Re: [PATCH] selftests/mm: do check_huge_anon() with a number been passed in

[Zi Yan <ziy@nvidia.com>]

On 8 Aug 2025, at 11:59, Zi Yan wrote:

> On 8 Aug 2025, at 11:56, Zi Yan wrote:
>
>> On 7 Aug 2025, at 4:27, Wei Yang wrote:
>>
>>> Currently it hard coded the number of hugepage to check for
>>> check_huge_anon(), but we already have the number passed in.
>>>
>>> Do the check based on the number of hugepage passed in is more
>>> reasonable.
>>>
>>> Signed-off-by: Wei Yang <richard.weiyang@gmail.com>
>>> ---
>>>  tools/testing/selftests/mm/split_huge_page_test.c | 2 +-
>>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/tools/testing/selftests/mm/split_huge_page_test.c b/tools/testing/selftests/mm/split_huge_page_test.c
>>> index 44a3f8a58806..bf40e6b121ab 100644
>>> --- a/tools/testing/selftests/mm/split_huge_page_test.c
>>> +++ b/tools/testing/selftests/mm/split_huge_page_test.c
>>> @@ -111,7 +111,7 @@ static void verify_rss_anon_split_huge_page_all_zeroes(char *one_page, int nr_hp
>>>  	unsigned long rss_anon_before, rss_anon_after;
>>>  	size_t i;
>>>
>>> -	if (!check_huge_anon(one_page, 4, pmd_pagesize))
>>> +	if (!check_huge_anon(one_page, nr_hpages, pmd_pagesize))
>>>  		ksft_exit_fail_msg("No THP is allocated\n");
>>>
>>>  	rss_anon_before = rss_anon();
>>
>> Which commit is this patch based on?
>>
>> In mm-new[1], verify_rss_anon_split_huge_page_all_zeroes() accepts
>> char *one_page and size_t len but no nr_hpages. This breaks
>> split_huge_page_test.c compilation.
>>
>> Are you forgetting some cleanup patches before this one?
>>
>> [1] https://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm.git/tree/tools/testing/selftests/mm/split_huge_page_test.c?h=mm-new#n110
>>
>>
>> Hi Andrew,
>>
>> Can you drop this one for now? Thanks.
>
> Wait. It seems that mm-new is different from Linus' tree[1]. I wonder
> what is going on.
>
> [1] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/tools/testing/selftests/mm/split_huge_page_test.c?h=v6.16#n109

Wei, it seems that you lost the race to Muhammad Usama Anjum's cleanup patch[1].
You might need to add nr_hpages back.

[1] https://lore.kernel.org/all/20250731160132.1795351-6-usama.anjum@collabora.com/T/#u

--
Best Regards,
Yan, Zi