Re: [v3 PATCH 2/5] mm: filemap: check if THP has hwpoisoned subpage for PMD page fault

[Yang Shi]

On Wed, Oct 6, 2021 at 12:54 PM Peter Xu <peterx@xxxxxxxxxx> wrote:
>
> On Mon, Oct 04, 2021 at 01:13:07PM -0700, Yang Shi wrote:
> > On Mon, Oct 4, 2021 at 12:41 PM Kirill A. Shutemov <kirill@xxxxxxxxxxxxx> wrote:
> > >
> > > On Mon, Oct 04, 2021 at 11:17:29AM -0700, Yang Shi wrote:
> > > > On Mon, Oct 4, 2021 at 7:06 AM Kirill A. Shutemov <kirill@xxxxxxxxxxxxx> wrote:
> > > > >
> > > > > On Thu, Sep 30, 2021 at 02:53:08PM -0700, Yang Shi wrote:
> > > > > > diff --git a/mm/filemap.c b/mm/filemap.c
> > > > > > index dae481293b5d..2acc2b977f66 100644
> > > > > > --- a/mm/filemap.c
> > > > > > +++ b/mm/filemap.c
> > > > > > @@ -3195,12 +3195,12 @@ static bool filemap_map_pmd(struct vm_fault *vmf, struct page *page)
> > > > > >       }
> > > > > >
> > > > > >       if (pmd_none(*vmf->pmd) && PageTransHuge(page)) {
> > > > > > -         vm_fault_t ret = do_set_pmd(vmf, page);
> > > > > > -         if (!ret) {
> > > > > > -                 /* The page is mapped successfully, reference consumed. */
> > > > > > -                 unlock_page(page);
> > > > > > -                 return true;
> > > > > > -         }
> > > > > > +             vm_fault_t ret = do_set_pmd(vmf, page);
> > > > > > +             if (!ret) {
> > > > > > +                     /* The page is mapped successfully, reference consumed. */
> > > > > > +                     unlock_page(page);
> > > > > > +                     return true;
> > > > > > +             }
> > > > > >       }
> > > > > >
> > > > > >       if (pmd_none(*vmf->pmd)) {
> > > > >
> > > > > Hm. Is it unrelated whitespace fix?
> > > >
> > > > It is a coding style clean up. I thought it may be overkilling to have
> > > > a separate patch. Do you prefer separate one?
> > >
> > > Maybe. I tried to find what changed here. It's confusing.
> >
> > Yeah, maybe. Anyway I will separate the real big fix and the cleanup
> > into two patches. This may be helpful for backporting too.
>
> Or maybe we just don't touch it until there's need for a functional change?  I
> feel it a pity to lose the git blame info for reindent-only patches, but no
> strong opinion, because I know many people don't think the same and I'm fine
> with either ways.

TBH I really don't think keeping old "git blame" info should be an
excuse to avoid any coding style cleanup.

>
> Another side note: perhaps a comment above pageflags enum on PG_has_hwpoisoned
> would be nice?  I saw that we've got a bunch of those already.

I was thinking about that, but it seems PG_double_map doesn't have
comment there either so I didn't add.

>
> Thanks,
>
> --
> Peter Xu
>