Re: Bug: get EXT3-fs error Allocating block in system zone

[Andrew Morton <akpm@linux-foundation.org>]

On Tue, 04 Dec 2007 21:54:57 +0100
Marco Gatti <marco@pianoinflames.de> wrote:

> Hello,
> 
> I have a brand new Fujitsu-Siemens Celsius W360 pc with a FSC D2587-A1 
> motherboard. It has a intel q35 chipset. In bios I have the sata 
> controller in pure AHCI mode (legacy pata disabled). On windows 
> everthing works fine. So mem is ok, too, tested with memtest+ overnight. 
> No failures with hdd or motherboard on windows. There's latest avaiable 
> BIOS version on the machine.
> 
> Specific system data see below.
> 
> I first tried self compiled kernel 2.6.24-rc3 and now at the end 
> 2.6.23.9 to have a stable version. I got compiled e1000 module for 
> network card onboard under latest gentoo stable version.
> 
> I first got acpi failure on boot, pci=nommconf worked first time, than I 
> patched with q35.pci.patch from 
> http://bugs.gentoo.org/show_bug.cgi?id=198810 and boot worked without 
> nommconf.

http://bugs.gentoo.org/attachment.cgi?id=135761&action=view

It disables PCI BARs during sizing.  ISTR Linus opining that this was the
wrong thing to do?

ISTR him having opinions on mmconfig too ;)

We should not require people to use obscure boot options to get their
machines working.

> But the effect is under every circumstances described above that I got 
> after an unspecific time EXT3-fs errors. I tried to use different 
> partitions, one for root and data, got errors on both.
> 
> Dec  3 15:05:34 adira EXT3-fs error (device sdb4): ext3_new_block: 
> Allocating block in system zone - blocks from 74907667, length 1
> 
> I tried different hard disks, different sizes of partitions, always the 
> same issue. I always saw that mem is fully cached (I have 8GB RAM!). 
> After that, the ext3 has severe faults fsck.ext3 repaired them, but an 
> amount of data was lost.
> 
> I also tried with different file systems (reiserfs3, xfs), also kernel 
> trace errors, so I got back to ext3.
> 
> Can't believe that it's a pure fs-error. Is it an ahci.c issue? Or a 
> problem with acpi and memory management?

At a guess I'd say the disk system is being bad.  it might be a hardware
failure too - it's a new system.

> Any ideas for fixing this?
> 

Nope, sorry.