Re: 2.1.96 - aic7xxx driver woes (bios)

Doug Ledford (dledford@dialnet.net)
Sat, 18 Apr 1998 03:09:59 -0500

Laszlo Vecsey wrote:
>
> Hi,
>
> With 2.1.96 (and I noticed this about 5 or 10 kernels back as well if I'm
> not mistaken) my AIC-7880 v1.26sP7 (onboard controller, intel r440lx
> motherboard) with the bios enabled, the linux aic7xxx driver reports the
> bios as disabled and I get the wrong number of drives identified. In fact,
> 15 devices show up when really I have 10 drives on the channel.
>
> I've backed down to 2.1.95 in the meantime, the 10 drives are identified
> and performance is pretty speedy, with the bios enabled. If I disable the
> bios I get similar behavior.. do I need to flash my controller to a more
> recent revision perhaps, with 2.1.96 and on?

Finding 15 devices when only 10 exist is almost always a problem with SCSI
IDs. In this case, I would have to think that the BIOS is doing SCSI Plug &
Play support (SCAM) and that the new driver in 2.1.96 is triggering the
devices into forgetting the SCAM settings set during the BIOS initialization
(possibly as part of the chip reset or bus reset). In any case, I would
think that configuring all of your SCSI devices with hard SCSI IDs (aka, use
the jumper to set the SCSI ID) and disabling SCAM support in the Adaptec
BIOS should help. However, as to the BIOS disabled/enabled message the
driver prints out, does the driver report that it's able to read the SEEPROM
if you bott with the command aic7xxx=verbose:0xffff? If the driver can read
the SEEPROM and still reports the BIOS as disabled, then that's a bug. If
the driver reports that it can't read the SEEPROM, then it will always say
the BIOS is disabled regardless of the actual setting (because we can't read
the bit that tells us whether it is or not).

-- 


 Doug Ledford  <dledford@dialnet.net>
  Opinions expressed are my own, but
     they should be everybody's.


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu