[x86] Stricter check in follow_huge_addr()

[x86] Stricter check in follow_huge_addr()

[Christoph Lameter]

From: Christoph Lameter <clameter@sgi.com>
Subject: Compound Pages: Stricter check in follow_huge_addr

The first page of the compound page is determined in follow_huge_addr()
but then PageCompound() only checks if the page is part of a compound page.
PageHead() allows checking if this is indeed the first page of the 
compound.

Cc: Jeremy Fitzhardinge <jeremy@goop.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Christoph Lameter <clameter@sgi.com>

---
 arch/x86/mm/hugetlbpage.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Index: linux-2.6.25-rc5-mm1/arch/x86/mm/hugetlbpage.c
===================================================================
--- linux-2.6.25-rc5-mm1.orig/arch/x86/mm/hugetlbpage.c	2008-03-26 20:43:56.796667401 -0700
+++ linux-2.6.25-rc5-mm1/arch/x86/mm/hugetlbpage.c	2008-03-26 20:44:55.526669888 -0700
@@ -178,7 +178,7 @@ follow_huge_addr(struct mm_struct *mm, u
 
 	page = &pte_page(*pte)[vpfn % (HPAGE_SIZE/PAGE_SIZE)];
 
-	WARN_ON(!PageCompound(page));
+	WARN_ON(!PageHead(page));
 
 	return page;
 }

Re: [x86] Stricter check in follow_huge_addr()

[Ingo Molnar]

* Christoph Lameter <clameter@sgi.com> wrote:

> From: Christoph Lameter <clameter@sgi.com>
> Subject: Compound Pages: Stricter check in follow_huge_addr
> 
> The first page of the compound page is determined in 
> follow_huge_addr() but then PageCompound() only checks if the page is 
> part of a compound page. PageHead() allows checking if this is indeed 
> the first page of the compound.

thanks Christoph, applied.

btw., what type of real-life problem does this address? Did this trigger 
somewhere? (if yes, do you have any URLs i could put into the 
changelog?)

	Ingo

Re: [x86] Stricter check in follow_huge_addr()

[Christoph Lameter]

On Thu, 27 Mar 2008, Ingo Molnar wrote:

> btw., what type of real-life problem does this address? Did this trigger 
> somewhere? (if yes, do you have any URLs i could put into the 
> changelog?)

Does not address any issues. I just saw it on review while thinking about 
how to optimize checks.