runs out of time, that is, consumed more than
REDIS_EXPIRELOOKUPS_TIME_PERC of CPU in a given cycle without being able
to take the number of already expired keys that are yet not collected
to a number smaller than 25% of the number of keys.

You can test this commit with different loads, but a simple way is to
use the following:

Extreme load with pipelining:

redis-benchmark -r 100000000 -n 100000000  \
        -P 32 set ele:rand:000000000000 foo ex 2

Remove the -P32 in order to avoid the pipelining for a more real-world
load.

In another terminal tab you can monitor the Redis behavior with:

redis-cli -i 0.1 -r -1 info keyspace

and

redis-cli --latency-history

Note: this commit will make Redis printing a lot of debug messages, it
is not a good idea to use it in production.
 2013-08-06 15:36:23 +02:00
antirez cc94697284 Redis 2.7.102 (2.8 Release Candidate 2). 2013-07-30 20:20:02 +02:00
antirez f8e43aba78 Test: regression test for issue #1221. 2013-07-29 17:40:18 +02:00
antirez 13f7ade551 Fix replicationFeedSlaves() off-by-one bug. 
This fixes issue #1221.
 2013-07-28 12:50:35 +02:00
antirez 78644f5e3e Remove dead variable bothsds from object.c. 
Thanks to @run and @badboy for spotting this.
Triva: clang was not able to provide me a warning about that when
compiling.

This closes #1024 and #1207, committing the change myself as the pull
requests no longer apply cleanly after other changes to the same
function.
 2013-07-28 11:02:20 +02:00
antirez aea8020554 Use latest sds.c in the hiredis library under deps. 2013-07-25 10:34:02 +02:00
antirez c3e1724a57 Ignore sdsrange return value. 2013-07-24 18:59:36 +02:00
antirez f899ab55ca sdsrange() does not need to return a value. 
Actaully the string is modified in-place and a reallocation is never
needed, so there is no need to return the new sds string pointer as
return value of the function, that is now just "void".
 2013-07-24 11:22:52 +02:00
antirez a1d37ba469 Inline protocol improved to accept quoted strings. 2013-07-24 10:38:05 +02:00
antirez 57c8e026b9 Every function inside sds.c is now commented. 2013-07-23 16:36:14 +02:00
antirez f3b44f291d Merge remote-tracking branch 'origin/2.8' into 2.8 2013-07-19 14:55:49 +02:00
antirez 5d0ad2f983 Fixed typo in 2.8 release notes. 2013-07-18 16:10:31 +02:00
Salvatore Sanfilippo 349f6349be Merge pull request #1203 from ronnix/patch-1 
Fix a few typos in release notes
 2013-07-18 03:27:41 -07:00
Ronan Amicel 5b19e9d433 Fix a few typos in release notes 2013-07-18 12:23:47 +02:00
antirez 3cb3714e99 Redis 2.7.101 (2.8 Release Candidate 1). 2013-07-18 11:26:53 +02:00
antirez a50635bb6c addReplyDouble(): format infinite in a libc agnostic way. 
There are systems that when printing +/- infinte with printf-family
functions will not use the usual "inf" "-inf", but different strings.
Handle that explicitly.

Fixes issue #930.
 2013-07-17 15:04:22 +02:00
antirez 2e449d42e0 Fixed typo in rio.h, simgle -> single. 2013-07-16 15:43:41 +02:00
yoav 9f6f436a51 Chunked loading of RDB to prevent redis from stalling reading very large keys. 2013-07-16 15:41:59 +02:00
antirez 112e763618 Make sure that ZADD can accept the full range of double values. 
This fixes issue #1194, that contains many details.

However in short, it was possible for ZADD to not accept as score values
that was however possible to obtain with multiple calls to ZINCRBY, like
in the following example:

redis 127.0.0.1:6379> zadd k 2.5e-308 m
(integer) 1
redis 127.0.0.1:6379> zincrby k -2.4e-308 m
"9.9999999999999694e-310"
redis 127.0.0.1:6379> zscore k m
"9.9999999999999694e-310"
redis 127.0.0.1:6379> zadd k 9.9999999999999694e-310 m1
(error) ERR value is not a valid float

The problem was due to strtod() returning ERANGE in the following case
specified by POSIX:

"If the correct value would cause an underflow, a value whose magnitude
is no greater than the smallest normalized positive number in the return
type shall be returned and errno set to [ERANGE].".

Now instead the returned value is accepted even when ERANGE is returned
as long as the return value of the function is not negative or positive
HUGE_VAL or zero.
 2013-07-16 15:08:57 +02:00
Ted Nyman fe04710908 Make sure the log standardizes on 'timeout' 2013-07-12 23:12:27 +02:00
antirez 98757b4a6e Use the environment locale for strcoll() collation. 2013-07-12 13:39:48 +02:00
antirez 18fabeb264 SORT ALPHA: use collation instead of binary comparison. 
Note that we only do it when STORE is not used, otherwise we want an
absolutely locale independent and binary safe sorting in order to ensure
AOF / replication consistency.

This is probably an unexpected behavior violating the least surprise
rule, but there is currently no other simple / good alternative.
 2013-07-12 13:39:44 +02:00
antirez d8fcbb6645 Fixed compareStringObject() and introduced collateStringObject(). 
compareStringObject was not always giving the same result when comparing
two exact strings, but encoded as integers or as sds strings, since it
switched to strcmp() when at least one of the strings were not sds
encoded.

For instance the two strings "123" and "123\x00456", where the first
string was integer encoded, would result into the old implementation of
compareStringObject() to return 0 as if the strings were equal, while
instead the second string is "greater" than the first in a binary
comparison.

The same compasion, but with "123" encoded as sds string, would instead
return a value < 0, as it is correct. It is not impossible that the
above caused some obscure bug, since the comparison was not always
deterministic, and compareStringObject() is used in the implementation
of skiplists, hash tables, and so forth.

At the same time, collateStringObject() was introduced by this commit, so
that can be used by SORT command to return sorted strings usign
collation instead of binary comparison. See next commit.
 2013-07-12 13:39:40 +02:00
Jan-Erik Rediger 1bcbb7a90c Wrap IPv6 in brackets in the prompt. 2013-07-12 10:58:47 +02:00
antirez 1e23848ed3 Sentinel: embed IPv6 address into [] when naming slave/sentinel instance. 2013-07-11 17:10:09 +02:00
antirez 076f6395b9 Sentinel: use comma as separator to publish hello messages. 
We use comma to play well with IPv6 addresses, but the implementation is
still able to parse the old messages separated by colons.
 2013-07-11 17:09:44 +02:00
antirez 4cfd70e903 hiredis: minimal IPv6 support. 2013-07-11 17:09:35 +02:00
antirez 98d0abcecd Sentinel: make sure published addr/id buffer is large enough. 
With ipv6 support we need more space, so we account for the IP address
max size plus what we need for the Run ID, port, flags.
 2013-07-11 17:09:30 +02:00
antirez 8669e70921 anet.c: save some vertical space. 2013-07-11 17:09:25 +02:00
antirez 9e089e7c5d anet.c: use SO_REUSEADDR when creating listening sockets. 
It used to be ok, but the socket option was removed when adding IPv6
support.
 2013-07-11 17:09:22 +02:00
antirez cdf801d92e Use getClientPeerId() for MONITOR implementation. 2013-07-11 17:09:18 +02:00
antirez 3472d045d8 getClientPeerId() refactored into two functions. 2013-07-11 17:09:14 +02:00
antirez da18366609 getClientPeerId() now reports errors. 
We now also use it in CLIENT KILL implementation.
 2013-07-11 17:09:10 +02:00
antirez 4fa68b2815 getClientPeerID introduced. 
The function returns an unique identifier for the client, as ip:port for
IPv4 and IPv6 clients, or as path:0 for Unix socket clients.

See the top comment in the function for more info.
 2013-07-11 17:09:04 +02:00
antirez a7451c1b6d All IP string repr buffers are now REDIS_IP_STR_LEN bytes. 2013-07-11 17:07:52 +02:00
antirez 2e75d3947c IPv6: bind IPv4 and IPv6 interfaces by default. 2013-07-11 17:07:41 +02:00
antirez b5423b099c Fix old anetPeerToString() API call in replication.c 2013-07-11 17:07:38 +02:00
antirez 1e7949227c Revert "Document port6 and bind6 config options." 
IPv6 support is not going to use IPv6 specific options, just it will be
possible to specify all the ipv4 / ipv6 addresses of the interfaces to
bind, otherwise connections will be accepted from all the interfaces in
both IPv4 and IPv6 addresses.

This reverts commit 93570e179e.
 2013-07-11 17:07:34 +02:00
Geoff Garside 0d8f254359 Add IPv6 support to sentinel.c. 
This has been done by exposing the anetSockName() function anet.c
to be used when the sentinel is publishing its existence to the masters.

This implementation is very unintelligent as it will likely break if used
with IPv6 as the nested colons will break any parsing of the PUBLISH string
by the master.
 2013-07-11 17:07:31 +02:00
Geoff Garside 5938126c98 Cleanup main() and BACKTRACE mistaken pulled while rebasing. 2013-07-11 17:07:27 +02:00
Geoff Garside d3bf85423a Fix calls to anetPeerToString() missing buffer size. 2013-07-11 17:07:22 +02:00
Geoff Garside 47620a3186 Document port6 and bind6 config options. 
Add commented port6 and bind6 options to default redis.conf file.
 2013-07-11 17:07:18 +02:00
Geoff Garside fd8a6ae7a6 Add anetTcp6Server() function. 
Refactor the common code from anetTcpServer into internal function which
can be used by both anetTcpServer and anetTcp6Server.
 2013-07-11 17:07:13 +02:00
Geoff Garside f638236902 Add static anetV6Only() function. 
This function sets the IPV6_V6ONLY option to 1 to use separate stack
IPv6 sockets.
 2013-07-11 17:07:07 +02:00
Geoff Garside 71795d4e7e Change anetTcpGenericConnect to use AF_UNSPEC. 
This allows anetTcpGenericConnect to try to connect to AF_INET6
addresses in addition to any resolved AF_INET addresses.
 2013-07-11 17:07:03 +02:00
Geoff Garside d37d006cd2 Mark places that might want changing for IPv6. 
Any places which I feel might want to be updated to work differently
with IPv6 have been marked with a comment starting "IPV6:".

Currently the only comments address places where an IP address is
combined with a port using the standard : separated form. These may want
to be changed when printing IPv6 addresses to wrap the address in []
such as

	[2001:db8::c0:ffee]:6379

instead of

	2001:db8::c0:ffee:6379

as the latter format is a technically valid IPv6 address and it is hard
to distinguish the IPv6 address component from the port unless you know
the port is supposed to be there.
 2013-07-11 17:06:55 +02:00
Geoff Garside 57e2b23dfb Expand ip char buffers which are too small for v6. 
Increase the size of character buffers being used to store printable IP
addresses so that they can safely store IPv6 addresses.
 2013-07-11 17:06:43 +02:00
Geoff Garside 39d57c057f Mark ip string buffers which could be reduced. 
In two places buffers have been created with a size of 128 bytes which
could be reduced to INET6_ADDRSTRLEN to still hold a full IP address.
These places have been marked as they are presently big enough to handle
the needs of storing a printable IPv6 address.
 2013-07-11 17:06:35 +02:00
Geoff Garside f1151f869e Update anetPeerToString to handle AF_INET6 addrs. 
Change the sockaddr_in to sockaddr_storage which is capable of storing
both AF_INET and AF_INET6 sockets. Uses the sockaddr_storage ss_family
to correctly return the printable IP address and port.
 2013-07-11 17:05:21 +02:00
Geoff Garside fefa14914e Update anetTcpAccept to handle AF_INET6 addresses. 
Change the sockaddr_in to sockaddr_storage which is capable of storing
both AF_INET and AF_INET6 sockets. Uses the sockaddr_storage ss_family
to correctly return the printable IP address and port.
 2013-07-11 17:05:16 +02:00
Geoff Garside 77ddec8582 Update anetResolve to resolve AF_INET6 as well. 
Change the getaddrinfo(3) hints family from AF_INET to AF_UNSPEC to
allow resolution of IPv6 addresses as well as IPv4 addresses. The
function will return the IP address of whichever address family is
preferenced by the operating system. Most current operating systems
will preference AF_INET6 over AF_INET.

Unfortunately without attempting to establish a connection to the
remote address we can't know if the host is capable of using the
returned IP address. It might be desirable to have anetResolve
accept an additional argument specifying the AF_INET/AF_INET6 address
the caller would like to receive. Currently though it does not appear
as though the anetResolve function is ever used within Redis.
 2013-07-11 17:05:12 +02:00
Geoff Garside 4b2e374e4a Update calls to anetResolve to include buffer size 2013-07-11 17:05:08 +02:00
Geoff Garside 3570411def Update calls to anetPeerToString to include ip_len. 2013-07-11 17:05:03 +02:00
Geoff Garside 68d72aa5b1 Add macro to define clusterNode.ip buffer size. 
Add REDIS_CLUSTER_IPLEN macro to define the size of the clusterNode ip
character array. Additionally use this macro in inet_ntop(3) calls where
the size of the array was being defined manually.

The REDIS_CLUSTER_IPLEN is defined as INET_ADDRSTRLEN which defines the
correct size of a buffer to store an IPv4 address in. The
INET_ADDRSTRLEN macro itself is defined in the <netinet/in.h> header
file and should be portable across the majority of systems.
 2013-07-11 17:04:32 +02:00
Geoff Garside 30836721f2 Update anetTcpAccept & anetPeerToString calls. 
Add the additional ip buffer length argument to function calls of
anetTcpAccept and anetPeerToString in network.c and cluster.c
 2013-07-11 17:02:29 +02:00
Geoff Garside 920ab4c926 Use inet_ntop(3) in anet. #apichange 
Replace inet_ntoa(3) calls with the more future proof inet_ntop(3)
function which is capable of handling additional address families.

API Change: anetTcpAccept() & anetPeerToString() additional argument
  additional argument required to specify the length of the character
  buffer the IP address is written to in order to comply with
  inet_ntop(3) function semantics.
 2013-07-11 17:02:12 +02:00
Geoff Garside aee5ee70fa Use getaddrinfo(3) in a anetTcpServer. 
Change anetTcpServer() function to use getaddrinfo(3) to perform
address resolution, socket creation and binding. Resolved addresses
are limited to those reachable by the AF_INET address family.
 2013-07-11 17:02:08 +02:00
Geoff Garside e50f2c93bc Use getaddrinfo(3) in anetTcpGenericConnect. 
Change anetTcpGenericConnect() function to use getaddrinfo(3) to
perform address resolution, socket creation and connection. Resolved
addresses are limited to those reachable by the AF_INET family.
 2013-07-11 17:02:04 +02:00
Geoff Garside d36dc1fc82 Add anetSetReuseAddr(err, fd) static function. 
Extract setting SO_REUSEADDR socket option into separate function
so the same code can be more easily used by anetCreateSocket and
other functions.
 2013-07-11 17:01:59 +02:00
Geoff Garside 855ff8df9f Use getaddrinfo(3) in anetResolve. #apichange 
Change anetResolve() function to use getaddrinfo(3) to resolve hostnames.
Resolved hostnames are limited to those reachable by the AF_INET address
family.

API Change: anetResolve requires additional argument.
  additional argument required to specify the length of the character
  buffer the IP address is written to in order to comply with
  inet_ntop(3) function semantics. inet_ntop(3) replaces inet_ntoa(3)
  as it has been designed to be compatible with more address families.
 2013-07-11 17:01:55 +02:00
antirez 33c5f0a243 Example redis.conf: bind to multiple interfaces documented. 2013-07-08 10:42:37 +02:00
antirez 323e5e2058 Remove cluster stuff from closeListeningSockets() for 2.8. 2013-07-08 10:31:49 +02:00
antirez fc022ca300 Binding multiple IPs done properly with multiple sockets. 2013-07-08 10:30:54 +02:00
antirez 00063bea33 Revert "anet.c: Allow creation of TCP listening sockets bound to N addresses." 
Bind() can't be called multiple times against the same socket, multiple
sockets are required to bind multiple interfaces, silly me.

This reverts commit bd234d62bb.
 2013-07-08 10:28:02 +02:00
antirez 6dabd34ad0 Ability to bind multiple addresses. 2013-07-08 10:27:21 +02:00
antirez 14857decfa anet.c: Allow creation of TCP listening sockets bound to N addresses. 2013-07-08 10:25:51 +02:00
antirez 2150cfebb7 sds.c: new function sdsjoin() to join strings. 2013-07-08 10:25:47 +02:00
antirez 3e3501262b redis-cli: introduced --pipe-timeout. 
When in --pipe mode, after all the data transfer to the server is
complete, now redis-cli waits at max the specified amount of
seconds (30 by default, use 0 to wait forever) without receiving any
reply at all from the server. After this time limit the operation is
aborted with an error.

That's related to issue #681.
 2013-07-03 12:18:55 +02:00
antirez 24b3799238 redis-cli --pipe: send final ECHO in a safer way. 
If the protocol read from stdin happened to contain grabage (invalid
random chars), in the previous implementation it was possible to end
with something like:

dksfjdksjflskfjl*2\r\n$4\r\nECHO....

That is invalid as the *2 should start into a new line. Now we prefix
the ECHO with a CRLF that has no effects on the server but prevents this
issues most of the times.

Of course if the offending wrong sequence is something like:

$3248772349\r\n

No one is going to save us as Redis will wait for data in the context of
a big argument, so this fix does not cover all the cases.

This partially fixes issue #681.
 2013-07-03 12:02:09 +02:00
antirez 2f6560f351 pqsort.c: remove the "switch to insertion sort" optimization. 
It causes catastrophic performance for certain inputs.

Relevant NetBSD commit:

http://cvsweb.netbsd.org/bsdweb.cgi/src/lib/libc/stdlib/qsort.c?rev=1.20&content-type=text/x-cvsweb-markup&only_with_tag=MAIN

This fixes issue #968.
 2013-07-02 17:48:49 +02:00
charsyam 847a40805f fix randstring bug 2013-07-02 12:25:30 +02:00
antirez aaf5623535 Only allow basenames for dbfilename and appendfilename. 
This fixes issue #1094.
 2013-07-02 12:24:12 +02:00
antirez 25a3ad7f74 pathIsBaseName() added to utils.c 
The function is used to test that the specified string looks like just
as the basename of a path, without any absolute or relative path.
 2013-07-02 12:24:08 +02:00
antirez 028fdbb70b getAbsolutePath() moved into utils.c 2013-07-02 12:24:05 +02:00
antirez 8090c37f71 CONFIG SET maxclients. 2013-07-01 10:44:08 +02:00
antirez 75dc48f04a ae.c event loop: API to resize the fd set size on the run. 2013-07-01 10:44:04 +02:00
antirez 0a0fd37e5b Allow SHUTDOWN in loading state. 2013-06-27 12:14:23 +02:00
antirez 7bfbc38068 Redis 2.7.3 2013-06-26 15:26:11 +02:00
antirez bb0d0fd479 function renamed: popcount_binary -> redisPopcount. 2013-06-26 15:24:58 +02:00
antirez cdf79c063f Don't disconnect pre PSYNC replication clients for timeout. 
Clients using SYNC to replicate are older implementations, such as
redis-cli --slave, and are not designed to acknowledge the master with
REPLCONF ACK commands, so we don't have any feedback and should not
disconnect them on timeout.
 2013-06-26 15:24:30 +02:00
antirez 7cca200603 Test: add some AOF testing to EVALSHA replication test. 2013-06-26 15:24:26 +02:00
antirez 29308077a8 Flush the replication script cache after SCRIPT FLUSH. 2013-06-26 15:24:22 +02:00
antirez 9faa6b9dc3 Test: EVALSHA replication. 2013-06-26 15:24:19 +02:00
antirez b0e0f3c5b2 Test: randomInt() behavior commented. 2013-06-26 15:24:16 +02:00
antirez 4e2b97bb89 Test: replication-3 test speedup in master-slave setup. 2013-06-26 15:24:12 +02:00
antirez d34c1ac208 Force propagation of SCRIPT LOAD to AOF. 2013-06-26 15:24:09 +02:00
antirez b7233f40fb SCRIPT FLUSH comment minor pedantic improvement. 2013-06-26 15:24:04 +02:00
antirez 8450d6e24d Move Replication Script Cache initialization in safer place. 
It should be called just one time at startup and not every time the Lua
scripting engine is re-initialized, otherwise memory is leaked.
 2013-06-26 15:23:52 +02:00
antirez 545fe0c318 Use the RSC to replicate EVALSHA unmodified. 
This commit uses the Replication Script Cache in order to avoid
translating EVALSHA into EVAL whenever possible for both the AOF and
slaves.
 2013-06-26 15:23:29 +02:00
antirez 9d894b1b8c Replication of scripts as EVALSHA: sha1 caching implemented. 
This code is only responsible to take an LRU-evicted fixed length cache
of SHA1 that we are sure all the slaves received.

In this commit only the implementation is provided, but the Redis core
does not use it to actually send EVALSHA to slaves when possible.
 2013-06-26 15:23:15 +02:00
antirez 8328d993e1 New API to force propagation. 
The old REDIS_CMD_FORCE_REPLICATION flag was removed from the
implementation of Redis, now there is a new API to force specific
executions of a command to be propagated to AOF / Replication link:

    void forceCommandPropagation(int flags);

The new API is also compatible with Lua scripting, so a script that will
execute commands that are forced to be propagated, will also be
propagated itself accordingly even if no change to data is operated.

As a side effect, this new design fixes the issue with scripts not able
to propagate PUBLISH to slaves (issue #873).
 2013-06-26 15:21:55 +02:00
antirez aeab473dfb Allow PUBSUB NUMSUB without channels. 
The result is an empty list but it is handy to call it programmatically.
 2013-06-26 15:21:12 +02:00
antirez a8f1474dfd PUBSUB command implemented. 
Currently it implements three subcommands:

PUBSUB CHANNELS [<pattern>]    List channels with non-zero subscribers.
PUBSUB NUMSUB [channel_1 ...]  List number of subscribers for channels.
PUBSUB NUMPAT                  Return number of subscribed patterns.
 2013-06-26 15:21:08 +02:00
antirez 72d3fbea65 Fix comment typo in integration/aof.tcl. 2013-06-26 15:21:05 +02:00
YAMAMOTO Takashi 3642557ae6 don't assume time_t == long 
time_t is always 64bit on recent versions of NetBSD.
 2013-06-26 15:20:59 +02:00
YAMAMOTO Takashi 30f3ae03a5 use nanosleep instead of usleep 
SUSv3 says that:
	The useconds argument shall be less than one million. If the value of
	useconds is 0, then the call has no effect.
and actually NetBSD's implementation rejects such a value with EINVAL.
use nanosleep which has no such a limitation instead.
 2013-06-26 15:20:56 +02:00
YAMAMOTO Takashi f27896a17d rename popcount to popcount_binary to avoid a conflict with NetBSD libc 
NetBSD-current's libc has a function named popcount.
hiding these extensions using feature macros is not possible because
redis uses other extensions covered by the same feature macro.
eg. inet_aton
 2013-06-26 15:20:32 +02:00
YAMAMOTO Takashi 9a7b09ed35 don't define _XOPEN_SOURCE for NetBSD 
on NetBSD, defining _XOPEN_SOURCE hides extensions
like inet_aton, strcasecmp, etc.
 2013-06-26 15:20:03 +02:00
Jan-Erik Rediger 90c3a14ac0 Initialize char* to NULL to remove compiler warning 2013-06-20 16:59:04 +02:00
antirez 9af8125c7d Sentinel: parse new INFO replication output correctly. 
Sentinel was not able to detect slaves when connected to a very recent
version of Redis master since a previos non-backward compatible change
to INFO broken the parsing of the slaves ip:port INFO output.

This fixes issue #1164
 2013-06-20 10:24:31 +02:00
antirez be8d5fd954 Test: regression test for #1163. 2013-06-19 18:53:14 +02:00
antirez 86147afd08 Allow writes from scripts called by AOF loading in read-only slaves. 
This fixes issue #1163
 2013-06-19 18:29:05 +02:00
antirez 449ec38fe4 Fix logStackTrace() when logging to stdout. 
When the semantics changed from logfile = NULL to logfile = "" to log
into standard output, no proper change was made to logStackTrace() to
make it able to work with the new setup.

This commit fixes the issue.
 2013-06-19 14:45:38 +02:00
antirez bf8e078aa2 Lua script errors format more unified. 
lua_pcall error handler now formats errors in a way more similar to
luaPushError() so that errors generated in different contexts look alike.
 2013-06-18 19:38:23 +02:00
antirez b3fbc84cc1 Lua scripting: improve error reporting. 
When calling Lua scripts we try to report not just the error but
information about the code line causing the error.
 2013-06-18 19:38:19 +02:00
Marc-Antoine Perennou 1cf839a216 test-server: only listen to 127.0.0.1 
Signed-off-by: Marc-Antoine Perennou <Marc-Antoine@Perennou.com>
 2013-06-18 19:38:16 +02:00
ioddly e34f00842f Try to report source of bad Lua API calls 2013-06-18 19:38:10 +02:00
antirez b435817b17 Binary safe dump of object content in redisLogObjectDebugInfo(). 2013-06-04 15:55:35 +02:00
antirez 054632f551 CONFIG SET: accept slave-priority zero, it is valid. 2013-05-31 19:31:51 +02:00
antirez d34c623e10 Test: avoid a false positive in min-slaves test. 2013-05-31 11:43:23 +02:00
antirez fe4b62fcaf Refresh good slaves count after CONFIG SET min-slaves-... 
This way just after the CONFIG SET enabling the min-slaves feature it is
possible to write to the database without delays.
 2013-05-31 07:20:19 +02:00
antirez 971217c7b4 New INFO field "min_slaves_good_slaves". 
When min-slaves-to-write feature is active, this field reports the
number of slaves considered good (online state, lag within the specified
range).
 2013-05-31 07:19:45 +02:00
antirez f5275da6e9 Refresh good slaves count when setting slave state as online. 2013-05-31 07:19:42 +02:00
antirez 9a7ea3dae9 Tests added for min-slaves feature. 2013-05-30 18:54:39 +02:00
antirez 4333e6ce20 Make tests compatible with new INFO replication output. 2013-05-30 11:43:53 +02:00
antirez 9d1bc3efcf min-replicas-to-write: only deny write commands. 
I guess I needed another coffee...
 2013-05-30 11:32:00 +02:00
antirez d0d67f8d42 min-slaves-to-write: don't accept writes with less than N replicas. 
This feature allows the user to specify the minimum number of
connected replicas having a lag less or equal than the specified
amount of seconds for writes to be accepted.
 2013-05-30 11:31:46 +02:00
antirez fc2587ec6e min-slaves-to-write: initial description of the feature in redis.conf 2013-05-30 11:30:34 +02:00
antirez 75fcf5c936 repl_offset field in INFO replication is now just offset. 2013-05-29 19:56:40 +02:00
antirez c64338aa8f Slaves list in INFO output: lag added, format changed. 
There is a new 'lag' information in the list of slaves, in the
"replication" section of the INFO output.

Also the format was changed in a backward incompatible way in order to
make it more easy to parse if new fields are added in the future, as the
new format is comma separated but has named fields (no longer positional
fields).
 2013-05-29 19:55:11 +02:00
antirez 889d39b5ba Accept REPLCONF in any state. 2013-05-28 15:26:55 +02:00
antirez 1b87b3ef38 A comment about BLPOP timeout did not reflected actual behavior. 2013-05-27 19:34:12 +02:00
antirez 2d497014ea Version bumped to 2.7.2. 2013-05-27 11:43:47 +02:00
antirez 75bf91abea redis.conf updated: repl-timeout now uesd by masters as well. 2013-05-27 11:43:21 +02:00
antirez 308940aa2c Close connection with timedout slaves. 
Now masters, using the time at which the last REPLCONF ACK was received,
are able to explicitly disconnect slaves that are no longer responding.

Previously the only chance was to see a very long output buffer, that
was highly suboptimal.
 2013-05-27 11:43:17 +02:00
antirez 45e6a4023a Send ACK to master once every second. 
ACKs can be also used as a base for synchronous replication. However in
that case they'll be explicitly requested by the master when the client
sends a request that needs to be replicated synchronously.
 2013-05-27 11:43:14 +02:00
antirez a74f8fe1ad Don't ACK the master after every command. 
Sending an ACK is now moved into the replicationSendAck() function.
 2013-05-27 11:43:10 +02:00
antirez 0000d5334d Make sure that REPLCONF ACK really has no return value. 2013-05-27 11:43:06 +02:00
antirez 146f1d7d86 Replication: send REPLCONF ACK to master. 2013-05-27 11:43:03 +02:00
antirez 1e77b77de4 REPLCONF ACK command. 
This special command is used by the slave to inform the master the
amount of replication stream it currently consumed.

it does not return anything so that we not need to consume additional
bandwidth needed by the master to reply something.

The master can do a number of things knowing the amount of stream
processed, such as understanding the "lag" in bytes of the slave, verify
if a given command was already processed by the slave, and so forth.
 2013-05-27 11:43:00 +02:00
antirez 0387f86947 Fixed a bug in no queueing replies to master. 2013-05-25 01:00:16 +02:00
antirez 7ae29d5d9e Replication: don't even queue replies to master commands. 
When master send commands, there is no need for the slave to reply.
Redis used to queue the reply in the output buffer and discard the reply
later, this is a waste of work and it is not clear why it was this way
(I sincerely don't remember).

This commit changes it in order to don't queue the reply at all.

All tests passing.
 2013-05-24 19:48:03 +02:00
antirez 302b0a5c7b Top comment for prepareClientToWrite() clarified. 
We don't write the output buffer to the client socket for slaves only if
the slave is not online.
 2013-05-24 18:42:18 +02:00
antirez 0b398b6ffa Don't stop reading redis.conf if line has no args. 
Should be "continue" and was "return".

This fixes issue #1110
 2013-05-18 16:22:24 +02:00
antirez 059018c4cd CONFIG REWRITE: remove cluster stuff for 2.8 back port. 2013-05-15 12:02:44 +02:00
antirez 414d5db85c Use memtoll() when parsing the backlog-size option. 2013-05-15 12:01:32 +02:00
antirez 604efb02ab CONFIG REWRITE: backlog size is a bytes option. 2013-05-15 12:01:27 +02:00
antirez 7d338fd982 CONFIG REWRITE: bindaddr -> bind. 2013-05-15 12:01:23 +02:00
antirez 2173f30836 CONFIG REWRITE: when rewriting amount of bytes use GB, MB, KB if possible. 2013-05-15 12:01:14 +02:00
antirez 901ceab20b CONFIG REWRITE: correctly escape the notify-keyspace-events option. 2013-05-15 12:01:06 +02:00
antirez b4ce02988c CONFIG REWRITE: "active-rehashing" -> "activerehashing". 2013-05-15 12:01:02 +02:00
antirez 2a29cf430a CONFIG REWRITE: fixed typo in AOF fsync policy. 2013-05-15 12:00:57 +02:00
antirez c95a60960b CONFIG REWRITE: repl-disable-tcp-nodelay is a boolean option. 2013-05-15 12:00:52 +02:00
antirez e8cc959300 Test default config modified to set logfile to empty string. 
Required because of recent changes in the way logfile is set to standard
output.
 2013-05-15 12:00:47 +02:00
antirez 180cfaae8e Added a define for most configuration defaults. 
Also the logfile option was modified to always have an explicit value
and to log to stdout when an empty string is used as log file.

Previously there was special handling of the string "stdout" that set
the logfile to NULL, this always required some special handling.
 2013-05-15 12:00:43 +02:00
antirez 417253afc3 CONFIG REWRITE: Use sane perms when creating redis.conf from scratch. 2013-05-15 12:00:38 +02:00
antirez 8feb3d261a CONFIG REWRITE: actually rewrite the config file, atomically. 2013-05-15 12:00:32 +02:00
antirez a461376ca9 CONFIG REWRITE: remove orphaned lines. 2013-05-15 12:00:28 +02:00
antirez 8cf0b66606 CONFIG REWRITE: strip multiple empty lines. 2013-05-15 12:00:24 +02:00
antirez cb8433f313 CONFIG REWRITE: support for client-output-buffer-limit. 2013-05-15 12:00:21 +02:00
antirez f62851aea9 CONFIG REWRITE: support for dir and slaveof. 2013-05-15 12:00:17 +02:00
antirez 91ef10ef45 CONFIG REWRITE: support for save and syslog-facility. 2013-05-15 12:00:13 +02:00
antirez 973f793b04 CONFIG REWRITE: Initial support code and design. 2013-05-15 12:00:03 +02:00
antirez 7a5d3d91a1 Obtain absoute path of configuration file, expose it in INFO. 2013-05-15 11:59:26 +02:00
antirez d8dd70c2dd redis-cli: help.h updated. 2013-05-14 11:22:58 +02:00
antirez 41babac68d Test: more PSYNC tests (backlog TTL). 2013-05-09 12:52:16 +02:00
antirez cb55ad991b Version bumped to 2.7.1 2013-05-08 15:24:54 +02:00
antirez c847c73e78 Test: check that replication partial sync works if we break the link. 
The test checks both successful syncs and unsuccessful ones by changing
the backlog size.
 2013-05-08 13:02:53 +02:00
antirez 2865bb7f1f Test: various issues with the replication-4.tcl test fixed. 
The test actually worked, but vars for master and slave were inverted
and sometimes used incorrectly.
 2013-05-08 11:59:42 +02:00
Jiahao Huang 3d93dffd5d in 32bit machine, popcount don't work with a input string length up to 512 MB, 
bitcount commant may return negtive integer with string length more than 256 MB
 2013-05-08 09:56:10 +02:00
antirez 85671d0e3b Added more info about 32 bit builds into README. 2013-05-08 09:48:46 +02:00
antirez b05d2516fe CONFIG SET server.masterauth aesthetic change. 
This is just to make the code exactly like the above instance used for
requirepass. No actual change nor the original code violated the Redis
coding style.
 2013-05-02 17:22:30 +02:00
Michel Martens b60465eb78 Reset masterauth if an empty string is configured. 2013-05-02 17:21:55 +02:00
charsyam a94db7908a Fix AOF bug: expire could be removed from key on AOF rewrite. 
There was a race condition in the AOF rewrite code that, with bad enough
timing, could cause a volatile key just about to expire to be turned
into a non-volatile key. The bug was never reported to cause actualy
issues, but was found analytically by an user in the Redis mailing list:

https://groups.google.com/forum/?fromgroups=#!topic/redis-db/Kvh2FAGK4Uk

This commit fixes issue #1079.
 2013-05-02 15:46:56 +02:00
antirez e7bcec829c Sentinel: changes to tilt mode. 
Tilt mode was too aggressive (not processing INFO output), this
resulted in a few problems:

1) Redirections were not followed when in tilt mode. This opened a
   window to misinform clients about the current master when a Sentinel
   was in tilt mode and a fail over happened during the time it was not
   able to update the state.

2) It was possible for a Sentinel exiting tilt mode to detect a false
   fail over start, if a slave rebooted with a wrong configuration
   about at the same time. This used to happen since in tilt mode we
   lose the information that the runid changed (reboot).

   Now instead the Sentinel in tilt mode will still remove the instance
   from the list of slaves if it changes state AND runid at the same
   time.

Both are edge conditions but the changes should overall improve the
reliability of Sentinel.
 2013-04-30 15:09:14 +02:00
antirez 4028a777b6 Sentinel: more sensible delay in master demote after tilt. 2013-04-30 15:09:10 +02:00
antirez 70845320cc Sentinel: only demote old master into slave under certain conditions. 
We used to always turn a master into a slave if the DEMOTE flag was set,
as this was a resurrecting master instance.

However the following race condition is possible for a Sentinel that
got partitioned or internal issues (tilt mode), and was not able to
refresh the state in the meantime:

1) Sentinel X is running, master is instance "A".
3) "A" fails, sentinels will promote slave "B" as master.
2) Sentinel X goes down because of a network partition.
4) "A" returns available, Sentinels will demote it as a slave.
5) "B" fails, other Sentinels will promote slave "A" as master.
6) At this point Sentinel X comes back.

When "X" comes back he thinks that:

"B" is the master.
"A" is the slave to demote.

We want to avoid that Sentinel "X" will demote "A" into a slave.
We also want that Sentinel "X" will detect that the conditions changed
and will reconfigure itself to monitor the right master.

There are two main ways for the Sentinel to reconfigure itself after
this event:

1) If "B" is reachable and already configured as a slave by other
sentinels, "X" will perform a redirection to "A".
2) If there are not the conditions to demote "A", the fact that "A"
reports to be a master will trigger a failover detection in "X", that
will end into a reconfiguraiton to monitor "A".

However if the Sentinel was not reachable, its state may not be updated,
so in case it titled, or was partiitoned from the master instance of the
slave to demote, the new implementation waits some time (enough to
guarantee we can detect the new INFO, and new DOWN conditions).

If after some time still there are not the right condiitons to demote
the instance, the DEMOTE flag is cleared.
 2013-04-30 15:09:06 +02:00
antirez 2b427a89fb Lua updated to version 5.1.5. 2013-04-29 12:02:01 +02:00
antirez d2ff5ed603 Sentinel: always redirect on master->slave transition. 
Sentinel redirected to the master if the instance changed runid or it
was the first time we got INFO, and a role change was detected from
master to slave.

While this is a good idea in case of slave->master, since otherwise we
could detect a failover without good reasons just after a reboot with a
slave with a wrong configuration, in the case of master->slave
transition is much better to always perform the redirection for the
following reasons:

1) A Sentinel may go down for some time. When it is back online there is
no other way to understand there was a failover.
2) Pointing clients to a slave seems to be always the wrong thing to do.
3) There is no good rationale about handling things differently once an
instance is rebooted (runid change) in that case.
 2013-04-24 11:34:02 +02:00
antirez b06f13e7b7 Config option to turn AOF rewrite incremental fsync on/off. 2013-04-24 10:57:35 +02:00
antirez 9ca306d874 AOF: sync data on disk every 32MB when rewriting. 
This prevents the kernel from putting too much stuff in the output
buffers, doing too heavy I/O all at once. So the goal of this commit is
to split the disk pressure due to the AOF rewrite process into smaller
spikes.

Please see issue #1019 for more information.
 2013-04-24 10:27:02 +02:00
antirez 5276996cf1 rio.c: added ability to fdatasync() from time to time while writing. 2013-04-24 10:26:58 +02:00
antirez e426ea5245 Test: fix RDB test checking file permissions. 
When the test is executed using the root account, setting the permission
to 222 does not work as expected, as root can read files with 222
permission.

Now we skip the test if root is detected.

This fixes issue #1034 and the duplicated #1040 issue.

Thanks to Jan-Erik Rediger (@badboy on Github) for finding a way to reproduce the issue.
 2013-04-23 14:16:47 +02:00
antirez d0c9a2a767 Sentinel: turn old master into a slave when it comes back. 2013-04-22 11:26:29 +02:00
antirez a500c120ba Test: split conceptually unrelated comments in RDB test. 2013-04-22 11:26:02 +02:00
antirez 3a865829a7 More explicit panic message on out of memory. 2013-04-19 15:11:55 +02:00
antirez 0e3baa5241 Redis/Jemalloc Gitignore were too aggressive. 
Redis gitignore was too aggressive since simply broken.

Jemalloc gitignore was too agressive because it is conceived to just
keep the files that allow to generate all the rest in development
environments (so for instance the "configure" file is excluded).
 2013-04-18 16:23:12 +02:00
antirez 42ba5d3788 redis-cli: raise error on bad command line switch. 
Previously redis-cli never tried to raise an error when an unrecognized
switch was encountered, as everything after the initial options is to be
transmitted to the server.

However this is too liberal, as there are no commands starting with "-".
So the new behavior is to produce an error if there is an unrecognized
switch starting with "-". This should not break past redis-cli usages
but should prevent broken options to be silently discarded.

As far the first token not starting with "-" is encountered, all the
rest is considered to be part of the command, so you cna still use
strings starting with "-" as values, like in:

    redis-cli --port 6380 set foo --my-value
 2013-04-11 13:19:52 +02:00
antirez 0d7b10ee6e redis-cli: --latency-history mode implemented. 2013-04-11 13:19:48 +02:00
antirez 8d81e68d18 Test: remove useless statements and comments from test default config. 2013-04-04 14:31:30 +02:00
antirez d83d72f380 Make rio.c comment 80-columns friendly. 2013-04-03 12:41:27 +02:00
antirez d6b0c18c51 Throttle BGSAVE attempt on saving error. 
When a BGSAVE fails, Redis used to flood itself trying to BGSAVE at
every next cron call, that is either 10 or 100 times per second
depending on configuration and server version.

This commit does not allow a new automatic BGSAVE attempt to be
performed before a few seconds delay (currently 5).

This avoids both the auto-flood problem and filling the disk with
logs at a serious rate.

The five seconds limit, considering a log entry of 200 bytes, will use
less than 4 MB of disk space per day that is reasonable, the sysadmin
should notice before of catastrofic events especially since by default
Redis will stop serving write queries after the first failed BGSAVE.

This fixes issue #849
 2013-04-02 14:12:28 +02:00
Johan Bergström 0bc7dbda1c use install as default installer (except on SunOS) 2013-04-02 11:54:20 +02:00
Johan Bergström 034d94193d Silence mkdir output 2013-04-02 11:30:16 +02:00
Johan Bergström 2094eb36e7 Only pass -rdynamic as linker option 2013-04-02 11:30:13 +02:00
Johan Bergström 31844fe138 Remove extra spaces 2013-04-02 11:29:56 +02:00
Johan Bergström c72459bf18 make check is a common naming convention for tests 2013-04-02 11:27:38 +02:00
Johan Bergström ca0b1c2436 Inherit CC for Lua 2013-04-02 11:26:50 +02:00
Johan Bergström 49eef7738e Spaces to tabs 2013-04-02 11:26:41 +02:00
Johan Bergström 0c8690d46e Slightly refactor CFLAGS/LDFLAGS/LIBS 
This way, we can avoid -rdynamic and -pthread warnings on darwin.
 2013-04-02 11:26:37 +02:00
charsyam cf8484b723 Support for case unsensitive SET options. 2013-03-29 10:37:29 +01:00
antirez 0c447d3506 Test: Extended SET tests. 2013-03-28 16:39:39 +01:00
antirez 24aa61785d Extended SET command implemented (issue #931). 2013-03-28 16:39:34 +01:00
antirez e6e07042a1 EXPIRE should not resurrect keys. Issue #1026. 2013-03-28 12:48:38 +01:00
antirez be832929c8 Test: regression test for issue #1026. 2013-03-28 12:48:35 +01:00
antirez 4de4080ea5 Better DEBUG error message when num of arguments is wrong. 2013-03-28 12:48:30 +01:00
antirez 978185bf67 Test: verify that lazy-expire works. 2013-03-28 12:48:26 +01:00
antirez 10d8e6a712 DEBUG set-active-expire added. 
We need the ability to disable the activeExpireCycle() (active
expired key collection) call for testing purposes.
 2013-03-28 12:48:19 +01:00
antirez c6a9a20d3b Test: test replication of MULTI/EXEC. 2013-03-27 12:05:22 +01:00
antirez 61b1d6da18 Test: Restore DB back to 9 after testing MULTI/EXEC with DB 5. 2013-03-27 12:05:19 +01:00
antirez 7009ff98b8 Test: new functions to capture and analyze the replication stream. 2013-03-27 12:05:15 +01:00
antirez a452b54849 TTL / PTTL commands: two bugs fixed. 
This commit fixes two corner cases for the TTL command.

1) When the key was already logically expired (expire time older
than current time) the command returned -1 instead of -2.

2) When the key was existing and the expire was found to be exactly 0
(the key was just about to expire), the command reported -1 (that is, no
expire) instead of a TTL of zero (that is, about to expire).
 2013-03-26 13:59:33 +01:00
antirez 10c6195e3f Flag PUBLISH as read-only in the command table. 2013-03-26 13:59:26 +01:00
antirez 2618101ad5 Transactions: propagate MULTI/EXEC only when needed. 
MULTI/EXEC is now propagated to the AOF / Slaves only once we encounter
the first command that is not a read-only one inside the transaction.

The old behavior was to always propagate an empty MULTI/EXEC block when
the transaction was composed just of read only commands, or even
completely empty. This created two problems:

1) It's a bandwidth waste in the replication link and a space waste
   inside the AOF file.

2) We used to always increment server.dirty to force the propagation of
   the EXEC command, resulting into triggering RDB saves more often
   than needed.

Note: even read-only commands may also trigger writes that will be
propagated, when we access a key that is found expired and Redis will
synthesize a DEL operation. However there is no need for this to stay
inside the transaction itself, but only to be ordered.

So for instance something like:

    MULTI
    GET foo
    SET key zap
    EXEC

May be propagated into:

    DEL foo
    MULTI
    SET key zap
    EXEC

While the DEL is outside the transaction, the commands are delivered in
the right order and it is not possible for other commands to be inserted
between DEL and MULTI.
 2013-03-26 13:59:21 +01:00
antirez 82516e8168 Transactions: use discardTransaction() in EXEC implementation. 2013-03-26 13:59:18 +01:00
antirez d91dfaf6dc Transactions: use the propagate() API to propagate MULTI. 
The behavior is the same, but the code is now cleaner and uses the
proper interface instead of dealing directly with AOF/replication
functions.
 2013-03-26 13:59:14 +01:00
antirez 150f7e43ca Allow SELECT while loading the DB. 
Fixes issue #1024.
 2013-03-26 13:59:05 +01:00
NanXiao 382420ec6e Update config.c 
Fix bug in configGetCommand function: get correct masterauth value.
 2013-03-25 19:32:20 +01:00
antirez 6b7562e204 Test: obuf-limits test false positive removed. 
Fixes #621.
 2013-03-25 11:56:31 +01:00
antirez ec18d4bfba redis-cli --stat, stolen from redis-tools. 
Redis-tools is a connection of tools no longer mantained that was
intented as a way to economically make sense of Redis in the pre-vmware
sponsorship era. However there was a nice redis-stat utility, this
commit imports one of the functionalities of this tool here in redis-cli
as it seems to be pretty useful.

Usage: redis-cli --stat

The output is similar to vmstat in the format, but with Redis specific
stuff of course.

From the point of view of the monitored instance, only INFO is used in
order to grab data.
 2013-03-25 11:52:25 +01:00
antirez 58708fa65a Replication: master_link_down_since_seconds initial value should be huge. 
server.repl_down_since used to be initialized to the current time at
startup. This is wrong since the replication never started. Clients
testing this filed to check if data is uptodate should never believe
data is recent if we never ever connected to our master.
 2013-03-13 12:55:02 +01:00
antirez ba8367ee49 Test: make sure broken RDB checksum is detected. 2013-03-13 11:15:31 +01:00
antirez 67a745515d Test: more RDB loading checks. 
A test for issue #1001 is included.
 2013-03-13 10:08:40 +01:00
antirez 7b6fb960cc Test: check that Redis starts empty without an RDB file. 2013-03-13 10:08:36 +01:00
antirez 64b91c1493 rdbLoad(): rework code to save vertical space. 2013-03-13 10:08:30 +01:00
Damian Janowski edb3c10361 Abort when opening the RDB file results in an error other than ENOENT. 
This fixes cases where the RDB file does exist but can't be accessed for
any reason. For instance, when the Redis process doesn't have enough
permissions on the file.
 2013-03-13 10:08:26 +01:00
antirez b2e42a90d3 Set default for stop_writes_on_bgsave_err in initServerConfig(). 
It was placed for error in initServer() that's called after the
configuation is already loaded, causing issue #1000.
 2013-03-12 18:36:02 +01:00
antirez 378fcaad55 redis-cli --bigkeys: don't crash with empty DBs. 2013-03-12 09:57:57 +01:00
antirez 8e7d962291 activeExpireCycle() smarter with many DBs and under expire pressure. 
activeExpireCycle() tries to test just a few DBs per iteration so that
it scales if there are many configured DBs in the Redis instance.
However this commit makes it a bit smarter when one a few of those DBs
are under expiration pressure and there are many many keys to expire.

What we do is to remember if in the last iteration had to return because
we ran out of time. In that case the next iteration we'll test all the
configured DBs so that we are sure we'll test again the DB under
pressure.

Before of this commit after some mass-expire in a given DB the function
tested just a few of the next DBs, possibly empty, a few per iteration,
so it took a long time for the function to reach again the DB under
pressure. This resulted in a lot of memory being used by already expired
keys and never accessed by clients.
 2013-03-11 11:30:13 +01:00
antirez c5afbb6e2a In databasesCron() never test more DBs than we have. 2013-03-11 11:30:09 +01:00
antirez 7e7ee815b0 Make comment name match var name in activeExpireCycle(). 2013-03-11 11:30:04 +01:00
antirez 6a8288766c Optimize inner loop of activeExpireCycle() for no-expires case. 2013-03-11 11:30:01 +01:00
antirez 9afb7789b3 REDIS_DBCRON_DBS_PER_SEC -> REDIS_DBCRON_DBS_PER_CALL 2013-03-11 11:29:57 +01:00
antirez 1a73260dce activeExpireCycle(): process only a small number of DBs per iteration. 
This small number of DBs is set to 16 so actually in the default
configuraiton Redis should behave exactly like in the past.
However the difference is that when the user configures a very large
number of DBs we don't do an O(N) operation, consuming a non trivial
amount of CPU per serverCron() iteration.
 2013-03-11 11:29:53 +01:00
antirez 8abfe5fe8b Use unsigned integers for DB ids, for defined wrap-to-zero. 2013-03-11 11:29:49 +01:00
antirez a4bb4b29fb Only resize/rehash a few databases per cron iteration. 
This is the first step to lower the CPU usage when many databases are
configured. The other is to also process a limited number of DBs per
call in the active expire cycle.
 2013-03-11 11:29:45 +01:00
antirez aa7f74c7e9 Actually call databasesCron() inside serverCron(). 2013-03-11 11:29:41 +01:00
antirez 21822fd0e9 Move Redis databases background processing to databasesCron(). 2013-03-11 11:29:34 +01:00
Gengliang Wang 6ae8fb477a Removed useless "return" statements in pubsub.c 
(original commit message edited)
 2013-03-06 16:50:53 +01:00
antirez bc1b2e8f96 API to lookup commands with their original name. 
A new server.orig_commands table was added to the server structure, this
contains a copy of the commant table unaffected by rename-command
statements in redis.conf.

A new API lookupCommandOrOriginal() was added that checks both tables,
new first, old later, so that rewriteClientCommandVector() and friends
can lookup commands with their new or original name in order to fix the
client->cmd pointer when the argument vector is renamed.

This fixes the segfault of issue #986, but does not fix a wider range of
problems resulting from renaming commands that actually operate on data
and are registered into the AOF file or propagated to slaves... That is
command renaming should be handled with care.
 2013-03-06 16:36:57 +01:00
antirez deccd104d8 Add a warning about command renaming in redis.conf. 2013-03-06 16:36:57 +01:00
antirez 5bcfa2ebbc Handle a non-impossible empty argv in loadServerConfigFromString(). 
Usually this does not happens since we trim for " \t\r\n", but if there
are other chars that return true with isspace(), we may end with an
empty argv. Better to handle the condition in an explicit way.
 2013-03-06 12:43:49 +01:00
antirez 00a5e86e44 redis-cli: use sdsfreesplitres() instead of hand-coding it. 2013-03-06 12:43:45 +01:00
antirez f72a3406a7 sds.c: sdssplitargs_free() removed as it was a duplicate. 2013-03-06 12:43:41 +01:00
antirez 68532cf01e More specific error message in loadServerConfigFromString(). 2013-03-06 12:43:37 +01:00
antirez a70e3ffb16 sdssplitargs(): on error set *argc to 0. 
This makes programs not checking the return value for NULL much safer
since with this change:

1) It is still possible to iterate the zero-length result without
crashes.
2) sdssplitargs_free will work against NULL and 0 count.
 2013-03-06 12:43:32 +01:00
antirez a6f557657c sdssplitargs(): now returns NULL only on error. 
An empty input string also resulted into the function returning NULL
making it harder for the caller to distinguish between error and empty
string without checking the original input string length.
 2013-03-06 12:43:27 +01:00
charsyam 5aeb1f71a3 Don't segfault on unbalanced quotes. 2013-03-06 12:43:22 +01:00
antirez aadb8aaa08 Allow AUTH while loading the DB in memory. 
While Redis is loading the AOF or RDB file in memory only a subset of
commands are allowed. This commit adds AUTH to this subset.
 2013-03-06 11:51:23 +01:00
antirez df3b492045 redis-cli: use keepalive socket option. 
This should improve things in two ways:

1) Prevent timeouts caused by the execution of long commands.
2) Improve detection of real connection errors.

This is mostly effective only on Linux because of the bogus default
keepalive settings. In Linux we have OS-specific calls to set the
keepalive interval to reasonable values.
 2013-03-04 11:16:21 +01:00
0x20h e618eb4fb3 suppress external diff program when using git diff. 2013-03-04 10:58:02 +01:00
Stam He a74056f1e0 point 2 of slave-serve-stale-data miss '-' between 'stale' and 'data' 2013-03-04 10:49:19 +01:00
antirez 02cd545f7d A comment in main() clarified. 2013-02-27 12:47:51 +01:00
antirez b148513fec Remove warning when printing redisBuildId(). 2013-02-27 12:47:38 +01:00
antirez 5e9e96999f Remove too agressive/spamming log in rdb.c. 2013-02-27 12:47:33 +01:00
antirez d2a37badc2 Use GCC printf format attribute for redisLog(). 
This commit also fixes redisLog() statements producing warnings.
 2013-02-27 12:47:16 +01:00
antirez 0d8d7f106a Better panic message for failed time event creation. 2013-02-27 12:00:51 +01:00
Stam He 438dfdd13d add a check for aeCreateTimeEvent 
1) Add a check for aeCreateTimeEvent in function initServer.
 2013-02-27 12:00:48 +01:00
Stam He 6f96ac1c1c Set proctitle: avoid the use of __attribute__((constructor)). 
This cased a segfault in some Linux system and was GCC-specific.

Commit modified by @antirez:

1) Stripped away the part to set the proc title via config for now.
2) Handle initialization of setproctitle only when the replacement
   is used.
3) Don't require GCC now that the attribute constructor is no
   longer used.
 2013-02-27 12:00:43 +01:00
antirez f69b0a0db8 setproctitle.c: declar tmp as static so valgrind will not detect a leak. 2013-02-26 15:51:15 +01:00
antirez ac3100bc3b Set process name in ps output to make operations safer. 
This commit allows Redis to set a process name that includes the binding
address and the port number in order to make operations simpler.

Redis children processes doing AOF rewrites or RDB saving change the
name into redis-aof-rewrite and redis-rdb-bgsave respectively.

This in general makes harder to kill the wrong process because of an
error and makes simpler to identify saving children.

This feature was suggested by Arnaud GRANAL in the Redis Google Group,
Arnaud also pointed me to the setproctitle.c implementation includeed in
this commit.

This feature should work on all the Linux, OSX, and all the three major
BSD systems.
 2013-02-26 12:03:48 +01:00
Arnaud Granal fedcd51185 Fix error "repl-backlog-size must be 1 or greater" 
The parameter repl-backlog-size is not parsed correctly in the configuration file. argv[0] is parsed instead of argv[1].
 2013-02-25 18:38:48 +01:00
antirez ff56772115 PSYNC: another change to unexpected reply from PSYNC. 2013-02-13 18:43:45 +01:00
antirez 964ee00de0 PSYNC: More robust handling of unexpected reply to PSYNC. 2013-02-13 18:34:18 +01:00
antirez 3b4c0d80d5 Avoid compiler warning by casting to match printf() specifier. 2013-02-13 13:38:32 +01:00
antirez d96a66f531 Replication: more strict error checking for master PING reply. 2013-02-12 16:59:27 +01:00
antirez 31f0a6ec50 Replication: added new stats counting full and partial resynchronizations. 2013-02-12 16:26:42 +01:00
antirez 5fe2577a19 Return a specific NOAUTH error if authentication is required. 2013-02-12 16:25:47 +01:00
antirez 9854252798 Test: avoid false positives in CLIENT SETNAME closed connection test. 2013-02-12 13:27:30 +01:00
Steven Penny 5a8cf340a4 Format to fit 80 columns 
This makes it readable on GitHub and editors without auto wrapping.
 2013-02-12 13:11:44 +01:00
antirez 0ae1a6b1c3 Add missing bracket removed for error after rebase of PSYNC. 2013-02-12 12:58:07 +01:00
antirez 2b1398e870 PSYNC: debugging printf() calls are now logs at DEBUG level. 2013-02-12 12:58:03 +01:00
antirez 5a4ef3e5b6 Remove harmless warning in slaveTryPartialResynchronization(). 2013-02-12 12:57:59 +01:00
antirez 4fe7672dfc PSYNC: don't use the client buffer to send +CONTINUE and +FULLRESYNC. 
When we are preparing an handshake with the slave we can't touch the
connection buffer as it'll be used to accumulate differences between
the sent RDB file and what arrives next from clients.

So in short we can't use addReply() family functions.

However we just use write(2) because we know that the socket buffer is
empty, since a prerequisite for SYNC to work is that the static buffer
and the output list are empty, and in general it is not expected that a
client SYNCs after doing some heavy I/O with the master.

However a short write connection is explicitly handled to avoid
fragility (we simply close the connection and the slave will retry).
 2013-02-12 12:57:56 +01:00
antirez 8529dd218f SYNC not allowed with pending data on the static output buffer. 2013-02-12 12:57:52 +01:00
antirez 6726ea5b53 Log the unexpected string received in place of the SYNC payload length. 2013-02-12 12:57:48 +01:00
antirez f5edd535d1 After SLAVEOF <newslave> don't allow chained slaves to PSYNC. 2013-02-12 12:57:44 +01:00
antirez 700e5eb4fc PSYNC: work in progress, preview #2, rebased to unstable. 2013-02-12 12:57:40 +01:00
antirez 4d8655cfd3 Use the new unified protocol to send SELECT to slaves. 
SELECT was still transmitted to slaves using the inline protocol, that
is conceived mostly for humans to type into telnet sessions, and is
notably not understood by redis-cli --slave.

Now the new protocol is used instead.
 2013-02-12 12:57:36 +01:00
antirez 01c21f9943 Use replicationFeedSlaves() to send PING to slaves. 
A Redis master sends PING commands to slaves from time to time: doing
this ensures that even if absence of writes, the master->slave channel
remains active and the slave can feel the master presence, instead of
closing the connection for timeout.

This commit changes the way PINGs are sent to slaves in order to use the
standard interface used to replicate all the other commands, that is,
the function replicationFeedSlaves().

With this change the stream of commands sent to every slave is exactly
the same regardless of their exact state (Transferring RDB for first
synchronization or slave already online). With the previous
implementation the PING was only sent to online slaves, with the result
that the output stream from master to slaves was not identical for all
the slaves: this is a problem if we want to implement partial resyncs in
the future using a global replication stream offset.

TL;DR: this commit should not change the behaviour in practical terms,
but is just something in preparation for partial resynchronization
support.
 2013-02-12 12:57:31 +01:00
antirez 5a35e485f9 Emit SELECT to slaves in a centralized way. 
Before this commit every Redis slave had its own selected database ID
state. This was not actually useful as the emitted stream of commands
is identical for all the slaves.

Now the the currently selected database is a global state that is set to
-1 when a new slave is attached, in order to force the SELECT command to
be re-emitted for all the slaves.

This change is useful in order to implement replication partial
resynchronization in the future, as makes sure that the stream of
commands received by slaves, including SELECT commands, are exactly the
same for every slave connected, at any time.

In this way we could have a global offset that can identify a specific
piece of the master -> slaves stream of commands.
 2013-02-12 12:57:26 +01:00
antirez c970816e4d Makefile: valgrind target added (forces -O0 and libc malloc). 2013-02-11 12:11:26 +01:00
antirez 8a22934c68 TCP keep-alive. Better documentation in redis.conf. 2013-02-11 11:44:35 +01:00
antirez b06b90b5d8 Tcp keep-alive: send three probes before detectin an error. 
Otherwise we end with less reliable connections because it's too easy
that a single packet gets lost.
 2013-02-11 11:44:31 +01:00
antirez 0a14a6542c tcp-keepalive option documented in redis.conf. 2013-02-11 11:44:27 +01:00
antirez 2ed3fc1502 Set SO_KEEPALIVE on client sockets if configured to do so. 2013-02-11 11:44:23 +01:00
antirez f2817cbd9e Add SO_KEEPALIVE support to anet.c. 2013-02-11 11:44:18 +01:00
antirez fcfdbda104 Sentinel: advertise the promoted slave address only after successful setup. 2013-02-11 11:44:14 +01:00
Pierre Chapuis 0d470b4c43 fix comments forgotten in #285 (zipmap -> ziplist) 2013-02-11 11:44:06 +01:00
antirez cc55a4525a Make all WATCHers dirty when the slave reloads the DB. 2013-02-08 10:27:21 +01:00
antirez 5eb0ec142d LASTSAVE is a "random" command. 2013-02-07 19:13:06 +01:00
antirez 5f7dff4d16 TCP_NODELAY after SYNC: changes to the implementation. 2013-02-05 12:05:39 +01:00
charsyam 1d80acae54 Turn off TCP_NODELAY on the slave socket after SYNC. 
Further details from @antirez:

It was reported by @StopForumSpam on Twitter that the Redis replication
link was strangely using multiple TCP packets for multiple commands.
This wastes a lot of bandwidth and is due to the TCP_NODELAY option we
enable on the socket after accepting a new connection.

However the master -> slave channel is a one-way channel since Redis
replication is asynchronous, so there is no point in trying to reduce
the latency, we should aim to reduce the bandwidth. For this reason this
commit introduces the ability to disable the nagle algorithm on the
socket after a successful SYNC.

This feature is off by default because the delay can be up to 40
milliseconds with normally configured Linux kernels.
 2013-02-05 12:05:24 +01:00
antirez 14daa01c52 Test: No clients timeout while testing. 2013-02-05 12:02:41 +01:00
Johan Bergström d2a8bca82b Use info nameofexectuable to find current executable 2013-02-05 11:44:55 +01:00
Johan Bergström 1124cda37a Enforce tcl 8.5 or newer 2013-02-05 11:44:51 +01:00
Johan Bergström e92edd3cee Check available tcl versions 2013-02-05 11:44:44 +01:00
Rock Li 2543fa6476 retval doesn't initalized 
If each if conditions are all fail, variable retval will under uninitlized
 2013-02-05 10:49:24 +01:00
David Celis 44d0eac568 Fix a few typos and improve grammar of redis.conf 
Make several edits to the example redis.conf configuration file for
improved flow and grammar.

Signed-off-by: David Celis <me@davidcel.is>
 2013-02-04 10:34:05 +01:00
Gengliang Wang 57e411c687 Fix a bug in srandmemberWithCountCommand() 
In CASE 2, the call sunionDiffGenericCommand will involve the string "srandmember"
> sadd foo one
(integer 1)
> sadd srandmember two
(integer 2)
> srandmember foo 3
1)"one"
2)"two"
 2013-02-04 10:33:21 +01:00
antirez 8024de7eb0 Z*STORE event fixed: generate del only if resulting sorted set is empty. 2013-01-29 13:50:09 +01:00
antirez ca8e7d4f8a Generate del events when S*STORE commands delete the destination key. 2013-01-29 13:43:24 +01:00
antirez 664de412af Tests for keyspace notifications. 2013-01-28 13:19:04 +01:00
antirez d3cf778d82 Send 'expired' events when a key expires by lookup. 2013-01-28 13:18:57 +01:00
antirez 9db8fbcfdd Keyspace notifications: fixed a leak and a bug introduced in the latest commit. 2013-01-28 13:18:53 +01:00
antirez b9bc4f9132 Keyspace events: it is now possible to select subclasses of events. 
When keyspace events are enabled, the overhead is not sever but
noticeable, so this commit introduces the ability to select subclasses
of events in order to avoid to generate events the user is not
interested in.

The events can be selected using redis.conf or CONFIG SET / GET.
 2013-01-28 13:18:36 +01:00
antirez 6de052cd77 decrRefCount -> decrRefCountVoid in list constructor. 2013-01-28 13:17:37 +01:00
antirez 81af92b023 Enable keyspace events notification when testing. 2013-01-28 13:17:33 +01:00
antirez 4ff138fbbe Keyspace events added for more commands. 2013-01-28 13:17:30 +01:00
antirez 2825f21fd8 Fix decrRefCount() prototype from void to robj pointer. 
decrRefCount used to get its argument as a void* pointer in order to be
used as destructor where a 'void free_object(void*)' prototype is
expected. However this made simpler to introduce bugs by freeing the
wrong pointer. This commit fixes the argument type and introduces a new
wrapper called decrRefCountVoid() that can be used when the void*
argument is needed.
 2013-01-28 13:17:26 +01:00
antirez a64383c861 notifyKeyspaceEvent(): release channel names using the right pointers. 2013-01-28 13:17:22 +01:00
antirez fdfb59beae Initial test events for the new keyspace notification API. 2013-01-28 13:17:18 +01:00
antirez dd2ce743ca Fixed over-80-cols comment in db.c 2013-01-28 13:17:15 +01:00
antirez 6e64525cc2 Two fixes to initial keyspace notifications API. 2013-01-28 13:17:11 +01:00
antirez 212edbc409 Keyspace events notification API. 2013-01-28 13:17:00 +01:00
antirez 153766e137 Lua struct library updated to version 0.2. 
There was a bug in the previous version of this library that caused a
crash under the circumstances described in issue #901.

The newer version of the library appears to be fixed (I tested it
manually with valgrind and everything seems fine now).

For more information about this library please visit this web site:

    http://www.inf.puc-rio.br/~roberto/struct/
 2013-01-23 10:53:31 +01:00
antirez 767a53aa84 redis-cli --bigkeys output is now simpler to understand. 2013-01-21 19:16:11 +01:00
antirez 3ff75e58e8 UNSUBSCRIBE and PUNSUBSCRIBE: always provide a reply. 
UNSUBSCRIBE and PUNSUBSCRIBE commands are designed to mass-unsubscribe
the client respectively all the channels and patters if called without
arguments.

However when these functions are called without arguments, but there are
no channels or patters we are subscribed to, the old behavior was to
don't reply at all.

This behavior is broken, as every command should always reply.
Also it is possible that we are no longer subscribed to a channels but we
are subscribed to patters or the other way around, and the client should
be notified with the correct number of subscriptions.

Also it is not pretty that sometimes we did not receive a reply at all
in a redis-cli session from these commands, blocking redis-cli trying
to read the reply.

This fixes issue #714.
 2013-01-21 19:02:24 +01:00
antirez 850117a8a5 Fixed a bug in memtest progress bar, that had no actual effects. 
This closes issue #859, thanks to @erbenmo.
 2013-01-21 12:38:46 +01:00
Bilal Husain 2f1318ab29 s/adiacent/adjacent/ 
fixed typo in a comment (step 2 memcheck)
 2013-01-21 12:20:07 +01:00
antirez f6f43d7da0 Not every __sun has backtrace(). 
I don't know how to test for Open Solaris that has support for
backtrace() so for now removing the #ifdef that breaks compilation under
other Solaris flavors.
 2013-01-21 12:07:48 +01:00
antirez 635c532c89 Additionally two typos fixed thanks to @jodal 2013-01-19 13:46:22 +01:00
antirez 39f0a33f78 Whitelist SIGUSR1 to avoid auto-triggering errors. 
This commit fixes issue #875 that was caused by the following events:

1) There is an active child doing BGSAVE.
2) flushall is called (or any other condition that makes Redis killing
the saving child process).
3) An error is sensed by Redis as the child exited with an error (killed
by a singal), that stops accepting write commands until a BGSAVE happens
to be executed with success.

Whitelisting SIGUSR1 and making sure Redis always uses this signal in
order to kill its own children fixes the issue.
 2013-01-19 13:30:46 +01:00
antirez 1e20c939fe Clear server.shutdown_asap on failed shutdown. 
When a SIGTERM is received Redis schedules a shutdown. However if it
fails to perform the shutdown it must be clear the shutdown_asap flag
otehrwise it will try again and again possibly making the server
unusable.
 2013-01-19 13:20:56 +01:00
antirez d766907cfb Slowlog: don't log EXEC but just the executed commands. 
The Redis Slow Log always used to log the slow commands executed inside
a MULTI/EXEC block. However also EXEC was logged at the end, which is
perfectly useless.

Now EXEC is no longer logged and a test was added to test this behavior.

This fixes issue #759.
 2013-01-19 12:55:15 +01:00
guiquanz 1caf09399e Fixed many typos. 
Conflicts fixed, mainly because 2.8 has no cluster support / files:
	00-RELEASENOTES
	src/cluster.c
	src/crc16.c
	src/redis-trib.rb
	src/redis.h
 2013-01-19 11:03:19 +01:00
charsyam ff1e4d22a6 redis-cli prompt bug fix 2013-01-19 10:34:22 +01:00
Jan-Erik Rediger a8f9cec16a Always exit if connection fails. 
This avoids unnecessary core dumps. Fixes antirez/redis#894
 2013-01-19 10:29:42 +01:00
bitterb aadcda9977 Fix an error reply for CLIENT command 2013-01-19 10:26:32 +01:00
Nathan Parry f2bc198d40 redis-cli --rdb fails if server sends a ping 
Redis pings slaves in "pre-synchronization stage" with newlines. (See
https://github.com/antirez/redis/blob/2.6.9/src/replication.c#L814)
However, redis-cli does not expect this - it sees the newline as the end
of the bulk length line, and ends up returning 0 as bulk the length.
This manifests as the following when running redis-cli:

    $ ./src/redis-cli --rdb some_file
    SYNC sent to master, writing 0 bytes to 'some_file'
    Transfer finished with success.

With this commit, we just ignore leading newlines while reading the bulk
length line.

To reproduce the problem, load enough data into Redis so that the
preparation of the RDB snapshot takes long enough for a ping to occur
while redis-cli is waiting for the data.
 2013-01-18 11:50:06 +01:00
antirez c48584b85d redis-cli: save an RDB dump from remote server to local file. 2013-01-16 19:48:30 +01:00
antirez 7897e0a0d6 Tests for CLIENT GETNAME/SETNAME. 2013-01-15 13:34:26 +01:00
antirez f31d10b9a6 Typo fixed, ASCI -> ASCII. 2013-01-15 13:34:24 +01:00
antirez 786bd3938e CLIENT GETNAME and CLIENT SETNAME introduced. 
Sometimes it is much simpler to debug complex Redis installations if it
is possible to assign clients a name that is displayed in the CLIENT
LIST output.

This is the case, for example, for "leaked" connections. The ability to
provide a name to the client makes it quite trivial to understand what
is the part of the code implementing the client not releasing the
resources appropriately.

Behavior:

    CLIENT SETNAME: set a name for the client, or remove the current
                    name if an empty name is set.
    CLIENT GETNAME: get the current name, or a nil.
    CLIENT LIST: now displays the client name if any.

Thanks to Mark Gravell for pushing this idea forward.
 2013-01-15 13:34:22 +01:00
antirez aa9497fdb3 Undo slave-master handshake when SLAVEOF sets a new slave. 
Issue #828 shows how Redis was not correctly undoing a non-blocking
connection attempt with the previous master when the master was set to a
new address using the SLAVEOF command.

This was also a result of lack of refactoring, so now there is a
function to cancel the non blocking handshake with the master.
The new function is now used when SLAVEOF NO ONE is called or when
SLAVEOF is used to set the master to a different address.
 2013-01-15 13:33:27 +01:00
antirez 64ee9c861f Makefile.dep updated. 2013-01-11 23:51:12 +01:00
antirez 8a70007f9e Comment in the call() function clarified a bit. 2013-01-10 12:04:24 +01:00
antirez e2e36cf19b Test: added regression for issue #872. 2013-01-10 12:04:19 +01:00
antirez 414e06653c Multiple fixes for EVAL (issue #872). 
1) The event handler was no restored after a timeout condition if the
   command was eventually executed with success.
2) The command was not converted to EVAL in case of errors in the middle
   of the execution.
3) Terrible duplication of code without any apparent reason.
 2013-01-10 12:04:15 +01:00
antirez cefff3a1b3 Better error reporting when fd event creation fails. 2013-01-03 14:32:32 +01:00
antirez 6570d2f89e ae.c: set errno when error is not a failing syscall. 
In this way the caller is able to perform better error checking or to
use strerror() without the risk of meaningless error messages being
displayed.
 2013-01-03 14:32:27 +01:00
antirez 858b5576ac Fix overflow in mstime() in redis-cli and benchmark. 
The problem does not exist in the Redis server implementation of mstime()
but is only limited to redis-cli and redis-benchmark.

Thix fixes issue #839.
 2012-12-20 15:21:40 +01:00
antirez 61fb0aacf5 CONTRIBUTING updated with request to add BSD license. 2012-12-17 11:17:55 +01:00
antirez 6fa2610ac9 Added missing license and copyright in deps/hiredis. 2012-12-17 11:11:36 +01:00
antirez a6d117b6c0 serverCron() frequency is now a runtime parameter (was REDIS_HZ). 
REDIS_HZ is the frequency our serverCron() function is called with.
A more frequent call to this function results into less latency when the
server is trying to handle very expansive background operations like
mass expires of a lot of keys at the same time.

Redis 2.4 used to have an HZ of 10. This was good enough with almost
every setup, but the incremental key expiration algorithm was working a
bit better under *extreme* pressure when HZ was set to 100 for Redis
2.6.

However for most users a latency spike of 30 milliseconds when million
of keys are expiring at the same time is acceptable, on the other hand a
default HZ of 100 in Redis 2.6 was causing idle instances to use some
CPU time compared to Redis 2.4. The CPU usage was in the order of 0.3%
for an idle instance, however this is a shame as more energy is consumed
by the server, if not important resources.

This commit introduces HZ as a runtime parameter, that can be queried by
INFO or CONFIG GET, and can be modified with CONFIG SET. At the same
time the default frequency is set back to 10.

In this way we default to a sane value of 10, but allows users to
easily switch to values up to 500 for near real-time applications if
needed and if they are willing to pay this small CPU usage penalty.
 2012-12-14 17:20:21 +01:00
Patrick TJ McPhee 5207c03c49 Define _XOPEN_SOURCE appropriately on NetBSD. 2012-12-12 18:42:41 +01:00
antirez bbee226ecd Fix config.h endianess detection to work on Linux / PPC64. 
Config.h performs endianess detection including OS-specific headers to
define the endianess macros, or when this is not possible, checking the
processor type via ifdefs.

Sometimes when the OS-specific macro is included, only __BYTE_ORDER is
defined, while BYTE_ORDER remains undefined. There is code at the end of
config.h endianess detection in order to define the macros without the
underscore, but it was not working correctly.

This commit fixes endianess detection fixing Redis on Linux / PPC64 and
possibly other systems.
 2012-12-11 17:19:54 +01:00
Brian J. McManus 17e31eb03a Issue 804 Add Default-Start and Default-Stop LSB tags for RedHat startup and update-rc.d compatability. 2012-12-03 12:31:12 +01:00
antirez 124cb6dda1 Memory leak fixed: release client's bpop->keys dictionary. 
Refactoring performed after issue #801 resolution (see commit
2f87cf8b01) introduced a memory leak that
is fixed by this commit.

I simply forgot to free the new allocated dictionary in the client
structure trusting the output of "make test" on OSX.

However due to changes in the "leaks" utility the test was no longer
testing memory leaks. This problem was also fixed.

Fortunately the CI test running at ci.redis.io spotted the bug in the
valgrind run.

The leak never ended into a stable release.
 2012-12-03 12:17:33 +01:00
antirez e39735214d Test: fixed osx "leaks" support in test. 
Due to changes in recent releases of osx leaks utility, the osx leak
detection no longer worked. Now it is fixed in a way that should be
backward compatible.
 2012-12-03 12:17:29 +01:00
antirez 07a9f85405 Blocking POP: use a dictionary to store keys clinet side. 
To store the keys we block for during a blocking pop operation, in the
case the client is blocked for more data to arrive, we used a simple
linear array of redis objects, in the blockingState structure:

    robj **keys;
    int count;

However in order to fix issue #801 we also use a dictionary in order to
avoid to end in the blocked clients queue for the same key multiple
times with the same client.

The dictionary was only temporary, just to avoid duplicates, but since
we create / destroy it there is no point in doing this duplicated work,
so this commit simply use a dictionary as the main structure to store
the keys we are blocked for. So instead of the previous fields we now
just have:

    dict *keys;

This simplifies the code and reduces the work done by the server during
a blocking POP operation.
 2012-12-02 20:42:59 +01:00
antirez fbf3e33d18 Test: regression for issue #801. 2012-12-02 20:42:56 +01:00
antirez 64e69c0d5f Client should not block multiple times on the same key. 
Sending a command like:

BLPOP foo foo foo foo 0

Resulted into a crash before this commit since the client ended being
inserted in the waiting list for this key multiple times.
This resulted into the function handleClientsBlockedOnLists() to fail
because we have code like that:

    if (de) {
        list *clients = dictGetVal(de);
        int numclients = listLength(clients);

        while(numclients--) {
            listNode *clientnode = listFirst(clients);

            /* server clients here... */
        }
    }

The code to serve clients used to remove the served client from the
waiting list, so if a client is blocking multiple times, eventually the
call to listFirst() will return NULL or worse will access random memory
since the list may no longer exist as it is removed by the function
unblockClientWaitingData() if there are no more clients waiting for this
list.

To avoid making the rest of the implementation more complex, this commit
modifies blockForKeys() so that a client will be put just a single time
into the waiting list for a given key.

Since it is Saturday, I hope this fixes issue #801.
 2012-12-02 20:42:51 +01:00
antirez c1eda786d6 SDIFF fuzz test added. 2012-11-30 17:14:08 +01:00
antirez ccc974d90b SDIFF is now able to select between two algorithms for speed. 
SDIFF used an algorithm that was O(N) where N is the total number
of elements of all the sets involved in the operation.

The algorithm worked like that:

ALGORITHM 1:

1) For the first set, add all the members to an auxiliary set.
2) For all the other sets, remove all the members of the set from the
auxiliary set.

So it is an O(N) algorithm where N is the total number of elements in
all the sets involved in the diff operation.

Cristobal Viedma suggested to modify the algorithm to the following:

ALGORITHM 2:

1) Iterate all the elements of the first set.
2) For every element, check if the element also exists in all the other
remaining sets.
3) Add the element to the auxiliary set only if it does not exist in any
of the other sets.

The complexity of this algorithm on the worst case is O(N*M) where N is
the size of the first set and M the total number of sets involved in the
operation.

However when there are elements in common, with this algorithm we stop
the computation for a given element as long as we find a duplicated
element into another set.

I (antirez) added an additional step to algorithm 2 to make it faster,
that is to sort the set to subtract from the biggest to the
smallest, so that it is more likely to find a duplicate in a larger sets
that are checked before the smaller ones.

WHAT IS BETTER?

None of course, for instance if the first set is much larger than the
other sets the second algorithm does a lot more work compared to the
first algorithm.

Similarly if the first set is much smaller than the other sets, the
original algorithm will less work.

So this commit makes Redis able to guess the number of operations
required by each algorithm, and select the best at runtime according
to the input received.

However, since the second algorithm has better constant times and can do
less work if there are duplicated elements, an advantage is given to the
second algorithm.
 2012-11-30 17:13:34 +01:00
antirez 2572bb13c6 redis-benchmark: seed the PRNG with time() at startup. 2012-11-30 17:13:25 +01:00
antirez e34c14df1a Make an EXEC test more latency proof. 2012-11-29 16:12:23 +01:00
antirez 3b71404d70 Introduced the Build ID in INFO and --version output. 
The idea is to be able to identify a build in a unique way, so for
instance after a bug report we can recognize that the build is the one
of a popular Linux distribution and perform the debugging in the same
environment.
 2012-11-29 14:22:15 +01:00
antirez cd99c14e38 On crash memory test rewrote so that it actaully works. 
1) We no longer test location by location, otherwise the CPU write cache
completely makes our business useless.
2) We still need a memory test that operates in steps from the first to
the last location in order to never hit the cache, but that is still
able to retain the memory content.

This was tested using a Linux box containing a bad memory module with a
zingle bit error (always zero).

So the final solution does has an error propagation step that is:

1) Invert bits at every location.
2) Swap adiacent locations.
3) Swap adiacent locations again.
4) Invert bits at every location.
5) Swap adiacent locations.
6) Swap adiacent locations again.

Before and after these steps, and after step 4, a CRC64 checksum is computed.
If the three CRC64 checksums don't match, a memory error was detected.
 2012-11-29 10:24:41 +01:00
antirez 053f1f357d Jemalloc updated to version 3.2.0. 2012-11-28 18:42:35 +01:00
charsyam 1f0d0b094a Remove unnecessary condition in _dictExpandIfNeeded (dict.c) 2012-11-28 11:45:52 +01:00
Matt Arsenault ccb500cf36 It's a watchdog, not a watchdong. 2012-11-28 11:42:24 +01:00
charsyam ed48550fe6 remove compile warning bioKillThreads 2012-11-28 11:42:20 +01:00
antirez d0570c9693 EVALSHA is now case insensitive. 
EVALSHA used to crash if the SHA1 was not lowercase (Issue #783).
Fixed using a case insensitive dictionary type for the sha -> script
map used for replication of scripts.
 2012-11-22 15:51:03 +01:00
antirez cc85ed41df Fix integer overflow in zunionInterGenericCommand(). 
This fixes issue #761.
 2012-11-22 15:28:48 +01:00
antirez ca9321afe2 Test: MULTI state is cleared after EXECABORT error. 2012-11-22 10:36:06 +01:00
antirez a527235f33 Test: make sure EXEC fails after previous transaction errors. 2012-11-22 10:36:01 +01:00
antirez 53600e34db Test: MULTI/EXEC tests moved into multi.tcl. 2012-11-22 10:35:50 +01:00
antirez 60f9dac672 Safer handling of MULTI/EXEC on errors. 
After the transcation starts with a MULIT, the previous behavior was to
return an error on problems such as maxmemory limit reached. But still
to execute the transaction with the subset of queued commands on EXEC.

While it is true that the client was able to check for errors
distinguish QUEUED by an error reply, MULTI/EXEC in most client
implementations uses pipelining for speed, so all the commands and EXEC
are sent without caring about replies.

With this change:

1) EXEC fails if at least one command was not queued because of an
error. The EXECABORT error is used.
2) A generic error is always reported on EXEC.
3) The client DISCARDs the MULTI state after a failed EXEC, otherwise
pipelining multiple transactions would be basically impossible:
After a failed EXEC the next transaction would be simply queued as
the tail of the previous transaction.
 2012-11-22 10:35:47 +01:00
antirez dff1ec4a36 Make bio.c threads killable ASAP if needed. 
We use this new bio.c feature in order to stop our I/O threads if there
is a memory test to do on crash. In this case we don't want anything
else than the main thread to run, otherwise the other threads may mess
with the heap and the memory test will report a false positive.
 2012-11-22 10:28:53 +01:00
antirez d4b5c4803b Fast memory test on Redis crash. 2012-11-22 10:28:48 +01:00
antirez dd905417ec Use more fine grained HAVE macros instead of HAVE_PROCFS. 2012-11-22 10:28:43 +01:00
antirez c342b075df Children creating AOF or RDB files now report memory used by COW. 
Finally Redis is able to report the amount of memory used by
copy-on-write while saving an RDB or writing an AOF file in background.

Note that this information is currently only logged (at NOTICE level)
and not shown in INFO because this is less trivial (but surely doable
with some minor form of interprocess communication).

The reason we can't capture this information on the parent before we
call wait3() is that the Linux kernel will release the child memory
ASAP, and only retain the minimal state for the process that is useful
to report the child termination to the parent.

The COW size is obtained by summing all the Private_Dirty fields found
in the "smap" file inside the proc filesystem for the process.

All this is Linux specific and is not available on other systems.
 2012-11-19 15:26:02 +01:00
antirez 1d8af6074c zmalloc_get_private_dirty() function added (Linux only). 
For non Linux systmes it just returns 0.

This function is useful to estimate copy-on-write because of childs
saving stuff on disk.
 2012-11-19 15:25:57 +01:00
antirez 1a706c729a zmalloc: kill unused __size parameter in update_zmalloc_stat_alloc() macro. 2012-11-14 13:07:54 +01:00
antirez 50c41de752 TTL API change: TTL returns -2 for non existing keys. 
The previous behavior was to return -1 if:

1) Existing key but without an expire set.
2) Non existing key.

Now the second case is handled in a different, and TTL will return -2
if the key does not exist at all.

PTTL follows the same behavior as well.
 2012-11-12 23:05:56 +01:00
antirez 388cc20418 Copyright date fixed in COPYING file. 2012-11-08 19:14:37 +01:00
antirez cda5c0ba06 Make clear that contributing code to the Redis project means to release it under the terms of the BSD license. 2012-11-08 18:43:41 +01:00
antirez 8ddb23b90c BSD license added to every C source and header file. 2012-11-08 18:34:04 +01:00
antirez 46c5d39660 Type mismatch errors are now prefixed with WRONGTYPE. 
So instead to reply with a generic error like:

-ERR ... wrong kind of value ...

now it replies with:

-WRONGTYPE ... wrong kind of value ...

This makes this particular error easy to check without resorting to
(fragile) pattern matching of the error string (however the error string
used to be consistent already).

Client libraries should return a specific exeption type for this error.

Most of the commit is about fixing unit tests.
 2012-11-06 20:28:15 +01:00
Runzhen Wang 2b28ef0e61 fix a typo in redis.h line 595 comment 2012-11-02 12:12:08 +01:00
antirez 64be5e365a More robust handling of AOF rewrite child. 
After the wait3() syscall we used to do something like that:

    if (pid == server.rdb_child_pid) {
        backgroundSaveDoneHandler(exitcode,bysignal);
    } else {
        ....
    }

So the AOF rewrite was handled in the else branch without actually
checking if the pid really matches. This commit makes the check explicit
and logs at WARNING level if the pid returned by wait3() does not match
neither the RDB or AOF rewrite child.
 2012-11-01 22:41:54 +01:00
Yecheng Fu fecc8797c2 fix typo in comments (redis.c, networking.c) 2012-11-01 22:26:49 +01:00
antirez d775b35715 Unix socket clients properly displayed in MONITOR and CLIENT LIST. 
This also fixes issue #745.
 2012-11-01 22:12:50 +01:00
antirez 655e6838fb 32 bit build fixed on Linux. 
It failed because of the way jemalloc was compiled (without passing the
right flags to make, but just to configure). Now the same set of flags
are also passed to the make command, fixing the issue.

This fixes issue #744
 2012-11-01 15:40:46 +01:00
YAMAMOTO Takashi 428b706308 fix a typo in a comment 2012-10-31 09:30:00 +01:00
antirez 9fab63b334 Invert two sides of if expression in SET to avoid a lookup. 
Because of the short circuit behavior of && inverting the two sides of
the if expression avoids an hash table lookup if the non-EX variant of
SET is called.

Thanks to Weibin Yao (@yaoweibin on github) for spotting this.
 2012-10-31 09:27:46 +01:00
antirez 537235bbbb No longer used macro rdbIsOpcode() removed. 2012-10-30 19:12:01 +01:00
antirez 38272f961a help.h update (adds bitop, bitcount, evalsha...) 2012-10-30 19:11:57 +01:00
antirez 6c7e957d17 Ctrl+w support in linenoise. 2012-10-30 19:11:46 +01:00
antirez 6bd105032c Marginally more robust glibc version test for sync_file_range detection. 2012-10-30 19:11:41 +01:00
charsyam 72f42cf40c patch config.h for sync_file_range 2012-10-30 19:11:34 +01:00
antirez 2a49aa1da2 Fix compilation on Linux kernels or glibc versions lacking sync_file_range(). 
This fixes issue #667.

Many thanks to Didier Spezia for the fix.
 2012-10-30 19:11:29 +01:00
antirez 24ee20e5ca Update memory peak stats while loading RDB / AOF. 2012-10-30 19:11:23 +01:00
antirez f7b42b1749 2.7.0 branch created as a fork of 2.6.0. 2012-10-22 23:32:26 +02:00
406 changed files with 81733 additions and 18883 deletions
Expand all files Collapse all files
.gitignore
+7 -29
View File
@@ -1,8 +1,13 @@
.*.swp
*.o
*.rdb
*.log
redis-*
dump.rdb
redis-benchmark
redis-check-aof
redis-check-dump
redis-cli
redis-sentinel
redis-server
doc-tools
release
misc/*
@@ -20,30 +25,3 @@ deps/lua/src/liblua.a
.make-*
.prerequisites
*.dSYM
*.user
*.exe
*.sdf
*.suo
msvs/Debug/
msvs/Release/
msvs/RedisBenchmark/Debug/
msvs/RedisBenchmark/Release/
msvs/RedisBenchmark/x64/
msvs/RedisCheckAof/Debug/
msvs/RedisCheckAof/Release/
msvs/RedisCheckAof/x64/
msvs/RedisCheckDump/Debug/
msvs/RedisCheckDump/Release/
msvs/RedisCheckDump/x64/
msvs/RedisCli/Debug/
msvs/RedisCli/Release/
msvs/RedisCli/x64/
msvs/hiredis/Debug/
msvs/hiredis/Release/
msvs/hiredis/x64/
msvs/lua/lua/Debug/
msvs/lua/lua/Release/
msvs/lua/lua/x64/
msvs/x64/
msvs/ipch
msvs/RedisServer.opensdf
00-RELEASENOTES
+633 -234
View File
@@ -1,9 +1,9 @@
Redis 2.6 release notes
Redis 2.8 release notes
=======================
** IMPORTANT ** Check the 'Migrating from 2.4 to 2.6' section at the end of
this file for information about what changed between 2.4 and
2.6 and how this may affect your application.
** IMPORTANT ** Check the 'Migrating from 2.6 to 2.8' section at the end of
this file for information about what changed between 2.6 and
2.8 and how this may affect your application.
--------------------------------------------------------------------------------
Upgrade urgency levels:
@@ -14,298 +14,697 @@ HIGH: There is a critical bug that may affect a subset of users. Upgrade!
CRITICAL: There is a critical bug affecting MOST USERS. Upgrade ASAP.
--------------------------------------------------------------------------------
--[ Redis 2.6.8 ]
--[ Redis 2.8.23 ] Release date: 15 Oct 2015
UPGRADE URGENCY: MODERATE if you use Lua scripting. Otherwise LOW.
Upgrade urgency: MODERATE, the most important thing is a fix in the replication
code that may make the slave hanging forever if the master
remains with an open socket even if it is no longer able to
reply.
* [BUGFIX] Multiple fixes for EVAL (issue #872).
* [BUGFIX] Fix overflow in mstime() in redis-cli and benchmark.
* [BUGFIX] Fix Linux / PPC64 behavior by correcting endianess detection.
* [BUGFIX] Fix NetBSD build by defining _XOPEN_SOURCE appropriately.
* [BUGFIX] Added missing license and copyright in a few places.
* [BUGFIX] Better error reporting when fd event creation fails.
* [FIX] MOVE now moves the TTL as well. A bug lasting forever... finally
fixed thanks to Andy Grunwald that reported it.
(reported by Andy Grunwald, fixed by Salvatore Sanfilippo)
* [FIX] Fix a false positive in HSTRLEN test.
* [FIX] Fix a bug in redis-cli --pipe mode that was not able to read back
replies from the server incrementally. Now a mass import will use
a lot less memory, and you can use --pipe to do incremental streaming.
(reported by Twitter user @fsaintjacques, fixed by Salvatore
Sanfilippo)
* [FIX] Slave detection of master timeout. (fixed by Kevin McGehee, refactoring
and regression test by Salvatore Sanfilippo)
--[ Redis 2.6.7 ]
--[ Redis 2.8.22 ] Release date: 8 Sep 2015
UPGRADE URGENCY: MODERATE (unless you BLPOP using the same key multiple times).
Upgrade urgency: HIGH for Redis and Sentinel. However note that in order to
fix certain replication bugs, the replication internals were
modified in a very heavy way. So while this release is
conceptually saner, it may contain regressions. For this
reason, before the release, QA activities were performed by
me (antirez) and Redis Labs and no evident bug was found.
* [BUGFIX] Don't crash if BLPOP & co are called with the same key repeated
multiple times (Issue #801).
* [FIX] A number of bugs related to replication PSYNC and the (yet experimental)
diskless replication feature were fixed. The bugs could lead to
inconsistency between masters and slaves. (Salvatore Sanfilippo, Oran
Agra fixed the issue found by Yuval Inbar)
* [FIX] A replication bug in the context of PSYNC partial resynchonization was
found and fixed. This bug happens even when diskless replication is off
in the case different slaves connect at different times while the master
is creating an RDB file, and later a partial resynchronization is
attempted by a slave that connected not as the first one. (Salvatore
Sanfilippo, Oran Agra)
* [FIX] Chained replication and PSYNC interactions leading to potential stale
chained slaves data set, see issue #2694. (Salvatore Sanfilippo fixed
an issue reported by "GeorgeBJ" user at Github)
* [FIX] redis-cli --scan iteration fixed when returned cursor overflows
32 bit signed integer. (Ofir Luzon, Yuval Inbar)
* [FIX] Sentinel: fixed a bug during the master switch process, where for a
failed conditional check, the new configuration is rewritten, during
a small window of time, in a corrupted way where the master is
also reported to be one of the slaves. This bug is rare to trigger
but apparently it happens in the wild, and the effect is to see
a replication loop where the master will try to replicate with itself.
A detailed explanation of the bug and its effects can be found in
the commit message here: https://github.com/antirez/redis/commit/c20218eb5770b2cafb12bc7092313b8358fedc0a.
The bug was found by Jan-Erik Rediger using a static analyzer and
fixed by Salvatore Sanfilippo.
* [FIX] Sentinel lack of arity checks for certain commands.
(Rogerio Goncalves, Salvatore Sanfilippo)
--[ Redis 2.6.6 ]
* [NEW] Replication internals rewritten in order to be more resistant to bugs.
The replication handshake in the slave side was rewritten as a non
blocking state machine. (Salvatore Sanfilippo, Oran Agra)
* [NEW] New "replication capabilities" feature introduced in order to signal
from the master to the slave what are the features supported, so that
the master can choose the kind of replication to start (diskless or
not) when master and slave are of different versions. (Oran Agra,
Salvatore Sanfilippo)
* [NEW] Log clients details when SLAVEOF command is received. (Salvatore
Sanfilippo with inputs from Nick Craver and Marc Gravell).
UPGRADE URGENCY: CRITICAL if you experienced one more more crashes.
MODERATE if Redis is running fine for you.
--[ Redis 2.8.21 ] Release date: 4 Jun 2015
* [BUGFIX] Jemalloc updated to 3.2.0.
Upgrade urgency: HIGH for Redis because of a security issue.
LOW for Sentinel.
--[ Redis 2.6.5 ]
* [FIX] Critical security issue fix by Ben Murphy: http://t.co/LpGTyZmfS7
* [FIX] SMOVE reply fixed when src and dst keys are the same. (Glenn Nethercutt)
* [FIX] Lua cmsgpack lib updated to support str8 type. (Sebastian Waisbrot)
UPGRADE URGENCY: MODERATE
* [NEW] Senitnel: CKQUORUM and FLUSHCONFIG commands back ported.
(Salvatore Sanfilippo and Bill Anderson)
Warning: this release of Redis introduces a different behavior in MULTI/EXEC
handling of errors. This was done because the new behavior is safer
compared to the old one, and should not break any code targeting
Redis 2.6 in a critical way.
--[ Redis 2.8.20 ] Release date: 5 May 2015
For more information check http://redis.io/topics/transactions
Upgrade urgency: LOW for Redis, MODERATE for Sentinel.
* [IMPROVED] RDB/AOF childern now log amount of additional memory used
because of copy on write.
* [BUGFIX] MIGRATE non critical fixes (see commits for details).
* [BUGFIX] MULTI/EXEC: now EXEC aborts on errors before EXEC.
* [BUGFIX] Fix integer overflow in zunionInterGenericCommand resulting
into Z[INTER|UNION][STORE] commands to crash under extremely
unlikely conditions (almost impossible in real world).
* [BUGFIX] EVALSHA is now case insensitive (and will not crash).
* [FIX] Sentinel memory leak due to hiredis fixed. (Salvatore Sanfilippo)
* [FIX] Sentinel memory leak on duplicated instance. (Charsyam)
* [FIX] Redis crash on Lua reaching output buffer limits. (Yossi Gottlieb)
* [FIX] Sentinel flushes config on +slave events. (Bill Anderson)
* [FIX] Fixes to diskless replication. (Oran Agra)
* [FIX] Redis (non clustered & clustered) replication bug involving blocking
operations: see issue #2473. (Salvatore Sanfilippo)
* [FIX] Config: missing activerehashing option support in CONFIG SET added.
(Salvatore Sanfilippo, thx to Bill Anderson)
* [FIX] AOF bug unlikely to happen in practice and mostly harmless: child
process segfaults when parent is not reachable via pipe. (Sun He)
* [FIX] Scripting engine now reports an error when misused with Lua debug
hooks, instead of crashing. (Salvatore Sanfilippo)
* [FIX] INFO loading stats: three fixes.
* [FIX] Fixed memory leaks in rdbSaveToSlavesSockets(). (Alon Diamant)
--[ Redis 2.6.4 ]
* [NEW] Redis-cli --latency-dist backported from unstable.
(Salvatore Sanfilippo)
UPGRADE URGENCY: LOW
--[ Redis 2.8.19 ] Release date: 16 Dec 2014
* [IMPROVED] BSD license and copyright notice added to every .c and .h file.
# UPGRADE URGENCY: LOW for both Redis and Sentinel. This release mostly
fixes small issues.
--[ Redis 2.6.2 ]
02d465c Don't log admin commands in MONITOR. (antirez)
4d8f426 List of commands flagged as admin commands modified. (antirez)
e47e460 Lua cmsgpack lib updated to latest version. (antirez)
5509c14 Add symlink to redis-sentinel during make install (Rhommel Lamas)
7de1ef7 SORT: Don't sort Set elements if not needed. (antirez)
e945a54 Fix zero-ordering SORT when called against lists (Matt Stancliff)
d81c383 Update redis_init_script.tpl (Ben Dowling)
dba57ea FIXED redis-benchmark's idle mode.With idle mode shouldn't create write event (zhanghailei)
888ea17 zipmap.c: update comments above (Sun He)
86ebc13 replaced // comments #2150 (Deepak Verma)
3d73f08 redis-benchmark AUTH command to be discarded after the first send #2150 (azure provisioned user)
76d53a6 sds.c: Correct two spelling mistakes in comments (Sun He)
4848cf9 sds.c/sdscatvprintf: set va_end to finish va_list cpy (Sun He)
d2f584f sds.c: Correct some comments (Sun He)
2ed3f09 Update whatisdoing.sh (Serghei Iakovlev)
77b997d Include stropts only if __sun is defined. (antirez)
d409371 Fix implicit declaration of ioctl on Solaris (Jan-Erik Rediger)
23b96c0 Silence _BSD_SOURCE warnings in glibc 2.20 and forward (Johan Bergström)
a47a042 Mark whatisdoing.sh as deprecated in top-comment. (antirez)
b5737d2 getting pid fixes (Serghei Iakovlev)
a598e08 sparkline.c: AddSample skip Empty label (Sun He)
7d480ab sparkline.c: mov label-ini into the AddSample Function (Sun He)
2f3c860 Only ignore sigpipe in interactive mode (Jan-Erik Rediger)
0c211a1 Simplify lua_cmsgpack macro and fix build on old Linux distros. (antirez)
UPGRADE URGENCY: LOW
--[ Redis 2.8.18 ] Release date: 4 Dec 2014
* [BUGFIX] The compilation fix for RHLE5 in 2.6.1 was broken. Fixed.
* [IMPROVED] Linenoise updated, now supports Ctrl+w.
# UPGRADE URGENCY: LOW for both Redis and Sentinel. This release mostly
adds new features to Redis, and contains non critical
fixes.
---[ Redis 2.6.1 ]
* [FIX] Linenoise updated to be more VT100 compatible. (Salvatore Sanfilippo)
* [FIX] A number of typos fixed inside comments. (Various authors)
* [FIX] redis-cli no longer quits after long timeouts. (Matt Stancliff)
* [FIX] Test framework improved to detect never terminating scripts, cleanup
instances on crashes. (Salvatore Sanfilippo)
* [FIX] PFCOUNT can be used on slaves now. (Salvatore Sanfilippo)
* [FIX] ZSCAN no longer report very small scores as 0. (Matt Stancliff,
Michael Grunder, Salvatore Sanfilippo)
* [FIX] Don't show the ASCII logo if syslog is enabled. Redis is now
an Enterprise Grade product. (Salvatore Sanfilippo)
UPGRADE URGENCY: LOW
* [NEW] EXPERIMENTAL: Diskless replication, for more info check the doc at
http://redis.io/topics/replication. (Salvatore Sanfilippo).
* [NEW] Transparent Huge Pages detection and reporting in logs and
LATENCY DOCTOR output. (Salvatore Sanfilippo)
* [NEW] Many Lua scripting enhancements: Bitops API, cjson upgrade and tests,
cmsgpack upgrade. (Matt Stancliff)
* [NEW] Total and instantaneous Network bandwidth tracking in INFO.
* [NEW] DEBUG POPULATE two args form implemented (old form still works).
The second argument is the key prefix. Default is "key:" (Salvatore
Sanfilippo)
* [NEW] Check that tcp-backlog is matched by /proc/sys/net/core/somaxconn, and
warn about it if not. (Salvatore Sanfilippo)
* [BUGFIX] Compilation on Linux < 2.6.17 or glibc < 2.6 fixed (RHLE5 & co).
--[ Redis 2.8.17 ] Release date: 19 Sep 2014
---[ Redis 2.6.0 ]
# UPGRADE URGENCY: HIGH for Redis Sentinel.
LOW for Redis Server (unmodified compared to 2.8.16).
UPGRADE URGENCY: HIGH
* [FIX] Resolved a memory leak in the hiredis library causing a memory leak
in Redis Sentinel when a monitored instance or another Sentinel is
unavailable. Every reconnection attempt will leak a small amount of
memory, but in the long run the process can reach a considerable size.
* [BUGFIX] Allow AUTH when server is in -BUSY state because of a slow script.
* [BUGFIX] MULTI/EXEC flow now makes sense when observed in MONITOR
* [BUGFIX] SCRIPT KILL now uses different error prefixes for different errors.
* [BUGFIX] Default memory limit for 32bit archs lowered from 3.5 to 3 GB.
* [BUGFIX] redis-check-dump is now compatible with RDB files generated by 2.6.
* [IMPROVED] New field in INFO: slave_read_only.
--[ Redis 2.8.16 ] Release date: 16 Sep 2014
---[ Redis 2.5.14 (2.6 Release Candidate 8) ]
# UPGRADE URGENCY: HIGH for Redis if you are using 2.8.15 + AOF.
LOW for Sentinel.
* [BUGFIX] Fixed compilation on FreeBSD.
* [IMPROVED] SRANDMEMBER <count> that returns multiple random elements.
* [IMPROVED] Sentinel backported to 2.6. It will be taken in sync with 2.8.
* [IMPROVED] Helper function for scripting to return errors and status replies.
* [IMPROVED] SORT by nosort [ASC|DESC] uses sorted set elements ordering.
* [BUGFIX] Better resistence to system clock skew.
* [IMPROVED] Warn the user when the configured maxmemory seems odd.
* [BUGFIX] Hashing function is now murmurhash2 for security purposes.
* [IMPROVED] Install script no longer uses a template but redis.conf itself.
* [FIX] The ability to load truncated AOF files introduced with Redis 2.8.15
contains a bug fixed in this release: after loading the file was not
truncated to the last valid command, so the new commands are appended
after a non well formed command. This means that:
---[ Redis 2.5.13 (2.6 Release Candidate 7) ]
1) The first AOF rewrite triggered by the server will automatically
fix the problem.
2) However, if the server is restarted before the rewrite, Redis may
not be able to load the file and you need to manually fix it.
UPGRADE URGENCY: HIGH
In order to fix a corrupted file you should start the redis-check-aof
utility WITHOUT the --fix option, just to check the offset where the
corruption is found. Around the offset reported by the check utility
you'll find, inside your AOF file, a command which is not complete
according to the Redis protocol. Just remove this incomplete command
leafing the file unaltered before and after the offending command,
and restart the server.
* [BUGFIX] Theoretical bug in ziplist fixed.
* [BUGFIX] Better out of memory handling (Log produced in log file).
* [BUGFIX] Incrementally flush RDB file on slave side while performing the
first synchronization with the master. This makes Redis less
blocking in environments where disk I/O is slow.
* [BUGFIX] Don't crash with Lua's redis.call() without arguments.
* [BUGFIX] Don't crash after a big number of Lua calls on 32 bit systems
because of a failed assertion.
* [BUGFIX] Fix SORT behaviour when called from scripting.
* [BUGFIX] Adjust slave PING period accordingly to REDIS_HZ define.
* [BUGFIX] BITCOUNT: fix crash on overflowing arguments.
* [BUGFIX] Return an error when SELECT argument is not an integer.
* [BUGFIX] Blocking operations on lists were completely reimplemented for
correctness. Now blocking list ops and pushes originated from
Lua scripts will play well together and will be replicated
and transmitted to the AOF correctly.
* [IMPROVED] Send async PING before starting replication to avoid blocking if
master allows us to connect but it is actually not able to reply.
* [IMPROVED] Support slave-priority for Redis Sentinel.
* [IMPROVED] Hiredis library updated.
IMPORTANT #1: Redis 2.8.15 is the only stable version of Redis with
this bug so probably no actual real-world problem happened since the
problem is automatically fixed at the first automatic AOF rewrite.
---[ Redis 2.5.12 (2.6 Release Candidate 6) ]
IMPORTANT #2: Before upgrading to Redis 2.8.16, if you are using Redis
2.8.15 with AOF enabled, make sure to trigger a manual AOF rewrite
using the BGREWRITEAOF command.
UPGRADE URGENCY: MODERATE.
* [FIX] SAVE is no longer propagated to AOF / slaves.
* [BUGFIX] Fixed a timing attack on AUTH (Issue #560).
* [BUGFIX] Don't assume that "char" is signed.
* [BUGFIX] Check that we have connection before enabling pipe mode.
* [BUGFIX] Use the optimized version of the function to convert a double to
its string representation. Compilation was disabled because of
a typo in the #if statement.
* [IMPROVED} REPLCONF internal command introduced, now INFO shows slaves with
correct port numbers. This makes 2.5.12 Redis Sentinel compatible.
* [IMPROVED] Truncate short write from the AOF for a cleaner restart. On short
writes (for instance out of space) Redis will now try to remove
the half-written data so that the next restart will work without
the need for the "redis-check-aof" utility.
* [IMPROVED] New in INFO: aof_last_bgrewrite_status
* [IMPROVED] Allow Pub/Sub in contexts where other commands are blocked.
* [BUGFIX] mark fd as writable when EPOLLERR or EPOLLHUP is returned by
epoll_wait.
--[ Redis 2.8.15 ] Release date: 12 Sep 2014
---[ Redis 2.5.11 (2.6 Release Candidate 5) ]
# UPGRADE URGENCY: LOW for Redis, HIGH for Sentinel.
UPGRADE URGENCY: HIGH.
* [FIX] Sentinel critical bug fixed: the absolute majority was computed in a
wrong way because of a programming error. Now the implementation does
what the specification says and the majority to authorize a failover
(that should not be confused with the ODOWN quorum) is the majority of
*all* the Sentinels ever seen for a given master, regardless of their
current state.
* [FIX] GETRANGE test no longer fails for 32 bit builds (Matt Stancliff).
* [FIX] Limit SCAN latency when the hash table is in an odd state (very few
populted buckets because rehashing is in progress). (Xiaost and
Salvatore Sanfilippo)
* [BUGFIX] Fixed Hash corruption when loading an RDB file generated by
previous versions of Redis that encoded hashes using
a different ziplist encoding format for small integers.
All the fileds that are integers in the range 0-255 may not
be recognized, or duplicated un updates, causing a crash
when the ziplist is converted to a real hash. (Issue #547).
* [BUGFIX] Fixed the count of memory used by output buffers in the
setDeferredMultiBulkLength() function.
* [NEW] Redis is now able to load truncated AOF files without requiring a
redis-check-aof utility run. The default now is to load truncated
(but apparently not corrupted) AOFs, you can change this in redis.conf.
(Salvatore Sanfilippo).
* [NEW] Sentinel: ability to announce itself with an arbitrary IP/port to work
in the context of natted networks. However this is probably still
not enough since there is no equivalent mechanism for slaves listed
in the master INFO output. (Dara Kong and Salvatore Sanfilippo)
---[ Redis 2.5.10 (2.6 Release Candidate 4) ]
--[ Redis 2.8.14 ] Release date: 1 Sep 2014
UPGRADE URGENCY: HIGH.
# UPGRADE URGENCY: HIGH for Lua scripting users, the server could crash because
of a bug introduced in Redis 2.8.10, otherwise LOW.
LOW for Redis Sentinel.
* [BUGFIX] Allow PREFIX to be overwritten on "make install".
* [BUGFIX] Run the test with just one client if the computer is slow.
* [BUGFIX] Event port support in our event driven libray.
* [BUGFIX] Jemalloc updated to 3.0.0. This fixes a possibly AOF rewrite issue.
See https://github.com/antirez/redis/issues/504 for info.
* [BUGFIX] Fixed issue #516: ZINTERSTORE / ZUNIONSTORE with mixed sets/zsets.
* [BUGFIX] Set fd to writable when poll(2) detects POLLERR or POLLHUP event.
* [BUGFIX] Fixed RESTORE hash failure (Issue #532).
* [IMPROVED] Allow an AOF rewrite buffer > 2GB (Related to issue #504).
* [IMPROVED] Server cron function frequency is now configurable (REDIS_HZ).
* [IMPROVED] Better, less blocking expired keys collection algorithm.
* [FEATURE] New commands: BITOP and BITCOUNT.
* [FEATURE] redis-cli --pipe for mass import.
* [FIX] Don't prevent use of shared integers if maxmemory policy is non-LRU.
(Salvatore Sanfilippo)
* [FIX] Fail SYNC if background save child aborted due to a signal.
(Yossi Gottlieb)
* [FIX] Different small redis-cli fixes. (Dov Murik, Charsyam, cubicdaiya,
Kashif Rasul, Jan-Erik Rediger, Matt Stancliff)
* [FIX] AIX compilation fixes. (Siah Lyimo)
* [FIX] A number of other smaller issues.
* [FIX] Improved SIGINT handling (Matt Stancliff, Salvatore Sanfilippo)
* [FIX] Use unsigned types in SDS header to raise limit to 4GB.
(Matt Stancliff, Salvatore Sanfilippo)
* [FIX] Handle signed/unsigned comparisons with more care around the code.
(Salvatore Sanfilippo)
* [FIX] Colorized test output fixed to don't change the background color.
(Mariano Pérez Rodríguez)
* [FIX] More Sentinel IPv6 fixes. (Eiichi Sato)
* [FIX] Deny CLIENT command in scripts. (Matt Stancliff)
* [FIX] Allow datasets with more than 2 billion of keys, initial work.
* [FIX] Fix a Lua scripting crash by storing the length of the static
argv when first allocated. (Paddy Byers)
What's new in Redis 2.5.9 (aka 2.6 Release Candidate 3)
=======================================================
* [NEW] Pub/Sub PING. (Salvatore Sanfilippo)
* [NEW] Much faster ZUNIONSTORE. (Kyle Hubert, Salvatore Sanfilippo)
* [NEW] Faster ll2string() implementation. (Salvatore Sanfilippo)
* [NEW] **WARNING, minor API change**: PUBSUB NUMSUB: return type modified
to integer. (Matt Stancliff)
* [NEW] redis-benchmark support for AUTH. (CharSyam)
UPGRADE URGENCY: critical, upgrade ASAP.
--[ Redis 2.8.13 ] Release date: 14 Jul 2014
* [BUGFIX] Fix for issue #500 (https://github.com/antirez/redis/pull/500).
Redis 2.6-RC1 and RC2 may corrupt ziplist-encoded sorted sets
produced by Redis 2.4.x.
* [BUGFIX] Fixed several bugs in init.d script.
* [BUGFIX] syncio.c functions modified for speed and correctness. On osx
(and possibly other BSD-based systems) the slave would block on
replication to send the SYNC command when the master was not
available. This is fixed now, but was not affecting Linux installs.
* Now when slave-serve-stale-data is set to yes and the master is down, instead
of reporting a generic error Redis replies with -MASTERDOWN.
# UPGRADE URGENCY: LOW for Redis and Sentinel, this is a features enhancement
release mostly. Since this release introduces the latency
monitoring feature, Redis 2.8 users experiencing latency
issues are strongly encouraged to upgrade.
What's new in Redis 2.5.8 (aka 2.6 Release Candidate 2)
=======================================================
* [FIX] CLIENT KILL minor backward compatibility fixes. (Salvatore Sanfilippo)
* [FIX] Enable HAVE_ATOMIC for PowerPC. (Matt Stancliff)
* [FIX] More robust PSYNC and AOF rewrites tests. (Salvatore Sanfilippo)
* [FIX] Solaris build fixed. (Matt Stancliff, Salvatore Sanfilippo)
UPGRADE URGENCY: high for all the users of the KEYS command, otherwise low.
* [NEW] The new latency monitoring feature, as documented at
http://redis.io/topics/latency-monitor (Salvatore Sanfilippo)
* [NEW] The COMMAND command, exposing the Redis command table
as an API. (Matt Stancliff)
* [NEW] Update used memory with C11 __atomic. (Matt Stancliff)
* [BUGFIX] Fix for KEYS command: if the DB contains keys with expires the KEYS
command may return the wrong output, having duplicated or missing
keys. See issue #487 and #488 on github for details.
--[ Redis 2.8.12 ] Release date: 23 Jun 2014
What's new in Redis 2.5.7 (aka 2.6 Release Candidate 1)
=======================================================
# UPGRADE URGENCY: HIGH for Redis, CRITICAL for Sentinel.
UPGRADE URGENCY: upgrade not recommended because this is an RC release.
# WARNING: This release contains a non backward compatible semantical change
to Lua script that should affect an almost zero number of users.
* This is the first release candidate for Redis 2.6. We are not aware of
bugs, but part of this code is young and was never tested in production
environments, so handle with care.
* [FIX / BREAKS BACKWARD COMPATIBILITY] Using SELECT inside Lua scripts no
longer makes the selected DB to be set in the calling client.
So Lua can still use SELECT, but the client calling the script will
remain set to the original DB. Thix fixes an issue with Redis
replication of Lua scripts that called SELECT without reverting the
selected DB to the original one. (Salvatore Sanfilippo)
* [FIX] Sentinel failover was instalbe if the master was detected as available
during the failover (especially during manual failovers) because
of an implementation error (lack of checking of
SRI_PROMOTED flag). (Salvatore Sanfilippo)
* [FIX] Cancel SHUTDOWN if initial AOF is being written. (Matt Stancliff)
* [FIX] Sentinel: bind source address for outcoming connections. (Matt
Stancliff).
* [FIX] Less timing sensitive Sentinel tests. (Salvatore Sanfilippo).
An overview of new features and changes in Redis 2.6.x
======================================================
* [NEW] redis-cli --intrinsic-latency stopped with SIGINT still reports
stats (Matt Stancliff)
* [NEW] Sentinels broadcast an HELLO message ASAP after a failover in order to
reach a consistent state faster (before it relied for periodic HELLO
messages). (Salvatore Sanfilippo).
* [NEW] Jemalloc updated to 3.6.0. (Salvatore Sanfilippo)
* [NEW] CLIENT LIST speedup. (Salvatore Sanfilippo)
* [NEW] CLIENT LIST new unique incremental ID to every client. (Salvatore
Sanfilippo)
* [NEW] ROLE command added. (Salvatore Sanfilippo)
* [NEW] CLIENT KILL new form to kill by client type and ID (see doc at
redis.io for more info). (Salvatore Sanfilippo)
* [NEW] Sentinel now disconnects clients when instances are reconfigured
(see http://redis.io/topics/sentinel-clients). (Salvatore Sanfilippo)
* [NEW] Hiredis update to latest version. (Matt Stancliff)
* Server side Lua scripting, see http://redis.io/commands/eval
* Virtual Memory removed (was deprecated in 2.4)
* Hardcoded limits about max number of clients removed.
* AOF low level semantics is generally more sane, and especially when used
in slaves.
* Milliseconds resolution expires, also added new commands with milliseconds
precision (PEXPIRE, PTTL, ...).
* Better memory usage for "small" lists, ziplists and hashes when fields or
values contain small integers.
* Read only slaves.
* New bit opeations: BITCOUNT and BITOP commands.
* Clients max output buffer soft and hard limits. You can specifiy different
limits for different classes of clients (normal,pubsub,slave).
* More incremental (less blocking) expired keys collection algorithm, in
practical terms this means that Redis is more responsive when a very
big number of keys expire about at the same time.
* AOF is now able to rewrite aggregate data types using variadic commands,
often producing an AOF that is faster to save, load, and is smaller in size.
* Every redis.conf directive is now accepted as a command line option for the
redis-server binary, with the same name and number of arguments.
* Hash table seed randomization for protection against collisions attacks.
* Performances improved when writing large objects to Redis.
* Integrated memory test, see redis-server --test-memory.
* INCRBYFLOAT and HINCRBYFLOAT commands.
* New DUMP, RESTORE, MIGRATE commands (back ported from Redis Cluster to 2.6).
* CRC64 checksump in RDB files.
* Better MONITOR output and behavior (now commands are logged before execution).
* "Software Watchdog" feature to debug latency issues.
* Significant parts of the core refactored or rewritten. New internal APIs
and core changes allowed to develop Redis Cluster on top of the new code,
however for 2.6 all the cluster code was removed, and will be released with
Redis 3.0 when it is more complete and stable.
* Redis ASCII art logo added at startup.
* Crash report on memory violation or failed asserts improved significantly
to make debugging of hard to catch bugs simpler.
* redis-benchmark improvements: ability to run selected tests,
CSV output, faster, better help.
* redis-cli improvements: --eval for comfortable development of Lua scripts.
* SHUTDOWN now supports two optional arguments: "SAVE" and "NOSAVE".
* INFO output split into sections, the command is now able to just show
pecific sections.
* New statistics about how many time a command was called, and how much
execution time it used (INFO commandstats).
* More predictable SORT behavior in edge cases.
* Better support for big endian and *BSD systems.
* Build system improved.
--[ Redis 2.8.11 ] Release date: 11 Jun 2014
Migrating from 2.4 to 2.6
# UPGRADE URGENCY: HIGH if you use Lua scripting, LOW otherwise.
* [FIX] A previous fix for Lua -> Redis numerical precision enhancement
introduced a new problem. In Redis 2.8.10 commands called from Lua
passing a string that "looks like" a very large number, may actually
use as argument the string converted as a float. This bug is now
fixed.
* [FIX] Now commands other than *PUSH* adding elements to a list will be able
to awake clients blocked in a blocking POP operation.
* [FIX] Cygwin compilation fixes.
--[ Redis 2.8.10 ] Release date: 5 Jun 2014
# UPGRADE URGENCY: HIGH if you use min-slaves-to-write option.
* [FIX] IMPORTANT! A min-slaves-to-write option active in a slave totally
prevented the slave from acception the master stream of commands.
This release includes testes for min-slaves-to-write, and a fix
for this issue.
* [FIX] Sometimes DEL returned 1 for already expired keys. Fixed.
* [FIX] Fix test false positive because new osx 'leaks' output.
* [FIX] PFCOUNT HLL cache invalidation fixed: no wrong value was reported
but the cache was not used at all, leading to lower performances.
* [FIX] Accept(2) multiple clients per readable-event invocation, and better
processing of I/O while loading or busy running a timedout script.
Basically now the LOADING / BUSY errors are reported at a decent
speed.
* [FIX] A softwaer watchdog crash fixed.
* [FIX] Fixed a Lua -> Redis numerical precision loss.
* [NEW] Lua scripting engine speed improved.
* [NEW] Sentinel generates one new event for humans to understand better
what is happening during a failover: +config-update-from.
Also the time at which a failover will be re-attempted is logged.
--[ Redis 2.8.9 ] Release date: 22 Apr 2014
# UPGRADE URGENCY: LOW, only new features introduced, no bugs fixed.
* [NEW] The HyperLogLog data structure. You can read more about it
in this blog post. http://antirez.com/news/75
* [NEW] The Sorted Set data type has now support for lexicographic range
queries, check the new commands ZRANGEBYLEX, ZLEXCOUNT and
ZREMRANGEBYLEX, which are documented at http://redis.io.
--[ Redis 2.8.8 ] Release date: 25 Mar 2014
# UPGRADE URGENCY: HIGH for Redis, LOW for Sentinel. There is a potentially
critical bug fix causing data loss in Redis but it requires
a combination of disk full and the use of the
SHUTDOWN command.
* [FIX] Fixed data loss when SHUTDOWN was used with a disk full condition.
* [FIX] Fixed a memory leak in the SORT syntax error processing.
* [FIX] When Sentinel down-after-milliseconds parameter is modified at runtime
now it gets propagated to all the slaves and sentinel instances
of the master.
* [FIX] `install_server.sh` script finally fixed.
* [FIX] Different fixes to maxclients handling.
* [NEW] Sentinels are now able to send update messages in a peer-to-peer
fashion even if no Redis instances are available. Now the Sentinel
liveness property that the most updated configuration in a given
partition is propagated to all the Sentinels is extended to partitions
without reachable instances.
* [NEW] Sentinel safety properties are now ensured in a crash-recovery system
model since some state is persisted on disk before replying to other
nodes, and reloaded at startup.
* [NEW] Sentinel now uses CLIENT SETNAME so that it is easy to identify
Sentinels using CLIENT LIST among other clients.
* [NEW] Sentinel failure detection and reconnection code improved.
* [NEW] Use all 24 bits (instead of 22) for the Redis objects LRU field.
Note that the new LRU algorithm using eviction pools was not backported
from unstable for safery / code maturity concerns.
* [NEW] Majory speedup for the INFO command (it is now 6 times faster).
* [NEW] More Sentinel unit tests.
* [NEW] New command DEBUG ERROR returns the specified error. Example:
DEBUG ERROR "LOADING database". This is handy to write Redis client
libraries unit tests.
* [NEW] redis-cli now supports multi-line editing via updated linenoise lib.
Thanks to Matt Stancliff and Jan-Erik Rediger for the work done in the context
of this release.
--[ Redis 2.8.7 ] Release date: 5 Mar 2014
# UPGRADE URGENCY: LOW for Redis, LOW for Sentinel. However this release adds
new features so users may want to upgrade in order to
exploit the new functionalities.
* [FIX] Sometimes the absolute config file path was obtained in a wrong way.
This happened when there was a "dir" directive inside the config file
and at the same time the configuration file was given as a relative
path to redis-server or redis-sentinel executables.
* [FIX] redis-cli: Automatically enter --slave mode when SYNC or PSYNC are
called during an interactive session.
* [FIX] Sentinel "IDONTKNOW" error removed as it does not made sense with the
new Sentinel design. This error was actually a fix for a design error
in the first implementation of Sentinel.
* [FIX] Sentinel: added a missing exit() call to abort after config file
checks at startup. This error was introduced with an improvement in
a previous 2.8 release.
* [FIX] BITCOUNT: fixed unaligned access causing issues in sparc and other
archs not capable of dealing with unaligned accesses. This also makes
the code faster in archs where unaligned accesses are allowed.
* [FIX] Sentinel: better nodes fail over start time desynchronization to avoid
split-brain during the voting process needed to get authorization to
fail over. This means the system is less likely to need to retry
and will fail over faster. No changes in behavior / correctness.
* [FIX] Force INFO used_memory_peak to match peak memory. This generated some
confusion among users even if it was not an actual bug.
* [NEW] Sentinel unit tests and framework. More tests needed and units must
be improved in order to have less false positives, but it is a start
and features a debugging console that is useful to fix tests or to
inspect bugs causing tests failures.
* [NEW] New Sentinel events: +/-monitor and +set used to monitor when an
instance to monitor is added or removed, or when a configuration
is modified via SENTINEL SET.
* [NEW] Redis-cli updated to use SCAN instead of random sampling via
RANDOMKEY in order to implement --bigkeys feature. Moreover the
implementation now supports pipelining and reports more information
at the end of the scan. Much faster, much better. A special thank
you to Michael Grunder for this improvement.
* [NEW] redis-cli now supports a new --intrinsic-latency mode that is able
to meter the latency of a system due to kernel / hypervisor.
How to use it is explained at http://redis.io/topics/latency.
* [NEW] New command BITPOS: find first bit set or clear in a bitmap.
* [NEW] CONFIG REWRITE calls are now logged.
--[ Redis 2.8.6 ] Release date: 13 Feb 2014
# UPGRADE URGENCY: HIGH for Redis, LOW for Sentinel. Redis users using Lua
scripts with EVALSHA and attached slaves and/or AOF
persistence should consider upgrading ASAP.
* [FIX] Fixed an critical EVALSHA script cache bug: scripts executed may not
propagate to AOF / Slaves correctly under certain conditions.
See issue #1549 at Github for more information.
* [FIX] Fixed multiple bugs resulting into closing the link with master or slave
during replication without good reasons. This will result in useless
resynchronizations, or infinite loops where the replication link can't
be established.
* [FIX] Don't count the time needed to populate the buffers of clients waiting
in MONITOR mode when populating the Slow Log entries.
* [NEW] AOF write errors (like no space on device) no longer abort Redis if the
fsync policy is none or every second. The database enters a read-only
mode where every write is refused with an error. Normal operations are
restored as soon as Redis is able to append again data to the AOF file.
* [NEW] Sentinel now accepts SHUTDOWN command.
--[ Redis 2.8.5 ] Release date: 4 Feb 2014
# UPGRADE URGENCY: HIGH for Redis, LOW for Sentinel. Redis users using Lua
scripts with expires, and Redis users relying on the
ability of Redis to block writes on RDB saving errors
should plan to upgrade ASAP.
* [FIX] Fixed a replication bug caused by Lua scripts + expired keys: keys could
expire in the middle of scripts causing non-deterministic behavior.
* [FIX] MISCONFIG error if condition fixed, the server was no longer able
to stop writes on RDB misconfiguration after this error was introduced.
* [FIX] REDIS_AOF_REWRITE_MIN_SIZE is now 64mb like example redis.conf default.
* [FIX] Perform fflush() before fsync() in rio.c (bug without actual effects).
* [FIX] Don't log MONITOR clients as disconnecting slaves.
* [FIX] SENTINEL MASTER arity check fixed. Crashed the Sentinel instance when
the command was given without arguments.
* [NEW] Allow CONFIG and SHUTDOWN while in stale-slave state.
* [NEW] Support for configurable TCP listen(2) backlog size.
* [NEW] redis-cli supports SCAN via the --scan and --pattern options.
* [NEW] SENTINEL SET master quorum via runtime API implemented.
--[ Redis 2.8.4 ] Release date: 13 Jan 2014
# UPGRADE URGENCY: MODERATE for Redis and Sentinel.
* [FIX] Makefile compatibility with non common make variants improved.
* [FIX] SDIFF crash in very unlikely to trigger state fixed.
* [FIX] Config rewriting fixed: don't wipe options unknown to the rewrite
process.
* [FIX] Set TCP port to 0 works again to disable TCP networking.
* [FIX] Fixed replication with old Redis instances as masters by not
sending REPLCONF ACK to them.
* [FIX] Fix keyspace notifications rewrite and CONFIG GET output.
* [FIX] Fix RESTORE TTL handling in 32 bit systems (32 bit overflow).
* [NEW] Sentinel now has a run time configuration API.
* [NEW] Log when we lost connection with master or slave.
* [NEW] When instance is turned from slave to master now inherits the
old master replication offset when possible. This improves the
Sentinel failover procedure.
--[ Redis 2.8.3 ] Release date: 11 Dec 2013
# UPGRADE URGENCY: MODERATE for Redis, HIGH for Sentinel.
* [FIX] Sentinel instance role sampling fixed, the system is now more
reliable during failover and when reconfiguring instances with
non matching configuration.
* [FIX] Inline requests are now handled even when terminated with just LF.
* [FIX] Replication timeout handling greatly improved, now the slave is able
to ping the master while removing the old data from memory, and while
loading the new RDB file. This avoid false timeouts sensed by
masters.
* [FIX] Fixed a replication bug involving 32 bit instances and big datasets
hard to compress that resulted into more than 2GB of RDB file sent.
* [FIX] Return error for inline requests with unbalanced quotes.
* [FIX] Publish the slave replication offset even when disconnected from the
master if there is still a cached master instance.
--[ Redis 2.8.2 ] Release date: 2 Dec 2013
# UPGRADE URGENCY: MODERATE for both Redis and Sentinel.
* [FIX] Sentinel better desynchronization to avoid split-brain elections
where no Sentinel managed to get elected.
* [FIX] Stop accepting writes on "MISCONF" error only if master, not slave.
* [FIX] Reply to PING with an error on "MISCONF" errors.
--[ Redis 2.8.1 ] Release date: 25 Nov 2013
# UPGRADE URGENCY: LOW for Redis, CRITICAL for Senitnel. You don't need to
upgrade your Redis instances but it is highly recommended
to upgrade and restart all the Sentinel processes.
* [FIX] Fixed a bug in "new Sentinel" config propagation.
* [FIX] Fixed a false positive in Redis tests.
--[ Redis 2.8.0 ] Release date: 22 Nov 2013
# UPGRADE URGENCY: LOW, unless you want to upgrade to new Sentinel code.
* [FIX] Fixed an error in rdbWriteRaw() that should have no practical impact.
* [NEW] Log the new master when SLAVEOF command is used.
* [NEW] Sentinel code synchronized with the unstable branch, the new Sentinel
is a reimplementation that uses more reliable algorithms.
--[ Redis 2.8 Release Candidate 6 (2.7.106) ] Release date: 6 Nov 2013
This is the 6th release candidate of Redis 2.8 (official version is 2.7.106).
# UPGRADE URGENCY: LOW, only new features back ported, no fixes.
* [NEW] SCAN, SSCAN, HSCAN, ZSCAN commands.
--[ Redis 2.8 Release Candidate 5 (2.7.105) ] Release date: 9 Oct 2013
This is the 5th release candidate of Redis 2.8 (official version is 2.7.105).
Important bugs fixed inside.
# UPGRADE URGENCY: HIGH because of many non critical replication bugs fixed.
* [FIX] redis-cli: don't crash with --bigkeys when the key no longer exist.
* [FIX] Allow AUTH / PING when disconnected from slave and serve-stale-data is no.
* [FIX] PSYNC: safer handling of PSYNC requests with offsets in the future.
* [FIX] Replication: Fix master timeout detection.
* [FIX] Replication: Correctly install the write handler after successful PSYNC.
--[ Redis 2.8 Release Candidate 4 (2.7.104) ] Release date: 30 Aug 2013
This is the fourth release candidate of Redis 2.8 (official version is 2.7.104).
Important bugs fixed inside.
# UPGRADE URGENCY: HIGH because of the EVAL memory leak.
* [FIX] Fixed a serious EVAL memory leak in the Lua stack.
* [FIX] Fixed server startup when no IPv6 address exists in any interface.
* [FIX] Send MISCONFIG error when BGSAVE fails because can't fork.
* [FIX] Memory efficiency with large (> a few kbytes) values improved considerably.
* [NEW] DEBUG SDSLEN for sds memory debugging.
--[ Redis 2.8 Release Candidate 3 (2.7.103) ] Release date: 19 Aug 2013
This is the third release candidate of Redis 2.8 (official version is 2.7.103).
Important bugs fixed inside.
# UPGRADE URGENCY: HIGH
* [FIX] Improved expired keys collection algorithm. Even under heavy load keys
to be expired can't accumulate because of lack of CPU time.
* [FIX] Replication speed regression fixed (issue #1238).
* [FIX] Fixed an hard to trigger PSYNC bug.
* [FIX] Fixed Issue #1240, ZUNIONSTORE could lead to wrong result.
* [NEW] Add per-db average TTL information in INFO output.
* [NEW] redis-benchmark improvements.
* [NEW] dict.c API wrong usage detection.
--[ Redis 2.8 Release Candidate 2 (2.7.102) ] Release date: 30 Jul 2013
This is the second release candidate of Redis 2.8 (official version is 2.7.102).
Important bugs fixed inside.
# UPGRADE URGENCY: HIGH
* [FIX] Fixed a critical replication bug, see issue #1221.
* [NEW] The new inline protocol now accepts quoted strings like, for example
you can now type in a telnet session: set 'foo bar' "hello world\n".
--[ Redis 2.8 Release Candidate 1 (2.7.101) ] Release date: 18 Jul 2013
This is the first release candidate of Redis 2.8 (official version is 2.7.101).
The following is a list of improvements in Redis 2.8, compared to Redis 2.6.
* [NEW] Slaves are now able to partially resynchronize with the master, so most
of the times a full resynchronization with the RDB creation in the master
side is not needed when the master-slave link is disconnected for a short
amount of time.
* [NEW] Experimental IPv6 support.
* [NEW] Slaves explicitly ping masters now, a master is able to detect a timed out
slave independently.
* [NEW] Masters can stop accepting writes if not enough slaves with a given
maximum latency are connected.
* [NEW] Keyspace changes notifications via Pub/Sub.
* [NEW] CONFIG SET maxclients is now available.
* [NEW] Ability to bind multiple IP addresses.
* [NEW] Set process names so that you can recognize, in the "ps" command output,
the listening port of an instance, or if it is a saving child.
* [NEW] Automatic memory check on crash.
* [NEW] CONFIG REWRITE is able to materialize the changes in the configuration
operated using CONFIG SET into the redis.conf file.
* [NEW] More NetBSD friendly code base.
* [NEW] PUBSUB command for Pub/Sub introspection capabilities.
* [NEW] EVALSHA can now be replicated as such, without requiring to be expanded
to a full EVAL for the replication link.
* [NEW] Better Lua scripts error reporting.
* [NEW] SDIFF performance improved.
* [FIX] A number of bugfixes.
Migrating from 2.6 to 2.8
=========================
Redis 2.4 is mostly a strict subset of 2.6. However there are a few things
Redis 2.6 is mostly a strict subset of 2.8. However there are a few things
that you should be aware of:
* You can't use .rdb and AOF files generated with 2.6 into a 2.4 instance.
* 2.6 slaves can be attached to 2.4 masters, but not the contrary, and only
for the time needed to perform the version upgrade.
The following commands changed behavior:
There are also a few API differences, that are unlikely to cause problems,
but it is better to keep them in mind:
* SORT now will refuse to sort in numerical mode elements that can't be parsed
as numbers.
* EXPIREs now all have millisecond resolution (but this is very unlikely to
break code that was not conceived exploting the previous resolution error
in some way.)
* INFO output is a bit different now, and contains empty lines and comments
starting with '#'. All the major clients should be already fixed to work
with the new INFO format.
* Slaves are only read-only by default (but you can change this easily
setting the "slave-read-only" configuration option to "no" editing your
redis.conf or using CONFIG SET.
The following INFO fields were renamed for consistency:
changes_since_last_save -> rdb_changes_since_last_save
bgsave_in_progress -> rdb_bgsave_in_progress
last_save_time -> rdb_last_save_time
last_bgsave_status -> rdb_last_bgsave_status
bgrewriteaof_in_progress -> aof_rewrite_in_progress
bgrewriteaof_scheduled -> aof_rewrite_scheduled
* The TTL and PTTL commands now return -2 if the key does not exist and
-1 if it exists but has no associated expire. Redis 2.6 and previous
versions used to return -1 for both the conditions.
* SORT with ALPHA now sorts according to local collation locale if no STORE
option is used.
* ZADD/ZINCRBY are now able to accept a bigger range of values as valid
scores, that is, all the values you may end having as a result of
calling ZINCRBY multiple times.
* Many errors are now prefixed by a more specific error code instead of
the generic -ERR, for example -WRONGTYPE, -NOAUTH, ...
* PUBLISH called inside Lua scripts is now correctly propagated to slaves.
The following redis.conf and CONFIG GET / SET parameters changed:
* hash-max-zipmap-entries, now replaced by hash-max-ziplist-entries
* hash-max-zipmap-value, now replaced by hash-max-ziplist-value
* glueoutputbuf option was now completely removed (was deprecated)
* logfile now uses the empty string in order to log to standard output,
so 'logfile stdout' is now invalid, use 'logfile ""' instead.
The following INFO fields changed format in a non-backward compatible way:
* The list of slaves in INFO is now in field=value format.
Replication:
Redis 2.8 can be used as slave for Redis 2.6, but doing this is only
a good idea for the short amount of time needed to upgrade your servers.
We suggest to update both master and slaves at about the same time.
--------------------------------------------------------------------------------
Credits: Where not specified the implementation and design are done by
Salvatore Sanfilippo and Pieter Noordhuis. Thanks to VMware for making all
Credits: Where not specified the implementation and design is done by
Salvatore Sanfilippo. Thanks to VMware and Pivotal for making all
this possible. Also many thanks to all the other contributors and the amazing
community we have.
BUGS
+1 -1
View File
@@ -1 +1 @@
Plese check https://github.com/antirez/redis/issues
Please check https://github.com/antirez/redis/issues
CONTRIBUTING
+1 -1
View File
@@ -22,7 +22,7 @@ each source file that you contribute.
1. Drop a message to the Redis Google Group with a proposal of semantics/API.
2. If in steps 1 you get an acknowledge from the project leaders, use the
2. If in step 1 you get an acknowledge from the project leaders, use the
following procedure to submit a patch:
a. Fork Redis on github ( http://help.github.com/fork-a-repo/ )
COPYING
+1 -1
View File
@@ -1,4 +1,4 @@
Copyright (c) 2006-2012, Salvatore Sanfilippo
Copyright (c) 2006-2015, Salvatore Sanfilippo
All rights reserved.
Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:
MANIFESTO
+55 -8
View File
@@ -4,17 +4,64 @@
Redis Manifesto
===============
1 - A DSL for Abstract Data Types. Redis is a DSL (Domain Specific Language) that manipulates abstract data types and implemented as a TCP daemon. Commands manipulate a key space where keys are binary-safe strings and values are different kinds of abstract data types. Every data type represents an abstract version of a fundamental data structure. For instance Redis Lists are an abstract representation of linked lists. In Redis, the essence of a data type isn't just the kind of operations that the data types support, but also the space and time complexity of the data type and the operations performed upon it.
1 - A DSL for Abstract Data Types. Redis is a DSL (Domain Specific Language)
that manipulates abstract data types and implemented as a TCP daemon.
Commands manipulate a key space where keys are binary-safe strings and
values are different kinds of abstract data types. Every data type
represents an abstract version of a fundamental data structure. For instance
Redis Lists are an abstract representation of linked lists. In Redis, the
essence of a data type isn't just the kind of operations that the data types
support, but also the space and time complexity of the data type and the
operations performed upon it.
2 - Memory storage is #1. The Redis data set, composed of defined key-value pairs, is primarily stored in the computer's memory. The amount of memory in all kinds of computers, including entry-level servers, is increasing significantly each year. Memory is fast, and allows Redis to have very predictable performance. Datasets composed of 10k or 40 millions keys will perform similarly. Complex data types like Redis Sorted Sets are easy to implement and manipulate in memory with good performance, making Redis very simple. Redis will continue to explore alternative options (where data can be optionally stored on disk, say) but the main goal of the project remains the development of an in-memory database.
2 - Memory storage is #1. The Redis data set, composed of defined key-value
pairs, is primarily stored in the computer's memory. The amount of memory in
all kinds of computers, including entry-level servers, is increasing
significantly each year. Memory is fast, and allows Redis to have very
predictable performance. Datasets composed of 10k or 40 millions keys will
perform similarly. Complex data types like Redis Sorted Sets are easy to
implement and manipulate in memory with good performance, making Redis very
simple. Redis will continue to explore alternative options (where data can
be optionally stored on disk, say) but the main goal of the project remains
the development of an in-memory database.
3 - Fundamental data structures for a fundamental API. The Redis API is a direct consequence of fundamental data structures. APIs can often be arbitrary but not an API that resembles the nature of fundamental data structures. If we ever meet intelligent life forms from another part of the universe, they'll likely know, understand and recognize the same basic data structures we have in our computer science books. Redis will avoid intermediate layers in API, so that the complexity is obvious and more complex operations can be performed as the sum of the basic operations.
3 - Fundamental data structures for a fundamental API. The Redis API is a direct
consequence of fundamental data structures. APIs can often be arbitrary but
not an API that resembles the nature of fundamental data structures. If we
ever meet intelligent life forms from another part of the universe, they'll
likely know, understand and recognize the same basic data structures we have
in our computer science books. Redis will avoid intermediate layers in API,
so that the complexity is obvious and more complex operations can be
performed as the sum of the basic operations.
4 - Code is like a poem; it's not just something we write to reach some practical result. Sometimes people that are far from the Redis philosophy suggest using other code written by other authors (frequently in other languages) in order to implement something Redis currently lacks. But to us this is like if Shakespeare decided to end Enrico IV using the Paradiso from the Divina Commedia. Is using any external code a bad idea? Not at all. Like in "One Thousand and One Nights" smaller self contained stories are embedded in a bigger story, we'll be happy to use beautiful self contained libraries when needed. At the same time, when writing the Redis story we're trying to write smaller stories that will fit in to other code.
4 - Code is like a poem; it's not just something we write to reach some
practical result. Sometimes people that are far from the Redis philosophy
suggest using other code written by other authors (frequently in other
languages) in order to implement something Redis currently lacks. But to us
this is like if Shakespeare decided to end Enrico IV using the Paradiso from
the Divina Commedia. Is using any external code a bad idea? Not at all. Like
in "One Thousand and One Nights" smaller self contained stories are embedded
in a bigger story, we'll be happy to use beautiful self contained libraries
when needed. At the same time, when writing the Redis story we're trying to
write smaller stories that will fit in to other code.
5 - We're against complexity. We believe designing systems is a fight against complexity. We'll accept to fight the complexity when it's worthwhile but we'll try hard to recognize when a small feature is not worth 1000s of lines of code. Most of the time the best way to fight complexity is by not creating it at all.
5 - We're against complexity. We believe designing systems is a fight against
complexity. We'll accept to fight the complexity when it's worthwhile but
we'll try hard to recognize when a small feature is not worth 1000s of lines
of code. Most of the time the best way to fight complexity is by not
creating it at all.
6 - Two levels of API. The Redis API has two levels: 1) a subset of the API fits naturally into a distributed version of Redis and 2) a more complex API that supports multi-key operations. Both are useful if used judiciously but there's no way to make the more complex multi-keys API distributed in an opaque way without violating our other principles. We don't want to provide the illusion of something that will work magically when actually it can't in all cases. Instead we'll provide commands to quickly migrate keys from one instance to another to perform multi-key operations and expose the tradeoffs to the user.
7 - We optimize for joy. We believe writing code is a lot of hard work, and the only way it can be worth is by enjoying it. When there is no longer joy in writing code, the best thing to do is stop. To prevent this, we'll avoid taking paths that will make Redis less of a joy to develop.
6 - Two levels of API. The Redis API has two levels: 1) a subset of the API fits
naturally into a distributed version of Redis and 2) a more complex API that
supports multi-key operations. Both are useful if used judiciously but
there's no way to make the more complex multi-keys API distributed in an
opaque way without violating our other principles. We don't want to provide
the illusion of something that will work magically when actually it can't in
all cases. Instead we'll provide commands to quickly migrate keys from one
instance to another to perform multi-key operations and expose the tradeoffs
to the user.
7 - We optimize for joy. We believe writing code is a lot of hard work, and the
only way it can be worth is by enjoying it. When there is no longer joy in
writing code, the best thing to do is stop. To prevent this, we'll avoid
taking paths that will make Redis less of a joy to develop.
README
+173
View File
@@ -0,0 +1,173 @@
Where to find complete Redis documentation?
-------------------------------------------
This README is just a fast "quick start" document. You can find more detailed
documentation at http://redis.io
Building Redis
--------------
Redis can be compiled and used on Linux, OSX, OpenBSD, NetBSD, FreeBSD.
We support big endian and little endian architectures.
It may compile on Solaris derived systems (for instance SmartOS) but our
support for this platform is "best effort" and Redis is not guaranteed to
work as well as in Linux, OSX, and *BSD there.
It is as simple as:
% make
You can run a 32 bit Redis binary using:
% make 32bit
After building Redis is a good idea to test it, using:
% make test
Fixing build problems with dependencies or cached build options
—--------
Redis has some dependencies which are included into the "deps" directory.
"make" does not rebuild dependencies automatically, even if something in the
source code of dependencies is changes.
When you update the source code with `git pull` or when code inside the
dependencies tree is modified in any other way, make sure to use the following
command in order to really clean everything and rebuild from scratch:
make distclean
This will clean: jemalloc, lua, hiredis, linenoise.
Also if you force certain build options like 32bit target, no C compiler
optimizations (for debugging purposes), and other similar build time options,
those options are cached indefinitely until you issue a "make distclean"
command.
Fixing problems building 32 bit binaries
---------
If after building Redis with a 32 bit target you need to rebuild it
with a 64 bit target, or the other way around, you need to perform a
"make distclean" in the root directory of the Redis distribution.
In case of build errors when trying to build a 32 bit binary of Redis, try
the following steps:
* Install the packages libc6-dev-i386 (also try g++-multilib).
* Try using the following command line instead of "make 32bit":
make CFLAGS="-m32 -march=native" LDFLAGS="-m32"
Allocator
---------
Selecting a non-default memory allocator when building Redis is done by setting
the `MALLOC` environment variable. Redis is compiled and linked against libc
malloc by default, with the exception of jemalloc being the default on Linux
systems. This default was picked because jemalloc has proven to have fewer
fragmentation problems than libc malloc.
To force compiling against libc malloc, use:
% make MALLOC=libc
To compile against jemalloc on Mac OS X systems, use:
% make MALLOC=jemalloc
Verbose build
-------------
Redis will build with a user friendly colorized output by default.
If you want to see a more verbose output use the following:
% make V=1
Running Redis
-------------
To run Redis with the default configuration just type:
% cd src
% ./redis-server
If you want to provide your redis.conf, you have to run it using an additional
parameter (the path of the configuration file):
% cd src
% ./redis-server /path/to/redis.conf
It is possible to alter the Redis configuration passing parameters directly
as options using the command line. Examples:
% ./redis-server --port 9999 --slaveof 127.0.0.1 6379
% ./redis-server /etc/redis/6379.conf --loglevel debug
All the options in redis.conf are also supported as options using the command
line, with exactly the same name.
Playing with Redis
------------------
You can use redis-cli to play with Redis. Start a redis-server instance,
then in another terminal try the following:
% cd src
% ./redis-cli
redis> ping
PONG
redis> set foo bar
OK
redis> get foo
"bar"
redis> incr mycounter
(integer) 1
redis> incr mycounter
(integer) 2
redis>
You can find the list of all the available commands here:
http://redis.io/commands
Installing Redis
-----------------
In order to install Redis binaries into /usr/local/bin just use:
% make install
You can use "make PREFIX=/some/other/directory install" if you wish to use a
different destination.
Make install will just install binaries in your system, but will not configure
init scripts and configuration files in the appropriate place. This is not
needed if you want just to play a bit with Redis, but if you are installing
it the proper way for a production system, we have a script doing this
for Ubuntu and Debian systems:
% cd utils
% ./install_server.sh
The script will ask you a few questions and will setup everything you need
to run Redis properly as a background daemon that will start again on
system reboots.
You'll be able to stop and start Redis using the script named
/etc/init.d/redis_<portnumber>, for instance /etc/init.d/redis_6379.
Code contributions
---
Note: by contributing code to the Redis project in any form, including sending
a pull request via Github, a code fragment or patch via private email or
public discussion groups, you agree to release your code under the terms
of the BSD license that you can find in the COPYING file included in the Redis
source distribution.
Please see the CONTRIBUTING file in this source distribution for more
information.
Enjoy!
README.md
-59
View File
@@ -1,59 +0,0 @@
Redis on Windows 2.6 prototype
===
## What's new in this release
- This is a port for Windows based on Redis 2.6.8.
- The port is similar to the port of Redis 2.4, including the same background saving technology.
- There is support for x64 version as well as 32 bit versions.
- For the 64 bit version, there is a limit of 2^32 objects in a structure, and a max length of 2^32 for any object
- Version number now 2.6.8-pre2 to indicate prerelease and to enable changing
- Version 2.6.8-pre2 fixes several failures that existed in 2.6.8-pre1. Most of these were related to handling opening and closing of non-blocking sockets.
- The unit/protocol desync test is updated to use nonblocking socket and now works on Windows
- The binaries (unsigned) have been moved to the root to make them easier to find. Previously they were under msvs.
- Signed binaries can be downloaded using Nuget.
##Acknowledgements
Special thanks to Dušan Majkic (https://github.com/dmajkic, https://github.com/dmajkic/redis/) for his project on GitHub that gave us the opportunity to quickly learn some on the intricacies of Redis code. His project also helped us to build our prototype quickly.
## Repo branches
- 2.6: This is the branch for the Windows Redis port based on Redis 2.6.
- 2.4: This branch has the Windows Redis port based on Redis 2.4.
## How to build Redis using Visual Studio
You can use the free Express Edition available at http://www.microsoft.com/visualstudio/en-us/products/2010-editions/visual-cpp-express.
- Open the solution file msvs\redisserver.sln in Visual Studio 10, select platform (win32 or x64) and build.
This should create the following executables in the msvs\$(Configuration) folder:
- redis-server.exe
- redis-benchmark.exe
- redis-cli.exe
- redis-check-dump.exe
- redis-check-aof.exe
For your convenience all binaries are be available in the msvs/bin/release|debug directories.
## RedisWatcher
So far the RedisWatcher is not carried over to 2.6. However this should not be affected by the Redis version, and the code in the 2.4 branch should work with the Redis 2.6 binaries.
## RedisWAInst
So far the RedisWAInst is not carried over to 2.6. However this should not be affected by the Redis version, and the code in the 2.4 branch should work with the Redis 2.6 binaries.
## Release Notes
The branch has been renamed from 2.6_alpha to 2.6 to indicate that the test pass has been completed.
This is a release version of the software.
To run the Redis test suite requires some manual work:
- The tests assume that the binaries are in the src folder, so you need to copy the binaries from the msvs folder to src.
- The tests make use of TCL. This must be installed separately.
- To run the tests you need to have a Unix shell on your machine. To execute the tests, run the following command: `tclsh8.5.exe tests/test_helper.tcl`.
If a Unix shell is not installed you may see the following error message: "couldn't execute "cat": no such file or directory".
## Known issues
None.
bin/.gitignore
-4
View File
@@ -1,4 +0,0 @@
!release
!debug
!release/*
!debug/*
bin/release/redisbin.zip
BIN
View File
Binary file not shown.
bin/release/redisbin64.zip
BIN
View File
Binary file not shown.
deps/Makefile
Vendored
+9 -4
View File
@@ -54,16 +54,21 @@ linenoise: .make-prerequisites
.PHONY: linenoise
ifeq ($(uname_S),SunOS)
# Make isinf() available
LUA_CFLAGS= -D__C99FEATURES__=1
# Make isinf() available
LUA_CFLAGS= -D__C99FEATURES__=1
endif
LUA_CFLAGS+= -O2 -Wall -DLUA_ANSI $(CFLAGS)
LUA_CFLAGS+= -O2 -Wall -DLUA_ANSI -DENABLE_CJSON_GLOBAL $(CFLAGS)
LUA_LDFLAGS+= $(LDFLAGS)
# lua's Makefile defines AR="ar rcu", which is unusual, and makes it more
# challenging to cross-compile lua (and redis). These defines make it easier
# to fit redis into cross-compilation environments, which typically set AR.
AR=ar
ARFLAGS=rcu
lua: .make-prerequisites
@printf '%b %b\n' $(MAKECOLOR)MAKE$(ENDCOLOR) $(BINCOLOR)$@$(ENDCOLOR)
cd lua/src && $(MAKE) all CFLAGS="$(LUA_CFLAGS)" MYLDFLAGS="$(LUA_LDFLAGS)"
cd lua/src && $(MAKE) all CFLAGS="$(LUA_CFLAGS)" MYLDFLAGS="$(LUA_LDFLAGS)" AR="$(AR) $(ARFLAGS)"
.PHONY: lua
deps/hiredis/.gitignore
+1 -1
View File
@@ -1,5 +1,5 @@
/hiredis-test
/hiredis-example*
/examples/hiredis-example*
/*.o
/*.so
/*.dylib
deps/hiredis/.travis.yml
Vendored
+6
View File
@@ -0,0 +1,6 @@
language: c
compiler:
- gcc
- clang
script: make && make check
deps/hiredis/CHANGELOG.md
Vendored
+8
View File
@@ -1,3 +1,11 @@
### 0.11.0
* Increase the maximum multi-bulk reply depth to 7.
* Increase the read buffer size from 2k to 16k.
* Use poll(2) instead of select(2) to support large fds (>= 1024).
### 0.10.1
* Makefile overhaul. Important to check out if you override one or more
deps/hiredis/Makefile
Vendored
+42 -24
View File
@@ -4,11 +4,24 @@
# This file is released under the BSD license, see the COPYING file
OBJ=net.o hiredis.o sds.o async.o
BINS=hiredis-example hiredis-test
EXAMPLES=hiredis-example hiredis-example-libevent hiredis-example-libev
TESTS=hiredis-test
LIBNAME=libhiredis
HIREDIS_MAJOR=0
HIREDIS_MINOR=10
HIREDIS_MINOR=11
# redis-server configuration used for testing
REDIS_PORT=56379
REDIS_SERVER=redis-server
define REDIS_TEST_CONFIG
daemonize yes
pidfile /tmp/hiredis-test-redis.pid
port $(REDIS_PORT)
bind 127.0.0.1
unixsocket /tmp/hiredis-test-redis.sock
endef
export REDIS_TEST_CONFIG
# Fallback to gcc when $CC is not in $PATH.
CC:=$(shell sh -c 'type $(CC) >/dev/null 2>/dev/null && echo $(CC) || echo gcc')
@@ -41,12 +54,11 @@ ifeq ($(uname_S),Darwin)
DYLIB_MAKE_CMD=$(CC) -shared -Wl,-install_name,$(DYLIB_MINOR_NAME) -o $(DYLIBNAME) $(LDFLAGS)
endif
all: $(DYLIBNAME) $(BINS)
all: $(DYLIBNAME)
# Deps (use make dep to generate this)
net.o: net.c fmacros.h net.h hiredis.h
async.o: async.c async.h hiredis.h sds.h dict.c dict.h
example.o: example.c hiredis.h
hiredis.o: hiredis.c fmacros.h hiredis.h net.h sds.h
sds.o: sds.c sds.h
test.o: test.c hiredis.h
@@ -61,36 +73,44 @@ dynamic: $(DYLIBNAME)
static: $(STLIBNAME)
# Binaries:
hiredis-example-libevent: example-libevent.c adapters/libevent.h $(STLIBNAME)
$(CC) -o $@ $(REAL_CFLAGS) $(REAL_LDFLAGS) -levent example-libevent.c $(STLIBNAME)
hiredis-example-libevent: examples/example-libevent.c adapters/libevent.h $(STLIBNAME)
$(CC) -o examples/$@ $(REAL_CFLAGS) $(REAL_LDFLAGS) -I. $< -levent $(STLIBNAME)
hiredis-example-libev: example-libev.c adapters/libev.h $(STLIBNAME)
$(CC) -o $@ $(REAL_CFLAGS) $(REAL_LDFLAGS) -lev example-libev.c $(STLIBNAME)
hiredis-example-libev: examples/example-libev.c adapters/libev.h $(STLIBNAME)
$(CC) -o examples/$@ $(REAL_CFLAGS) $(REAL_LDFLAGS) -I. $< -lev $(STLIBNAME)
ifndef AE_DIR
hiredis-example-ae:
@echo "Please specify AE_DIR (e.g. <redis repository>/src)"
@false
else
hiredis-example-ae: example-ae.c adapters/ae.h $(STLIBNAME)
$(CC) -o $@ $(REAL_CFLAGS) $(REAL_LDFLAGS) -I$(AE_DIR) $(AE_DIR)/ae.o $(AE_DIR)/zmalloc.o example-ae.c $(STLIBNAME)
hiredis-example-ae: examples/example-ae.c adapters/ae.h $(STLIBNAME)
$(CC) -o examples/$@ $(REAL_CFLAGS) $(REAL_LDFLAGS) -I. -I$(AE_DIR) $< $(AE_DIR)/ae.o $(AE_DIR)/zmalloc.o $(AE_DIR)/../deps/jemalloc/lib/libjemalloc.a -pthread $(STLIBNAME)
endif
hiredis-%: %.o $(STLIBNAME)
ifndef LIBUV_DIR
hiredis-example-libuv:
@echo "Please specify LIBUV_DIR (e.g. ../libuv/)"
@false
else
hiredis-example-libuv: examples/example-libuv.c adapters/libuv.h $(STLIBNAME)
$(CC) -o examples/$@ $(REAL_CFLAGS) $(REAL_LDFLAGS) -I. -I$(LIBUV_DIR)/include $< $(LIBUV_DIR)/.libs/libuv.a -lpthread $(STLIBNAME)
endif
hiredis-example: examples/example.c $(STLIBNAME)
$(CC) -o examples/$@ $(REAL_CFLAGS) $(REAL_LDFLAGS) -I. $< $(STLIBNAME)
examples: $(EXAMPLES)
hiredis-test: test.o $(STLIBNAME)
$(CC) -o $@ $(REAL_LDFLAGS) $< $(STLIBNAME)
test: hiredis-test
./hiredis-test
check: hiredis-test
echo \
"daemonize yes\n" \
"pidfile /tmp/hiredis-test-redis.pid\n" \
"port 56379\n" \
"bind 127.0.0.1\n" \
"unixsocket /tmp/hiredis-test-redis.sock" \
| redis-server -
./hiredis-test -h 127.0.0.1 -p 56379 -s /tmp/hiredis-test-redis.sock || \
@echo "$$REDIS_TEST_CONFIG" | $(REDIS_SERVER) -
./hiredis-test -h 127.0.0.1 -p $(REDIS_PORT) -s /tmp/hiredis-test-redis.sock || \
( kill `cat /tmp/hiredis-test-redis.pid` && false )
kill `cat /tmp/hiredis-test-redis.pid`
@@ -98,17 +118,15 @@ check: hiredis-test
$(CC) -std=c99 -pedantic -c $(REAL_CFLAGS) $<
clean:
rm -rf $(DYLIBNAME) $(STLIBNAME) $(BINS) hiredis-example* *.o *.gcda *.gcno *.gcov
rm -rf $(DYLIBNAME) $(STLIBNAME) $(TESTS) examples/hiredis-example* *.o *.gcda *.gcno *.gcov
dep:
$(CC) -MM *.c
# Installation related variables and target
PREFIX?=/usr/local
INCLUDE_PATH?=include/hiredis
LIBRARY_PATH?=lib
INSTALL_INCLUDE_PATH= $(PREFIX)/$(INCLUDE_PATH)
INSTALL_LIBRARY_PATH= $(PREFIX)/$(LIBRARY_PATH)
INSTALL_INCLUDE_PATH= $(PREFIX)/include/hiredis
INSTALL_LIBRARY_PATH= $(PREFIX)/lib
ifeq ($(uname_S),SunOS)
INSTALL?= cp -r
deps/hiredis/README.md
Vendored
+7 -2
View File
@@ -1,3 +1,5 @@
[![Build Status](https://travis-ci.org/redis/hiredis.png)](https://travis-ci.org/redis/hiredis)
# HIREDIS
Hiredis is a minimalistic C client library for the [Redis](http://redis.io/) database.
@@ -44,7 +46,7 @@ After trying to connect to Redis using `redisConnect` you should
check the `err` field to see if establishing the connection was successful:
redisContext *c = redisConnect("127.0.0.1", 6379);
if (c->err) {
if (c != NULL && c->err) {
printf("Error: %s\n", c->errstr);
// handle error
}
@@ -66,7 +68,7 @@ When you need to pass binary safe strings in a command, the `%b` specifier can b
used. Together with a pointer to the string, it requires a `size_t` length argument
of the string:
reply = redisCommand(context, "SET foo %b", value, valuelen);
reply = redisCommand(context, "SET foo %b", value, (size_t) valuelen);
Internally, Hiredis splits the command in different arguments and will
convert it to the protocol used to communicate with Redis.
@@ -337,6 +339,9 @@ and a reply object (as described above) via `void **reply`. The returned status
can be either `REDIS_OK` or `REDIS_ERR`, where the latter means something went
wrong (either a protocol error, or an out of memory error).
The parser limits the level of nesting for multi bulk payloads to 7. If the
multi bulk nesting level is higher than this, the parser returns an error.
### Customizing replies
The function `redisReaderGetReply` creates `redisReply` and makes the function
deps/hiredis/adapters/libuv.h
Vendored
+121
View File
@@ -0,0 +1,121 @@
#ifndef __HIREDIS_LIBUV_H__
#define __HIREDIS_LIBUV_H__
#include <uv.h>
#include "../hiredis.h"
#include "../async.h"
#include <string.h>
typedef struct redisLibuvEvents {
redisAsyncContext* context;
uv_poll_t handle;
int events;
} redisLibuvEvents;
int redisLibuvAttach(redisAsyncContext*, uv_loop_t*);
static void redisLibuvPoll(uv_poll_t* handle, int status, int events) {
redisLibuvEvents* p = (redisLibuvEvents*)handle->data;
if (status != 0) {
return;
}
if (events & UV_READABLE) {
redisAsyncHandleRead(p->context);
}
if (events & UV_WRITABLE) {
redisAsyncHandleWrite(p->context);
}
}
static void redisLibuvAddRead(void *privdata) {
redisLibuvEvents* p = (redisLibuvEvents*)privdata;
p->events |= UV_READABLE;
uv_poll_start(&p->handle, p->events, redisLibuvPoll);
}
static void redisLibuvDelRead(void *privdata) {
redisLibuvEvents* p = (redisLibuvEvents*)privdata;
p->events &= ~UV_READABLE;
if (p->events) {
uv_poll_start(&p->handle, p->events, redisLibuvPoll);
} else {
uv_poll_stop(&p->handle);
}
}
static void redisLibuvAddWrite(void *privdata) {
redisLibuvEvents* p = (redisLibuvEvents*)privdata;
p->events |= UV_WRITABLE;
uv_poll_start(&p->handle, p->events, redisLibuvPoll);
}
static void redisLibuvDelWrite(void *privdata) {
redisLibuvEvents* p = (redisLibuvEvents*)privdata;
p->events &= ~UV_WRITABLE;
if (p->events) {
uv_poll_start(&p->handle, p->events, redisLibuvPoll);
} else {
uv_poll_stop(&p->handle);
}
}
static void on_close(uv_handle_t* handle) {
redisLibuvEvents* p = (redisLibuvEvents*)handle->data;
free(p);
}
static void redisLibuvCleanup(void *privdata) {
redisLibuvEvents* p = (redisLibuvEvents*)privdata;
uv_close((uv_handle_t*)&p->handle, on_close);
}
static int redisLibuvAttach(redisAsyncContext* ac, uv_loop_t* loop) {
redisContext *c = &(ac->c);
if (ac->ev.data != NULL) {
return REDIS_ERR;
}
ac->ev.addRead = redisLibuvAddRead;
ac->ev.delRead = redisLibuvDelRead;
ac->ev.addWrite = redisLibuvAddWrite;
ac->ev.delWrite = redisLibuvDelWrite;
ac->ev.cleanup = redisLibuvCleanup;
redisLibuvEvents* p = (redisLibuvEvents*)malloc(sizeof(*p));
if (!p) {
return REDIS_ERR;
}
memset(p, 0, sizeof(*p));
if (uv_poll_init(loop, &p->handle, c->fd) != 0) {
return REDIS_ERR;
}
ac->ev.data = p;
p->handle.data = p;
p->context = ac;
return REDIS_OK;
}
#endif
deps/hiredis/async.c
Vendored
+52 -22
View File
@@ -32,9 +32,7 @@
#include "fmacros.h"
#include <stdlib.h>
#include <string.h>
#ifndef _WIN32
#include <strings.h>
#endif
#include <assert.h>
#include <ctype.h>
#include <errno.h>
@@ -42,9 +40,6 @@
#include "net.h"
#include "dict.c"
#include "sds.h"
#ifdef _WIN32
#include "../../src/win32fixes.h"
#endif
#define _EL_ADD_READ(ctx) do { \
if ((ctx)->ev.addRead) (ctx)->ev.addRead((ctx)->ev.data); \
@@ -62,32 +57,28 @@
if ((ctx)->ev.cleanup) (ctx)->ev.cleanup((ctx)->ev.data); \
} while(0);
#ifdef _WIN32
#define strcasecmp _stricmp
#define strncasecmp _strnicmp
#endif
/* Forward declaration of function in hiredis.c */
void __redisAppendCommand(redisContext *c, char *cmd, size_t len);
/* Functions managing dictionary of callbacks for pub/sub. */
static unsigned int callbackHash(const void *key) {
return dictGenHashFunction((unsigned char*)key,(int)sdslen((char*)key));
return dictGenHashFunction((const unsigned char *)key,
sdslen((const sds)key));
}
static void *callbackValDup(void *privdata, const void *src) {
redisCallback *dup = malloc(sizeof(*dup));
((void) privdata);
redisCallback *dup = malloc(sizeof(*dup));
memcpy(dup,src,sizeof(*dup));
return dup;
}
static int callbackKeyCompare(void *privdata, const void *key1, const void *key2) {
size_t l1, l2;
int l1, l2;
((void) privdata);
l1 = sdslen((sds)key1);
l2 = sdslen((sds)key2);
l1 = sdslen((const sds)key1);
l2 = sdslen((const sds)key2);
if (l1 != l2) return 0;
return memcmp(key1,key2,l1) == 0;
}
@@ -112,7 +103,12 @@ static dictType callbackDict = {
};
static redisAsyncContext *redisAsyncInitialize(redisContext *c) {
redisAsyncContext *ac = realloc(c,sizeof(redisAsyncContext));
redisAsyncContext *ac;
ac = realloc(c,sizeof(redisAsyncContext));
if (ac == NULL)
return NULL;
c = &(ac->c);
/* The regular connect functions will always set the flag REDIS_CONNECTED.
@@ -152,15 +148,45 @@ static void __redisAsyncCopyError(redisAsyncContext *ac) {
}
redisAsyncContext *redisAsyncConnect(const char *ip, int port) {
redisContext *c = redisConnectNonBlock(ip,port);
redisContext *c;
redisAsyncContext *ac;
c = redisConnectNonBlock(ip,port);
if (c == NULL)
return NULL;
ac = redisAsyncInitialize(c);
if (ac == NULL) {
redisFree(c);
return NULL;
}
__redisAsyncCopyError(ac);
return ac;
}
redisAsyncContext *redisAsyncConnectBind(const char *ip, int port,
const char *source_addr) {
redisContext *c = redisConnectBindNonBlock(ip,port,source_addr);
redisAsyncContext *ac = redisAsyncInitialize(c);
__redisAsyncCopyError(ac);
return ac;
}
redisAsyncContext *redisAsyncConnectUnix(const char *path) {
redisContext *c = redisConnectUnixNonBlock(path);
redisAsyncContext *ac = redisAsyncInitialize(c);
redisContext *c;
redisAsyncContext *ac;
c = redisConnectUnixNonBlock(path);
if (c == NULL)
return NULL;
ac = redisAsyncInitialize(c);
if (ac == NULL) {
redisFree(c);
return NULL;
}
__redisAsyncCopyError(ac);
return ac;
}
@@ -192,6 +218,9 @@ static int __redisPushCallback(redisCallbackList *list, redisCallback *source) {
/* Copy callback from stack to heap */
cb = malloc(sizeof(*cb));
if (cb == NULL)
return REDIS_ERR_OOM;
if (source != NULL) {
memcpy(cb,source,sizeof(*cb));
cb->next = NULL;
@@ -370,7 +399,7 @@ static int __redisGetSubscribeCallback(redisAsyncContext *ac, redisReply *reply,
void redisProcessCallbacks(redisAsyncContext *ac) {
redisContext *c = &(ac->c);
redisCallback cb;
redisCallback cb = {NULL, NULL, NULL};
void *reply = NULL;
int status;
@@ -382,7 +411,7 @@ void redisProcessCallbacks(redisAsyncContext *ac) {
__redisAsyncDisconnect(ac);
return;
}
/* If monitor mode, repush callback */
if(c->flags & REDIS_MONITORING) {
__redisPushCallback(&ac->replies,&cb);
@@ -414,6 +443,7 @@ void redisProcessCallbacks(redisAsyncContext *ac) {
if (((redisReply*)reply)->type == REDIS_REPLY_ERROR) {
c->err = REDIS_ERR_OTHER;
snprintf(c->errstr,sizeof(c->errstr),"%s",((redisReply*)reply)->str);
c->reader->fn->freeObject(reply);
__redisAsyncDisconnect(ac);
return;
}
@@ -452,7 +482,7 @@ void redisProcessCallbacks(redisAsyncContext *ac) {
static int __redisAsyncHandleConnect(redisAsyncContext *ac) {
redisContext *c = &(ac->c);
if (redisCheckSocketError(c,(int)c->fd) == REDIS_ERR) {
if (redisCheckSocketError(c) == REDIS_ERR) {
/* Try again later when connect(2) is still in progress. */
if (errno == EINPROGRESS)
return REDIS_OK;
deps/hiredis/async.h
Vendored
+1
View File
@@ -102,6 +102,7 @@ typedef struct redisAsyncContext {
/* Functions that proxy to hiredis */
redisAsyncContext *redisAsyncConnect(const char *ip, int port);
redisAsyncContext *redisAsyncConnectBind(const char *ip, int port, const char *source_addr);
redisAsyncContext *redisAsyncConnectUnix(const char *path);
int redisAsyncSetConnectCallback(redisAsyncContext *ac, redisConnectCallback *fn);
int redisAsyncSetDisconnectCallback(redisAsyncContext *ac, redisDisconnectCallback *fn);
deps/hiredis/dict.c
Vendored
-65
View File
@@ -42,11 +42,7 @@
/* -------------------------- private prototypes ---------------------------- */
static int _dictExpandIfNeeded(dict *ht);
#ifdef _WIN32
static size_t _dictNextPower(size_t size);
#else
static unsigned long _dictNextPower(unsigned long size);
#endif
static int _dictKeyIndex(dict *ht, const void *key);
static int _dictInit(dict *ht, dictType *type, void *privDataPtr);
@@ -89,53 +85,6 @@ static int _dictInit(dict *ht, dictType *type, void *privDataPtr) {
}
/* Expand or create the hashtable */
#ifdef _WIN32
static int dictExpand(dict *ht, size_t size) {
dict n; /* the new hashtable */
size_t realsize = _dictNextPower(size), i;
/* the size is invalid if it is smaller than the number of
* elements already inside the hashtable */
if (ht->used > size)
return DICT_ERR;
_dictInit(&n, ht->type, ht->privdata);
n.size = realsize;
n.sizemask = realsize-1;
n.table = calloc(realsize,sizeof(dictEntry*));
/* Copy all the elements from the old to the new table:
* note that if the old hash table is empty ht->size is zero,
* so dictExpand just creates an hash table. */
n.used = ht->used;
for (i = 0; i < ht->size && ht->used > 0; i++) {
dictEntry *he, *nextHe;
if (ht->table[i] == NULL) continue;
/* For each hash entry on this slot... */
he = ht->table[i];
while(he) {
unsigned int h;
nextHe = he->next;
/* Get the new element index */
h = dictHashKey(ht, he->key) & n.sizemask;
he->next = n.table[h];
n.table[h] = he;
ht->used--;
/* Pass to the next element */
he = nextHe;
}
}
assert(ht->used == 0);
free(ht->table);
/* Remap the new hashtable in the old */
*ht = n;
return DICT_OK;
}
#else
static int dictExpand(dict *ht, unsigned long size) {
dict n; /* the new hashtable */
unsigned long realsize = _dictNextPower(size), i;
@@ -181,7 +130,6 @@ static int dictExpand(dict *ht, unsigned long size) {
*ht = n;
return DICT_OK;
}
#endif
/* Add an element to the target hash table */
static int dictAdd(dict *ht, void *key, void *val) {
@@ -355,18 +303,6 @@ static int _dictExpandIfNeeded(dict *ht) {
}
/* Our hash table capability is a power of two */
#ifdef _WIN32
static size_t _dictNextPower(size_t size) {
size_t i = DICT_HT_INITIAL_SIZE;
if (size >= LONG_MAX) return LONG_MAX;
while(1) {
if (i >= size)
return i;
i *= 2;
}
}
#else
static unsigned long _dictNextPower(unsigned long size) {
unsigned long i = DICT_HT_INITIAL_SIZE;
@@ -377,7 +313,6 @@ static unsigned long _dictNextPower(unsigned long size) {
i *= 2;
}
}
#endif
/* Returns the index of a free slot that can be populated with
* an hash entry for the given 'key'.
deps/hiredis/dict.h
Vendored
-10
View File
@@ -60,15 +60,9 @@ typedef struct dictType {
typedef struct dict {
dictEntry **table;
dictType *type;
#ifdef _WIN32
size_t size;
size_t sizemask;
size_t used;
#else
unsigned long size;
unsigned long sizemask;
unsigned long used;
#endif
void *privdata;
} dict;
@@ -119,11 +113,7 @@ typedef struct dictIterator {
/* API */
static unsigned int dictGenHashFunction(const unsigned char *buf, int len);
static dict *dictCreate(dictType *type, void *privDataPtr);
#ifdef _WIN32
static int dictExpand(dict *ht, size_t size);
#else
static int dictExpand(dict *ht, unsigned long size);
#endif
static int dictAdd(dict *ht, void *key, void *val);
static int dictReplace(dict *ht, void *key, void *val);
static int dictDelete(dict *ht, const void *key);
deps/hiredis/example-ae.c → deps/hiredis/examples/example-ae.c
Vendored
+10 -4
View File
@@ -2,9 +2,10 @@
#include <stdlib.h>
#include <string.h>
#include <signal.h>
#include "hiredis.h"
#include "async.h"
#include "adapters/ae.h"
#include <hiredis.h>
#include <async.h>
#include <adapters/ae.h>
/* Put event loop in the global scope, so it can be explicitly stopped */
static aeEventLoop *loop;
@@ -21,17 +22,22 @@ void getCallback(redisAsyncContext *c, void *r, void *privdata) {
void connectCallback(const redisAsyncContext *c, int status) {
if (status != REDIS_OK) {
printf("Error: %s\n", c->errstr);
aeStop(loop);
return;
}
printf("Connected...\n");
}
void disconnectCallback(const redisAsyncContext *c, int status) {
if (status != REDIS_OK) {
printf("Error: %s\n", c->errstr);
aeStop(loop);
return;
}
printf("Disconnected...\n");
aeStop(loop);
}
int main (int argc, char **argv) {
@@ -44,7 +50,7 @@ int main (int argc, char **argv) {
return 1;
}
loop = aeCreateEventLoop();
loop = aeCreateEventLoop(64);
redisAeAttach(loop, c);
redisAsyncSetConnectCallback(c,connectCallback);
redisAsyncSetDisconnectCallback(c,disconnectCallback);
deps/hiredis/example-libev.c → deps/hiredis/examples/example-libev.c
Vendored
+4 -3
View File
@@ -2,9 +2,10 @@
#include <stdlib.h>
#include <string.h>
#include <signal.h>
#include "hiredis.h"
#include "async.h"
#include "adapters/libev.h"
#include <hiredis.h>
#include <async.h>
#include <adapters/libev.h>
void getCallback(redisAsyncContext *c, void *r, void *privdata) {
redisReply *reply = r;
deps/hiredis/example-libevent.c → deps/hiredis/examples/example-libevent.c
Vendored
+4 -3
View File
@@ -2,9 +2,10 @@
#include <stdlib.h>
#include <string.h>
#include <signal.h>
#include "hiredis.h"
#include "async.h"
#include "adapters/libevent.h"
#include <hiredis.h>
#include <async.h>
#include <adapters/libevent.h>
void getCallback(redisAsyncContext *c, void *r, void *privdata) {
redisReply *reply = r;
deps/hiredis/examples/example-libuv.c
Vendored
+53
View File
@@ -0,0 +1,53 @@
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <signal.h>
#include <hiredis.h>
#include <async.h>
#include <adapters/libuv.h>
void getCallback(redisAsyncContext *c, void *r, void *privdata) {
redisReply *reply = r;
if (reply == NULL) return;
printf("argv[%s]: %s\n", (char*)privdata, reply->str);
/* Disconnect after receiving the reply to GET */
redisAsyncDisconnect(c);
}
void connectCallback(const redisAsyncContext *c, int status) {
if (status != REDIS_OK) {
printf("Error: %s\n", c->errstr);
return;
}
printf("Connected...\n");
}
void disconnectCallback(const redisAsyncContext *c, int status) {
if (status != REDIS_OK) {
printf("Error: %s\n", c->errstr);
return;
}
printf("Disconnected...\n");
}
int main (int argc, char **argv) {
signal(SIGPIPE, SIG_IGN);
uv_loop_t* loop = uv_default_loop();
redisAsyncContext *c = redisAsyncConnect("127.0.0.1", 6379);
if (c->err) {
/* Let *c leak for now... */
printf("Error: %s\n", c->errstr);
return 1;
}
redisLibuvAttach(c,loop);
redisAsyncSetConnectCallback(c,connectCallback);
redisAsyncSetDisconnectCallback(c,disconnectCallback);
redisAsyncCommand(c, NULL, NULL, "SET key %b", argv[argc-1], strlen(argv[argc-1]));
redisAsyncCommand(c, getCallback, (char*)"end-1", "GET key");
uv_run(loop, UV_RUN_DEFAULT);
return 0;
}
deps/hiredis/example.c → deps/hiredis/examples/example.c
Vendored
+16 -6
View File
@@ -2,17 +2,24 @@
#include <stdlib.h>
#include <string.h>
#include "hiredis.h"
#include <hiredis.h>
int main(void) {
int main(int argc, char **argv) {
unsigned int j;
redisContext *c;
redisReply *reply;
const char *hostname = (argc > 1) ? argv[1] : "127.0.0.1";
int port = (argc > 2) ? atoi(argv[2]) : 6379;
struct timeval timeout = { 1, 500000 }; // 1.5 seconds
c = redisConnectWithTimeout((char*)"127.0.0.2", 6379, timeout);
if (c->err) {
printf("Connection error: %s\n", c->errstr);
c = redisConnectWithTimeout(hostname, port, timeout);
if (c == NULL || c->err) {
if (c) {
printf("Connection error: %s\n", c->errstr);
redisFree(c);
} else {
printf("Connection error: can't allocate redis context\n");
}
exit(1);
}
@@ -27,7 +34,7 @@ int main(void) {
freeReplyObject(reply);
/* Set a key using binary safe API */
reply = redisCommand(c,"SET %b %b", "bar", 3, "hello", 5);
reply = redisCommand(c,"SET %b %b", "bar", (size_t) 3, "hello", (size_t) 5);
printf("SET (binary API): %s\n", reply->str);
freeReplyObject(reply);
@@ -64,5 +71,8 @@ int main(void) {
}
freeReplyObject(reply);
/* Disconnects and frees the context */
redisFree(c);
return 0;
}
deps/hiredis/fmacros.h
Vendored
+8
View File
@@ -5,6 +5,10 @@
#define _BSD_SOURCE
#endif
#if defined(_AIX)
#define _ALL_SOURCE
#endif
#if defined(__sun__)
#define _POSIX_C_SOURCE 200112L
#elif defined(__linux__) || defined(__OpenBSD__) || defined(__NetBSD__)
@@ -13,4 +17,8 @@
#define _XOPEN_SOURCE
#endif
#if __APPLE__ && __MACH__
#define _OSX
#endif
#endif
deps/hiredis/hiredis.c
Vendored
+100 -104
View File
@@ -32,9 +32,7 @@
#include "fmacros.h"
#include <string.h>
#include <stdlib.h>
#ifndef _WIN32
#include <unistd.h>
#endif
#include <unistd.h>
#include <assert.h>
#include <errno.h>
#include <ctype.h>
@@ -42,9 +40,6 @@
#include "hiredis.h"
#include "net.h"
#include "sds.h"
#ifdef _WIN32
#include "../../src/win32fixes.h"
#endif
static redisReply *createReplyObject(int type);
static void *createStringObject(const redisReadTask *task, char *str, size_t len);
@@ -121,7 +116,7 @@ static void *createStringObject(const redisReadTask *task, char *str, size_t len
memcpy(buf,str,len);
buf[len] = '\0';
r->str = buf;
r->len = (int)len;
r->len = len;
if (task->parent) {
parent = task->parent->obj;
@@ -264,7 +259,7 @@ static char *readBytes(redisReader *r, unsigned int bytes) {
/* Find pointer to \r\n. */
static char *seekNewline(char *s, size_t len) {
int pos = 0;
int _len = (int)(len-1);
int _len = len-1;
/* Position should be < len-1 because the character at "pos" should be
* followed by a \n. Note that strchr cannot be used because it doesn't
@@ -324,7 +319,7 @@ static char *readLine(redisReader *r, int *_len) {
p = r->buf+r->pos;
s = seekNewline(p,(r->len-r->pos));
if (s != NULL) {
len = (int)(s-(r->buf+r->pos));
len = s-(r->buf+r->pos);
r->pos += len+2; /* skip \r\n */
if (_len) *_len = len;
return p;
@@ -395,11 +390,7 @@ static int processBulkItem(redisReader *r) {
redisReadTask *cur = &(r->rstack[r->ridx]);
void *obj = NULL;
char *p, *s;
#ifdef _WIN32
long long len;
#else
long len;
#endif
unsigned long bytelen;
int success = 0;
@@ -407,7 +398,7 @@ static int processBulkItem(redisReader *r) {
s = seekNewline(p,r->len-r->pos);
if (s != NULL) {
p = r->buf+r->pos;
bytelen = (int)(s-(r->buf+r->pos)+2); /* include \r\n */
bytelen = s-(r->buf+r->pos)+2; /* include \r\n */
len = readLongLong(p);
if (len < 0) {
@@ -419,10 +410,10 @@ static int processBulkItem(redisReader *r) {
success = 1;
} else {
/* Only continue when the buffer contains the entire bulk item. */
bytelen += (unsigned long)len+2; /* include \r\n */
bytelen += len+2; /* include \r\n */
if (r->pos+bytelen <= r->len) {
if (r->fn && r->fn->createString)
obj = r->fn->createString(cur,s+2,(size_t)len);
obj = r->fn->createString(cur,s+2,len);
else
obj = (void*)REDIS_REPLY_STRING;
success = 1;
@@ -452,11 +443,7 @@ static int processMultiBulkItem(redisReader *r) {
redisReadTask *cur = &(r->rstack[r->ridx]);
void *obj;
char *p;
#ifdef _WIN32
long long elements;
#else
long elements;
#endif
int root = 0;
/* Set error for nested multi bulks with depth > 7 */
@@ -484,7 +471,7 @@ static int processMultiBulkItem(redisReader *r) {
moveToNextTask(r);
} else {
if (r->fn && r->fn->createArray)
obj = r->fn->createArray(cur,(int)elements);
obj = r->fn->createArray(cur,elements);
else
obj = (void*)REDIS_REPLY_ARRAY;
@@ -495,7 +482,7 @@ static int processMultiBulkItem(redisReader *r) {
/* Modify task stack when there are more than 0 elements. */
if (elements > 0) {
cur->elements = (int)elements;
cur->elements = elements;
cur->obj = obj;
r->ridx++;
r->rstack[r->ridx].type = -1;
@@ -663,7 +650,7 @@ int redisReaderGetReply(redisReader *r, void **reply) {
/* Discard part of the buffer when we've consumed at least 1k, to avoid
* doing unnecessary calls to memmove() in sds.c. */
if (r->pos >= 1024) {
r->buf = sdsrange(r->buf,(int)r->pos,-1);
sdsrange(r->buf,r->pos,-1);
r->pos = 0;
r->len = sdslen(r->buf);
}
@@ -693,7 +680,7 @@ static int intlen(int i) {
/* Helper that calculates the bulk length given a certain string length. */
static size_t bulklen(size_t len) {
return (size_t)(1+intlen((int)len)+2+(int)len+2);
return 1+intlen(len)+2+len+2;
}
int redisvFormatCommand(char **target, const char *format, va_list ap) {
@@ -724,7 +711,7 @@ int redisvFormatCommand(char **target, const char *format, va_list ap) {
if (newargv == NULL) goto err;
curargv = newargv;
curargv[argc++] = curarg;
totlen += (int)bulklen(sdslen(curarg));
totlen += bulklen(sdslen(curarg));
/* curarg is put in argv so it can be overwritten. */
curarg = sdsempty();
@@ -876,7 +863,7 @@ int redisvFormatCommand(char **target, const char *format, va_list ap) {
if (newargv == NULL) goto err;
curargv = newargv;
curargv[argc++] = curarg;
totlen += (int)bulklen(sdslen(curarg));
totlen += bulklen(sdslen(curarg));
} else {
sdsfree(curarg);
}
@@ -888,18 +875,14 @@ int redisvFormatCommand(char **target, const char *format, va_list ap) {
totlen += 1+intlen(argc)+2;
/* Build the command at protocol level */
cmd = (char *)malloc(totlen+1);
cmd = malloc(totlen+1);
if (cmd == NULL) goto err;
pos = sprintf(cmd,"*%d\r\n",argc);
for (j = 0; j < argc; j++) {
#ifdef _WIN32
pos += sprintf(cmd+pos,"$%llu\r\n",(unsigned long long)sdslen(curargv[j]));
#else
pos += sprintf(cmd+pos,"$%zu\r\n",sdslen(curargv[j]));
#endif
memcpy(cmd+pos,curargv[j],sdslen(curargv[j]));
pos += (int)sdslen(curargv[j]);
pos += sdslen(curargv[j]);
sdsfree(curargv[j]);
cmd[pos++] = '\r';
cmd[pos++] = '\n';
@@ -934,7 +917,7 @@ err:
* %b represents a binary safe string
*
* When using %b you need to provide both the pointer to the string
* and the length in bytes. Examples:
* and the length in bytes as a size_t. Examples:
*
* len = redisFormatCommand(target, "GET %s", mykey);
* len = redisFormatCommand(target, "SET %s %b", mykey, myval, myvallen);
@@ -963,7 +946,7 @@ int redisFormatCommandArgv(char **target, int argc, const char **argv, const siz
totlen = 1+intlen(argc)+2;
for (j = 0; j < argc; j++) {
len = argvlen ? argvlen[j] : strlen(argv[j]);
totlen += (int)bulklen(len);
totlen += bulklen(len);
}
/* Build the command at protocol level */
@@ -974,13 +957,9 @@ int redisFormatCommandArgv(char **target, int argc, const char **argv, const siz
pos = sprintf(cmd,"*%d\r\n",argc);
for (j = 0; j < argc; j++) {
len = argvlen ? argvlen[j] : strlen(argv[j]);
#ifdef _WIN32
pos += sprintf(cmd+pos,"$%llu\r\n",(unsigned long long)len);
#else
pos += sprintf(cmd+pos,"$%zu\r\n",len);
#endif
memcpy(cmd+pos,argv[j],len);
pos += (int)len;
pos += len;
cmd[pos++] = '\r';
cmd[pos++] = '\n';
}
@@ -1023,11 +1002,7 @@ static redisContext *redisContextInit(void) {
void redisFree(redisContext *c) {
if (c->fd > 0)
#ifdef _WIN32
closesocket(c->fd);
#else
close(c->fd);
#endif
if (c->obuf != NULL)
sdsfree(c->obuf);
if (c->reader != NULL)
@@ -1035,73 +1010,122 @@ void redisFree(redisContext *c) {
free(c);
}
int redisFreeKeepFd(redisContext *c) {
int fd = c->fd;
c->fd = -1;
redisFree(c);
return fd;
}
/* Connect to a Redis instance. On error the field error in the returned
* context will be set to the return value of the error function.
* When no set of reply functions is given, the default set will be used. */
redisContext *redisConnect(const char *ip, int port) {
redisContext *c = redisContextInit();
redisContext *c;
c = redisContextInit();
if (c == NULL)
return NULL;
c->flags |= REDIS_BLOCK;
redisContextConnectTcp(c,ip,port,NULL);
return c;
}
redisContext *redisConnectWithTimeout(const char *ip, int port, struct timeval tv) {
redisContext *c = redisContextInit();
redisContext *redisConnectWithTimeout(const char *ip, int port, const struct timeval tv) {
redisContext *c;
c = redisContextInit();
if (c == NULL)
return NULL;
c->flags |= REDIS_BLOCK;
redisContextConnectTcp(c,ip,port,&tv);
return c;
}
redisContext *redisConnectNonBlock(const char *ip, int port) {
redisContext *c = redisContextInit();
redisContext *c;
c = redisContextInit();
if (c == NULL)
return NULL;
c->flags &= ~REDIS_BLOCK;
redisContextConnectTcp(c,ip,port,NULL);
return c;
}
redisContext *redisConnectUnix(const char *path) {
redisContext *redisConnectBindNonBlock(const char *ip, int port,
const char *source_addr) {
redisContext *c = redisContextInit();
c->flags &= ~REDIS_BLOCK;
redisContextConnectBindTcp(c,ip,port,NULL,source_addr);
return c;
}
redisContext *redisConnectUnix(const char *path) {
redisContext *c;
c = redisContextInit();
if (c == NULL)
return NULL;
c->flags |= REDIS_BLOCK;
redisContextConnectUnix(c,path,NULL);
return c;
}
redisContext *redisConnectUnixWithTimeout(const char *path, struct timeval tv) {
redisContext *c = redisContextInit();
redisContext *redisConnectUnixWithTimeout(const char *path, const struct timeval tv) {
redisContext *c;
c = redisContextInit();
if (c == NULL)
return NULL;
c->flags |= REDIS_BLOCK;
redisContextConnectUnix(c,path,&tv);
return c;
}
redisContext *redisConnectUnixNonBlock(const char *path) {
redisContext *c = redisContextInit();
redisContext *c;
c = redisContextInit();
if (c == NULL)
return NULL;
c->flags &= ~REDIS_BLOCK;
redisContextConnectUnix(c,path,NULL);
return c;
}
/* initializers if caller handles connection */
redisContext *redisConnected() {
redisContext *c = redisContextInit();
c->fd = -1;
c->flags |= REDIS_BLOCK;
return c;
}
redisContext *redisConnectFd(int fd) {
redisContext *c;
redisContext *redisConnectedNonBlock() {
redisContext *c = redisContextInit();
c->fd = -1;
c->flags &= ~REDIS_BLOCK;
c = redisContextInit();
if (c == NULL)
return NULL;
c->fd = fd;
c->flags |= REDIS_BLOCK | REDIS_CONNECTED;
return c;
}
/* Set read/write timeout on a blocking socket. */
int redisSetTimeout(redisContext *c, struct timeval tv) {
int redisSetTimeout(redisContext *c, const struct timeval tv) {
if (c->flags & REDIS_BLOCK)
return redisContextSetTimeout(c,tv);
return REDIS_ERR;
}
/* Enable connection KeepAlive. */
int redisEnableKeepAlive(redisContext *c) {
if (redisKeepAlive(c, REDIS_KEEPALIVE_INTERVAL) != REDIS_OK)
return REDIS_ERR;
return REDIS_OK;
}
/* Use this function to handle a read event on the descriptor. It will try
* and read some bytes from the socket and feed them to the reply parser.
*
@@ -1115,18 +1139,9 @@ int redisBufferRead(redisContext *c) {
if (c->err)
return REDIS_ERR;
#ifdef _WIN32
nread = recv((SOCKET)c->fd,buf,sizeof(buf),0);
if (nread == -1) {
errno = WSAGetLastError();
if ((errno == ENOENT) || (errno == WSAEWOULDBLOCK))
errno = EAGAIN;
}
#else
nread = read(c->fd,buf,sizeof(buf));
#endif
if (nread == -1) {
if (errno == EAGAIN && !(c->flags & REDIS_BLOCK)) {
if ((errno == EAGAIN && !(c->flags & REDIS_BLOCK)) || (errno == EINTR)) {
/* Try again later */
} else {
__redisSetError(c,REDIS_ERR_IO,NULL);
@@ -1144,25 +1159,6 @@ int redisBufferRead(redisContext *c) {
return REDIS_OK;
}
/* Use this function if the caller has already read the data. It will
* feed bytes to the reply parser.
*
* After this function is called, you may use redisContextReadReply to
* see if there is a reply available. */
int redisBufferReadDone(redisContext *c, char *buf, int nread) {
if (nread == 0) {
__redisSetError(c,REDIS_ERR_EOF, sdsnew("Server closed the connection"));
return REDIS_ERR;
} else {
if (redisReaderFeed(c->reader,buf,nread) != REDIS_OK) {
__redisSetError(c,c->reader->err,c->reader->errstr);
return REDIS_ERR;
}
}
return REDIS_OK;
}
/* Write the output buffer to the socket.
*
* Returns REDIS_OK when the buffer is empty, or (a part of) the buffer was
@@ -1180,18 +1176,9 @@ int redisBufferWrite(redisContext *c, int *done) {
return REDIS_ERR;
if (sdslen(c->obuf) > 0) {
#ifdef _WIN32
nwritten = send((SOCKET)c->fd,c->obuf,(int)sdslen(c->obuf),0);
if (nwritten == -1) {
errno = WSAGetLastError();
if ((errno == ENOENT) || (errno == WSAEWOULDBLOCK))
errno = EAGAIN;
}
#else
nwritten = write(c->fd,c->obuf,sdslen(c->obuf));
#endif
if (nwritten == -1) {
if (errno == EAGAIN && !(c->flags & REDIS_BLOCK)) {
if ((errno == EAGAIN && !(c->flags & REDIS_BLOCK)) || (errno == EINTR)) {
/* Try again later */
} else {
__redisSetError(c,REDIS_ERR_IO,NULL);
@@ -1202,7 +1189,7 @@ int redisBufferWrite(redisContext *c, int *done) {
sdsfree(c->obuf);
c->obuf = sdsempty();
} else {
c->obuf = sdsrange(c->obuf,nwritten,-1);
sdsrange(c->obuf,nwritten,-1);
}
}
}
@@ -1257,7 +1244,7 @@ int redisGetReply(redisContext *c, void **reply) {
* is used, you need to call redisGetReply yourself to retrieve
* the reply (or replies in pub/sub).
*/
int __redisAppendCommand(redisContext *c, char *cmd, size_t len) {
int __redisAppendCommand(redisContext *c, const char *cmd, size_t len) {
sds newbuf;
newbuf = sdscatlen(c->obuf,cmd,len);
@@ -1270,6 +1257,15 @@ int __redisAppendCommand(redisContext *c, char *cmd, size_t len) {
return REDIS_OK;
}
int redisAppendFormattedCommand(redisContext *c, const char *cmd, size_t len) {
if (__redisAppendCommand(c, cmd, len) != REDIS_OK) {
return REDIS_ERR;
}
return REDIS_OK;
}
int redisvAppendCommand(redisContext *c, const char *format, va_list ap) {
char *cmd;
int len;
deps/hiredis/hiredis.h
Vendored
+15 -25
View File
@@ -33,24 +33,11 @@
#define __HIREDIS_H
#include <stdio.h> /* for size_t */
#include <stdarg.h> /* for va_list */
#ifndef _WIN32
#include <sys/time.h> /* for struct timeval */
#endif
#ifdef _WIN32
#ifndef FD_SETSIZE
#define FD_SETSIZE 16000
#endif
#include <winsock2.h>
#include <windows.h>
#ifndef va_copy
#define va_copy(d,s) d = (s)
#endif
#endif
#define HIREDIS_MAJOR 0
#define HIREDIS_MINOR 10
#define HIREDIS_PATCH 1
#define HIREDIS_MINOR 11
#define HIREDIS_PATCH 0
#define REDIS_ERR -1
#define REDIS_OK 0
@@ -101,6 +88,8 @@
#define REDIS_READER_MAX_BUF (1024*16) /* Default max unused reader buffer. */
#define REDIS_KEEPALIVE_INTERVAL 15 /* seconds */
#ifdef __cplusplus
extern "C" {
#endif
@@ -177,29 +166,26 @@ int redisFormatCommandArgv(char **target, int argc, const char **argv, const siz
typedef struct redisContext {
int err; /* Error flags, 0 when there is no error */
char errstr[128]; /* String representation of error when applicable */
#ifdef _WIN32
SOCKET fd;
#else
int fd;
#endif
int flags;
char *obuf; /* Write buffer */
redisReader *reader; /* Protocol reader */
} redisContext;
redisContext *redisConnect(const char *ip, int port);
redisContext *redisConnectWithTimeout(const char *ip, int port, struct timeval tv);
redisContext *redisConnectWithTimeout(const char *ip, int port, const struct timeval tv);
redisContext *redisConnectNonBlock(const char *ip, int port);
redisContext *redisConnectBindNonBlock(const char *ip, int port, const char *source_addr);
redisContext *redisConnectUnix(const char *path);
redisContext *redisConnectUnixWithTimeout(const char *path, struct timeval tv);
redisContext *redisConnectUnixWithTimeout(const char *path, const struct timeval tv);
redisContext *redisConnectUnixNonBlock(const char *path);
redisContext *redisConnected();
redisContext *redisConnectedNonBlock();
int redisSetTimeout(redisContext *c, struct timeval tv);
redisContext *redisConnectFd(int fd);
int redisSetTimeout(redisContext *c, const struct timeval tv);
int redisEnableKeepAlive(redisContext *c);
void redisFree(redisContext *c);
int redisFreeKeepFd(redisContext *c);
int redisBufferRead(redisContext *c);
int redisBufferWrite(redisContext *c, int *done);
int redisBufferReadDone(redisContext *c, char *buf, int nread);
/* In a blocking context, this function first checks if there are unconsumed
* replies to return and returns one if so. Otherwise, it flushes the output
@@ -208,6 +194,10 @@ int redisBufferReadDone(redisContext *c, char *buf, int nread);
int redisGetReply(redisContext *c, void **reply);
int redisGetReplyFromReader(redisContext *c, void **reply);
/* Write a formatted command to the output buffer. Use these functions in blocking mode
* to get a pipeline of commands. */
int redisAppendFormattedCommand(redisContext *c, const char *cmd, size_t len);
/* Write a command to the output buffer. Use these functions in blocking mode
* to get a pipeline of commands. */
int redisvAppendCommand(redisContext *c, const char *format, va_list ap);
deps/hiredis/net.c
Vendored
+131 -257
View File
@@ -32,11 +32,6 @@
#include "fmacros.h"
#include <sys/types.h>
#ifdef _WIN32
#ifndef FD_SETSIZE
#define FD_SETSIZE 16000
#endif
#else
#include <sys/socket.h>
#include <sys/select.h>
#include <sys/un.h>
@@ -44,29 +39,30 @@
#include <netinet/tcp.h>
#include <arpa/inet.h>
#include <unistd.h>
#include <netdb.h>
#endif
#include <fcntl.h>
#include <string.h>
#include <netdb.h>
#include <errno.h>
#include <stdarg.h>
#include <stdio.h>
#ifndef _WIN32
#include <poll.h>
#endif
#include <limits.h>
#include "net.h"
#include "sds.h"
#ifdef _WIN32
#include "../../src/win32fixes.h"
#endif
/* Defined in hiredis.c */
void __redisSetError(redisContext *c, int type, const char *str);
static void redisContextCloseFd(redisContext *c) {
if (c && c->fd >= 0) {
close(c->fd);
c->fd = -1;
}
}
static void __redisSetErrorFromErrno(redisContext *c, int type, const char *prefix) {
char buf[128];
char buf[128] = { 0 };
size_t len = 0;
if (prefix != NULL)
@@ -75,90 +71,40 @@ static void __redisSetErrorFromErrno(redisContext *c, int type, const char *pref
__redisSetError(c,type,buf);
}
static int redisSetReuseAddr(redisContext *c, int fd) {
static int redisSetReuseAddr(redisContext *c) {
int on = 1;
if (setsockopt(fd, SOL_SOCKET, SO_REUSEADDR, &on, sizeof(on)) == -1) {
if (setsockopt(c->fd, SOL_SOCKET, SO_REUSEADDR, &on, sizeof(on)) == -1) {
__redisSetErrorFromErrno(c,REDIS_ERR_IO,NULL);
close(fd);
redisContextCloseFd(c);
return REDIS_ERR;
}
return REDIS_OK;
}
#ifdef _WIN32
static int redisCreateSocket(redisContext *c, int type) {
SOCKET s;
int on=1;
s = socket(type, SOCK_STREAM, IPPROTO_TCP);
if (s == INVALID_SOCKET) {
__redisSetError(c,REDIS_ERR_IO,sdscatprintf(sdsempty(), "socket error: %d\n", WSAGetLastError()));
return REDIS_ERR;
}
if (type == AF_INET) {
LINGER l;
l.l_onoff = 1;
l.l_linger = 2;
setsockopt((int)s, SOL_SOCKET, SO_LINGER, (const char *) &l, sizeof(l));
if (setsockopt((int)s, SOL_SOCKET, SO_REUSEADDR, (const char *) &on, sizeof(on)) == -1) {
__redisSetError(c,REDIS_ERR_IO,NULL);
closesocket(s);
return REDIS_ERR;
}
}
return (int)s;
}
#else
static int redisCreateSocket(redisContext *c, int type) {
int s;
if ((s = socket(type, SOCK_STREAM, 0)) == -1) {
__redisSetErrorFromErrno(c,REDIS_ERR_IO,NULL);
return REDIS_ERR;
}
c->fd = s;
if (type == AF_INET) {
if (redisSetReuseAddr(c,s) == REDIS_ERR) {
if (redisSetReuseAddr(c) == REDIS_ERR) {
return REDIS_ERR;
}
}
return s;
}
}
#endif
#ifdef _WIN32
static int redisSetBlocking(redisContext *c, int fd, int blocking) {
/* If iMode = 0, blocking is enabled; */
/* If iMode != 0, non-blocking mode is enabled. */
u_long flags;
if (blocking)
flags = (u_long)0;
else
flags = (u_long)1;
if (ioctlsocket((SOCKET)fd, FIONBIO, &flags) == SOCKET_ERROR) {
errno = WSAGetLastError();
__redisSetError(c,REDIS_ERR_IO,
sdscatprintf(sdsempty(), "ioctlsocket(FIONBIO): %d\n", errno));
closesocket(fd);
return REDIS_ERR;
};
return REDIS_OK;
}
#else
static int redisSetBlocking(redisContext *c, int fd, int blocking) {
static int redisSetBlocking(redisContext *c, int blocking) {
int flags;
/* Set the socket nonblocking.
* Note that fcntl(2) for F_GETFL and F_SETFL can't be
* interrupted by a signal. */
if ((flags = fcntl(fd, F_GETFL)) == -1) {
if ((flags = fcntl(c->fd, F_GETFL)) == -1) {
__redisSetErrorFromErrno(c,REDIS_ERR_IO,"fcntl(F_GETFL)");
close(fd);
redisContextCloseFd(c);
return REDIS_ERR;
}
@@ -167,92 +113,81 @@ static int redisSetBlocking(redisContext *c, int fd, int blocking) {
else
flags |= O_NONBLOCK;
if (fcntl(fd, F_SETFL, flags) == -1) {
if (fcntl(c->fd, F_SETFL, flags) == -1) {
__redisSetErrorFromErrno(c,REDIS_ERR_IO,"fcntl(F_SETFL)");
close(fd);
redisContextCloseFd(c);
return REDIS_ERR;
}
return REDIS_OK;
}
int redisKeepAlive(redisContext *c, int interval) {
int val = 1;
int fd = c->fd;
if (setsockopt(fd, SOL_SOCKET, SO_KEEPALIVE, &val, sizeof(val)) == -1){
__redisSetError(c,REDIS_ERR_OTHER,strerror(errno));
return REDIS_ERR;
}
val = interval;
#ifdef _OSX
if (setsockopt(fd, IPPROTO_TCP, TCP_KEEPALIVE, &val, sizeof(val)) < 0) {
__redisSetError(c,REDIS_ERR_OTHER,strerror(errno));
return REDIS_ERR;
}
#else
#ifndef __sun
val = interval;
if (setsockopt(fd, IPPROTO_TCP, TCP_KEEPIDLE, &val, sizeof(val)) < 0) {
__redisSetError(c,REDIS_ERR_OTHER,strerror(errno));
return REDIS_ERR;
}
val = interval/3;
if (val == 0) val = 1;
if (setsockopt(fd, IPPROTO_TCP, TCP_KEEPINTVL, &val, sizeof(val)) < 0) {
__redisSetError(c,REDIS_ERR_OTHER,strerror(errno));
return REDIS_ERR;
}
val = 3;
if (setsockopt(fd, IPPROTO_TCP, TCP_KEEPCNT, &val, sizeof(val)) < 0) {
__redisSetError(c,REDIS_ERR_OTHER,strerror(errno));
return REDIS_ERR;
}
#endif
#endif
#ifdef _WIN32
static int redisSetTcpNoDelay(redisContext *c, int fd) {
int yes = 1;
if (setsockopt((SOCKET)fd, IPPROTO_TCP, TCP_NODELAY, (const char *)&yes, sizeof(yes)) == -1) {
__redisSetError(c,REDIS_ERR_IO,
sdscatprintf(sdsempty(), "setsockopt(TCP_NODELAY): %d", (int)GetLastError()));
closesocket(fd);
return REDIS_ERR;
}
return REDIS_OK;
}
#else
static int redisSetTcpNoDelay(redisContext *c, int fd) {
static int redisSetTcpNoDelay(redisContext *c) {
int yes = 1;
if (setsockopt(fd, IPPROTO_TCP, TCP_NODELAY, &yes, sizeof(yes)) == -1) {
if (setsockopt(c->fd, IPPROTO_TCP, TCP_NODELAY, &yes, sizeof(yes)) == -1) {
__redisSetErrorFromErrno(c,REDIS_ERR_IO,"setsockopt(TCP_NODELAY)");
close(fd);
redisContextCloseFd(c);
return REDIS_ERR;
}
return REDIS_OK;
}
#endif
#ifdef _WIN32
static int redisContextWaitReady(redisContext *c, int fd, const struct timeval *timeout) {
struct timeval to;
struct timeval *toptr = NULL;
fd_set wfd;
/* Only use timeout when not NULL. */
if (timeout != NULL) {
to = *timeout;
toptr = &to;
}
if (errno == EINPROGRESS) {
FD_ZERO(&wfd);
FD_SET((SOCKET)fd, &wfd);
if (select(FD_SETSIZE, NULL, &wfd, NULL, toptr) == -1) {
__redisSetErrorFromErrno(c,REDIS_ERR_IO,"select(2)");
closesocket(fd);
return REDIS_ERR;
}
if (!FD_ISSET(fd, &wfd)) {
errno = WSAGetLastError();
__redisSetErrorFromErrno(c,REDIS_ERR_IO,NULL);
closesocket(fd);
return REDIS_ERR;
}
if (redisCheckSocketError(c, fd) != REDIS_OK)
return REDIS_ERR;
return REDIS_OK;
}
__redisSetErrorFromErrno(c,REDIS_ERR_IO,NULL);
closesocket(fd);
return REDIS_ERR;
}
#else
#define __MAX_MSEC (((LONG_MAX) - 999) / 1000)
static int redisContextWaitReady(redisContext *c, int fd, const struct timeval *timeout) {
static int redisContextWaitReady(redisContext *c, const struct timeval *timeout) {
struct pollfd wfd[1];
long msec;
msec = -1;
wfd[0].fd = fd;
wfd[0].fd = c->fd;
wfd[0].events = POLLOUT;
/* Only use timeout when not NULL. */
if (timeout != NULL) {
if (timeout->tv_usec > 1000000 || timeout->tv_sec > __MAX_MSEC) {
close(fd);
__redisSetErrorFromErrno(c, REDIS_ERR_IO, NULL);
redisContextCloseFd(c);
return REDIS_ERR;
}
@@ -268,78 +203,45 @@ static int redisContextWaitReady(redisContext *c, int fd, const struct timeval *
if ((res = poll(wfd, 1, msec)) == -1) {
__redisSetErrorFromErrno(c, REDIS_ERR_IO, "poll(2)");
close(fd);
redisContextCloseFd(c);
return REDIS_ERR;
} else if (res == 0) {
errno = ETIMEDOUT;
__redisSetErrorFromErrno(c,REDIS_ERR_IO,NULL);
close(fd);
redisContextCloseFd(c);
return REDIS_ERR;
}
if (redisCheckSocketError(c, fd) != REDIS_OK)
if (redisCheckSocketError(c) != REDIS_OK)
return REDIS_ERR;
return REDIS_OK;
}
__redisSetErrorFromErrno(c,REDIS_ERR_IO,NULL);
close(fd);
redisContextCloseFd(c);
return REDIS_ERR;
}
#endif
int redisCheckSocketError(redisContext *c, int fd) {
int redisCheckSocketError(redisContext *c) {
int err = 0;
socklen_t errlen = sizeof(err);
#ifdef _WIN32
if (getsockopt(fd, SOL_SOCKET, SO_ERROR, (char *)&err, &errlen) == SOCKET_ERROR) {
errno = WSAGetLastError();
if (getsockopt(c->fd, SOL_SOCKET, SO_ERROR, &err, &errlen) == -1) {
__redisSetErrorFromErrno(c,REDIS_ERR_IO,"getsockopt(SO_ERROR)");
closesocket(fd);
return REDIS_ERR;
}
#else
if (getsockopt(fd, SOL_SOCKET, SO_ERROR, &err, &errlen) == -1) {
__redisSetErrorFromErrno(c,REDIS_ERR_IO,"getsockopt(SO_ERROR)");
close(fd);
return REDIS_ERR;
}
#endif
if (err) {
errno = err;
__redisSetErrorFromErrno(c,REDIS_ERR_IO,NULL);
#ifdef _WIN32
closesocket(fd);
#else
close(fd);
#endif
return REDIS_ERR;
}
return REDIS_OK;
}
#ifdef _WIN32
int redisContextSetTimeout(redisContext *c, struct timeval tv) {
DWORD ms = (tv.tv_sec * 1000) + (tv.tv_usec / 1000);
if (setsockopt((int)c->fd,SOL_SOCKET,SO_RCVTIMEO,(const char *)&ms,sizeof(ms)) == SOCKET_ERROR ) {
errno = WSAGetLastError();
__redisSetErrorFromErrno(c,REDIS_ERR_IO,"setsockopt(SO_RCVTIMEO)");
return REDIS_ERR;
}
if (setsockopt((int)c->fd,SOL_SOCKET,SO_SNDTIMEO,(const char *)&ms,sizeof(ms)) == SOCKET_ERROR ) {
errno = WSAGetLastError();
__redisSetErrorFromErrno(c,REDIS_ERR_IO,"setsockopt(SO_SNDTIMEO)");
return REDIS_ERR;
}
return REDIS_OK;
}
#else
int redisContextSetTimeout(redisContext *c, struct timeval tv) {
int redisContextSetTimeout(redisContext *c, const struct timeval tv) {
if (setsockopt(c->fd,SOL_SOCKET,SO_RCVTIMEO,&tv,sizeof(tv)) == -1) {
__redisSetErrorFromErrno(c,REDIS_ERR_IO,"setsockopt(SO_RCVTIMEO)");
return REDIS_ERR;
@@ -350,70 +252,13 @@ int redisContextSetTimeout(redisContext *c, struct timeval tv) {
}
return REDIS_OK;
}
#endif
#ifdef _WIN32
int redisContextConnectTcp(redisContext *c, const char *addr, int port, struct timeval *timeout) {
int s;
int blocking = (c->flags & REDIS_BLOCK);
struct sockaddr_in sa;
unsigned long inAddress;
if ((s = redisCreateSocket(c,AF_INET)) < 0)
return REDIS_ERR;
sa.sin_family = AF_INET;
sa.sin_port = htons(port);
inAddress = inet_addr(addr);
if (inAddress == INADDR_NONE || inAddress == INADDR_ANY) {
struct hostent *he;
he = gethostbyname(addr);
if (he == NULL) {
__redisSetError(c,REDIS_ERR_OTHER,
sdscatprintf(sdsempty(),"can't resolve: %s\n", addr));
closesocket(s);
return REDIS_ERR;
}
memcpy(&sa.sin_addr, he->h_addr, sizeof(struct in_addr));
}
else {
sa.sin_addr.s_addr = inAddress;
}
if (redisSetTcpNoDelay(c,s) != REDIS_OK)
return REDIS_ERR;
if (connect((SOCKET)s, (struct sockaddr*)&sa, sizeof(sa)) == -1) {
errno = WSAGetLastError();
if ((errno == WSAEINVAL) || (errno == WSAEWOULDBLOCK))
errno = EINPROGRESS;
if (errno == EINPROGRESS && !blocking) {
/* This is ok. */
} else {
if (redisContextWaitReady(c,s,timeout) != REDIS_OK)
return REDIS_ERR;
}
}
if (blocking) {
if (redisSetBlocking(c,s,1) != REDIS_OK)
return REDIS_ERR;
} else {
if (redisSetBlocking(c,s,0) != REDIS_OK)
return REDIS_ERR;
}
c->fd = s;
c->flags |= REDIS_CONNECTED;
return REDIS_OK;
}
#else
int redisContextConnectTcp(redisContext *c, const char *addr, int port, struct timeval *timeout) {
static int _redisContextConnectTcp(redisContext *c, const char *addr, int port,
const struct timeval *timeout,
const char *source_addr) {
int s, rv;
char _port[6]; /* strlen("65535"); */
struct addrinfo hints, *servinfo, *p;
struct addrinfo hints, *servinfo, *bservinfo, *p, *b;
int blocking = (c->flags & REDIS_BLOCK);
snprintf(_port, 6, "%d", port);
@@ -421,33 +266,64 @@ int redisContextConnectTcp(redisContext *c, const char *addr, int port, struct t
hints.ai_family = AF_INET;
hints.ai_socktype = SOCK_STREAM;
/* Try with IPv6 if no IPv4 address was found. We do it in this order since
* in a Redis client you can't afford to test if you have IPv6 connectivity
* as this would add latency to every connect. Otherwise a more sensible
* route could be: Use IPv6 if both addresses are available and there is IPv6
* connectivity. */
if ((rv = getaddrinfo(addr,_port,&hints,&servinfo)) != 0) {
__redisSetError(c,REDIS_ERR_OTHER,gai_strerror(rv));
return REDIS_ERR;
hints.ai_family = AF_INET6;
if ((rv = getaddrinfo(addr,_port,&hints,&servinfo)) != 0) {
__redisSetError(c,REDIS_ERR_OTHER,gai_strerror(rv));
return REDIS_ERR;
}
}
for (p = servinfo; p != NULL; p = p->ai_next) {
if ((s = socket(p->ai_family,p->ai_socktype,p->ai_protocol)) == -1)
continue;
if (redisSetBlocking(c,s,0) != REDIS_OK)
c->fd = s;
if (redisSetBlocking(c,0) != REDIS_OK)
goto error;
if (source_addr) {
int bound = 0;
/* Using getaddrinfo saves us from self-determining IPv4 vs IPv6 */
if ((rv = getaddrinfo(source_addr, NULL, &hints, &bservinfo)) != 0) {
char buf[128];
snprintf(buf,sizeof(buf),"Can't get addr: %s",gai_strerror(rv));
__redisSetError(c,REDIS_ERR_OTHER,buf);
goto error;
}
for (b = bservinfo; b != NULL; b = b->ai_next) {
if (bind(s,b->ai_addr,b->ai_addrlen) != -1) {
bound = 1;
break;
}
}
freeaddrinfo(bservinfo);
if (!bound) {
char buf[128];
snprintf(buf,sizeof(buf),"Can't bind socket: %s",strerror(errno));
__redisSetError(c,REDIS_ERR_OTHER,buf);
goto error;
}
}
if (connect(s,p->ai_addr,p->ai_addrlen) == -1) {
if (errno == EHOSTUNREACH) {
close(s);
redisContextCloseFd(c);
continue;
} else if (errno == EINPROGRESS && !blocking) {
/* This is ok. */
} else {
if (redisContextWaitReady(c,s,timeout) != REDIS_OK)
if (redisContextWaitReady(c,timeout) != REDIS_OK)
goto error;
}
}
if (blocking && redisSetBlocking(c,s,1) != REDIS_OK)
if (blocking && redisSetBlocking(c,1) != REDIS_OK)
goto error;
if (redisSetTcpNoDelay(c,s) != REDIS_OK)
if (redisSetTcpNoDelay(c) != REDIS_OK)
goto error;
c->fd = s;
c->flags |= REDIS_CONNECTED;
rv = REDIS_OK;
goto end;
@@ -465,44 +341,42 @@ end:
freeaddrinfo(servinfo);
return rv; // Need to return REDIS_OK if alright
}
#endif
#ifdef _WIN32
int redisContextConnectUnix(redisContext *c, const char *path, struct timeval *timeout) {
(void) timeout;
__redisSetError(c,REDIS_ERR_IO,
sdscatprintf(sdsempty(),"Unix sockets are not suported on Windows platform. (%s)\n", path));
return REDIS_ERR;
int redisContextConnectTcp(redisContext *c, const char *addr, int port,
const struct timeval *timeout) {
return _redisContextConnectTcp(c, addr, port, timeout, NULL);
}
#else
int redisContextConnectUnix(redisContext *c, const char *path, struct timeval *timeout) {
int s;
int redisContextConnectBindTcp(redisContext *c, const char *addr, int port,
const struct timeval *timeout,
const char *source_addr) {
return _redisContextConnectTcp(c, addr, port, timeout, source_addr);
}
int redisContextConnectUnix(redisContext *c, const char *path, const struct timeval *timeout) {
int blocking = (c->flags & REDIS_BLOCK);
struct sockaddr_un sa;
if ((s = redisCreateSocket(c,AF_LOCAL)) < 0)
if (redisCreateSocket(c,AF_LOCAL) < 0)
return REDIS_ERR;
if (redisSetBlocking(c,s,0) != REDIS_OK)
if (redisSetBlocking(c,0) != REDIS_OK)
return REDIS_ERR;
sa.sun_family = AF_LOCAL;
strncpy(sa.sun_path,path,sizeof(sa.sun_path)-1);
if (connect(s, (struct sockaddr*)&sa, sizeof(sa)) == -1) {
if (connect(c->fd, (struct sockaddr*)&sa, sizeof(sa)) == -1) {
if (errno == EINPROGRESS && !blocking) {
/* This is ok. */
} else {
if (redisContextWaitReady(c,s,timeout) != REDIS_OK)
if (redisContextWaitReady(c,timeout) != REDIS_OK)
return REDIS_ERR;
}
}
/* Reset socket to be blocking after connect(2). */
if (blocking && redisSetBlocking(c,s,1) != REDIS_OK)
if (blocking && redisSetBlocking(c,1) != REDIS_OK)
return REDIS_ERR;
c->fd = s;
c->flags |= REDIS_CONNECTED;
return REDIS_OK;
}
#endif
deps/hiredis/net.h
Vendored
+9 -5
View File
@@ -35,13 +35,17 @@
#include "hiredis.h"
#if defined(__sun)
#if defined(__sun) || defined(_AIX)
#define AF_LOCAL AF_UNIX
#endif
int redisCheckSocketError(redisContext *c, int fd);
int redisContextSetTimeout(redisContext *c, struct timeval tv);
int redisContextConnectTcp(redisContext *c, const char *addr, int port, struct timeval *timeout);
int redisContextConnectUnix(redisContext *c, const char *path, struct timeval *timeout);
int redisCheckSocketError(redisContext *c);
int redisContextSetTimeout(redisContext *c, const struct timeval tv);
int redisContextConnectTcp(redisContext *c, const char *addr, int port, const struct timeval *timeout);
int redisContextConnectBindTcp(redisContext *c, const char *addr, int port,
const struct timeval *timeout,
const char *source_addr);
int redisContextConnectUnix(redisContext *c, const char *path, const struct timeval *timeout);
int redisKeepAlive(redisContext *c, int interval);
#endif
deps/hiredis/sds.c
Vendored
+653 -155
View File
File diff suppressed because it is too large Load Diff
deps/hiredis/sds.h
Vendored
+25 -17
View File
@@ -31,29 +31,26 @@
#ifndef __SDS_H
#define __SDS_H
#define SDS_MAX_PREALLOC (1024*1024)
#include <sys/types.h>
#include <stdarg.h>
#ifdef _WIN32
#define inline __inline
#define va_copy(d,s) d = (s)
#endif
typedef char *sds;
struct sdshdr {
int len;
int free;
unsigned int len;
unsigned int free;
char buf[];
};
static inline size_t sdslen(const sds s) {
struct sdshdr *sh = (struct sdshdr *)(s-(sizeof(struct sdshdr)));
struct sdshdr *sh = (void*)(s-(sizeof(struct sdshdr)));
return sh->len;
}
static inline size_t sdsavail(const sds s) {
struct sdshdr *sh = (struct sdshdr *)(s-(sizeof(struct sdshdr)));
struct sdshdr *sh = (void*)(s-(sizeof(struct sdshdr)));
return sh->free;
}
@@ -63,12 +60,13 @@ sds sdsempty(void);
size_t sdslen(const sds s);
sds sdsdup(const sds s);
void sdsfree(sds s);
size_t sdsavail(sds s);
size_t sdsavail(const sds s);
sds sdsgrowzero(sds s, size_t len);
sds sdscatlen(sds s, const void *t, size_t len);
sds sdscat(sds s, const char *t);
sds sdscpylen(sds s, char *t, size_t len);
sds sdscpy(sds s, char *t);
sds sdscatsds(sds s, const sds t);
sds sdscpylen(sds s, const char *t, size_t len);
sds sdscpy(sds s, const char *t);
sds sdscatvprintf(sds s, const char *fmt, va_list ap);
#ifdef __GNUC__
@@ -78,16 +76,26 @@ sds sdscatprintf(sds s, const char *fmt, ...)
sds sdscatprintf(sds s, const char *fmt, ...);
#endif
sds sdscatfmt(sds s, char const *fmt, ...);
sds sdstrim(sds s, const char *cset);
sds sdsrange(sds s, int start, int end);
void sdsrange(sds s, int start, int end);
void sdsupdatelen(sds s);
int sdscmp(sds s1, sds s2);
sds *sdssplitlen(char *s, int len, char *sep, int seplen, int *count);
void sdsclear(sds s);
int sdscmp(const sds s1, const sds s2);
sds *sdssplitlen(const char *s, int len, const char *sep, int seplen, int *count);
void sdsfreesplitres(sds *tokens, int count);
void sdstolower(sds s);
void sdstoupper(sds s);
sds sdsfromlonglong(long long value);
sds sdscatrepr(sds s, char *p, size_t len);
sds *sdssplitargs(char *line, int *argc);
sds sdscatrepr(sds s, const char *p, size_t len);
sds *sdssplitargs(const char *line, int *argc);
sds sdsmapchars(sds s, const char *from, const char *to, size_t setlen);
sds sdsjoin(char **argv, int argc, char *sep);
/* Low level functions exposed to the user API */
sds sdsMakeRoomFor(sds s, size_t addlen);
void sdsIncrLen(sds s, int incr);
sds sdsRemoveFreeSpace(sds s);
size_t sdsAllocSize(sds s);
#endif
deps/hiredis/test.c
Vendored
+96 -17
View File
@@ -8,12 +8,14 @@
#include <unistd.h>
#include <signal.h>
#include <errno.h>
#include <limits.h>
#include "hiredis.h"
enum connection_type {
CONN_TCP,
CONN_UNIX
CONN_UNIX,
CONN_FD
};
struct config {
@@ -22,6 +24,7 @@ struct config {
struct {
const char *host;
int port;
struct timeval timeout;
} tcp;
struct {
@@ -48,7 +51,7 @@ static redisContext *select_database(redisContext *c) {
assert(reply != NULL);
freeReplyObject(reply);
/* Make sure the DB is emtpy */
/* Make sure the DB is empty */
reply = redisCommand(c,"DBSIZE");
assert(reply != NULL);
if (reply->type == REDIS_REPLY_INTEGER && reply->integer == 0) {
@@ -62,7 +65,7 @@ static redisContext *select_database(redisContext *c) {
return c;
}
static void disconnect(redisContext *c) {
static int disconnect(redisContext *c, int keep_fd) {
redisReply *reply;
/* Make sure we're on DB 9. */
@@ -73,8 +76,11 @@ static void disconnect(redisContext *c) {
assert(reply != NULL);
freeReplyObject(reply);
/* Free the context as well. */
/* Free the context as well, but keep the fd if requested. */
if (keep_fd)
return redisFreeKeepFd(c);
redisFree(c);
return -1;
}
static redisContext *connect(struct config config) {
@@ -84,11 +90,22 @@ static redisContext *connect(struct config config) {
c = redisConnect(config.tcp.host, config.tcp.port);
} else if (config.type == CONN_UNIX) {
c = redisConnectUnix(config.unix.path);
} else if (config.type == CONN_FD) {
/* Create a dummy connection just to get an fd to inherit */
redisContext *dummy_ctx = redisConnectUnix(config.unix.path);
if (dummy_ctx) {
int fd = disconnect(dummy_ctx, 1);
printf("Connecting to inherited fd %d\n", fd);
c = redisConnectFd(fd);
}
} else {
assert(NULL);
}
if (c->err) {
if (c == NULL) {
printf("Connection error: can't allocate redis context\n");
exit(1);
} else if (c->err) {
printf("Connection error: %s\n", c->errstr);
exit(1);
}
@@ -125,13 +142,13 @@ static void test_format_commands(void) {
free(cmd);
test("Format command with %%b string interpolation: ");
len = redisFormatCommand(&cmd,"SET %b %b","foo",3,"b\0r",3);
len = redisFormatCommand(&cmd,"SET %b %b","foo",(size_t)3,"b\0r",(size_t)3);
test_cond(strncmp(cmd,"*3\r\n$3\r\nSET\r\n$3\r\nfoo\r\n$3\r\nb\0r\r\n",len) == 0 &&
len == 4+4+(3+2)+4+(3+2)+4+(3+2));
free(cmd);
test("Format command with %%b and an empty string: ");
len = redisFormatCommand(&cmd,"SET %b %b","foo",3,"",0);
len = redisFormatCommand(&cmd,"SET %b %b","foo",(size_t)3,"",(size_t)0);
test_cond(strncmp(cmd,"*3\r\n$3\r\nSET\r\n$3\r\nfoo\r\n$0\r\n\r\n",len) == 0 &&
len == 4+4+(3+2)+4+(3+2)+4+(0+2));
free(cmd);
@@ -177,7 +194,7 @@ static void test_format_commands(void) {
FLOAT_WIDTH_TEST(double);
test("Format command with invalid printf format: ");
len = redisFormatCommand(&cmd,"key:%08p %b",(void*)1234,"foo",3);
len = redisFormatCommand(&cmd,"key:%08p %b",(void*)1234,"foo",(size_t)3);
test_cond(len == -1);
const char *argv[3];
@@ -200,10 +217,33 @@ static void test_format_commands(void) {
free(cmd);
}
static void test_append_formatted_commands(struct config config) {
redisContext *c;
redisReply *reply;
char *cmd;
int len;
c = connect(config);
test("Append format command: ");
len = redisFormatCommand(&cmd, "SET foo bar");
test_cond(redisAppendFormattedCommand(c, cmd, len) == REDIS_OK);
assert(redisGetReply(c, (void*)&reply) == REDIS_OK);
free(cmd);
freeReplyObject(reply);
disconnect(c, 0);
}
static void test_reply_reader(void) {
redisReader *reader;
void *reply;
int ret;
int i;
test("Error handling in reply parser: ");
reader = redisReaderCreate();
@@ -225,12 +265,13 @@ static void test_reply_reader(void) {
strcasecmp(reader->errstr,"Protocol error, got \"@\" as reply type byte") == 0);
redisReaderFree(reader);
test("Set error on nested multi bulks with depth > 2: ");
test("Set error on nested multi bulks with depth > 7: ");
reader = redisReaderCreate();
redisReaderFeed(reader,(char*)"*1\r\n",4);
redisReaderFeed(reader,(char*)"*1\r\n",4);
redisReaderFeed(reader,(char*)"*1\r\n",4);
redisReaderFeed(reader,(char*)"*1\r\n",4);
for (i = 0; i < 9; i++) {
redisReaderFeed(reader,(char*)"*1\r\n",4);
}
ret = redisReaderGetReply(reader,NULL);
test_cond(ret == REDIS_ERR &&
strncasecmp(reader->errstr,"No support for",14) == 0);
@@ -284,7 +325,10 @@ static void test_blocking_connection_errors(void) {
c = redisConnect((char*)"idontexist.local", 6379);
test_cond(c->err == REDIS_ERR_OTHER &&
(strcmp(c->errstr,"Name or service not known") == 0 ||
strcmp(c->errstr,"Can't resolve: idontexist.local") == 0));
strcmp(c->errstr,"Can't resolve: idontexist.local") == 0 ||
strcmp(c->errstr,"nodename nor servname provided, or not known") == 0 ||
strcmp(c->errstr,"No address associated with hostname") == 0 ||
strcmp(c->errstr,"no address associated with name") == 0));
redisFree(c);
test("Returns error when the port is not open: ");
@@ -326,7 +370,7 @@ static void test_blocking_connection(struct config config) {
freeReplyObject(reply);
test("%%b String interpolation works: ");
reply = redisCommand(c,"SET %b %b","foo",3,"hello\x00world",11);
reply = redisCommand(c,"SET %b %b","foo",(size_t)3,"hello\x00world",(size_t)11);
freeReplyObject(reply);
reply = redisCommand(c,"GET foo");
test_cond(reply->type == REDIS_REPLY_STRING &&
@@ -374,7 +418,7 @@ static void test_blocking_connection(struct config config) {
strcasecmp(reply->element[1]->str,"pong") == 0);
freeReplyObject(reply);
disconnect(c);
disconnect(c, 0);
}
static void test_blocking_io_errors(struct config config) {
@@ -428,6 +472,30 @@ static void test_blocking_io_errors(struct config config) {
redisFree(c);
}
static void test_invalid_timeout_errors(struct config config) {
redisContext *c;
test("Set error when an invalid timeout usec value is given to redisConnectWithTimeout: ");
config.tcp.timeout.tv_sec = 0;
config.tcp.timeout.tv_usec = 10000001;
c = redisConnectWithTimeout(config.tcp.host, config.tcp.port, config.tcp.timeout);
test_cond(c->err == REDIS_ERR_IO);
test("Set error when an invalid timeout sec value is given to redisConnectWithTimeout: ");
config.tcp.timeout.tv_sec = (((LONG_MAX) - 999) / 1000) + 1;
config.tcp.timeout.tv_usec = 0;
c = redisConnectWithTimeout(config.tcp.host, config.tcp.port, config.tcp.timeout);
test_cond(c->err == REDIS_ERR_IO);
redisFree(c);
}
static void test_throughput(struct config config) {
redisContext *c = connect(config);
redisReply **replies;
@@ -490,7 +558,7 @@ static void test_throughput(struct config config) {
free(replies);
printf("\t(%dx LRANGE with 500 elements (pipelined): %.3fs)\n", num, (t2-t1)/1000000.0);
disconnect(c);
disconnect(c, 0);
}
// static long __test_callback_flags = 0;
@@ -603,6 +671,7 @@ int main(int argc, char **argv) {
}
};
int throughput = 1;
int test_inherit_fd = 1;
/* Ignore broken pipe signal (for I/O error tests). */
signal(SIGPIPE, SIG_IGN);
@@ -621,6 +690,8 @@ int main(int argc, char **argv) {
cfg.unix.path = argv[0];
} else if (argc >= 1 && !strcmp(argv[0],"--skip-throughput")) {
throughput = 0;
} else if (argc >= 1 && !strcmp(argv[0],"--skip-inherit-fd")) {
test_inherit_fd = 0;
} else {
fprintf(stderr, "Invalid argument: %s\n", argv[0]);
exit(1);
@@ -636,6 +707,8 @@ int main(int argc, char **argv) {
cfg.type = CONN_TCP;
test_blocking_connection(cfg);
test_blocking_io_errors(cfg);
test_invalid_timeout_errors(cfg);
test_append_formatted_commands(cfg);
if (throughput) test_throughput(cfg);
printf("\nTesting against Unix socket connection (%s):\n", cfg.unix.path);
@@ -644,6 +717,12 @@ int main(int argc, char **argv) {
test_blocking_io_errors(cfg);
if (throughput) test_throughput(cfg);
if (test_inherit_fd) {
printf("\nTesting against inherited fd (%s):\n", cfg.unix.path);
cfg.type = CONN_FD;
test_blocking_connection(cfg);
}
if (fails) {
printf("*** %d TESTS FAILED ***\n", fails);
return 1;
deps/hiredis/zmalloc.h
Vendored
+13
View File
@@ -0,0 +1,13 @@
/* Drop in replacement for zmalloc.h in order to just use libc malloc without
* any wrappering. */
#ifndef ZMALLOC_H
#define ZMALLOC_H
#define zmalloc malloc
#define zrealloc realloc
#define zcalloc(x) calloc(x,1)
#define zfree free
#define zstrdup strdup
#endif
deps/jemalloc/.gitignore
+72
View File
@@ -0,0 +1,72 @@
/*.gcov.*
/autom4te.cache/
/bin/jemalloc.sh
/config.stamp
/config.log
/config.status
/configure
/doc/html.xsl
/doc/manpages.xsl
/doc/jemalloc.xml
/doc/jemalloc.html
/doc/jemalloc.3
/lib/
/Makefile
/include/jemalloc/internal/jemalloc_internal.h
/include/jemalloc/internal/jemalloc_internal_defs.h
/include/jemalloc/internal/private_namespace.h
/include/jemalloc/internal/private_unnamespace.h
/include/jemalloc/internal/public_namespace.h
/include/jemalloc/internal/public_symbols.txt
/include/jemalloc/internal/public_unnamespace.h
/include/jemalloc/internal/size_classes.h
/include/jemalloc/jemalloc.h
/include/jemalloc/jemalloc_defs.h
/include/jemalloc/jemalloc_macros.h
/include/jemalloc/jemalloc_mangle.h
/include/jemalloc/jemalloc_mangle_jet.h
/include/jemalloc/jemalloc_protos.h
/include/jemalloc/jemalloc_protos_jet.h
/include/jemalloc/jemalloc_rename.h
/src/*.[od]
/src/*.gcda
/src/*.gcno
/test/test.sh
test/include/test/jemalloc_test.h
test/include/test/jemalloc_test_defs.h
/test/integration/[A-Za-z]*
!/test/integration/[A-Za-z]*.*
/test/integration/*.[od]
/test/integration/*.gcda
/test/integration/*.gcno
/test/integration/*.out
/test/src/*.[od]
/test/src/*.gcda
/test/src/*.gcno
/test/stress/[A-Za-z]*
!/test/stress/[A-Za-z]*.*
/test/stress/*.[od]
/test/stress/*.gcda
/test/stress/*.gcno
/test/stress/*.out
/test/unit/[A-Za-z]*
!/test/unit/[A-Za-z]*.*
/test/unit/*.[od]
/test/unit/*.gcda
/test/unit/*.gcno
/test/unit/*.out
/VERSION
deps/jemalloc/COPYING
Vendored
+27
View File
@@ -0,0 +1,27 @@
Unless otherwise specified, files in the jemalloc source distribution are
subject to the following license:
--------------------------------------------------------------------------------
Copyright (C) 2002-2014 Jason Evans <jasone@canonware.com>.
All rights reserved.
Copyright (C) 2007-2012 Mozilla Foundation. All rights reserved.
Copyright (C) 2009-2014 Facebook, Inc. All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are met:
1. Redistributions of source code must retain the above copyright notice(s),
this list of conditions and the following disclaimer.
2. Redistributions in binary form must reproduce the above copyright notice(s),
this list of conditions and the following disclaimer in the documentation
and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDER(S) ``AS IS'' AND ANY EXPRESS
OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO
EVENT SHALL THE COPYRIGHT HOLDER(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE
OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
--------------------------------------------------------------------------------
deps/jemalloc/ChangeLog
Vendored
+548
View File
@@ -0,0 +1,548 @@
Following are change highlights associated with official releases. Important
bug fixes are all mentioned, but internal enhancements are omitted here for
brevity (even though they are more fun to write about). Much more detail can be
found in the git revision history:
https://github.com/jemalloc/jemalloc
* 3.6.0 (March 31, 2014)
This version contains a critical bug fix for a regression present in 3.5.0 and
3.5.1.
Bug fixes:
- Fix a regression in arena_chunk_alloc() that caused crashes during
small/large allocation if chunk allocation failed. In the absence of this
bug, chunk allocation failure would result in allocation failure, e.g. NULL
return from malloc(). This regression was introduced in 3.5.0.
- Fix backtracing for gcc intrinsics-based backtracing by specifying
-fno-omit-frame-pointer to gcc. Note that the application (and all the
libraries it links to) must also be compiled with this option for
backtracing to be reliable.
- Use dss allocation precedence for huge allocations as well as small/large
allocations.
- Fix test assertion failure message formatting. This bug did not manifect on
x86_64 systems because of implementation subtleties in va_list.
- Fix inconsequential test failures for hash and SFMT code.
New features:
- Support heap profiling on FreeBSD. This feature depends on the proc
filesystem being mounted during heap profile dumping.
* 3.5.1 (February 25, 2014)
This version primarily addresses minor bugs in test code.
Bug fixes:
- Configure Solaris/Illumos to use MADV_FREE.
- Fix junk filling for mremap(2)-based huge reallocation. This is only
relevant if configuring with the --enable-mremap option specified.
- Avoid compilation failure if 'restrict' C99 keyword is not supported by the
compiler.
- Add a configure test for SSE2 rather than assuming it is usable on i686
systems. This fixes test compilation errors, especially on 32-bit Linux
systems.
- Fix mallctl argument size mismatches (size_t vs. uint64_t) in the stats unit
test.
- Fix/remove flawed alignment-related overflow tests.
- Prevent compiler optimizations that could change backtraces in the
prof_accum unit test.
* 3.5.0 (January 22, 2014)
This version focuses on refactoring and automated testing, though it also
includes some non-trivial heap profiling optimizations not mentioned below.
New features:
- Add the *allocx() API, which is a successor to the experimental *allocm()
API. The *allocx() functions are slightly simpler to use because they have
fewer parameters, they directly return the results of primary interest, and
mallocx()/rallocx() avoid the strict aliasing pitfall that
allocm()/rallocm() share with posix_memalign(). Note that *allocm() is
slated for removal in the next non-bugfix release.
- Add support for LinuxThreads.
Bug fixes:
- Unless heap profiling is enabled, disable floating point code and don't link
with libm. This, in combination with e.g. EXTRA_CFLAGS=-mno-sse on x64
systems, makes it possible to completely disable floating point register
use. Some versions of glibc neglect to save/restore caller-saved floating
point registers during dynamic lazy symbol loading, and the symbol loading
code uses whatever malloc the application happens to have linked/loaded
with, the result being potential floating point register corruption.
- Report ENOMEM rather than EINVAL if an OOM occurs during heap profiling
backtrace creation in imemalign(). This bug impacted posix_memalign() and
aligned_alloc().
- Fix a file descriptor leak in a prof_dump_maps() error path.
- Fix prof_dump() to close the dump file descriptor for all relevant error
paths.
- Fix rallocm() to use the arena specified by the ALLOCM_ARENA(s) flag for
allocation, not just deallocation.
- Fix a data race for large allocation stats counters.
- Fix a potential infinite loop during thread exit. This bug occurred on
Solaris, and could affect other platforms with similar pthreads TSD
implementations.
- Don't junk-fill reallocations unless usable size changes. This fixes a
violation of the *allocx()/*allocm() semantics.
- Fix growing large reallocation to junk fill new space.
- Fix huge deallocation to junk fill when munmap is disabled.
- Change the default private namespace prefix from empty to je_, and change
--with-private-namespace-prefix so that it prepends an additional prefix
rather than replacing je_. This reduces the likelihood of applications
which statically link jemalloc experiencing symbol name collisions.
- Add missing private namespace mangling (relevant when
--with-private-namespace is specified).
- Add and use JEMALLOC_INLINE_C so that static inline functions are marked as
static even for debug builds.
- Add a missing mutex unlock in a malloc_init_hard() error path. In practice
this error path is never executed.
- Fix numerous bugs in malloc_strotumax() error handling/reporting. These
bugs had no impact except for malformed inputs.
- Fix numerous bugs in malloc_snprintf(). These bugs were not exercised by
existing calls, so they had no impact.
* 3.4.1 (October 20, 2013)
Bug fixes:
- Fix a race in the "arenas.extend" mallctl that could cause memory corruption
of internal data structures and subsequent crashes.
- Fix Valgrind integration flaws that caused Valgrind warnings about reads of
uninitialized memory in:
+ arena chunk headers
+ internal zero-initialized data structures (relevant to tcache and prof
code)
- Preserve errno during the first allocation. A readlink(2) call during
initialization fails unless /etc/malloc.conf exists, so errno was typically
set during the first allocation prior to this fix.
- Fix compilation warnings reported by gcc 4.8.1.
* 3.4.0 (June 2, 2013)
This version is essentially a small bugfix release, but the addition of
aarch64 support requires that the minor version be incremented.
Bug fixes:
- Fix race-triggered deadlocks in chunk_record(). These deadlocks were
typically triggered by multiple threads concurrently deallocating huge
objects.
New features:
- Add support for the aarch64 architecture.
* 3.3.1 (March 6, 2013)
This version fixes bugs that are typically encountered only when utilizing
custom run-time options.
Bug fixes:
- Fix a locking order bug that could cause deadlock during fork if heap
profiling were enabled.
- Fix a chunk recycling bug that could cause the allocator to lose track of
whether a chunk was zeroed. On FreeBSD, NetBSD, and OS X, it could cause
corruption if allocating via sbrk(2) (unlikely unless running with the
"dss:primary" option specified). This was completely harmless on Linux
unless using mlockall(2) (and unlikely even then, unless the
--disable-munmap configure option or the "dss:primary" option was
specified). This regression was introduced in 3.1.0 by the
mlockall(2)/madvise(2) interaction fix.
- Fix TLS-related memory corruption that could occur during thread exit if the
thread never allocated memory. Only the quarantine and prof facilities were
susceptible.
- Fix two quarantine bugs:
+ Internal reallocation of the quarantined object array leaked the old
array.
+ Reallocation failure for internal reallocation of the quarantined object
array (very unlikely) resulted in memory corruption.
- Fix Valgrind integration to annotate all internally allocated memory in a
way that keeps Valgrind happy about internal data structure access.
- Fix building for s390 systems.
* 3.3.0 (January 23, 2013)
This version includes a few minor performance improvements in addition to the
listed new features and bug fixes.
New features:
- Add clipping support to lg_chunk option processing.
- Add the --enable-ivsalloc option.
- Add the --without-export option.
- Add the --disable-zone-allocator option.
Bug fixes:
- Fix "arenas.extend" mallctl to output the number of arenas.
- Fix chunk_recycle() to unconditionally inform Valgrind that returned memory
is undefined.
- Fix build break on FreeBSD related to alloca.h.
* 3.2.0 (November 9, 2012)
In addition to a couple of bug fixes, this version modifies page run
allocation and dirty page purging algorithms in order to better control
page-level virtual memory fragmentation.
Incompatible changes:
- Change the "opt.lg_dirty_mult" default from 5 to 3 (32:1 to 8:1).
Bug fixes:
- Fix dss/mmap allocation precedence code to use recyclable mmap memory only
after primary dss allocation fails.
- Fix deadlock in the "arenas.purge" mallctl. This regression was introduced
in 3.1.0 by the addition of the "arena.<i>.purge" mallctl.
* 3.1.0 (October 16, 2012)
New features:
- Auto-detect whether running inside Valgrind, thus removing the need to
manually specify MALLOC_CONF=valgrind:true.
- Add the "arenas.extend" mallctl, which allows applications to create
manually managed arenas.
- Add the ALLOCM_ARENA() flag for {,r,d}allocm().
- Add the "opt.dss", "arena.<i>.dss", and "stats.arenas.<i>.dss" mallctls,
which provide control over dss/mmap precedence.
- Add the "arena.<i>.purge" mallctl, which obsoletes "arenas.purge".
- Define LG_QUANTUM for hppa.
Incompatible changes:
- Disable tcache by default if running inside Valgrind, in order to avoid
making unallocated objects appear reachable to Valgrind.
- Drop const from malloc_usable_size() argument on Linux.
Bug fixes:
- Fix heap profiling crash if sampled object is freed via realloc(p, 0).
- Remove const from __*_hook variable declarations, so that glibc can modify
them during process forking.
- Fix mlockall(2)/madvise(2) interaction.
- Fix fork(2)-related deadlocks.
- Fix error return value for "thread.tcache.enabled" mallctl.
* 3.0.0 (May 11, 2012)
Although this version adds some major new features, the primary focus is on
internal code cleanup that facilitates maintainability and portability, most
of which is not reflected in the ChangeLog. This is the first release to
incorporate substantial contributions from numerous other developers, and the
result is a more broadly useful allocator (see the git revision history for
contribution details). Note that the license has been unified, thanks to
Facebook granting a license under the same terms as the other copyright
holders (see COPYING).
New features:
- Implement Valgrind support, redzones, and quarantine.
- Add support for additional platforms:
+ FreeBSD
+ Mac OS X Lion
+ MinGW
+ Windows (no support yet for replacing the system malloc)
- Add support for additional architectures:
+ MIPS
+ SH4
+ Tilera
- Add support for cross compiling.
- Add nallocm(), which rounds a request size up to the nearest size class
without actually allocating.
- Implement aligned_alloc() (blame C11).
- Add the "thread.tcache.enabled" mallctl.
- Add the "opt.prof_final" mallctl.
- Update pprof (from gperftools 2.0).
- Add the --with-mangling option.
- Add the --disable-experimental option.
- Add the --disable-munmap option, and make it the default on Linux.
- Add the --enable-mremap option, which disables use of mremap(2) by default.
Incompatible changes:
- Enable stats by default.
- Enable fill by default.
- Disable lazy locking by default.
- Rename the "tcache.flush" mallctl to "thread.tcache.flush".
- Rename the "arenas.pagesize" mallctl to "arenas.page".
- Change the "opt.lg_prof_sample" default from 0 to 19 (1 B to 512 KiB).
- Change the "opt.prof_accum" default from true to false.
Removed features:
- Remove the swap feature, including the "config.swap", "swap.avail",
"swap.prezeroed", "swap.nfds", and "swap.fds" mallctls.
- Remove highruns statistics, including the
"stats.arenas.<i>.bins.<j>.highruns" and
"stats.arenas.<i>.lruns.<j>.highruns" mallctls.
- As part of small size class refactoring, remove the "opt.lg_[qc]space_max",
"arenas.cacheline", "arenas.subpage", "arenas.[tqcs]space_{min,max}", and
"arenas.[tqcs]bins" mallctls.
- Remove the "arenas.chunksize" mallctl.
- Remove the "opt.lg_prof_tcmax" option.
- Remove the "opt.lg_prof_bt_max" option.
- Remove the "opt.lg_tcache_gc_sweep" option.
- Remove the --disable-tiny option, including the "config.tiny" mallctl.
- Remove the --enable-dynamic-page-shift configure option.
- Remove the --enable-sysv configure option.
Bug fixes:
- Fix a statistics-related bug in the "thread.arena" mallctl that could cause
invalid statistics and crashes.
- Work around TLS deallocation via free() on Linux. This bug could cause
write-after-free memory corruption.
- Fix a potential deadlock that could occur during interval- and
growth-triggered heap profile dumps.
- Fix large calloc() zeroing bugs due to dropping chunk map unzeroed flags.
- Fix chunk_alloc_dss() to stop claiming memory is zeroed. This bug could
cause memory corruption and crashes with --enable-dss specified.
- Fix fork-related bugs that could cause deadlock in children between fork
and exec.
- Fix malloc_stats_print() to honor 'b' and 'l' in the opts parameter.
- Fix realloc(p, 0) to act like free(p).
- Do not enforce minimum alignment in memalign().
- Check for NULL pointer in malloc_usable_size().
- Fix an off-by-one heap profile statistics bug that could be observed in
interval- and growth-triggered heap profiles.
- Fix the "epoch" mallctl to update cached stats even if the passed in epoch
is 0.
- Fix bin->runcur management to fix a layout policy bug. This bug did not
affect correctness.
- Fix a bug in choose_arena_hard() that potentially caused more arenas to be
initialized than necessary.
- Add missing "opt.lg_tcache_max" mallctl implementation.
- Use glibc allocator hooks to make mixed allocator usage less likely.
- Fix build issues for --disable-tcache.
- Don't mangle pthread_create() when --with-private-namespace is specified.
* 2.2.5 (November 14, 2011)
Bug fixes:
- Fix huge_ralloc() race when using mremap(2). This is a serious bug that
could cause memory corruption and/or crashes.
- Fix huge_ralloc() to maintain chunk statistics.
- Fix malloc_stats_print(..., "a") output.
* 2.2.4 (November 5, 2011)
Bug fixes:
- Initialize arenas_tsd before using it. This bug existed for 2.2.[0-3], as
well as for --disable-tls builds in earlier releases.
- Do not assume a 4 KiB page size in test/rallocm.c.
* 2.2.3 (August 31, 2011)
This version fixes numerous bugs related to heap profiling.
Bug fixes:
- Fix a prof-related race condition. This bug could cause memory corruption,
but only occurred in non-default configurations (prof_accum:false).
- Fix off-by-one backtracing issues (make sure that prof_alloc_prep() is
excluded from backtraces).
- Fix a prof-related bug in realloc() (only triggered by OOM errors).
- Fix prof-related bugs in allocm() and rallocm().
- Fix prof_tdata_cleanup() for --disable-tls builds.
- Fix a relative include path, to fix objdir builds.
* 2.2.2 (July 30, 2011)
Bug fixes:
- Fix a build error for --disable-tcache.
- Fix assertions in arena_purge() (for real this time).
- Add the --with-private-namespace option. This is a workaround for symbol
conflicts that can inadvertently arise when using static libraries.
* 2.2.1 (March 30, 2011)
Bug fixes:
- Implement atomic operations for x86/x64. This fixes compilation failures
for versions of gcc that are still in wide use.
- Fix an assertion in arena_purge().
* 2.2.0 (March 22, 2011)
This version incorporates several improvements to algorithms and data
structures that tend to reduce fragmentation and increase speed.
New features:
- Add the "stats.cactive" mallctl.
- Update pprof (from google-perftools 1.7).
- Improve backtracing-related configuration logic, and add the
--disable-prof-libgcc option.
Bug fixes:
- Change default symbol visibility from "internal", to "hidden", which
decreases the overhead of library-internal function calls.
- Fix symbol visibility so that it is also set on OS X.
- Fix a build dependency regression caused by the introduction of the .pic.o
suffix for PIC object files.
- Add missing checks for mutex initialization failures.
- Don't use libgcc-based backtracing except on x64, where it is known to work.
- Fix deadlocks on OS X that were due to memory allocation in
pthread_mutex_lock().
- Heap profiling-specific fixes:
+ Fix memory corruption due to integer overflow in small region index
computation, when using a small enough sample interval that profiling
context pointers are stored in small run headers.
+ Fix a bootstrap ordering bug that only occurred with TLS disabled.
+ Fix a rallocm() rsize bug.
+ Fix error detection bugs for aligned memory allocation.
* 2.1.3 (March 14, 2011)
Bug fixes:
- Fix a cpp logic regression (due to the "thread.{de,}allocatedp" mallctl fix
for OS X in 2.1.2).
- Fix a "thread.arena" mallctl bug.
- Fix a thread cache stats merging bug.
* 2.1.2 (March 2, 2011)
Bug fixes:
- Fix "thread.{de,}allocatedp" mallctl for OS X.
- Add missing jemalloc.a to build system.
* 2.1.1 (January 31, 2011)
Bug fixes:
- Fix aligned huge reallocation (affected allocm()).
- Fix the ALLOCM_LG_ALIGN macro definition.
- Fix a heap dumping deadlock.
- Fix a "thread.arena" mallctl bug.
* 2.1.0 (December 3, 2010)
This version incorporates some optimizations that can't quite be considered
bug fixes.
New features:
- Use Linux's mremap(2) for huge object reallocation when possible.
- Avoid locking in mallctl*() when possible.
- Add the "thread.[de]allocatedp" mallctl's.
- Convert the manual page source from roff to DocBook, and generate both roff
and HTML manuals.
Bug fixes:
- Fix a crash due to incorrect bootstrap ordering. This only impacted
--enable-debug --enable-dss configurations.
- Fix a minor statistics bug for mallctl("swap.avail", ...).
* 2.0.1 (October 29, 2010)
Bug fixes:
- Fix a race condition in heap profiling that could cause undefined behavior
if "opt.prof_accum" were disabled.
- Add missing mutex unlocks for some OOM error paths in the heap profiling
code.
- Fix a compilation error for non-C99 builds.
* 2.0.0 (October 24, 2010)
This version focuses on the experimental *allocm() API, and on improved
run-time configuration/introspection. Nonetheless, numerous performance
improvements are also included.
New features:
- Implement the experimental {,r,s,d}allocm() API, which provides a superset
of the functionality available via malloc(), calloc(), posix_memalign(),
realloc(), malloc_usable_size(), and free(). These functions can be used to
allocate/reallocate aligned zeroed memory, ask for optional extra memory
during reallocation, prevent object movement during reallocation, etc.
- Replace JEMALLOC_OPTIONS/JEMALLOC_PROF_PREFIX with MALLOC_CONF, which is
more human-readable, and more flexible. For example:
JEMALLOC_OPTIONS=AJP
is now:
MALLOC_CONF=abort:true,fill:true,stats_print:true
- Port to Apple OS X. Sponsored by Mozilla.
- Make it possible for the application to control thread-->arena mappings via
the "thread.arena" mallctl.
- Add compile-time support for all TLS-related functionality via pthreads TSD.
This is mainly of interest for OS X, which does not support TLS, but has a
TSD implementation with similar performance.
- Override memalign() and valloc() if they are provided by the system.
- Add the "arenas.purge" mallctl, which can be used to synchronously purge all
dirty unused pages.
- Make cumulative heap profiling data optional, so that it is possible to
limit the amount of memory consumed by heap profiling data structures.
- Add per thread allocation counters that can be accessed via the
"thread.allocated" and "thread.deallocated" mallctls.
Incompatible changes:
- Remove JEMALLOC_OPTIONS and malloc_options (see MALLOC_CONF above).
- Increase default backtrace depth from 4 to 128 for heap profiling.
- Disable interval-based profile dumps by default.
Bug fixes:
- Remove bad assertions in fork handler functions. These assertions could
cause aborts for some combinations of configure settings.
- Fix strerror_r() usage to deal with non-standard semantics in GNU libc.
- Fix leak context reporting. This bug tended to cause the number of contexts
to be underreported (though the reported number of objects and bytes were
correct).
- Fix a realloc() bug for large in-place growing reallocation. This bug could
cause memory corruption, but it was hard to trigger.
- Fix an allocation bug for small allocations that could be triggered if
multiple threads raced to create a new run of backing pages.
- Enhance the heap profiler to trigger samples based on usable size, rather
than request size.
- Fix a heap profiling bug due to sometimes losing track of requested object
size for sampled objects.
* 1.0.3 (August 12, 2010)
Bug fixes:
- Fix the libunwind-based implementation of stack backtracing (used for heap
profiling). This bug could cause zero-length backtraces to be reported.
- Add a missing mutex unlock in library initialization code. If multiple
threads raced to initialize malloc, some of them could end up permanently
blocked.
* 1.0.2 (May 11, 2010)
Bug fixes:
- Fix junk filling of large objects, which could cause memory corruption.
- Add MAP_NORESERVE support for chunk mapping, because otherwise virtual
memory limits could cause swap file configuration to fail. Contributed by
Jordan DeLong.
* 1.0.1 (April 14, 2010)
Bug fixes:
- Fix compilation when --enable-fill is specified.
- Fix threads-related profiling bugs that affected accuracy and caused memory
to be leaked during thread exit.
- Fix dirty page purging race conditions that could cause crashes.
- Fix crash in tcache flushing code during thread destruction.
* 1.0.0 (April 11, 2010)
This release focuses on speed and run-time introspection. Numerous
algorithmic improvements make this release substantially faster than its
predecessors.
New features:
- Implement autoconf-based configuration system.
- Add mallctl*(), for the purposes of introspection and run-time
configuration.
- Make it possible for the application to manually flush a thread's cache, via
the "tcache.flush" mallctl.
- Base maximum dirty page count on proportion of active memory.
- Compute various addtional run-time statistics, including per size class
statistics for large objects.
- Expose malloc_stats_print(), which can be called repeatedly by the
application.
- Simplify the malloc_message() signature to only take one string argument,
and incorporate an opaque data pointer argument for use by the application
in combination with malloc_stats_print().
- Add support for allocation backed by one or more swap files, and allow the
application to disable over-commit if swap files are in use.
- Implement allocation profiling and leak checking.
Removed features:
- Remove the dynamic arena rebalancing code, since thread-specific caching
reduces its utility.
Bug fixes:
- Modify chunk allocation to work when address space layout randomization
(ASLR) is in use.
- Fix thread cleanup bugs related to TLS destruction.
- Handle 0-size allocation requests in posix_memalign().
- Fix a chunk leak. The leaked chunks were never touched, so this impacted
virtual memory usage, but not physical memory usage.
* linux_2008082[78]a (August 27/28, 2008)
These snapshot releases are the simple result of incorporating Linux-specific
support into the FreeBSD malloc sources.
--------------------------------------------------------------------------------
vim:filetype=text:textwidth=80
deps/jemalloc/INSTALL
Vendored
+306
View File
@@ -0,0 +1,306 @@
Building and installing jemalloc can be as simple as typing the following while
in the root directory of the source tree:
./configure
make
make install
=== Advanced configuration =====================================================
The 'configure' script supports numerous options that allow control of which
functionality is enabled, where jemalloc is installed, etc. Optionally, pass
any of the following arguments (not a definitive list) to 'configure':
--help
Print a definitive list of options.
--prefix=<install-root-dir>
Set the base directory in which to install. For example:
./configure --prefix=/usr/local
will cause files to be installed into /usr/local/include, /usr/local/lib,
and /usr/local/man.
--with-rpath=<colon-separated-rpath>
Embed one or more library paths, so that libjemalloc can find the libraries
it is linked to. This works only on ELF-based systems.
--with-mangling=<map>
Mangle public symbols specified in <map> which is a comma-separated list of
name:mangled pairs.
For example, to use ld's --wrap option as an alternative method for
overriding libc's malloc implementation, specify something like:
--with-mangling=malloc:__wrap_malloc,free:__wrap_free[...]
Note that mangling happens prior to application of the prefix specified by
--with-jemalloc-prefix, and mangled symbols are then ignored when applying
the prefix.
--with-jemalloc-prefix=<prefix>
Prefix all public APIs with <prefix>. For example, if <prefix> is
"prefix_", API changes like the following occur:
malloc() --> prefix_malloc()
malloc_conf --> prefix_malloc_conf
/etc/malloc.conf --> /etc/prefix_malloc.conf
MALLOC_CONF --> PREFIX_MALLOC_CONF
This makes it possible to use jemalloc at the same time as the system
allocator, or even to use multiple copies of jemalloc simultaneously.
By default, the prefix is "", except on OS X, where it is "je_". On OS X,
jemalloc overlays the default malloc zone, but makes no attempt to actually
replace the "malloc", "calloc", etc. symbols.
--without-export
Don't export public APIs. This can be useful when building jemalloc as a
static library, or to avoid exporting public APIs when using the zone
allocator on OSX.
--with-private-namespace=<prefix>
Prefix all library-private APIs with <prefix>je_. For shared libraries,
symbol visibility mechanisms prevent these symbols from being exported, but
for static libraries, naming collisions are a real possibility. By
default, <prefix> is empty, which results in a symbol prefix of je_ .
--with-install-suffix=<suffix>
Append <suffix> to the base name of all installed files, such that multiple
versions of jemalloc can coexist in the same installation directory. For
example, libjemalloc.so.0 becomes libjemalloc<suffix>.so.0.
--enable-cc-silence
Enable code that silences non-useful compiler warnings. This is helpful
when trying to tell serious warnings from those due to compiler
limitations, but it potentially incurs a performance penalty.
--enable-debug
Enable assertions and validation code. This incurs a substantial
performance hit, but is very useful during application development.
Implies --enable-ivsalloc.
--enable-code-coverage
Enable code coverage support, for use during jemalloc test development.
Additional testing targets are available if this option is enabled:
coverage
coverage_unit
coverage_integration
coverage_stress
These targets do not clear code coverage results from previous runs, and
there are interactions between the various coverage targets, so it is
usually advisable to run 'make clean' between repeated code coverage runs.
--enable-ivsalloc
Enable validation code, which verifies that pointers reside within
jemalloc-owned chunks before dereferencing them. This incurs a substantial
performance hit.
--disable-stats
Disable statistics gathering functionality. See the "opt.stats_print"
option documentation for usage details.
--enable-prof
Enable heap profiling and leak detection functionality. See the "opt.prof"
option documentation for usage details. When enabled, there are several
approaches to backtracing, and the configure script chooses the first one
in the following list that appears to function correctly:
+ libunwind (requires --enable-prof-libunwind)
+ libgcc (unless --disable-prof-libgcc)
+ gcc intrinsics (unless --disable-prof-gcc)
--enable-prof-libunwind
Use the libunwind library (http://www.nongnu.org/libunwind/) for stack
backtracing.
--disable-prof-libgcc
Disable the use of libgcc's backtracing functionality.
--disable-prof-gcc
Disable the use of gcc intrinsics for backtracing.
--with-static-libunwind=<libunwind.a>
Statically link against the specified libunwind.a rather than dynamically
linking with -lunwind.
--disable-tcache
Disable thread-specific caches for small objects. Objects are cached and
released in bulk, thus reducing the total number of mutex operations. See
the "opt.tcache" option for usage details.
--enable-mremap
Enable huge realloc() via mremap(2). mremap() is disabled by default
because the flavor used is specific to Linux, which has a quirk in its
virtual memory allocation algorithm that causes semi-permanent VM map holes
under normal jemalloc operation.
--disable-munmap
Disable virtual memory deallocation via munmap(2); instead keep track of
the virtual memory for later use. munmap() is disabled by default (i.e.
--disable-munmap is implied) on Linux, which has a quirk in its virtual
memory allocation algorithm that causes semi-permanent VM map holes under
normal jemalloc operation.
--enable-dss
Enable support for page allocation/deallocation via sbrk(2), in addition to
mmap(2).
--disable-fill
Disable support for junk/zero filling of memory, quarantine, and redzones.
See the "opt.junk", "opt.zero", "opt.quarantine", and "opt.redzone" option
documentation for usage details.
--disable-valgrind
Disable support for Valgrind.
--disable-experimental
Disable support for the experimental API (*allocm()).
--disable-zone-allocator
Disable zone allocator for Darwin. This means jemalloc won't be hooked as
the default allocator on OSX/iOS.
--enable-utrace
Enable utrace(2)-based allocation tracing. This feature is not broadly
portable (FreeBSD has it, but Linux and OS X do not).
--enable-xmalloc
Enable support for optional immediate termination due to out-of-memory
errors, as is commonly implemented by "xmalloc" wrapper function for malloc.
See the "opt.xmalloc" option documentation for usage details.
--enable-lazy-lock
Enable code that wraps pthread_create() to detect when an application
switches from single-threaded to multi-threaded mode, so that it can avoid
mutex locking/unlocking operations while in single-threaded mode. In
practice, this feature usually has little impact on performance unless
thread-specific caching is disabled.
--disable-tls
Disable thread-local storage (TLS), which allows for fast access to
thread-local variables via the __thread keyword. If TLS is available,
jemalloc uses it for several purposes.
--with-xslroot=<path>
Specify where to find DocBook XSL stylesheets when building the
documentation.
The following environment variables (not a definitive list) impact configure's
behavior:
CFLAGS="?"
Pass these flags to the compiler. You probably shouldn't define this unless
you know what you are doing. (Use EXTRA_CFLAGS instead.)
EXTRA_CFLAGS="?"
Append these flags to CFLAGS. This makes it possible to add flags such as
-Werror, while allowing the configure script to determine what other flags
are appropriate for the specified configuration.
The configure script specifically checks whether an optimization flag (-O*)
is specified in EXTRA_CFLAGS, and refrains from specifying an optimization
level if it finds that one has already been specified.
CPPFLAGS="?"
Pass these flags to the C preprocessor. Note that CFLAGS is not passed to
'cpp' when 'configure' is looking for include files, so you must use
CPPFLAGS instead if you need to help 'configure' find header files.
LD_LIBRARY_PATH="?"
'ld' uses this colon-separated list to find libraries.
LDFLAGS="?"
Pass these flags when linking.
PATH="?"
'configure' uses this to find programs.
=== Advanced compilation =======================================================
To build only parts of jemalloc, use the following targets:
build_lib_shared
build_lib_static
build_lib
build_doc_html
build_doc_man
build_doc
To install only parts of jemalloc, use the following targets:
install_bin
install_include
install_lib_shared
install_lib_static
install_lib
install_doc_html
install_doc_man
install_doc
To clean up build results to varying degrees, use the following make targets:
clean
distclean
relclean
=== Advanced installation ======================================================
Optionally, define make variables when invoking make, including (not
exclusively):
INCLUDEDIR="?"
Use this as the installation prefix for header files.
LIBDIR="?"
Use this as the installation prefix for libraries.
MANDIR="?"
Use this as the installation prefix for man pages.
DESTDIR="?"
Prepend DESTDIR to INCLUDEDIR, LIBDIR, DATADIR, and MANDIR. This is useful
when installing to a different path than was specified via --prefix.
CC="?"
Use this to invoke the C compiler.
CFLAGS="?"
Pass these flags to the compiler.
CPPFLAGS="?"
Pass these flags to the C preprocessor.
LDFLAGS="?"
Pass these flags when linking.
PATH="?"
Use this to search for programs used during configuration and building.
=== Development ================================================================
If you intend to make non-trivial changes to jemalloc, use the 'autogen.sh'
script rather than 'configure'. This re-generates 'configure', enables
configuration dependency rules, and enables re-generation of automatically
generated source files.
The build system supports using an object directory separate from the source
tree. For example, you can create an 'obj' directory, and from within that
directory, issue configuration and build commands:
autoconf
mkdir obj
cd obj
../configure --enable-autogen
make
=== Documentation ==============================================================
The manual page is generated in both html and roff formats. Any web browser
can be used to view the html manual. The roff manual page can be formatted
prior to installation via the following command:
nroff -man -t doc/jemalloc.3
deps/jemalloc/Makefile.in
Vendored
+438
View File
@@ -0,0 +1,438 @@
# Clear out all vpaths, then set just one (default vpath) for the main build
# directory.
vpath
vpath % .
# Clear the default suffixes, so that built-in rules are not used.
.SUFFIXES :
SHELL := /bin/sh
CC := @CC@
# Configuration parameters.
DESTDIR =
BINDIR := $(DESTDIR)@BINDIR@
INCLUDEDIR := $(DESTDIR)@INCLUDEDIR@
LIBDIR := $(DESTDIR)@LIBDIR@
DATADIR := $(DESTDIR)@DATADIR@
MANDIR := $(DESTDIR)@MANDIR@
srcroot := @srcroot@
objroot := @objroot@
abs_srcroot := @abs_srcroot@
abs_objroot := @abs_objroot@
# Build parameters.
CPPFLAGS := @CPPFLAGS@ -I$(srcroot)include -I$(objroot)include
CFLAGS := @CFLAGS@
LDFLAGS := @LDFLAGS@
EXTRA_LDFLAGS := @EXTRA_LDFLAGS@
LIBS := @LIBS@
RPATH_EXTRA := @RPATH_EXTRA@
SO := @so@
IMPORTLIB := @importlib@
O := @o@
A := @a@
EXE := @exe@
LIBPREFIX := @libprefix@
REV := @rev@
install_suffix := @install_suffix@
ABI := @abi@
XSLTPROC := @XSLTPROC@
AUTOCONF := @AUTOCONF@
_RPATH = @RPATH@
RPATH = $(if $(1),$(call _RPATH,$(1)))
cfghdrs_in := @cfghdrs_in@
cfghdrs_out := @cfghdrs_out@
cfgoutputs_in := @cfgoutputs_in@
cfgoutputs_out := @cfgoutputs_out@
enable_autogen := @enable_autogen@
enable_code_coverage := @enable_code_coverage@
enable_experimental := @enable_experimental@
enable_zone_allocator := @enable_zone_allocator@
DSO_LDFLAGS = @DSO_LDFLAGS@
SOREV = @SOREV@
PIC_CFLAGS = @PIC_CFLAGS@
CTARGET = @CTARGET@
LDTARGET = @LDTARGET@
MKLIB = @MKLIB@
AR = @AR@
ARFLAGS = @ARFLAGS@
CC_MM = @CC_MM@
ifeq (macho, $(ABI))
TEST_LIBRARY_PATH := DYLD_FALLBACK_LIBRARY_PATH="$(objroot)lib"
else
ifeq (pecoff, $(ABI))
TEST_LIBRARY_PATH := PATH="$(PATH):$(objroot)lib"
else
TEST_LIBRARY_PATH :=
endif
endif
LIBJEMALLOC := $(LIBPREFIX)jemalloc$(install_suffix)
# Lists of files.
BINS := $(srcroot)bin/pprof $(objroot)bin/jemalloc.sh
C_HDRS := $(objroot)include/jemalloc/jemalloc$(install_suffix).h
C_SRCS := $(srcroot)src/jemalloc.c $(srcroot)src/arena.c \
$(srcroot)src/atomic.c $(srcroot)src/base.c $(srcroot)src/bitmap.c \
$(srcroot)src/chunk.c $(srcroot)src/chunk_dss.c \
$(srcroot)src/chunk_mmap.c $(srcroot)src/ckh.c $(srcroot)src/ctl.c \
$(srcroot)src/extent.c $(srcroot)src/hash.c $(srcroot)src/huge.c \
$(srcroot)src/mb.c $(srcroot)src/mutex.c $(srcroot)src/prof.c \
$(srcroot)src/quarantine.c $(srcroot)src/rtree.c $(srcroot)src/stats.c \
$(srcroot)src/tcache.c $(srcroot)src/util.c $(srcroot)src/tsd.c
ifeq ($(enable_zone_allocator), 1)
C_SRCS += $(srcroot)src/zone.c
endif
ifeq ($(IMPORTLIB),$(SO))
STATIC_LIBS := $(objroot)lib/$(LIBJEMALLOC).$(A)
endif
ifdef PIC_CFLAGS
STATIC_LIBS += $(objroot)lib/$(LIBJEMALLOC)_pic.$(A)
else
STATIC_LIBS += $(objroot)lib/$(LIBJEMALLOC)_s.$(A)
endif
DSOS := $(objroot)lib/$(LIBJEMALLOC).$(SOREV)
ifneq ($(SOREV),$(SO))
DSOS += $(objroot)lib/$(LIBJEMALLOC).$(SO)
endif
MAN3 := $(objroot)doc/jemalloc$(install_suffix).3
DOCS_XML := $(objroot)doc/jemalloc$(install_suffix).xml
DOCS_HTML := $(DOCS_XML:$(objroot)%.xml=$(srcroot)%.html)
DOCS_MAN3 := $(DOCS_XML:$(objroot)%.xml=$(srcroot)%.3)
DOCS := $(DOCS_HTML) $(DOCS_MAN3)
C_TESTLIB_SRCS := $(srcroot)test/src/math.c $(srcroot)test/src/mtx.c \
$(srcroot)test/src/SFMT.c $(srcroot)test/src/test.c \
$(srcroot)test/src/thd.c
C_UTIL_INTEGRATION_SRCS := $(srcroot)src/util.c
TESTS_UNIT := $(srcroot)test/unit/bitmap.c \
$(srcroot)test/unit/ckh.c \
$(srcroot)test/unit/hash.c \
$(srcroot)test/unit/junk.c \
$(srcroot)test/unit/mallctl.c \
$(srcroot)test/unit/math.c \
$(srcroot)test/unit/mq.c \
$(srcroot)test/unit/mtx.c \
$(srcroot)test/unit/prof_accum.c \
$(srcroot)test/unit/prof_gdump.c \
$(srcroot)test/unit/prof_idump.c \
$(srcroot)test/unit/ql.c \
$(srcroot)test/unit/qr.c \
$(srcroot)test/unit/quarantine.c \
$(srcroot)test/unit/rb.c \
$(srcroot)test/unit/rtree.c \
$(srcroot)test/unit/SFMT.c \
$(srcroot)test/unit/stats.c \
$(srcroot)test/unit/tsd.c \
$(srcroot)test/unit/util.c \
$(srcroot)test/unit/zero.c
TESTS_UNIT_AUX := $(srcroot)test/unit/prof_accum_a.c \
$(srcroot)test/unit/prof_accum_b.c
TESTS_INTEGRATION := $(srcroot)test/integration/aligned_alloc.c \
$(srcroot)test/integration/allocated.c \
$(srcroot)test/integration/mallocx.c \
$(srcroot)test/integration/mremap.c \
$(srcroot)test/integration/posix_memalign.c \
$(srcroot)test/integration/rallocx.c \
$(srcroot)test/integration/thread_arena.c \
$(srcroot)test/integration/thread_tcache_enabled.c \
$(srcroot)test/integration/xallocx.c
ifeq ($(enable_experimental), 1)
TESTS_INTEGRATION += $(srcroot)test/integration/allocm.c \
$(srcroot)test/integration/MALLOCX_ARENA.c \
$(srcroot)test/integration/rallocm.c
endif
TESTS_STRESS :=
TESTS := $(TESTS_UNIT) $(TESTS_INTEGRATION) $(TESTS_STRESS)
C_OBJS := $(C_SRCS:$(srcroot)%.c=$(objroot)%.$(O))
C_PIC_OBJS := $(C_SRCS:$(srcroot)%.c=$(objroot)%.pic.$(O))
C_JET_OBJS := $(C_SRCS:$(srcroot)%.c=$(objroot)%.jet.$(O))
C_TESTLIB_UNIT_OBJS := $(C_TESTLIB_SRCS:$(srcroot)%.c=$(objroot)%.unit.$(O))
C_TESTLIB_INTEGRATION_OBJS := $(C_TESTLIB_SRCS:$(srcroot)%.c=$(objroot)%.integration.$(O))
C_UTIL_INTEGRATION_OBJS := $(C_UTIL_INTEGRATION_SRCS:$(srcroot)%.c=$(objroot)%.integration.$(O))
C_TESTLIB_STRESS_OBJS := $(C_TESTLIB_SRCS:$(srcroot)%.c=$(objroot)%.stress.$(O))
C_TESTLIB_OBJS := $(C_TESTLIB_UNIT_OBJS) $(C_TESTLIB_INTEGRATION_OBJS) $(C_UTIL_INTEGRATION_OBJS) $(C_TESTLIB_STRESS_OBJS)
TESTS_UNIT_OBJS := $(TESTS_UNIT:$(srcroot)%.c=$(objroot)%.$(O))
TESTS_UNIT_AUX_OBJS := $(TESTS_UNIT_AUX:$(srcroot)%.c=$(objroot)%.$(O))
TESTS_INTEGRATION_OBJS := $(TESTS_INTEGRATION:$(srcroot)%.c=$(objroot)%.$(O))
TESTS_STRESS_OBJS := $(TESTS_STRESS:$(srcroot)%.c=$(objroot)%.$(O))
TESTS_OBJS := $(TESTS_UNIT_OBJS) $(TESTS_UNIT_AUX_OBJS) $(TESTS_INTEGRATION_OBJS) $(TESTS_STRESS_OBJS)
.PHONY: all dist build_doc_html build_doc_man build_doc
.PHONY: install_bin install_include install_lib
.PHONY: install_doc_html install_doc_man install_doc install
.PHONY: tests check clean distclean relclean
.SECONDARY : $(TESTS_OBJS)
# Default target.
all: build_lib
dist: build_doc
$(srcroot)doc/%.html : $(objroot)doc/%.xml $(srcroot)doc/stylesheet.xsl $(objroot)doc/html.xsl
$(XSLTPROC) -o $@ $(objroot)doc/html.xsl $<
$(srcroot)doc/%.3 : $(objroot)doc/%.xml $(srcroot)doc/stylesheet.xsl $(objroot)doc/manpages.xsl
$(XSLTPROC) -o $@ $(objroot)doc/manpages.xsl $<
build_doc_html: $(DOCS_HTML)
build_doc_man: $(DOCS_MAN3)
build_doc: $(DOCS)
#
# Include generated dependency files.
#
ifdef CC_MM
-include $(C_OBJS:%.$(O)=%.d)
-include $(C_PIC_OBJS:%.$(O)=%.d)
-include $(C_JET_OBJS:%.$(O)=%.d)
-include $(C_TESTLIB_OBJS:%.$(O)=%.d)
-include $(TESTS_OBJS:%.$(O)=%.d)
endif
$(C_OBJS): $(objroot)src/%.$(O): $(srcroot)src/%.c
$(C_PIC_OBJS): $(objroot)src/%.pic.$(O): $(srcroot)src/%.c
$(C_PIC_OBJS): CFLAGS += $(PIC_CFLAGS)
$(C_JET_OBJS): $(objroot)src/%.jet.$(O): $(srcroot)src/%.c
$(C_JET_OBJS): CFLAGS += -DJEMALLOC_JET
$(C_TESTLIB_UNIT_OBJS): $(objroot)test/src/%.unit.$(O): $(srcroot)test/src/%.c
$(C_TESTLIB_UNIT_OBJS): CPPFLAGS += -DJEMALLOC_UNIT_TEST
$(C_TESTLIB_INTEGRATION_OBJS): $(objroot)test/src/%.integration.$(O): $(srcroot)test/src/%.c
$(C_TESTLIB_INTEGRATION_OBJS): CPPFLAGS += -DJEMALLOC_INTEGRATION_TEST
$(C_UTIL_INTEGRATION_OBJS): $(objroot)src/%.integration.$(O): $(srcroot)src/%.c
$(C_TESTLIB_STRESS_OBJS): $(objroot)test/src/%.stress.$(O): $(srcroot)test/src/%.c
$(C_TESTLIB_STRESS_OBJS): CPPFLAGS += -DJEMALLOC_STRESS_TEST -DJEMALLOC_STRESS_TESTLIB
$(C_TESTLIB_OBJS): CPPFLAGS += -I$(srcroot)test/include -I$(objroot)test/include
$(TESTS_UNIT_OBJS): CPPFLAGS += -DJEMALLOC_UNIT_TEST
$(TESTS_UNIT_AUX_OBJS): CPPFLAGS += -DJEMALLOC_UNIT_TEST
define make-unit-link-dep
$(1): TESTS_UNIT_LINK_OBJS += $(2)
$(1): $(2)
endef
$(foreach test, $(TESTS_UNIT:$(srcroot)test/unit/%.c=$(objroot)test/unit/%$(EXE)), $(eval $(call make-unit-link-dep,$(test),$(filter $(test:%=%_a.$(O)) $(test:%=%_b.$(O)),$(TESTS_UNIT_AUX_OBJS)))))
$(TESTS_INTEGRATION_OBJS): CPPFLAGS += -DJEMALLOC_INTEGRATION_TEST
$(TESTS_STRESS_OBJS): CPPFLAGS += -DJEMALLOC_STRESS_TEST
$(TESTS_OBJS): $(objroot)test/%.$(O): $(srcroot)test/%.c
$(TESTS_OBJS): CPPFLAGS += -I$(srcroot)test/include -I$(objroot)test/include
ifneq ($(IMPORTLIB),$(SO))
$(C_OBJS): CPPFLAGS += -DDLLEXPORT
endif
ifndef CC_MM
# Dependencies.
HEADER_DIRS = $(srcroot)include/jemalloc/internal \
$(objroot)include/jemalloc $(objroot)include/jemalloc/internal
HEADERS = $(wildcard $(foreach dir,$(HEADER_DIRS),$(dir)/*.h))
$(C_OBJS) $(C_PIC_OBJS) $(C_JET_OBJS) $(C_TESTLIB_OBJS) $(TESTS_OBJS): $(HEADERS)
$(TESTS_OBJS): $(objroot)test/unit/jemalloc_test.h
endif
$(C_OBJS) $(C_PIC_OBJS) $(C_JET_OBJS) $(C_TESTLIB_OBJS) $(TESTS_OBJS): %.$(O):
@mkdir -p $(@D)
$(CC) $(CFLAGS) -c $(CPPFLAGS) $(CTARGET) $<
ifdef CC_MM
@$(CC) -MM $(CPPFLAGS) -MT $@ -o $(@:%.$(O)=%.d) $<
endif
ifneq ($(SOREV),$(SO))
%.$(SO) : %.$(SOREV)
@mkdir -p $(@D)
ln -sf $(<F) $@
endif
$(objroot)lib/$(LIBJEMALLOC).$(SOREV) : $(if $(PIC_CFLAGS),$(C_PIC_OBJS),$(C_OBJS))
@mkdir -p $(@D)
$(CC) $(DSO_LDFLAGS) $(call RPATH,$(RPATH_EXTRA)) $(LDTARGET) $+ $(LDFLAGS) $(LIBS) $(EXTRA_LDFLAGS)
$(objroot)lib/$(LIBJEMALLOC)_pic.$(A) : $(C_PIC_OBJS)
$(objroot)lib/$(LIBJEMALLOC).$(A) : $(C_OBJS)
$(objroot)lib/$(LIBJEMALLOC)_s.$(A) : $(C_OBJS)
$(STATIC_LIBS):
@mkdir -p $(@D)
$(AR) $(ARFLAGS)@AROUT@ $+
$(objroot)test/unit/%$(EXE): $(objroot)test/unit/%.$(O) $(TESTS_UNIT_LINK_OBJS) $(C_JET_OBJS) $(C_TESTLIB_UNIT_OBJS)
@mkdir -p $(@D)
$(CC) $(LDTARGET) $(filter %.$(O),$^) $(call RPATH,$(objroot)lib) $(LDFLAGS) $(filter-out -lm,$(LIBS)) -lm $(EXTRA_LDFLAGS)
$(objroot)test/integration/%$(EXE): $(objroot)test/integration/%.$(O) $(C_TESTLIB_INTEGRATION_OBJS) $(C_UTIL_INTEGRATION_OBJS) $(objroot)lib/$(LIBJEMALLOC).$(IMPORTLIB)
@mkdir -p $(@D)
$(CC) $(LDTARGET) $(filter %.$(O),$^) $(call RPATH,$(objroot)lib) $(objroot)lib/$(LIBJEMALLOC).$(IMPORTLIB) $(LDFLAGS) $(filter-out -lm,$(filter -lpthread,$(LIBS))) -lm $(EXTRA_LDFLAGS)
$(objroot)test/stress/%$(EXE): $(objroot)test/stress/%.$(O) $(C_JET_OBJS) $(C_TESTLIB_STRESS_OBJS) $(objroot)lib/$(LIBJEMALLOC).$(IMPORTLIB)
@mkdir -p $(@D)
$(CC) $(LDTARGET) $(filter %.$(O),$^) $(call RPATH,$(objroot)lib) $(objroot)lib/$(LIBJEMALLOC).$(IMPORTLIB) $(LDFLAGS) $(filter-out -lm,$(LIBS)) -lm $(EXTRA_LDFLAGS)
build_lib_shared: $(DSOS)
build_lib_static: $(STATIC_LIBS)
build_lib: build_lib_shared build_lib_static
install_bin:
install -d $(BINDIR)
@for b in $(BINS); do \
echo "install -m 755 $$b $(BINDIR)"; \
install -m 755 $$b $(BINDIR); \
done
install_include:
install -d $(INCLUDEDIR)/jemalloc
@for h in $(C_HDRS); do \
echo "install -m 644 $$h $(INCLUDEDIR)/jemalloc"; \
install -m 644 $$h $(INCLUDEDIR)/jemalloc; \
done
install_lib_shared: $(DSOS)
install -d $(LIBDIR)
install -m 755 $(objroot)lib/$(LIBJEMALLOC).$(SOREV) $(LIBDIR)
ifneq ($(SOREV),$(SO))
ln -sf $(LIBJEMALLOC).$(SOREV) $(LIBDIR)/$(LIBJEMALLOC).$(SO)
endif
install_lib_static: $(STATIC_LIBS)
install -d $(LIBDIR)
@for l in $(STATIC_LIBS); do \
echo "install -m 755 $$l $(LIBDIR)"; \
install -m 755 $$l $(LIBDIR); \
done
install_lib: install_lib_shared install_lib_static
install_doc_html:
install -d $(DATADIR)/doc/jemalloc$(install_suffix)
@for d in $(DOCS_HTML); do \
echo "install -m 644 $$d $(DATADIR)/doc/jemalloc$(install_suffix)"; \
install -m 644 $$d $(DATADIR)/doc/jemalloc$(install_suffix); \
done
install_doc_man:
install -d $(MANDIR)/man3
@for d in $(DOCS_MAN3); do \
echo "install -m 644 $$d $(MANDIR)/man3"; \
install -m 644 $$d $(MANDIR)/man3; \
done
install_doc: install_doc_html install_doc_man
install: install_bin install_include install_lib install_doc
tests_unit: $(TESTS_UNIT:$(srcroot)%.c=$(objroot)%$(EXE))
tests_integration: $(TESTS_INTEGRATION:$(srcroot)%.c=$(objroot)%$(EXE))
tests_stress: $(TESTS_STRESS:$(srcroot)%.c=$(objroot)%$(EXE))
tests: tests_unit tests_integration tests_stress
check_unit_dir:
@mkdir -p $(objroot)test/unit
check_integration_dir:
@mkdir -p $(objroot)test/integration
check_stress_dir:
@mkdir -p $(objroot)test/stress
check_dir: check_unit_dir check_integration_dir check_stress_dir
check_unit: tests_unit check_unit_dir
$(SHELL) $(objroot)test/test.sh $(TESTS_UNIT:$(srcroot)%.c=$(objroot)%)
check_integration: tests_integration check_integration_dir
$(SHELL) $(objroot)test/test.sh $(TESTS_INTEGRATION:$(srcroot)%.c=$(objroot)%)
check_stress: tests_stress check_stress_dir
$(SHELL) $(objroot)test/test.sh $(TESTS_STRESS:$(srcroot)%.c=$(objroot)%)
check: tests check_dir
$(SHELL) $(objroot)test/test.sh $(TESTS:$(srcroot)%.c=$(objroot)%)
ifeq ($(enable_code_coverage), 1)
coverage_unit: check_unit
$(SHELL) $(srcroot)coverage.sh $(srcroot)src jet $(C_JET_OBJS)
$(SHELL) $(srcroot)coverage.sh $(srcroot)test/src unit $(C_TESTLIB_UNIT_OBJS)
$(SHELL) $(srcroot)coverage.sh $(srcroot)test/unit unit $(TESTS_UNIT_OBJS)
coverage_integration: check_integration
$(SHELL) $(srcroot)coverage.sh $(srcroot)src pic $(C_PIC_OBJS)
$(SHELL) $(srcroot)coverage.sh $(srcroot)src integration $(C_UTIL_INTEGRATION_OBJS)
$(SHELL) $(srcroot)coverage.sh $(srcroot)test/src integration $(C_TESTLIB_INTEGRATION_OBJS)
$(SHELL) $(srcroot)coverage.sh $(srcroot)test/integration integration $(TESTS_INTEGRATION_OBJS)
coverage_stress: check_stress
$(SHELL) $(srcroot)coverage.sh $(srcroot)src pic $(C_PIC_OBJS)
$(SHELL) $(srcroot)coverage.sh $(srcroot)src jet $(C_JET_OBJS)
$(SHELL) $(srcroot)coverage.sh $(srcroot)test/src stress $(C_TESTLIB_STRESS_OBJS)
$(SHELL) $(srcroot)coverage.sh $(srcroot)test/stress stress $(TESTS_STRESS_OBJS)
coverage: check
$(SHELL) $(srcroot)coverage.sh $(srcroot)src pic $(C_PIC_OBJS)
$(SHELL) $(srcroot)coverage.sh $(srcroot)src jet $(C_JET_OBJS)
$(SHELL) $(srcroot)coverage.sh $(srcroot)src integration $(C_UTIL_INTEGRATION_OBJS)
$(SHELL) $(srcroot)coverage.sh $(srcroot)test/src unit $(C_TESTLIB_UNIT_OBJS)
$(SHELL) $(srcroot)coverage.sh $(srcroot)test/src integration $(C_TESTLIB_INTEGRATION_OBJS)
$(SHELL) $(srcroot)coverage.sh $(srcroot)test/src stress $(C_TESTLIB_STRESS_OBJS)
$(SHELL) $(srcroot)coverage.sh $(srcroot)test/unit unit $(TESTS_UNIT_OBJS) $(TESTS_UNIT_AUX_OBJS)
$(SHELL) $(srcroot)coverage.sh $(srcroot)test/integration integration $(TESTS_INTEGRATION_OBJS)
$(SHELL) $(srcroot)coverage.sh $(srcroot)test/stress integration $(TESTS_STRESS_OBJS)
endif
clean:
rm -f $(C_OBJS)
rm -f $(C_PIC_OBJS)
rm -f $(C_JET_OBJS)
rm -f $(C_TESTLIB_OBJS)
rm -f $(C_OBJS:%.$(O)=%.d)
rm -f $(C_OBJS:%.$(O)=%.gcda)
rm -f $(C_OBJS:%.$(O)=%.gcno)
rm -f $(C_PIC_OBJS:%.$(O)=%.d)
rm -f $(C_PIC_OBJS:%.$(O)=%.gcda)
rm -f $(C_PIC_OBJS:%.$(O)=%.gcno)
rm -f $(C_JET_OBJS:%.$(O)=%.d)
rm -f $(C_JET_OBJS:%.$(O)=%.gcda)
rm -f $(C_JET_OBJS:%.$(O)=%.gcno)
rm -f $(C_TESTLIB_OBJS:%.$(O)=%.d)
rm -f $(C_TESTLIB_OBJS:%.$(O)=%.gcda)
rm -f $(C_TESTLIB_OBJS:%.$(O)=%.gcno)
rm -f $(TESTS_OBJS:%.$(O)=%$(EXE))
rm -f $(TESTS_OBJS)
rm -f $(TESTS_OBJS:%.$(O)=%.d)
rm -f $(TESTS_OBJS:%.$(O)=%.gcda)
rm -f $(TESTS_OBJS:%.$(O)=%.gcno)
rm -f $(TESTS_OBJS:%.$(O)=%.out)
rm -f $(DSOS) $(STATIC_LIBS)
rm -f $(objroot)*.gcov.*
distclean: clean
rm -rf $(objroot)autom4te.cache
rm -f $(objroot)bin/jemalloc.sh
rm -f $(objroot)config.log
rm -f $(objroot)config.status
rm -f $(objroot)config.stamp
rm -f $(cfghdrs_out)
rm -f $(cfgoutputs_out)
relclean: distclean
rm -f $(objroot)configure
rm -f $(srcroot)VERSION
rm -f $(DOCS_HTML)
rm -f $(DOCS_MAN3)
#===============================================================================
# Re-configuration rules.
ifeq ($(enable_autogen), 1)
$(srcroot)configure : $(srcroot)configure.ac
cd ./$(srcroot) && $(AUTOCONF)
$(objroot)config.status : $(srcroot)configure
./$(objroot)config.status --recheck
$(srcroot)config.stamp.in : $(srcroot)configure.ac
echo stamp > $(srcroot)config.stamp.in
$(objroot)config.stamp : $(cfgoutputs_in) $(cfghdrs_in) $(srcroot)configure
./$(objroot)config.status
@touch $@
# There must be some action in order for make to re-read Makefile when it is
# out of date.
$(cfgoutputs_out) $(cfghdrs_out) : $(objroot)config.stamp
@true
endif
deps/jemalloc/README
Vendored
+20
View File
@@ -0,0 +1,20 @@
jemalloc is a general purpose malloc(3) implementation that emphasizes
fragmentation avoidance and scalable concurrency support. jemalloc first came
into use as the FreeBSD libc allocator in 2005, and since then it has found its
way into numerous applications that rely on its predictable behavior. In 2010
jemalloc development efforts broadened to include developer support features
such as heap profiling, Valgrind integration, and extensive monitoring/tuning
hooks. Modern jemalloc releases continue to be integrated back into FreeBSD,
and therefore versatility remains critical. Ongoing development efforts trend
toward making jemalloc among the best allocators for a broad range of demanding
applications, and eliminating/mitigating weaknesses that have practical
repercussions for real world applications.
The COPYING file contains copyright and licensing information.
The INSTALL file contains information on how to configure, build, and install
jemalloc.
The ChangeLog file contains a brief summary of changes for each release.
URL: http://www.canonware.com/jemalloc/
deps/jemalloc/VERSION
Vendored
+1
View File
@@ -0,0 +1 @@
3.6.0-0-g46c0af68bd248b04df75e4f92d5fb804c3d75340
deps/jemalloc/autogen.sh
Vendored Executable
+17
View File
@@ -0,0 +1,17 @@
#!/bin/sh
for i in autoconf; do
echo "$i"
$i
if [ $? -ne 0 ]; then
echo "Error $? in $i"
exit 1
fi
done
echo "./configure --enable-autogen $@"
./configure --enable-autogen $@
if [ $? -ne 0 ]; then
echo "Error $? in ./configure"
exit 1
fi
deps/jemalloc/bin/jemalloc.sh.in
Vendored
+9
View File
@@ -0,0 +1,9 @@
#!/bin/sh
prefix=@prefix@
exec_prefix=@exec_prefix@
libdir=@libdir@
@LD_PRELOAD_VAR@=${libdir}/libjemalloc.@SOREV@
export @LD_PRELOAD_VAR@
exec "$@"
deps/jemalloc/bin/pprof
Vendored Executable
+5372
View File
File diff suppressed because it is too large Load Diff
deps/jemalloc/config.guess
Vendored Executable
+1558
View File
File diff suppressed because it is too large Load Diff
deps/jemalloc/config.stamp.in
Vendored
View File
deps/jemalloc/config.sub
Vendored Executable
+1793
View File
File diff suppressed because it is too large Load Diff
deps/jemalloc/configure
Vendored Executable
+8962
View File
File diff suppressed because it is too large Load Diff
deps/jemalloc/configure.ac
Vendored
+1487
View File
File diff suppressed because it is too large Load Diff
deps/jemalloc/coverage.sh
Vendored Executable
+16
View File
@@ -0,0 +1,16 @@
#!/bin/sh
set -e
objdir=$1
suffix=$2
shift 2
objs=$@
gcov -b -p -f -o "${objdir}" ${objs}
# Move gcov outputs so that subsequent gcov invocations won't clobber results
# for the same sources with different compilation flags.
for f in `find . -maxdepth 1 -type f -name '*.gcov'` ; do
mv "${f}" "${f}.${suffix}"
done
deps/jemalloc/doc/html.xsl.in
Vendored
+4
View File
@@ -0,0 +1,4 @@
<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="1.0">
<xsl:import href="@XSLROOT@/html/docbook.xsl"/>
<xsl:import href="@abs_srcroot@doc/stylesheet.xsl"/>
</xsl:stylesheet>
deps/jemalloc/doc/jemalloc.3
Vendored
+1630
View File
File diff suppressed because it is too large Load Diff
deps/jemalloc/doc/jemalloc.html
Vendored
+1508
View File
File diff suppressed because one or more lines are too long
deps/jemalloc/doc/jemalloc.xml.in
Vendored
+2345
View File
File diff suppressed because it is too large Load Diff
deps/jemalloc/doc/manpages.xsl.in
Vendored
+4
View File
@@ -0,0 +1,4 @@
<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="1.0">
<xsl:import href="@XSLROOT@/manpages/docbook.xsl"/>
<xsl:import href="@abs_srcroot@doc/stylesheet.xsl"/>
</xsl:stylesheet>
deps/jemalloc/doc/stylesheet.xsl
Vendored
+7
View File
@@ -0,0 +1,7 @@
<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="1.0">
<xsl:param name="funcsynopsis.style">ansi</xsl:param>
<xsl:param name="function.parens" select="1"/>
<xsl:template match="mallctl">
"<xsl:call-template name="inline.monoseq"/>"
</xsl:template>
</xsl:stylesheet>
deps/jemalloc/include/jemalloc/internal/arena.h
Vendored
+1063
View File
File diff suppressed because it is too large Load Diff
deps/jemalloc/include/jemalloc/internal/atomic.h
Vendored
+304
View File
@@ -0,0 +1,304 @@
/******************************************************************************/
#ifdef JEMALLOC_H_TYPES
#endif /* JEMALLOC_H_TYPES */
/******************************************************************************/
#ifdef JEMALLOC_H_STRUCTS
#endif /* JEMALLOC_H_STRUCTS */
/******************************************************************************/
#ifdef JEMALLOC_H_EXTERNS
#define atomic_read_uint64(p) atomic_add_uint64(p, 0)
#define atomic_read_uint32(p) atomic_add_uint32(p, 0)
#define atomic_read_z(p) atomic_add_z(p, 0)
#define atomic_read_u(p) atomic_add_u(p, 0)
#endif /* JEMALLOC_H_EXTERNS */
/******************************************************************************/
#ifdef JEMALLOC_H_INLINES
#ifndef JEMALLOC_ENABLE_INLINE
uint64_t atomic_add_uint64(uint64_t *p, uint64_t x);
uint64_t atomic_sub_uint64(uint64_t *p, uint64_t x);
uint32_t atomic_add_uint32(uint32_t *p, uint32_t x);
uint32_t atomic_sub_uint32(uint32_t *p, uint32_t x);
size_t atomic_add_z(size_t *p, size_t x);
size_t atomic_sub_z(size_t *p, size_t x);
unsigned atomic_add_u(unsigned *p, unsigned x);
unsigned atomic_sub_u(unsigned *p, unsigned x);
#endif
#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_ATOMIC_C_))
/******************************************************************************/
/* 64-bit operations. */
#if (LG_SIZEOF_PTR == 3 || LG_SIZEOF_INT == 3)
# ifdef __GCC_HAVE_SYNC_COMPARE_AND_SWAP_8
JEMALLOC_INLINE uint64_t
atomic_add_uint64(uint64_t *p, uint64_t x)
{
return (__sync_add_and_fetch(p, x));
}
JEMALLOC_INLINE uint64_t
atomic_sub_uint64(uint64_t *p, uint64_t x)
{
return (__sync_sub_and_fetch(p, x));
}
#elif (defined(_MSC_VER))
JEMALLOC_INLINE uint64_t
atomic_add_uint64(uint64_t *p, uint64_t x)
{
return (InterlockedExchangeAdd64(p, x));
}
JEMALLOC_INLINE uint64_t
atomic_sub_uint64(uint64_t *p, uint64_t x)
{
return (InterlockedExchangeAdd64(p, -((int64_t)x)));
}
#elif (defined(JEMALLOC_OSATOMIC))
JEMALLOC_INLINE uint64_t
atomic_add_uint64(uint64_t *p, uint64_t x)
{
return (OSAtomicAdd64((int64_t)x, (int64_t *)p));
}
JEMALLOC_INLINE uint64_t
atomic_sub_uint64(uint64_t *p, uint64_t x)
{
return (OSAtomicAdd64(-((int64_t)x), (int64_t *)p));
}
# elif (defined(__amd64__) || defined(__x86_64__))
JEMALLOC_INLINE uint64_t
atomic_add_uint64(uint64_t *p, uint64_t x)
{
asm volatile (
"lock; xaddq %0, %1;"
: "+r" (x), "=m" (*p) /* Outputs. */
: "m" (*p) /* Inputs. */
);
return (x);
}
JEMALLOC_INLINE uint64_t
atomic_sub_uint64(uint64_t *p, uint64_t x)
{
x = (uint64_t)(-(int64_t)x);
asm volatile (
"lock; xaddq %0, %1;"
: "+r" (x), "=m" (*p) /* Outputs. */
: "m" (*p) /* Inputs. */
);
return (x);
}
# elif (defined(JEMALLOC_ATOMIC9))
JEMALLOC_INLINE uint64_t
atomic_add_uint64(uint64_t *p, uint64_t x)
{
/*
* atomic_fetchadd_64() doesn't exist, but we only ever use this
* function on LP64 systems, so atomic_fetchadd_long() will do.
*/
assert(sizeof(uint64_t) == sizeof(unsigned long));
return (atomic_fetchadd_long(p, (unsigned long)x) + x);
}
JEMALLOC_INLINE uint64_t
atomic_sub_uint64(uint64_t *p, uint64_t x)
{
assert(sizeof(uint64_t) == sizeof(unsigned long));
return (atomic_fetchadd_long(p, (unsigned long)(-(long)x)) - x);
}
# elif (defined(JE_FORCE_SYNC_COMPARE_AND_SWAP_8))
JEMALLOC_INLINE uint64_t
atomic_add_uint64(uint64_t *p, uint64_t x)
{
return (__sync_add_and_fetch(p, x));
}
JEMALLOC_INLINE uint64_t
atomic_sub_uint64(uint64_t *p, uint64_t x)
{
return (__sync_sub_and_fetch(p, x));
}
# else
# error "Missing implementation for 64-bit atomic operations"
# endif
#endif
/******************************************************************************/
/* 32-bit operations. */
#ifdef __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4
JEMALLOC_INLINE uint32_t
atomic_add_uint32(uint32_t *p, uint32_t x)
{
return (__sync_add_and_fetch(p, x));
}
JEMALLOC_INLINE uint32_t
atomic_sub_uint32(uint32_t *p, uint32_t x)
{
return (__sync_sub_and_fetch(p, x));
}
#elif (defined(_MSC_VER))
JEMALLOC_INLINE uint32_t
atomic_add_uint32(uint32_t *p, uint32_t x)
{
return (InterlockedExchangeAdd(p, x));
}
JEMALLOC_INLINE uint32_t
atomic_sub_uint32(uint32_t *p, uint32_t x)
{
return (InterlockedExchangeAdd(p, -((int32_t)x)));
}
#elif (defined(JEMALLOC_OSATOMIC))
JEMALLOC_INLINE uint32_t
atomic_add_uint32(uint32_t *p, uint32_t x)
{
return (OSAtomicAdd32((int32_t)x, (int32_t *)p));
}
JEMALLOC_INLINE uint32_t
atomic_sub_uint32(uint32_t *p, uint32_t x)
{
return (OSAtomicAdd32(-((int32_t)x), (int32_t *)p));
}
#elif (defined(__i386__) || defined(__amd64__) || defined(__x86_64__))
JEMALLOC_INLINE uint32_t
atomic_add_uint32(uint32_t *p, uint32_t x)
{
asm volatile (
"lock; xaddl %0, %1;"
: "+r" (x), "=m" (*p) /* Outputs. */
: "m" (*p) /* Inputs. */
);
return (x);
}
JEMALLOC_INLINE uint32_t
atomic_sub_uint32(uint32_t *p, uint32_t x)
{
x = (uint32_t)(-(int32_t)x);
asm volatile (
"lock; xaddl %0, %1;"
: "+r" (x), "=m" (*p) /* Outputs. */
: "m" (*p) /* Inputs. */
);
return (x);
}
#elif (defined(JEMALLOC_ATOMIC9))
JEMALLOC_INLINE uint32_t
atomic_add_uint32(uint32_t *p, uint32_t x)
{
return (atomic_fetchadd_32(p, x) + x);
}
JEMALLOC_INLINE uint32_t
atomic_sub_uint32(uint32_t *p, uint32_t x)
{
return (atomic_fetchadd_32(p, (uint32_t)(-(int32_t)x)) - x);
}
#elif (defined(JE_FORCE_SYNC_COMPARE_AND_SWAP_4))
JEMALLOC_INLINE uint32_t
atomic_add_uint32(uint32_t *p, uint32_t x)
{
return (__sync_add_and_fetch(p, x));
}
JEMALLOC_INLINE uint32_t
atomic_sub_uint32(uint32_t *p, uint32_t x)
{
return (__sync_sub_and_fetch(p, x));
}
#else
# error "Missing implementation for 32-bit atomic operations"
#endif
/******************************************************************************/
/* size_t operations. */
JEMALLOC_INLINE size_t
atomic_add_z(size_t *p, size_t x)
{
#if (LG_SIZEOF_PTR == 3)
return ((size_t)atomic_add_uint64((uint64_t *)p, (uint64_t)x));
#elif (LG_SIZEOF_PTR == 2)
return ((size_t)atomic_add_uint32((uint32_t *)p, (uint32_t)x));
#endif
}
JEMALLOC_INLINE size_t
atomic_sub_z(size_t *p, size_t x)
{
#if (LG_SIZEOF_PTR == 3)
return ((size_t)atomic_add_uint64((uint64_t *)p,
(uint64_t)-((int64_t)x)));
#elif (LG_SIZEOF_PTR == 2)
return ((size_t)atomic_add_uint32((uint32_t *)p,
(uint32_t)-((int32_t)x)));
#endif
}
/******************************************************************************/
/* unsigned operations. */
JEMALLOC_INLINE unsigned
atomic_add_u(unsigned *p, unsigned x)
{
#if (LG_SIZEOF_INT == 3)
return ((unsigned)atomic_add_uint64((uint64_t *)p, (uint64_t)x));
#elif (LG_SIZEOF_INT == 2)
return ((unsigned)atomic_add_uint32((uint32_t *)p, (uint32_t)x));
#endif
}
JEMALLOC_INLINE unsigned
atomic_sub_u(unsigned *p, unsigned x)
{
#if (LG_SIZEOF_INT == 3)
return ((unsigned)atomic_add_uint64((uint64_t *)p,
(uint64_t)-((int64_t)x)));
#elif (LG_SIZEOF_INT == 2)
return ((unsigned)atomic_add_uint32((uint32_t *)p,
(uint32_t)-((int32_t)x)));
#endif
}
/******************************************************************************/
#endif
#endif /* JEMALLOC_H_INLINES */
/******************************************************************************/
deps/jemalloc/include/jemalloc/internal/base.h
Vendored
+26
View File
@@ -0,0 +1,26 @@
/******************************************************************************/
#ifdef JEMALLOC_H_TYPES
#endif /* JEMALLOC_H_TYPES */
/******************************************************************************/
#ifdef JEMALLOC_H_STRUCTS
#endif /* JEMALLOC_H_STRUCTS */
/******************************************************************************/
#ifdef JEMALLOC_H_EXTERNS
void *base_alloc(size_t size);
void *base_calloc(size_t number, size_t size);
extent_node_t *base_node_alloc(void);
void base_node_dealloc(extent_node_t *node);
bool base_boot(void);
void base_prefork(void);
void base_postfork_parent(void);
void base_postfork_child(void);
#endif /* JEMALLOC_H_EXTERNS */
/******************************************************************************/
#ifdef JEMALLOC_H_INLINES
#endif /* JEMALLOC_H_INLINES */
/******************************************************************************/
deps/jemalloc/include/jemalloc/internal/bitmap.h
Vendored
+184
View File
@@ -0,0 +1,184 @@
/******************************************************************************/
#ifdef JEMALLOC_H_TYPES
/* Maximum bitmap bit count is 2^LG_BITMAP_MAXBITS. */
#define LG_BITMAP_MAXBITS LG_RUN_MAXREGS
typedef struct bitmap_level_s bitmap_level_t;
typedef struct bitmap_info_s bitmap_info_t;
typedef unsigned long bitmap_t;
#define LG_SIZEOF_BITMAP LG_SIZEOF_LONG
/* Number of bits per group. */
#define LG_BITMAP_GROUP_NBITS (LG_SIZEOF_BITMAP + 3)
#define BITMAP_GROUP_NBITS (ZU(1) << LG_BITMAP_GROUP_NBITS)
#define BITMAP_GROUP_NBITS_MASK (BITMAP_GROUP_NBITS-1)
/* Maximum number of levels possible. */
#define BITMAP_MAX_LEVELS \
(LG_BITMAP_MAXBITS / LG_SIZEOF_BITMAP) \
+ !!(LG_BITMAP_MAXBITS % LG_SIZEOF_BITMAP)
#endif /* JEMALLOC_H_TYPES */
/******************************************************************************/
#ifdef JEMALLOC_H_STRUCTS
struct bitmap_level_s {
/* Offset of this level's groups within the array of groups. */
size_t group_offset;
};
struct bitmap_info_s {
/* Logical number of bits in bitmap (stored at bottom level). */
size_t nbits;
/* Number of levels necessary for nbits. */
unsigned nlevels;
/*
* Only the first (nlevels+1) elements are used, and levels are ordered
* bottom to top (e.g. the bottom level is stored in levels[0]).
*/
bitmap_level_t levels[BITMAP_MAX_LEVELS+1];
};
#endif /* JEMALLOC_H_STRUCTS */
/******************************************************************************/
#ifdef JEMALLOC_H_EXTERNS
void bitmap_info_init(bitmap_info_t *binfo, size_t nbits);
size_t bitmap_info_ngroups(const bitmap_info_t *binfo);
size_t bitmap_size(size_t nbits);
void bitmap_init(bitmap_t *bitmap, const bitmap_info_t *binfo);
#endif /* JEMALLOC_H_EXTERNS */
/******************************************************************************/
#ifdef JEMALLOC_H_INLINES
#ifndef JEMALLOC_ENABLE_INLINE
bool bitmap_full(bitmap_t *bitmap, const bitmap_info_t *binfo);
bool bitmap_get(bitmap_t *bitmap, const bitmap_info_t *binfo, size_t bit);
void bitmap_set(bitmap_t *bitmap, const bitmap_info_t *binfo, size_t bit);
size_t bitmap_sfu(bitmap_t *bitmap, const bitmap_info_t *binfo);
void bitmap_unset(bitmap_t *bitmap, const bitmap_info_t *binfo, size_t bit);
#endif
#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_BITMAP_C_))
JEMALLOC_INLINE bool
bitmap_full(bitmap_t *bitmap, const bitmap_info_t *binfo)
{
unsigned rgoff = binfo->levels[binfo->nlevels].group_offset - 1;
bitmap_t rg = bitmap[rgoff];
/* The bitmap is full iff the root group is 0. */
return (rg == 0);
}
JEMALLOC_INLINE bool
bitmap_get(bitmap_t *bitmap, const bitmap_info_t *binfo, size_t bit)
{
size_t goff;
bitmap_t g;
assert(bit < binfo->nbits);
goff = bit >> LG_BITMAP_GROUP_NBITS;
g = bitmap[goff];
return (!(g & (1LU << (bit & BITMAP_GROUP_NBITS_MASK))));
}
JEMALLOC_INLINE void
bitmap_set(bitmap_t *bitmap, const bitmap_info_t *binfo, size_t bit)
{
size_t goff;
bitmap_t *gp;
bitmap_t g;
assert(bit < binfo->nbits);
assert(bitmap_get(bitmap, binfo, bit) == false);
goff = bit >> LG_BITMAP_GROUP_NBITS;
gp = &bitmap[goff];
g = *gp;
assert(g & (1LU << (bit & BITMAP_GROUP_NBITS_MASK)));
g ^= 1LU << (bit & BITMAP_GROUP_NBITS_MASK);
*gp = g;
assert(bitmap_get(bitmap, binfo, bit));
/* Propagate group state transitions up the tree. */
if (g == 0) {
unsigned i;
for (i = 1; i < binfo->nlevels; i++) {
bit = goff;
goff = bit >> LG_BITMAP_GROUP_NBITS;
gp = &bitmap[binfo->levels[i].group_offset + goff];
g = *gp;
assert(g & (1LU << (bit & BITMAP_GROUP_NBITS_MASK)));
g ^= 1LU << (bit & BITMAP_GROUP_NBITS_MASK);
*gp = g;
if (g != 0)
break;
}
}
}
/* sfu: set first unset. */
JEMALLOC_INLINE size_t
bitmap_sfu(bitmap_t *bitmap, const bitmap_info_t *binfo)
{
size_t bit;
bitmap_t g;
unsigned i;
assert(bitmap_full(bitmap, binfo) == false);
i = binfo->nlevels - 1;
g = bitmap[binfo->levels[i].group_offset];
bit = ffsl(g) - 1;
while (i > 0) {
i--;
g = bitmap[binfo->levels[i].group_offset + bit];
bit = (bit << LG_BITMAP_GROUP_NBITS) + (ffsl(g) - 1);
}
bitmap_set(bitmap, binfo, bit);
return (bit);
}
JEMALLOC_INLINE void
bitmap_unset(bitmap_t *bitmap, const bitmap_info_t *binfo, size_t bit)
{
size_t goff;
bitmap_t *gp;
bitmap_t g;
bool propagate;
assert(bit < binfo->nbits);
assert(bitmap_get(bitmap, binfo, bit));
goff = bit >> LG_BITMAP_GROUP_NBITS;
gp = &bitmap[goff];
g = *gp;
propagate = (g == 0);
assert((g & (1LU << (bit & BITMAP_GROUP_NBITS_MASK))) == 0);
g ^= 1LU << (bit & BITMAP_GROUP_NBITS_MASK);
*gp = g;
assert(bitmap_get(bitmap, binfo, bit) == false);
/* Propagate group state transitions up the tree. */
if (propagate) {
unsigned i;
for (i = 1; i < binfo->nlevels; i++) {
bit = goff;
goff = bit >> LG_BITMAP_GROUP_NBITS;
gp = &bitmap[binfo->levels[i].group_offset + goff];
g = *gp;
propagate = (g == 0);
assert((g & (1LU << (bit & BITMAP_GROUP_NBITS_MASK)))
== 0);
g ^= 1LU << (bit & BITMAP_GROUP_NBITS_MASK);
*gp = g;
if (propagate == false)
break;
}
}
}
#endif
#endif /* JEMALLOC_H_INLINES */
/******************************************************************************/
deps/jemalloc/include/jemalloc/internal/chunk.h
Vendored
+63
View File
@@ -0,0 +1,63 @@
/******************************************************************************/
#ifdef JEMALLOC_H_TYPES
/*
* Size and alignment of memory chunks that are allocated by the OS's virtual
* memory system.
*/
#define LG_CHUNK_DEFAULT 22
/* Return the chunk address for allocation address a. */
#define CHUNK_ADDR2BASE(a) \
((void *)((uintptr_t)(a) & ~chunksize_mask))
/* Return the chunk offset of address a. */
#define CHUNK_ADDR2OFFSET(a) \
((size_t)((uintptr_t)(a) & chunksize_mask))
/* Return the smallest chunk multiple that is >= s. */
#define CHUNK_CEILING(s) \
(((s) + chunksize_mask) & ~chunksize_mask)
#endif /* JEMALLOC_H_TYPES */
/******************************************************************************/
#ifdef JEMALLOC_H_STRUCTS
#endif /* JEMALLOC_H_STRUCTS */
/******************************************************************************/
#ifdef JEMALLOC_H_EXTERNS
extern size_t opt_lg_chunk;
extern const char *opt_dss;
/* Protects stats_chunks; currently not used for any other purpose. */
extern malloc_mutex_t chunks_mtx;
/* Chunk statistics. */
extern chunk_stats_t stats_chunks;
extern rtree_t *chunks_rtree;
extern size_t chunksize;
extern size_t chunksize_mask; /* (chunksize - 1). */
extern size_t chunk_npages;
extern size_t map_bias; /* Number of arena chunk header pages. */
extern size_t arena_maxclass; /* Max size class for arenas. */
void *chunk_alloc(size_t size, size_t alignment, bool base, bool *zero,
dss_prec_t dss_prec);
void chunk_unmap(void *chunk, size_t size);
void chunk_dealloc(void *chunk, size_t size, bool unmap);
bool chunk_boot(void);
void chunk_prefork(void);
void chunk_postfork_parent(void);
void chunk_postfork_child(void);
#endif /* JEMALLOC_H_EXTERNS */
/******************************************************************************/
#ifdef JEMALLOC_H_INLINES
#endif /* JEMALLOC_H_INLINES */
/******************************************************************************/
#include "jemalloc/internal/chunk_dss.h"
#include "jemalloc/internal/chunk_mmap.h"
deps/jemalloc/include/jemalloc/internal/chunk_dss.h
Vendored
+38
View File
@@ -0,0 +1,38 @@
/******************************************************************************/
#ifdef JEMALLOC_H_TYPES
typedef enum {
dss_prec_disabled = 0,
dss_prec_primary = 1,
dss_prec_secondary = 2,
dss_prec_limit = 3
} dss_prec_t;
#define DSS_PREC_DEFAULT dss_prec_secondary
#define DSS_DEFAULT "secondary"
#endif /* JEMALLOC_H_TYPES */
/******************************************************************************/
#ifdef JEMALLOC_H_STRUCTS
extern const char *dss_prec_names[];
#endif /* JEMALLOC_H_STRUCTS */
/******************************************************************************/
#ifdef JEMALLOC_H_EXTERNS
dss_prec_t chunk_dss_prec_get(void);
bool chunk_dss_prec_set(dss_prec_t dss_prec);
void *chunk_alloc_dss(size_t size, size_t alignment, bool *zero);
bool chunk_in_dss(void *chunk);
bool chunk_dss_boot(void);
void chunk_dss_prefork(void);
void chunk_dss_postfork_parent(void);
void chunk_dss_postfork_child(void);
#endif /* JEMALLOC_H_EXTERNS */
/******************************************************************************/
#ifdef JEMALLOC_H_INLINES
#endif /* JEMALLOC_H_INLINES */
/******************************************************************************/
deps/jemalloc/include/jemalloc/internal/chunk_mmap.h
Vendored
+22
View File
@@ -0,0 +1,22 @@
/******************************************************************************/
#ifdef JEMALLOC_H_TYPES
#endif /* JEMALLOC_H_TYPES */
/******************************************************************************/
#ifdef JEMALLOC_H_STRUCTS
#endif /* JEMALLOC_H_STRUCTS */
/******************************************************************************/
#ifdef JEMALLOC_H_EXTERNS
bool pages_purge(void *addr, size_t length);
void *chunk_alloc_mmap(size_t size, size_t alignment, bool *zero);
bool chunk_dealloc_mmap(void *chunk, size_t size);
#endif /* JEMALLOC_H_EXTERNS */
/******************************************************************************/
#ifdef JEMALLOC_H_INLINES
#endif /* JEMALLOC_H_INLINES */
/******************************************************************************/
deps/jemalloc/include/jemalloc/internal/ckh.h
Vendored
+88
View File
@@ -0,0 +1,88 @@
/******************************************************************************/
#ifdef JEMALLOC_H_TYPES
typedef struct ckh_s ckh_t;
typedef struct ckhc_s ckhc_t;
/* Typedefs to allow easy function pointer passing. */
typedef void ckh_hash_t (const void *, size_t[2]);
typedef bool ckh_keycomp_t (const void *, const void *);
/* Maintain counters used to get an idea of performance. */
/* #define CKH_COUNT */
/* Print counter values in ckh_delete() (requires CKH_COUNT). */
/* #define CKH_VERBOSE */
/*
* There are 2^LG_CKH_BUCKET_CELLS cells in each hash table bucket. Try to fit
* one bucket per L1 cache line.
*/
#define LG_CKH_BUCKET_CELLS (LG_CACHELINE - LG_SIZEOF_PTR - 1)
#endif /* JEMALLOC_H_TYPES */
/******************************************************************************/
#ifdef JEMALLOC_H_STRUCTS
/* Hash table cell. */
struct ckhc_s {
const void *key;
const void *data;
};
struct ckh_s {
#ifdef CKH_COUNT
/* Counters used to get an idea of performance. */
uint64_t ngrows;
uint64_t nshrinks;
uint64_t nshrinkfails;
uint64_t ninserts;
uint64_t nrelocs;
#endif
/* Used for pseudo-random number generation. */
#define CKH_A 1103515241
#define CKH_C 12347
uint32_t prng_state;
/* Total number of items. */
size_t count;
/*
* Minimum and current number of hash table buckets. There are
* 2^LG_CKH_BUCKET_CELLS cells per bucket.
*/
unsigned lg_minbuckets;
unsigned lg_curbuckets;
/* Hash and comparison functions. */
ckh_hash_t *hash;
ckh_keycomp_t *keycomp;
/* Hash table with 2^lg_curbuckets buckets. */
ckhc_t *tab;
};
#endif /* JEMALLOC_H_STRUCTS */
/******************************************************************************/
#ifdef JEMALLOC_H_EXTERNS
bool ckh_new(ckh_t *ckh, size_t minitems, ckh_hash_t *hash,
ckh_keycomp_t *keycomp);
void ckh_delete(ckh_t *ckh);
size_t ckh_count(ckh_t *ckh);
bool ckh_iter(ckh_t *ckh, size_t *tabind, void **key, void **data);
bool ckh_insert(ckh_t *ckh, const void *key, const void *data);
bool ckh_remove(ckh_t *ckh, const void *searchkey, void **key,
void **data);
bool ckh_search(ckh_t *ckh, const void *seachkey, void **key, void **data);
void ckh_string_hash(const void *key, size_t r_hash[2]);
bool ckh_string_keycomp(const void *k1, const void *k2);
void ckh_pointer_hash(const void *key, size_t r_hash[2]);
bool ckh_pointer_keycomp(const void *k1, const void *k2);
#endif /* JEMALLOC_H_EXTERNS */
/******************************************************************************/
#ifdef JEMALLOC_H_INLINES
#endif /* JEMALLOC_H_INLINES */
/******************************************************************************/
deps/jemalloc/include/jemalloc/internal/ctl.h
Vendored
+117
View File
@@ -0,0 +1,117 @@
/******************************************************************************/
#ifdef JEMALLOC_H_TYPES
typedef struct ctl_node_s ctl_node_t;
typedef struct ctl_named_node_s ctl_named_node_t;
typedef struct ctl_indexed_node_s ctl_indexed_node_t;
typedef struct ctl_arena_stats_s ctl_arena_stats_t;
typedef struct ctl_stats_s ctl_stats_t;
#endif /* JEMALLOC_H_TYPES */
/******************************************************************************/
#ifdef JEMALLOC_H_STRUCTS
struct ctl_node_s {
bool named;
};
struct ctl_named_node_s {
struct ctl_node_s node;
const char *name;
/* If (nchildren == 0), this is a terminal node. */
unsigned nchildren;
const ctl_node_t *children;
int (*ctl)(const size_t *, size_t, void *, size_t *,
void *, size_t);
};
struct ctl_indexed_node_s {
struct ctl_node_s node;
const ctl_named_node_t *(*index)(const size_t *, size_t, size_t);
};
struct ctl_arena_stats_s {
bool initialized;
unsigned nthreads;
const char *dss;
size_t pactive;
size_t pdirty;
arena_stats_t astats;
/* Aggregate stats for small size classes, based on bin stats. */
size_t allocated_small;
uint64_t nmalloc_small;
uint64_t ndalloc_small;
uint64_t nrequests_small;
malloc_bin_stats_t bstats[NBINS];
malloc_large_stats_t *lstats; /* nlclasses elements. */
};
struct ctl_stats_s {
size_t allocated;
size_t active;
size_t mapped;
struct {
size_t current; /* stats_chunks.curchunks */
uint64_t total; /* stats_chunks.nchunks */
size_t high; /* stats_chunks.highchunks */
} chunks;
struct {
size_t allocated; /* huge_allocated */
uint64_t nmalloc; /* huge_nmalloc */
uint64_t ndalloc; /* huge_ndalloc */
} huge;
unsigned narenas;
ctl_arena_stats_t *arenas; /* (narenas + 1) elements. */
};
#endif /* JEMALLOC_H_STRUCTS */
/******************************************************************************/
#ifdef JEMALLOC_H_EXTERNS
int ctl_byname(const char *name, void *oldp, size_t *oldlenp, void *newp,
size_t newlen);
int ctl_nametomib(const char *name, size_t *mibp, size_t *miblenp);
int ctl_bymib(const size_t *mib, size_t miblen, void *oldp, size_t *oldlenp,
void *newp, size_t newlen);
bool ctl_boot(void);
void ctl_prefork(void);
void ctl_postfork_parent(void);
void ctl_postfork_child(void);
#define xmallctl(name, oldp, oldlenp, newp, newlen) do { \
if (je_mallctl(name, oldp, oldlenp, newp, newlen) \
!= 0) { \
malloc_printf( \
"<jemalloc>: Failure in xmallctl(\"%s\", ...)\n", \
name); \
abort(); \
} \
} while (0)
#define xmallctlnametomib(name, mibp, miblenp) do { \
if (je_mallctlnametomib(name, mibp, miblenp) != 0) { \
malloc_printf("<jemalloc>: Failure in " \
"xmallctlnametomib(\"%s\", ...)\n", name); \
abort(); \
} \
} while (0)
#define xmallctlbymib(mib, miblen, oldp, oldlenp, newp, newlen) do { \
if (je_mallctlbymib(mib, miblen, oldp, oldlenp, newp, \
newlen) != 0) { \
malloc_write( \
"<jemalloc>: Failure in xmallctlbymib()\n"); \
abort(); \
} \
} while (0)
#endif /* JEMALLOC_H_EXTERNS */
/******************************************************************************/
#ifdef JEMALLOC_H_INLINES
#endif /* JEMALLOC_H_INLINES */
/******************************************************************************/
deps/jemalloc/include/jemalloc/internal/extent.h
Vendored
+46
View File
@@ -0,0 +1,46 @@
/******************************************************************************/
#ifdef JEMALLOC_H_TYPES
typedef struct extent_node_s extent_node_t;
#endif /* JEMALLOC_H_TYPES */
/******************************************************************************/
#ifdef JEMALLOC_H_STRUCTS
/* Tree of extents. */
struct extent_node_s {
/* Linkage for the size/address-ordered tree. */
rb_node(extent_node_t) link_szad;
/* Linkage for the address-ordered tree. */
rb_node(extent_node_t) link_ad;
/* Profile counters, used for huge objects. */
prof_ctx_t *prof_ctx;
/* Pointer to the extent that this tree node is responsible for. */
void *addr;
/* Total region size. */
size_t size;
/* True if zero-filled; used by chunk recycling code. */
bool zeroed;
};
typedef rb_tree(extent_node_t) extent_tree_t;
#endif /* JEMALLOC_H_STRUCTS */
/******************************************************************************/
#ifdef JEMALLOC_H_EXTERNS
rb_proto(, extent_tree_szad_, extent_tree_t, extent_node_t)
rb_proto(, extent_tree_ad_, extent_tree_t, extent_node_t)
#endif /* JEMALLOC_H_EXTERNS */
/******************************************************************************/
#ifdef JEMALLOC_H_INLINES
#endif /* JEMALLOC_H_INLINES */
/******************************************************************************/
deps/jemalloc/include/jemalloc/internal/hash.h
Vendored
+335
View File
@@ -0,0 +1,335 @@
/*
* The following hash function is based on MurmurHash3, placed into the public
* domain by Austin Appleby. See http://code.google.com/p/smhasher/ for
* details.
*/
/******************************************************************************/
#ifdef JEMALLOC_H_TYPES
#endif /* JEMALLOC_H_TYPES */
/******************************************************************************/
#ifdef JEMALLOC_H_STRUCTS
#endif /* JEMALLOC_H_STRUCTS */
/******************************************************************************/
#ifdef JEMALLOC_H_EXTERNS
#endif /* JEMALLOC_H_EXTERNS */
/******************************************************************************/
#ifdef JEMALLOC_H_INLINES
#ifndef JEMALLOC_ENABLE_INLINE
uint32_t hash_x86_32(const void *key, int len, uint32_t seed);
void hash_x86_128(const void *key, const int len, uint32_t seed,
uint64_t r_out[2]);
void hash_x64_128(const void *key, const int len, const uint32_t seed,
uint64_t r_out[2]);
void hash(const void *key, size_t len, const uint32_t seed,
size_t r_hash[2]);
#endif
#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_HASH_C_))
/******************************************************************************/
/* Internal implementation. */
JEMALLOC_INLINE uint32_t
hash_rotl_32(uint32_t x, int8_t r)
{
return (x << r) | (x >> (32 - r));
}
JEMALLOC_INLINE uint64_t
hash_rotl_64(uint64_t x, int8_t r)
{
return (x << r) | (x >> (64 - r));
}
JEMALLOC_INLINE uint32_t
hash_get_block_32(const uint32_t *p, int i)
{
return (p[i]);
}
JEMALLOC_INLINE uint64_t
hash_get_block_64(const uint64_t *p, int i)
{
return (p[i]);
}
JEMALLOC_INLINE uint32_t
hash_fmix_32(uint32_t h)
{
h ^= h >> 16;
h *= 0x85ebca6b;
h ^= h >> 13;
h *= 0xc2b2ae35;
h ^= h >> 16;
return (h);
}
JEMALLOC_INLINE uint64_t
hash_fmix_64(uint64_t k)
{
k ^= k >> 33;
k *= QU(0xff51afd7ed558ccdLLU);
k ^= k >> 33;
k *= QU(0xc4ceb9fe1a85ec53LLU);
k ^= k >> 33;
return (k);
}
JEMALLOC_INLINE uint32_t
hash_x86_32(const void *key, int len, uint32_t seed)
{
const uint8_t *data = (const uint8_t *) key;
const int nblocks = len / 4;
uint32_t h1 = seed;
const uint32_t c1 = 0xcc9e2d51;
const uint32_t c2 = 0x1b873593;
/* body */
{
const uint32_t *blocks = (const uint32_t *) (data + nblocks*4);
int i;
for (i = -nblocks; i; i++) {
uint32_t k1 = hash_get_block_32(blocks, i);
k1 *= c1;
k1 = hash_rotl_32(k1, 15);
k1 *= c2;
h1 ^= k1;
h1 = hash_rotl_32(h1, 13);
h1 = h1*5 + 0xe6546b64;
}
}
/* tail */
{
const uint8_t *tail = (const uint8_t *) (data + nblocks*4);
uint32_t k1 = 0;
switch (len & 3) {
case 3: k1 ^= tail[2] << 16;
case 2: k1 ^= tail[1] << 8;
case 1: k1 ^= tail[0]; k1 *= c1; k1 = hash_rotl_32(k1, 15);
k1 *= c2; h1 ^= k1;
}
}
/* finalization */
h1 ^= len;
h1 = hash_fmix_32(h1);
return (h1);
}
UNUSED JEMALLOC_INLINE void
hash_x86_128(const void *key, const int len, uint32_t seed,
uint64_t r_out[2])
{
const uint8_t * data = (const uint8_t *) key;
const int nblocks = len / 16;
uint32_t h1 = seed;
uint32_t h2 = seed;
uint32_t h3 = seed;
uint32_t h4 = seed;
const uint32_t c1 = 0x239b961b;
const uint32_t c2 = 0xab0e9789;
const uint32_t c3 = 0x38b34ae5;
const uint32_t c4 = 0xa1e38b93;
/* body */
{
const uint32_t *blocks = (const uint32_t *) (data + nblocks*16);
int i;
for (i = -nblocks; i; i++) {
uint32_t k1 = hash_get_block_32(blocks, i*4 + 0);
uint32_t k2 = hash_get_block_32(blocks, i*4 + 1);
uint32_t k3 = hash_get_block_32(blocks, i*4 + 2);
uint32_t k4 = hash_get_block_32(blocks, i*4 + 3);
k1 *= c1; k1 = hash_rotl_32(k1, 15); k1 *= c2; h1 ^= k1;
h1 = hash_rotl_32(h1, 19); h1 += h2;
h1 = h1*5 + 0x561ccd1b;
k2 *= c2; k2 = hash_rotl_32(k2, 16); k2 *= c3; h2 ^= k2;
h2 = hash_rotl_32(h2, 17); h2 += h3;
h2 = h2*5 + 0x0bcaa747;
k3 *= c3; k3 = hash_rotl_32(k3, 17); k3 *= c4; h3 ^= k3;
h3 = hash_rotl_32(h3, 15); h3 += h4;
h3 = h3*5 + 0x96cd1c35;
k4 *= c4; k4 = hash_rotl_32(k4, 18); k4 *= c1; h4 ^= k4;
h4 = hash_rotl_32(h4, 13); h4 += h1;
h4 = h4*5 + 0x32ac3b17;
}
}
/* tail */
{
const uint8_t *tail = (const uint8_t *) (data + nblocks*16);
uint32_t k1 = 0;
uint32_t k2 = 0;
uint32_t k3 = 0;
uint32_t k4 = 0;
switch (len & 15) {
case 15: k4 ^= tail[14] << 16;
case 14: k4 ^= tail[13] << 8;
case 13: k4 ^= tail[12] << 0;
k4 *= c4; k4 = hash_rotl_32(k4, 18); k4 *= c1; h4 ^= k4;
case 12: k3 ^= tail[11] << 24;
case 11: k3 ^= tail[10] << 16;
case 10: k3 ^= tail[ 9] << 8;
case 9: k3 ^= tail[ 8] << 0;
k3 *= c3; k3 = hash_rotl_32(k3, 17); k3 *= c4; h3 ^= k3;
case 8: k2 ^= tail[ 7] << 24;
case 7: k2 ^= tail[ 6] << 16;
case 6: k2 ^= tail[ 5] << 8;
case 5: k2 ^= tail[ 4] << 0;
k2 *= c2; k2 = hash_rotl_32(k2, 16); k2 *= c3; h2 ^= k2;
case 4: k1 ^= tail[ 3] << 24;
case 3: k1 ^= tail[ 2] << 16;
case 2: k1 ^= tail[ 1] << 8;
case 1: k1 ^= tail[ 0] << 0;
k1 *= c1; k1 = hash_rotl_32(k1, 15); k1 *= c2; h1 ^= k1;
}
}
/* finalization */
h1 ^= len; h2 ^= len; h3 ^= len; h4 ^= len;
h1 += h2; h1 += h3; h1 += h4;
h2 += h1; h3 += h1; h4 += h1;
h1 = hash_fmix_32(h1);
h2 = hash_fmix_32(h2);
h3 = hash_fmix_32(h3);
h4 = hash_fmix_32(h4);
h1 += h2; h1 += h3; h1 += h4;
h2 += h1; h3 += h1; h4 += h1;
r_out[0] = (((uint64_t) h2) << 32) | h1;
r_out[1] = (((uint64_t) h4) << 32) | h3;
}
UNUSED JEMALLOC_INLINE void
hash_x64_128(const void *key, const int len, const uint32_t seed,
uint64_t r_out[2])
{
const uint8_t *data = (const uint8_t *) key;
const int nblocks = len / 16;
uint64_t h1 = seed;
uint64_t h2 = seed;
const uint64_t c1 = QU(0x87c37b91114253d5LLU);
const uint64_t c2 = QU(0x4cf5ad432745937fLLU);
/* body */
{
const uint64_t *blocks = (const uint64_t *) (data);
int i;
for (i = 0; i < nblocks; i++) {
uint64_t k1 = hash_get_block_64(blocks, i*2 + 0);
uint64_t k2 = hash_get_block_64(blocks, i*2 + 1);
k1 *= c1; k1 = hash_rotl_64(k1, 31); k1 *= c2; h1 ^= k1;
h1 = hash_rotl_64(h1, 27); h1 += h2;
h1 = h1*5 + 0x52dce729;
k2 *= c2; k2 = hash_rotl_64(k2, 33); k2 *= c1; h2 ^= k2;
h2 = hash_rotl_64(h2, 31); h2 += h1;
h2 = h2*5 + 0x38495ab5;
}
}
/* tail */
{
const uint8_t *tail = (const uint8_t*)(data + nblocks*16);
uint64_t k1 = 0;
uint64_t k2 = 0;
switch (len & 15) {
case 15: k2 ^= ((uint64_t)(tail[14])) << 48;
case 14: k2 ^= ((uint64_t)(tail[13])) << 40;
case 13: k2 ^= ((uint64_t)(tail[12])) << 32;
case 12: k2 ^= ((uint64_t)(tail[11])) << 24;
case 11: k2 ^= ((uint64_t)(tail[10])) << 16;
case 10: k2 ^= ((uint64_t)(tail[ 9])) << 8;
case 9: k2 ^= ((uint64_t)(tail[ 8])) << 0;
k2 *= c2; k2 = hash_rotl_64(k2, 33); k2 *= c1; h2 ^= k2;
case 8: k1 ^= ((uint64_t)(tail[ 7])) << 56;
case 7: k1 ^= ((uint64_t)(tail[ 6])) << 48;
case 6: k1 ^= ((uint64_t)(tail[ 5])) << 40;
case 5: k1 ^= ((uint64_t)(tail[ 4])) << 32;
case 4: k1 ^= ((uint64_t)(tail[ 3])) << 24;
case 3: k1 ^= ((uint64_t)(tail[ 2])) << 16;
case 2: k1 ^= ((uint64_t)(tail[ 1])) << 8;
case 1: k1 ^= ((uint64_t)(tail[ 0])) << 0;
k1 *= c1; k1 = hash_rotl_64(k1, 31); k1 *= c2; h1 ^= k1;
}
}
/* finalization */
h1 ^= len; h2 ^= len;
h1 += h2;
h2 += h1;
h1 = hash_fmix_64(h1);
h2 = hash_fmix_64(h2);
h1 += h2;
h2 += h1;
r_out[0] = h1;
r_out[1] = h2;
}
/******************************************************************************/
/* API. */
JEMALLOC_INLINE void
hash(const void *key, size_t len, const uint32_t seed, size_t r_hash[2])
{
#if (LG_SIZEOF_PTR == 3 && !defined(JEMALLOC_BIG_ENDIAN))
hash_x64_128(key, len, seed, (uint64_t *)r_hash);
#else
uint64_t hashes[2];
hash_x86_128(key, len, seed, hashes);
r_hash[0] = (size_t)hashes[0];
r_hash[1] = (size_t)hashes[1];
#endif
}
#endif
#endif /* JEMALLOC_H_INLINES */
/******************************************************************************/
deps/jemalloc/include/jemalloc/internal/huge.h
Vendored
+46
View File
@@ -0,0 +1,46 @@
/******************************************************************************/
#ifdef JEMALLOC_H_TYPES
#endif /* JEMALLOC_H_TYPES */
/******************************************************************************/
#ifdef JEMALLOC_H_STRUCTS
#endif /* JEMALLOC_H_STRUCTS */
/******************************************************************************/
#ifdef JEMALLOC_H_EXTERNS
/* Huge allocation statistics. */
extern uint64_t huge_nmalloc;
extern uint64_t huge_ndalloc;
extern size_t huge_allocated;
/* Protects chunk-related data structures. */
extern malloc_mutex_t huge_mtx;
void *huge_malloc(size_t size, bool zero, dss_prec_t dss_prec);
void *huge_palloc(size_t size, size_t alignment, bool zero,
dss_prec_t dss_prec);
bool huge_ralloc_no_move(void *ptr, size_t oldsize, size_t size,
size_t extra);
void *huge_ralloc(void *ptr, size_t oldsize, size_t size, size_t extra,
size_t alignment, bool zero, bool try_tcache_dalloc, dss_prec_t dss_prec);
#ifdef JEMALLOC_JET
typedef void (huge_dalloc_junk_t)(void *, size_t);
extern huge_dalloc_junk_t *huge_dalloc_junk;
#endif
void huge_dalloc(void *ptr, bool unmap);
size_t huge_salloc(const void *ptr);
dss_prec_t huge_dss_prec_get(arena_t *arena);
prof_ctx_t *huge_prof_ctx_get(const void *ptr);
void huge_prof_ctx_set(const void *ptr, prof_ctx_t *ctx);
bool huge_boot(void);
void huge_prefork(void);
void huge_postfork_parent(void);
void huge_postfork_child(void);
#endif /* JEMALLOC_H_EXTERNS */
/******************************************************************************/
#ifdef JEMALLOC_H_INLINES
#endif /* JEMALLOC_H_INLINES */
/******************************************************************************/
deps/jemalloc/include/jemalloc/internal/jemalloc_internal.h.in
Vendored
+1028
View File
File diff suppressed because it is too large Load Diff
deps/jemalloc/include/jemalloc/internal/jemalloc_internal_defs.h.in
Vendored
+205
View File
@@ -0,0 +1,205 @@
#ifndef JEMALLOC_INTERNAL_DEFS_H_
#define JEMALLOC_INTERNAL_DEFS_H_
/*
* If JEMALLOC_PREFIX is defined via --with-jemalloc-prefix, it will cause all
* public APIs to be prefixed. This makes it possible, with some care, to use
* multiple allocators simultaneously.
*/
#undef JEMALLOC_PREFIX
#undef JEMALLOC_CPREFIX
/*
* JEMALLOC_PRIVATE_NAMESPACE is used as a prefix for all library-private APIs.
* For shared libraries, symbol visibility mechanisms prevent these symbols
* from being exported, but for static libraries, naming collisions are a real
* possibility.
*/
#undef JEMALLOC_PRIVATE_NAMESPACE
/*
* Hyper-threaded CPUs may need a special instruction inside spin loops in
* order to yield to another virtual CPU.
*/
#undef CPU_SPINWAIT
/* Defined if the equivalent of FreeBSD's atomic(9) functions are available. */
#undef JEMALLOC_ATOMIC9
/*
* Defined if OSAtomic*() functions are available, as provided by Darwin, and
* documented in the atomic(3) manual page.
*/
#undef JEMALLOC_OSATOMIC
/*
* Defined if __sync_add_and_fetch(uint32_t *, uint32_t) and
* __sync_sub_and_fetch(uint32_t *, uint32_t) are available, despite
* __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4 not being defined (which means the
* functions are defined in libgcc instead of being inlines)
*/
#undef JE_FORCE_SYNC_COMPARE_AND_SWAP_4
/*
* Defined if __sync_add_and_fetch(uint64_t *, uint64_t) and
* __sync_sub_and_fetch(uint64_t *, uint64_t) are available, despite
* __GCC_HAVE_SYNC_COMPARE_AND_SWAP_8 not being defined (which means the
* functions are defined in libgcc instead of being inlines)
*/
#undef JE_FORCE_SYNC_COMPARE_AND_SWAP_8
/*
* Defined if OSSpin*() functions are available, as provided by Darwin, and
* documented in the spinlock(3) manual page.
*/
#undef JEMALLOC_OSSPIN
/*
* Defined if _malloc_thread_cleanup() exists. At least in the case of
* FreeBSD, pthread_key_create() allocates, which if used during malloc
* bootstrapping will cause recursion into the pthreads library. Therefore, if
* _malloc_thread_cleanup() exists, use it as the basis for thread cleanup in
* malloc_tsd.
*/
#undef JEMALLOC_MALLOC_THREAD_CLEANUP
/*
* Defined if threaded initialization is known to be safe on this platform.
* Among other things, it must be possible to initialize a mutex without
* triggering allocation in order for threaded allocation to be safe.
*/
#undef JEMALLOC_THREADED_INIT
/*
* Defined if the pthreads implementation defines
* _pthread_mutex_init_calloc_cb(), in which case the function is used in order
* to avoid recursive allocation during mutex initialization.
*/
#undef JEMALLOC_MUTEX_INIT_CB
/* Defined if sbrk() is supported. */
#undef JEMALLOC_HAVE_SBRK
/* Non-empty if the tls_model attribute is supported. */
#undef JEMALLOC_TLS_MODEL
/* JEMALLOC_CC_SILENCE enables code that silences unuseful compiler warnings. */
#undef JEMALLOC_CC_SILENCE
/* JEMALLOC_CODE_COVERAGE enables test code coverage analysis. */
#undef JEMALLOC_CODE_COVERAGE
/*
* JEMALLOC_DEBUG enables assertions and other sanity checks, and disables
* inline functions.
*/
#undef JEMALLOC_DEBUG
/* JEMALLOC_STATS enables statistics calculation. */
#undef JEMALLOC_STATS
/* JEMALLOC_PROF enables allocation profiling. */
#undef JEMALLOC_PROF
/* Use libunwind for profile backtracing if defined. */
#undef JEMALLOC_PROF_LIBUNWIND
/* Use libgcc for profile backtracing if defined. */
#undef JEMALLOC_PROF_LIBGCC
/* Use gcc intrinsics for profile backtracing if defined. */
#undef JEMALLOC_PROF_GCC
/*
* JEMALLOC_TCACHE enables a thread-specific caching layer for small objects.
* This makes it possible to allocate/deallocate objects without any locking
* when the cache is in the steady state.
*/
#undef JEMALLOC_TCACHE
/*
* JEMALLOC_DSS enables use of sbrk(2) to allocate chunks from the data storage
* segment (DSS).
*/
#undef JEMALLOC_DSS
/* Support memory filling (junk/zero/quarantine/redzone). */
#undef JEMALLOC_FILL
/* Support utrace(2)-based tracing. */
#undef JEMALLOC_UTRACE
/* Support Valgrind. */
#undef JEMALLOC_VALGRIND
/* Support optional abort() on OOM. */
#undef JEMALLOC_XMALLOC
/* Support lazy locking (avoid locking unless a second thread is launched). */
#undef JEMALLOC_LAZY_LOCK
/* One page is 2^STATIC_PAGE_SHIFT bytes. */
#undef STATIC_PAGE_SHIFT
/*
* If defined, use munmap() to unmap freed chunks, rather than storing them for
* later reuse. This is disabled by default on Linux because common sequences
* of mmap()/munmap() calls will cause virtual memory map holes.
*/
#undef JEMALLOC_MUNMAP
/*
* If defined, use mremap(...MREMAP_FIXED...) for huge realloc(). This is
* disabled by default because it is Linux-specific and it will cause virtual
* memory map holes, much like munmap(2) does.
*/
#undef JEMALLOC_MREMAP
/* TLS is used to map arenas and magazine caches to threads. */
#undef JEMALLOC_TLS
/*
* JEMALLOC_IVSALLOC enables ivsalloc(), which verifies that pointers reside
* within jemalloc-owned chunks before dereferencing them.
*/
#undef JEMALLOC_IVSALLOC
/*
* Darwin (OS X) uses zones to work around Mach-O symbol override shortcomings.
*/
#undef JEMALLOC_ZONE
#undef JEMALLOC_ZONE_VERSION
/*
* Methods for purging unused pages differ between operating systems.
*
* madvise(..., MADV_DONTNEED) : On Linux, this immediately discards pages,
* such that new pages will be demand-zeroed if
* the address region is later touched.
* madvise(..., MADV_FREE) : On FreeBSD and Darwin, this marks pages as being
* unused, such that they will be discarded rather
* than swapped out.
*/
#undef JEMALLOC_PURGE_MADVISE_DONTNEED
#undef JEMALLOC_PURGE_MADVISE_FREE
/*
* Define if operating system has alloca.h header.
*/
#undef JEMALLOC_HAS_ALLOCA_H
/* C99 restrict keyword supported. */
#undef JEMALLOC_HAS_RESTRICT
/* For use by hash code. */
#undef JEMALLOC_BIG_ENDIAN
/* sizeof(int) == 2^LG_SIZEOF_INT. */
#undef LG_SIZEOF_INT
/* sizeof(long) == 2^LG_SIZEOF_LONG. */
#undef LG_SIZEOF_LONG
/* sizeof(intmax_t) == 2^LG_SIZEOF_INTMAX_T. */
#undef LG_SIZEOF_INTMAX_T
#endif /* JEMALLOC_INTERNAL_DEFS_H_ */
deps/jemalloc/include/jemalloc/internal/jemalloc_internal_macros.h
Vendored
+51
View File
@@ -0,0 +1,51 @@
/*
* JEMALLOC_ALWAYS_INLINE and JEMALLOC_INLINE are used within header files for
* functions that are static inline functions if inlining is enabled, and
* single-definition library-private functions if inlining is disabled.
*
* JEMALLOC_ALWAYS_INLINE_C and JEMALLOC_INLINE_C are for use in .c files, in
* which case the denoted functions are always static, regardless of whether
* inlining is enabled.
*/
#if defined(JEMALLOC_DEBUG) || defined(JEMALLOC_CODE_COVERAGE)
/* Disable inlining to make debugging/profiling easier. */
# define JEMALLOC_ALWAYS_INLINE
# define JEMALLOC_ALWAYS_INLINE_C static
# define JEMALLOC_INLINE
# define JEMALLOC_INLINE_C static
# define inline
#else
# define JEMALLOC_ENABLE_INLINE
# ifdef JEMALLOC_HAVE_ATTR
# define JEMALLOC_ALWAYS_INLINE \
static inline JEMALLOC_ATTR(unused) JEMALLOC_ATTR(always_inline)
# define JEMALLOC_ALWAYS_INLINE_C \
static inline JEMALLOC_ATTR(always_inline)
# else
# define JEMALLOC_ALWAYS_INLINE static inline
# define JEMALLOC_ALWAYS_INLINE_C static inline
# endif
# define JEMALLOC_INLINE static inline
# define JEMALLOC_INLINE_C static inline
# ifdef _MSC_VER
# define inline _inline
# endif
#endif
#ifdef JEMALLOC_CC_SILENCE
# define UNUSED JEMALLOC_ATTR(unused)
#else
# define UNUSED
#endif
#define ZU(z) ((size_t)z)
#define QU(q) ((uint64_t)q)
#define QI(q) ((int64_t)q)
#ifndef __DECONST
# define __DECONST(type, var) ((type)(uintptr_t)(const void *)(var))
#endif
#ifndef JEMALLOC_HAS_RESTRICT
# define restrict
#endif
deps/jemalloc/include/jemalloc/internal/mb.h
Vendored
+115
View File
@@ -0,0 +1,115 @@
/******************************************************************************/
#ifdef JEMALLOC_H_TYPES
#endif /* JEMALLOC_H_TYPES */
/******************************************************************************/
#ifdef JEMALLOC_H_STRUCTS
#endif /* JEMALLOC_H_STRUCTS */
/******************************************************************************/
#ifdef JEMALLOC_H_EXTERNS
#endif /* JEMALLOC_H_EXTERNS */
/******************************************************************************/
#ifdef JEMALLOC_H_INLINES
#ifndef JEMALLOC_ENABLE_INLINE
void mb_write(void);
#endif
#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_MB_C_))
#ifdef __i386__
/*
* According to the Intel Architecture Software Developer's Manual, current
* processors execute instructions in order from the perspective of other
* processors in a multiprocessor system, but 1) Intel reserves the right to
* change that, and 2) the compiler's optimizer could re-order instructions if
* there weren't some form of barrier. Therefore, even if running on an
* architecture that does not need memory barriers (everything through at least
* i686), an "optimizer barrier" is necessary.
*/
JEMALLOC_INLINE void
mb_write(void)
{
# if 0
/* This is a true memory barrier. */
asm volatile ("pusha;"
"xor %%eax,%%eax;"
"cpuid;"
"popa;"
: /* Outputs. */
: /* Inputs. */
: "memory" /* Clobbers. */
);
#else
/*
* This is hopefully enough to keep the compiler from reordering
* instructions around this one.
*/
asm volatile ("nop;"
: /* Outputs. */
: /* Inputs. */
: "memory" /* Clobbers. */
);
#endif
}
#elif (defined(__amd64__) || defined(__x86_64__))
JEMALLOC_INLINE void
mb_write(void)
{
asm volatile ("sfence"
: /* Outputs. */
: /* Inputs. */
: "memory" /* Clobbers. */
);
}
#elif defined(__powerpc__)
JEMALLOC_INLINE void
mb_write(void)
{
asm volatile ("eieio"
: /* Outputs. */
: /* Inputs. */
: "memory" /* Clobbers. */
);
}
#elif defined(__sparc64__)
JEMALLOC_INLINE void
mb_write(void)
{
asm volatile ("membar #StoreStore"
: /* Outputs. */
: /* Inputs. */
: "memory" /* Clobbers. */
);
}
#elif defined(__tile__)
JEMALLOC_INLINE void
mb_write(void)
{
__sync_synchronize();
}
#else
/*
* This is much slower than a simple memory barrier, but the semantics of mutex
* unlock make this work.
*/
JEMALLOC_INLINE void
mb_write(void)
{
malloc_mutex_t mtx;
malloc_mutex_init(&mtx);
malloc_mutex_lock(&mtx);
malloc_mutex_unlock(&mtx);
}
#endif
#endif
#endif /* JEMALLOC_H_INLINES */
/******************************************************************************/
deps/jemalloc/include/jemalloc/internal/mutex.h
Vendored
+99
View File
@@ -0,0 +1,99 @@
/******************************************************************************/
#ifdef JEMALLOC_H_TYPES
typedef struct malloc_mutex_s malloc_mutex_t;
#ifdef _WIN32
# define MALLOC_MUTEX_INITIALIZER
#elif (defined(JEMALLOC_OSSPIN))
# define MALLOC_MUTEX_INITIALIZER {0}
#elif (defined(JEMALLOC_MUTEX_INIT_CB))
# define MALLOC_MUTEX_INITIALIZER {PTHREAD_MUTEX_INITIALIZER, NULL}
#else
# if (defined(PTHREAD_MUTEX_ADAPTIVE_NP) && \
defined(PTHREAD_ADAPTIVE_MUTEX_INITIALIZER_NP))
# define MALLOC_MUTEX_TYPE PTHREAD_MUTEX_ADAPTIVE_NP
# define MALLOC_MUTEX_INITIALIZER {PTHREAD_ADAPTIVE_MUTEX_INITIALIZER_NP}
# else
# define MALLOC_MUTEX_TYPE PTHREAD_MUTEX_DEFAULT
# define MALLOC_MUTEX_INITIALIZER {PTHREAD_MUTEX_INITIALIZER}
# endif
#endif
#endif /* JEMALLOC_H_TYPES */
/******************************************************************************/
#ifdef JEMALLOC_H_STRUCTS
struct malloc_mutex_s {
#ifdef _WIN32
CRITICAL_SECTION lock;
#elif (defined(JEMALLOC_OSSPIN))
OSSpinLock lock;
#elif (defined(JEMALLOC_MUTEX_INIT_CB))
pthread_mutex_t lock;
malloc_mutex_t *postponed_next;
#else
pthread_mutex_t lock;
#endif
};
#endif /* JEMALLOC_H_STRUCTS */
/******************************************************************************/
#ifdef JEMALLOC_H_EXTERNS
#ifdef JEMALLOC_LAZY_LOCK
extern bool isthreaded;
#else
# undef isthreaded /* Undo private_namespace.h definition. */
# define isthreaded true
#endif
bool malloc_mutex_init(malloc_mutex_t *mutex);
void malloc_mutex_prefork(malloc_mutex_t *mutex);
void malloc_mutex_postfork_parent(malloc_mutex_t *mutex);
void malloc_mutex_postfork_child(malloc_mutex_t *mutex);
bool mutex_boot(void);
#endif /* JEMALLOC_H_EXTERNS */
/******************************************************************************/
#ifdef JEMALLOC_H_INLINES
#ifndef JEMALLOC_ENABLE_INLINE
void malloc_mutex_lock(malloc_mutex_t *mutex);
void malloc_mutex_unlock(malloc_mutex_t *mutex);
#endif
#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_MUTEX_C_))
JEMALLOC_INLINE void
malloc_mutex_lock(malloc_mutex_t *mutex)
{
if (isthreaded) {
#ifdef _WIN32
EnterCriticalSection(&mutex->lock);
#elif (defined(JEMALLOC_OSSPIN))
OSSpinLockLock(&mutex->lock);
#else
pthread_mutex_lock(&mutex->lock);
#endif
}
}
JEMALLOC_INLINE void
malloc_mutex_unlock(malloc_mutex_t *mutex)
{
if (isthreaded) {
#ifdef _WIN32
LeaveCriticalSection(&mutex->lock);
#elif (defined(JEMALLOC_OSSPIN))
OSSpinLockUnlock(&mutex->lock);
#else
pthread_mutex_unlock(&mutex->lock);
#endif
}
}
#endif
#endif /* JEMALLOC_H_INLINES */
/******************************************************************************/
deps/jemalloc/include/jemalloc/internal/private_namespace.sh
Vendored Executable
+5
View File
@@ -0,0 +1,5 @@
#!/bin/sh
for symbol in `cat $1` ; do
echo "#define ${symbol} JEMALLOC_N(${symbol})"
done
deps/jemalloc/include/jemalloc/internal/private_symbols.txt
Vendored
+413
View File
@@ -0,0 +1,413 @@
a0calloc
a0free
a0malloc
arena_alloc_junk_small
arena_bin_index
arena_bin_info
arena_boot
arena_dalloc
arena_dalloc_bin
arena_dalloc_bin_locked
arena_dalloc_junk_large
arena_dalloc_junk_small
arena_dalloc_large
arena_dalloc_large_locked
arena_dalloc_small
arena_dss_prec_get
arena_dss_prec_set
arena_malloc
arena_malloc_large
arena_malloc_small
arena_mapbits_allocated_get
arena_mapbits_binind_get
arena_mapbits_dirty_get
arena_mapbits_get
arena_mapbits_large_binind_set
arena_mapbits_large_get
arena_mapbits_large_set
arena_mapbits_large_size_get
arena_mapbits_small_runind_get
arena_mapbits_small_set
arena_mapbits_unallocated_set
arena_mapbits_unallocated_size_get
arena_mapbits_unallocated_size_set
arena_mapbits_unzeroed_get
arena_mapbits_unzeroed_set
arena_mapbitsp_get
arena_mapbitsp_read
arena_mapbitsp_write
arena_mapp_get
arena_maxclass
arena_new
arena_palloc
arena_postfork_child
arena_postfork_parent
arena_prefork
arena_prof_accum
arena_prof_accum_impl
arena_prof_accum_locked
arena_prof_ctx_get
arena_prof_ctx_set
arena_prof_promoted
arena_ptr_small_binind_get
arena_purge_all
arena_quarantine_junk_small
arena_ralloc
arena_ralloc_junk_large
arena_ralloc_no_move
arena_redzone_corruption
arena_run_regind
arena_salloc
arena_stats_merge
arena_tcache_fill_small
arenas
arenas_booted
arenas_cleanup
arenas_extend
arenas_initialized
arenas_lock
arenas_tls
arenas_tsd
arenas_tsd_boot
arenas_tsd_cleanup_wrapper
arenas_tsd_get
arenas_tsd_get_wrapper
arenas_tsd_init_head
arenas_tsd_set
atomic_add_u
atomic_add_uint32
atomic_add_uint64
atomic_add_z
atomic_sub_u
atomic_sub_uint32
atomic_sub_uint64
atomic_sub_z
base_alloc
base_boot
base_calloc
base_node_alloc
base_node_dealloc
base_postfork_child
base_postfork_parent
base_prefork
bitmap_full
bitmap_get
bitmap_info_init
bitmap_info_ngroups
bitmap_init
bitmap_set
bitmap_sfu
bitmap_size
bitmap_unset
bt_init
buferror
choose_arena
choose_arena_hard
chunk_alloc
chunk_alloc_dss
chunk_alloc_mmap
chunk_boot
chunk_dealloc
chunk_dealloc_mmap
chunk_dss_boot
chunk_dss_postfork_child
chunk_dss_postfork_parent
chunk_dss_prec_get
chunk_dss_prec_set
chunk_dss_prefork
chunk_in_dss
chunk_npages
chunk_postfork_child
chunk_postfork_parent
chunk_prefork
chunk_unmap
chunks_mtx
chunks_rtree
chunksize
chunksize_mask
ckh_bucket_search
ckh_count
ckh_delete
ckh_evict_reloc_insert
ckh_insert
ckh_isearch
ckh_iter
ckh_new
ckh_pointer_hash
ckh_pointer_keycomp
ckh_rebuild
ckh_remove
ckh_search
ckh_string_hash
ckh_string_keycomp
ckh_try_bucket_insert
ckh_try_insert
ctl_boot
ctl_bymib
ctl_byname
ctl_nametomib
ctl_postfork_child
ctl_postfork_parent
ctl_prefork
dss_prec_names
extent_tree_ad_first
extent_tree_ad_insert
extent_tree_ad_iter
extent_tree_ad_iter_recurse
extent_tree_ad_iter_start
extent_tree_ad_last
extent_tree_ad_new
extent_tree_ad_next
extent_tree_ad_nsearch
extent_tree_ad_prev
extent_tree_ad_psearch
extent_tree_ad_remove
extent_tree_ad_reverse_iter
extent_tree_ad_reverse_iter_recurse
extent_tree_ad_reverse_iter_start
extent_tree_ad_search
extent_tree_szad_first
extent_tree_szad_insert
extent_tree_szad_iter
extent_tree_szad_iter_recurse
extent_tree_szad_iter_start
extent_tree_szad_last
extent_tree_szad_new
extent_tree_szad_next
extent_tree_szad_nsearch
extent_tree_szad_prev
extent_tree_szad_psearch
extent_tree_szad_remove
extent_tree_szad_reverse_iter
extent_tree_szad_reverse_iter_recurse
extent_tree_szad_reverse_iter_start
extent_tree_szad_search
get_errno
hash
hash_fmix_32
hash_fmix_64
hash_get_block_32
hash_get_block_64
hash_rotl_32
hash_rotl_64
hash_x64_128
hash_x86_128
hash_x86_32
huge_allocated
huge_boot
huge_dalloc
huge_dalloc_junk
huge_dss_prec_get
huge_malloc
huge_mtx
huge_ndalloc
huge_nmalloc
huge_palloc
huge_postfork_child
huge_postfork_parent
huge_prefork
huge_prof_ctx_get
huge_prof_ctx_set
huge_ralloc
huge_ralloc_no_move
huge_salloc
iallocm
icalloc
icalloct
idalloc
idalloct
imalloc
imalloct
ipalloc
ipalloct
iqalloc
iqalloct
iralloc
iralloct
iralloct_realign
isalloc
isthreaded
ivsalloc
ixalloc
jemalloc_postfork_child
jemalloc_postfork_parent
jemalloc_prefork
malloc_cprintf
malloc_mutex_init
malloc_mutex_lock
malloc_mutex_postfork_child
malloc_mutex_postfork_parent
malloc_mutex_prefork
malloc_mutex_unlock
malloc_printf
malloc_snprintf
malloc_strtoumax
malloc_tsd_boot
malloc_tsd_cleanup_register
malloc_tsd_dalloc
malloc_tsd_malloc
malloc_tsd_no_cleanup
malloc_vcprintf
malloc_vsnprintf
malloc_write
map_bias
mb_write
mutex_boot
narenas_auto
narenas_total
narenas_total_get
ncpus
nhbins
opt_abort
opt_dss
opt_junk
opt_lg_chunk
opt_lg_dirty_mult
opt_lg_prof_interval
opt_lg_prof_sample
opt_lg_tcache_max
opt_narenas
opt_prof
opt_prof_accum
opt_prof_active
opt_prof_final
opt_prof_gdump
opt_prof_leak
opt_prof_prefix
opt_quarantine
opt_redzone
opt_stats_print
opt_tcache
opt_utrace
opt_valgrind
opt_xmalloc
opt_zero
p2rz
pages_purge
pow2_ceil
prof_backtrace
prof_boot0
prof_boot1
prof_boot2
prof_bt_count
prof_ctx_get
prof_ctx_set
prof_dump_open
prof_free
prof_gdump
prof_idump
prof_interval
prof_lookup
prof_malloc
prof_mdump
prof_postfork_child
prof_postfork_parent
prof_prefork
prof_promote
prof_realloc
prof_sample_accum_update
prof_sample_threshold_update
prof_tdata_booted
prof_tdata_cleanup
prof_tdata_get
prof_tdata_init
prof_tdata_initialized
prof_tdata_tls
prof_tdata_tsd
prof_tdata_tsd_boot
prof_tdata_tsd_cleanup_wrapper
prof_tdata_tsd_get
prof_tdata_tsd_get_wrapper
prof_tdata_tsd_init_head
prof_tdata_tsd_set
quarantine
quarantine_alloc_hook
quarantine_boot
quarantine_booted
quarantine_cleanup
quarantine_init
quarantine_tls
quarantine_tsd
quarantine_tsd_boot
quarantine_tsd_cleanup_wrapper
quarantine_tsd_get
quarantine_tsd_get_wrapper
quarantine_tsd_init_head
quarantine_tsd_set
register_zone
rtree_delete
rtree_get
rtree_get_locked
rtree_new
rtree_postfork_child
rtree_postfork_parent
rtree_prefork
rtree_set
s2u
sa2u
set_errno
small_size2bin
stats_cactive
stats_cactive_add
stats_cactive_get
stats_cactive_sub
stats_chunks
stats_print
tcache_alloc_easy
tcache_alloc_large
tcache_alloc_small
tcache_alloc_small_hard
tcache_arena_associate
tcache_arena_dissociate
tcache_bin_flush_large
tcache_bin_flush_small
tcache_bin_info
tcache_boot0
tcache_boot1
tcache_booted
tcache_create
tcache_dalloc_large
tcache_dalloc_small
tcache_destroy
tcache_enabled_booted
tcache_enabled_get
tcache_enabled_initialized
tcache_enabled_set
tcache_enabled_tls
tcache_enabled_tsd
tcache_enabled_tsd_boot
tcache_enabled_tsd_cleanup_wrapper
tcache_enabled_tsd_get
tcache_enabled_tsd_get_wrapper
tcache_enabled_tsd_init_head
tcache_enabled_tsd_set
tcache_event
tcache_event_hard
tcache_flush
tcache_get
tcache_initialized
tcache_maxclass
tcache_salloc
tcache_stats_merge
tcache_thread_cleanup
tcache_tls
tcache_tsd
tcache_tsd_boot
tcache_tsd_cleanup_wrapper
tcache_tsd_get
tcache_tsd_get_wrapper
tcache_tsd_init_head
tcache_tsd_set
thread_allocated_booted
thread_allocated_initialized
thread_allocated_tls
thread_allocated_tsd
thread_allocated_tsd_boot
thread_allocated_tsd_cleanup_wrapper
thread_allocated_tsd_get
thread_allocated_tsd_get_wrapper
thread_allocated_tsd_init_head
thread_allocated_tsd_set
tsd_init_check_recursion
tsd_init_finish
u2rz
deps/jemalloc/include/jemalloc/internal/private_unnamespace.sh
Vendored Executable
+5
View File
@@ -0,0 +1,5 @@
#!/bin/sh
for symbol in `cat $1` ; do
echo "#undef ${symbol}"
done
deps/jemalloc/include/jemalloc/internal/prng.h
Vendored
+60
View File
@@ -0,0 +1,60 @@
/******************************************************************************/
#ifdef JEMALLOC_H_TYPES
/*
* Simple linear congruential pseudo-random number generator:
*
* prng(y) = (a*x + c) % m
*
* where the following constants ensure maximal period:
*
* a == Odd number (relatively prime to 2^n), and (a-1) is a multiple of 4.
* c == Odd number (relatively prime to 2^n).
* m == 2^32
*
* See Knuth's TAOCP 3rd Ed., Vol. 2, pg. 17 for details on these constraints.
*
* This choice of m has the disadvantage that the quality of the bits is
* proportional to bit position. For example. the lowest bit has a cycle of 2,
* the next has a cycle of 4, etc. For this reason, we prefer to use the upper
* bits.
*
* Macro parameters:
* uint32_t r : Result.
* unsigned lg_range : (0..32], number of least significant bits to return.
* uint32_t state : Seed value.
* const uint32_t a, c : See above discussion.
*/
#define prng32(r, lg_range, state, a, c) do { \
assert(lg_range > 0); \
assert(lg_range <= 32); \
\
r = (state * (a)) + (c); \
state = r; \
r >>= (32 - lg_range); \
} while (false)
/* Same as prng32(), but 64 bits of pseudo-randomness, using uint64_t. */
#define prng64(r, lg_range, state, a, c) do { \
assert(lg_range > 0); \
assert(lg_range <= 64); \
\
r = (state * (a)) + (c); \
state = r; \
r >>= (64 - lg_range); \
} while (false)
#endif /* JEMALLOC_H_TYPES */
/******************************************************************************/
#ifdef JEMALLOC_H_STRUCTS
#endif /* JEMALLOC_H_STRUCTS */
/******************************************************************************/
#ifdef JEMALLOC_H_EXTERNS
#endif /* JEMALLOC_H_EXTERNS */
/******************************************************************************/
#ifdef JEMALLOC_H_INLINES
#endif /* JEMALLOC_H_INLINES */
/******************************************************************************/
deps/jemalloc/include/jemalloc/internal/prof.h
Vendored
+613
View File
@@ -0,0 +1,613 @@
/******************************************************************************/
#ifdef JEMALLOC_H_TYPES
typedef struct prof_bt_s prof_bt_t;
typedef struct prof_cnt_s prof_cnt_t;
typedef struct prof_thr_cnt_s prof_thr_cnt_t;
typedef struct prof_ctx_s prof_ctx_t;
typedef struct prof_tdata_s prof_tdata_t;
/* Option defaults. */
#ifdef JEMALLOC_PROF
# define PROF_PREFIX_DEFAULT "jeprof"
#else
# define PROF_PREFIX_DEFAULT ""
#endif
#define LG_PROF_SAMPLE_DEFAULT 19
#define LG_PROF_INTERVAL_DEFAULT -1
/*
* Hard limit on stack backtrace depth. The version of prof_backtrace() that
* is based on __builtin_return_address() necessarily has a hard-coded number
* of backtrace frame handlers, and should be kept in sync with this setting.
*/
#define PROF_BT_MAX 128
/* Maximum number of backtraces to store in each per thread LRU cache. */
#define PROF_TCMAX 1024
/* Initial hash table size. */
#define PROF_CKH_MINITEMS 64
/* Size of memory buffer to use when writing dump files. */
#define PROF_DUMP_BUFSIZE 65536
/* Size of stack-allocated buffer used by prof_printf(). */
#define PROF_PRINTF_BUFSIZE 128
/*
* Number of mutexes shared among all ctx's. No space is allocated for these
* unless profiling is enabled, so it's okay to over-provision.
*/
#define PROF_NCTX_LOCKS 1024
/*
* prof_tdata pointers close to NULL are used to encode state information that
* is used for cleaning up during thread shutdown.
*/
#define PROF_TDATA_STATE_REINCARNATED ((prof_tdata_t *)(uintptr_t)1)
#define PROF_TDATA_STATE_PURGATORY ((prof_tdata_t *)(uintptr_t)2)
#define PROF_TDATA_STATE_MAX PROF_TDATA_STATE_PURGATORY
#endif /* JEMALLOC_H_TYPES */
/******************************************************************************/
#ifdef JEMALLOC_H_STRUCTS
struct prof_bt_s {
/* Backtrace, stored as len program counters. */
void **vec;
unsigned len;
};
#ifdef JEMALLOC_PROF_LIBGCC
/* Data structure passed to libgcc _Unwind_Backtrace() callback functions. */
typedef struct {
prof_bt_t *bt;
unsigned nignore;
unsigned max;
} prof_unwind_data_t;
#endif
struct prof_cnt_s {
/*
* Profiling counters. An allocation/deallocation pair can operate on
* different prof_thr_cnt_t objects that are linked into the same
* prof_ctx_t cnts_ql, so it is possible for the cur* counters to go
* negative. In principle it is possible for the *bytes counters to
* overflow/underflow, but a general solution would require something
* like 128-bit counters; this implementation doesn't bother to solve
* that problem.
*/
int64_t curobjs;
int64_t curbytes;
uint64_t accumobjs;
uint64_t accumbytes;
};
struct prof_thr_cnt_s {
/* Linkage into prof_ctx_t's cnts_ql. */
ql_elm(prof_thr_cnt_t) cnts_link;
/* Linkage into thread's LRU. */
ql_elm(prof_thr_cnt_t) lru_link;
/*
* Associated context. If a thread frees an object that it did not
* allocate, it is possible that the context is not cached in the
* thread's hash table, in which case it must be able to look up the
* context, insert a new prof_thr_cnt_t into the thread's hash table,
* and link it into the prof_ctx_t's cnts_ql.
*/
prof_ctx_t *ctx;
/*
* Threads use memory barriers to update the counters. Since there is
* only ever one writer, the only challenge is for the reader to get a
* consistent read of the counters.
*
* The writer uses this series of operations:
*
* 1) Increment epoch to an odd number.
* 2) Update counters.
* 3) Increment epoch to an even number.
*
* The reader must assure 1) that the epoch is even while it reads the
* counters, and 2) that the epoch doesn't change between the time it
* starts and finishes reading the counters.
*/
unsigned epoch;
/* Profiling counters. */
prof_cnt_t cnts;
};
struct prof_ctx_s {
/* Associated backtrace. */
prof_bt_t *bt;
/* Protects nlimbo, cnt_merged, and cnts_ql. */
malloc_mutex_t *lock;
/*
* Number of threads that currently cause this ctx to be in a state of
* limbo due to one of:
* - Initializing per thread counters associated with this ctx.
* - Preparing to destroy this ctx.
* - Dumping a heap profile that includes this ctx.
* nlimbo must be 1 (single destroyer) in order to safely destroy the
* ctx.
*/
unsigned nlimbo;
/* Temporary storage for summation during dump. */
prof_cnt_t cnt_summed;
/* When threads exit, they merge their stats into cnt_merged. */
prof_cnt_t cnt_merged;
/*
* List of profile counters, one for each thread that has allocated in
* this context.
*/
ql_head(prof_thr_cnt_t) cnts_ql;
/* Linkage for list of contexts to be dumped. */
ql_elm(prof_ctx_t) dump_link;
};
typedef ql_head(prof_ctx_t) prof_ctx_list_t;
struct prof_tdata_s {
/*
* Hash of (prof_bt_t *)-->(prof_thr_cnt_t *). Each thread keeps a
* cache of backtraces, with associated thread-specific prof_thr_cnt_t
* objects. Other threads may read the prof_thr_cnt_t contents, but no
* others will ever write them.
*
* Upon thread exit, the thread must merge all the prof_thr_cnt_t
* counter data into the associated prof_ctx_t objects, and unlink/free
* the prof_thr_cnt_t objects.
*/
ckh_t bt2cnt;
/* LRU for contents of bt2cnt. */
ql_head(prof_thr_cnt_t) lru_ql;
/* Backtrace vector, used for calls to prof_backtrace(). */
void **vec;
/* Sampling state. */
uint64_t prng_state;
uint64_t threshold;
uint64_t accum;
/* State used to avoid dumping while operating on prof internals. */
bool enq;
bool enq_idump;
bool enq_gdump;
};
#endif /* JEMALLOC_H_STRUCTS */
/******************************************************************************/
#ifdef JEMALLOC_H_EXTERNS
extern bool opt_prof;
/*
* Even if opt_prof is true, sampling can be temporarily disabled by setting
* opt_prof_active to false. No locking is used when updating opt_prof_active,
* so there are no guarantees regarding how long it will take for all threads
* to notice state changes.
*/
extern bool opt_prof_active;
extern size_t opt_lg_prof_sample; /* Mean bytes between samples. */
extern ssize_t opt_lg_prof_interval; /* lg(prof_interval). */
extern bool opt_prof_gdump; /* High-water memory dumping. */
extern bool opt_prof_final; /* Final profile dumping. */
extern bool opt_prof_leak; /* Dump leak summary at exit. */
extern bool opt_prof_accum; /* Report cumulative bytes. */
extern char opt_prof_prefix[
/* Minimize memory bloat for non-prof builds. */
#ifdef JEMALLOC_PROF
PATH_MAX +
#endif
1];
/*
* Profile dump interval, measured in bytes allocated. Each arena triggers a
* profile dump when it reaches this threshold. The effect is that the
* interval between profile dumps averages prof_interval, though the actual
* interval between dumps will tend to be sporadic, and the interval will be a
* maximum of approximately (prof_interval * narenas).
*/
extern uint64_t prof_interval;
/*
* If true, promote small sampled objects to large objects, since small run
* headers do not have embedded profile context pointers.
*/
extern bool prof_promote;
void bt_init(prof_bt_t *bt, void **vec);
void prof_backtrace(prof_bt_t *bt, unsigned nignore);
prof_thr_cnt_t *prof_lookup(prof_bt_t *bt);
#ifdef JEMALLOC_JET
size_t prof_bt_count(void);
typedef int (prof_dump_open_t)(bool, const char *);
extern prof_dump_open_t *prof_dump_open;
#endif
void prof_idump(void);
bool prof_mdump(const char *filename);
void prof_gdump(void);
prof_tdata_t *prof_tdata_init(void);
void prof_tdata_cleanup(void *arg);
void prof_boot0(void);
void prof_boot1(void);
bool prof_boot2(void);
void prof_prefork(void);
void prof_postfork_parent(void);
void prof_postfork_child(void);
#endif /* JEMALLOC_H_EXTERNS */
/******************************************************************************/
#ifdef JEMALLOC_H_INLINES
#define PROF_ALLOC_PREP(nignore, size, ret) do { \
prof_tdata_t *prof_tdata; \
prof_bt_t bt; \
\
assert(size == s2u(size)); \
\
prof_tdata = prof_tdata_get(true); \
if ((uintptr_t)prof_tdata <= (uintptr_t)PROF_TDATA_STATE_MAX) { \
if (prof_tdata != NULL) \
ret = (prof_thr_cnt_t *)(uintptr_t)1U; \
else \
ret = NULL; \
break; \
} \
\
if (opt_prof_active == false) { \
/* Sampling is currently inactive, so avoid sampling. */\
ret = (prof_thr_cnt_t *)(uintptr_t)1U; \
} else if (opt_lg_prof_sample == 0) { \
/* Don't bother with sampling logic, since sampling */\
/* interval is 1. */\
bt_init(&bt, prof_tdata->vec); \
prof_backtrace(&bt, nignore); \
ret = prof_lookup(&bt); \
} else { \
if (prof_tdata->threshold == 0) { \
/* Initialize. Seed the prng differently for */\
/* each thread. */\
prof_tdata->prng_state = \
(uint64_t)(uintptr_t)&size; \
prof_sample_threshold_update(prof_tdata); \
} \
\
/* Determine whether to capture a backtrace based on */\
/* whether size is enough for prof_accum to reach */\
/* prof_tdata->threshold. However, delay updating */\
/* these variables until prof_{m,re}alloc(), because */\
/* we don't know for sure that the allocation will */\
/* succeed. */\
/* */\
/* Use subtraction rather than addition to avoid */\
/* potential integer overflow. */\
if (size >= prof_tdata->threshold - \
prof_tdata->accum) { \
bt_init(&bt, prof_tdata->vec); \
prof_backtrace(&bt, nignore); \
ret = prof_lookup(&bt); \
} else \
ret = (prof_thr_cnt_t *)(uintptr_t)1U; \
} \
} while (0)
#ifndef JEMALLOC_ENABLE_INLINE
malloc_tsd_protos(JEMALLOC_ATTR(unused), prof_tdata, prof_tdata_t *)
prof_tdata_t *prof_tdata_get(bool create);
void prof_sample_threshold_update(prof_tdata_t *prof_tdata);
prof_ctx_t *prof_ctx_get(const void *ptr);
void prof_ctx_set(const void *ptr, size_t usize, prof_ctx_t *ctx);
bool prof_sample_accum_update(size_t size);
void prof_malloc(const void *ptr, size_t usize, prof_thr_cnt_t *cnt);
void prof_realloc(const void *ptr, size_t usize, prof_thr_cnt_t *cnt,
size_t old_usize, prof_ctx_t *old_ctx);
void prof_free(const void *ptr, size_t size);
#endif
#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_PROF_C_))
/* Thread-specific backtrace cache, used to reduce bt2ctx contention. */
malloc_tsd_externs(prof_tdata, prof_tdata_t *)
malloc_tsd_funcs(JEMALLOC_INLINE, prof_tdata, prof_tdata_t *, NULL,
prof_tdata_cleanup)
JEMALLOC_INLINE prof_tdata_t *
prof_tdata_get(bool create)
{
prof_tdata_t *prof_tdata;
cassert(config_prof);
prof_tdata = *prof_tdata_tsd_get();
if (create && prof_tdata == NULL)
prof_tdata = prof_tdata_init();
return (prof_tdata);
}
JEMALLOC_INLINE void
prof_sample_threshold_update(prof_tdata_t *prof_tdata)
{
/*
* The body of this function is compiled out unless heap profiling is
* enabled, so that it is possible to compile jemalloc with floating
* point support completely disabled. Avoiding floating point code is
* important on memory-constrained systems, but it also enables a
* workaround for versions of glibc that don't properly save/restore
* floating point registers during dynamic lazy symbol loading (which
* internally calls into whatever malloc implementation happens to be
* integrated into the application). Note that some compilers (e.g.
* gcc 4.8) may use floating point registers for fast memory moves, so
* jemalloc must be compiled with such optimizations disabled (e.g.
* -mno-sse) in order for the workaround to be complete.
*/
#ifdef JEMALLOC_PROF
uint64_t r;
double u;
cassert(config_prof);
/*
* Compute sample threshold as a geometrically distributed random
* variable with mean (2^opt_lg_prof_sample).
*
* __ __
* | log(u) | 1
* prof_tdata->threshold = | -------- |, where p = -------------------
* | log(1-p) | opt_lg_prof_sample
* 2
*
* For more information on the math, see:
*
* Non-Uniform Random Variate Generation
* Luc Devroye
* Springer-Verlag, New York, 1986
* pp 500
* (http://luc.devroye.org/rnbookindex.html)
*/
prng64(r, 53, prof_tdata->prng_state,
UINT64_C(6364136223846793005), UINT64_C(1442695040888963407));
u = (double)r * (1.0/9007199254740992.0L);
prof_tdata->threshold = (uint64_t)(log(u) /
log(1.0 - (1.0 / (double)((uint64_t)1U << opt_lg_prof_sample))))
+ (uint64_t)1U;
#endif
}
JEMALLOC_INLINE prof_ctx_t *
prof_ctx_get(const void *ptr)
{
prof_ctx_t *ret;
arena_chunk_t *chunk;
cassert(config_prof);
assert(ptr != NULL);
chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ptr);
if (chunk != ptr) {
/* Region. */
ret = arena_prof_ctx_get(ptr);
} else
ret = huge_prof_ctx_get(ptr);
return (ret);
}
JEMALLOC_INLINE void
prof_ctx_set(const void *ptr, size_t usize, prof_ctx_t *ctx)
{
arena_chunk_t *chunk;
cassert(config_prof);
assert(ptr != NULL);
chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ptr);
if (chunk != ptr) {
/* Region. */
arena_prof_ctx_set(ptr, usize, ctx);
} else
huge_prof_ctx_set(ptr, ctx);
}
JEMALLOC_INLINE bool
prof_sample_accum_update(size_t size)
{
prof_tdata_t *prof_tdata;
cassert(config_prof);
/* Sampling logic is unnecessary if the interval is 1. */
assert(opt_lg_prof_sample != 0);
prof_tdata = prof_tdata_get(false);
if ((uintptr_t)prof_tdata <= (uintptr_t)PROF_TDATA_STATE_MAX)
return (true);
/* Take care to avoid integer overflow. */
if (size >= prof_tdata->threshold - prof_tdata->accum) {
prof_tdata->accum -= (prof_tdata->threshold - size);
/* Compute new sample threshold. */
prof_sample_threshold_update(prof_tdata);
while (prof_tdata->accum >= prof_tdata->threshold) {
prof_tdata->accum -= prof_tdata->threshold;
prof_sample_threshold_update(prof_tdata);
}
return (false);
} else {
prof_tdata->accum += size;
return (true);
}
}
JEMALLOC_INLINE void
prof_malloc(const void *ptr, size_t usize, prof_thr_cnt_t *cnt)
{
cassert(config_prof);
assert(ptr != NULL);
assert(usize == isalloc(ptr, true));
if (opt_lg_prof_sample != 0) {
if (prof_sample_accum_update(usize)) {
/*
* Don't sample. For malloc()-like allocation, it is
* always possible to tell in advance how large an
* object's usable size will be, so there should never
* be a difference between the usize passed to
* PROF_ALLOC_PREP() and prof_malloc().
*/
assert((uintptr_t)cnt == (uintptr_t)1U);
}
}
if ((uintptr_t)cnt > (uintptr_t)1U) {
prof_ctx_set(ptr, usize, cnt->ctx);
cnt->epoch++;
/*********/
mb_write();
/*********/
cnt->cnts.curobjs++;
cnt->cnts.curbytes += usize;
if (opt_prof_accum) {
cnt->cnts.accumobjs++;
cnt->cnts.accumbytes += usize;
}
/*********/
mb_write();
/*********/
cnt->epoch++;
/*********/
mb_write();
/*********/
} else
prof_ctx_set(ptr, usize, (prof_ctx_t *)(uintptr_t)1U);
}
JEMALLOC_INLINE void
prof_realloc(const void *ptr, size_t usize, prof_thr_cnt_t *cnt,
size_t old_usize, prof_ctx_t *old_ctx)
{
prof_thr_cnt_t *told_cnt;
cassert(config_prof);
assert(ptr != NULL || (uintptr_t)cnt <= (uintptr_t)1U);
if (ptr != NULL) {
assert(usize == isalloc(ptr, true));
if (opt_lg_prof_sample != 0) {
if (prof_sample_accum_update(usize)) {
/*
* Don't sample. The usize passed to
* PROF_ALLOC_PREP() was larger than what
* actually got allocated, so a backtrace was
* captured for this allocation, even though
* its actual usize was insufficient to cross
* the sample threshold.
*/
cnt = (prof_thr_cnt_t *)(uintptr_t)1U;
}
}
}
if ((uintptr_t)old_ctx > (uintptr_t)1U) {
told_cnt = prof_lookup(old_ctx->bt);
if (told_cnt == NULL) {
/*
* It's too late to propagate OOM for this realloc(),
* so operate directly on old_cnt->ctx->cnt_merged.
*/
malloc_mutex_lock(old_ctx->lock);
old_ctx->cnt_merged.curobjs--;
old_ctx->cnt_merged.curbytes -= old_usize;
malloc_mutex_unlock(old_ctx->lock);
told_cnt = (prof_thr_cnt_t *)(uintptr_t)1U;
}
} else
told_cnt = (prof_thr_cnt_t *)(uintptr_t)1U;
if ((uintptr_t)told_cnt > (uintptr_t)1U)
told_cnt->epoch++;
if ((uintptr_t)cnt > (uintptr_t)1U) {
prof_ctx_set(ptr, usize, cnt->ctx);
cnt->epoch++;
} else if (ptr != NULL)
prof_ctx_set(ptr, usize, (prof_ctx_t *)(uintptr_t)1U);
/*********/
mb_write();
/*********/
if ((uintptr_t)told_cnt > (uintptr_t)1U) {
told_cnt->cnts.curobjs--;
told_cnt->cnts.curbytes -= old_usize;
}
if ((uintptr_t)cnt > (uintptr_t)1U) {
cnt->cnts.curobjs++;
cnt->cnts.curbytes += usize;
if (opt_prof_accum) {
cnt->cnts.accumobjs++;
cnt->cnts.accumbytes += usize;
}
}
/*********/
mb_write();
/*********/
if ((uintptr_t)told_cnt > (uintptr_t)1U)
told_cnt->epoch++;
if ((uintptr_t)cnt > (uintptr_t)1U)
cnt->epoch++;
/*********/
mb_write(); /* Not strictly necessary. */
}
JEMALLOC_INLINE void
prof_free(const void *ptr, size_t size)
{
prof_ctx_t *ctx = prof_ctx_get(ptr);
cassert(config_prof);
if ((uintptr_t)ctx > (uintptr_t)1) {
prof_thr_cnt_t *tcnt;
assert(size == isalloc(ptr, true));
tcnt = prof_lookup(ctx->bt);
if (tcnt != NULL) {
tcnt->epoch++;
/*********/
mb_write();
/*********/
tcnt->cnts.curobjs--;
tcnt->cnts.curbytes -= size;
/*********/
mb_write();
/*********/
tcnt->epoch++;
/*********/
mb_write();
/*********/
} else {
/*
* OOM during free() cannot be propagated, so operate
* directly on cnt->ctx->cnt_merged.
*/
malloc_mutex_lock(ctx->lock);
ctx->cnt_merged.curobjs--;
ctx->cnt_merged.curbytes -= size;
malloc_mutex_unlock(ctx->lock);
}
}
}
#endif
#endif /* JEMALLOC_H_INLINES */
/******************************************************************************/
deps/jemalloc/include/jemalloc/internal/public_namespace.sh
Vendored Executable
+6
View File
@@ -0,0 +1,6 @@
#!/bin/sh
for nm in `cat $1` ; do
n=`echo ${nm} |tr ':' ' ' |awk '{print $1}'`
echo "#define je_${n} JEMALLOC_N(${n})"
done
deps/jemalloc/include/jemalloc/internal/public_unnamespace.sh
Vendored Executable
+6
View File
@@ -0,0 +1,6 @@
#!/bin/sh
for nm in `cat $1` ; do
n=`echo ${nm} |tr ':' ' ' |awk '{print $1}'`
echo "#undef je_${n}"
done
deps/jemalloc/include/jemalloc/internal/ql.h
Vendored
+83
View File
@@ -0,0 +1,83 @@
/*
* List definitions.
*/
#define ql_head(a_type) \
struct { \
a_type *qlh_first; \
}
#define ql_head_initializer(a_head) {NULL}
#define ql_elm(a_type) qr(a_type)
/* List functions. */
#define ql_new(a_head) do { \
(a_head)->qlh_first = NULL; \
} while (0)
#define ql_elm_new(a_elm, a_field) qr_new((a_elm), a_field)
#define ql_first(a_head) ((a_head)->qlh_first)
#define ql_last(a_head, a_field) \
((ql_first(a_head) != NULL) \
? qr_prev(ql_first(a_head), a_field) : NULL)
#define ql_next(a_head, a_elm, a_field) \
((ql_last(a_head, a_field) != (a_elm)) \
? qr_next((a_elm), a_field) : NULL)
#define ql_prev(a_head, a_elm, a_field) \
((ql_first(a_head) != (a_elm)) ? qr_prev((a_elm), a_field) \
: NULL)
#define ql_before_insert(a_head, a_qlelm, a_elm, a_field) do { \
qr_before_insert((a_qlelm), (a_elm), a_field); \
if (ql_first(a_head) == (a_qlelm)) { \
ql_first(a_head) = (a_elm); \
} \
} while (0)
#define ql_after_insert(a_qlelm, a_elm, a_field) \
qr_after_insert((a_qlelm), (a_elm), a_field)
#define ql_head_insert(a_head, a_elm, a_field) do { \
if (ql_first(a_head) != NULL) { \
qr_before_insert(ql_first(a_head), (a_elm), a_field); \
} \
ql_first(a_head) = (a_elm); \
} while (0)
#define ql_tail_insert(a_head, a_elm, a_field) do { \
if (ql_first(a_head) != NULL) { \
qr_before_insert(ql_first(a_head), (a_elm), a_field); \
} \
ql_first(a_head) = qr_next((a_elm), a_field); \
} while (0)
#define ql_remove(a_head, a_elm, a_field) do { \
if (ql_first(a_head) == (a_elm)) { \
ql_first(a_head) = qr_next(ql_first(a_head), a_field); \
} \
if (ql_first(a_head) != (a_elm)) { \
qr_remove((a_elm), a_field); \
} else { \
ql_first(a_head) = NULL; \
} \
} while (0)
#define ql_head_remove(a_head, a_type, a_field) do { \
a_type *t = ql_first(a_head); \
ql_remove((a_head), t, a_field); \
} while (0)
#define ql_tail_remove(a_head, a_type, a_field) do { \
a_type *t = ql_last(a_head, a_field); \
ql_remove((a_head), t, a_field); \
} while (0)
#define ql_foreach(a_var, a_head, a_field) \
qr_foreach((a_var), ql_first(a_head), a_field)
#define ql_reverse_foreach(a_var, a_head, a_field) \
qr_reverse_foreach((a_var), ql_first(a_head), a_field)
deps/jemalloc/include/jemalloc/internal/qr.h
Vendored
+67
View File
@@ -0,0 +1,67 @@
/* Ring definitions. */
#define qr(a_type) \
struct { \
a_type *qre_next; \
a_type *qre_prev; \
}
/* Ring functions. */
#define qr_new(a_qr, a_field) do { \
(a_qr)->a_field.qre_next = (a_qr); \
(a_qr)->a_field.qre_prev = (a_qr); \
} while (0)
#define qr_next(a_qr, a_field) ((a_qr)->a_field.qre_next)
#define qr_prev(a_qr, a_field) ((a_qr)->a_field.qre_prev)
#define qr_before_insert(a_qrelm, a_qr, a_field) do { \
(a_qr)->a_field.qre_prev = (a_qrelm)->a_field.qre_prev; \
(a_qr)->a_field.qre_next = (a_qrelm); \
(a_qr)->a_field.qre_prev->a_field.qre_next = (a_qr); \
(a_qrelm)->a_field.qre_prev = (a_qr); \
} while (0)
#define qr_after_insert(a_qrelm, a_qr, a_field) \
do \
{ \
(a_qr)->a_field.qre_next = (a_qrelm)->a_field.qre_next; \
(a_qr)->a_field.qre_prev = (a_qrelm); \
(a_qr)->a_field.qre_next->a_field.qre_prev = (a_qr); \
(a_qrelm)->a_field.qre_next = (a_qr); \
} while (0)
#define qr_meld(a_qr_a, a_qr_b, a_field) do { \
void *t; \
(a_qr_a)->a_field.qre_prev->a_field.qre_next = (a_qr_b); \
(a_qr_b)->a_field.qre_prev->a_field.qre_next = (a_qr_a); \
t = (a_qr_a)->a_field.qre_prev; \
(a_qr_a)->a_field.qre_prev = (a_qr_b)->a_field.qre_prev; \
(a_qr_b)->a_field.qre_prev = t; \
} while (0)
/* qr_meld() and qr_split() are functionally equivalent, so there's no need to
* have two copies of the code. */
#define qr_split(a_qr_a, a_qr_b, a_field) \
qr_meld((a_qr_a), (a_qr_b), a_field)
#define qr_remove(a_qr, a_field) do { \
(a_qr)->a_field.qre_prev->a_field.qre_next \
= (a_qr)->a_field.qre_next; \
(a_qr)->a_field.qre_next->a_field.qre_prev \
= (a_qr)->a_field.qre_prev; \
(a_qr)->a_field.qre_next = (a_qr); \
(a_qr)->a_field.qre_prev = (a_qr); \
} while (0)
#define qr_foreach(var, a_qr, a_field) \
for ((var) = (a_qr); \
(var) != NULL; \
(var) = (((var)->a_field.qre_next != (a_qr)) \
? (var)->a_field.qre_next : NULL))
#define qr_reverse_foreach(var, a_qr, a_field) \
for ((var) = ((a_qr) != NULL) ? qr_prev(a_qr, a_field) : NULL; \
(var) != NULL; \
(var) = (((var) != (a_qr)) \
? (var)->a_field.qre_prev : NULL))
deps/jemalloc/include/jemalloc/internal/quarantine.h
Vendored
+67
View File
@@ -0,0 +1,67 @@
/******************************************************************************/
#ifdef JEMALLOC_H_TYPES
typedef struct quarantine_obj_s quarantine_obj_t;
typedef struct quarantine_s quarantine_t;
/* Default per thread quarantine size if valgrind is enabled. */
#define JEMALLOC_VALGRIND_QUARANTINE_DEFAULT (ZU(1) << 24)
#endif /* JEMALLOC_H_TYPES */
/******************************************************************************/
#ifdef JEMALLOC_H_STRUCTS
struct quarantine_obj_s {
void *ptr;
size_t usize;
};
struct quarantine_s {
size_t curbytes;
size_t curobjs;
size_t first;
#define LG_MAXOBJS_INIT 10
size_t lg_maxobjs;
quarantine_obj_t objs[1]; /* Dynamically sized ring buffer. */
};
#endif /* JEMALLOC_H_STRUCTS */
/******************************************************************************/
#ifdef JEMALLOC_H_EXTERNS
quarantine_t *quarantine_init(size_t lg_maxobjs);
void quarantine(void *ptr);
void quarantine_cleanup(void *arg);
bool quarantine_boot(void);
#endif /* JEMALLOC_H_EXTERNS */
/******************************************************************************/
#ifdef JEMALLOC_H_INLINES
#ifndef JEMALLOC_ENABLE_INLINE
malloc_tsd_protos(JEMALLOC_ATTR(unused), quarantine, quarantine_t *)
void quarantine_alloc_hook(void);
#endif
#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_QUARANTINE_C_))
malloc_tsd_externs(quarantine, quarantine_t *)
malloc_tsd_funcs(JEMALLOC_ALWAYS_INLINE, quarantine, quarantine_t *, NULL,
quarantine_cleanup)
JEMALLOC_ALWAYS_INLINE void
quarantine_alloc_hook(void)
{
quarantine_t *quarantine;
assert(config_fill && opt_quarantine);
quarantine = *quarantine_tsd_get();
if (quarantine == NULL)
quarantine_init(LG_MAXOBJS_INIT);
}
#endif
#endif /* JEMALLOC_H_INLINES */
/******************************************************************************/
deps/jemalloc/include/jemalloc/internal/rb.h
Vendored
+969
View File
@@ -0,0 +1,969 @@
/*-
*******************************************************************************
*
* cpp macro implementation of left-leaning 2-3 red-black trees. Parent
* pointers are not used, and color bits are stored in the least significant
* bit of right-child pointers (if RB_COMPACT is defined), thus making node
* linkage as compact as is possible for red-black trees.
*
* Usage:
*
* #include <stdint.h>
* #include <stdbool.h>
* #define NDEBUG // (Optional, see assert(3).)
* #include <assert.h>
* #define RB_COMPACT // (Optional, embed color bits in right-child pointers.)
* #include <rb.h>
* ...
*
*******************************************************************************
*/
#ifndef RB_H_
#define RB_H_
#ifdef RB_COMPACT
/* Node structure. */
#define rb_node(a_type) \
struct { \
a_type *rbn_left; \
a_type *rbn_right_red; \
}
#else
#define rb_node(a_type) \
struct { \
a_type *rbn_left; \
a_type *rbn_right; \
bool rbn_red; \
}
#endif
/* Root structure. */
#define rb_tree(a_type) \
struct { \
a_type *rbt_root; \
a_type rbt_nil; \
}
/* Left accessors. */
#define rbtn_left_get(a_type, a_field, a_node) \
((a_node)->a_field.rbn_left)
#define rbtn_left_set(a_type, a_field, a_node, a_left) do { \
(a_node)->a_field.rbn_left = a_left; \
} while (0)
#ifdef RB_COMPACT
/* Right accessors. */
#define rbtn_right_get(a_type, a_field, a_node) \
((a_type *) (((intptr_t) (a_node)->a_field.rbn_right_red) \
& ((ssize_t)-2)))
#define rbtn_right_set(a_type, a_field, a_node, a_right) do { \
(a_node)->a_field.rbn_right_red = (a_type *) (((uintptr_t) a_right) \
| (((uintptr_t) (a_node)->a_field.rbn_right_red) & ((size_t)1))); \
} while (0)
/* Color accessors. */
#define rbtn_red_get(a_type, a_field, a_node) \
((bool) (((uintptr_t) (a_node)->a_field.rbn_right_red) \
& ((size_t)1)))
#define rbtn_color_set(a_type, a_field, a_node, a_red) do { \
(a_node)->a_field.rbn_right_red = (a_type *) ((((intptr_t) \
(a_node)->a_field.rbn_right_red) & ((ssize_t)-2)) \
| ((ssize_t)a_red)); \
} while (0)
#define rbtn_red_set(a_type, a_field, a_node) do { \
(a_node)->a_field.rbn_right_red = (a_type *) (((uintptr_t) \
(a_node)->a_field.rbn_right_red) | ((size_t)1)); \
} while (0)
#define rbtn_black_set(a_type, a_field, a_node) do { \
(a_node)->a_field.rbn_right_red = (a_type *) (((intptr_t) \
(a_node)->a_field.rbn_right_red) & ((ssize_t)-2)); \
} while (0)
#else
/* Right accessors. */
#define rbtn_right_get(a_type, a_field, a_node) \
((a_node)->a_field.rbn_right)
#define rbtn_right_set(a_type, a_field, a_node, a_right) do { \
(a_node)->a_field.rbn_right = a_right; \
} while (0)
/* Color accessors. */
#define rbtn_red_get(a_type, a_field, a_node) \
((a_node)->a_field.rbn_red)
#define rbtn_color_set(a_type, a_field, a_node, a_red) do { \
(a_node)->a_field.rbn_red = (a_red); \
} while (0)
#define rbtn_red_set(a_type, a_field, a_node) do { \
(a_node)->a_field.rbn_red = true; \
} while (0)
#define rbtn_black_set(a_type, a_field, a_node) do { \
(a_node)->a_field.rbn_red = false; \
} while (0)
#endif
/* Node initializer. */
#define rbt_node_new(a_type, a_field, a_rbt, a_node) do { \
rbtn_left_set(a_type, a_field, (a_node), &(a_rbt)->rbt_nil); \
rbtn_right_set(a_type, a_field, (a_node), &(a_rbt)->rbt_nil); \
rbtn_red_set(a_type, a_field, (a_node)); \
} while (0)
/* Tree initializer. */
#define rb_new(a_type, a_field, a_rbt) do { \
(a_rbt)->rbt_root = &(a_rbt)->rbt_nil; \
rbt_node_new(a_type, a_field, a_rbt, &(a_rbt)->rbt_nil); \
rbtn_black_set(a_type, a_field, &(a_rbt)->rbt_nil); \
} while (0)
/* Internal utility macros. */
#define rbtn_first(a_type, a_field, a_rbt, a_root, r_node) do { \
(r_node) = (a_root); \
if ((r_node) != &(a_rbt)->rbt_nil) { \
for (; \
rbtn_left_get(a_type, a_field, (r_node)) != &(a_rbt)->rbt_nil;\
(r_node) = rbtn_left_get(a_type, a_field, (r_node))) { \
} \
} \
} while (0)
#define rbtn_last(a_type, a_field, a_rbt, a_root, r_node) do { \
(r_node) = (a_root); \
if ((r_node) != &(a_rbt)->rbt_nil) { \
for (; rbtn_right_get(a_type, a_field, (r_node)) != \
&(a_rbt)->rbt_nil; (r_node) = rbtn_right_get(a_type, a_field, \
(r_node))) { \
} \
} \
} while (0)
#define rbtn_rotate_left(a_type, a_field, a_node, r_node) do { \
(r_node) = rbtn_right_get(a_type, a_field, (a_node)); \
rbtn_right_set(a_type, a_field, (a_node), \
rbtn_left_get(a_type, a_field, (r_node))); \
rbtn_left_set(a_type, a_field, (r_node), (a_node)); \
} while (0)
#define rbtn_rotate_right(a_type, a_field, a_node, r_node) do { \
(r_node) = rbtn_left_get(a_type, a_field, (a_node)); \
rbtn_left_set(a_type, a_field, (a_node), \
rbtn_right_get(a_type, a_field, (r_node))); \
rbtn_right_set(a_type, a_field, (r_node), (a_node)); \
} while (0)
/*
* The rb_proto() macro generates function prototypes that correspond to the
* functions generated by an equivalently parameterized call to rb_gen().
*/
#define rb_proto(a_attr, a_prefix, a_rbt_type, a_type) \
a_attr void \
a_prefix##new(a_rbt_type *rbtree); \
a_attr a_type * \
a_prefix##first(a_rbt_type *rbtree); \
a_attr a_type * \
a_prefix##last(a_rbt_type *rbtree); \
a_attr a_type * \
a_prefix##next(a_rbt_type *rbtree, a_type *node); \
a_attr a_type * \
a_prefix##prev(a_rbt_type *rbtree, a_type *node); \
a_attr a_type * \
a_prefix##search(a_rbt_type *rbtree, a_type *key); \
a_attr a_type * \
a_prefix##nsearch(a_rbt_type *rbtree, a_type *key); \
a_attr a_type * \
a_prefix##psearch(a_rbt_type *rbtree, a_type *key); \
a_attr void \
a_prefix##insert(a_rbt_type *rbtree, a_type *node); \
a_attr void \
a_prefix##remove(a_rbt_type *rbtree, a_type *node); \
a_attr a_type * \
a_prefix##iter(a_rbt_type *rbtree, a_type *start, a_type *(*cb)( \
a_rbt_type *, a_type *, void *), void *arg); \
a_attr a_type * \
a_prefix##reverse_iter(a_rbt_type *rbtree, a_type *start, \
a_type *(*cb)(a_rbt_type *, a_type *, void *), void *arg);
/*
* The rb_gen() macro generates a type-specific red-black tree implementation,
* based on the above cpp macros.
*
* Arguments:
*
* a_attr : Function attribute for generated functions (ex: static).
* a_prefix : Prefix for generated functions (ex: ex_).
* a_rb_type : Type for red-black tree data structure (ex: ex_t).
* a_type : Type for red-black tree node data structure (ex: ex_node_t).
* a_field : Name of red-black tree node linkage (ex: ex_link).
* a_cmp : Node comparison function name, with the following prototype:
* int (a_cmp *)(a_type *a_node, a_type *a_other);
* ^^^^^^
* or a_key
* Interpretation of comparision function return values:
* -1 : a_node < a_other
* 0 : a_node == a_other
* 1 : a_node > a_other
* In all cases, the a_node or a_key macro argument is the first
* argument to the comparison function, which makes it possible
* to write comparison functions that treat the first argument
* specially.
*
* Assuming the following setup:
*
* typedef struct ex_node_s ex_node_t;
* struct ex_node_s {
* rb_node(ex_node_t) ex_link;
* };
* typedef rb_tree(ex_node_t) ex_t;
* rb_gen(static, ex_, ex_t, ex_node_t, ex_link, ex_cmp)
*
* The following API is generated:
*
* static void
* ex_new(ex_t *tree);
* Description: Initialize a red-black tree structure.
* Args:
* tree: Pointer to an uninitialized red-black tree object.
*
* static ex_node_t *
* ex_first(ex_t *tree);
* static ex_node_t *
* ex_last(ex_t *tree);
* Description: Get the first/last node in tree.
* Args:
* tree: Pointer to an initialized red-black tree object.
* Ret: First/last node in tree, or NULL if tree is empty.
*
* static ex_node_t *
* ex_next(ex_t *tree, ex_node_t *node);
* static ex_node_t *
* ex_prev(ex_t *tree, ex_node_t *node);
* Description: Get node's successor/predecessor.
* Args:
* tree: Pointer to an initialized red-black tree object.
* node: A node in tree.
* Ret: node's successor/predecessor in tree, or NULL if node is
* last/first.
*
* static ex_node_t *
* ex_search(ex_t *tree, ex_node_t *key);
* Description: Search for node that matches key.
* Args:
* tree: Pointer to an initialized red-black tree object.
* key : Search key.
* Ret: Node in tree that matches key, or NULL if no match.
*
* static ex_node_t *
* ex_nsearch(ex_t *tree, ex_node_t *key);
* static ex_node_t *
* ex_psearch(ex_t *tree, ex_node_t *key);
* Description: Search for node that matches key. If no match is found,
* return what would be key's successor/predecessor, were
* key in tree.
* Args:
* tree: Pointer to an initialized red-black tree object.
* key : Search key.
* Ret: Node in tree that matches key, or if no match, hypothetical node's
* successor/predecessor (NULL if no successor/predecessor).
*
* static void
* ex_insert(ex_t *tree, ex_node_t *node);
* Description: Insert node into tree.
* Args:
* tree: Pointer to an initialized red-black tree object.
* node: Node to be inserted into tree.
*
* static void
* ex_remove(ex_t *tree, ex_node_t *node);
* Description: Remove node from tree.
* Args:
* tree: Pointer to an initialized red-black tree object.
* node: Node in tree to be removed.
*
* static ex_node_t *
* ex_iter(ex_t *tree, ex_node_t *start, ex_node_t *(*cb)(ex_t *,
* ex_node_t *, void *), void *arg);
* static ex_node_t *
* ex_reverse_iter(ex_t *tree, ex_node_t *start, ex_node *(*cb)(ex_t *,
* ex_node_t *, void *), void *arg);
* Description: Iterate forward/backward over tree, starting at node. If
* tree is modified, iteration must be immediately
* terminated by the callback function that causes the
* modification.
* Args:
* tree : Pointer to an initialized red-black tree object.
* start: Node at which to start iteration, or NULL to start at
* first/last node.
* cb : Callback function, which is called for each node during
* iteration. Under normal circumstances the callback function
* should return NULL, which causes iteration to continue. If a
* callback function returns non-NULL, iteration is immediately
* terminated and the non-NULL return value is returned by the
* iterator. This is useful for re-starting iteration after
* modifying tree.
* arg : Opaque pointer passed to cb().
* Ret: NULL if iteration completed, or the non-NULL callback return value
* that caused termination of the iteration.
*/
#define rb_gen(a_attr, a_prefix, a_rbt_type, a_type, a_field, a_cmp) \
a_attr void \
a_prefix##new(a_rbt_type *rbtree) { \
rb_new(a_type, a_field, rbtree); \
} \
a_attr a_type * \
a_prefix##first(a_rbt_type *rbtree) { \
a_type *ret; \
rbtn_first(a_type, a_field, rbtree, rbtree->rbt_root, ret); \
if (ret == &rbtree->rbt_nil) { \
ret = NULL; \
} \
return (ret); \
} \
a_attr a_type * \
a_prefix##last(a_rbt_type *rbtree) { \
a_type *ret; \
rbtn_last(a_type, a_field, rbtree, rbtree->rbt_root, ret); \
if (ret == &rbtree->rbt_nil) { \
ret = NULL; \
} \
return (ret); \
} \
a_attr a_type * \
a_prefix##next(a_rbt_type *rbtree, a_type *node) { \
a_type *ret; \
if (rbtn_right_get(a_type, a_field, node) != &rbtree->rbt_nil) { \
rbtn_first(a_type, a_field, rbtree, rbtn_right_get(a_type, \
a_field, node), ret); \
} else { \
a_type *tnode = rbtree->rbt_root; \
assert(tnode != &rbtree->rbt_nil); \
ret = &rbtree->rbt_nil; \
while (true) { \
int cmp = (a_cmp)(node, tnode); \
if (cmp < 0) { \
ret = tnode; \
tnode = rbtn_left_get(a_type, a_field, tnode); \
} else if (cmp > 0) { \
tnode = rbtn_right_get(a_type, a_field, tnode); \
} else { \
break; \
} \
assert(tnode != &rbtree->rbt_nil); \
} \
} \
if (ret == &rbtree->rbt_nil) { \
ret = (NULL); \
} \
return (ret); \
} \
a_attr a_type * \
a_prefix##prev(a_rbt_type *rbtree, a_type *node) { \
a_type *ret; \
if (rbtn_left_get(a_type, a_field, node) != &rbtree->rbt_nil) { \
rbtn_last(a_type, a_field, rbtree, rbtn_left_get(a_type, \
a_field, node), ret); \
} else { \
a_type *tnode = rbtree->rbt_root; \
assert(tnode != &rbtree->rbt_nil); \
ret = &rbtree->rbt_nil; \
while (true) { \
int cmp = (a_cmp)(node, tnode); \
if (cmp < 0) { \
tnode = rbtn_left_get(a_type, a_field, tnode); \
} else if (cmp > 0) { \
ret = tnode; \
tnode = rbtn_right_get(a_type, a_field, tnode); \
} else { \
break; \
} \
assert(tnode != &rbtree->rbt_nil); \
} \
} \
if (ret == &rbtree->rbt_nil) { \
ret = (NULL); \
} \
return (ret); \
} \
a_attr a_type * \
a_prefix##search(a_rbt_type *rbtree, a_type *key) { \
a_type *ret; \
int cmp; \
ret = rbtree->rbt_root; \
while (ret != &rbtree->rbt_nil \
&& (cmp = (a_cmp)(key, ret)) != 0) { \
if (cmp < 0) { \
ret = rbtn_left_get(a_type, a_field, ret); \
} else { \
ret = rbtn_right_get(a_type, a_field, ret); \
} \
} \
if (ret == &rbtree->rbt_nil) { \
ret = (NULL); \
} \
return (ret); \
} \
a_attr a_type * \
a_prefix##nsearch(a_rbt_type *rbtree, a_type *key) { \
a_type *ret; \
a_type *tnode = rbtree->rbt_root; \
ret = &rbtree->rbt_nil; \
while (tnode != &rbtree->rbt_nil) { \
int cmp = (a_cmp)(key, tnode); \
if (cmp < 0) { \
ret = tnode; \
tnode = rbtn_left_get(a_type, a_field, tnode); \
} else if (cmp > 0) { \
tnode = rbtn_right_get(a_type, a_field, tnode); \
} else { \
ret = tnode; \
break; \
} \
} \
if (ret == &rbtree->rbt_nil) { \
ret = (NULL); \
} \
return (ret); \
} \
a_attr a_type * \
a_prefix##psearch(a_rbt_type *rbtree, a_type *key) { \
a_type *ret; \
a_type *tnode = rbtree->rbt_root; \
ret = &rbtree->rbt_nil; \
while (tnode != &rbtree->rbt_nil) { \
int cmp = (a_cmp)(key, tnode); \
if (cmp < 0) { \
tnode = rbtn_left_get(a_type, a_field, tnode); \
} else if (cmp > 0) { \
ret = tnode; \
tnode = rbtn_right_get(a_type, a_field, tnode); \
} else { \
ret = tnode; \
break; \
} \
} \
if (ret == &rbtree->rbt_nil) { \
ret = (NULL); \
} \
return (ret); \
} \
a_attr void \
a_prefix##insert(a_rbt_type *rbtree, a_type *node) { \
struct { \
a_type *node; \
int cmp; \
} path[sizeof(void *) << 4], *pathp; \
rbt_node_new(a_type, a_field, rbtree, node); \
/* Wind. */ \
path->node = rbtree->rbt_root; \
for (pathp = path; pathp->node != &rbtree->rbt_nil; pathp++) { \
int cmp = pathp->cmp = a_cmp(node, pathp->node); \
assert(cmp != 0); \
if (cmp < 0) { \
pathp[1].node = rbtn_left_get(a_type, a_field, \
pathp->node); \
} else { \
pathp[1].node = rbtn_right_get(a_type, a_field, \
pathp->node); \
} \
} \
pathp->node = node; \
/* Unwind. */ \
for (pathp--; (uintptr_t)pathp >= (uintptr_t)path; pathp--) { \
a_type *cnode = pathp->node; \
if (pathp->cmp < 0) { \
a_type *left = pathp[1].node; \
rbtn_left_set(a_type, a_field, cnode, left); \
if (rbtn_red_get(a_type, a_field, left)) { \
a_type *leftleft = rbtn_left_get(a_type, a_field, left);\
if (rbtn_red_get(a_type, a_field, leftleft)) { \
/* Fix up 4-node. */ \
a_type *tnode; \
rbtn_black_set(a_type, a_field, leftleft); \
rbtn_rotate_right(a_type, a_field, cnode, tnode); \
cnode = tnode; \
} \
} else { \
return; \
} \
} else { \
a_type *right = pathp[1].node; \
rbtn_right_set(a_type, a_field, cnode, right); \
if (rbtn_red_get(a_type, a_field, right)) { \
a_type *left = rbtn_left_get(a_type, a_field, cnode); \
if (rbtn_red_get(a_type, a_field, left)) { \
/* Split 4-node. */ \
rbtn_black_set(a_type, a_field, left); \
rbtn_black_set(a_type, a_field, right); \
rbtn_red_set(a_type, a_field, cnode); \
} else { \
/* Lean left. */ \
a_type *tnode; \
bool tred = rbtn_red_get(a_type, a_field, cnode); \
rbtn_rotate_left(a_type, a_field, cnode, tnode); \
rbtn_color_set(a_type, a_field, tnode, tred); \
rbtn_red_set(a_type, a_field, cnode); \
cnode = tnode; \
} \
} else { \
return; \
} \
} \
pathp->node = cnode; \
} \
/* Set root, and make it black. */ \
rbtree->rbt_root = path->node; \
rbtn_black_set(a_type, a_field, rbtree->rbt_root); \
} \
a_attr void \
a_prefix##remove(a_rbt_type *rbtree, a_type *node) { \
struct { \
a_type *node; \
int cmp; \
} *pathp, *nodep, path[sizeof(void *) << 4]; \
/* Wind. */ \
nodep = NULL; /* Silence compiler warning. */ \
path->node = rbtree->rbt_root; \
for (pathp = path; pathp->node != &rbtree->rbt_nil; pathp++) { \
int cmp = pathp->cmp = a_cmp(node, pathp->node); \
if (cmp < 0) { \
pathp[1].node = rbtn_left_get(a_type, a_field, \
pathp->node); \
} else { \
pathp[1].node = rbtn_right_get(a_type, a_field, \
pathp->node); \
if (cmp == 0) { \
/* Find node's successor, in preparation for swap. */ \
pathp->cmp = 1; \
nodep = pathp; \
for (pathp++; pathp->node != &rbtree->rbt_nil; \
pathp++) { \
pathp->cmp = -1; \
pathp[1].node = rbtn_left_get(a_type, a_field, \
pathp->node); \
} \
break; \
} \
} \
} \
assert(nodep->node == node); \
pathp--; \
if (pathp->node != node) { \
/* Swap node with its successor. */ \
bool tred = rbtn_red_get(a_type, a_field, pathp->node); \
rbtn_color_set(a_type, a_field, pathp->node, \
rbtn_red_get(a_type, a_field, node)); \
rbtn_left_set(a_type, a_field, pathp->node, \
rbtn_left_get(a_type, a_field, node)); \
/* If node's successor is its right child, the following code */\
/* will do the wrong thing for the right child pointer. */\
/* However, it doesn't matter, because the pointer will be */\
/* properly set when the successor is pruned. */\
rbtn_right_set(a_type, a_field, pathp->node, \
rbtn_right_get(a_type, a_field, node)); \
rbtn_color_set(a_type, a_field, node, tred); \
/* The pruned leaf node's child pointers are never accessed */\
/* again, so don't bother setting them to nil. */\
nodep->node = pathp->node; \
pathp->node = node; \
if (nodep == path) { \
rbtree->rbt_root = nodep->node; \
} else { \
if (nodep[-1].cmp < 0) { \
rbtn_left_set(a_type, a_field, nodep[-1].node, \
nodep->node); \
} else { \
rbtn_right_set(a_type, a_field, nodep[-1].node, \
nodep->node); \
} \
} \
} else { \
a_type *left = rbtn_left_get(a_type, a_field, node); \
if (left != &rbtree->rbt_nil) { \
/* node has no successor, but it has a left child. */\
/* Splice node out, without losing the left child. */\
assert(rbtn_red_get(a_type, a_field, node) == false); \
assert(rbtn_red_get(a_type, a_field, left)); \
rbtn_black_set(a_type, a_field, left); \
if (pathp == path) { \
rbtree->rbt_root = left; \
} else { \
if (pathp[-1].cmp < 0) { \
rbtn_left_set(a_type, a_field, pathp[-1].node, \
left); \
} else { \
rbtn_right_set(a_type, a_field, pathp[-1].node, \
left); \
} \
} \
return; \
} else if (pathp == path) { \
/* The tree only contained one node. */ \
rbtree->rbt_root = &rbtree->rbt_nil; \
return; \
} \
} \
if (rbtn_red_get(a_type, a_field, pathp->node)) { \
/* Prune red node, which requires no fixup. */ \
assert(pathp[-1].cmp < 0); \
rbtn_left_set(a_type, a_field, pathp[-1].node, \
&rbtree->rbt_nil); \
return; \
} \
/* The node to be pruned is black, so unwind until balance is */\
/* restored. */\
pathp->node = &rbtree->rbt_nil; \
for (pathp--; (uintptr_t)pathp >= (uintptr_t)path; pathp--) { \
assert(pathp->cmp != 0); \
if (pathp->cmp < 0) { \
rbtn_left_set(a_type, a_field, pathp->node, \
pathp[1].node); \
assert(rbtn_red_get(a_type, a_field, pathp[1].node) \
== false); \
if (rbtn_red_get(a_type, a_field, pathp->node)) { \
a_type *right = rbtn_right_get(a_type, a_field, \
pathp->node); \
a_type *rightleft = rbtn_left_get(a_type, a_field, \
right); \
a_type *tnode; \
if (rbtn_red_get(a_type, a_field, rightleft)) { \
/* In the following diagrams, ||, //, and \\ */\
/* indicate the path to the removed node. */\
/* */\
/* || */\
/* pathp(r) */\
/* // \ */\
/* (b) (b) */\
/* / */\
/* (r) */\
/* */\
rbtn_black_set(a_type, a_field, pathp->node); \
rbtn_rotate_right(a_type, a_field, right, tnode); \
rbtn_right_set(a_type, a_field, pathp->node, tnode);\
rbtn_rotate_left(a_type, a_field, pathp->node, \
tnode); \
} else { \
/* || */\
/* pathp(r) */\
/* // \ */\
/* (b) (b) */\
/* / */\
/* (b) */\
/* */\
rbtn_rotate_left(a_type, a_field, pathp->node, \
tnode); \
} \
/* Balance restored, but rotation modified subtree */\
/* root. */\
assert((uintptr_t)pathp > (uintptr_t)path); \
if (pathp[-1].cmp < 0) { \
rbtn_left_set(a_type, a_field, pathp[-1].node, \
tnode); \
} else { \
rbtn_right_set(a_type, a_field, pathp[-1].node, \
tnode); \
} \
return; \
} else { \
a_type *right = rbtn_right_get(a_type, a_field, \
pathp->node); \
a_type *rightleft = rbtn_left_get(a_type, a_field, \
right); \
if (rbtn_red_get(a_type, a_field, rightleft)) { \
/* || */\
/* pathp(b) */\
/* // \ */\
/* (b) (b) */\
/* / */\
/* (r) */\
a_type *tnode; \
rbtn_black_set(a_type, a_field, rightleft); \
rbtn_rotate_right(a_type, a_field, right, tnode); \
rbtn_right_set(a_type, a_field, pathp->node, tnode);\
rbtn_rotate_left(a_type, a_field, pathp->node, \
tnode); \
/* Balance restored, but rotation modified */\
/* subree root, which may actually be the tree */\
/* root. */\
if (pathp == path) { \
/* Set root. */ \
rbtree->rbt_root = tnode; \
} else { \
if (pathp[-1].cmp < 0) { \
rbtn_left_set(a_type, a_field, \
pathp[-1].node, tnode); \
} else { \
rbtn_right_set(a_type, a_field, \
pathp[-1].node, tnode); \
} \
} \
return; \
} else { \
/* || */\
/* pathp(b) */\
/* // \ */\
/* (b) (b) */\
/* / */\
/* (b) */\
a_type *tnode; \
rbtn_red_set(a_type, a_field, pathp->node); \
rbtn_rotate_left(a_type, a_field, pathp->node, \
tnode); \
pathp->node = tnode; \
} \
} \
} else { \
a_type *left; \
rbtn_right_set(a_type, a_field, pathp->node, \
pathp[1].node); \
left = rbtn_left_get(a_type, a_field, pathp->node); \
if (rbtn_red_get(a_type, a_field, left)) { \
a_type *tnode; \
a_type *leftright = rbtn_right_get(a_type, a_field, \
left); \
a_type *leftrightleft = rbtn_left_get(a_type, a_field, \
leftright); \
if (rbtn_red_get(a_type, a_field, leftrightleft)) { \
/* || */\
/* pathp(b) */\
/* / \\ */\
/* (r) (b) */\
/* \ */\
/* (b) */\
/* / */\
/* (r) */\
a_type *unode; \
rbtn_black_set(a_type, a_field, leftrightleft); \
rbtn_rotate_right(a_type, a_field, pathp->node, \
unode); \
rbtn_rotate_right(a_type, a_field, pathp->node, \
tnode); \
rbtn_right_set(a_type, a_field, unode, tnode); \
rbtn_rotate_left(a_type, a_field, unode, tnode); \
} else { \
/* || */\
/* pathp(b) */\
/* / \\ */\
/* (r) (b) */\
/* \ */\
/* (b) */\
/* / */\
/* (b) */\
assert(leftright != &rbtree->rbt_nil); \
rbtn_red_set(a_type, a_field, leftright); \
rbtn_rotate_right(a_type, a_field, pathp->node, \
tnode); \
rbtn_black_set(a_type, a_field, tnode); \
} \
/* Balance restored, but rotation modified subtree */\
/* root, which may actually be the tree root. */\
if (pathp == path) { \
/* Set root. */ \
rbtree->rbt_root = tnode; \
} else { \
if (pathp[-1].cmp < 0) { \
rbtn_left_set(a_type, a_field, pathp[-1].node, \
tnode); \
} else { \
rbtn_right_set(a_type, a_field, pathp[-1].node, \
tnode); \
} \
} \
return; \
} else if (rbtn_red_get(a_type, a_field, pathp->node)) { \
a_type *leftleft = rbtn_left_get(a_type, a_field, left);\
if (rbtn_red_get(a_type, a_field, leftleft)) { \
/* || */\
/* pathp(r) */\
/* / \\ */\
/* (b) (b) */\
/* / */\
/* (r) */\
a_type *tnode; \
rbtn_black_set(a_type, a_field, pathp->node); \
rbtn_red_set(a_type, a_field, left); \
rbtn_black_set(a_type, a_field, leftleft); \
rbtn_rotate_right(a_type, a_field, pathp->node, \
tnode); \
/* Balance restored, but rotation modified */\
/* subtree root. */\
assert((uintptr_t)pathp > (uintptr_t)path); \
if (pathp[-1].cmp < 0) { \
rbtn_left_set(a_type, a_field, pathp[-1].node, \
tnode); \
} else { \
rbtn_right_set(a_type, a_field, pathp[-1].node, \
tnode); \
} \
return; \
} else { \
/* || */\
/* pathp(r) */\
/* / \\ */\
/* (b) (b) */\
/* / */\
/* (b) */\
rbtn_red_set(a_type, a_field, left); \
rbtn_black_set(a_type, a_field, pathp->node); \
/* Balance restored. */ \
return; \
} \
} else { \
a_type *leftleft = rbtn_left_get(a_type, a_field, left);\
if (rbtn_red_get(a_type, a_field, leftleft)) { \
/* || */\
/* pathp(b) */\
/* / \\ */\
/* (b) (b) */\
/* / */\
/* (r) */\
a_type *tnode; \
rbtn_black_set(a_type, a_field, leftleft); \
rbtn_rotate_right(a_type, a_field, pathp->node, \
tnode); \
/* Balance restored, but rotation modified */\
/* subtree root, which may actually be the tree */\
/* root. */\
if (pathp == path) { \
/* Set root. */ \
rbtree->rbt_root = tnode; \
} else { \
if (pathp[-1].cmp < 0) { \
rbtn_left_set(a_type, a_field, \
pathp[-1].node, tnode); \
} else { \
rbtn_right_set(a_type, a_field, \
pathp[-1].node, tnode); \
} \
} \
return; \
} else { \
/* || */\
/* pathp(b) */\
/* / \\ */\
/* (b) (b) */\
/* / */\
/* (b) */\
rbtn_red_set(a_type, a_field, left); \
} \
} \
} \
} \
/* Set root. */ \
rbtree->rbt_root = path->node; \
assert(rbtn_red_get(a_type, a_field, rbtree->rbt_root) == false); \
} \
a_attr a_type * \
a_prefix##iter_recurse(a_rbt_type *rbtree, a_type *node, \
a_type *(*cb)(a_rbt_type *, a_type *, void *), void *arg) { \
if (node == &rbtree->rbt_nil) { \
return (&rbtree->rbt_nil); \
} else { \
a_type *ret; \
if ((ret = a_prefix##iter_recurse(rbtree, rbtn_left_get(a_type, \
a_field, node), cb, arg)) != &rbtree->rbt_nil \
|| (ret = cb(rbtree, node, arg)) != NULL) { \
return (ret); \
} \
return (a_prefix##iter_recurse(rbtree, rbtn_right_get(a_type, \
a_field, node), cb, arg)); \
} \
} \
a_attr a_type * \
a_prefix##iter_start(a_rbt_type *rbtree, a_type *start, a_type *node, \
a_type *(*cb)(a_rbt_type *, a_type *, void *), void *arg) { \
int cmp = a_cmp(start, node); \
if (cmp < 0) { \
a_type *ret; \
if ((ret = a_prefix##iter_start(rbtree, start, \
rbtn_left_get(a_type, a_field, node), cb, arg)) != \
&rbtree->rbt_nil || (ret = cb(rbtree, node, arg)) != NULL) { \
return (ret); \
} \
return (a_prefix##iter_recurse(rbtree, rbtn_right_get(a_type, \
a_field, node), cb, arg)); \
} else if (cmp > 0) { \
return (a_prefix##iter_start(rbtree, start, \
rbtn_right_get(a_type, a_field, node), cb, arg)); \
} else { \
a_type *ret; \
if ((ret = cb(rbtree, node, arg)) != NULL) { \
return (ret); \
} \
return (a_prefix##iter_recurse(rbtree, rbtn_right_get(a_type, \
a_field, node), cb, arg)); \
} \
} \
a_attr a_type * \
a_prefix##iter(a_rbt_type *rbtree, a_type *start, a_type *(*cb)( \
a_rbt_type *, a_type *, void *), void *arg) { \
a_type *ret; \
if (start != NULL) { \
ret = a_prefix##iter_start(rbtree, start, rbtree->rbt_root, \
cb, arg); \
} else { \
ret = a_prefix##iter_recurse(rbtree, rbtree->rbt_root, cb, arg);\
} \
if (ret == &rbtree->rbt_nil) { \
ret = NULL; \
} \
return (ret); \
} \
a_attr a_type * \
a_prefix##reverse_iter_recurse(a_rbt_type *rbtree, a_type *node, \
a_type *(*cb)(a_rbt_type *, a_type *, void *), void *arg) { \
if (node == &rbtree->rbt_nil) { \
return (&rbtree->rbt_nil); \
} else { \
a_type *ret; \
if ((ret = a_prefix##reverse_iter_recurse(rbtree, \
rbtn_right_get(a_type, a_field, node), cb, arg)) != \
&rbtree->rbt_nil || (ret = cb(rbtree, node, arg)) != NULL) { \
return (ret); \
} \
return (a_prefix##reverse_iter_recurse(rbtree, \
rbtn_left_get(a_type, a_field, node), cb, arg)); \
} \
} \
a_attr a_type * \
a_prefix##reverse_iter_start(a_rbt_type *rbtree, a_type *start, \
a_type *node, a_type *(*cb)(a_rbt_type *, a_type *, void *), \
void *arg) { \
int cmp = a_cmp(start, node); \
if (cmp > 0) { \
a_type *ret; \
if ((ret = a_prefix##reverse_iter_start(rbtree, start, \
rbtn_right_get(a_type, a_field, node), cb, arg)) != \
&rbtree->rbt_nil || (ret = cb(rbtree, node, arg)) != NULL) { \
return (ret); \
} \
return (a_prefix##reverse_iter_recurse(rbtree, \
rbtn_left_get(a_type, a_field, node), cb, arg)); \
} else if (cmp < 0) { \
return (a_prefix##reverse_iter_start(rbtree, start, \
rbtn_left_get(a_type, a_field, node), cb, arg)); \
} else { \
a_type *ret; \
if ((ret = cb(rbtree, node, arg)) != NULL) { \
return (ret); \
} \
return (a_prefix##reverse_iter_recurse(rbtree, \
rbtn_left_get(a_type, a_field, node), cb, arg)); \
} \
} \
a_attr a_type * \
a_prefix##reverse_iter(a_rbt_type *rbtree, a_type *start, \
a_type *(*cb)(a_rbt_type *, a_type *, void *), void *arg) { \
a_type *ret; \
if (start != NULL) { \
ret = a_prefix##reverse_iter_start(rbtree, start, \
rbtree->rbt_root, cb, arg); \
} else { \
ret = a_prefix##reverse_iter_recurse(rbtree, rbtree->rbt_root, \
cb, arg); \
} \
if (ret == &rbtree->rbt_nil) { \
ret = NULL; \
} \
return (ret); \
}
#endif /* RB_H_ */
deps/jemalloc/include/jemalloc/internal/rtree.h
Vendored
+172
View File
@@ -0,0 +1,172 @@
/*
* This radix tree implementation is tailored to the singular purpose of
* tracking which chunks are currently owned by jemalloc. This functionality
* is mandatory for OS X, where jemalloc must be able to respond to object
* ownership queries.
*
*******************************************************************************
*/
#ifdef JEMALLOC_H_TYPES
typedef struct rtree_s rtree_t;
/*
* Size of each radix tree node (must be a power of 2). This impacts tree
* depth.
*/
#define RTREE_NODESIZE (1U << 16)
typedef void *(rtree_alloc_t)(size_t);
typedef void (rtree_dalloc_t)(void *);
#endif /* JEMALLOC_H_TYPES */
/******************************************************************************/
#ifdef JEMALLOC_H_STRUCTS
struct rtree_s {
rtree_alloc_t *alloc;
rtree_dalloc_t *dalloc;
malloc_mutex_t mutex;
void **root;
unsigned height;
unsigned level2bits[1]; /* Dynamically sized. */
};
#endif /* JEMALLOC_H_STRUCTS */
/******************************************************************************/
#ifdef JEMALLOC_H_EXTERNS
rtree_t *rtree_new(unsigned bits, rtree_alloc_t *alloc, rtree_dalloc_t *dalloc);
void rtree_delete(rtree_t *rtree);
void rtree_prefork(rtree_t *rtree);
void rtree_postfork_parent(rtree_t *rtree);
void rtree_postfork_child(rtree_t *rtree);
#endif /* JEMALLOC_H_EXTERNS */
/******************************************************************************/
#ifdef JEMALLOC_H_INLINES
#ifndef JEMALLOC_ENABLE_INLINE
#ifdef JEMALLOC_DEBUG
uint8_t rtree_get_locked(rtree_t *rtree, uintptr_t key);
#endif
uint8_t rtree_get(rtree_t *rtree, uintptr_t key);
bool rtree_set(rtree_t *rtree, uintptr_t key, uint8_t val);
#endif
#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_RTREE_C_))
#define RTREE_GET_GENERATE(f) \
/* The least significant bits of the key are ignored. */ \
JEMALLOC_INLINE uint8_t \
f(rtree_t *rtree, uintptr_t key) \
{ \
uint8_t ret; \
uintptr_t subkey; \
unsigned i, lshift, height, bits; \
void **node, **child; \
\
RTREE_LOCK(&rtree->mutex); \
for (i = lshift = 0, height = rtree->height, node = rtree->root;\
i < height - 1; \
i++, lshift += bits, node = child) { \
bits = rtree->level2bits[i]; \
subkey = (key << lshift) >> ((ZU(1) << (LG_SIZEOF_PTR + \
3)) - bits); \
child = (void**)node[subkey]; \
if (child == NULL) { \
RTREE_UNLOCK(&rtree->mutex); \
return (0); \
} \
} \
\
/* \
* node is a leaf, so it contains values rather than node \
* pointers. \
*/ \
bits = rtree->level2bits[i]; \
subkey = (key << lshift) >> ((ZU(1) << (LG_SIZEOF_PTR+3)) - \
bits); \
{ \
uint8_t *leaf = (uint8_t *)node; \
ret = leaf[subkey]; \
} \
RTREE_UNLOCK(&rtree->mutex); \
\
RTREE_GET_VALIDATE \
return (ret); \
}
#ifdef JEMALLOC_DEBUG
# define RTREE_LOCK(l) malloc_mutex_lock(l)
# define RTREE_UNLOCK(l) malloc_mutex_unlock(l)
# define RTREE_GET_VALIDATE
RTREE_GET_GENERATE(rtree_get_locked)
# undef RTREE_LOCK
# undef RTREE_UNLOCK
# undef RTREE_GET_VALIDATE
#endif
#define RTREE_LOCK(l)
#define RTREE_UNLOCK(l)
#ifdef JEMALLOC_DEBUG
/*
* Suppose that it were possible for a jemalloc-allocated chunk to be
* munmap()ped, followed by a different allocator in another thread re-using
* overlapping virtual memory, all without invalidating the cached rtree
* value. The result would be a false positive (the rtree would claim that
* jemalloc owns memory that it had actually discarded). This scenario
* seems impossible, but the following assertion is a prudent sanity check.
*/
# define RTREE_GET_VALIDATE \
assert(rtree_get_locked(rtree, key) == ret);
#else
# define RTREE_GET_VALIDATE
#endif
RTREE_GET_GENERATE(rtree_get)
#undef RTREE_LOCK
#undef RTREE_UNLOCK
#undef RTREE_GET_VALIDATE
JEMALLOC_INLINE bool
rtree_set(rtree_t *rtree, uintptr_t key, uint8_t val)
{
uintptr_t subkey;
unsigned i, lshift, height, bits;
void **node, **child;
malloc_mutex_lock(&rtree->mutex);
for (i = lshift = 0, height = rtree->height, node = rtree->root;
i < height - 1;
i++, lshift += bits, node = child) {
bits = rtree->level2bits[i];
subkey = (key << lshift) >> ((ZU(1) << (LG_SIZEOF_PTR+3)) -
bits);
child = (void**)node[subkey];
if (child == NULL) {
size_t size = ((i + 1 < height - 1) ? sizeof(void *)
: (sizeof(uint8_t))) << rtree->level2bits[i+1];
child = (void**)rtree->alloc(size);
if (child == NULL) {
malloc_mutex_unlock(&rtree->mutex);
return (true);
}
memset(child, 0, size);
node[subkey] = child;
}
}
/* node is a leaf, so it contains values rather than node pointers. */
bits = rtree->level2bits[i];
subkey = (key << lshift) >> ((ZU(1) << (LG_SIZEOF_PTR+3)) - bits);
{
uint8_t *leaf = (uint8_t *)node;
leaf[subkey] = val;
}
malloc_mutex_unlock(&rtree->mutex);
return (false);
}
#endif
#endif /* JEMALLOC_H_INLINES */
/******************************************************************************/
deps/jemalloc/include/jemalloc/internal/size_classes.sh
Vendored Executable
+122
View File
@@ -0,0 +1,122 @@
#!/bin/sh
# The following limits are chosen such that they cover all supported platforms.
# Range of quanta.
lg_qmin=3
lg_qmax=4
# The range of tiny size classes is [2^lg_tmin..2^(lg_q-1)].
lg_tmin=3
# Range of page sizes.
lg_pmin=12
lg_pmax=16
pow2() {
e=$1
pow2_result=1
while [ ${e} -gt 0 ] ; do
pow2_result=$((${pow2_result} + ${pow2_result}))
e=$((${e} - 1))
done
}
cat <<EOF
/* This file was automatically generated by size_classes.sh. */
/******************************************************************************/
#ifdef JEMALLOC_H_TYPES
EOF
lg_q=${lg_qmin}
while [ ${lg_q} -le ${lg_qmax} ] ; do
lg_t=${lg_tmin}
while [ ${lg_t} -le ${lg_q} ] ; do
lg_p=${lg_pmin}
while [ ${lg_p} -le ${lg_pmax} ] ; do
echo "#if (LG_TINY_MIN == ${lg_t} && LG_QUANTUM == ${lg_q} && LG_PAGE == ${lg_p})"
echo "#define SIZE_CLASSES_DEFINED"
pow2 ${lg_q}; q=${pow2_result}
pow2 ${lg_t}; t=${pow2_result}
pow2 ${lg_p}; p=${pow2_result}
bin=0
psz=0
sz=${t}
delta=$((${sz} - ${psz}))
echo "/* SIZE_CLASS(bin, delta, sz) */"
echo "#define SIZE_CLASSES \\"
# Tiny size classes.
while [ ${sz} -lt ${q} ] ; do
echo " SIZE_CLASS(${bin}, ${delta}, ${sz}) \\"
bin=$((${bin} + 1))
psz=${sz}
sz=$((${sz} + ${sz}))
delta=$((${sz} - ${psz}))
done
# Quantum-multiple size classes. For each doubling of sz, as many as 4
# size classes exist. Their spacing is the greater of:
# - q
# - sz/4, where sz is a power of 2
while [ ${sz} -lt ${p} ] ; do
if [ ${sz} -ge $((${q} * 4)) ] ; then
i=$((${sz} / 4))
else
i=${q}
fi
next_2pow=$((${sz} * 2))
while [ ${sz} -lt $next_2pow ] ; do
echo " SIZE_CLASS(${bin}, ${delta}, ${sz}) \\"
bin=$((${bin} + 1))
psz=${sz}
sz=$((${sz} + ${i}))
delta=$((${sz} - ${psz}))
done
done
echo
echo "#define NBINS ${bin}"
echo "#define SMALL_MAXCLASS ${psz}"
echo "#endif"
echo
lg_p=$((${lg_p} + 1))
done
lg_t=$((${lg_t} + 1))
done
lg_q=$((${lg_q} + 1))
done
cat <<EOF
#ifndef SIZE_CLASSES_DEFINED
# error "No size class definitions match configuration"
#endif
#undef SIZE_CLASSES_DEFINED
/*
* The small_size2bin lookup table uses uint8_t to encode each bin index, so we
* cannot support more than 256 small size classes. Further constrain NBINS to
* 255 to support prof_promote, since all small size classes, plus a "not
* small" size class must be stored in 8 bits of arena_chunk_map_t's bits
* field.
*/
#if (NBINS > 255)
# error "Too many small size classes"
#endif
#endif /* JEMALLOC_H_TYPES */
/******************************************************************************/
#ifdef JEMALLOC_H_STRUCTS
#endif /* JEMALLOC_H_STRUCTS */
/******************************************************************************/
#ifdef JEMALLOC_H_EXTERNS
#endif /* JEMALLOC_H_EXTERNS */
/******************************************************************************/
#ifdef JEMALLOC_H_INLINES
#endif /* JEMALLOC_H_INLINES */
/******************************************************************************/
EOF
deps/jemalloc/include/jemalloc/internal/stats.h
Vendored
+173
View File
@@ -0,0 +1,173 @@
/******************************************************************************/
#ifdef JEMALLOC_H_TYPES
typedef struct tcache_bin_stats_s tcache_bin_stats_t;
typedef struct malloc_bin_stats_s malloc_bin_stats_t;
typedef struct malloc_large_stats_s malloc_large_stats_t;
typedef struct arena_stats_s arena_stats_t;
typedef struct chunk_stats_s chunk_stats_t;
#endif /* JEMALLOC_H_TYPES */
/******************************************************************************/
#ifdef JEMALLOC_H_STRUCTS
struct tcache_bin_stats_s {
/*
* Number of allocation requests that corresponded to the size of this
* bin.
*/
uint64_t nrequests;
};
struct malloc_bin_stats_s {
/*
* Current number of bytes allocated, including objects currently
* cached by tcache.
*/
size_t allocated;
/*
* Total number of allocation/deallocation requests served directly by
* the bin. Note that tcache may allocate an object, then recycle it
* many times, resulting many increments to nrequests, but only one
* each to nmalloc and ndalloc.
*/
uint64_t nmalloc;
uint64_t ndalloc;
/*
* Number of allocation requests that correspond to the size of this
* bin. This includes requests served by tcache, though tcache only
* periodically merges into this counter.
*/
uint64_t nrequests;
/* Number of tcache fills from this bin. */
uint64_t nfills;
/* Number of tcache flushes to this bin. */
uint64_t nflushes;
/* Total number of runs created for this bin's size class. */
uint64_t nruns;
/*
* Total number of runs reused by extracting them from the runs tree for
* this bin's size class.
*/
uint64_t reruns;
/* Current number of runs in this bin. */
size_t curruns;
};
struct malloc_large_stats_s {
/*
* Total number of allocation/deallocation requests served directly by
* the arena. Note that tcache may allocate an object, then recycle it
* many times, resulting many increments to nrequests, but only one
* each to nmalloc and ndalloc.
*/
uint64_t nmalloc;
uint64_t ndalloc;
/*
* Number of allocation requests that correspond to this size class.
* This includes requests served by tcache, though tcache only
* periodically merges into this counter.
*/
uint64_t nrequests;
/* Current number of runs of this size class. */
size_t curruns;
};
struct arena_stats_s {
/* Number of bytes currently mapped. */
size_t mapped;
/*
* Total number of purge sweeps, total number of madvise calls made,
* and total pages purged in order to keep dirty unused memory under
* control.
*/
uint64_t npurge;
uint64_t nmadvise;
uint64_t purged;
/* Per-size-category statistics. */
size_t allocated_large;
uint64_t nmalloc_large;
uint64_t ndalloc_large;
uint64_t nrequests_large;
/*
* One element for each possible size class, including sizes that
* overlap with bin size classes. This is necessary because ipalloc()
* sometimes has to use such large objects in order to assure proper
* alignment.
*/
malloc_large_stats_t *lstats;
};
struct chunk_stats_s {
/* Number of chunks that were allocated. */
uint64_t nchunks;
/* High-water mark for number of chunks allocated. */
size_t highchunks;
/*
* Current number of chunks allocated. This value isn't maintained for
* any other purpose, so keep track of it in order to be able to set
* highchunks.
*/
size_t curchunks;
};
#endif /* JEMALLOC_H_STRUCTS */
/******************************************************************************/
#ifdef JEMALLOC_H_EXTERNS
extern bool opt_stats_print;
extern size_t stats_cactive;
void stats_print(void (*write)(void *, const char *), void *cbopaque,
const char *opts);
#endif /* JEMALLOC_H_EXTERNS */
/******************************************************************************/
#ifdef JEMALLOC_H_INLINES
#ifndef JEMALLOC_ENABLE_INLINE
size_t stats_cactive_get(void);
void stats_cactive_add(size_t size);
void stats_cactive_sub(size_t size);
#endif
#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_STATS_C_))
JEMALLOC_INLINE size_t
stats_cactive_get(void)
{
return (atomic_read_z(&stats_cactive));
}
JEMALLOC_INLINE void
stats_cactive_add(size_t size)
{
atomic_add_z(&stats_cactive, size);
}
JEMALLOC_INLINE void
stats_cactive_sub(size_t size)
{
atomic_sub_z(&stats_cactive, size);
}
#endif
#endif /* JEMALLOC_H_INLINES */
/******************************************************************************/
deps/jemalloc/include/jemalloc/internal/tcache.h
Vendored
+443
View File
@@ -0,0 +1,443 @@
/******************************************************************************/
#ifdef JEMALLOC_H_TYPES
typedef struct tcache_bin_info_s tcache_bin_info_t;
typedef struct tcache_bin_s tcache_bin_t;
typedef struct tcache_s tcache_t;
/*
* tcache pointers close to NULL are used to encode state information that is
* used for two purposes: preventing thread caching on a per thread basis and
* cleaning up during thread shutdown.
*/
#define TCACHE_STATE_DISABLED ((tcache_t *)(uintptr_t)1)
#define TCACHE_STATE_REINCARNATED ((tcache_t *)(uintptr_t)2)
#define TCACHE_STATE_PURGATORY ((tcache_t *)(uintptr_t)3)
#define TCACHE_STATE_MAX TCACHE_STATE_PURGATORY
/*
* Absolute maximum number of cache slots for each small bin in the thread
* cache. This is an additional constraint beyond that imposed as: twice the
* number of regions per run for this size class.
*
* This constant must be an even number.
*/
#define TCACHE_NSLOTS_SMALL_MAX 200
/* Number of cache slots for large size classes. */
#define TCACHE_NSLOTS_LARGE 20
/* (1U << opt_lg_tcache_max) is used to compute tcache_maxclass. */
#define LG_TCACHE_MAXCLASS_DEFAULT 15
/*
* TCACHE_GC_SWEEP is the approximate number of allocation events between
* full GC sweeps. Integer rounding may cause the actual number to be
* slightly higher, since GC is performed incrementally.
*/
#define TCACHE_GC_SWEEP 8192
/* Number of tcache allocation/deallocation events between incremental GCs. */
#define TCACHE_GC_INCR \
((TCACHE_GC_SWEEP / NBINS) + ((TCACHE_GC_SWEEP / NBINS == 0) ? 0 : 1))
#endif /* JEMALLOC_H_TYPES */
/******************************************************************************/
#ifdef JEMALLOC_H_STRUCTS
typedef enum {
tcache_enabled_false = 0, /* Enable cast to/from bool. */
tcache_enabled_true = 1,
tcache_enabled_default = 2
} tcache_enabled_t;
/*
* Read-only information associated with each element of tcache_t's tbins array
* is stored separately, mainly to reduce memory usage.
*/
struct tcache_bin_info_s {
unsigned ncached_max; /* Upper limit on ncached. */
};
struct tcache_bin_s {
tcache_bin_stats_t tstats;
int low_water; /* Min # cached since last GC. */
unsigned lg_fill_div; /* Fill (ncached_max >> lg_fill_div). */
unsigned ncached; /* # of cached objects. */
void **avail; /* Stack of available objects. */
};
struct tcache_s {
ql_elm(tcache_t) link; /* Used for aggregating stats. */
uint64_t prof_accumbytes;/* Cleared after arena_prof_accum() */
arena_t *arena; /* This thread's arena. */
unsigned ev_cnt; /* Event count since incremental GC. */
unsigned next_gc_bin; /* Next bin to GC. */
tcache_bin_t tbins[1]; /* Dynamically sized. */
/*
* The pointer stacks associated with tbins follow as a contiguous
* array. During tcache initialization, the avail pointer in each
* element of tbins is initialized to point to the proper offset within
* this array.
*/
};
#endif /* JEMALLOC_H_STRUCTS */
/******************************************************************************/
#ifdef JEMALLOC_H_EXTERNS
extern bool opt_tcache;
extern ssize_t opt_lg_tcache_max;
extern tcache_bin_info_t *tcache_bin_info;
/*
* Number of tcache bins. There are NBINS small-object bins, plus 0 or more
* large-object bins.
*/
extern size_t nhbins;
/* Maximum cached size class. */
extern size_t tcache_maxclass;
size_t tcache_salloc(const void *ptr);
void tcache_event_hard(tcache_t *tcache);
void *tcache_alloc_small_hard(tcache_t *tcache, tcache_bin_t *tbin,
size_t binind);
void tcache_bin_flush_small(tcache_bin_t *tbin, size_t binind, unsigned rem,
tcache_t *tcache);
void tcache_bin_flush_large(tcache_bin_t *tbin, size_t binind, unsigned rem,
tcache_t *tcache);
void tcache_arena_associate(tcache_t *tcache, arena_t *arena);
void tcache_arena_dissociate(tcache_t *tcache);
tcache_t *tcache_create(arena_t *arena);
void tcache_destroy(tcache_t *tcache);
void tcache_thread_cleanup(void *arg);
void tcache_stats_merge(tcache_t *tcache, arena_t *arena);
bool tcache_boot0(void);
bool tcache_boot1(void);
#endif /* JEMALLOC_H_EXTERNS */
/******************************************************************************/
#ifdef JEMALLOC_H_INLINES
#ifndef JEMALLOC_ENABLE_INLINE
malloc_tsd_protos(JEMALLOC_ATTR(unused), tcache, tcache_t *)
malloc_tsd_protos(JEMALLOC_ATTR(unused), tcache_enabled, tcache_enabled_t)
void tcache_event(tcache_t *tcache);
void tcache_flush(void);
bool tcache_enabled_get(void);
tcache_t *tcache_get(bool create);
void tcache_enabled_set(bool enabled);
void *tcache_alloc_easy(tcache_bin_t *tbin);
void *tcache_alloc_small(tcache_t *tcache, size_t size, bool zero);
void *tcache_alloc_large(tcache_t *tcache, size_t size, bool zero);
void tcache_dalloc_small(tcache_t *tcache, void *ptr, size_t binind);
void tcache_dalloc_large(tcache_t *tcache, void *ptr, size_t size);
#endif
#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_TCACHE_C_))
/* Map of thread-specific caches. */
malloc_tsd_externs(tcache, tcache_t *)
malloc_tsd_funcs(JEMALLOC_ALWAYS_INLINE, tcache, tcache_t *, NULL,
tcache_thread_cleanup)
/* Per thread flag that allows thread caches to be disabled. */
malloc_tsd_externs(tcache_enabled, tcache_enabled_t)
malloc_tsd_funcs(JEMALLOC_ALWAYS_INLINE, tcache_enabled, tcache_enabled_t,
tcache_enabled_default, malloc_tsd_no_cleanup)
JEMALLOC_INLINE void
tcache_flush(void)
{
tcache_t *tcache;
cassert(config_tcache);
tcache = *tcache_tsd_get();
if ((uintptr_t)tcache <= (uintptr_t)TCACHE_STATE_MAX)
return;
tcache_destroy(tcache);
tcache = NULL;
tcache_tsd_set(&tcache);
}
JEMALLOC_INLINE bool
tcache_enabled_get(void)
{
tcache_enabled_t tcache_enabled;
cassert(config_tcache);
tcache_enabled = *tcache_enabled_tsd_get();
if (tcache_enabled == tcache_enabled_default) {
tcache_enabled = (tcache_enabled_t)opt_tcache;
tcache_enabled_tsd_set(&tcache_enabled);
}
return ((bool)tcache_enabled);
}
JEMALLOC_INLINE void
tcache_enabled_set(bool enabled)
{
tcache_enabled_t tcache_enabled;
tcache_t *tcache;
cassert(config_tcache);
tcache_enabled = (tcache_enabled_t)enabled;
tcache_enabled_tsd_set(&tcache_enabled);
tcache = *tcache_tsd_get();
if (enabled) {
if (tcache == TCACHE_STATE_DISABLED) {
tcache = NULL;
tcache_tsd_set(&tcache);
}
} else /* disabled */ {
if (tcache > TCACHE_STATE_MAX) {
tcache_destroy(tcache);
tcache = NULL;
}
if (tcache == NULL) {
tcache = TCACHE_STATE_DISABLED;
tcache_tsd_set(&tcache);
}
}
}
JEMALLOC_ALWAYS_INLINE tcache_t *
tcache_get(bool create)
{
tcache_t *tcache;
if (config_tcache == false)
return (NULL);
if (config_lazy_lock && isthreaded == false)
return (NULL);
tcache = *tcache_tsd_get();
if ((uintptr_t)tcache <= (uintptr_t)TCACHE_STATE_MAX) {
if (tcache == TCACHE_STATE_DISABLED)
return (NULL);
if (tcache == NULL) {
if (create == false) {
/*
* Creating a tcache here would cause
* allocation as a side effect of free().
* Ordinarily that would be okay since
* tcache_create() failure is a soft failure
* that doesn't propagate. However, if TLS
* data are freed via free() as in glibc,
* subtle corruption could result from setting
* a TLS variable after its backing memory is
* freed.
*/
return (NULL);
}
if (tcache_enabled_get() == false) {
tcache_enabled_set(false); /* Memoize. */
return (NULL);
}
return (tcache_create(choose_arena(NULL)));
}
if (tcache == TCACHE_STATE_PURGATORY) {
/*
* Make a note that an allocator function was called
* after tcache_thread_cleanup() was called.
*/
tcache = TCACHE_STATE_REINCARNATED;
tcache_tsd_set(&tcache);
return (NULL);
}
if (tcache == TCACHE_STATE_REINCARNATED)
return (NULL);
not_reached();
}
return (tcache);
}
JEMALLOC_ALWAYS_INLINE void
tcache_event(tcache_t *tcache)
{
if (TCACHE_GC_INCR == 0)
return;
tcache->ev_cnt++;
assert(tcache->ev_cnt <= TCACHE_GC_INCR);
if (tcache->ev_cnt == TCACHE_GC_INCR)
tcache_event_hard(tcache);
}
JEMALLOC_ALWAYS_INLINE void *
tcache_alloc_easy(tcache_bin_t *tbin)
{
void *ret;
if (tbin->ncached == 0) {
tbin->low_water = -1;
return (NULL);
}
tbin->ncached--;
if ((int)tbin->ncached < tbin->low_water)
tbin->low_water = tbin->ncached;
ret = tbin->avail[tbin->ncached];
return (ret);
}
JEMALLOC_ALWAYS_INLINE void *
tcache_alloc_small(tcache_t *tcache, size_t size, bool zero)
{
void *ret;
size_t binind;
tcache_bin_t *tbin;
binind = SMALL_SIZE2BIN(size);
assert(binind < NBINS);
tbin = &tcache->tbins[binind];
size = arena_bin_info[binind].reg_size;
ret = tcache_alloc_easy(tbin);
if (ret == NULL) {
ret = tcache_alloc_small_hard(tcache, tbin, binind);
if (ret == NULL)
return (NULL);
}
assert(tcache_salloc(ret) == arena_bin_info[binind].reg_size);
if (zero == false) {
if (config_fill) {
if (opt_junk) {
arena_alloc_junk_small(ret,
&arena_bin_info[binind], false);
} else if (opt_zero)
memset(ret, 0, size);
}
VALGRIND_MAKE_MEM_UNDEFINED(ret, size);
} else {
if (config_fill && opt_junk) {
arena_alloc_junk_small(ret, &arena_bin_info[binind],
true);
}
VALGRIND_MAKE_MEM_UNDEFINED(ret, size);
memset(ret, 0, size);
}
if (config_stats)
tbin->tstats.nrequests++;
if (config_prof)
tcache->prof_accumbytes += arena_bin_info[binind].reg_size;
tcache_event(tcache);
return (ret);
}
JEMALLOC_ALWAYS_INLINE void *
tcache_alloc_large(tcache_t *tcache, size_t size, bool zero)
{
void *ret;
size_t binind;
tcache_bin_t *tbin;
size = PAGE_CEILING(size);
assert(size <= tcache_maxclass);
binind = NBINS + (size >> LG_PAGE) - 1;
assert(binind < nhbins);
tbin = &tcache->tbins[binind];
ret = tcache_alloc_easy(tbin);
if (ret == NULL) {
/*
* Only allocate one large object at a time, because it's quite
* expensive to create one and not use it.
*/
ret = arena_malloc_large(tcache->arena, size, zero);
if (ret == NULL)
return (NULL);
} else {
if (config_prof && prof_promote && size == PAGE) {
arena_chunk_t *chunk =
(arena_chunk_t *)CHUNK_ADDR2BASE(ret);
size_t pageind = (((uintptr_t)ret - (uintptr_t)chunk) >>
LG_PAGE);
arena_mapbits_large_binind_set(chunk, pageind,
BININD_INVALID);
}
if (zero == false) {
if (config_fill) {
if (opt_junk)
memset(ret, 0xa5, size);
else if (opt_zero)
memset(ret, 0, size);
}
VALGRIND_MAKE_MEM_UNDEFINED(ret, size);
} else {
VALGRIND_MAKE_MEM_UNDEFINED(ret, size);
memset(ret, 0, size);
}
if (config_stats)
tbin->tstats.nrequests++;
if (config_prof)
tcache->prof_accumbytes += size;
}
tcache_event(tcache);
return (ret);
}
JEMALLOC_ALWAYS_INLINE void
tcache_dalloc_small(tcache_t *tcache, void *ptr, size_t binind)
{
tcache_bin_t *tbin;
tcache_bin_info_t *tbin_info;
assert(tcache_salloc(ptr) <= SMALL_MAXCLASS);
if (config_fill && opt_junk)
arena_dalloc_junk_small(ptr, &arena_bin_info[binind]);
tbin = &tcache->tbins[binind];
tbin_info = &tcache_bin_info[binind];
if (tbin->ncached == tbin_info->ncached_max) {
tcache_bin_flush_small(tbin, binind, (tbin_info->ncached_max >>
1), tcache);
}
assert(tbin->ncached < tbin_info->ncached_max);
tbin->avail[tbin->ncached] = ptr;
tbin->ncached++;
tcache_event(tcache);
}
JEMALLOC_ALWAYS_INLINE void
tcache_dalloc_large(tcache_t *tcache, void *ptr, size_t size)
{
size_t binind;
tcache_bin_t *tbin;
tcache_bin_info_t *tbin_info;
assert((size & PAGE_MASK) == 0);
assert(tcache_salloc(ptr) > SMALL_MAXCLASS);
assert(tcache_salloc(ptr) <= tcache_maxclass);
binind = NBINS + (size >> LG_PAGE) - 1;
if (config_fill && opt_junk)
memset(ptr, 0x5a, size);
tbin = &tcache->tbins[binind];
tbin_info = &tcache_bin_info[binind];
if (tbin->ncached == tbin_info->ncached_max) {
tcache_bin_flush_large(tbin, binind, (tbin_info->ncached_max >>
1), tcache);
}
assert(tbin->ncached < tbin_info->ncached_max);
tbin->avail[tbin->ncached] = ptr;
tbin->ncached++;
tcache_event(tcache);
}
#endif
#endif /* JEMALLOC_H_INLINES */
/******************************************************************************/
deps/jemalloc/include/jemalloc/internal/tsd.h
Vendored
+434
View File
@@ -0,0 +1,434 @@
/******************************************************************************/
#ifdef JEMALLOC_H_TYPES
/* Maximum number of malloc_tsd users with cleanup functions. */
#define MALLOC_TSD_CLEANUPS_MAX 8
typedef bool (*malloc_tsd_cleanup_t)(void);
#if (!defined(JEMALLOC_MALLOC_THREAD_CLEANUP) && !defined(JEMALLOC_TLS) && \
!defined(_WIN32))
typedef struct tsd_init_block_s tsd_init_block_t;
typedef struct tsd_init_head_s tsd_init_head_t;
#endif
/*
* TLS/TSD-agnostic macro-based implementation of thread-specific data. There
* are four macros that support (at least) three use cases: file-private,
* library-private, and library-private inlined. Following is an example
* library-private tsd variable:
*
* In example.h:
* typedef struct {
* int x;
* int y;
* } example_t;
* #define EX_INITIALIZER JEMALLOC_CONCAT({0, 0})
* malloc_tsd_protos(, example, example_t *)
* malloc_tsd_externs(example, example_t *)
* In example.c:
* malloc_tsd_data(, example, example_t *, EX_INITIALIZER)
* malloc_tsd_funcs(, example, example_t *, EX_INITIALIZER,
* example_tsd_cleanup)
*
* The result is a set of generated functions, e.g.:
*
* bool example_tsd_boot(void) {...}
* example_t **example_tsd_get() {...}
* void example_tsd_set(example_t **val) {...}
*
* Note that all of the functions deal in terms of (a_type *) rather than
* (a_type) so that it is possible to support non-pointer types (unlike
* pthreads TSD). example_tsd_cleanup() is passed an (a_type *) pointer that is
* cast to (void *). This means that the cleanup function needs to cast *and*
* dereference the function argument, e.g.:
*
* void
* example_tsd_cleanup(void *arg)
* {
* example_t *example = *(example_t **)arg;
*
* [...]
* if ([want the cleanup function to be called again]) {
* example_tsd_set(&example);
* }
* }
*
* If example_tsd_set() is called within example_tsd_cleanup(), it will be
* called again. This is similar to how pthreads TSD destruction works, except
* that pthreads only calls the cleanup function again if the value was set to
* non-NULL.
*/
/* malloc_tsd_protos(). */
#define malloc_tsd_protos(a_attr, a_name, a_type) \
a_attr bool \
a_name##_tsd_boot(void); \
a_attr a_type * \
a_name##_tsd_get(void); \
a_attr void \
a_name##_tsd_set(a_type *val);
/* malloc_tsd_externs(). */
#ifdef JEMALLOC_MALLOC_THREAD_CLEANUP
#define malloc_tsd_externs(a_name, a_type) \
extern __thread a_type a_name##_tls; \
extern __thread bool a_name##_initialized; \
extern bool a_name##_booted;
#elif (defined(JEMALLOC_TLS))
#define malloc_tsd_externs(a_name, a_type) \
extern __thread a_type a_name##_tls; \
extern pthread_key_t a_name##_tsd; \
extern bool a_name##_booted;
#elif (defined(_WIN32))
#define malloc_tsd_externs(a_name, a_type) \
extern DWORD a_name##_tsd; \
extern bool a_name##_booted;
#else
#define malloc_tsd_externs(a_name, a_type) \
extern pthread_key_t a_name##_tsd; \
extern tsd_init_head_t a_name##_tsd_init_head; \
extern bool a_name##_booted;
#endif
/* malloc_tsd_data(). */
#ifdef JEMALLOC_MALLOC_THREAD_CLEANUP
#define malloc_tsd_data(a_attr, a_name, a_type, a_initializer) \
a_attr __thread a_type JEMALLOC_TLS_MODEL \
a_name##_tls = a_initializer; \
a_attr __thread bool JEMALLOC_TLS_MODEL \
a_name##_initialized = false; \
a_attr bool a_name##_booted = false;
#elif (defined(JEMALLOC_TLS))
#define malloc_tsd_data(a_attr, a_name, a_type, a_initializer) \
a_attr __thread a_type JEMALLOC_TLS_MODEL \
a_name##_tls = a_initializer; \
a_attr pthread_key_t a_name##_tsd; \
a_attr bool a_name##_booted = false;
#elif (defined(_WIN32))
#define malloc_tsd_data(a_attr, a_name, a_type, a_initializer) \
a_attr DWORD a_name##_tsd; \
a_attr bool a_name##_booted = false;
#else
#define malloc_tsd_data(a_attr, a_name, a_type, a_initializer) \
a_attr pthread_key_t a_name##_tsd; \
a_attr tsd_init_head_t a_name##_tsd_init_head = { \
ql_head_initializer(blocks), \
MALLOC_MUTEX_INITIALIZER \
}; \
a_attr bool a_name##_booted = false;
#endif
/* malloc_tsd_funcs(). */
#ifdef JEMALLOC_MALLOC_THREAD_CLEANUP
#define malloc_tsd_funcs(a_attr, a_name, a_type, a_initializer, \
a_cleanup) \
/* Initialization/cleanup. */ \
a_attr bool \
a_name##_tsd_cleanup_wrapper(void) \
{ \
\
if (a_name##_initialized) { \
a_name##_initialized = false; \
a_cleanup(&a_name##_tls); \
} \
return (a_name##_initialized); \
} \
a_attr bool \
a_name##_tsd_boot(void) \
{ \
\
if (a_cleanup != malloc_tsd_no_cleanup) { \
malloc_tsd_cleanup_register( \
&a_name##_tsd_cleanup_wrapper); \
} \
a_name##_booted = true; \
return (false); \
} \
/* Get/set. */ \
a_attr a_type * \
a_name##_tsd_get(void) \
{ \
\
assert(a_name##_booted); \
return (&a_name##_tls); \
} \
a_attr void \
a_name##_tsd_set(a_type *val) \
{ \
\
assert(a_name##_booted); \
a_name##_tls = (*val); \
if (a_cleanup != malloc_tsd_no_cleanup) \
a_name##_initialized = true; \
}
#elif (defined(JEMALLOC_TLS))
#define malloc_tsd_funcs(a_attr, a_name, a_type, a_initializer, \
a_cleanup) \
/* Initialization/cleanup. */ \
a_attr bool \
a_name##_tsd_boot(void) \
{ \
\
if (a_cleanup != malloc_tsd_no_cleanup) { \
if (pthread_key_create(&a_name##_tsd, a_cleanup) != 0) \
return (true); \
} \
a_name##_booted = true; \
return (false); \
} \
/* Get/set. */ \
a_attr a_type * \
a_name##_tsd_get(void) \
{ \
\
assert(a_name##_booted); \
return (&a_name##_tls); \
} \
a_attr void \
a_name##_tsd_set(a_type *val) \
{ \
\
assert(a_name##_booted); \
a_name##_tls = (*val); \
if (a_cleanup != malloc_tsd_no_cleanup) { \
if (pthread_setspecific(a_name##_tsd, \
(void *)(&a_name##_tls))) { \
malloc_write("<jemalloc>: Error" \
" setting TSD for "#a_name"\n"); \
if (opt_abort) \
abort(); \
} \
} \
}
#elif (defined(_WIN32))
#define malloc_tsd_funcs(a_attr, a_name, a_type, a_initializer, \
a_cleanup) \
/* Data structure. */ \
typedef struct { \
bool initialized; \
a_type val; \
} a_name##_tsd_wrapper_t; \
/* Initialization/cleanup. */ \
a_attr bool \
a_name##_tsd_cleanup_wrapper(void) \
{ \
a_name##_tsd_wrapper_t *wrapper; \
\
wrapper = (a_name##_tsd_wrapper_t *) TlsGetValue(a_name##_tsd); \
if (wrapper == NULL) \
return (false); \
if (a_cleanup != malloc_tsd_no_cleanup && \
wrapper->initialized) { \
a_type val = wrapper->val; \
a_type tsd_static_data = a_initializer; \
wrapper->initialized = false; \
wrapper->val = tsd_static_data; \
a_cleanup(&val); \
if (wrapper->initialized) { \
/* Trigger another cleanup round. */ \
return (true); \
} \
} \
malloc_tsd_dalloc(wrapper); \
return (false); \
} \
a_attr bool \
a_name##_tsd_boot(void) \
{ \
\
a_name##_tsd = TlsAlloc(); \
if (a_name##_tsd == TLS_OUT_OF_INDEXES) \
return (true); \
if (a_cleanup != malloc_tsd_no_cleanup) { \
malloc_tsd_cleanup_register( \
&a_name##_tsd_cleanup_wrapper); \
} \
a_name##_booted = true; \
return (false); \
} \
/* Get/set. */ \
a_attr a_name##_tsd_wrapper_t * \
a_name##_tsd_get_wrapper(void) \
{ \
a_name##_tsd_wrapper_t *wrapper = (a_name##_tsd_wrapper_t *) \
TlsGetValue(a_name##_tsd); \
\
if (wrapper == NULL) { \
wrapper = (a_name##_tsd_wrapper_t *) \
malloc_tsd_malloc(sizeof(a_name##_tsd_wrapper_t)); \
if (wrapper == NULL) { \
malloc_write("<jemalloc>: Error allocating" \
" TSD for "#a_name"\n"); \
abort(); \
} else { \
static a_type tsd_static_data = a_initializer; \
wrapper->initialized = false; \
wrapper->val = tsd_static_data; \
} \
if (!TlsSetValue(a_name##_tsd, (void *)wrapper)) { \
malloc_write("<jemalloc>: Error setting" \
" TSD for "#a_name"\n"); \
abort(); \
} \
} \
return (wrapper); \
} \
a_attr a_type * \
a_name##_tsd_get(void) \
{ \
a_name##_tsd_wrapper_t *wrapper; \
\
assert(a_name##_booted); \
wrapper = a_name##_tsd_get_wrapper(); \
return (&wrapper->val); \
} \
a_attr void \
a_name##_tsd_set(a_type *val) \
{ \
a_name##_tsd_wrapper_t *wrapper; \
\
assert(a_name##_booted); \
wrapper = a_name##_tsd_get_wrapper(); \
wrapper->val = *(val); \
if (a_cleanup != malloc_tsd_no_cleanup) \
wrapper->initialized = true; \
}
#else
#define malloc_tsd_funcs(a_attr, a_name, a_type, a_initializer, \
a_cleanup) \
/* Data structure. */ \
typedef struct { \
bool initialized; \
a_type val; \
} a_name##_tsd_wrapper_t; \
/* Initialization/cleanup. */ \
a_attr void \
a_name##_tsd_cleanup_wrapper(void *arg) \
{ \
a_name##_tsd_wrapper_t *wrapper = (a_name##_tsd_wrapper_t *)arg;\
\
if (a_cleanup != malloc_tsd_no_cleanup && \
wrapper->initialized) { \
wrapper->initialized = false; \
a_cleanup(&wrapper->val); \
if (wrapper->initialized) { \
/* Trigger another cleanup round. */ \
if (pthread_setspecific(a_name##_tsd, \
(void *)wrapper)) { \
malloc_write("<jemalloc>: Error" \
" setting TSD for "#a_name"\n"); \
if (opt_abort) \
abort(); \
} \
return; \
} \
} \
malloc_tsd_dalloc(wrapper); \
} \
a_attr bool \
a_name##_tsd_boot(void) \
{ \
\
if (pthread_key_create(&a_name##_tsd, \
a_name##_tsd_cleanup_wrapper) != 0) \
return (true); \
a_name##_booted = true; \
return (false); \
} \
/* Get/set. */ \
a_attr a_name##_tsd_wrapper_t * \
a_name##_tsd_get_wrapper(void) \
{ \
a_name##_tsd_wrapper_t *wrapper = (a_name##_tsd_wrapper_t *) \
pthread_getspecific(a_name##_tsd); \
\
if (wrapper == NULL) { \
tsd_init_block_t block; \
wrapper = tsd_init_check_recursion( \
&a_name##_tsd_init_head, &block); \
if (wrapper) \
return (wrapper); \
wrapper = (a_name##_tsd_wrapper_t *) \
malloc_tsd_malloc(sizeof(a_name##_tsd_wrapper_t)); \
block.data = wrapper; \
if (wrapper == NULL) { \
malloc_write("<jemalloc>: Error allocating" \
" TSD for "#a_name"\n"); \
abort(); \
} else { \
static a_type tsd_static_data = a_initializer; \
wrapper->initialized = false; \
wrapper->val = tsd_static_data; \
} \
if (pthread_setspecific(a_name##_tsd, \
(void *)wrapper)) { \
malloc_write("<jemalloc>: Error setting" \
" TSD for "#a_name"\n"); \
abort(); \
} \
tsd_init_finish(&a_name##_tsd_init_head, &block); \
} \
return (wrapper); \
} \
a_attr a_type * \
a_name##_tsd_get(void) \
{ \
a_name##_tsd_wrapper_t *wrapper; \
\
assert(a_name##_booted); \
wrapper = a_name##_tsd_get_wrapper(); \
return (&wrapper->val); \
} \
a_attr void \
a_name##_tsd_set(a_type *val) \
{ \
a_name##_tsd_wrapper_t *wrapper; \
\
assert(a_name##_booted); \
wrapper = a_name##_tsd_get_wrapper(); \
wrapper->val = *(val); \
if (a_cleanup != malloc_tsd_no_cleanup) \
wrapper->initialized = true; \
}
#endif
#endif /* JEMALLOC_H_TYPES */
/******************************************************************************/
#ifdef JEMALLOC_H_STRUCTS
#if (!defined(JEMALLOC_MALLOC_THREAD_CLEANUP) && !defined(JEMALLOC_TLS) && \
!defined(_WIN32))
struct tsd_init_block_s {
ql_elm(tsd_init_block_t) link;
pthread_t thread;
void *data;
};
struct tsd_init_head_s {
ql_head(tsd_init_block_t) blocks;
malloc_mutex_t lock;
};
#endif
#endif /* JEMALLOC_H_STRUCTS */
/******************************************************************************/
#ifdef JEMALLOC_H_EXTERNS
void *malloc_tsd_malloc(size_t size);
void malloc_tsd_dalloc(void *wrapper);
void malloc_tsd_no_cleanup(void *);
void malloc_tsd_cleanup_register(bool (*f)(void));
void malloc_tsd_boot(void);
#if (!defined(JEMALLOC_MALLOC_THREAD_CLEANUP) && !defined(JEMALLOC_TLS) && \
!defined(_WIN32))
void *tsd_init_check_recursion(tsd_init_head_t *head,
tsd_init_block_t *block);
void tsd_init_finish(tsd_init_head_t *head, tsd_init_block_t *block);
#endif
#endif /* JEMALLOC_H_EXTERNS */
/******************************************************************************/
#ifdef JEMALLOC_H_INLINES
#endif /* JEMALLOC_H_INLINES */
/******************************************************************************/
deps/jemalloc/include/jemalloc/internal/util.h
Vendored
+162
View File
@@ -0,0 +1,162 @@
/******************************************************************************/
#ifdef JEMALLOC_H_TYPES
/* Size of stack-allocated buffer passed to buferror(). */
#define BUFERROR_BUF 64
/*
* Size of stack-allocated buffer used by malloc_{,v,vc}printf(). This must be
* large enough for all possible uses within jemalloc.
*/
#define MALLOC_PRINTF_BUFSIZE 4096
/*
* Wrap a cpp argument that contains commas such that it isn't broken up into
* multiple arguments.
*/
#define JEMALLOC_ARG_CONCAT(...) __VA_ARGS__
/*
* Silence compiler warnings due to uninitialized values. This is used
* wherever the compiler fails to recognize that the variable is never used
* uninitialized.
*/
#ifdef JEMALLOC_CC_SILENCE
# define JEMALLOC_CC_SILENCE_INIT(v) = v
#else
# define JEMALLOC_CC_SILENCE_INIT(v)
#endif
/*
* Define a custom assert() in order to reduce the chances of deadlock during
* assertion failure.
*/
#ifndef assert
#define assert(e) do { \
if (config_debug && !(e)) { \
malloc_printf( \
"<jemalloc>: %s:%d: Failed assertion: \"%s\"\n", \
__FILE__, __LINE__, #e); \
abort(); \
} \
} while (0)
#endif
#ifndef not_reached
#define not_reached() do { \
if (config_debug) { \
malloc_printf( \
"<jemalloc>: %s:%d: Unreachable code reached\n", \
__FILE__, __LINE__); \
abort(); \
} \
} while (0)
#endif
#ifndef not_implemented
#define not_implemented() do { \
if (config_debug) { \
malloc_printf("<jemalloc>: %s:%d: Not implemented\n", \
__FILE__, __LINE__); \
abort(); \
} \
} while (0)
#endif
#ifndef assert_not_implemented
#define assert_not_implemented(e) do { \
if (config_debug && !(e)) \
not_implemented(); \
} while (0)
#endif
/* Use to assert a particular configuration, e.g., cassert(config_debug). */
#define cassert(c) do { \
if ((c) == false) \
not_reached(); \
} while (0)
#endif /* JEMALLOC_H_TYPES */
/******************************************************************************/
#ifdef JEMALLOC_H_STRUCTS
#endif /* JEMALLOC_H_STRUCTS */
/******************************************************************************/
#ifdef JEMALLOC_H_EXTERNS
int buferror(int err, char *buf, size_t buflen);
uintmax_t malloc_strtoumax(const char *restrict nptr,
char **restrict endptr, int base);
void malloc_write(const char *s);
/*
* malloc_vsnprintf() supports a subset of snprintf(3) that avoids floating
* point math.
*/
int malloc_vsnprintf(char *str, size_t size, const char *format,
va_list ap);
int malloc_snprintf(char *str, size_t size, const char *format, ...)
JEMALLOC_ATTR(format(printf, 3, 4));
void malloc_vcprintf(void (*write_cb)(void *, const char *), void *cbopaque,
const char *format, va_list ap);
void malloc_cprintf(void (*write)(void *, const char *), void *cbopaque,
const char *format, ...) JEMALLOC_ATTR(format(printf, 3, 4));
void malloc_printf(const char *format, ...)
JEMALLOC_ATTR(format(printf, 1, 2));
#endif /* JEMALLOC_H_EXTERNS */
/******************************************************************************/
#ifdef JEMALLOC_H_INLINES
#ifndef JEMALLOC_ENABLE_INLINE
size_t pow2_ceil(size_t x);
void set_errno(int errnum);
int get_errno(void);
#endif
#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_UTIL_C_))
/* Compute the smallest power of 2 that is >= x. */
JEMALLOC_INLINE size_t
pow2_ceil(size_t x)
{
x--;
x |= x >> 1;
x |= x >> 2;
x |= x >> 4;
x |= x >> 8;
x |= x >> 16;
#if (LG_SIZEOF_PTR == 3)
x |= x >> 32;
#endif
x++;
return (x);
}
/* Sets error code */
JEMALLOC_INLINE void
set_errno(int errnum)
{
#ifdef _WIN32
SetLastError(errnum);
#else
errno = errnum;
#endif
}
/* Get last error code */
JEMALLOC_INLINE int
get_errno(void)
{
#ifdef _WIN32
return (GetLastError());
#else
return (errno);
#endif
}
#endif
#endif /* JEMALLOC_H_INLINES */
/******************************************************************************/
deps/jemalloc/include/jemalloc/jemalloc.sh
Vendored Executable
+28
View File
@@ -0,0 +1,28 @@
#!/bin/sh
objroot=$1
cat <<EOF
#ifndef JEMALLOC_H_
#define JEMALLOC_H_
#ifdef __cplusplus
extern "C" {
#endif
EOF
for hdr in jemalloc_defs.h jemalloc_rename.h jemalloc_macros.h \
jemalloc_protos.h jemalloc_mangle.h ; do
cat "${objroot}include/jemalloc/${hdr}" \
| grep -v 'Generated from .* by configure\.' \
| sed -e 's/^#define /#define /g' \
| sed -e 's/ $//g'
echo
done
cat <<EOF
#ifdef __cplusplus
};
#endif
#endif /* JEMALLOC_H_ */
EOF
deps/jemalloc/include/jemalloc/jemalloc_defs.h.in
Vendored
+24
View File
@@ -0,0 +1,24 @@
/* Defined if __attribute__((...)) syntax is supported. */
#undef JEMALLOC_HAVE_ATTR
/* Support the experimental API. */
#undef JEMALLOC_EXPERIMENTAL
/*
* Define overrides for non-standard allocator-related functions if they are
* present on the system.
*/
#undef JEMALLOC_OVERRIDE_MEMALIGN
#undef JEMALLOC_OVERRIDE_VALLOC
/*
* At least Linux omits the "const" in:
*
* size_t malloc_usable_size(const void *ptr);
*
* Match the operating system's prototype.
*/
#undef JEMALLOC_USABLE_SIZE_CONST
/* sizeof(void *) == 2^LG_SIZEOF_PTR. */
#undef LG_SIZEOF_PTR
deps/jemalloc/include/jemalloc/jemalloc_macros.h.in
Vendored
+61
View File
@@ -0,0 +1,61 @@
#include <limits.h>
#include <strings.h>
#define JEMALLOC_VERSION "@jemalloc_version@"
#define JEMALLOC_VERSION_MAJOR @jemalloc_version_major@
#define JEMALLOC_VERSION_MINOR @jemalloc_version_minor@
#define JEMALLOC_VERSION_BUGFIX @jemalloc_version_bugfix@
#define JEMALLOC_VERSION_NREV @jemalloc_version_nrev@
#define JEMALLOC_VERSION_GID "@jemalloc_version_gid@"
# define MALLOCX_LG_ALIGN(la) (la)
# if LG_SIZEOF_PTR == 2
# define MALLOCX_ALIGN(a) (ffs(a)-1)
# else
# define MALLOCX_ALIGN(a) \
((a < (size_t)INT_MAX) ? ffs(a)-1 : ffs(a>>32)+31)
# endif
# define MALLOCX_ZERO ((int)0x40)
/* Bias arena index bits so that 0 encodes "MALLOCX_ARENA() unspecified". */
# define MALLOCX_ARENA(a) ((int)(((a)+1) << 8))
#ifdef JEMALLOC_EXPERIMENTAL
# define ALLOCM_LG_ALIGN(la) (la)
# if LG_SIZEOF_PTR == 2
# define ALLOCM_ALIGN(a) (ffs(a)-1)
# else
# define ALLOCM_ALIGN(a) \
((a < (size_t)INT_MAX) ? ffs(a)-1 : ffs(a>>32)+31)
# endif
# define ALLOCM_ZERO ((int)0x40)
# define ALLOCM_NO_MOVE ((int)0x80)
/* Bias arena index bits so that 0 encodes "ALLOCM_ARENA() unspecified". */
# define ALLOCM_ARENA(a) ((int)(((a)+1) << 8))
# define ALLOCM_SUCCESS 0
# define ALLOCM_ERR_OOM 1
# define ALLOCM_ERR_NOT_MOVED 2
#endif
#ifdef JEMALLOC_HAVE_ATTR
# define JEMALLOC_ATTR(s) __attribute__((s))
# define JEMALLOC_EXPORT JEMALLOC_ATTR(visibility("default"))
# define JEMALLOC_ALIGNED(s) JEMALLOC_ATTR(aligned(s))
# define JEMALLOC_SECTION(s) JEMALLOC_ATTR(section(s))
# define JEMALLOC_NOINLINE JEMALLOC_ATTR(noinline)
#elif _MSC_VER
# define JEMALLOC_ATTR(s)
# ifdef DLLEXPORT
# define JEMALLOC_EXPORT __declspec(dllexport)
# else
# define JEMALLOC_EXPORT __declspec(dllimport)
# endif
# define JEMALLOC_ALIGNED(s) __declspec(align(s))
# define JEMALLOC_SECTION(s) __declspec(allocate(s))
# define JEMALLOC_NOINLINE __declspec(noinline)
#else
# define JEMALLOC_ATTR(s)
# define JEMALLOC_EXPORT
# define JEMALLOC_ALIGNED(s)
# define JEMALLOC_SECTION(s)
# define JEMALLOC_NOINLINE
#endif
deps/jemalloc/include/jemalloc/jemalloc_mangle.sh
Vendored Executable
+45
View File
@@ -0,0 +1,45 @@
#!/bin/sh
public_symbols_txt=$1
symbol_prefix=$2
cat <<EOF
/*
* By default application code must explicitly refer to mangled symbol names,
* so that it is possible to use jemalloc in conjunction with another allocator
* in the same application. Define JEMALLOC_MANGLE in order to cause automatic
* name mangling that matches the API prefixing that happened as a result of
* --with-mangling and/or --with-jemalloc-prefix configuration settings.
*/
#ifdef JEMALLOC_MANGLE
# ifndef JEMALLOC_NO_DEMANGLE
# define JEMALLOC_NO_DEMANGLE
# endif
EOF
for nm in `cat ${public_symbols_txt}` ; do
n=`echo ${nm} |tr ':' ' ' |awk '{print $1}'`
echo "# define ${n} ${symbol_prefix}${n}"
done
cat <<EOF
#endif
/*
* The ${symbol_prefix}* macros can be used as stable alternative names for the
* public jemalloc API if JEMALLOC_NO_DEMANGLE is defined. This is primarily
* meant for use in jemalloc itself, but it can be used by application code to
* provide isolation from the name mangling specified via --with-mangling
* and/or --with-jemalloc-prefix.
*/
#ifndef JEMALLOC_NO_DEMANGLE
EOF
for nm in `cat ${public_symbols_txt}` ; do
n=`echo ${nm} |tr ':' ' ' |awk '{print $1}'`
echo "# undef ${symbol_prefix}${n}"
done
cat <<EOF
#endif
EOF
deps/jemalloc/include/jemalloc/jemalloc_protos.h.in
Vendored
+58
View File
@@ -0,0 +1,58 @@
/*
* The @je_@ prefix on the following public symbol declarations is an artifact
* of namespace management, and should be omitted in application code unless
* JEMALLOC_NO_DEMANGLE is defined (see jemalloc_mangle@install_suffix@.h).
*/
extern JEMALLOC_EXPORT const char *@je_@malloc_conf;
extern JEMALLOC_EXPORT void (*@je_@malloc_message)(void *cbopaque,
const char *s);
JEMALLOC_EXPORT void *@je_@malloc(size_t size) JEMALLOC_ATTR(malloc);
JEMALLOC_EXPORT void *@je_@calloc(size_t num, size_t size)
JEMALLOC_ATTR(malloc);
JEMALLOC_EXPORT int @je_@posix_memalign(void **memptr, size_t alignment,
size_t size) JEMALLOC_ATTR(nonnull(1));
JEMALLOC_EXPORT void *@je_@aligned_alloc(size_t alignment, size_t size)
JEMALLOC_ATTR(malloc);
JEMALLOC_EXPORT void *@je_@realloc(void *ptr, size_t size);
JEMALLOC_EXPORT void @je_@free(void *ptr);
JEMALLOC_EXPORT void *@je_@mallocx(size_t size, int flags);
JEMALLOC_EXPORT void *@je_@rallocx(void *ptr, size_t size, int flags);
JEMALLOC_EXPORT size_t @je_@xallocx(void *ptr, size_t size, size_t extra,
int flags);
JEMALLOC_EXPORT size_t @je_@sallocx(const void *ptr, int flags);
JEMALLOC_EXPORT void @je_@dallocx(void *ptr, int flags);
JEMALLOC_EXPORT size_t @je_@nallocx(size_t size, int flags);
JEMALLOC_EXPORT int @je_@mallctl(const char *name, void *oldp,
size_t *oldlenp, void *newp, size_t newlen);
JEMALLOC_EXPORT int @je_@mallctlnametomib(const char *name, size_t *mibp,
size_t *miblenp);
JEMALLOC_EXPORT int @je_@mallctlbymib(const size_t *mib, size_t miblen,
void *oldp, size_t *oldlenp, void *newp, size_t newlen);
JEMALLOC_EXPORT void @je_@malloc_stats_print(void (*write_cb)(void *,
const char *), void *@je_@cbopaque, const char *opts);
JEMALLOC_EXPORT size_t @je_@malloc_usable_size(
JEMALLOC_USABLE_SIZE_CONST void *ptr);
#ifdef JEMALLOC_OVERRIDE_MEMALIGN
JEMALLOC_EXPORT void * @je_@memalign(size_t alignment, size_t size)
JEMALLOC_ATTR(malloc);
#endif
#ifdef JEMALLOC_OVERRIDE_VALLOC
JEMALLOC_EXPORT void * @je_@valloc(size_t size) JEMALLOC_ATTR(malloc);
#endif
#ifdef JEMALLOC_EXPERIMENTAL
JEMALLOC_EXPORT int @je_@allocm(void **ptr, size_t *rsize, size_t size,
int flags) JEMALLOC_ATTR(nonnull(1));
JEMALLOC_EXPORT int @je_@rallocm(void **ptr, size_t *rsize, size_t size,
size_t extra, int flags) JEMALLOC_ATTR(nonnull(1));
JEMALLOC_EXPORT int @je_@sallocm(const void *ptr, size_t *rsize, int flags)
JEMALLOC_ATTR(nonnull(1));
JEMALLOC_EXPORT int @je_@dallocm(void *ptr, int flags)
JEMALLOC_ATTR(nonnull(1));
JEMALLOC_EXPORT int @je_@nallocm(size_t *rsize, size_t size, int flags);
#endif
deps/jemalloc/include/jemalloc/jemalloc_rename.sh
Vendored Executable
+22
View File
@@ -0,0 +1,22 @@
#!/bin/sh
public_symbols_txt=$1
cat <<EOF
/*
* Name mangling for public symbols is controlled by --with-mangling and
* --with-jemalloc-prefix. With default settings the je_ prefix is stripped by
* these macro definitions.
*/
#ifndef JEMALLOC_NO_RENAME
EOF
for nm in `cat ${public_symbols_txt}` ; do
n=`echo ${nm} |tr ':' ' ' |awk '{print $1}'`
m=`echo ${nm} |tr ':' ' ' |awk '{print $2}'`
echo "# define je_${n} ${m}"
done
cat <<EOF
#endif
EOF
deps/jemalloc/include/msvc_compat/inttypes.h
Vendored
+313
View File
@@ -0,0 +1,313 @@
// ISO C9x compliant inttypes.h for Microsoft Visual Studio
// Based on ISO/IEC 9899:TC2 Committee draft (May 6, 2005) WG14/N1124
//
// Copyright (c) 2006 Alexander Chemeris
//
// Redistribution and use in source and binary forms, with or without
// modification, are permitted provided that the following conditions are met:
//
// 1. Redistributions of source code must retain the above copyright notice,
// this list of conditions and the following disclaimer.
//
// 2. Redistributions in binary form must reproduce the above copyright
// notice, this list of conditions and the following disclaimer in the
// documentation and/or other materials provided with the distribution.
//
// 3. The name of the author may be used to endorse or promote products
// derived from this software without specific prior written permission.
//
// THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR IMPLIED
// WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
// MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO
// EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
// PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
// OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
// WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
// OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
// ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
//
///////////////////////////////////////////////////////////////////////////////
#ifndef _MSC_VER // [
#error "Use this header only with Microsoft Visual C++ compilers!"
#endif // _MSC_VER ]
#ifndef _MSC_INTTYPES_H_ // [
#define _MSC_INTTYPES_H_
#if _MSC_VER > 1000
#pragma once
#endif
#include "stdint.h"
// 7.8 Format conversion of integer types
typedef struct {
intmax_t quot;
intmax_t rem;
} imaxdiv_t;
// 7.8.1 Macros for format specifiers
#if !defined(__cplusplus) || defined(__STDC_FORMAT_MACROS) // [ See footnote 185 at page 198
#ifdef _WIN64
# define __PRI64_PREFIX "l"
# define __PRIPTR_PREFIX "l"
#else
# define __PRI64_PREFIX "ll"
# define __PRIPTR_PREFIX
#endif
// The fprintf macros for signed integers are:
#define PRId8 "d"
#define PRIi8 "i"
#define PRIdLEAST8 "d"
#define PRIiLEAST8 "i"
#define PRIdFAST8 "d"
#define PRIiFAST8 "i"
#define PRId16 "hd"
#define PRIi16 "hi"
#define PRIdLEAST16 "hd"
#define PRIiLEAST16 "hi"
#define PRIdFAST16 "hd"
#define PRIiFAST16 "hi"
#define PRId32 "d"
#define PRIi32 "i"
#define PRIdLEAST32 "d"
#define PRIiLEAST32 "i"
#define PRIdFAST32 "d"
#define PRIiFAST32 "i"
#define PRId64 __PRI64_PREFIX "d"
#define PRIi64 __PRI64_PREFIX "i"
#define PRIdLEAST64 __PRI64_PREFIX "d"
#define PRIiLEAST64 __PRI64_PREFIX "i"
#define PRIdFAST64 __PRI64_PREFIX "d"
#define PRIiFAST64 __PRI64_PREFIX "i"
#define PRIdMAX __PRI64_PREFIX "d"
#define PRIiMAX __PRI64_PREFIX "i"
#define PRIdPTR __PRIPTR_PREFIX "d"
#define PRIiPTR __PRIPTR_PREFIX "i"
// The fprintf macros for unsigned integers are:
#define PRIo8 "o"
#define PRIu8 "u"
#define PRIx8 "x"
#define PRIX8 "X"
#define PRIoLEAST8 "o"
#define PRIuLEAST8 "u"
#define PRIxLEAST8 "x"
#define PRIXLEAST8 "X"
#define PRIoFAST8 "o"
#define PRIuFAST8 "u"
#define PRIxFAST8 "x"
#define PRIXFAST8 "X"
#define PRIo16 "ho"
#define PRIu16 "hu"
#define PRIx16 "hx"
#define PRIX16 "hX"
#define PRIoLEAST16 "ho"
#define PRIuLEAST16 "hu"
#define PRIxLEAST16 "hx"
#define PRIXLEAST16 "hX"
#define PRIoFAST16 "ho"
#define PRIuFAST16 "hu"
#define PRIxFAST16 "hx"
#define PRIXFAST16 "hX"
#define PRIo32 "o"
#define PRIu32 "u"
#define PRIx32 "x"
#define PRIX32 "X"
#define PRIoLEAST32 "o"
#define PRIuLEAST32 "u"
#define PRIxLEAST32 "x"
#define PRIXLEAST32 "X"
#define PRIoFAST32 "o"
#define PRIuFAST32 "u"
#define PRIxFAST32 "x"
#define PRIXFAST32 "X"
#define PRIo64 __PRI64_PREFIX "o"
#define PRIu64 __PRI64_PREFIX "u"
#define PRIx64 __PRI64_PREFIX "x"
#define PRIX64 __PRI64_PREFIX "X"
#define PRIoLEAST64 __PRI64_PREFIX "o"
#define PRIuLEAST64 __PRI64_PREFIX "u"
#define PRIxLEAST64 __PRI64_PREFIX "x"
#define PRIXLEAST64 __PRI64_PREFIX "X"
#define PRIoFAST64 __PRI64_PREFIX "o"
#define PRIuFAST64 __PRI64_PREFIX "u"
#define PRIxFAST64 __PRI64_PREFIX "x"
#define PRIXFAST64 __PRI64_PREFIX "X"
#define PRIoMAX __PRI64_PREFIX "o"
#define PRIuMAX __PRI64_PREFIX "u"
#define PRIxMAX __PRI64_PREFIX "x"
#define PRIXMAX __PRI64_PREFIX "X"
#define PRIoPTR __PRIPTR_PREFIX "o"
#define PRIuPTR __PRIPTR_PREFIX "u"
#define PRIxPTR __PRIPTR_PREFIX "x"
#define PRIXPTR __PRIPTR_PREFIX "X"
// The fscanf macros for signed integers are:
#define SCNd8 "d"
#define SCNi8 "i"
#define SCNdLEAST8 "d"
#define SCNiLEAST8 "i"
#define SCNdFAST8 "d"
#define SCNiFAST8 "i"
#define SCNd16 "hd"
#define SCNi16 "hi"
#define SCNdLEAST16 "hd"
#define SCNiLEAST16 "hi"
#define SCNdFAST16 "hd"
#define SCNiFAST16 "hi"
#define SCNd32 "ld"
#define SCNi32 "li"
#define SCNdLEAST32 "ld"
#define SCNiLEAST32 "li"
#define SCNdFAST32 "ld"
#define SCNiFAST32 "li"
#define SCNd64 "I64d"
#define SCNi64 "I64i"
#define SCNdLEAST64 "I64d"
#define SCNiLEAST64 "I64i"
#define SCNdFAST64 "I64d"
#define SCNiFAST64 "I64i"
#define SCNdMAX "I64d"
#define SCNiMAX "I64i"
#ifdef _WIN64 // [
# define SCNdPTR "I64d"
# define SCNiPTR "I64i"
#else // _WIN64 ][
# define SCNdPTR "ld"
# define SCNiPTR "li"
#endif // _WIN64 ]
// The fscanf macros for unsigned integers are:
#define SCNo8 "o"
#define SCNu8 "u"
#define SCNx8 "x"
#define SCNX8 "X"
#define SCNoLEAST8 "o"
#define SCNuLEAST8 "u"
#define SCNxLEAST8 "x"
#define SCNXLEAST8 "X"
#define SCNoFAST8 "o"
#define SCNuFAST8 "u"
#define SCNxFAST8 "x"
#define SCNXFAST8 "X"
#define SCNo16 "ho"
#define SCNu16 "hu"
#define SCNx16 "hx"
#define SCNX16 "hX"
#define SCNoLEAST16 "ho"
#define SCNuLEAST16 "hu"
#define SCNxLEAST16 "hx"
#define SCNXLEAST16 "hX"
#define SCNoFAST16 "ho"
#define SCNuFAST16 "hu"
#define SCNxFAST16 "hx"
#define SCNXFAST16 "hX"
#define SCNo32 "lo"
#define SCNu32 "lu"
#define SCNx32 "lx"
#define SCNX32 "lX"
#define SCNoLEAST32 "lo"
#define SCNuLEAST32 "lu"
#define SCNxLEAST32 "lx"
#define SCNXLEAST32 "lX"
#define SCNoFAST32 "lo"
#define SCNuFAST32 "lu"
#define SCNxFAST32 "lx"
#define SCNXFAST32 "lX"
#define SCNo64 "I64o"
#define SCNu64 "I64u"
#define SCNx64 "I64x"
#define SCNX64 "I64X"
#define SCNoLEAST64 "I64o"
#define SCNuLEAST64 "I64u"
#define SCNxLEAST64 "I64x"
#define SCNXLEAST64 "I64X"
#define SCNoFAST64 "I64o"
#define SCNuFAST64 "I64u"
#define SCNxFAST64 "I64x"
#define SCNXFAST64 "I64X"
#define SCNoMAX "I64o"
#define SCNuMAX "I64u"
#define SCNxMAX "I64x"
#define SCNXMAX "I64X"
#ifdef _WIN64 // [
# define SCNoPTR "I64o"
# define SCNuPTR "I64u"
# define SCNxPTR "I64x"
# define SCNXPTR "I64X"
#else // _WIN64 ][
# define SCNoPTR "lo"
# define SCNuPTR "lu"
# define SCNxPTR "lx"
# define SCNXPTR "lX"
#endif // _WIN64 ]
#endif // __STDC_FORMAT_MACROS ]
// 7.8.2 Functions for greatest-width integer types
// 7.8.2.1 The imaxabs function
#define imaxabs _abs64
// 7.8.2.2 The imaxdiv function
// This is modified version of div() function from Microsoft's div.c found
// in %MSVC.NET%\crt\src\div.c
#ifdef STATIC_IMAXDIV // [
static
#else // STATIC_IMAXDIV ][
_inline
#endif // STATIC_IMAXDIV ]
imaxdiv_t __cdecl imaxdiv(intmax_t numer, intmax_t denom)
{
imaxdiv_t result;
result.quot = numer / denom;
result.rem = numer % denom;
if (numer < 0 && result.rem > 0) {
// did division wrong; must fix up
++result.quot;
result.rem -= denom;
}
return result;
}
// 7.8.2.3 The strtoimax and strtoumax functions
#define strtoimax _strtoi64
#define strtoumax _strtoui64
// 7.8.2.4 The wcstoimax and wcstoumax functions
#define wcstoimax _wcstoi64
#define wcstoumax _wcstoui64
#endif // _MSC_INTTYPES_H_ ]
deps/jemalloc/include/msvc_compat/stdbool.h
Vendored
+16
View File
@@ -0,0 +1,16 @@
#ifndef stdbool_h
#define stdbool_h
#include <wtypes.h>
/* MSVC doesn't define _Bool or bool in C, but does have BOOL */
/* Note this doesn't pass autoconf's test because (bool) 0.5 != true */
typedef BOOL _Bool;
#define bool _Bool
#define true 1
#define false 0
#define __bool_true_false_are_defined 1
#endif /* stdbool_h */

Some files were not shown because too many files have changed in this diff Show More