linux-kernel - Re: [GIT PULL] SLAB changes for v3.10

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]

Date:	Wed, 8 May 2013 08:45:19 -0700
From:	Tony Lindgren <tony@...mide.com>
To:	Christoph Lameter <cl@...ux.com>
Cc:	Pekka Enberg <penberg@...nel.org>,
	Linus Torvalds <torvalds@...ux-foundation.org>,
	Andrew Morton <akpm@...ux-foundation.org>,
	linux-kernel@...r.kernel.org
Subject: Re: [GIT PULL] SLAB changes for v3.10

* Christoph Lameter <cl@...ux.com> [130508 07:01]:
> On Tue, 7 May 2013, Tony Lindgren wrote:
> 
> > OK got it narrowed down to CONFIG_DEBUG_SPINLOCK=y causing the problem
> > with commit 8a965b3b. Ain't nothing like bisecting and booting and then
> > diffing .config files on top of that.
> >
> > > Without reverting 8a965b3b I'm getting:
> 
> The patch (commit 8a965b3baa89ffedc73c0fbc750006c631012ced) merely changed the sequence of
> slab creation to address an issue in SLAB.
> 
> Hmmm.. But if KMALLOC_SHIFT_LOW is higher than 6 or 7 then the creation of
> the non-power of two slab could be skipped as a result of the patch. But
> we should not need them in those cases.
> 
> Can I see the kernel config?

Attached are minimal defconfig-bad and defconfig-good with pretty
much everything disabled. These boot vexpress in qemu using earlyprintk
for the output.
 
> What is the value of KMALLOC_SHIFT_LOW?

It's 6.
 
> Can you figure out which kernel slab the function is trying to access?

I can certainly debug it further, but it's also pretty easy to reproduce:

1. Download and build qemu-linaro from:

   https://git.linaro.org/gitweb?p=qemu/qemu-linaro.git

   The last time I tried the stock qemu I could not get vexpress to
   boot with it, don't know if that's still the case. But the above
   works for me.

2. Cross compile kernel with the attached defconfig-bad

   $ ARCH=arm CROSS_COMPILE=... make zImage

3. Try to boot the kernel in qemu

   $ qemu-system-arm -machine vexpress-a9 -m 1024 -curses -net nic \
	-net user -serial stdio -append "console=ttyAMA0,115200n8 \
	root=/dev/mmcblk0p2 ro rootwait physmap.enabled=0 \
	debug earlyprintk" -kernel zImage

Regards,

Tony

View attachment "defconfig-bad" of type "text/plain" (1922 bytes)

View attachment "defconfig-good" of type "text/plain" (1922 bytes)