[PATCH v2 0/2] selftests/mm: uffd-stress fixes

[PATCH v2 0/2] selftests/mm: uffd-stress fixes

[Dev Jain]

This patchset ensures that the number of hugepages is correctly set in the
system so that the uffd-stress test does not fail due to the racy nature of
the test. Patch 1 changes the hugepage constraint in the run_vmtests.sh
script, whereas patch 2 changes the constraint in the test itself.

---
Based on 6.17-rc5.

Dev Jain (2):
  selftests/mm/uffd-stress: make test operate on less hugetlb memory
  selftests/mm/uffd-stress: stricten constraint on free hugepages needed
    before the test

 tools/testing/selftests/mm/run_vmtests.sh | 10 +++++++---
 tools/testing/selftests/mm/uffd-stress.c  | 17 +++++++++++------
 2 files changed, 18 insertions(+), 9 deletions(-)

-- 
2.30.2

[PATCH v2 1/2] selftests/mm/uffd-stress: make test operate on less hugetlb memory

[Dev Jain]

We observed uffd-stress selftest failure on arm64 and intermittent failures
on x86 too:
running ./uffd-stress hugetlb-private 128 32

bounces: 17, mode: rnd read, ERROR: UFFDIO_COPY error: -12 (errno=12, @uffd-common.c:617) [FAIL]
not ok 18 uffd-stress hugetlb-private 128 32 # exit=1

For this particular case, the number of free hugepages from run_vmtests.sh
will be 128, and the test will allocate 64 hugepages in the source
location. The stress() function will start spawning threads which will
operate on the destination location, triggering uffd-operations like
UFFDIO_COPY from src to dst, which means that we will require 64 more
hugepages for the dst location.

Let us observe the locking_thread() function. It will lock the mutex kept
at dst, triggering uffd-copy. Suppose that 127 (64 for src and 63 for dst)
hugepages have been reserved. In case of BOUNCE_RANDOM, it may happen that
two threads trying to lock the mutex at dst, try to do so at the same
hugepage number. If one thread succeeds in reserving the last hugepage,
then the other thread may fail in alloc_hugetlb_folio(), returning -ENOMEM.
I can confirm that this is indeed the case by this hacky patch:

diff --git a/mm/hugetlb.c b/mm/hugetlb.c
index 753f99b4c718..39eb21d8a91b 100644
--- a/mm/hugetlb.c
+++ b/mm/hugetlb.c
@@ -6929,6 +6929,11 @@ int hugetlb_mfill_atomic_pte(pte_t *dst_pte,
 
 		folio = alloc_hugetlb_folio(dst_vma, dst_addr, false);
 		if (IS_ERR(folio)) {
+			pte_t *actual_pte = hugetlb_walk(dst_vma, dst_addr, PMD_SIZE);
+			if (actual_pte) {
+				ret = -EEXIST;
+				goto out;
+			}
 			ret = -ENOMEM;
 			goto out;
 		}

This code path gets triggered indicating that the PMD at which one thread
is trying to map a hugepage, gets filled by a racing thread.

Therefore, instead of using freepgs to compute the amount of memory,
use freepgs - (min(32, nr_cpus) - 1), so that the test still has some extra
hugepages to use. The adjustment is a function of min(32, nr_cpus) - the
value of nr_parallel in the test - because in the worst case, nr_parallel
number of threads will try to map a hugepage on the same PMD, one will
win the allocation race, and the other nr_parallel - 1 threads will fail,
so we need extra nr_parallel - 1 hugepages to satisfy this request.
Note that, in case the adjusted value underflows, there is a check for
the number of free hugepages in the test itself, which will fail:
get_free_hugepages() < bytes / page_size
A negative value will be passed on to bytes which is of type size_t, thus
the RHS will become a large value and the check will fail,
so we are safe.

Signed-off-by: Dev Jain <dev.jain@arm.com>
---
 tools/testing/selftests/mm/run_vmtests.sh | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/tools/testing/selftests/mm/run_vmtests.sh b/tools/testing/selftests/mm/run_vmtests.sh
index 471e539d82b8..b9c73ef81c03 100755
--- a/tools/testing/selftests/mm/run_vmtests.sh
+++ b/tools/testing/selftests/mm/run_vmtests.sh
@@ -322,11 +322,15 @@ CATEGORY="gup_test" run_test ./gup_longterm
 CATEGORY="userfaultfd" run_test ./uffd-unit-tests
 uffd_stress_bin=./uffd-stress
 CATEGORY="userfaultfd" run_test ${uffd_stress_bin} anon 20 16
-# Hugetlb tests require source and destination huge pages. Pass in half
-# the size of the free pages we have, which is used for *each*.
+# Hugetlb tests require source and destination huge pages. Pass in almost half
+# the size of the free pages we have, which is used for *each*. An adjustment
+# of (nr_parallel - 1) is done (see nr_parallel in uffd-stress.c) to have some
+# extra hugepages - this is done to prevent the test from failing by racily
+# reserving more hugepages than strictly required.
 # uffd-stress expects a region expressed in MiB, so we adjust
 # half_ufd_size_MB accordingly.
-half_ufd_size_MB=$(((freepgs * hpgsize_KB) / 1024 / 2))
+adjustment=$(( (31 < (nr_cpus - 1)) ? 31 : (nr_cpus - 1) ))
+half_ufd_size_MB=$((((freepgs - adjustment) * hpgsize_KB) / 1024 / 2))
 CATEGORY="userfaultfd" run_test ${uffd_stress_bin} hugetlb "$half_ufd_size_MB" 32
 CATEGORY="userfaultfd" run_test ${uffd_stress_bin} hugetlb-private "$half_ufd_size_MB" 32
 CATEGORY="userfaultfd" run_test ${uffd_stress_bin} shmem 20 16
-- 
2.30.2

[PATCH v2 2/2] selftests/mm/uffd-stress: stricten constraint on free hugepages needed before the test

[Dev Jain]

The test requires at least 2 * (bytes/page_size) hugetlb memory, since
we require identical number of hugepages for src and dst location. Fix
this.

Along with the above, as explained in patch "selftests/mm/uffd-stress:
Make test operate on less hugetlb memory", the racy nature of the test
requires that we have some extra number of hugepages left beyond what is
required. Therefore, stricten this constraint.

Fixes: 5a6aa60d1823 ("selftests/mm: skip uffd hugetlb tests with insufficient hugepages")
Signed-off-by: Dev Jain <dev.jain@arm.com>
---
 tools/testing/selftests/mm/uffd-stress.c | 17 +++++++++++------
 1 file changed, 11 insertions(+), 6 deletions(-)

diff --git a/tools/testing/selftests/mm/uffd-stress.c b/tools/testing/selftests/mm/uffd-stress.c
index 40af7f67c407..3f6213ae0dd5 100644
--- a/tools/testing/selftests/mm/uffd-stress.c
+++ b/tools/testing/selftests/mm/uffd-stress.c
@@ -448,12 +448,6 @@ 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);
 	if (nr_cpus > 32) {
 		/* Don't let calculation below go to zero. */
@@ -464,6 +458,17 @@ int main(int argc, char **argv)
 		nr_parallel = nr_cpus;
 	}
 
+	/*
+	 * src and dst each require bytes / page_size number of hugepages.
+	 * Ensure nr_parallel - 1 hugepages on top of that to account
+	 * for racy extra reservation of hugepages.
+	 */
+	if (test_type == TEST_HUGETLB &&
+	   get_free_hugepages() < 2 * (bytes / page_size) + nr_parallel - 1) {
+		printf("skip: Skipping userfaultfd... not enough hugepages\n");
+		return KSFT_SKIP;
+	}
+
 	nr_pages_per_cpu = bytes / page_size / nr_parallel;
 	if (!nr_pages_per_cpu) {
 		_err("pages_per_cpu = 0, cannot test (%lu / %lu / %lu)",
-- 
2.30.2

Re: [PATCH v2 0/2] selftests/mm: uffd-stress fixes

[Dev Jain]

On 09/09/25 11:45 am, Dev Jain wrote:
> This patchset ensures that the number of hugepages is correctly set in the
> system so that the uffd-stress test does not fail due to the racy nature of
> the test. Patch 1 changes the hugepage constraint in the run_vmtests.sh
> script, whereas patch 2 changes the constraint in the test itself.
>
> ---

Missed the changelog:

v1->v2:
Compute adjustment as a function of the number of threads

v1:
https://lore.kernel.org/all/20250826070705.53841-1-dev.jain@arm.com/

> Based on 6.17-rc5.
>
> Dev Jain (2):
>    selftests/mm/uffd-stress: make test operate on less hugetlb memory
>    selftests/mm/uffd-stress: stricten constraint on free hugepages needed
>      before the test
>
>   tools/testing/selftests/mm/run_vmtests.sh | 10 +++++++---
>   tools/testing/selftests/mm/uffd-stress.c  | 17 +++++++++++------
>   2 files changed, 18 insertions(+), 9 deletions(-)
>