>I tried new EISA config stuff to config AHA-2842 VL SCSI card which is
>almost software compatible with AHA-2742 EISA SCSI card, and encounterd
>two problems.
>
>  (1)	new EISA config stuff checks whether EISA ROM BIOS is exist or
>	not, and if not exist, then doesn't probe EISA bus subdevices.
>	so that AHA-2842 is never probed.
>
>  (2)	(at least) AHA-2842 requires additional priming to quiry
>	vendor and product ID. (I didn't try AHA-2742, though)
>
>so, which is suitable solution to write AHA-2842 probe & attach code ?
>
>  (a)	apply following patches, and share code with EISA config stuff.
>
>  (b)	duplicate EISA config stuff to AHA-2842	specific probe routine.

You may want to look at the various arguments for and against this approach
that came up in the FreeBSD-current mailing lists.  The 2842 probe is
non-invasive to real EISA devices, and since the EISA ID register space is
"reserved" I haven't found a non-EISA device (other than the 2842) that
lives in ID address space or is upset by probing the full EISA ID address
space regardless of the presence of an EISA mainboard.  The question mainly
comes down to one of "purity", but since the 2842 was designed to mimic an
EISA device (much like ISA cards like the 3c509 can mimic an EISA device) I
decided to make eisaconf in FreeBSD work with the 2842. YMMV.

>--
>soda@sra.co.jp		Software Research Associates, Inc., Japan
>(Noriyuki Soda)		   software tools and technology group

--
Justin T. Gibbs
===========================================
  FreeBSD: Turning PCs into workstations
===========================================