2.2.X+scsi+ide=timeout mystery

[Doug Ledford]

Steve Cahill wrote:
> 
> Well Folks,
> 
> This one is a real mystery.....
> 
> I cannot believe that my setup is unique, but hey... stranger things can
> happen I suppose. I have an Adaptec AHA-2940U2W with several devices ( see
> 2.0.38 dmesg out below), and an IDE primary drive where my boot record lives
> with a (bootable) DOS partition as well as some other Linux partitions:
> 
> hda: hda1 hda2 hda3 hda4 < hda5 hda6 hda7 hda8 >
> hda: WDC AC26400B, 6149MB w/512kB Cache, CHS=784/255/63
> 
> My main (mother) board is an FIC PA-2013 with a 350MHz AMD K6-2
> CPU, Millenium G200 video adapter, 128MB (PC100) SDRAM, sound card and modem
> reside in the 2 ISA slots and a 3COM 3c905B Cyclone ethernet card in one of
> the PCI slots.
> (aka very typical setup??)
> 
> When booting any Linux 2.2.X kernel with this IDE drive present I get a
> timeout message (which I have posted before, but will include here for
> completeness):
> 
> scsi : aborting command due to timeout : pid 60, scsi0, channel 0, id 1, lun
> 0 0x28 00 00 70 c2 ce 00 00 02 00
> 
> The timeout can last anywhere from 25 to 40 seconds.  Another symptom is
> LILO will hang (i.e. executing lilo) or take a very long time to execute.
> 
> When I boot any of the 2.2.X kernels without this drive present the timeout
> disappears. In fact when I boot with this drive present in any of the newer
> 2.0.3X kernels which support this adapter there is NO timeout and lilo
> behaves normally!
> 
> So why don't I just shutup and just use the 2.0.38 kernel? Well I do, but
> I would like to have some of the newer 2.2.X kernel features not present in
> the 2.0.3X kernels.
> 
> If anyone has a similiar hardware setup (at the very least has the 2940U2W)
> that mixes both scsi and ide in the 2.2.X kernels and get no timeouts or
> lilo troubles, please let me know. I have discussed this at length with Doug
> Ledford and I am sure he would be interested in knowing also. It is possible
> that there is something in my BIOS (and/or BIOS settings) that the 2.0.3X
> kernels (as well as WINDOZE :) overlooks or doesn't care about.

We had a machine in here in the test lab with the same problem.  After
upgrading the kernel to a 2.2.14 or later kernel, the problem myseriously went
away.  We attribute it to IDE changes in the 2.2.14 kernel patch.  But, I'm
not an IDE expert....

-- 

 Doug Ledford <dledford at redhat.com>  http://people.redhat.com/dledford
      Please check my web site for aic7xxx updates/answers before
                      e-mailing me about problems


To Unsubscribe: send mail to majordomo at FreeBSD.org
with "unsubscribe aic7xxx" in the body of the message