[sane-devel] Questions on debugging fujitsu backend (for SCSI fi-4750C)
Mike Wirth
mwirth at gmail.com
Mon Sep 23 21:32:03 UTC 2013
Folks,
I have a Fujitsu fi-4750C SCSI-scanner (Actually, two of them, one recently
for spare parts, but it may be good enough for use. Will know soon. Ping
me directly if interested.)
Works fine using the standard Fujitsu WinXP TWAIN driver, using Acrobat 10
as the scanning app, but slow (running on a wimpy old ThinkPad + dock),
including ADF duplex scanning.
Built and installed SANE on a Mac G5 machine (Note: PPC processor) running
Mac OS X (latest PPC-compatible version). Works great, faster, including
ADF Front Side. BUT NOT anything involving ADF Back Side (either back side
alone, or duplex). I get a:
do_scsi_cmd: return 'Error during device I/O'
In trying to chase this down, I set the SANE_DEBUG_FUJITSU environment
variable at its noisiest level, 35, and ran both front side ADF
(successful) and back side ADF (failed) scans and compared the (huge!)
logs. They are identical, up to the point of error (at line 9,122), except
for the following lines:
- started=0, side=0, source=*1* <-- for ADF front, *2* for back
- do_scsi_cmd: start … *<writes 10 bytes, then 72 bytes, in the 0-9th
byte of the 72…>*
- 00…xx…*00* *<for front>*
- 00…xx…*80* *<for back>*
- do_scsi_cmd: start … *<writes 10 bytes, in the 0…6th byte of the 10…>*
- 28…xx…*00* *<for front>*
- 28…xx…*80* *<for back>*
- setup_buffers: free buffer *1* <-- for ADF front, *0* for back
- do_scsi_cmd: start … *<writes 6 bytes, then 1 byte, which are…>*
- 1b 00 00 00 01 00, then *00* *<for front>*
- 1b 00 00 00 01 00, then *80* *<for back>*
The changes from source = 1 to 2, and in "00" to "80" appear *plausible* for
changing ADF front to back, the difference in 'free buffer" less so. But,
not knowing the SCSI command structure for the scanner, I can't proceed.
Thus, a couple questions:
1. Can anyone provide me with the SCSI command listing and docs for the
Fujitsu fi-4750C?
2. Anybody have a maintenance manual for this or similar scanner?
3. Any other hints on how to proceed?
Would like to provide a fix for this before your next release comes out :-)
TIA,
Mike
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.alioth.debian.org/pipermail/sane-devel/attachments/20130923/c93d8856/attachment-0001.html>
More information about the sane-devel
mailing list