2.4.12-ac5: IDE-SCSI kernel panic
on a CD with a read error, using IDE-SCSI, there has always been lots of problems (at least for me).
After tons of errors in the logs (medium error/unrecovered error), I tried to kill-9 the process reading the mounted CD (xine) without success (one thread stuck in lock_page, the other <defunct>). I then tried to reboot, but even before init started kill-TERM processes, I got this (hand-copied from screen):
scsi0: ERROR on channel 0, id 1, lun 0, CDB: Request Sense 00 00 00 40 00 Info fld=0x437ea, Current sd0b:00: sense key Medium Error Additional sense indicates Unrecovered read error
I/O error: dev 0b:00, sector 1105832
Incorrect segment count at 0xc01e4342nr_segments is 3f counted segments is 19
Flags 0 0
Segment 0xd92e86c0, blocks 4, addr 0x1f983fff
Segment 0xd92e8660, blocks 4, addr 0x1f9847ff
[I'm not copying them all, around 25 of them]
Kernel panic: Ththththaats all folks. Too dangerous to continue.
Xav
Quoted 4 lines
scsi0: ERROR on channel 0, id 1, lun 0, CDB: Request Sense 00 00 00 40 00 Info fld=0x437ea, Current sd0b:00: sense key Medium Error Additional sense indicates Unrecovered read error
I/O error: dev 0b:00, sector 1105832
This stuff is ok.
Quoted 7 lines
Incorrect segment count at 0xc01e4342nr_segments is 3f counted segments is 19
Flags 0 0
Segment0xd92e86c0, blocks 4, addr0x1f983fffSegment0xd92e8660, blocks 4, addr0x1f9847ff[I'm not copying them all, around 25 of them]Kernel panic: Ththththaats all folks. Too dangerous to continue.
Someone did a partial I/O and then it seems failed to properly update the descriptor when retrying half of it. Ugly. This looks like a CD trace so I assume this was sr.c that barfed. It could also be a generic scsilib bug