Commit fa8cc33d authored by Peter Xu's avatar Peter Xu Committed by Jinjiang Tu
Browse files

mm/hugetlb: fix missing hugetlb_lock for resv uncharge 

mainline inclusion
from mainline-v6.9-rc6
commit b76b46902c2d0395488c8412e1116c2486cdfcb2
category: bugfix
bugzilla: https://gitee.com/src-openeuler/kernel/issues/I9QR5M
CVE: CVE-2024-36000

Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=b76b46902c2d0395488c8412e1116c2486cdfcb2

--------------------------------

There is a recent report on UFFDIO_COPY over hugetlb:

https://lore.kernel.org/all/000000000000ee06de0616177560@google.com/

350:	lockdep_assert_held(&hugetlb_lock);

Should be an issue in hugetlb but triggered in an userfault context, where
it goes into the unlikely path where two threads modifying the resv map
together.  Mike has a fix in that path for resv uncharge but it looks like
the locking criteria was overlooked: hugetlb_cgroup_uncharge_folio_rsvd()
will update the cgroup pointer, so it requires to be called with the lock
held.

Link: https://lkml.kernel.org/r/20240417211836.2742593-3-peterx@redhat.com


Fixes: 79aa925b ("hugetlb_cgroup: fix reservation accounting")
Signed-off-by: default avatarPeter Xu <peterx@redhat.com>
Reported-by: default avatar <syzbot+4b8077a5fccc61c385a1@syzkaller.appspotmail.com>
Reviewed-by: default avatarMina Almasry <almasrymina@google.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>

Conflicts:
	mm/hugetlb.c
[Conflicts due to page converts to folio.]
Signed-off-by: default avatarJinjiang Tu <tujinjiang@huawei.com>
parent 3b9084bc

mm/hugetlb.c

+4 −1
Original line number Diff line number Diff line
@@ -2709,9 +2709,12 @@ struct page *alloc_huge_page(struct vm_area_struct *vma, 


		rsv_adjust = hugepage_subpool_put_pages(spool, 1, info);

		hugetlb_acct_memory(h, -rsv_adjust, info);

		if (deferred_reserve)

		if (deferred_reserve) {

			spin_lock_irq(&hugetlb_lock);

			hugetlb_cgroup_uncharge_page_rsvd(hstate_index(h),

					pages_per_huge_page(h), page);

			spin_unlock_irq(&hugetlb_lock);

		}

	}

	return page;