-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
(I originally posted this to linux-tape, but as I realized this might
be more of an IDE problem - so here goes...)
I've spent the past few days trying to get my new OnStream DI-30 IDE
tape drive working, with little success. I thought I might have a
hardware problem, but the drive works fine under windows (backup and
compare of a 2G partition works correctly consistently). For good
measure, I even replaced the IDE cable with the one supplied with the
drive - that made no difference.
Although I've seen other problems, the most consistent one seems to be
a problem with verifying a tape. I can do a tar of a 1.2G partition,
and that will work fine. However, a compare of that same partition
aborts relatively quickly, with messages like this:
May 22 00:09:20 lydia kernel: ide-tape: ht0: skipping frame,
logical_blk_num 3015 (expected 8)
May 22 00:09:20 lydia kernel: ide-tape: ht0: couldn't find logical
block 8, aborting (block 3015 found)
May 22 00:09:20 lydia kernel: ide-tape: ht0: skipping frame,
logical_blk_num 3015 (expected 9)
May 22 00:09:20 lydia kernel: ide-tape: ht0: couldn't find logical
block 9, aborting (block 3015 found)
The weird part is the references to block 3015 are very, very common -
even with different tapes. I've gotten this error with 2, possibly 3
tapes, with different kernels, over 3 days of testing.
One other big glitch - my X server (Mach64 3.3.6) has a tendency to
lock up completely when the tape is being accessed, if you do things
like switch virtual screens around. The lock is total - keyboard and
mouse death, uninterruptible sleep. The machine is still up, though
(remote logins work). Even under heavy load, my X server has never
acted this badly before.
I'm using the patches from the OnStream web site: ide.2.2.14.20000124
and ide_2_2_15_20000504. I just did a diff between
ide_2_2_15_20000504 and the latest stable IDE patch,
ide.2.2.15.20000509, and they seem virtually identical, so I expect my
problems to be the same (I can check to make sure if anyone suspects
otherwise).
I've played with patched versions of both 2.2.14 and 2.2.15 (both of
which are stable for me otherwise). I've tried building the kernel
using gcc 2.95.2 and 2.7.2.3. I've built it both with and without
chipset-specific support (PIIX).
I've also seen messages about lost interrupts and DMA problems;
however, my hard drive has worked consistently with Ultra DMA access.
Also, disabling DMA on hdd using hdparm doesn't fix things.
My system is an HP Pavilion 8260, which has an Intel 440LX chipset.
My IDE configuration is as follows:
primary master: Maxtor 91020D6 10G
primary slave: none
secondary master: Toshiba CD-ROM XM-6402B
secondary slave: OnStream DI-30
The DI-30 firmware is reported as version 1.06. My system is running
Debian 2.2 (potato). My current kernel (ide-patched 2.2.14) uses the
following options in the block section:
CONFIG_BLK_DEV_FD=y
CONFIG_BLK_DEV_IDE=y
CONFIG_BLK_DEV_IDEDISK=y
CONFIG_BLK_DEV_IDECD=y
CONFIG_BLK_DEV_IDETAPE=m
CONFIG_BLK_DEV_IDEPCI=y
CONFIG_BLK_DEV_IDEDMA=y
CONFIG_IDEDMA_AUTO=y
IDEDMA_PCI_EXPERIMENTAL=y
CONFIG_BLK_DEV_PIIX=y
CONFIG_BLK_DEV_PIIX_TUNING=y
CONFIG_BLK_DEV_LOOP=m
CONFIG_BLK_DEV_RAM=m
CONFIG_PARIDE_PARPORT=m
CONFIG_BLK_DEV_IDE_MODES=y
I can think of more tests to try, but right now I'd just be shooting
in the dark. Anyone have a clue about what's going on?
Thanks!
--Anil
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.1 (GNU/Linux)
iEYEARECAAYFAjkpokQACgkQXOpXEmNZ3SevUACghtjlWJh0rtUODT0qoLb3piWn
TtoAnApA2eLVgTgZw+n5lkce7gzKO+Tg
=ud3X
-----END PGP SIGNATURE-----
-
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/
This archive was generated by hypermail 2b29 : Mon May 22 2000 - 17:31:13 EDT