[tlug] dma timeouts

[Botond Botyanszki <9915104t@example.com>]

Hello Tlug,

I get dma timeouts on 3 out of my 3 machines.

This is my desktop machine (2.4.21-pre1 kernel) hdd is an asus dvd
reader.
The dvd then turns itself off (led goes out), and nothing but a reboot
can help. After the reboot it works perfectly with the same cd:
Jan 26 13:31:57 mephisto kernel: hdd: timeout waiting for DMA
Jan 26 13:31:57 mephisto kernel: hdd: timeout waiting for DMA
Jan 26 13:31:57 mephisto kernel: hdd: (__ide_dma_test_irq) called while
not waiting
Jan 26 13:31:57 mephisto kernel: hdd: status timeout: status=0xd0 { Busy
}
Jan 26 13:31:57 mephisto kernel: hdd: status timeout: error=0x00
Jan 26 13:31:57 mephisto kernel: hdd: drive not ready for command
Jan 26 13:31:57 mephisto kernel: hdd: ATAPI reset complete
Jan 26 13:32:12 mephisto kernel: hdd: timeout waiting for DMA


Next one, this is a dell inspiron 5000 laptop, also with kernel
2.4.21-pre1. The harddisk gets these about 1-2 times a week, then starts
emitting a strange noise as if the fan is hitting a cable (not clicking
loudly though).  After the reboot, everything is fine  (no lost data so
far) until a few days after it starts again. Sometimes the machine locks
up, sometimes it gives just errors after failing to read from the drive.
I can't even remotely reboot through ssh right now:
[root@example.com /]# reboot
bash: /sbin/reboot: Input/output error
[root@example.com /]# halt
bash: /sbin/halt: Input/output error

logs:
Feb 11 20:03:57 mick kernel: hda: dma_timer_expiry: dma status == 0x20
Feb 11 20:03:57 mick kernel: hda: timeout waiting for DMA
Feb 11 20:03:57 mick kernel: hda: timeout waiting for DMA
Feb 11 20:03:57 mick kernel: hda: (__ide_dma_test_irq) called while not
waiting
Feb 11 20:03:57 mick kernel: hda: status timeout: status=0xd0 { Busy }
Feb 11 20:03:57 mick kernel:
Feb 11 20:03:57 mick kernel: hda: drive not ready for command
Feb 11 20:04:32 mick kernel: ide0: reset timed-out, status=0x80
Feb 11 20:04:32 mick kernel: hda: status timeout: status=0x80 { Busy }
Feb 11 20:04:32 mick kernel:
Feb 11 20:04:32 mick kernel: hda: drive not ready for command
Feb 11 20:05:02 mick kernel: ide0: reset timed-out, status=0x80
Feb 11 20:05:02 mick kernel: end_request: I/O error, dev 03:07 (hda),
sector


Third one is also a dell, but an older inspiron (4100).
May 22 20:59:28 purple kernel: hda: timeout waiting for DMA
May 22 20:59:28 purple kernel: ide_dmaproc: chipset supported
ide_dma_timeout func only: 14
May 22 20:59:28 purple kernel: hda: status timeout: status=0xd0 { Busy }

May 22 20:59:28 purple kernel: hda: drive not ready for command
May 22 20:59:28 purple kernel: ide0: reset: success
These are very rare though and after a few seconds of blocking i/o, it
works just as before.
Rather then dma timeouts, this  harddisk gets read errors(?) a bit more
frequently, but I didn't notice any data corruption or bad sectors:
Nov 30 20:22:20 purple kernel: end_request: I/O error, dev 03:04 (hda),
sector 1315128
Nov 30 20:22:24 purple kernel: hda: dma_intr: status=0x51 { DriveReady
SeekComplete Error }
Nov 30 20:22:24 purple kernel: hda: dma_intr: error=0x40 {
UncorrectableError }, LBAsect=10183247, sector=131513

Now, my question is: am I having just bad luck (3 out of 3 machines)
with bad hardware or is this some nasty kernel bug or both? Are there
any diagnostic tools that I should run maybe? Any help would be
appreciated.