(cloudberry) 24/38: Remove palloc() call in ResWaitOnLock (#13701)

[ma...@apache.org]

This is an automated email from the ASF dual-hosted git repository.

maxyang pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/cloudberry.git

commit 73da0f32f3860cb9ca128221761bc49929586deb
Author: dh-cloud <60...@users.noreply.github.com>
AuthorDate: Wed Jul 20 09:02:52 2022 +0800

    Remove palloc() call in ResWaitOnLock (#13701)
    
    If palloc() failed in ResWaitOnLock due to OOM, it will trigger
    a PANIC "Waiting on lock already held" because partitionLock is
    not released correctly.
    
    The first attempt is to add a try/catch surrounding for this
    palloc() call, and release the partitionLock before rethrowing
    error. But it is not good as ResIncrementRemove is not called.
    
    So use local char new_status[160], length 160 is enough.
    
    ----
    
    Steps to reproduce:
    
    Add a fault at the palloc() position:
        SIMPLE_FAULT_INJECTOR("reslock_acquire_before_wait");
    
    ```sql
    -- superuser:
    
    -- test user session 1:
    > SELECT pg_sleep(120);
    
    -- test user session 2:
    > SELECT version();
    PANIC:  Waiting on lock already held! (lwlock.c:668)
    server closed the connection unexpectedly
    ```
---
 src/backend/utils/resscheduler/resqueue.c | 14 ++++++++------
 1 file changed, 8 insertions(+), 6 deletions(-)

diff --git a/src/backend/utils/resscheduler/resqueue.c b/src/backend/utils/resscheduler/resqueue.c
index fa3c985b89..7ee6afbe6f 100644
--- a/src/backend/utils/resscheduler/resqueue.c
+++ b/src/backend/utils/resscheduler/resqueue.c
@@ -1085,18 +1085,21 @@ ResWaitOnLock(LOCALLOCK *locallock, ResourceOwner owner, ResPortalIncrement *inc
 {
 	uint32		hashcode = locallock->hashcode;
 	LWLockId	partitionLock = LockHashPartitionLock(hashcode);
+	char		new_status[160];
 	const char *old_status;
-	char	   *new_status = NULL;
 	int			len;
 
 	/* Report change to waiting status */
 	if (update_process_title)
 	{
+		/* We should avoid using palloc() here */
 		old_status = get_real_act_ps_display(&len);
-		new_status = (char *) palloc(len + 8 + 1);
-		memcpy(new_status, old_status, len);
-		strcpy(new_status + len, " queuing");
-		set_ps_display(new_status);		/* truncate off " queuing" */
+		len = Min(len, sizeof(new_status) - 9);
+		snprintf(new_status, sizeof(new_status), "%.*s queuing",
+				 len, old_status);
+		set_ps_display(new_status);
+
+		/* Truncate off " queuing" */
 		new_status[len] = '\0';
 	}
 
@@ -1123,7 +1126,6 @@ ResWaitOnLock(LOCALLOCK *locallock, ResourceOwner owner, ResPortalIncrement *inc
 	if (update_process_title)
 	{
 		set_ps_display(new_status);
-		pfree(new_status);
 	}
 
 	return;


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@cloudberry.apache.org
For additional commands, e-mail: commits-help@cloudberry.apache.org