Ack, phhfft. Ignore this. Literally right after I hit "Send", one of the 
sane developers told me that the device is sufficiently different from 
other HP models that it won't work with the HP driver.

/me goes back to the drawing board.

On 5/26/24 4:47 PM, Ian Darwin wrote:
> Scanimage (and xsane) give generic failure message instead of 
> actionable perror output.
>
> This is on amd64, and has failed on 7.4, 7.5 and -current. Sane config 
> files are stock except adding the USB id for this scanner and 
> commenting out all the unused devices in /etc/sane.d/dll.conf.
>
> $ scanimage -L
> device `hpaio:/usb/DeskJet_4100_series?serial=CN23JFG31P' is a 
> Hewlett-Packard DeskJet_4100_series all-in-one
> $ scanimage > /tmp/id.pnm
> Output format is not set, using pnm as a default.
> scanimage: open of device 
> hpaio:/usb/DeskJet_4100_series?serial=CN23JFG31P failed: Error during 
> device I/O
> $ ktrace scanimage > /tmp/id.pnm
> Output format is not set, using pnm as a default.
> scanimage: open of device 
> hpaio:/usb/DeskJet_4100_series?serial=CN23JFG31P failed: Error during 
> device I/O
> $ kdump > /tmp/id.ktrace
>
> Relevant bit from ktrace follows. Some time after it has successfully 
> opened and read from /usr/local/lib/libhmud.so.2.0, it blows with 
> this, seemingly looking in /usr/lib:
>
>  81078 scanimage CALL  sendsyslog(0x7c360a32b4c0,87,0<>)
>  81078 scanimage GIO   fd -1 wrote 87 bytes
>        "<11>scanimage: common/utils.c 246: unable to load library 
> libhpmud.so.0: File not found"
>  81078 scanimage RET   sendsyslog 0
>
> By "relevant" I mean" right after this message, it starts unloading 
> and closing and giving the ever-helpful "failed: Error during device 
> I/O" cop-out.
>
> All the syslog output from one run:
>
> 2636:May 26 16:32:46 dalai scanimage: scan/sane/escl.c 434: 
> escl_open() session=0x0
> 2637:May 26 16:32:46 dalai scanimage: io/hpmud/musb.c 428: Found 
> interface conf=0, iface=1, altset=0, index=1
> 2638:May 26 16:32:46 dalai scanimage: io/hpmud/musb.c 390: Active 
> kernel driver on interface=1 ret=-12
> 2639:May 26 16:32:46 dalai scanimage: io/hpmud/musb.c 536: claimed 
> 7/1/2 interface
> 2640:May 26 16:32:46 dalai scanimage: io/hpmud/musb.c 781: read actual 
> device_id successfully fd=1 len=299
> 2641:May 26 16:32:46 dalai scanimage: io/hpmud/musb.c 562: released 
> 7/1/2 interface
> 2642:May 26 16:32:46 dalai scanimage: common/utils.c 246: unable to 
> load library libhpmud.so.0: File not found
> 2643:May 26 16:32:46 dalai scanimage: common/utils.c 246: unable to 
> load library libhpmud.so.0: File not found
> 2644:May 26 16:32:46 dalai scanimage: scan/sane/escl.c 89: Calling 
> escl bb_unload:
>
> Full 6MB ktrace at https://darwinsys.com/tmp/scanimage-kdump.txt
>
> $ cd /usr/lib
> $ ls -lL libhpmud*
> -rw-r--r--  1 root  bin  226380 May 11 22:25 libhpmud.a
> -rw-r--r--  1 root  bin    1121 May 11 22:25 libhpmud.la
> -rw-r--r--  1 root  bin  140272 May 11 22:25 libhpmud.so.2.0
>
> All these packages are up-to-date:
>
> cups-2.4.8 foomatic-db-4.0.20210611            hplip-3.23.12p0
> cups-browsed-2.0.0 foomatic-db-engine-4.0.13p4 hplip-common-3.23.12p0
> cups-filters-2.0.0p0 hpaio-3.23.12p0 sane-backends-1.3.1
> cups-libs-2.4.8                     hpcups-3.23.12p0
>
> This worked a year or so ago when I last used it, though I'm not 100% 
> sure if it was this scanner or its predecessor, which isn't much to go 
> on, but any thoughts?
>
> The recent commit to sane-backends is not the cause as it was failing 
> before that.
>