[lkml]   [1996]   [Sep]   [30]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
Subjectproblems with IDE driver
I'm seeing a very unexpected kind of behavior from the IDE driver.  When
I enable 32-bit I/O (using hdparm) I see a massive degredation of the
drive throughput (as reported by hdparm -t). It goes from about 3 MB/sec
to about .3 MB/sec. Also I strt seeing these show up in my syslog:

hdb: irq timeout: status=0x58 { DriveReady SeekComplete DataRequest }
ide0: reset: success

I think I remember some threads a while back where these were attributed
to a drive about to fail, but both drives (hda and hdb) produce them
only after I enable 32-bit I/O.

This behavior occurs no matter what other changes I make in the driver
parameters (unmasking IRQs, changing the readahead, changing the PIO

I wrote Wolfram Podien (the maintainer of the umc8672 driver) and he
replied that the problem was not with his driver, as the card specific
driver code is never called after startup.

I've been seeing this behavior since very late in the 1.3.xx series, and
it has continued up through 2.0.21. I have not seen this behavior
using WinNT (3.51 and 4.0) nor with Win95. I have the 32-bit drivers
for the card for each of these. I suppose they might eat the error
messages, but there is no speed degredation with them either.

I also did not see this behavior with a dtc2278 card in the same system.

If I have managed to miss some FAQ about this, then I'm sorry for wasting
time on this list. If it is a known bug with 32-bit I/O and the umc8672
card, and it is not in a FAQ of the ide.txt file, then maybe it should be


System info:
32 mb
VLB umc8672
WDC AC31000H (hda)
WDC AC2420H (hdb)
various and sundry kernel setups

 \ /
  Last update: 2005-03-22 13:38    [W:0.036 / U:1.380 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site