linux-kernel - Re: 3.1.3 2.6.37.6 2.6.33.4 kernel bug in pata

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]

Date:	Sat, 17 Dec 2011 01:15:17 +0100 (CET)
From:	Stefano Laporta <Stefano.Laporta@...infn.it>
To:	Alan Cox <alan@...rguk.ukuu.org.uk>
cc:	Stefano Laporta <Stefano.Laporta@...infn.it>,
	linux-kernel@...r.kernel.org,
	Bartlomiej Zolnierkiewicz <bzolnier@...il.com>
Subject: Re: 3.1.3 2.6.37.6 2.6.33.4 kernel bug in pata_oldpiix driver

I have been able to find another similar PC still working, a Pentium 90MHz
with Intel i430FX chipset, 32MB of RAM, an 1.2GB HD and a CD-ROM.
The kernels which crash on my Pentium 133MHz  crash even on the Pentium 90,
  giving exactly the same error messages.

So it seems not to be a problem of a single machine.

I switched on the debug options in kernel configuration, but this makes 
the bug (apparently) disappear. Memory overwrites are always difficult 
to catch.


On Mon, 28 Nov 2011, Alan Cox wrote:

>> BUG: unable to handle kernel paging request at 45230451
>
> That looks like junk from somewhere.
>
>> IP: [<c17bb1f3>] oldpiix_set_dmamode+0xb3/0x120
>
>>   [<c17bb3b0>] ? oldpiix_qc_issue+0x40/0x60
>>   [<c1788d8a>] ? ata_qc_issue+0x1ea/0x330
>
> So we were issuing a command, we called the oldpiix code, it decided that
> the timings were not loaded for this drive (ie it was those loaded for
> the other device on the bus)
>
> It will then have gone into oldpiix_set_piomode which ran succesfully
> implying that ap and adev were valid at that point but seem to have
> mysteriously gone invalid later.
>
> No idea, and the fact that changing things like the pre-empt behaviour or
> build options makes me wonder if it was scribbled on from elsewhere.
>
>
>
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/