Re: [PATCH v1 4/4] mm/memory-failure: check __PG_HWPOISON separately from PAGE_FLAGS_CHECK_AT_*

[Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>]

On Thu, Jul 23, 2015 at 01:37:02PM -0700, Andrew Morton wrote:
> On Thu, 16 Jul 2015 01:41:56 +0000 Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> wrote:
> 
> > The race condition addressed in commit add05cecef80 ("mm: soft-offline: don't
> > free target page in successful page migration") was not closed completely,
> > because that can happen not only for soft-offline, but also for hard-offline.
> > Consider that a slab page is about to be freed into buddy pool, and then an
> > uncorrected memory error hits the page just after entering __free_one_page(),
> > then VM_BUG_ON_PAGE(page->flags & PAGE_FLAGS_CHECK_AT_PREP) is triggered,
> > despite the fact that it's not necessary because the data on the affected
> > page is not consumed.
> > 
> > To solve it, this patch drops __PG_HWPOISON from page flag checks at
> > allocation/free time. I think it's justified because __PG_HWPOISON flags is
> > defined to prevent the page from being reused and setting it outside the
> > page's alloc-free cycle is a designed behavior (not a bug.)
> > 
> > And the patch reverts most of the changes from commit add05cecef80 about
> > the new refcounting rule of soft-offlined pages, which is no longer necessary.
> > 
> > ...
> >
> > --- v4.2-rc2.orig/mm/memory-failure.c
> > +++ v4.2-rc2/mm/memory-failure.c
> > @@ -1723,6 +1723,9 @@ int soft_offline_page(struct page *page, int flags)
> >  
> >  	get_online_mems();
> >  
> > +	if (get_pageblock_migratetype(page) != MIGRATE_ISOLATE)
> > +		set_migratetype_isolate(page, true);
> > +
> >  	ret = get_any_page(page, pfn, flags);
> >  	put_online_mems();
> >  	if (ret > 0) { /* for in-use pages */
> 
> This patch gets build-broken by your
> mm-page_isolation-make-set-unset_migratetype_isolate-file-local.patch,
> which I shall drop.

I apologize this build failure. At first I planned to add another hwpoison patch
after this to remove this migratetype thing separately, but I was not 100% sure
of the correctness, so I did not include it in this version.
But Vlastimil's cleanup patch showed me that using MIGRATE_ISOLATE at free time
(, which is what soft offline code does now,) is wrong (or not an expected usage).
So I shouldn't have reverted the above part.

So I want the patch "mm, page_isolation: make set/unset_migratetype_isolate()
file-local" to be merged first, and I'd like to update this hwpoison before
going into mmotm. Could you drop this series from your tree for now?
I'll repost the next version probably next week.

Thanks,
Naoya Horiguchi
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>