Subject: Re: si driver
To: None <mouse@Collatz.McRCIM.McGill.EDU>
From: Gordon W. Ross <>
List: port-sun3
Date: 11/29/1995 13:57:22
> Date: Sat, 25 Nov 1995 21:13:14 -0500
> From: der Mouse <mouse@Collatz.McRCIM.McGill.EDU>
> Cc:
> My Sun-3/150 just exhibited the "short transfer: wanted 10 got 6."
> misbehavior.  Here's what happened; I was doing a network comparison
> that would have been doing a lot of reads, but no writes except
> possibly for inode access time updates:
> [Daily-Planet] 92> si0: SCSI bus busy, resetting...
> ncr5380_command: short transfer: wanted 10 got 6.
> si(0:0:0)-28,0,0,4,e6,40,0,0,4,0-
> Stopped at      _Debugger+0x6:  unlk    a6
> db> 
> So I remembered what someone said, and typed "call ncr5380_show_state",
> which produced
> I didn't try the show_trace call because I had no way to save the
> voluminous result.  Instead, I typed "cont", which resulted in
> panic: ncr5380_scsi_cmd: polled request, abort failed
> Stopped at      _Debugger+0x6:  unlk    a6
> db> 
> So I just let it crash and reboot...
> The disk is reported as
> si0 at vmes0 addr 0xff200000 level 2 vector 0x40
> scsibus0 at si0
> si0 targ 0 lun 0: <SEAGATE, ST15230N, 0498> SCSI2 0/direct fixed
> sd0 at scsibus0: 4095MB, 3992 cyl, 19 head, 110 sec, 512 bytes/sec
> and I will be glad to provide any further info desired.  I may even be
> able to make it happen again, if I'm (un)lucky.

I'm not sure why it could not "abort" the command, but maybe it was
due to being stopped in the debugger for a while.  Turn off the
ncr5380_debug variable (as described in previous mail) so the driver
will not stop on short command transfers.  Then, if this problem
persists, try increasing the value of ncr5380_wait_req_timo
(double it, perhaps) and see if that helps.