mirrors/haproxy
1
0
Fork 0
mirror of https://git.haproxy.org/git/haproxy.git/ synced 2025-09-29 01:31:19 +02:00
Code Issues Projects Releases Wiki Activity

BUG/MINOR: tasks: only requeue a task if it was already in the queue

Browse Source 
Commit 0742c314c3 ("BUG/MEDIUM: tasks: Make sure we switch wait queues
in task_set_affinity().") had a slight side effect on expired timeouts,
which is that when used before a timeout is updated, it will cause an
existing task to be requeued earlier than its expected timeout when done
before being updated, resulting in the next poll wakup timeout too early
or even instantly if the previous wake up was done on a timeout. This is
visible in strace when health checks are enabled because there are two
poll calls, one of which has a short or zero delay. The correct solution
is to only requeue a task if it was already in the queue.

This can be backported to all branches having the fix above.
This commit is contained in:
Willy Tarreau 2019-12-11 09:11:58 +01:00
parent 4ac36d691a
commit 440d09b244
1 changed files with 5 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
include/proto/task.h
View File

@ -155,12 +155,14 @@ static inline void task_wakeup(struct task *t, unsigned int f)
/* change the thread affinity of a task to <thread_mask> */ /* change the thread affinity of a task to <thread_mask> */
static inline void task_set_affinity(struct task *t, unsigned long thread_mask) static inline void task_set_affinity(struct task *t, unsigned long thread_mask)
{ {
if (task_in_wq(t)) if (unlikely(task_in_wq(t))) {
task_unlink_wq(t); task_unlink_wq(t);
t->thread_mask = thread_mask; t->thread_mask = thread_mask;
if (t->expire != TICK_ETERNITY)
task_queue(t); task_queue(t);
} }
else
t->thread_mask = thread_mask;
}
/* /*
* Unlink the task from the wait queue, and possibly update the last_timer * Unlink the task from the wait queue, and possibly update the last_timer