Re: [PATCH v2 3/3] mm/hugetlb: add migration entry check in hugetlb_change_protection

[Hugh Dickins]

On Fri, 1 Aug 2014, Naoya Horiguchi wrote:

> There is a race condition between hugepage migration and change_protection(),
> where hugetlb_change_protection() doesn't care about migration entries and
> wrongly overwrites them. That causes unexpected results like kernel crash.
> 
> This patch adds is_hugetlb_entry_(migration|hwpoisoned) check in this
> function and skip all such entries.
> 
> Signed-off-by: Naoya Horiguchi <n-horiguchi@xxxxxxxxxxxxx>
> Cc: <stable@xxxxxxxxxxxxxxx>  # [3.12+]
> ---
>  mm/hugetlb.c | 8 +++++++-
>  1 file changed, 7 insertions(+), 1 deletion(-)
> 
> diff --git mmotm-2014-07-22-15-58.orig/mm/hugetlb.c mmotm-2014-07-22-15-58/mm/hugetlb.c
> index 863f45f63cd5..1da7ca2e2a02 100644
> --- mmotm-2014-07-22-15-58.orig/mm/hugetlb.c
> +++ mmotm-2014-07-22-15-58/mm/hugetlb.c
> @@ -3355,7 +3355,13 @@ unsigned long hugetlb_change_protection(struct vm_area_struct *vma,
>  			spin_unlock(ptl);
>  			continue;
>  		}
> -		if (!huge_pte_none(huge_ptep_get(ptep))) {
> +		pte = huge_ptep_get(ptep);
> +		if (unlikely(is_hugetlb_entry_migration(pte) ||
> +			     is_hugetlb_entry_hwpoisoned(pte))) {

Another instance of this pattern.  Oh well, perhaps we have to continue
this way while backporting fixes, but the repetition irritates me.
Or use is_swap_pte() as follow_hugetlb_page() does?

More importantly, the regular change_pte_range() has to
make_migration_entry_read() if is_migration_entry_write():
why is that not necessary here?

And have you compared hugetlb codepaths with normal codepaths, to see
if there are other huge places which need to check for a migration entry
now?  If you have checked, please reassure us in the commit message:
we would prefer not to have these fixes coming in one by one.

(I first thought __unmap_hugepage_range() would need it, but since
zap_pte_range() only checks it for rss stats, and hugetlb does not
participate in rss stats, it looks like no need.)

Hugh

> +			spin_unlock(ptl);
> +			continue;
> +		}
> +		if (!huge_pte_none(pte)) {
>  			pte = huge_ptep_get_and_clear(mm, address, ptep);
>  			pte = pte_mkhuge(huge_pte_modify(pte, newprot));
>  			pte = arch_make_huge_pte(pte, vma, NULL, 0);
> -- 
> 1.9.3
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/