[sane-devel] canon_pp backend problem

[Matthew Duggan]

On Fri, Dec 17, 2004 at 11:23:59AM +0100, Thorsten Scherf wrote:
> hi,

G'day,

> I installed the lasted sane-backend tarball (1.0.15) and the
> xsane-frontend (0.96). I commented out the canon_pp line in dll.conf
> and build the binary files.

Wait, do you mean you commented it out or un-commented it?  It's
commented out by default, or it should be because it's not safe to
always run the detection (it's slow and might interfere with printers
and scanners).

> the canon_pp.conf look like:
> 
> ieee1284 parport0
> calibrate ~/.sane/canon_pp-calibration-pp0 parport0
> init_mode FB620P parport0 #(tried it also with AUTO)

I assume you're trying to use an FB620P or FB320P then?

> after setting SANE_DEBUG_CANON_PP=100 and starting xsane,  I got the
> following messages:
> 
> [root@kermit downloads]# /usr/local/bin/xsane
...
> [canon_pp] sane_init: Finding scanner on port 'parport0'
> [canon_pp] sane_init: Found!

I really should change this debug message, it's doing nothing of the
sort.  It's finding the scanner data structure associated with that port
name.

> [canon_pp] detect_mode: Using ECP-S Mode

This is always a worry.. I've had a lot of trouble with ports in ECP-S
mode.  What kernel and motherboard chipset are you running out of
interest?

> [canon_pp] sane_init: >> initialise
> [canon_pp] WARNING: Don't know how to reset an FBx20P, you may have to
> power cycle

This only helps if they're ever replying.  The FB620P and FB320P are
very dumb and difficult to reset when they're not listening.  I've seen
the Windows driver do some stupid things, and they would have had the
full spec.. I have no idea what I'm doing, so I'm afraid it's power
cycles whenever the scanner hasn't been shut down properly :(

> [canon_pp] Timeout: Scanner wakeup reply 1 (0x03 in 0x1f) - Status =
> 0x1f
... (no response)
> [canon_pp] initialise: could not wake scanner
> [canon_pp] sane_init: Couldn't contact scanner on port parport0.
> Probably no scanner there?
> [root@kermit downloads]#
> 
> and a gui popped up, that no scanner was found.
> 
> manually reseting the scanner does not help.

It seems from that output that the driver's not getting any response on
the port at all when it sends the wakeup signals.  If you have a printer
on the passthrough port, try removing it.  If your parallel port
supports DMA transfers, try giving parport_pc a dma argument.  Like I
said before, ECP-S seems to cause trouble on some chipsets.  I wish I
had one of them handy so I could try to work out why, but both my
machines behave well.

Hope that's of some use.

Cheers,

- Matthew Duggan