Commit 9966b579 authored May 22, 2024 by Mike Kravetz Committed by Liu Shixin May 22, 2024

hugetlb: check for hugetlb folio before vmemmap_restore

mainline inclusion
from mainline-v6.7-rc1
commit 30a89adf872d2e46323840964c95dc0ae3bb5843
category: bugfix
bugzilla: https://gitee.com/openeuler/kernel/issues/I9R3AY
CVE: NA

Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=30a89adf872d2e46323840964c95dc0ae3bb5843

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

In commit d8f5f7e445f0 ("hugetlb: set hugetlb page flag before
optimizing vmemmap") checks were added to print a warning if
hugetlb_vmemmap_restore was called on a non-hugetlb page.

This was mostly due to ordering issues in the hugetlb page set up and tear
down sequencees.  One place missed was the routine
dissolve_free_huge_page.

Naoya Horiguchi noted: "I saw that VM_WARN_ON_ONCE() in
hugetlb_vmemmap_restore is triggered when memory_failure() is called on a
free hugetlb page with vmemmap optimization disabled (the warning is not
triggered if vmemmap optimization is enabled).  I think that we need check
folio_test_hugetlb() before dissolve_free_huge_page() calls
hugetlb_vmemmap_restore_folio()."

Perform the check as suggested by Naoya.

Link: https://lkml.kernel.org/r/20231017032140.GA3680@monkey


Fixes: d8f5f7e445f0 ("hugetlb: set hugetlb page flag before optimizing vmemmap")
Signed-off-by: Mike Kravetz <mike.kravetz@oracle.com>
Suggested-by: Naoya Horiguchi <naoya.horiguchi@linux.dev>
Tested-by: Naoya Horiguchi <naoya.horiguchi@linux.dev>
Cc: Anshuman Khandual <anshuman.khandual@arm.com>
Cc: Barry Song <song.bao.hua@hisilicon.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Joao Martins <joao.m.martins@oracle.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Miaohe Lin <linmiaohe@huawei.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Muchun Song <songmuchun@bytedance.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Xiongchun Duan <duanxiongchun@bytedance.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Liu Shixin <liushixin2@huawei.com>

parent fb24953c

mm/hugetlb.c

+15 −9

Original line number	Diff line number	Diff line
		@@ -2411,17 +2411,23 @@ int dissolve_free_huge_page(struct page *page)
		* need to adjust max_huge_pages if the page is not freed.
		* Attempt to allocate vmemmmap here so that we can take
		* appropriate action on failure.
		*
		* The folio_test_hugetlb check here is because
		* remove_hugetlb_folio will clear hugetlb folio flag for
		* non-vmemmap optimized hugetlb folios.
		*/
		if (folio_test_hugetlb(folio)) {
		rc = hugetlb_vmemmap_restore(h, &folio->page);
		if (!rc) {
		update_and_free_hugetlb_folio(h, folio, false);
		} else {
		if (rc) {
		spin_lock_irq(&hugetlb_lock);
		add_hugetlb_folio(h, folio, false);
		h->max_huge_pages++;
		spin_unlock_irq(&hugetlb_lock);
		goto out;
		}
		} else
		rc = 0;

		update_and_free_hugetlb_folio(h, folio, false);
		return rc;
		}
		out: