Re: [PATCH 2/2] [BUGFIX] pagemap: fix pfn calculation for hugepage
From: Daisuke Nishimura
Date: Fri Mar 19 2010 - 04:59:55 EST
On Fri, 19 Mar 2010 16:10:23 +0900, KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx> wrote:
> On Fri, 19 Mar 2010 15:26:36 +0900
> Naoya Horiguchi <n-horiguchi@xxxxxxxxxxxxx> wrote:
>
> > When we look into pagemap using page-types with option -p, the value
> > of pfn for hugepages looks wrong (see below.)
> > This is because pte was evaluated only once for one vma
> > although it should be updated for each hugepage. This patch fixes it.
> >
> > $ page-types -p 3277 -Nl -b huge
> > voffset offset len flags
> > 7f21e8a00 11e400 1 ___U___________H_G________________
> > 7f21e8a01 11e401 1ff ________________TG________________
> > 7f21e8c00 11e400 1 ___U___________H_G________________
> > 7f21e8c01 11e401 1ff ________________TG________________
> > ^^^
> > should not be the same
> >
> > With this patch applied:
> >
> > $ page-types -p 3386 -Nl -b huge
> > voffset offset len flags
> > 7fec7a600 112c00 1 ___UD__________H_G________________
> > 7fec7a601 112c01 1ff ________________TG________________
> > 7fec7a800 113200 1 ___UD__________H_G________________
> > 7fec7a801 113201 1ff ________________TG________________
> > ^^^
> > OK
> >
> Hmm. Is this bug ? To me, it's just shown in hugepage's pagesize, by design.
>
> _And_, Doesn't this patch change behavior of walk_pagemap_range() implicitly ?
> No influence to other users ? (as memcontrol.c. in mmotm. Ask Nishimura-san ;)
>
>From the view point of memcg, this change in walk_pagemap_range() has no influence,
because memcg does "if (is_vm_hugetlb_page(vma)) break" before calling walk_pagemap_range().
But we must check other callers too.
Thanks,
Daisuke Nishimura.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/