[Cooker] Bug in cooker tulip driver.

2001-08-22 Thread Ram'on Garc'ia Fern'andez

Hello,

We are having problems with the tulip driver of cooker kernel rpms.

After installing any of the cooker RPMS (starting from
kernel-smp-2.4.4, and not fixed in 2.4.8) the network does not work
any longer. The kernel shows messages about trying to change the media
of the card, but it always tries the same media :-).

We could fix the problem by forcing the media by adding a line
options tulip options=20
to the file /etc/modules.conf

In this way the card worked somewhat, but it was not reliable. We
expienced NFS problems using a SGI IRIX machine as NFS client.

All this problems were fixed by compiling the tulip driver in kernel
2.4.3 against the current kernel.

BTW, is the bug tracking system actively monitored? This problem
was reported there, but no response was ever received.

Ramon




[Cooker] Bug in cooker kernel rpm: kernel hang completely

2001-08-22 Thread Ram'on Garc'ia Fern'andez

Hello,

This crash was produced when mounting the Zip drive.

(This causes modprobe to load the ide-scsi module)

RPM: kernel-smp-2.4.8-4mdk
2 Pentium III processors 800 MHz kernel

After that the kernel hang completely. No ping. NumLock keys do not
respond. For this reason, I had to write by hand all the kernel
messages. Please forgive mistakes.

Ramon

EIP: C0186B6A (account_io_start + 0x26)
*PDE = 0
Flags = 10046
EAX = 0
EBX = 393946
ECX = 2
EDX = 0
ESI = 2
EDI = 0
EBP = C7F33018
ESP = C78E3CC0
DS = ES = SS = 18
process modprobe
Stack: C75C2C60 C0186C02 00393946 C75C2C60 0 2 2 C40F0DA0
0  39346 0 C0187466 C75C2C60 0 2 C40F0DA0
Call trace:
C0186C02 req_new_io+0x36
C0187466 __make_request+0x606
C01877A1 generic_make_request+0xe1
C0187855 submit_bh+0x55
C0187A6B ll_rw_block+0x1fb
C0139287 bread+0x33
C015BA86 validate_partition_table+0x16
C015BBA1 ldm_partition+0x8d
C015A07C check_partition+0xe0
C0117CEE release_console_sem+0x8e
C015A4E3 grok_partitions+0x7b
C015A450 register_disk+0x28
C8929CE5 ?? (probably from ide-scsi module; unfortunately I have no record).
C892AB68 ??
C892AB20 ??
C88114B1 ??
Code: 01 4B 30 EB 0B 90 85 D2 74 03 FF 43 38 01 4B 40
(the symbol buffer is the start of this piece of code)
0:   01 4b 30add%ecx,0x30(%ebx)
3:   eb 0b   jmp10 buffer+0x10
5:   90  nop
6:   85 d2   test   %edx,%edx
8:   74 03   je d buffer+0xd
a:   ff 43 38incl   0x38(%ebx)
d:   01 4b 40add%ecx,0x40(%ebx)