Skip to contentSkip to conversation

linux-2.2.2 kernel hangs

DISCUSSION Alois Fertl 2 participants 2 messages Mar 1, 1999

lore
Share & tools
lore
#0 lore raw

On Saturday I managed to update my kernel sources to 2.2.2 from vger/linux
at cvs.samba.org. There was some restructuring on samba and I had to find
out how to update my Repository and Root CVS files.

This 2.2.2 version has severe problems on a Motorola RiscPC. Actually the
kernel hangs within a short time if cpio'ing files from a NFS tree to the
local scsi drive. It starts with a message about the nfs server not responding and than endlessly repeats scsi timeouts. Due to my problems with CVS I do not have any of the versions between 2.1.130 and 2.2.2. The
2.1.130 version did not show the problem.

Is this a known problem.
Did anybody else experiance the same problem.

Here is a sample output:

nfs: server dino not responding, still trying
scsi : aborting command due to timeout : pid 3940, scsi0, channel 0, id 0, lun 0 0x0a 02 ff c0 80 00
ncr53c8xx_abort: pid=3940 serial_number=3947 serial_number_at_timeout=3947
scsi : aborting command due to timeout : pid 3941, scsi0, channel 0, id 0, lun 0 0x0a 03 00 5a 02 00
ncr53c8xx_abort: pid=3941 serial_number=3948 serial_number_at_timeout=3948
ncr53c810-0: abort ccb=c0195020 (cancel)
SCSI host 0 abort (pid 3941) timed out - resetting
SCSI bus is being reset for host 0 channel 0.
ncr53c8xx_reset: pid=3941 reset_flags=2 serial_number=3948 serial_number_at_timeout=3948
SCSI host 0 abort (pid 3941) timed out - resetting
SCSI bus is being reset for host 0 channel 0.
ncr53c8xx_reset: pid=3941 reset_flags=2 serial_number=3949 serial_number_at_timeout=3949
SCSI host 0 abort (pid 3941) timed out - resetting
SCSI bus is being reset for host 0 channel 0.
ncr53c8xx_reset: pid=3941 reset_flags=2 serial_number=3951 serial_number_at_timeout=3951
SCSI host 0 abort (pid 3941) timed out - resetting
SCSI bus is being reset for host 0 channel 0.
ncr53c8xx_reset: pid=3941 reset_flags=2 serial_number=3953 serial_number_at_timeout=3953

[[ This message was sent via the linuxppc-dev mailing list. Replies are ]] [[ not forced back to the list, so be sure to Cc linuxppc-dev if your ]] [[ reply is of general interest. To unsubscribe from linuxppc-dev, send ]] [[ the message 'unsubscribe' to [email protected] ]]

#1 lore raw
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

"Alois" == Alois Fertl <[email protected]> writes:

Alois> This 2.2.2 version has severe problems on a Motorola RiscPC. Actually Alois> the
Alois> kernel hangs within a short time if cpio'ing files from a NFS tree to Alois> the
Alois> local scsi drive. It starts with a message about the nfs server not Alois> responding and than endlessly repeats scsi timeouts. Due to my problems Alois> with CVS I do not have any of the versions between 2.1.130 and 2.2.2. Alois> The
Alois> 2.1.130 version did not show the problem.

Alois> Is this a known problem.

Yes. But it is not related specifically to PPC. I had the same problem with adaptec 2940 on Intel (and was not alone). The problem is that apparently, the scsi driver tries to transfer byte at a rate that cannot be handled as UW2 chips are now working... (20 Mb/s instead of 10Mb...).

To fix it on PC, I had to upgrade the adaptec BIOS and use the new per-device speed setup feature... I do not know how you will handled this is a ppc... May be there is a compilation option to set the default speed but I haven't checked...

BTW : I said that your patch for IDE will not work on MCP750.
This was wrong. It works and otherwyse you cannot boot using pload. Sorry for the mistake. The bad news, is that you must reformat your disk if you were using IDE before using the patch...

Signature · 13 lines
   __                 
  /  `                   	Eric Valette
 /--   __  o _.          	Canon CRF
(___, / (_(_(__         	Rue de la touche lambert
				35517 Cesson-Sevigne  Cedex
				FRANCE
Tel: +33 (0)2 99 87 68 91	Fax: +33 (0)2 99 84 11 30
E-mail: [email protected]

[[ This message was sent via the linuxppc-dev mailing list. Replies are ]]
[[ not forced back to the list, so be sure to  Cc linuxppc-dev  if your ]]
[[ reply is of general interest. To unsubscribe from linuxppc-dev, send ]]
[[ the message 'unsubscribe' to [email protected] ]]