On Thu, Aug 03, 2023 at 12:00:47PM -0400, Lucas Karpinski wrote: > test_kmem_basic creates 100,000 negative dentries, with each one mapping > to a slab object. After memory.high is set, these are reclaimed through > the shrink_slab function call which reclaims all 100,000 entries. The > test passes the majority of the time because when slab1 is calculated, > it is often above 0, however, 0 is also an acceptable value. > > Signed-off-by: Lucas Karpinski <lkarpins@xxxxxxxxxx> Acked-by: Johannes Weiner <hannes@xxxxxxxxxxx> > @@ -71,7 +71,7 @@ static int test_kmem_basic(const char *root) > > cg_write(cg, "memory.high", "1M"); > slab1 = cg_read_key_long(cg, "memory.stat", "slab "); > - if (slab1 <= 0) > + if (slab1 < 0) > goto cleanup; This conflicts with a recent patch already queued up in -mm: selftests: cgroup: fix test_kmem_basic false positives which adds a sleep(1) between cg_write() and cg_read_key_long(). Can you please rebase on top of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm.git mm-unstable and re-send the patch with To: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> ? Thanks
