[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#404950: Installation Report [mips] [rc1] [Failure] SWARM (Broadcom BCM91250a)



Karsten, can you check if this problem is still there with current
kernels?  i.e. 2.6.23 or 2.6.24


* Karsten Merker <merker@debian.org> [2007-01-05 01:22]:
> On Thu, Jan 04, 2007 at 08:30:32PM +0100, Karsten Merker wrote:
> > On Thu, Jan 04, 2007 at 04:48:09PM +0100, Frans Pop wrote:
> 
> > > The device that is passed to partman is:
> > > /dev/ide/host0/bus0/target0/lun0/part
> > > 
> > > The code in partman that interprets this expects either:
> > > /dev/ide/host0/bus0/target0/lun0/disc
> > > /dev/ide/host0/bus0/target0/lun0/part1
> > > 
> > > But in this case, "part" without a partition number seems to be used for 
> > > the whole disk, which perfectly explains the display as a partition 
> > > without partition number.
> > > We could support this in the relevant partman function, but I'd like to 
> > > know if this use of "part" to indicate the whole disk is valid first. It 
> > > could also be a bug, either in the kernel or in udev's devfs 
> > > compatibility code.
> > > 
> > > So, my follow-up questions:
> > > - what is the output of 'cd /dev/ide/host0/bus0/target0/lun0/; ls -l'?
> > 
> > ~ # cd /dev/ide/host0/bus0/target0/lun0/; ls -l
> > brw-rw----    1 root     root       3,   0 Jan  4 19:07 part
> > brw-rw----    1 root     root       3,   1 Jan  4 19:07 part1
> 
> > > - what is the non-devfs representation of this disk (e.g. /dev/hda*) and
> > >   what devices are present for that?
> > 
> > ~ # ls -l /dev/hda*
> > lrwxrwxrwx    1 root     root           32 Jan  4 19:07 /dev/hda ->
> > ide/host0/bus0/target0/lun0/part
> > lrwxrwxrwx    1 root     root           33 Jan  4 19:07 /dev/hda1 ->
> > ide/host0/bus0/target0/lun0/part1
> 
> To narrow down the possible reasons for this behaviour I have done some
> further tests:
> 
> - Configuring the machine to little endian mode (mipsel)
>   does not change anything, so it is not endianess-related.
> 
> - On another mipsel machine (Cobalt NASRaQ) which uses the
>   integrated PCI IDE controller of its VIA vt82c586a southbridge,
>   everything looks normal:
> 
>   ~ # cd /dev/ide/host0/bus0/target0/lun0/; ls -l
>   brw-rw----    1 root     root       3,   0 Jan  1 00:00 disc
>   brw-rw----    1 root     root       3,   1 Jan  1 00:00 part1
> 
> Besides the different IDE controllers and their drivers there is
> one other difference between the two: the SWARM uses a 64bit
> kernel while the Cobalt uses a 32bit kernel. Userland is 32bit in
> both cases.
> 
> Regards,
> Karsten
> -- 
> #include <standard_disclaimer>
> Nach Paragraph 28 Abs. 3 Bundesdatenschutzgesetz widerspreche ich der Nutzung
> oder Uebermittlung meiner Daten fuer Werbezwecke oder fuer die Markt- oder
> Meinungsforschung.
> 
> 

-- 
Martin Michlmayr
http://www.cyrius.com/



Reply to: