i had the same problem with version 1.17 of the ncr53c8xx driver
on an 825 board.
version 1.18 will fix this specific problem and is available
at linux.wauug.org/pub/roudier.
it is supposedly available in all kernels from 2.1.33 forward
but i've not had much luck in configuring it in 2.1.35.
it is an astonishing performance improvement for the 825.
i suspect this is due to 16 bit transfer and configurable
tag queueing.
bill
On 23 Apr 1997, Thomas Ackermann wrote:
> I constantly keep getting such SCSI errors:
>
> Apr 19 22:50:31 final kernel: scsi : aborting command due to timeout : pid 4537,
> scsi0, channel 0, id 0, lun 0 Write (10) 00 00 25 33 cf 00 00 02 00
> Apr 19 22:50:31 final kernel: scsi0 : found command 4537 in Linux issue queue
>
> This is true due to all kernel up to 2.0.29 - and i suspect it in future
> kernels too ...
>
> I use the ncr53c8xx driver, not the BSD version - and would like to hear if
> anybody else is getting such messages.
>
> I use an ASUS SP3G 486 PCI Board (Rev. 1.8, Saturn II) with the NCR810
> controller.
>
> As this configuration is reported to be very stable (a grown up motherboard
> and such), i suspect problems with the NCR driver ...
>
> I too constantly get error at boot-time - talking about "deleted inode xyz has
> zero dtime. Set dtime? Yes" ...
>
> Anybody else had this experience?
>
> (Further information: 32MB RAM, 2GM IBM DORS, 1 GB IBM DPES)
>
> Byebye,
> -
> Thomas Ackermann | <tgm@math.uni-bonn.de> | Tel. 0228/73-7773|631369
> LINUX, not Micro$oft! | no unnecessary MIME extended mails!
> no unnecessary Graphics in HTML-Document! | GURPS, not T$R!
> put your HTML stuff in a ftp-accessible directory!
>