[PATCH] support: Also run malloc hugetlb2 tests when transparent hugepage is 'always'

[Yury Khrustalev]

On Mon, Jun 01, 2026 at 10:17:29AM -0300, Adhemerval Zanella wrote:
> The glibc.malloc.hugetlb=1 is redundant with kernel is n 'always'
> mode, but the madvise does work, and the tunable should not fail.

Thanks, just a couple of typos below.

> diff --git a/support/support_check_hugetlb.c b/support/support_check_hugetlb.c
> index be7e10e80d6..2ea78e4fa39 100644
> --- a/support/support_check_hugetlb.c
> +++ b/support/support_check_hugetlb.c
> 
> ...
> 
>  #define MODE_MADVISE "always [madvise] never\n"
> +#define MODE_ALWAYS  "[always] madvise never\n"
>  
>    char str[sizeof(MODE_MADVISE)];
>    ssize_t s = read (fd, str, sizeof (str));
> @@ -42,7 +43,7 @@ support_thp_is_madvise (void)
>    if (s != sizeof (str) - 1)
>      return false;
>    str[s] = '\0';
> -  return strcmp (str, MODE_MADVISE) == 0;
> +  return strcmp (str, MODE_MADVISE) == 0 || strcmp (str, MODE_ALWAYS);

It should be: strcmp (str, MODE_ALWAYS) == 0


>  }
>  
>  bool
> @@ -69,7 +70,7 @@ support_check_malloc_hugetlb (void)
>      return;
>  
>    size_t hugetlb = TUNABLE_GET_FULL (glibc, malloc, hugetlb, size_t, NULL);
> -  if (hugetlb == 1 && !support_thp_is_madvise ())
> +  if (hugetlb == 1 && !support_thp_work_madvise ())
>      FAIL_UNSUPPORTED ("glibc.malloc.hugetlb=1 requires"
>                        " /sys/kernel/mm/transparent_hugepage/enabled"
>                        " = madvise");

Nit: perhaps rephrase the message to something like

  glibc.malloc.hugetlb=1 requires that transparent_hugepage is either
  madvise or always

To include both accepted values?