On Thu, Mar 10, 2016 at 11:52:54AM -0800, Andrew Morton wrote:
> On Thu, 10 Mar 2016 17:54:06 +0300 "Kirill A. Shutemov" <kirill.shutemov@xxxxxxxxxxxxxxx> wrote:
>
> > split_huge_pmd() tries to munlock page with munlock_vma_page(). That
> > requires the page to locked.
> >
> > If the is locked by caller, we would get a deadlock:
> >
> > ...
> >
> > I don't think the deadlock is triggerable without split_huge_page()
> > simplifilcation patchset.
> >
> > But munlock_vma_page() here is wrong: we want to munlock the page
> > unconditionally, no need in rmap lookup, that munlock_vma_page() does.
> >
> > Let's use clear_page_mlock() instead. It can be called under ptl.
> >
> > Signed-off-by: Kirill A. Shutemov <kirill.shutemov@xxxxxxxxxxxxxxx>
> > Fixes: ee0b79212791 ("thp: allow mlocked THP again")
>
> This is the incorrect hash (or something weird happened at my end).
> I'm seeing
>
> commit e90309c9f7722db4ff5bce3b9e6e04d1460f2553
> Author: Kirill A. Shutemov <kirill.shutemov@xxxxxxxxxxxxxxx>
> Date: Fri Jan 15 16:54:33 2016 -0800
>
> thp: allow mlocked THP again
>
> That's the second time this has happened this week so please
> double-check whatever you're doing here?
Ouch. Sorry. It's hash from mhocko/mm.git since-4.4, not mainline.
> The patch itself doesn't apply to mainline, which is a bit strange
> given that it "Fixes" a bug in an already-mainlined patch. The patch
> as-sent depends upon
> thp-rewrite-freeze_page-unfreeze_page-with-generic-rmap-walkers.patch,
> so I have queued it after that patch.
I think that's fine.
--
Kirill A. Shutemov
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxx. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>