Re: [PATCH v8 3/4] kselftest: mm: integrate huge page checks

[David Hildenbrand (Arm)]

On 9/29/26 12:01, Yeoreum Yun wrote:
> On Tue, Sep 29, 2026 at 10:07:37AM +0200, David Hildenbrand (Arm) wrote:
>> On 9/24/26 21:11, Yeoreum Yun wrote:
>>> check_large_folios() only checks for large folios without distinguishing
>>> between anonymous and file-backed huge pages.
>>>
>>> To add huge page type checking, integrate the huge page checks into
>>> __check_huge():
>>>
>>>   1. If hpage_size == pmd_pagesize, check PAGE_IS_HUGE instead of using
>>>      check_large_folios(), since only the mapping type matters. This
>>>      identifies PMD-mapped huge pages.
>>>
>>>   2. Otherwise, use check_large_folios() to detect large folios. This
>>>      covers mTHP cases.
>>>
>>>   3. Check the folio flags according to the huge page type.
>>>
>>> Suggested-by: David Hildenbrand (Arm) <david@xxxxxxxxxx>
>>> Suggested-by: Zi Yan <ziy@xxxxxxxxxx>
>>> Signed-off-by: Yeoreum Yun <yeoreum.yun@xxxxxxx>
>>> ---
>>
>> Instead of merging both things (detecting mapping vs. detecting anon vs. file),
>> could we simply perform the anon vs. file change separately?
>>
>> Doing another pagemap walk that focuses on that should end up with something
>> that is easier to read.
> 
> Okay. I'll change like below in next-spin:
> 
> -------&<-------
> 
> @@ -411,57 +400,78 @@ static bool check_huge_type(uint64_t categories, enum check_huge_type type)
>         return false;
>  }
> 
> -static bool __check_pmd_huge(void *addr, size_t len, int nr_hpages,
> -                 uint64_t hpage_size, enum check_huge_type type)
> +static bool __check_huge(void *addr, size_t len, int nr_hpages,
> +               uint64_t hpage_size, enum check_huge_type type)
>  {
>  {
> -       int pagemap_fd;
> +       bool ret = false;
> +       int pagemap_fd, kpageflags_fd;
>         int nr_pmd_mappings = 0;
> +       uint64_t pmd_pagesize, scan_mapping_size;
>         uint64_t categories;
> +       unsigned long pfn;
> +       bool check_pmd_mapping, allow_nonpresent;
>         char *start = addr;
>         char *end = start + len;
> 
> +       pmd_pagesize = read_pmd_pagesize();
> +       if (!pmd_pagesize)
> +               ksft_exit_fail_msg("reading PMD pagesize failed\n");
> +
> +       check_pmd_mapping = hpage_size == pmd_pagesize;
> +       scan_mapping_size = (nr_hpages > 0) ? hpage_size : psize();
> +       /* Some mTHP tests check a partially populated PMD-sized range. */
> +       allow_nonpresent = (uint64_t)nr_hpages * hpage_size < len;
> +
>         pagemap_fd = open(PAGEMAP_PATH, O_RDONLY);
>         if (pagemap_fd < 0)
>                 ksft_exit_fail_perror("open pagemap");
> 
> -       for (; start < end; start += hpage_size) {
> +       kpageflags_fd = open(KPAGEFLAGS_PATH, O_RDONLY);
> +       if (kpageflags_fd < 0)
> +               ksft_exit_fail_perror("open kpageflags");
> +
> +       for (; start < end; start += scan_mapping_size) {
>                 categories = pagemap_scan_get_categories(pagemap_fd, start);
> -               if (!(categories & PAGE_IS_HUGE))
> +               pfn = pagemap_get_pfn(pagemap_fd, start);
> +               if (pfn == -1UL) {
> +                       if (!allow_nonpresent)
> +                               goto out;
>                         continue;
> -               if (check_huge_type(categories, type))
> +               }
> +               if (!check_huge_type(categories, type))
> +                       goto out;
> +       }
> +
> +       if (!check_pmd_mapping) {
> +               ret = check_large_folios(pagemap_fd, kpageflags_fd,
> +                               addr, len, nr_hpages, hpage_size);
> +               goto out;
> +       }
> +
> +       for (start = addr; start < end; start += scan_mapping_size) {
> +               categories = pagemap_scan_get_categories(pagemap_fd, start);
> +               if (categories & PAGE_IS_HUGE)
>                         nr_pmd_mappings++;
>         }
> -       close(pagemap_fd);
> 
> -       return nr_hpages == nr_pmd_mappings;
> +       if (nr_pmd_mappings != nr_hpages)
> +               goto out;
> +       ret = true;
> +
> +out:
> +       close(pagemap_fd);
> +       close(kpageflags_fd);
> +       return ret;
>  }
> 

I'd leave existing __check_huge() mostly alone, and instead have an additional
function that checks the type.

Essentially a __check_type() or sth that we run after the large folio / pmd check.

-- 
Cheers,

David