Re: [LKP] [mm] ac5b2c1891: vm-scalability.throughput -61.3% regression
From: David Rientjes
Date: Wed Dec 05 2018 - 17:12:41 EST
On Wed, 5 Dec 2018, Linus Torvalds wrote:
> > So ultimately we decided that the saner behavior that gives the least
> > risk of regression for the short term, until we can do something
> > better, was the one that is already applied upstream.
>
> You're ignoring the fact that people *did* report things regressed.
>
> That's the part I find unacceptable. You're saying "we picked
> something that minimized regressions".
>
> No it didn't. The regression is present and real, and is on a real
> load, not a benchmark.
>
> So that argument is clearly bogus.
>
> I'm going to revert the commit since people apparently seem to be
> ignoring this fundamental issue.
>
> Real workloads regressed. The regressions got reported. Ignoring that
> isn't acceptable.
>
Please allow me to prepare my v2 because it's not a clean revert due to
the follow-up 89c83fb539f9 ("mm, thp: consolidate THP gfp handling into
alloc_hugepage_direct_gfpmask") and will incorporate the feedback from
Michal to not change anything outside of the thp fault path.