problems with Seagate 120 GB drives when mutlwrite = 16
From: kris
Date: Sun Oct 19 2003 - 10:57:08 EST
Hi,
I have noticed some problems with recent large drives, connected to a
variety of controllers.
I tested with nforce ide controller, a CMD649 based controller and an
Intel 870 cghipset. all have same or similar symptoms.
Linux 2.4.22 kernel : (Linux borg-cube 2.4.22-xfs #2 SMP Tue Oct 7
20:53:04 CEST 2003 i686 unknown)
Oct 6 15:52:12 borg-cube kernel: hdg: dma_timer_expiry: dma status ==
0x21
Oct 6 15:52:22 borg-cube kernel: hdg: timeout waiting for DMA
Oct 6 15:52:22 borg-cube kernel: hdg: timeout waiting for DMA
Oct 6 15:52:22 borg-cube kernel: hdg: status error: status=0x58 {
DriveReady SeekComplete DataRequest }
Oct 6 15:52:22 borg-cube kernel:
Oct 6 15:52:22 borg-cube kernel: hdg: drive not ready for command
Oct 6 15:52:22 borg-cube kernel: hdg: status timeout: status=0xd0 {
Busy }
Oct 6 15:52:22 borg-cube kernel:
Oct 6 15:52:22 borg-cube kernel: hdg: no DRQ after issuing WRITE
Oct 6 15:52:22 borg-cube kernel: ide3: reset: success
same in 2.4.20 ( kernel from Suse)
2.6.0-test6 :
Oct 9 09:43:09 borg-cube kernel: hdg: dma_timer_expiry: dma status ==
0x21
Oct 9 09:43:18 borg-cube kernel:
Oct 9 09:43:19 borg-cube kernel: hdg: DMA timeout error
Oct 9 09:43:19 borg-cube kernel: hdg: dma timeout error: status=0x58 {
DriveReady SeekComplete DataRequest }
Oct 9 09:43:19 borg-cube kernel:
Oct 9 09:43:19 borg-cube kernel: hdg: status timeout: status=0xd0 {
Busy }
Oct 9 09:43:19 borg-cube kernel:
Oct 9 09:43:19 borg-cube kernel: hdg: no DRQ after issuing MULTWRITE
Oct 9 09:43:20 borg-cube kernel: ide3: reset: success
same in 2.6.0-test8
so behaviour is consistent.
I can avoid this with either turning off dma access or disabling the
multwrite ( hdparm -m0 /dev/hdg)
is this a known bug, or should I file one ?
kind regards, Kris
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/