From: David Gibson <david@gibson.dropbear.id.au>
To: Andrew Morton <akpm@osdl.org>, Linus Torvalds <torvalds@osdl.org>
Cc: linuxppc64-dev@ozlabs.org, linux-kernel@vger.kernel.org
Subject: powerpc: More hugepage boundary case fixes
Date: Thu, 24 Nov 2005 13:34:56 +1100 [thread overview]
Message-ID: <20051124023456.GA3024@localhost.localdomain> (raw)
Please apply,
Blah. The patch [0] I recently sent fixing errors with
in_hugepage_area() and prepare_hugepage_range() for powerpc itself has
an off-by-one bug. Furthermore, the related functions
touches_hugepage_*_range() and within_hugepage_*_range() are also
buggy. Some of the bugs, like those addressed in [0] originated with
commit 7d24f0b8a53261709938ffabe3e00f88f6498df9 where we tweaked the
semantics of where hugepages are allowed. Other bugs have been there
essentially forever, and are due to the undefined behaviour of '<<'
with shift counts greater than the type width (LOW_ESID_MASK could
return non-zero for high ranges with the right congruences).
The good news is that I now have a testsuite which should pick up
things like this if they creep in again.
[0] "powerpc-fix-for-hugepage-areas-straddling-4gb-boundary" in -mm
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Index: working-2.6/include/asm-powerpc/page_64.h
===================================================================
--- working-2.6.orig/include/asm-powerpc/page_64.h 2005-11-24 11:15:38.000000000 +1100
+++ working-2.6/include/asm-powerpc/page_64.h 2005-11-24 13:09:28.000000000 +1100
@@ -103,8 +103,9 @@ extern unsigned int HPAGE_SHIFT;
#define HTLB_AREA_SIZE (1UL << HTLB_AREA_SHIFT)
#define GET_HTLB_AREA(x) ((x) >> HTLB_AREA_SHIFT)
-#define LOW_ESID_MASK(addr, len) (((1U << (GET_ESID(addr+len-1)+1)) \
- - (1U << GET_ESID(addr))) & 0xffff)
+#define LOW_ESID_MASK(addr, len) \
+ (((1U << (GET_ESID(min((addr)+(len)-1, 0x100000000UL))+1)) \
+ - (1U << GET_ESID(min((addr), 0x100000000UL)))) & 0xffff)
#define HTLB_AREA_MASK(addr, len) (((1U << (GET_HTLB_AREA(addr+len-1)+1)) \
- (1U << GET_HTLB_AREA(addr))) & 0xffff)
@@ -113,17 +114,21 @@ extern unsigned int HPAGE_SHIFT;
#define ARCH_HAS_SETCLEAR_HUGE_PTE
#define touches_hugepage_low_range(mm, addr, len) \
- (LOW_ESID_MASK((addr), (len)) & (mm)->context.low_htlb_areas)
+ (((addr) < 0x100000000UL) \
+ && (LOW_ESID_MASK((addr), (len)) & (mm)->context.low_htlb_areas))
#define touches_hugepage_high_range(mm, addr, len) \
- (HTLB_AREA_MASK((addr), (len)) & (mm)->context.high_htlb_areas)
+ ((((addr) + (len)) > 0x100000000UL) \
+ && (HTLB_AREA_MASK((addr), (len)) & (mm)->context.high_htlb_areas))
#define __within_hugepage_low_range(addr, len, segmask) \
- ((LOW_ESID_MASK((addr), (len)) | (segmask)) == (segmask))
+ ( (((addr)+(len)) <= 0x100000000UL) \
+ && ((LOW_ESID_MASK((addr), (len)) | (segmask)) == (segmask)))
#define within_hugepage_low_range(addr, len) \
__within_hugepage_low_range((addr), (len), \
current->mm->context.low_htlb_areas)
#define __within_hugepage_high_range(addr, len, zonemask) \
- ((HTLB_AREA_MASK((addr), (len)) | (zonemask)) == (zonemask))
+ ( ((addr) >= 0x100000000UL) \
+ && ((HTLB_AREA_MASK((addr), (len)) | (zonemask)) == (zonemask)))
#define within_hugepage_high_range(addr, len) \
__within_hugepage_high_range((addr), (len), \
current->mm->context.high_htlb_areas)
Index: working-2.6/arch/powerpc/mm/hugetlbpage.c
===================================================================
--- working-2.6.orig/arch/powerpc/mm/hugetlbpage.c 2005-11-24 11:15:38.000000000 +1100
+++ working-2.6/arch/powerpc/mm/hugetlbpage.c 2005-11-24 13:18:41.000000000 +1100
@@ -295,7 +295,7 @@ int prepare_hugepage_range(unsigned long
if (addr < 0x100000000UL)
err = open_low_hpage_areas(current->mm,
LOW_ESID_MASK(addr, len));
- if ((addr + len) >= 0x100000000UL)
+ if ((addr + len) > 0x100000000UL)
err = open_high_hpage_areas(current->mm,
HTLB_AREA_MASK(addr, len));
if (err) {
--
David Gibson | I'll have my music baroque, and my code
david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_
| _way_ _around_!
http://www.ozlabs.org/~dgibson
reply other threads:[~2005-11-24 2:35 UTC|newest]
Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20051124023456.GA3024@localhost.localdomain \
--to=david@gibson.dropbear.id.au \
--cc=akpm@osdl.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linuxppc64-dev@ozlabs.org \
--cc=torvalds@osdl.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.