linux-kernel - Re: [3.2+] BUG: scheduling while atomic: swapper/0/0/0x10000100

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] [day] [month] [year] [list]

Message-ID: <1327913505.2285.3.camel@scapa>
Date:	Mon, 30 Jan 2012 09:51:45 +0100
From:	Yves-Alexis Perez <corsac@...ian.org>
To:	linux-kernel@...r.kernel.org, Greg KH <greg@...ah.com>
Subject: Re: [3.2+] BUG: scheduling while atomic: swapper/0/0/0x10000100

(dropping linux-usb and relevant people from CC:)

On jeu., 2012-01-26 at 16:27 -0800, Kuninori Morimoto wrote:
> Hi All
> 
> > > > Specially when you consider the fact that
> > > > it's very recently that we dropped the ARCH dependency with commit
> > > > d242c110c43b55aaf3ebb3ceac1eeab0d452a177
> > > > 
> > > > > Kuninori, any thoughts about this?  Should I revert this from the
> > > > > tree?
> > > > 
> > > > I don't think so, I think this was a bad bisection.
> > > > 
> > > Maybe (especially since the first bad commit is a merge commit, so I'm
> > > not completely sure I did correctly the second bisect). But I'm puzzled
> > > why reverting the commit on top of 3.2.1 fixes that. And now that Felipe
> > > mentions it, USB Gadget is not even compiled, so I'm completely lost.
> > 
> > Hehehe :-) That's fun... Maybe it's some other instability and it just
> > didn't kick when you reverted that commit. It could be that the problem
> > is somewhere completely different and you just didn't fully debug it yet
> > :-)
> 
> I was very much surprised at this issue.
> As Felipe explained, renesas_usbhs was for SuperH system.
> x86 can not use it,
> but, if your PC has "renesas_usbhs.ko" as kernel module,
> and loaded it somehow...
> 

I redid the bisect a bit more thoroughly, using the (original) exact
same config as the Debian kernel. I get:

git bisect bad
ec7ae517537ae5c7b0b2cd7f562dfa3e7a05b954 is the first bad commit

with:

corsac@...pa: git bisect log
# bad: [805a6af8dba5dfdd35ec35dc52ec0122400b2610] Linux 3.2
# good: [c3b92c8787367a8bb53d57d9789b558f1295cc96] Linux 3.1
git bisect start 'v3.2' 'v3.1'
# good: [68d99b2c8efcb6ed3807a55569300c53b5f88be5] Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound
git bisect good 68d99b2c8efcb6ed3807a55569300c53b5f88be5
# bad: [0de9adf284ec20454ecf37ffd98e7e98ba7292d6] MAINTAINERS: Update eCryptfs maintainer address
git bisect bad 0de9adf284ec20454ecf37ffd98e7e98ba7292d6
# bad: [b4beb4bf9934d151bf4581a54ae028927374cb2a] Merge branch 'for-linus/i2c-3.2' of git://git.fluff.org/bjdooks/linux
git bisect bad b4beb4bf9934d151bf4581a54ae028927374cb2a
# bad: [df9d6985be2a7e7683c46e4c6ea608fc69f02b45] mm: do not drain pagevecs for mlockall(MCL_FUTURE)
git bisect bad df9d6985be2a7e7683c46e4c6ea608fc69f02b45
# good: [b57287ba497b63a0d87a058631bbddfed9392e9f] video: platinumfb: Add __devexit_p at necessary place
git bisect good b57287ba497b63a0d87a058631bbddfed9392e9f
# bad: [41684f67af75b04152a1714e1a5375dfb00ee3da] Merge branch 'gpio/next' of git://git.secretlab.ca/git/linux-2.6
git bisect bad 41684f67af75b04152a1714e1a5375dfb00ee3da
# good: [ffaac8f45bfb2dffb78179baa5740de34058eef8] [SCSI] libsas: Allow expander T-T attachments
git bisect good ffaac8f45bfb2dffb78179baa5740de34058eef8
# good: [d4a759a954072dcb0a0bbd64ec4bc5252182d032] [SCSI] megaraid_sas: Fix mismatch in megasas_reset_fusion() mutex lock-unlock
git bisect good d4a759a954072dcb0a0bbd64ec4bc5252182d032
# good: [590134fa78fbdbe5fea78c7ae0b2c3364bc9572f] [SCSI] qla4xxx: export address/port of connection (fix udev disk names)
git bisect good 590134fa78fbdbe5fea78c7ae0b2c3364bc9572f
# good: [76c05c8a0d56faf210cb9681786bb3e17cd59793] gpio: pl061: add DT binding support
git bisect good 76c05c8a0d56faf210cb9681786bb3e17cd59793
# good: [83eaea22bdfc9e1cec88f81be5b64f30f6c37e8b] Revert "ceph: don't truncate dirty pages in invalidate work thread"
git bisect good 83eaea22bdfc9e1cec88f81be5b64f30f6c37e8b
# good: [3310f7541f0c991b51324a7712db51fb8f912601] ceph: fix 32-bit ino numbers
git bisect good 3310f7541f0c991b51324a7712db51fb8f912601
# good: [97d2eb13a019ec09cc1a7ea2d3705c0b117b3c0d] Merge branch 'for-linus' of git://ceph.newdream.net/git/ceph-client
git bisect good 97d2eb13a019ec09cc1a7ea2d3705c0b117b3c0d
# good: [d92ef29a6fa971d9e314e412cd9c09757906411a] h8300: Move gpio.h to gpio-internal.h
git bisect good d92ef29a6fa971d9e314e412cd9c09757906411a
# bad: [ec7ae517537ae5c7b0b2cd7f562dfa3e7a05b954] Merge git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-misc-2.6
git bisect bad ec7ae517537ae5c7b0b2cd7f562dfa3e7a05b954

As the first bad commit is again a merge commit, I'll redo the bisect
with both ends commit (so 97d2eb1 590134f afaict).

Regards,
-- 
Yves-Alexis

Download attachment "signature.asc" of type "application/pgp-signature" (837 bytes)