On Fri, Mar 13, 2020 at 3:39 PM Mina Almasry <almasrymina@xxxxxxxxxx> wrote:
>
> This appears to be a mistake in commit faced7e0806cf ("mm: hugetlb
> controller for cgroups v2"). Essentially that commit does
> a hugetlb_cgroup_from_counter assuming that page_counter_try_charge has
> initialized counter, but if page_counter_try_charge has failed then it
> seems it does not initialize counter, so
> hugetlb_cgroup_from_counter(counter) ends up pointing to random memory,
> causing kasan to complain.
>
> Solution, simply use h_cg, instead of
> hugetlb_cgroup_from_counter(counter), since that is a reference to the
> hugetlb_cgroup anyway. After this change kasan ceases to complain.
>
> Signed-off-by: Mina Almasry <almasrymina@xxxxxxxxxx>
> Reported-by: syzbot+cac0c4e204952cf449b1@xxxxxxxxxxxxxxxxxxxxxxxxx
> Fixes: commit faced7e0806cf ("mm: hugetlb controller for cgroups v2")
> Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
> Cc: linux-mm@xxxxxxxxx
> Cc: linux-kernel@xxxxxxxxxxxxxxx
> Cc: Giuseppe Scrivano <gscrivan@xxxxxxxxxx>
> Cc: Tejun Heo <tj@xxxxxxxxxx>
> Cc: mike.kravetz@xxxxxxxxxx
> Cc: rientjes@xxxxxxxxxx
>
> ---
> mm/hugetlb_cgroup.c | 3 +--
> 1 file changed, 1 insertion(+), 2 deletions(-)
>
> diff --git a/mm/hugetlb_cgroup.c b/mm/hugetlb_cgroup.c
> index 7994eb8a2a0b4..aabf65d4d91ba 100644
> --- a/mm/hugetlb_cgroup.c
> +++ b/mm/hugetlb_cgroup.c
> @@ -259,8 +259,7 @@ static int __hugetlb_cgroup_charge_cgroup(int idx, unsigned long nr_pages,
> __hugetlb_cgroup_counter_from_cgroup(h_cg, idx, rsvd),
> nr_pages, &counter)) {
> ret = -ENOMEM;
> - hugetlb_event(hugetlb_cgroup_from_counter(counter, idx), idx,
> - HUGETLB_MAX);
> + hugetlb_event(h_cg, idx, HUGETLB_MAX);
> css_put(&h_cg->css);
> goto done;
> }
> --
> 2.25.1.481.gfbce0eb801-goog
The patch this fixes is in linus's tree, but isn't in 5.5 stable yet.
