sane mustek mfs-12000sp rev. 1.07

Thomas Wenzke (thomas@twenzke.nacamar.de)
Mon, 23 Feb 1998 00:43:36 +0100 (MET)

Sorry forgot the mail-subject in previous mail

-------------------------------------------------------------

Hello

I have a Mustek Paragon 1200 SP since last week and try it
on my Linux-i386.
It's an old Pentium-133 with EIDE-Disk, SCSI-CDROM and
SCSI-Tandberg-Tape.
There is a NCR-810 Symbios-Controller and a NE-2000 in the Box.
This works well all the days.

I disabled all the SCSI2-Goodies, like disconnect, tagged command
queueing, synchronous transfers and applied the timeout-patch
in the driver.

The Scanner-Type is
MFS-12000SP

The Firmware-Revisiontest tells
Rev. 1.07
(and I think that's my problem ((8-( )

Scanning with the default-resolution works fine - i think.
But altering the resolution to more than 35 dpi results in horizontal
stripes and color-shift away from red in direction green and blue and
grayish colors.

My /usr/local/etc/sane.d/mustek.conf contains
strip-height 1.5 # several values tested from absent to 5.0
/dev/scanner
linedistance-fix # eliminates horizontal strips in default resolution
lineart-fix # no effect in color-mode (i think)

After some tricky testing with flag MUSTEK_FLAG_LD_MFS
(makes more weird colors and stripes)
with sane-0.70 and sane-0.67 - i recognize that flag MUSTEK_FLAG_LD_MFS
has also a positive effect.
without this flag my syslog says at start of a scan :

Feb 22 02:19:12 twenzke1 kernel: ncr53c810-0-<target 5, lun 0>:
extraneous data discarded.
Feb 22 02:19:12 twenzke1 kernel: ncr53c810-0-<target 5, lun 0>:
COMMAND FAILED (9 0) @00ffe810.

further more - without that flag in the original source-versions
i see always the same initial line-distance-factor of 235 over all
resolutions (and i feel serious).
Unfortunatly i am new with scanners (it's my first one).

So i recompile the original source-version of sane-0.67 for debugging.

i think the scanner-command that leeds to this syslog-messages is
distance[] = 17 0 0 0 5 0

my gdb traceback follows :

#0 sanei_scsi_req_wait (id=0x8072760) at sanei_scsi.c:595
#1 0x80653d6 in sanei_scsi_cmd (fd=11, src=0x806a30a,
src_size=6, dst=0xbfff7614,
dst_size=0xbfff7608)
at sanei_scsi.c:678
#2 0x8054811 in dev_cmd (s=0x80925f0, src=0x806a30a, src_size=6,
dst=0xbfff7614, dst_size=0xbfff7608)
at mustek.c:324
#3 0x8055bbc in line_distance (s=0x80925f0)
at mustek.c:1048

after line 594 in sanei_scsi.c :

issue (req); /* ensure the command is running */

the syslog-messages happens.

after that i'm looking for req :

(gdb) p *req
$45 = {fd = 11, running = 1, done = 0, status = SANE_STATUS_GOOD,
dst_len = 0xbfff7608, dst = 0xbfff7614,
cdb = {hdr = {pack_len = 42, reply_len = 41, pack_id = 9,
result = 0, twelve_byte = 0, other_flags = 0,
sense_buffer = '\000' <repeats 15 times>},
data = "\021\000\000\000\005\000\000\000\000\005
\002\000\000\000\000\000d\000",
' ' <repeats 13 times>,
"1.07MUSTEK MFS-12000SP
\000\001\000\000\000\001\003\001\001",
'\000' <repeats 27 times>, "",
'\000' <repeats 130463 times>},
next = 0x0}

the toplevel-function dev_cmd() in line_distance returns
SANE_STATUS_GOOD
result_len = 5
result is always 235 0 0 0 5

Is this a new (wonderful) firmware-revision from mustek ?

What can i do ? Please help !

best regards

Thomas Wenzke

PS: If you reply, you can send a CC to my office.
The EMail-Adress is wenzke@nacamar.net

--
Source code, list archive, and docs: http://www.mostang.com/sane/
To unsubscribe: echo unsubscribe sane-devel | mail majordomo@mostang.com