Re: [PATCH v2 0/2] mm: soft-offline: fix race against page allocation

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

On Fri, Oct 26, 2018 at 10:46:36AM +0200, Michal Hocko wrote:
> On Wed 22-08-18 10:00:25, Michal Hocko wrote:
> > On Wed 22-08-18 01:37:48, Naoya Horiguchi wrote:
> > > On Wed, Aug 15, 2018 at 03:43:34PM -0700, Andrew Morton wrote:
> > > > On Tue, 17 Jul 2018 14:32:30 +0900 Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> wrote:
> > > > 
> > > > > I've updated the patchset based on feedbacks:
> > > > > 
> > > > > - updated comments (from Andrew),
> > > > > - moved calling set_hwpoison_free_buddy_page() from mm/migrate.c to mm/memory-failure.c,
> > > > >   which is necessary to check the return code of set_hwpoison_free_buddy_page(),
> > > > > - lkp bot reported a build error when only 1/2 is applied.
> > > > > 
> > > > >   >    mm/memory-failure.c: In function 'soft_offline_huge_page':
> > > > >   > >> mm/memory-failure.c:1610:8: error: implicit declaration of function
> > > > >   > 'set_hwpoison_free_buddy_page'; did you mean 'is_free_buddy_page'?
> > > > >   > [-Werror=implicit-function-declaration]
> > > > >   >        if (set_hwpoison_free_buddy_page(page))
> > > > >   >            ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > > > >   >            is_free_buddy_page
> > > > >   >    cc1: some warnings being treated as errors
> > > > > 
> > > > >   set_hwpoison_free_buddy_page() is defined in 2/2, so we can't use it
> > > > >   in 1/2. Simply doing s/set_hwpoison_free_buddy_page/!TestSetPageHWPoison/
> > > > >   will fix this.
> > > > > 
> > > > > v1: https://lkml.org/lkml/2018/7/12/968
> > > > > 
> > > > 
> > > > Quite a bit of discussion on these two, but no actual acks or
> > > > review-by's?
> > > 
> > > Really sorry for late response.
> > > Xishi provided feedback on previous version, but no final ack/reviewed-by.
> > > This fix should work on the reported issue, but rewriting soft-offlining
> > > without PageHWPoison flag would be the better fix (no actual patch yet.)
> > 
> > If we can go with the later the I would obviously prefer that. I cannot
> > promise to work on the patch though. I can help with reviewing of
> > course.
> > 
> > If this is important enough that people are hitting the issue in normal
> > workloads then sure, let's go with the simple fix and continue on top of
> > that.
> 
> Naoya, did you have any chance to look at this or have any plans to look?
> I am willing to review and help with the overal design but I cannot
> really promise to work on the code.

I have a draft version of a patch to isolate a page in buddy-friendly manner
without PageHWPoison flag (that was written weeks ago, but I couldn't finish
because my other project interrupted me ...).
I'll post it after testing, especially confirming that hotplug code properly
reset the isolated page.

Thanks,
Naoya Horiguchi