linux-kernel - Re: Linux 2.6.35.2

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <AANLkTikfjGfqGFkgyRKnTCi6gPSvW3Td1O1ViQUoeZVF@mail.gmail.com>
Date:	Sat, 14 Aug 2010 22:34:13 +0200
From:	Sedat Dilek <sedat.dilek@...glemail.com>
To:	Linus Torvalds <torvalds@...ux-foundation.org>
Cc:	Ed Tomlinson <edt@....ca>, Greg KH <gregkh@...e.de>,
	LKML <linux-kernel@...r.kernel.org>, penberg@...nel.org,
	arthur@....ro, wylda@...ny.cz
Subject: Re: Linux 2.6.35.2

On Sat, Aug 14, 2010 at 8:47 PM, Linus Torvalds
<torvalds@...ux-foundation.org> wrote:
> On Sat, Aug 14, 2010 at 11:27 AM, Ed Tomlinson <edt@....ca> wrote:
>>
>> It seems to be helping here.  Without it 35.2 would either stall in boot or give lots of tracebacks.  With it things
>> start normally and my background stuff works as expected.
>
> Ok, thanks. I've committed it to the -git tree.
>
> Just to verify: you have CONFIG_HIGHPTE enabled on a x86-32 kernel,
> right? I'd have expected this to not ever show up anywhere else, and
> I'm just verifying that there isn't anything else going on. I'm
> appending the commit message, the patch hasn't changed.
>
>                            Linus
> ---
>
> Author: Linus Torvalds <torvalds@...ux-foundation.org>
> Date:   Sat Aug 14 11:44:56 2010 -0700
>
>    mm: fix page table unmap for stack guard page properly
>
>    We do in fact need to unmap the page table _before_ doing the whole
>    stack guard page logic, because if it is needed (mainly 32-bit x86 with
>    PAE and CONFIG_HIGHPTE, but other architectures may use it too) then it
>    will do a kmap_atomic/kunmap_atomic.
>
>    And those kmaps will create an atomic region that we cannot do
>    allocations in.  However, the whole stack expand code will need to do
>    anon_vma_prepare() and vma_lock_anon_vma() and they cannot do that in an
>    atomic region.
>
>    Now, a better model might actually be to do the anon_vma_prepare() when
>    _creating_ a VM_GROWSDOWN segment, and not have to worry about any of
>    this at page fault time.  But in the meantime, this is the
>    straightforward fix for the issue.
>
>    See https://bugzilla.kernel.org/show_bug.cgi?id=16588 for details.
>
>    Reported-by: Wylda <wylda@...ny.cz>
>    Reported-by: Sedat Dilek <sedat.dilek@...il.com>
>    Reported-by: Mike Pagano <mpagano@...too.org>
>    Reported-by: François Valenduc <francois.valenduc@...ablenet.be>
>    Tested-by: Ed Tomlinson <edt@....ca>
>    Cc: Pekka Enberg <penberg@...nel.org>
>    Cc: Greg KH <gregkh@...e.de>
>    Cc: stable@...nel.org
>    Signed-off-by: Linus Torvalds <torvalds@...ux-foundation.org>
>
>  mm/memory.c |   13 ++++++-------
>  1 files changed, 6 insertions(+), 7 deletions(-)
>

OK, I applied this patch on top of 2.6.35-git15.

Then, stressed the new kernel with a mesa-from-git build.
Looks fine on first sight.

Feel free to add:

     Tested-by: Sedat Dilek <sedat.dilek@...il.com>

- Sedat -

P.S.: Attached dmesg and kernel-config files

$ grep CONFIG_HIGHPTE /boot/config-$(uname -r)
CONFIG_HIGHPTE=y

$ cat /proc/version
Linux version 2.6.35-git15.sd.2-686 (Debian
2.6.35-6~git15.sd.2+dileks.1) (sedat.dilek@...il.com) (gcc version
4.4.5 20100728 (prerelease) (Debian 4.4.4-8) ) #2 SMP Sat Aug 14
22:12:55 CEST 2010

View attachment "dmesg.txt" of type "text/plain" (48116 bytes)

Download attachment "config-2.6.35-git15.sd.2-686" of type "application/octet-stream" (120398 bytes)