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