I posted some info about this a while ago ... now I have more trouble.
Could it be that sd.c, once it detects a bad block and (apparently)
has the drive reassigning it, doesn't reset the "bad" state and reports
EVERY block following this as BAD?
...
I have a
Vendor: FUJITSU Model: M2513A Rev: 1700
Type: Direct-Access ANSI SCSI revision: 02
on my
Chip NCR53C875, device id 0xf, revision id 0x3
IO port address 0xe800, IRQ number 12
Using memory mapped IO at virtual address 0xc4800000
Synchronous period factor 12, max commands per lun 32
Debug flags 0x0, verbosity level 2
(Tekram DC-390F, actually). Write cache enabled, Error correction enabled.
Now, when I do an
mke2fs -b 2048 -m0 /dev/sdb
(yes, whole device - I don't partition removables...)
I get a properly setup partition, no problem. Then I do a cp -av ~ /mod.
After a while, this starts:
Jan 16 20:33:34 debian kernel: scsi : aborting command due to timeout : pid
3283259, scsi0, channel 0, id 2, lun 0 Read (6) 02 80 03 01 00
Jan 16 20:33:34 debian kernel: ncr53c8xx_abort: pid=3283259
serial_number=3283318 serial_number_at_timeout=3283318
Jan 16 20:33:38 debian kernel: SCSI host 0 abort (pid 3283259) timed out -
resetting
Jan 16 20:33:38 debian kernel: SCSI bus is being reset for host 0 channel 0.
Jan 16 20:33:38 debian kernel: ncr53c8xx_reset: pid=3283259 reset_flags=2
serial_number=3283318 serial_number_at_timeout=3283318
Jan 16 20:33:38 debian kernel: ncr53c875-0: resetting, command processing
suspended for 2 seconds
Jan 16 20:33:38 debian kernel: ncr53c875-0: restart (scsi reset).
Jan 16 20:33:38 debian kernel: ncr53c875-0: enabling clock multiplier
Jan 16 20:33:38 debian kernel: ncr53c875-0: Downloading SCSI SCRIPTS.
Jan 16 20:33:38 debian kernel: ncr53c875-0: command processing resumed
Jan 16 20:33:38 debian kernel: ncr53c875-0-<2,*>: FAST-10 SCSI 10.0 MB/s
(100 ns, offset 10)
Jan 16 20:33:38 debian kernel: scsi0: MEDIUM ERROR on channel 0, id 2, lun
0, CDB: Read (6) 02 80 03 01 00
Jan 16 20:33:38 debian kernel: Deferred error sd08:10: sense key Medium
Error
Jan 16 20:33:38 debian kernel: Additional sense indicates Write error - auto
reallocation failed
Jan 16 20:33:38 debian kernel: scsidisk I/O error: dev 08:10, sector 655372
:
:
:
This keeps happening, say 20 times, until "cp" complains about "I/O error"
and commits suicide.
OK, a read error. Should be reallocated, shouldn't it? scsi-config tells me
it is. scsiinfo as well. Well, scsi-config also tells me there are about
1500 entries in the "grown defects" list, which I CANNOT clear (neither via
low level format by BIOS nor via linux utilities)
OK, so the medium is bad in some places. Lets do a mke2fs -c.
Bullshit:
Jan 16 21:39:59 debian kernel: sdb: Write Protect is off
Jan 16 21:39:59 debian kernel: sdb:sd.c:Bad block number requestedscsidisk
I/O error: dev 08:10, sector 0
Jan 16 21:39:59 debian kernel: Dev 08:10: unable to read RDB block 0
Jan 16 21:39:59 debian kernel: sd.c:Bad block number requestedscsidisk I/O
error: dev 08:10, sector 0
Jan 16 21:39:59 debian kernel: 08:10: error reading partition table
Jan 16 21:48:42 debian kernel: sd.c:Bad block number requestedscsidisk I/O
error: dev 08:10, sector 5626
Jan 16 21:48:42 debian kernel: sd.c:Bad block number requestedscsidisk I/O
error: dev 08:10, sector 5628
Jan 16 21:48:42 debian kernel: sd.c:Bad block number requestedscsidisk I/O
error: dev 08:10, sector 5630
Jan 16 21:48:42 debian kernel: sd.c:Bad block number requestedscsidisk I/O
error: dev 08:10, sector 5632
Jan 16 21:48:42 debian kernel: sd.c:Bad block number requestedscsidisk I/O
error: dev 08:10, sector 5634
Jan 16 21:48:42 debian kernel: sd.c:Bad block number requestedscsidisk I/O
error: dev 08:10, sector 5636
Jan 16 21:48:42 debian kernel: sd.c:Bad block number requestedscsidisk I/O
error: dev 08:10, sector 5638
Jan 16 21:48:42 debian kernel: sd.c:Bad block number requestedscsidisk I/O
error: dev 08:10, sector 5640
Jan 16 21:48:42 debian kernel: sd.c:Bad block number requestedscsidisk I/O
error: dev 08:10, sector 5642
Jan 16 21:48:42 debian kernel: sd.c:Bad block number requestedscsidisk I/O
error: dev 08:10, sector 5644
Jan 16 21:48:42 debian kernel: sd.c:Bad block number requestedscsidisk I/O
error
:
:
and so on, the whole time the disk is being "badblock'ed". Takes a little
longer, because every message is logged, of course.
The point is, I have successfully VFAT formatted this medium and have had
_absolutely_ no trouble, only was it painfully slow (<300-400k/s reading). I
have installed Windows games on it - no problem.
Any ideas, hints, patches? Or is 2.1.131 already too antique in respect to
ncr8c53xx SCSI code? :)
--
_ciao, Jens_______________________________ http://www.pinguin.conetix.de
cat /dev/boiler/water | tea | sieve > /cup
mount -t hdev /dev/human/mouth01 /mouth ; cat /cup >/mouth/gulp
-
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/