Subject: RE: More info on the Dual 2940UW problem
To: 'firstname.lastname@example.org' <email@example.com>
From: Scott Burns <Scott.Burns@Netcontech.Com>
Date: 11/19/1998 15:03:41
I a couple of other BSD boot disks and here are the results:
- Hangs once it gets to the card, prints out one line of detail on the
card then stops probing at that point.
- Sorry, I don't think there is a msgbuf log written to disk because of
- Seems to boot into some type of fancy installation system which I am not
used to ;-) Obviously either the install doesn't load up all the
controllers, or it is working ok. Can I get down to a shell and issue a
couple of mounts and see what happens ?
- The circuit boards seem to say Rev C on the back, on the end which is
furthest from the external connector so I am not sure if it is the Rev B.
problem Justin reported ?
- Tried to move around the cards so that we are not using IRQ 5
- Booting the diskette locks up as well with the two controllers so since
disk locks up I guess it is not the problem of not being able to find the
because NetBSD is assigning the sdxxx drives in a different order than I
- All PCI slots are reported in CMOS as being capable of mastering
- There is no PNP on this machine that I have seen, the isapnp support
does not load
in the floppy kernel
I will begin and try changing around slots if we think there is value here
and hard wiring IRQ's etc rather than using the auto assignment stuff if we
think that will help. Any value in calling Adaptec if I say I am using
From: Justin T. Gibbs[SMTP:firstname.lastname@example.org]
Sent: Wednesday, November 18, 1998 11:37 PM
Cc: 'Justin T. Gibbs'
Subject: Re: More info on the Dual 2940UW problem
>I will quickly try this boot disk to see if the problem is resolved with
>it. Should it not work I will report all details so you can hopefully
>improve your driver. Thanks for the interesting info on the Rev B bug. I
>will check my cards.
Based on the BIOS revision, they are at least rev B cards. I don't
believe that Adaptec has revved these parts since B and from what I've
seen, Rev B cards is all you can get from the channel.
>Unfortunately at this time, because of the large scale
>use of NetBSD in our office, it is not practical for us to switch to
>Free/OpenBSD at this time.
Then your only recourse may be to use a different controller, like a
BT-958. Not only is the 2940 a much better performer than the BT-958,
but the difference in I/O performance for CAM vs. Julian E. SCSI code is
very noticeable for typical server workloads. NetBSD should adopt CAM.
The divergence between my current driver and the one in NetBSD is so
large and my time so limited, that I can't really fix bugs in the NetBSD