Re: [PATCH 3/3] memory-hotplug: bug fix race between isolation and allocation

[Mel Gorman <mgorman@suse.de>]

On Thu, Sep 06, 2012 at 01:49:03PM +0900, Minchan Kim wrote:
> > > __offline_isolated_pages
> > > /*
> > >  * BUG_ON hit or offline page
> > >  * which is used by someone
> > >  */
> > > BUG_ON(!PageBuddy(page A));
> > > 
> > 
> > offline_page calling BUG_ON because someone allocated the page is
> > ridiculous. I did not spot where that check is but it should be changed. The
> > correct action is to retry the isolation.
> 
> It is where __offline_isolated_pges.
> 
> ..
>         while (pfn < end_pfn) {
>                 if (!pfn_valid(pfn)) {
>                         pfn++;
>                         continue;
>                 }    
>                 page = pfn_to_page(pfn);
>                 BUG_ON(page_count(page));
>                 BUG_ON(!PageBuddy(page)); <---- HERE
>                 order = page_order(page);
> ...
> 
> Comment of offline_isolated_pages says following as.
> 
>         We cannot do rollback at this point
> 
> So if the comment is true, BUG_ON does make sense to me.

It's massive overkill. I see no reason why it cannot return EBUSY all the
way back up to offline_pages() and retry with the migration step.  It would
both remove that BUG_ON and improve reliability of memory hot-remove.

> But I don't see why we can't retry it as I look thorugh code.
> Anyway, It's another story which isn't related to this patch.
> 

True.

> > 
> > > Signed-off-by: Minchan Kim <minchan@kernel.org>
> > 
> > At no point in the changelog do you actually say what he patch does :/
> 
> Argh, I will do.
> 
> > 
> > > ---
> > >  mm/page_isolation.c |    5 ++++-
> > >  1 file changed, 4 insertions(+), 1 deletion(-)
> > > 
> > > diff --git a/mm/page_isolation.c b/mm/page_isolation.c
> > > index acf65a7..4699d1f 100644
> > > --- a/mm/page_isolation.c
> > > +++ b/mm/page_isolation.c
> > > @@ -196,8 +196,11 @@ __test_page_isolated_in_pageblock(unsigned long pfn, unsigned long end_pfn)
> > >  			continue;
> > >  		}
> > >  		page = pfn_to_page(pfn);
> > > -		if (PageBuddy(page))
> > > +		if (PageBuddy(page)) {
> > > +			if (get_page_migratetype(page) != MIGRATE_ISOLATE)
> > > +				break;
> > >  			pfn += 1 << page_order(page);
> > > +		}
> > 
> > It is possible the page is moved to the MIGRATE_ISOLATE list between when
> > the page was freed to the buddy allocator and this check was made. The
> > page->index information is stale and the impact is that the hotplug
> > operation fails when it could have succeeded. That said, I think it is a
> > very unlikely race that will never happen in practice.
> 
> I understand you mean move_freepages which I have missed. Right?

Yes.

> Then, I will fix it, too.
> 
> > 
> > More importantly, the effect of this path is that EBUSY gets bubbled all
> > the way up and the hotplug operations fails. This is fine but as the page
> > is free at the time this problem is detected you also have the option
> > of moving the PageBuddy page to the MIGRATE_ISOLATE list at this time
> > if you take the zone lock. This will mean you need to change the name of
> > test_pages_isolated() of course.
> 
> Sorry, I can't get your point. Could you elaborate it more?

You detect a PageBuddy page but it's on the wrong list. Instead of returning
and failing memory-hotremove, move the free page to the correct list at
the time it is detected.

> Is it related to this patch?

No, it's not important and was a suggestion on how it could be made
better. However, retrying hot-remove would be even better again. I'm not
suggesting it be done as part of this series.

-- 
Mel Gorman
SUSE Labs