Attempt to fix a rare crash in cluster tests. (#10265)

The theory is that a replica gets disconnected from within REPLCONF ACK,
so when we go up the stack, we'll crash when attempting to access
c->cmd->flags

(cherry picked from commit aa9beaca77)

src/replication.c

@@ -1224,6 +1224,9 @@ void updateSlavesWaitingBgsave(int bgsaveerr, int type) {
     int mincapa = -1;
     listIter li;
 
+    /* Note: there's a chance we got here from within the REPLCONF ACK command
+     * so we must avoid using freeClient, otherwise we'll crash on our way up. */
+
     listRewind(server.slaves,&li);
     while((ln = listNext(&li))) {
         client *slave = ln->value;
@@ -1236,7 +1239,7 @@ void updateSlavesWaitingBgsave(int bgsaveerr, int type) {
             struct redis_stat buf;
 
             if (bgsaveerr != C_OK) {
-                freeClient(slave);
+                freeClientAsync(slave);
                 serverLog(LL_WARNING,"SYNC failed. BGSAVE child returned an error");
                 continue;
             }
@@ -1281,7 +1284,7 @@ void updateSlavesWaitingBgsave(int bgsaveerr, int type) {
             } else {
                 if ((slave->repldbfd = open(server.rdb_filename,O_RDONLY)) == -1 ||
                     redis_fstat(slave->repldbfd,&buf) == -1) {
-                    freeClient(slave);
+                    freeClientAsync(slave);
                     serverLog(LL_WARNING,"SYNC failed. Can't open/stat DB after BGSAVE: %s", strerror(errno));
                     continue;
                 }
@@ -1293,7 +1296,7 @@ void updateSlavesWaitingBgsave(int bgsaveerr, int type) {
 
                 connSetWriteHandler(slave->conn,NULL);
                 if (connSetWriteHandler(slave->conn,sendBulkToSlave) == C_ERR) {
-                    freeClient(slave);
+                    freeClientAsync(slave);
                     continue;
                 }
             }