On 27.08.25 09:52, Chunyu Hu wrote:
> The nr_hugepgs variable is used to keep the original nr_hugepages at the
> hugepage setup step at test beginning. After userfaultfd test, a cleaup is
> executed, both /sys/kernel/mm/hugepages/hugepages-*/nr_hugepages and
> /proc/sys//vm/nr_hugepages are reset to 'original' value before userfaultfd
> test starts.
>
> Issue here is the value used to restore /proc/sys/vm/nr_hugepages is
> nr_hugepgs which is the initial value before the vm_runtests.sh runs, not
> the value before userfaultfd test starts. 'va_high_addr_swith.sh' tests
> runs after that will possibly see no hugepages available for test, and got
> EINVAL when mmap(HUGETLB), making the result invalid.
>
> And before pkey tests, nr_hugepgs is changed to be used as a temp variable
> to save nr_hugepages before pkey test, and restore it after pkey tests
> finish. The original nr_hugepages value is not tracked anymore, so no way
> to restore it after all tests finish.
>
> Add a new variable nr_hugepgs_origin to save the original nr_hugepages, and
> and restore it to nr_hugepages after all tests finish. And change to use
> the nr_hugepgs variable to save the /proc/sys/vm/nr_hugeages after hugepage
> setup, it's also the value before userfaultfd test starts, and the correct
> value to be restored after userfaultfd finishes. The va_high_addr_switch.sh
> broken will be resolved.
>
> Signed-off-by: Chunyu Hu <chuhu@redhat.com>
> ---
>   tools/testing/selftests/mm/run_vmtests.sh | 9 +++++++--
>   1 file changed, 7 insertions(+), 2 deletions(-)
>
> diff --git a/tools/testing/selftests/mm/run_vmtests.sh b/tools/testing/selftests/mm/run_vmtests.sh
> index 471e539d82b8..f1a7ad3ec6a7 100755
> --- a/tools/testing/selftests/mm/run_vmtests.sh
> +++ b/tools/testing/selftests/mm/run_vmtests.sh
> @@ -172,13 +172,13 @@ fi
>   
>   # set proper nr_hugepages
>   if [ -n "$freepgs" ] && [ -n "$hpgsize_KB" ]; then
> -     nr_hugepgs=$(cat /proc/sys/vm/nr_hugepages)
> +     nr_hugepgs_origin=$(cat /proc/sys/vm/nr_hugepages)

I'd call this "orig_nr_hugepgs".

But it's a shame that the naming is then out of sync with nr_size_hugepgs?

>       needpgs=$((needmem_KB / hpgsize_KB))
>       tries=2
>       while [ "$tries" -gt 0 ] && [ "$freepgs" -lt "$needpgs" ]; do
>               lackpgs=$((needpgs - freepgs))
>               echo 3 > /proc/sys/vm/drop_caches
> -             if ! echo $((lackpgs + nr_hugepgs)) > /proc/sys/vm/nr_hugepages; then
> +             if ! echo $((lackpgs + nr_hugepgs_origin)) > /proc/sys/vm/nr_hugepages; then
>                       echo "Please run this test as root"
>                       exit $ksft_skip
>               fi
> @@ -189,6 +189,7 @@ if [ -n "$freepgs" ] && [ -n "$hpgsize_KB" ]; then
>               done < /proc/meminfo
>               tries=$((tries - 1))
>       done
> +     nr_hugepgs=$(cat /proc/sys/vm/nr_hugepages)
>       if [ "$freepgs" -lt "$needpgs" ]; then
>               printf "Not enough huge pages available (%d < %d)\n" \
>                      "$freepgs" "$needpgs"
> @@ -532,6 +533,10 @@ CATEGORY="page_frag" run_test ./test_page_frag.sh aligned
>   
>   CATEGORY="page_frag" run_test ./test_page_frag.sh nonaligned
>   
> +if [ "${HAVE_HUGEPAGES}" = 1 ]; then
> +     echo "$nr_hugepgs_origin" > /proc/sys/vm/nr_hugepages
> +fi

FWIW, I think the tests should maybe be doing that
(save+configure+restore) themselves, like we do with THP settings through.

thp_save_settings()
thp_write_settings()

and friends.

This is not really something run_vmtests.sh should bother with.

A bigger rework, though ...

--
Cheers

David / dhildenb