mirrors/haproxy
1
0
Fork 0
mirror of https://git.haproxy.org/git/haproxy.git/ synced 2025-09-20 13:21:29 +02:00
Code Issues Projects Releases Wiki Activity

OPTIM: sink: reduce contention on sink_announce_dropped()

Browse Source 
perf top shows that sink_announce_dropped() consumes most of the CPU
on a 128-thread x86 system. Digging further reveals that the atomic
fetch_or() on the dropped field used to detect the presence of another
thread is entirely responsible for this. Indeed, the compiler implements
it using a CAS that loops without relaxing and makes all threads wait
until they can synchronize on this one, only to discover later that
another thread is there and they need to give up.

Let's just replace this with a hand-crafted CAS loop that will detect
*before* attempting the CAS if another thread is there. Doing so
achieves the same goal without forcing threads to agree. With this
simple change, the sustained request rate on h1 with all traces on
bumped from 110k/s to 244k/s!

This should be backported to stable releases where it's often needed
to help debugging.
This commit is contained in:
Willy Tarreau 2025-09-18 08:38:34 +02:00
parent 361c227465
commit 4431e3bd26
1 changed files with 7 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
src/sink.c
View File

@ -298,11 +298,13 @@ int sink_announce_dropped(struct sink *sink, struct log_header hdr)
* another thread is already on them and we can just pass and * another thread is already on them and we can just pass and
* count another drop (hence add 2). * count another drop (hence add 2).
*/ */
dropped = HA_ATOMIC_FETCH_OR(&sink->ctx.dropped, 1); dropped = HA_ATOMIC_LOAD(&sink->ctx.dropped);
if (dropped & 1) { do {
/* another thread was already on it */ if (dropped & 1) {
goto leave; /* another thread was already on it */
} goto leave;
}
} while (!_HA_ATOMIC_CAS(&sink->ctx.dropped, &dropped, dropped | 1));
last_dropped = 0; last_dropped = 0;
dropped >>= 1; dropped >>= 1;