Oops in 2.1.127 w/ide-scsi

David B. Rees (crichton@gimp.org)
Thu, 12 Nov 1998 14:01:42 -0500 (EST)


Hello,

I've been trying to use the ide-scsi stuff for my ATAPI cdr.
However, I've found a nice bug in the kernel every time I try
to use the ide-scsi emulation.

ide-scsi is compiled into the kernel (2.1.127), and at boot, right
after detecting all my scsi stuff, I get the following calltrace
from an oops: (sorry, no offsets, this is before init starts up)

scsi_request_sense
scsi_old_done
aic7xxx_done
idescsi_end_request
atapi_output_bytes
...
all the way back to symbol L6.

I've alos managed to disassemble the code portion of the oops:
8b 57 34 81 c2 c8 00 00 00 0c cd 1f c0 39 d0 73 51 a1 0c

0x804900c <str>: movl 0x34(%edi),%edx
0x804900f <str+3>: addl $0xc8,%edx
0x8049015 <str+9>: movl 0xc01fcd0c,%eax
0x804901a <str+14>: cmpl %edx,%eax
0x804901c <str+16>: jae 0x804906f <__FRAME_BEGIN__+75>
0x804901e <str+18>: movl 0xc,%eax
0x8049023 <str+23>: addb %dl,(%eax)

>From some printk'ing on my own, I found that this function is
in scsi_obsolete. Some things I found odd:

1) scsi_obsolete and scsi_error both have scsi_request_sense defined.
There are two scsi_request_sense symbols in my System.map. Is scsi_error
not done yet?

2) the code from scsi_obsolete doesn't match up with the above code
at all :( I'm somewhat new to this, so I could be getting it wrong.

My system has an Adaptec 2940UW controller and CD-RW CRX100E, ATAPI
CDROM on the second IDE channel. If you need any more info I'd be
glad to help.

Since I'm not on this list, could you please send responses to
mcrichto@purdue.edu.

Thanks,
Mark Crichton

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/