| Message ID | 20230113184447.1707316-1-mjguzik@gmail.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | lockref: stop doing cpu_relax in the cmpxchg loop | expand |
> diff --git a/lib/lockref.c b/lib/lockref.c > index 45e93ece8ba0..2afe4c5d8919 100644 > --- a/lib/lockref.c > +++ b/lib/lockref.c > @@ -23,7 +23,6 @@ > } \ > if (!--retry) \ > break; \ > - cpu_relax(); \ > } \ > } while (0) The computer necrophiliacs at Debian and Gentoo seem determined to keep ia64 alive. So perhaps this should s/cpu_relax/soemt_relax/ where soemt_relax is a no-op everywhere except ia64, which can define it as cpu_relax. The ia64 case is quite painful if one thread on a core is spinning in that loop, while the other thread on the same core is the one that needs to run to update that value so that the cmpxchg can succeed. -Tony
On Fri, Jan 13, 2023 at 3:47 PM Luck, Tony <tony.luck@intel.com> wrote: > > The computer necrophiliacs at Debian and Gentoo seem determined > to keep ia64 alive. > > So perhaps this should s/cpu_relax/soemt_relax/ where soemt_relax > is a no-op everywhere except ia64, which can define it as cpu_relax. Heh. I already took your earlier "$ git rm -r arch/ia64" comment as an ack for not really caring about ia64. I suspect nobody will notice, and if ia64 is the only reason to do this, I really don't think it would be worth it. Linus
diff --git a/lib/lockref.c b/lib/lockref.c index 45e93ece8ba0..2afe4c5d8919 100644 --- a/lib/lockref.c +++ b/lib/lockref.c @@ -23,7 +23,6 @@ } \ if (!--retry) \ break; \ - cpu_relax(); \ } \ } while (0)
On the x86-64 architecture even a failing cmpxchg grants exclusive access to the cacheline, making it preferable to retry the failed op immediately instead of stalling with the pause instruction. To illustrate the impact, below are benchmark results obtained by running various will-it-scale tests on top of the 6.2-rc3 kernel and Cascade Lake (2 sockets * 24 cores * 2 threads) CPU. All results in ops/s. Note there is some variance in re-runs, but the code is consistently faster when contention is present. open3 ("Same file open/close"): proc stock no-pause 1 805603 814942 (+%1) 2 1054980 1054781 (-0%) 8 1544802 1822858 (+18%) 24 1191064 2199665 (+84%) 48 851582 1469860 (+72%) 96 609481 1427170 (+134%) fstat2 ("Same file fstat"): proc stock no-pause 1 3013872 3047636 (+1%) 2 4284687 4400421 (+2%) 8 3257721 5530156 (+69%) 24 2239819 5466127 (+144%) 48 1701072 5256609 (+209%) 96 1269157 6649326 (+423%) Additionally, a kernel with a private patch to help access() scalability: access2 ("Same file access"): proc stock patched patched+nopause 24 2378041 2005501 5370335 (-15% / +125%) That is, fixing the problems in access itself *reduces* scalability after the cacheline ping-pong only happens in lockref with the pause instruction. Note that fstat and access benchmarks are not currently integrated into will-it-scale, but interested parties can find them in pull requests to said project. Code at hand has a rather tortured history. First modification showed up in d472d9d98b463dd7 ("lockref: Relax in cmpxchg loop"), written with Itanium in mind. Later it got patched up to use an arch-dependent macro to stop doing it on s390 where it caused a significant regression. Said macro had undergone revisions and was ultimately eliminated later, going back to cpu_relax. While I intended to only remove cpu_relax for x86-64, I got the following comment from Linus: > I would actually prefer just removing it entirely and see if somebody > else hollers. You have the numbers to prove it hurts on real hardware, > and I don't think we have any numbers to the contrary. > So I think it's better to trust the numbers and remove it as a > failure, than say "let's just remove it on x86-64 and leave everybody >else with the potentially broken code" Additionally, Will Deacon (maintainer of the arm64 port, one of the architectures previously benchmarked): > So, from the arm64 side of the fence, I'm perfectly happy just removing > the cpu_relax() calls from lockref. As such, come back full circle in history and whack it altogether. Signed-off-by: Mateusz Guzik <mjguzik@gmail.com> --- lib/lockref.c | 1 - 1 file changed, 1 deletion(-)