On 11 Sep 2026, at 8:45, Jaeyeon Lee wrote:

> The XFS setup for ./khugepaged all:file checks that the kernel supports
> XFS but not that mkfs.xfs is installed. When CONFIG_XFS_FS=y and
> xfsprogs is missing, mkfs.xfs and mount both fail, but
> SPLIT_HUGE_PAGE_TEST_XFS_PATH is assigned from mktemp -d and stays set.
> The test then runs against plain tmpfs and six subtests fail.
>
> The script already has a skip path for this test, but it never runs
> because the path variable is always set. Check that mkfs.xfs exists
> before attempting the XFS setup.
>
> Assisted-by: claude.ai:claude-opus-5
> Signed-off-by: Jaeyeon Lee <[email protected]>
> ---
>  tools/testing/selftests/mm/run_vmtests.sh | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/tools/testing/selftests/mm/run_vmtests.sh 
> b/tools/testing/selftests/mm/run_vmtests.sh
> index d09f9f6a384e..6315b67b6733 100755
> --- a/tools/testing/selftests/mm/run_vmtests.sh
> +++ b/tools/testing/selftests/mm/run_vmtests.sh
> @@ -415,7 +415,8 @@ CATEGORY="thp" run_test ./khugepaged -c 4 
> mthp_khugepaged:anon
>  # Try to create XFS if not provided
>  if [ -z "${SPLIT_HUGE_PAGE_TEST_XFS_PATH}" ]; then
>      if test_selected "thp"; then
> -     if grep xfs /proc/filesystems &>/dev/null; then
> +     if grep xfs /proc/filesystems &>/dev/null &&
> +        command -v mkfs.xfs &>/dev/null; then
>           XFS_IMG=$(mktemp /tmp/xfs_img_XXXXXX)
>           SPLIT_HUGE_PAGE_TEST_XFS_PATH=$(mktemp -d /tmp/xfs_dir_XXXXXX)
>           truncate -s 314572800 ${XFS_IMG}
>

This might be more robust, since it can detect mkfs.xfs and mount failures.

  if grep xfs /proc/filesystems &>/dev/null; then
      XFS_IMG=$(mktemp /tmp/xfs_img_XXXXXX)
      XFS_DIR=$(mktemp -d /tmp/xfs_dir_XXXXXX)
      truncate -s 314572800 ${XFS_IMG}
      if mkfs.xfs -q ${XFS_IMG} && mount -t xfs -o loop ${XFS_IMG} ${XFS_DIR}; 
then
          SPLIT_HUGE_PAGE_TEST_XFS_PATH=${XFS_DIR}
          MOUNTED_XFS=1
      else
          rmdir ${XFS_DIR}
          rm -f ${XFS_IMG}
      fi
  fi



Best Regards,
Yan, Zi

Reply via email to