Messages in this thread | | | Date | Fri, 10 Apr 2015 13:38:59 -0700 | From | Andrew Morton <> | Subject | Re: [PATCH] mm/hugetlb: use pmd_page() in follow_huge_pmd() |
| |
On Fri, 10 Apr 2015 10:08:49 +0200 Gerald Schaefer <gerald.schaefer@de.ibm.com> wrote:
> On Thu, 9 Apr 2015 12:41:47 -0700 (PDT) > David Rientjes <rientjes@google.com> wrote: > > > On Thu, 9 Apr 2015, Gerald Schaefer wrote: > > > > > commit 61f77eda "mm/hugetlb: reduce arch dependent code around > > > follow_huge_*" broke follow_huge_pmd() on s390, where pmd and pte > > > layout differ and using pte_page() on a huge pmd will return wrong > > > results. Using pmd_page() instead fixes this. > > > > > > All architectures that were touched by commit 61f77eda have > > > pmd_page() defined, so this should not break anything on other > > > architectures. > > > > > > Signed-off-by: Gerald Schaefer <gerald.schaefer@de.ibm.com> > > > Cc: stable@vger.kernel.org # v3.12 > > > > Acked-by: David Rientjes <rientjes@google.com> > > > > I'm not sure where the stable cc came from, though: commit 61f77eda > > makes s390 use a generic version of follow_huge_pmd() and that > > generic version is buggy for s930 because of commit e66f17ff7177 > > ("mm/hugetlb: take page table lock in follow_huge_pmd()"). Both of > > those are 4.0 material, though, so why is this needed for stable 3.12? > > Both commits 61f77eda and e66f17ff already made it into the 3.12 stable > tree, probably because of SLES 12 (actually that's how I noticed them). > > But I guess I screwed up the stable CC, stable@vger.kernel.org.#.v3.12 > somehow doesn't look right, not sure if the CC in the patch header > suffices. Looks like Jiri Slaby added the patches to 3.12, putting him > on CC now.
hm. I think I'll make it
Fixes: 61f77eda "mm/hugetlb: reduce arch dependent code around follow_huge_*" ... Cc: <stable@vger.kernel.org>
There's enough info here for the various tree maintainers to work out whether their kernel needs this fix.
| |