lkml.org 
[lkml]   [2012]   [Sep]   [25]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
From
SubjectRe: [PATCH] pagemap: fix wrong KPF_THP on slab pages
Date
On Tue, Sep 25, 2012 at 05:20:48PM -0700, David Rientjes wrote:
> On Tue, 25 Sep 2012, Naoya Horiguchi wrote:
>
> > KPF_THP can be set on non-huge compound pages like slab pages, because
> > PageTransCompound only sees PG_head and PG_tail. Obviously this is a bug
> > and breaks user space applications which look for thp via /proc/kpageflags.
> > Currently thp is constructed only on anonymous pages, so this patch makes
> > KPF_THP be set when both of PageAnon and PageTransCompound are true.
> >
> > Changelog in v2:
> > - add a comment in code
> >
> > Signed-off-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
>
> Wouldn't PageTransCompound(page) && !PageHuge(page) && !PageSlab(page) be
> better for a future extension of thp support?

Yes, this saves us an additional change when thp starts handling pagecaches.
Andrew, can you replace the previous version in -mm tree with new one below?

Thanks,
Naoya
---
From: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Date: Tue, 25 Sep 2012 21:30:25 -0400
Subject: [PATCH v3] kpageflags: fix wrong KPF_THP on slab pages

KPF_THP can be set on non-huge compound pages like slab pages, because
PageTransCompound only sees PG_head and PG_tail. Obviously this is a bug
and breaks user space applications which look for thp via /proc/kpageflags.
This patch rules out setting KPF_THP wrongly by additional PageSlab check.

Changelog in v3:
- check PageSlab instead of PageAnon
- fix patch subject

Changelog in v2:
- add a comment in code

Signed-off-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
---
fs/proc/page.c | 7 ++++++-
1 file changed, 6 insertions(+), 1 deletion(-)

diff --git a/fs/proc/page.c b/fs/proc/page.c
index 7fcd0d6..e36d1f3 100644
--- a/fs/proc/page.c
+++ b/fs/proc/page.c
@@ -115,7 +115,12 @@ u64 stable_page_flags(struct page *page)
u |= 1 << KPF_COMPOUND_TAIL;
if (PageHuge(page))
u |= 1 << KPF_HUGE;
- else if (PageTransCompound(page))
+ /*
+ * PageTransCompound can be true for slab pages because it just sees
+ * PG_head/PG_head, so we need to check PageSlab to make sure the given
+ * page is a thp, not a non-huge compound page.
+ */
+ else if (PageTransCompound(page) && !PageSlab(page))
u |= 1 << KPF_THP;

/*
--
1.7.11.4


\
 
 \ /
  Last update: 2012-09-26 04:41    [W:0.074 / U:0.612 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site