Re: mm: change huge_ptep_clear_flush() to return the original pte

Re: mm: change huge_ptep_clear_flush() to return the original pte

[Naresh Kamboju]

Following build failures noticed on arm64 on Linux mainline tree.

Which was detected and discussed at this email thread.
https://lore.kernel.org/all/20220512193855.4f6ce32f@canb.auug.org.au/T/

Reported-by: Linux Kernel Functional Testing <lkft@linaro.org>

Build errors:
arch/arm64/mm/hugetlbpage.c: In function 'huge_ptep_clear_flush':
arch/arm64/mm/hugetlbpage.c:515:16: error: implicit declaration of
function 'get_clear_flush'; did you mean 'ptep_clear_flush'?
[-Werror=implicit-function-declaration]
  515 |         return get_clear_flush(vma->vm_mm, addr, ptep, pgsize, ncontig);
      |                ^~~~~~~~~~~~~~~
      |                ptep_clear_flush
arch/arm64/mm/hugetlbpage.c:515:16: error: incompatible types when
returning type 'int' but 'pte_t' was expected
  515 |         return get_clear_flush(vma->vm_mm, addr, ptep, pgsize, ncontig);
      |                ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
arch/arm64/mm/hugetlbpage.c:516:1: error: control reaches end of
non-void function [-Werror=return-type]
  516 | }
      | ^
cc1: some warnings being treated as errors

metadata:
    git_repo: https://gitlab.com/Linaro/lkft/mirrors/torvalds/linux-mainline
    git_describe: v5.18-9901-g7f50d4dfe816
    git_sha: 7f50d4dfe816dd916a7cbf39039674825c2b388b
    git_short_log: 7f50d4dfe816 (\Merge tag 'for-linus' of
https://github.com/openrisc/linux\)
    target_arch: arm64
    toolchain: gcc-11


--
Linaro LKFT
https://lkft.linaro.org

Re: mm: change huge_ptep_clear_flush() to return the original pte

[Baolin Wang]

On 5/27/2022 10:56 AM, Naresh Kamboju wrote:
> Following build failures noticed on arm64 on Linux mainline tree.
> 
> Which was detected and discussed at this email thread.
> https://lore.kernel.org/all/20220512193855.4f6ce32f@canb.auug.org.au/T/
> 
> Reported-by: Linux Kernel Functional Testing <lkft@linaro.org>
> 
> Build errors:
> arch/arm64/mm/hugetlbpage.c: In function 'huge_ptep_clear_flush':
> arch/arm64/mm/hugetlbpage.c:515:16: error: implicit declaration of
> function 'get_clear_flush'; did you mean 'ptep_clear_flush'?
> [-Werror=implicit-function-declaration]
>    515 |         return get_clear_flush(vma->vm_mm, addr, ptep, pgsize, ncontig);
>        |                ^~~~~~~~~~~~~~~
>        |                ptep_clear_flush
> arch/arm64/mm/hugetlbpage.c:515:16: error: incompatible types when
> returning type 'int' but 'pte_t' was expected
>    515 |         return get_clear_flush(vma->vm_mm, addr, ptep, pgsize, ncontig);
>        |                ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> arch/arm64/mm/hugetlbpage.c:516:1: error: control reaches end of
> non-void function [-Werror=return-type]
>    516 | }
>        | ^
> cc1: some warnings being treated as errors
> 
> metadata:
>      git_repo: https://gitlab.com/Linaro/lkft/mirrors/torvalds/linux-mainline
>      git_describe: v5.18-9901-g7f50d4dfe816
>      git_sha: 7f50d4dfe816dd916a7cbf39039674825c2b388b
>      git_short_log: 7f50d4dfe816 (\Merge tag 'for-linus' of
> https://github.com/openrisc/linux\)
>      target_arch: arm64
>      toolchain: gcc-11

That is caused by commit:

00df1f1a133b ("mm: change huge_ptep_clear_flush() to return the original 
pte")

interacting with commit:

fb396bb459c1 ("arm64/hugetlb: Drop TLB flush from get_clear_flush()")

And Catalin has fixed the conflict with below changes [1], not sure why 
it is not merged.

Linus, should I send a proper patch to fix this conflict for you?

diff --git a/arch/arm64/mm/hugetlbpage.c b/arch/arm64/mm/hugetlbpage.c
index 30f5b76aabe9..9a999550df8e 100644
--- a/arch/arm64/mm/hugetlbpage.c
+++ b/arch/arm64/mm/hugetlbpage.c
@@ -485,12 +485,15 @@ pte_t huge_ptep_clear_flush(struct vm_area_struct 
*vma,
  {
  	size_t pgsize;
  	int ncontig;
+	pte_t orig_pte;

  	if (!pte_cont(READ_ONCE(*ptep)))
  		return ptep_clear_flush(vma, addr, ptep);

  	ncontig = find_num_contig(vma->vm_mm, addr, ptep, &pgsize);
-	return get_clear_contig(vma->vm_mm, addr, ptep, pgsize, ncontig);
+	orig_pte = get_clear_contig(vma->vm_mm, addr, ptep, pgsize, ncontig);
+	flush_tlb_range(vma, addr, addr + pgsize * ncontig);
+	return orig_pte;
  }

[1] https://lore.kernel.org/all/YnzqffV7STYS24Yn@arm.com/

Re: mm: change huge_ptep_clear_flush() to return the original pte

[Linus Torvalds]

On Thu, May 26, 2022 at 8:36 PM Baolin Wang
<baolin.wang@linux.alibaba.com> wrote:
>
> That is caused by commit:
>
> 00df1f1a133b ("mm: change huge_ptep_clear_flush() to return the original
> pte")
>
> interacting with commit:
>
> fb396bb459c1 ("arm64/hugetlb: Drop TLB flush from get_clear_flush()")
>
> And Catalin has fixed the conflict with below changes [1], not sure why
> it is not merged.

Argh. Nobody informed me (or if they did, I clearly missed it).

> Linus, should I send a proper patch to fix this conflict for you?

Yeah, somebody please send me a proper tested fix-up patch with commit
message etc.

Thanks,

              Linus

Re: mm: change huge_ptep_clear_flush() to return the original pte

[Baolin Wang]

On 5/27/2022 11:42 AM, Linus Torvalds wrote:
> On Thu, May 26, 2022 at 8:36 PM Baolin Wang
> <baolin.wang@linux.alibaba.com> wrote:
>>
>> That is caused by commit:
>>
>> 00df1f1a133b ("mm: change huge_ptep_clear_flush() to return the original
>> pte")
>>
>> interacting with commit:
>>
>> fb396bb459c1 ("arm64/hugetlb: Drop TLB flush from get_clear_flush()")
>>
>> And Catalin has fixed the conflict with below changes [1], not sure why
>> it is not merged.
> 
> Argh. Nobody informed me (or if they did, I clearly missed it).
> 
>> Linus, should I send a proper patch to fix this conflict for you?
> 
> Yeah, somebody please send me a proper tested fix-up patch with commit
> message etc.

Sure. I send out a fix patch [1] which is already tested in linux-next 
for several days.

[1] 
https://lore.kernel.org/all/814e20c19b110209ee12ecae7cb05f8a78d021c8.1653625820.git.baolin.wang@linux.alibaba.com/