Re: I did a test on a DEC AlphaStation 600 and the SCSI driver seems borked

Dennis Clarke <[email protected]> Mon, 16 Feb 2026 15:15:11 -0500
Newsgroups gmane.os.netbsd.ports.alpha
Organization GENUNIX
Message-ID <[email protected]>
On 2/16/26 1:26 PM, Magnus Lindholm wrote:
>>
>> This looks to be a different SCSI controller type :
>>
> 
> yeah, I'm booting from an adaptec controller
> 

Hrmmmm ... I have no internal SCSI disks at all. Unless we count the
CDROM which seems to work just fine. I did discover over lost hours
that one must use CD-R media and *not* CD-RW. The old CDROM works
fine with the CD-R stuff if the burn is very very slow.

.
.
.


>>
>> How are these disks connected to the SCSI controller(s) ?
>>
>>
> 
> the root disk is connected to the adaptec internal 68-pin
> connector, the disk attached to the qlogic ISP controller
> is connected via the external scsi connector via a
> disc cabinet (DEC)
> 


Well here I am at a loss because I am using an external Sun SCSI storage
array which may be old but it is a battle axe. Works flawlessly on every
machine I ever tried and that includes this DEC AlphaStation 600. Just
not with the NetBSD isp driver.

Looking in sys/dev/ic/isp/isp.c where I know we need to deal with an
error etype at line 5279. However, I can not handle the error which
seems to be unknown and we fell into the default :

         default:
             isp_print_bytes(isp, "Unhandled Response Type", QENTRY_LEN, 
qe);
             if (XS_NOERR(xs)) {
                 XS_SETERR(xs, HBA_BOTCH);
             }
             break;

So then wherefore do we get this isp0: Unhandled Response Type 0x9 ?

This all starts with the LeeLoo Dallas "Multi-Pass" at line 5053 :

     buddaboom = req_status_flags = req_state_flags = 0;


Clearly, yes ..... big buddaboom. Big.

I would need to make this verbose as hell and rebuild isp and see what
sort of really Big Budda Boom noises are made.


--
Dennis Clarke
RISC-V/SPARC/PPC/ARM/CISC
UNIX and Linux spoken