Re: [RFC PATCH v2 01/16] mm,hwpoison: cleanup unused PageHuge() check

[Naoya Horiguchi]

On Fri, Oct 18, 2019 at 01:48:32PM +0200, Michal Hocko wrote:
> On Thu 17-10-19 16:21:08, Oscar Salvador wrote:
> > From: Naoya Horiguchi <n-horiguchi@xxxxxxxxxxxxx>
> > 
> > Drop the PageHuge check since memory_failure forks into memory_failure_hugetlb()
> > for hugetlb pages.
> > 
> > Signed-off-by: Oscar Salvador <osalvador@xxxxxxx>
> > Signed-off-by: Naoya Horiguchi <n-horiguchi@xxxxxxxxxxxxx>
> 
> s-o-b chain is reversed.
> 
> The code is a bit confusing. Doesn't this check aim for THP?

No, PageHuge() is false for thp, so this if branch is just dead code.

> AFAICS
> PageTransHuge(hpage) will split the THP or fail so PageTransHuge
> shouldn't be possible anymore, right?

Yes, that's right.

> But why does hwpoison_user_mappings
> still work with hpage then?

hwpoison_user_mappings() is called both from memory_failure() and
from memory_failure_hugetlb(), so it need handle both cases.

Thanks,
Naoya Horiguchi

> 
> > ---
> >  mm/memory-failure.c | 5 +----
> >  1 file changed, 1 insertion(+), 4 deletions(-)
> > 
> > diff --git a/mm/memory-failure.c b/mm/memory-failure.c
> > index 05c8c6df25e6..2cbadb58c7df 100644
> > --- a/mm/memory-failure.c
> > +++ b/mm/memory-failure.c
> > @@ -1345,10 +1345,7 @@ int memory_failure(unsigned long pfn, int flags)
> >  	 * page_remove_rmap() in try_to_unmap_one(). So to determine page status
> >  	 * correctly, we save a copy of the page flags at this time.
> >  	 */
> > -	if (PageHuge(p))
> > -		page_flags = hpage->flags;
> > -	else
> > -		page_flags = p->flags;
> > +	page_flags = p->flags;
> >  
> >  	/*
> >  	 * unpoison always clear PG_hwpoison inside page lock
> > -- 
> > 2.12.3
> 
> -- 
> Michal Hocko
> SUSE Labs
>