From: Ralf Baechle <ralf@linux-mips.org>
To: Linus Torvalds <torvalds@osdl.org>, Andrew Morton <akpm@osdl.org>
Cc: linux-kernel@vger.kernel.org,
Atsushi Nemoto <anemo@mba.ocn.ne.jp>,
Ralf Baechle <ralf@linux-mips.org>
Subject: [PATCH 1/3] Fix COW D-cache aliasing on fork
Date: Thu, 19 Oct 2006 17:35:46 +0100 [thread overview]
Message-ID: <1161275748231-git-send-email-ralf@linux-mips.org> (raw)
From: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
Problem:
1. There is a process containing two thread (T1 and T2). The
thread T1 calls fork(). Then dup_mmap() function called on T1 context.
static inline int dup_mmap(struct mm_struct *mm, struct mm_struct *oldmm)
...
flush_cache_mm(current->mm);
... /* A */
(write-protect all Copy-On-Write pages)
... /* B */
flush_tlb_mm(current->mm);
...
2. When preemption happens between A and B (or on SMP kernel), the
thread T2 can run and modify data on COW pages without page fault
(modified data will stay in cache).
3. Some time after fork() completed, the thread T2 may cause a page
fault by write-protect on a COW page.
4. Then data of the COW page will be copied to newly allocated
physical page (copy_cow_page()). It reads data via kernel mapping.
The kernel mapping can have different 'color' with user space
mapping of the thread T2 (dcache aliasing). Therefore
copy_cow_page() will copy stale data. Then the modified data in
cache will be lost.
In order to allow architecture code to deal with this problem allow
architecture code to override copy_user_highpage() by defining
__HAVE_ARCH_COPY_USER_HIGHPAGE in <asm/page.h>.
The main part of this patch was originally written by Ralf Baechle;
Atushi Nemoto did the the debugging.
Signed-off-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
include/linux/highmem.h | 4 ++++
1 file changed, 4 insertions(+)
Index: upstream-linus/include/linux/highmem.h
===================================================================
--- upstream-linus.orig/include/linux/highmem.h 2006-10-17 00:14:43.000000000 +0100
+++ upstream-linus/include/linux/highmem.h 2006-10-17 00:15:21.000000000 +0100
@@ -94,6 +94,8 @@ static inline void memclear_highpage_flu
kunmap_atomic(kaddr, KM_USER0);
}
+#ifndef __HAVE_ARCH_COPY_USER_HIGHPAGE
+
static inline void copy_user_highpage(struct page *to, struct page *from, unsigned long vaddr)
{
char *vfrom, *vto;
@@ -107,6 +109,8 @@ static inline void copy_user_highpage(st
smp_wmb();
}
+#endif
+
static inline void copy_highpage(struct page *to, struct page *from)
{
char *vfrom, *vto;
next reply other threads:[~2006-10-19 16:35 UTC|newest]
Thread overview: 46+ messages / expand[flat|nested] mbox.gz Atom feed top
2006-10-19 16:35 Ralf Baechle [this message]
2006-10-19 16:35 ` [PATCH 2/3] Pass vma argument to copy_user_highpage() Ralf Baechle
2006-10-19 16:35 ` [PATCH 3/3] MIPS: Fix COW D-cache aliasing on fork Ralf Baechle
2006-10-22 2:32 ` Ralf Baechle
2006-10-19 17:46 ` [PATCH 1/3] " Nick Piggin
2006-10-19 18:13 ` Ralf Baechle
2006-10-19 18:48 ` Nick Piggin
2006-10-19 22:59 ` David Miller
2006-10-20 14:39 ` Nick Piggin
2006-10-20 15:49 ` Linus Torvalds
2006-10-20 15:57 ` Nick Piggin
2006-10-20 16:36 ` Linus Torvalds
2006-10-20 16:47 ` Nick Piggin
2006-10-20 17:16 ` Linus Torvalds
2006-10-20 17:37 ` Nick Piggin
2006-10-21 0:46 ` Ralf Baechle
2006-10-20 19:36 ` David Miller
2006-10-20 19:54 ` Linus Torvalds
2006-10-20 19:58 ` David Miller
2006-10-20 20:10 ` Linus Torvalds
2006-10-20 20:59 ` Russell King
2006-10-20 21:06 ` David Miller
2006-10-20 21:17 ` Russell King
2006-10-20 21:30 ` David Miller
2006-10-20 21:12 ` Linus Torvalds
2006-10-20 21:28 ` Russell King
2006-10-20 21:41 ` Ralf Baechle
2006-10-21 16:28 ` Atsushi Nemoto
2006-10-20 21:49 ` Ralf Baechle
2006-10-20 22:02 ` Linus Torvalds
2006-10-20 22:22 ` David Miller
2006-10-20 22:51 ` Ralf Baechle
2006-10-20 23:28 ` Linus Torvalds
2006-10-21 0:06 ` Ralf Baechle
2006-10-21 0:38 ` Linus Torvalds
2006-10-21 1:29 ` Paul Mackerras
2006-10-21 2:11 ` David Miller
2006-10-21 2:37 ` Linus Torvalds
2006-10-21 2:46 ` David Miller
2006-10-21 18:27 ` Ralf Baechle
2006-10-22 1:34 ` Ralf Baechle
2006-12-02 9:49 ` Russell King
2006-10-23 8:50 ` Martin Schwidefsky
2006-10-20 16:05 ` Ralf Baechle
2006-10-20 16:30 ` Nick Piggin
2006-10-20 19:23 ` David Miller
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=1161275748231-git-send-email-ralf@linux-mips.org \
--to=ralf@linux-mips.org \
--cc=akpm@osdl.org \
--cc=anemo@mba.ocn.ne.jp \
--cc=linux-kernel@vger.kernel.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox