This series addresses issues related to hugepage requirements in the MM selftests, ensuring tests are skipped rather than failing when the necessary hugepage count is not met.
This adjustment allows for a more graceful handling for systems with insufficient hugepages, preventing unnecessary test failures and improving the overall robustness of the test suite.
Nico Pache (3): selftests/mm: Dont fail testsuite due to a lack of hugepages selftests/mm: Skip uffd hugetlb tests with insufficient hugepages selftests/mm: Skip the hugetlb-madvise tests on unmet hugepage requirements
Changes from v1: - Added checks to skip tests when hugepage requirements are not met, rather than exiting with a failure.
tools/testing/selftests/mm/hugetlb-madvise.c | 2 +- tools/testing/selftests/mm/run_vmtests.sh | 1 - tools/testing/selftests/mm/uffd-stress.c | 6 ++++++ 3 files changed, 7 insertions(+), 2 deletions(-)
On systems that have large core counts and large page sizes, but limited memory, the userfaultfd test hugepage requirement is too large.
Exiting early due to missing one test's requirements is a rather aggressive strategy, and prevents a lot of other tests from running. Remove the early exit to prevent this.
Fixes: ee00479d6702 ("selftests: vm: Try harder to allocate huge pages") Signed-off-by: Nico Pache npache@redhat.com --- tools/testing/selftests/mm/run_vmtests.sh | 1 - 1 file changed, 1 deletion(-)
diff --git a/tools/testing/selftests/mm/run_vmtests.sh b/tools/testing/selftests/mm/run_vmtests.sh index 246d53a5d7f28..727ea22ba408e 100755 --- a/tools/testing/selftests/mm/run_vmtests.sh +++ b/tools/testing/selftests/mm/run_vmtests.sh @@ -173,7 +173,6 @@ if [ -n "$freepgs" ] && [ -n "$hpgsize_KB" ]; then if [ "$freepgs" -lt "$needpgs" ]; then printf "Not enough huge pages available (%d < %d)\n" \ "$freepgs" "$needpgs" - exit 1 fi else echo "no hugetlbfs support in kernel?"
Now that run_vmtests.sh does not guarantee that the correct hugepage count is available, add a check inside the userfaultfd hugetlb test to verify the nr_hugepages count before continuing.
Signed-off-by: Nico Pache npache@redhat.com --- tools/testing/selftests/mm/uffd-stress.c | 6 ++++++ 1 file changed, 6 insertions(+)
diff --git a/tools/testing/selftests/mm/uffd-stress.c b/tools/testing/selftests/mm/uffd-stress.c index 7e83829bbb335..1165446f10883 100644 --- a/tools/testing/selftests/mm/uffd-stress.c +++ b/tools/testing/selftests/mm/uffd-stress.c @@ -441,6 +441,12 @@ int main(int argc, char **argv) parse_test_type_arg(argv[1]); bytes = atol(argv[2]) * 1024 * 1024;
+ if (test_type == TEST_HUGETLB && + get_free_hugepages() < bytes / page_size) { + printf("skip: Skipping userfaultfd... not enough hugepages\n"); + return KSFT_SKIP; + } + nr_cpus = sysconf(_SC_NPROCESSORS_ONLN);
nr_pages_per_cpu = bytes / page_size / nr_cpus;
Now that run_vmtests.sh does not guarantee that the correct hugepage count is available, skip the hugetlb-madvise test if the requirements are not met rather than failing.
Signed-off-by: Nico Pache npache@redhat.com --- tools/testing/selftests/mm/hugetlb-madvise.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/tools/testing/selftests/mm/hugetlb-madvise.c b/tools/testing/selftests/mm/hugetlb-madvise.c index f32d99565c5ea..e74107185324f 100644 --- a/tools/testing/selftests/mm/hugetlb-madvise.c +++ b/tools/testing/selftests/mm/hugetlb-madvise.c @@ -19,6 +19,7 @@ #include <sys/mman.h> #include <fcntl.h> #include "vm_util.h" +#include "../kselftest.h"
#define MIN_FREE_PAGES 20 #define NR_HUGE_PAGES 10 /* common number of pages to map/allocate */ @@ -78,7 +79,7 @@ int main(int argc, char **argv) free_hugepages = get_free_hugepages(); if (free_hugepages < MIN_FREE_PAGES) { printf("Not enough free huge pages to test, exiting!\n"); - exit(1); + exit(KSFT_SKIP); }
fd = memfd_create(argv[0], MFD_HUGETLB);
linux-kselftest-mirror@lists.linaro.org
-
Nico Pache