sane 0.66 bug report

David Banz (banz@athene.informatik.uni-bonn.de)
Mon, 03 Nov 1997 10:21:19 +0100

Scanner access not *that* easy... :-(

My system:
P133, 32 MB RAM, Adaptec AHA-2940 Bios 1.21, Microtek Scanmaker E6,
Linux 2.0.25 (SCSI drivers compiled as modules), 2 IDE-HDs, 1 SCSI-HD,
1 ATAPI-CDROM, Xfree 3.3, Afterstep 1.0 / Kwm (KDE Window Manager),
SBpro compatibe soundcard on A220,I5,D1, Tseng ET4000/W32p PCI video
card,
"gtk+971025" from ftp.gimp.org

SG_BIG_BUF in sg.h increased to legal maximum, as explained in
SCSI-manpage.
I assumed since the SCSI drivers are all modules, I just need to do a
"make modules", and not a complete kernel-recompilation. Is this
correct?

I use the following commands (in that order) to insert SCSI support:
insmod scsi_mod
insmod aic7xxx
insmod sd_mod
insmod sr_mod
insmod sg
My Conner SCSI-HD works perfectly with this.

The following things went wrong:
- scanimage quits with a segmentation fault right when the scan is about
to
begin, if I want to scan with about 280 dpi or more (the E6 is capable
of
600 dpi scans)
- xscanimage immediately quits with "ERROR sigsegv caught", if I want to
use the Microtek-backend, I can see the main window for half a second
or
so (the pnm-backend however, works perfectly)

Scanning via TWAIN under Windows 3.1 works perfectly in any resolution.
I hope you can help me.

David Banz

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