From: Steven Rostedt (Google) <[email protected]>
Running my yearly branch profiler to see where likely/unlikely annotation
may be added or removed, I discovered this:
correct incorrect % Function File Line
------- --------- - -------- ---- ----
0 457918 100 page_try_dup_anon_rmap rmap.h 264
[..]
458021 0 0 page_try_dup_anon_rmap rmap.h 265
I thought it was interesting that line 264 of rmap.h had a 100% incorrect
annotation, but the line directly below it was 100% correct. Looking at the
code:
if (likely(!is_device_private_page(page) &&
unlikely(page_needs_cow_for_dma(vma, page))))
It didn't make sense. The "likely()" was around the entire if statement
(not just the "!is_device_private_page(page)"), which also included the
"unlikely()" portion of that if condition.
If the unlikely portion is unlikely to be true, that would make the entire
if condition unlikely to be true, so it made no sense at all to say the
entire if condition is true.
What is more likely to be likely is just the first part of the if statement
before the && operation. It's likely to be a misplaced parenthesis. And
after making the if condition broken into a likely() && unlikely(), both
now appear to be correct!
Cc: [email protected]
Fixes:fb3d824d1a46c ("mm/rmap: split page_dup_rmap() into page_dup_file_rmap() and page_try_dup_anon_rmap()")
Signed-off-by: Steven Rostedt (Google) <[email protected]>
---
diff --git a/include/linux/rmap.h b/include/linux/rmap.h
index b26fe858fd44..3c2fc291b071 100644
--- a/include/linux/rmap.h
+++ b/include/linux/rmap.h
@@ -261,8 +261,8 @@ static inline int page_try_dup_anon_rmap(struct page *page, bool compound,
* guarantee the pinned page won't be randomly replaced in the
* future on write faults.
*/
- if (likely(!is_device_private_page(page) &&
- unlikely(page_needs_cow_for_dma(vma, page))))
+ if (likely(!is_device_private_page(page)) &&
+ unlikely(page_needs_cow_for_dma(vma, page)))
return -EBUSY;
ClearPageAnonExclusive(page);
On 01.12.23 20:59, Steven Rostedt wrote:
> From: Steven Rostedt (Google) <[email protected]>
>
> Running my yearly branch profiler to see where likely/unlikely annotation
> may be added or removed, I discovered this:
>
> correct incorrect % Function File Line
> ------- --------- - -------- ---- ----
> 0 457918 100 page_try_dup_anon_rmap rmap.h 264
> [..]
> 458021 0 0 page_try_dup_anon_rmap rmap.h 265
>
That looks like a handy tool!
> I thought it was interesting that line 264 of rmap.h had a 100% incorrect
> annotation, but the line directly below it was 100% correct. Looking at the
> code:
>
> if (likely(!is_device_private_page(page) &&
> unlikely(page_needs_cow_for_dma(vma, page))))
>
> It didn't make sense. The "likely()" was around the entire if statement
> (not just the "!is_device_private_page(page)"), which also included the
> "unlikely()" portion of that if condition.
Yes, that was clearly misplaced.
>
> If the unlikely portion is unlikely to be true, that would make the entire
> if condition unlikely to be true, so it made no sense at all to say the
> entire if condition is true.
>
> What is more likely to be likely is just the first part of the if statement
> before the && operation. It's likely to be a misplaced parenthesis. And
> after making the if condition broken into a likely() && unlikely(), both
> now appear to be correct!
>
Reviewed-by: David Hildenbrand <[email protected]>
But
> Cc: [email protected]
stable, really? Why?
> Fixes:fb3d824d1a46c ("mm/rmap: split page_dup_rmap() into page_dup_file_rmap() and page_try_dup_anon_rmap()")
and does it even fix a real bug?
--
Cheers,
David / dhildenb
On Fri, 1 Dec 2023 21:06:22 +0100
David Hildenbrand <[email protected]> wrote:
> But
>
> > Cc: [email protected]
>
> stable, really? Why?
>
> > Fixes:fb3d824d1a46c ("mm/rmap: split page_dup_rmap() into page_dup_file_rmap() and page_try_dup_anon_rmap()")
>
> and does it even fix a real bug?
As a performance person, who measures likely and unlikely results (the
ftrace ring buffer was sped up by over 50% with strategically placed
likely/unlikely annotation). I find this to be a real bug, and something I
would want backported to the kernels we maintain in ChromeOS (which uses
upstream stable kernels).
-- Steve
On 01.12.23 21:15, Steven Rostedt wrote:
> On Fri, 1 Dec 2023 21:06:22 +0100
> David Hildenbrand <[email protected]> wrote:
>
>> But
>>
>>> Cc: [email protected]
>>
>> stable, really? Why?
>>
>>> Fixes:fb3d824d1a46c ("mm/rmap: split page_dup_rmap() into page_dup_file_rmap() and page_try_dup_anon_rmap()")
>>
>> and does it even fix a real bug?
>
> As a performance person, who measures likely and unlikely results (the
> ftrace ring buffer was sped up by over 50% with strategically placed
> likely/unlikely annotation). I find this to be a real bug, and something I
> would want backported to the kernels we maintain in ChromeOS (which uses
> upstream stable kernels).
Okay, I don't care that much about a "Fixes" annotation.
But I am *pretty* sure that this is not stable material, looking once
again at the stable rules.
Anyhow, thanks for catching this!
--
Cheers,
David / dhildenb
On 12/1/23 20:59, Steven Rostedt wrote:
> From: Steven Rostedt (Google) <[email protected]>
>
> Running my yearly branch profiler to see where likely/unlikely annotation
> may be added or removed, I discovered this:
>
> correct incorrect % Function File Line
> ------- --------- - -------- ---- ----
> 0 457918 100 page_try_dup_anon_rmap rmap.h 264
> [..]
> 458021 0 0 page_try_dup_anon_rmap rmap.h 265
>
> I thought it was interesting that line 264 of rmap.h had a 100% incorrect
> annotation, but the line directly below it was 100% correct. Looking at the
> code:
>
> if (likely(!is_device_private_page(page) &&
> unlikely(page_needs_cow_for_dma(vma, page))))
>
> It didn't make sense. The "likely()" was around the entire if statement
> (not just the "!is_device_private_page(page)"), which also included the
> "unlikely()" portion of that if condition.
>
> If the unlikely portion is unlikely to be true, that would make the entire
> if condition unlikely to be true, so it made no sense at all to say the
> entire if condition is true.
>
> What is more likely to be likely is just the first part of the if statement
> before the && operation. It's likely to be a misplaced parenthesis. And
> after making the if condition broken into a likely() && unlikely(), both
> now appear to be correct!
>
> Cc: [email protected]
> Fixes:fb3d824d1a46c ("mm/rmap: split page_dup_rmap() into page_dup_file_rmap() and page_try_dup_anon_rmap()")
> Signed-off-by: Steven Rostedt (Google) <[email protected]>
Acked-by: Vlastimil Babka <[email protected]>
Pragmatically speaking, stable maintainers haven't been following the stable
rules for a long time, and a commit with Fixes and without Cc: stable is
often backported on the assumption people forget Cc: stable, and "Fixes:"
implies there's a bug to fix, and it's good to have bugs fixed in stable...
We have (repeatedly...) had mm extempted from this and Cc: stable is
required, which is good. So if Steven thinks there are reasons to backport,
then I'd rather let him keep the Cc: stable, instead of this later becoming
an argument to question the mm extemption again :)
> ---
> diff --git a/include/linux/rmap.h b/include/linux/rmap.h
> index b26fe858fd44..3c2fc291b071 100644
> --- a/include/linux/rmap.h
> +++ b/include/linux/rmap.h
> @@ -261,8 +261,8 @@ static inline int page_try_dup_anon_rmap(struct page *page, bool compound,
> * guarantee the pinned page won't be randomly replaced in the
> * future on write faults.
> */
> - if (likely(!is_device_private_page(page) &&
> - unlikely(page_needs_cow_for_dma(vma, page))))
> + if (likely(!is_device_private_page(page)) &&
> + unlikely(page_needs_cow_for_dma(vma, page)))
> return -EBUSY;
>
> ClearPageAnonExclusive(page);