lkml.org 
[lkml]   [1999]   [Feb]   [25]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectIDE driver bug
Hello,

I was told that I may have discovered a bug in the IDE driver. It was
suggested to me that you would rather have the logs than a long
description, so I will try to be as succinct as possible.

I have been trying to get a CastleWood
(http://www.castlewoodsystems.com) ORB drive to work with Linux. The
kernel is 2.0.36 (something; whatever comes with RedHat 5.2). The ORB
is a 2.2Gb removable media drive with hard drive-like specs. Transfers
are at 12Mb sustained, and access times are 10ms. The disks are low
cost ($30) and are the same size as Zip media. The drives are low cost
($179) and small as well, fitting into a half-height floppy bay.

The drive is recognized by the kernel, and I can partition and make a
filesystem (ext2) on the disk. I can copy to and from the disk, run
fsck -f, unmount, eject, reinsert, and remount the disk as many times as
I want. The only fault at this point are an occasional "Busy" timeout:

Feb 24 22:10:24 ender kernel: hdc: status timeout: status=0xd0 { Busy }
Feb 24 22:10:24 ender kernel: hdc: no DRQ after issuing WRITE
Feb 24 22:10:25 ender kernel: ide1: reset: success

This happens at least once during mke2fs, and sometimes during a write.
I have dumped 800Mb to the disk, and found a single 50K file that was
corrupt.

The real problems start when I reboot the machine. Once the machine is
up, fdisk recognizes the correct drive geometry and sees the correct
partition. However, mounting the drive after a reboot causes a massive
dump of errors of the type:

Feb 24 22:15:19 ender kernel: EXT2-fs error (device 16:01):
read_block_bitmap: Cannot read block bitmap - block_group = 224,
block_bitmap = 1835019
Feb 24 22:15:19 ender kernel: attempt to access beyond end of device
Feb 24 22:15:19 ender kernel: 16:01: rw=0, want=1843212, limit=1056352

and fsck tells me:

bash# fsck /dev/hdc1
Parallelizing fsck version 1.12 (9-Jul-98)
e2fsck 1.12, 9-Jul-98 for EXT2 FS 0.5b, 95/08/09
The filesystem size (according to the superblock) is 2153560 blocks
The physical size of the device is 1056352 blocks
Either the superblock or the partition table is likely to be corrupt!
Abort<y>?

If I don't abort, I have an (unending) stream of inode problems, which
doesn't suprise me. By now, the drive is near useless... unless I
delete the partition, make a new one, and make a new filesystem. Then
the drive works like a charm... until I reboot.

I've attached a segment of my syslog, starting at the first glitch,
encountered while making a filesystem on a new partition. I've also
included a dump of 'hdparm -i /dev/hdc'.

My machine is entirely new:

FIC 2013 motherboard @ 100Mhz, 1Mb cache
AMD K2-300 CPU
128Mb 8ns 100Mhz DRAMM
BusLogic PCI SCSI controller
Matrox G200 8Mb AGP
Soundblaster 16 ISA PnP
5Gb EIDE HD (Western Digital?)
External Zip on the SCSI card
Creative Labs CDROM (24x or some such)
Random ISA modem (56K v/d/f)
Printer on da printer port, Mouse on da mouse port, keyboard
on da keyboard port (motherboard all)
Runnng RedHat Linux 5.2, kernel out-of-the-box

Please let me know if there is any other information I can get to you,
and what your opinion(s) are.

Thank you *very* much,

Sean Russell
(also at ser@efn.org)

--
|.. --------------------- Sean Russell ----------------------
<|> ser@n-link.net ::: http://javalab.uoregon.edu/ser
/|\ ------- [ Software Developer ] --------
/| [ n-Link ]



/dev/hdc:

Model=CASTLEWOOD ORB2-E, FwRev=USR-3.4, SerialNo=19A26A18
Config={ Fixed }
RawCHS=4273/16/63, TrkSize=0, SectSize=0, ECCbytes=38
BuffType=0(?), BuffSize=0kB, MaxMultSect=16, MultSect=off
DblWordIO=no, maxPIO=240(?), DMA=yes, maxDMA=0(slow)
CurCHS=4273/16/63, CurSects=4307184, LBA=yes, LBAsects=4307184
tDMA={min:120,rec:120}, DMA modes: *mword2
IORDY=yes, tPIO={min:120,w/IORDY:120}, PIO modes: mode3 mode4

[unhandled content-type:application/x-gzip]begin:vcard
n:Russell;Sean
tel;fax:(541)617-0997
tel;home:(541)617-0997
tel;work:(541)383-4028
x-mozilla-html:TRUE
org:S&K Electronics;
adr:;;P.O. Box 6044;Bend;OR;97708-6044;USA
version:2.1
email;internet:ser@efn.org
title:Sr. Software Developer
note:http://javalab.uoregon.edu/ser
x-mozilla-cpt:;-21984
fn:Sean Russell
end:vcard
\
 
 \ /
  Last update: 2005-03-22 13:50    [W:0.034 / U:0.724 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site