Re: Fwd: Debian Hurd Installer problems

[Alex Sassmannshausen]

Hi Thomas,

I hope the information below is what you were looking for. Let me know
if it is not.

I have tested:

- gnumach-image-1.3.99-486_1.3.99.dfsg.git20120219-1_i386.deb
(installed by downloading file from snapshot.debian.org, extracting,
then copying across to /hurd/boot/gnumach.gz)
CHS: 1023/240/63 CTL=8 from BIOS
Fails with: 
start ext2fs: hd0: irq timeout: {original error pattern, eventually hanging 
with:}
end_request: I/O error, dev 03:04, sector 2

- gnumach-image-1.3.99-486_1.3.99.dfsg.git20120610-1_i386.deb
(installed through aptitude install gnumach on Wheezy, then copied
across to /hurd/boot/gnumach.gz)
CHS: 1023/240/63 CTL=8 from BIOS
Fails with:
start ext2fs: Hurd server bootstrap: ext2fs[device:hd0s4] exec inithd0:
irq timeout: {original error pattern, eventually hanging with:}
end_request: I/O error, dev 03:04, sector 3328880

- gnumach-image-1.3.99-486_1.3.99.dfsg.git20120710-1_i386.deb
(installed by downloading file from snapshot.debian.org, extracting,
then copying across to /hurd/boot/gnumach.gz)
CHS: 1023/240/63 CTL=8 from BIOS
Fails with: 
start ext2fs: Hurd server bootstrap: ext2fs[device:hd0s4] exechd0:
irq timeout: {original error pattern, eventually hanging with:}
end_request: I/O error, dev 03:04, sector 2130304

- gnumach-image-1.3.99-486_1.3.99.dfsg.git20121227-2_i386.deb
(installed by downloading file from snapshot.debian.org, extracting,
then copying across to /hurd/boot/gnumach.gz)
CHS: 1023/240/63 CTL=8 from BIOS
Fails with: 
start ext2fs: Hurd server bootstrap: ext2fs[device:hd0s4] exec init prochd0:
irq timeout: {original error pattern, eventually hanging with:}
end_request: I/O error, dev 03:04, sector 2130000
Upon restart (x2, quick succession):
start ext2fs: hd0:
irq timeout: {original error pattern, eventually hanging with:}
end_request: I/O error, dev 03:04, sector 0
Upon restart again (after a 5 minute wait):
start ext2fs: Hurd server bootstrap: ext2fs[device:hd0s4] exec inithd0:
irq timeout: {original error pattern, eventually hanging with:}
end_request: I/O error, dev 03:04, sector 3328880
irq timeout: {original error pattern, eventually hanging with:}
end_request: I/O error, dev 03:04, sector 2139840

I originally thought there was a pattern, whereby different version got
to different stages in the boot process, and failing reads on different
sectors; but the different results for the latest build seem to
contradict this…

I'm happy to continue testing in any direction if you think it's
valuable or worth it.
The CHS values seem to be consistent. 

Best wishes,

Alex


Thomas Schwinge <thomas@codesourcery.com> writes:

> Hi!
>
> On Fri, 1 Feb 2013 19:34:13 +0100, Samuel Thibault <samuel.thibault@gnu.org> 
> wrote:
>> Not immediately. IDE disks were working last time I tried on a real
>> machine, but perhaps these disks have bad timings.
>
> Wondering if it might be related to the changes I have recently done for
> solid-state disks?  But that only enlarged the timeout (and the "C/H/S"
> changes should be completely unrelated?), so I don't see any direct
> connection.  But Alex, would you please give a slightly older Debian
> gnumach-* package a try?  (You can retrieve one via
> <http://snapshot.debian.org/>, unpack the *.deb file and point the GRUB
> config to that kernel image.)  Just for a sanity check, please also see
> whether there are any changes in the C/H/S layout GNU Mach reports during
> booting.
>
>
> Grüße,
>  Thomas