Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 19 May 1999 02:53:18 -0700 (PDT)
From:      Piet Delaney <piet@cup.hp.com>
To:        piet@cup.hp.com, davidh@progmatics.com.au, hasty@netcom.com, davidh@wr.com.au
Cc:        freebsd-ports@freebsd.org, obrien@nuxi.com, piet@hpfsvr02.cup.hp.com, piet@piet.net, sane-devel@mostang.com
Subject:   Re: Sane 1.0.1 - Looks like /usr/local/lib/libsaine-net.so.1 wasn't installed.
Message-ID:  <199905190953.CAA11098@hpfsvr02.cup.hp.com>

next in thread | raw e-mail | index | archive | help
> 
> Well, I have checked my config here. My dmesg reports:
> 
> pt0 at ncr0 bus 0 target 5 lun 0
> pt0: <HP C5100A R032> Fixed Processor SCSI-2 device
> pt0: 3.300MB/s transfers

Perhaps your using on an old kernel. Mine comes up with pass0 and pass1
so it's using the new (brooken?) CAM interface.


> 
> My hp.conf file has in it:
> 
> scsi HP
> /dev/scanner
> 
> My /dev reports:
> 
> [4 ishtar: etc/sane.d] % ls -l /dev/scanner
> lrwxr-xr-x  1 root  wheel  3 Feb  3 07:23 /dev/scanner@ -> pt0
> [5 ishtar: etc/sane.d] % ls -l /dev/pt0
> crw-rw-rw-  1 root  sysprog   61,   0 Feb  1 08:30 /dev/pt0

So symbolic pointers aren't a problem. I had a feeling that was the case.


> 
> My kernel is configured to have:
> 
> device pt0 at scbus?                    # SCSI processor type
> device sctarg0 at scbus?                # SCSI target

Interesting. My GENERIC kernel config doesn't have this.
I'm using version 1.169 of GENERIC. What version are you using?

Does your have a line:

	device pass0




> 
> For some reason, when I run scanimage --list-devices,

In my case I think scanimage is looking for the wrong type of
objects in /usr/local/lib/sane. I have all .a and .dl files.
What are .dl files (dynamic link I assume). Seems sane is
looking for .so.1 files (shared objects).


> mine doesn't
> appear, but if I type
> 
> xscanimage hp:/dev/pt0

I don't recall the docs suggesting this : format? Where was that?
That style is usually used by X. So scanimage and xscanimage have
a different syntax for the device even though it's the same executable.
Boy talking about under-documented features. I think some OS specific
examples would be helpfull.



> 
> everything runs as expected.

Sure does, but I'm still stuck. Not surprise Dave O'Brien gave up
on this turkey. It's not very close to the Microsoft plug and play
paradigm. 

I'll read the SANE 1.01 manual when I get home; might help.
Seems like the Indian Blind man story where every observer
has a different view of reality. I suppose that's being in-sane.


-piet

> 
> Hope that helps.
> On Tue, May 18, 1999 at 11:58:23PM -0700, Piet Delaney wrote:
> > 
> > > The scsi pass thru on FreeBSD is strange :(
> > > 
> > > Do a man on :
> > > pass
> > > camcontrol
> > > 
> > > ls -ald /dev/xpt0 /dev/pass1
> > > crw-------  1 hasty  operator  104,   0 Nov 22 23:38 /dev/xpt0
> > > 
> > > crw-rw-r--  1 root  wheel   31,   3 Apr 19 11:20 /dev/pass1
> > 
> > I think using pass1 with minor device 1 would be less confusing.
> > 
> > 
> > > 
> > > pass1 is my umax scanner -- the minor device in /dev/pass1 denotes the 
> > > pass device printed with dmes:
> > > {hasty} dmesg | grep pass
> > > pass3 at ahc1 bus 0 target 5 lun 0
> > > pass3: <UMAX UMAX S-12 V2.0> Fixed Scanner SCSI-2 device 
> > > pass3: 3.300MB/s transfers
> > > 
> > > cd /usr/local/etc/sane.d
> > > {hasty} ls
> > > abaton.conf     dc210.conf      hp.conf         ricoh.conf      tamarack.conf
> > > agfafocus.conf  dc25.conf       microtek.conf   s9036.conf      umax.conf
> > > apple.conf      dll.conf        microtek2.conf  saned.conf
> > > artec.conf      dmc.conf        mustek.conf     sharp.conf
> > > coolscan.conf   epson.conf      net.conf        snapscan.conf
> > > {hasty} cat umax.conf
> > > /dev/pass1
> > 
> > My umax file has more stuff in it:
> > 
> > scsi UMAX * Scanner
> > scsi LinoHell Office
> > .
> > .
> > .
> > /dev/scanner
> > 
> > 
> > 
> > > Make doubly sure that your *.conf entry where is required does point to 
> > > a string /dev/pass[0-9] otherwise the user cam interface will fail
> > > to talk to your scsi device --- for instance if I change umax.conf
> > > to point to /dev/scanner which is a symlink -- the scsi interface
> > > library will fail:
> > 
> > I changed from using the symbolic link (Like the docs recomend!),
> > but it didn't change anything. Seems kinda absurd that the guys
> > that wrote this don't fix it. There must be a rational reason for this.
> > 
> > 
> > > 
> > > Don't forget to create the /dev/xpt0 device or again the cam (scsi) 
> > > interface library will fail to access your scsi device.
> > > That should do it. I found all this out by doing a ktrace on "sane" 
> > > and examining the devices which it opened :(
> > 
> > I ran scanimage with ktrace and it seems to be stopping when scanimage
> > tried to access /usr/local/lib/libsane-net.so.1 which doesn't exist.
> > Perhaps the port makefile is missing the install of libsane-net?
> > I briefly looked at the makefile and didn't see where libsain-net
> > is installed. 
> > 
> > -piet
> > 
> >  
> 
> -- 
> Cheers,
> david		-- http://www.angelfire.com/wa/hobley
> 


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-ports" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199905190953.CAA11098>