linux-kernel - Re: unexplainable read errors, copy/diff-issue

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: <20061110135649.16cccca0.vsu@altlinux.ru>
Date:	Fri, 10 Nov 2006 13:56:49 +0300
From:	Sergey Vlasov <vsu@...linux.ru>
To:	Christoph Anton Mitterer <calestyo@...entia.net>
Cc:	linux-kernel@...r.kernel.org
Subject: Re: unexplainable read errors, copy/diff-issue

On Fri, 10 Nov 2006 03:01:52 +0100 Christoph Anton Mitterer wrote:

> The original post may be found under:
> http://marc.theaimsgroup.com/?t=116291314500001&r=1&w=2
>
> First of all my software/hardware data:
> CPU: 2x DualCore Opteron 275
> Mainboard: Tyan Thunder K8WE (S2895)
> Chipsets: Nvidia nForce professional 2200, Nvidia nForce professional
> 2050, AMD 8131
> Memory: Kingston ValueRAM 4x 1GB Registered ECC
> Harddisks: 1x PATA IBM/Hitachi, 2x SATA IBM/Hitachi
> Additional Devices/Drives: Plextor PX760A DVD/CD, TerraTec Aureon 7.1
> Universe soundcard, Hauppage Nova 500T DualDVB-T card.
> Distribution: Debian sid
> Kernel: self-compiled 2.6.18.2 (see below for .config) with applied EDAC
> patches
>
> The system should be cooled enough so I don't think that this comes from
> overheating issues. Nothing is overclocked.
>
> Ok the problem is the following:
> I copy large amounts of data (>15GB in several files each about 80MB,
> but it happens even for smaller amounts so size does not seem to be
> related) from one location to another.
> Then I diff the two locations.
> There are differences found!!!
> When I do diff again I find again differences but mostly in other files
> (so not always in the same place).
> Sometimes there are not differences (but this is rare as I'm testing
> with > 15GB)

So you have 4GB RAM, and most likely some memory is remapped above the
4GB address boundary.  Could you show the full dmesg output after boot?

Other things you can try:

 - Boot with mem=3072M (or some larger value which is still less than
   the amount of RAM below the 4GB boundary - the exact value could be
   found from the dmesg output) and check whether you can reproduce the
   corruption in this configuration.

 - Look in the BIOS setup for memory remapping options (Google indicates
   that it may be called "Hammer Configuration/Memory Hole Mapping" on
   this board).  Maybe you need to try different values (AFAIR there
   were some complaints about unstabilities with software remapping;
   cannot find the exact page now).

Content of type "application/pgp-signature" skipped