[791] in linux-scsi channel archive
Re: SCSI scanner project
daemon@ATHENA.MIT.EDU (Drew Eckhardt)
Thu Nov 9 22:44:05 1995
To: Kenneth Stailey <kstailey@owl.dol-esa.gov>
cc: johnsonm@nigel.vnet.net, linux-scsi@vger.rutgers.edu
In-reply-to: Your message of "Wed, 08 Nov 1995 11:58:49 EST."
<199511081658.LAA05381@owl.dol-esa.gov>
Date: Thu, 09 Nov 1995 10:56:46 -0700
From: Drew Eckhardt <drew@poohsticks.org>
In message <199511081658.LAA05381@owl.dol-esa.gov>, kstailey@owl.dol-esa.gov wr
ites:
>The UMAX UC-630 scanner does some nasty undocumented stuff too. It has a
>behavior I called "falling off the bus" where after sending certain
>commands it would return a SCSI message (or something) that suggested it
>was ready for another command when this simply wasn't the case.
Probably COMMAND COMPLETE.
>To work around this I added a routine to loop doing TEST UNIT READY
>and sleeping a couple of seconds until TEST UNIT READY returns OK.
>My guess is that they were trying to do a bogus version of SCSI disconnect
>since the scanner didn't support disconnect,
No "disconnect" of any sort; devices like this simply return COMMAND COMPLETE
after the command has been parsed and started exeucting.
>Has anyone ever seen another SCSI peripheral that does this?
Many of the slow (Sequential access rewind, retension, file space
forward and reverse; direct access format) SCSI commands have an
IMMEDIATE bit, where if set the device returns COMMAND COMPLETE
real soon.
