Re: [linus:master] [mm] c0bff412e6: stress-ng.clone.ops_per_sec -2.9% regression

[Mateusz Guzik]

On Wed, Aug 14, 2024 at 11:45 AM David Hildenbrand <david@xxxxxxxxxx> wrote:
>
> On 14.08.24 06:10, Mateusz Guzik wrote:
> > On Wed, Aug 14, 2024 at 5:02 AM Yin Fengwei <fengwei.yin@xxxxxxxxx> wrote:
> >>
> >> On 8/13/24 03:14, Mateusz Guzik wrote:
> >>> would you mind benchmarking the change which merely force-inlines _compund_page?
> >>>
> >>> https://lore.kernel.org/linux-mm/66c4fcc5-47f6-438c-a73a-3af6e19c3200@xxxxxxxxxx/
> >> This change can resolve the regression also:
> >
> > Great, thanks.
> >
> > David, I guess this means it would be fine to inline the entire thing
> > at least from this bench standpoint. Given that this is your idea I
> > guess you should do the needful(tm)? :)
>
> Testing
>
> diff --git a/include/linux/page-flags.h b/include/linux/page-flags.h
> index 5769fe6e4950..25e25b34f4a0 100644
> --- a/include/linux/page-flags.h
> +++ b/include/linux/page-flags.h
> @@ -235,7 +235,7 @@ static __always_inline int page_is_fake_head(const struct page *page)
>          return page_fixed_fake_head(page) != page;
>   }
>
> -static inline unsigned long _compound_head(const struct page *page)
> +static __always_inline unsigned long _compound_head(const struct page *page)
>   {
>          unsigned long head = READ_ONCE(page->compound_head);
>
>
> With a kernel-config based on something derived from Fedora
> config-6.8.9-100.fc38.x86_64 for convenience with
>
> CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP=y
>
> add/remove: 15/14 grow/shrink: 79/87 up/down: 12836/-13917 (-1081)
[snip]
> Total: Before=32786363, After=32785282, chg -0.00%

I guess there should be no opposition then?

Given that this is your patch I presume you are going to see this through.

I don't want any mention or cc on the patch, thanks for understanding :)

-- 
Mateusz Guzik <mjguzik gmail.com>