mm/hugetlb_cgroup: call page_counter_set_max() outside VM_BUG_ON()
Subsystems (3)
PATCH 1/1 mm/hugetlb_cgroup: call page_counter_set_max() outside VM_BUG_ON()
hugetlb_cgroup_css_alloc() rounds the counter limit down to a multiple of the huge page size and then applies it inside an assertion:
VM_BUG_ON(page_counter_set_max(fault, limit)); VM_BUG_ON(page_counter_set_max(rsvd, limit));
With CONFIG_DEBUG_VM=n, VM_BUG_ON(cond) is BUILD_BUG_ON_INVALID(cond), i.e. ((void)(sizeof((__force long)(cond)))), whose operand is never evaluated. page_counter_set_max() is not a predicate - it performs xchg(&counter->max, nr_pages) - so on every non-debug kernel the limit is never applied and the counters keep page_counter_init()'s PAGE_COUNTER_MAX.
That is user-visible, because hugetlb_cgroup_read_u64_max() recomputes the same rounded value and uses equality as its "unlimited" sentinel. PAGE_COUNTER_MAX is LONG_MAX / PAGE_SIZE = 2251799813685247, which is odd, so round_down() really does change it and the two sides disagree. With CONFIG_DEBUG_VM=n:
$ cat /sys/fs/cgroup/t/hugetlb.2MB.max 9223372036854771712
and with this patch:
$ cat /sys/fs/cgroup/t/hugetlb.2MB.max max
A debug option should not change cgroup output.
Call the function, then assert the result, as v6.12 did. Use VM_WARN_ON_ONCE() rather than restoring VM_BUG_ON(): the two are identical under CONFIG_DEBUG_VM=n, and checkpatch asks that new code not use BUG() variants.
--- mm/hugetlb_cgroup.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-)
mm/hugetlb_cgroup.c +5 -2
@@ -97,6 +97,7 @@ static void hugetlb_cgroup_init(struct hugetlb_cgroup *h_cgroup,struct page_counter *fault, *fault_parent =NULL ;struct page_counter *rsvd, *rsvd_parent =NULL ;unsigned long limit; +int ret;if (parent_h_cgroup) { fault_parent = hugetlb_cgroup_counter_from_cgroup(
@@ -118,8 +119,10 @@ static void hugetlb_cgroup_init(struct hugetlb_cgroup *h_cgroup, limit =round_down (PAGE_COUNTER_MAX ,pages_per_huge_page (&hstates[idx])); -VM_BUG_ON (page_counter_set_max (fault, limit)); -VM_BUG_ON (page_counter_set_max (rsvd, limit)); + ret =page_counter_set_max (fault, limit); +VM_WARN_ON_ONCE (ret); + ret =page_counter_set_max (rsvd, limit); +VM_WARN_ON_ONCE (ret); } }
Signature · 1 line
2.43.0
Quoted 33 lines · Narek Jilavyan
hugetlb_cgroup_css_alloc()rounds the counter limit down to a multiple of the huge page size and then applies it inside an assertion:
VM_BUG_ON(page_counter_set_max(fault, limit));VM_BUG_ON(page_counter_set_max(rsvd, limit));With
CONFIG_DEBUG_VM=n,VM_BUG_ON(cond)isBUILD_BUG_ON_INVALID(cond), i.e. ((void)(sizeof((__force long)(cond)))), whose operand is never evaluated.page_counter_set_max()is not a predicate - it performsxchg(&counter->max, nr_pages)- so on every non-debug kernel the limit is never applied and the counters keep page_counter_init()'sPAGE_COUNTER_MAX.That is user-visible, because
hugetlb_cgroup_read_u64_max()recomputes the same rounded value and uses equality as its "unlimited" sentinel.PAGE_COUNTER_MAXisLONG_MAX/PAGE_SIZE= 2251799813685247, which is odd, soround_down()really does change it and the two sides disagree. WithCONFIG_DEBUG_VM=n:$ cat /sys/fs/cgroup/t/hugetlb.2MB.max 9223372036854771712and with this patch:
$ cat /sys/fs/cgroup/t/hugetlb.2MB.max maxA debug option should not change cgroup output.
Quoted 7 lines
Call the function, then assert the result, as v6.12 did. Use
VM_WARN_ON_ONCE()rather than restoringVM_BUG_ON(): the two are identical underCONFIG_DEBUG_VM=n, and checkpatch asks that new code not useBUG()variants.Fixes 0e2759afcaf9 ("page_counter: track failcnt only for legacy cgroups")Signed-off-by Narek Jilavyan <[email protected]>
Thanks.
Andrew Morton
maintainerQuoted 27 lines · On Mon, 17 Aug 2026 10:34:33 +0000 Narek Jilavyan <[email protected]> wrote:
hugetlb_cgroup_css_alloc()rounds the counter limit down to a multiple of the huge page size and then applies it inside an assertion:VM_BUG_ON(page_counter_set_max(fault, limit)); VM_BUG_ON(page_counter_set_max(rsvd, limit));With
CONFIG_DEBUG_VM=n,VM_BUG_ON(cond)isBUILD_BUG_ON_INVALID(cond), i.e. ((void)(sizeof((__force long)(cond)))), whose operand is never evaluated.page_counter_set_max()is not a predicate - it performsxchg(&counter->max, nr_pages)- so on every non-debug kernel the limit is never applied and the counters keep page_counter_init()'sPAGE_COUNTER_MAX.That is user-visible, because
hugetlb_cgroup_read_u64_max()recomputes the same rounded value and uses equality as its "unlimited" sentinel.PAGE_COUNTER_MAXisLONG_MAX/PAGE_SIZE= 2251799813685247, which is odd, soround_down()really does change it and the two sides disagree. WithCONFIG_DEBUG_VM=n:$ cat /sys/fs/cgroup/t/hugetlb.2MB.max 9223372036854771712and with this patch:
$ cat /sys/fs/cgroup/t/hugetlb.2MB.max max
Quoted 6 lines
A debug option should not change cgroup output.
Call the function, then assert the result, as v6.12 did. Use
VM_WARN_ON_ONCE()rather than restoringVM_BUG_ON(): the two are identical underCONFIG_DEBUG_VM=n, and checkpatch asks that new code not useBUG()variants.
Nice, thanks, I'll add cc:stable to this, to help ensure that users of earlier kernels get to enjoy it.