3. I had to add a cludge to the cmd out phase where the
sequence of commands is delayed between bytes.
You may or may not need this, this appeared to
be bus speed dependent. I basically looped for
2000 clock ticks before outputting sequential
command bytes.
--- drew's comment ---
I changed the command out phase so that REQ is checked before a byte
is sent, so this sbould be unecessary
---
4. I compiled with -DOVERRIDE=0xc8000 since I do not
(nor do I want) have a Future Domain BIOS chip.
5. In scsi.h change the reference to a host target id of 7 to 6.
Its ORed into the lun in a define I think.
The only problem I have encountered is that on first power-up
the READ CAPACITY (and for tapes READ BLOCK LIMITS) fail. For
--- drew again ---
This has been kludged arround - the error handling code in version
1 of the SCSI drivers doesn't work very well, and won't be fixed
because version 2 handles things fine.
In the init code, if something fails, it gets retried - this seems
to have worked in 99.99% of the cases
---
now all I do is a soft reset and that goes away. Also, the
driver appears to get stuck in a non-error timeout loop sometimes
so that disk access appears extremely slow.
If you need a description of the ports for the TMC I can spend
some time and post it, seems hardly worth it though as this board
is so slow. I only use it for SCSI tapes, disk access is crummy.
Louis Birk.
birk_lj@rso.dnet.ge.com
--
Microsoft is responsible for propogating the evils it calls DOS and Windows,
IBM for AIX (appropriately called Aches by those having to administer it), but neither is as bad as AT&T. Boycott AT&T, and let them know how you feel.