[sane-devel] HP 7450C, hpusbscsi, permissions in Fedora Core 3

Thomas Frayne tomf@sjpc.org
01 Jan 2005 22:11:30 -0800


I am running on an up to date Fedora Core 3 system, and have installed
sane-backends 1.0.15-1.4 and xsane 0.92-13 via synaptic.  Over the last
18 months, I have periodically tried to get my HP 7450C scanner working
under SANE, but was never completely successful.  I am now trying again,
and am getting similar problems to those I have had before.

The problems I have seen so far:

1.  hpusbscsi vs libusb.

sane-find-scanner finds the scanner at libusb:002:004.  The hardware
browser and  scanimage -L in root mode finds `avision:/dev/sg2' is a
Hewlett-Packard ScanJet 7400c flatbed scanner.  In the past, I bypassed
this problem by removing the hpusbscsi module and putting something
about libusb into fstab.  However, I was never able to stop the system
from loading hpusbscsi whenever I rebooted and I don't remember seeing
an answer to my queries about this.  Is it still necessary to remove
hpusbscsi?  Is there a way to fix this problem so that the fix will
survive rebooting?  Is there a central place to find the procedure
necessary to do this, or do I have to search through old posts to the
mailing list?

2. User mode vs root mode

xsane wants to run in user mode, but returns "no device found" or gets a
segmentation fault when I try it.  I need to fix the permissions.  In
root mode xsane hangs while it is scanning for devices.

http://www.sane-project.org/README.linux hints at how to permanently
change permissions by changing "libsane.usermap" and "libusbscanner",
and there is more info in the libsane.usermap file.

There was no entry in libsane.usermap for the 7400 or 7450C, so I took
the vender and product from sane-find-scanner, and added 
# Hewlett-Packard|ScanJet 7450C
libusbscanner             0x0003      0x03f0   0x0801    0x0000      
0x0000       0x00         0x00

Please add this entry in the next release of SANE.

I have not added my userid to the scanner group and tested whether this
takes care of the permissions problem.  I'll send a follow-up when I do,
but I might need a fix for the hpusbscsi problem first.