[sane-devel] Subsequent scans timeout with Canon imageClass MF4150

[Rolf Bensch]

Hi Dan,

It could be helpful if you sniff the usb protocols on a Windows machine.
Then I can compare them with Sane and patch it. Please use this program:
 http://www.sane-project.org/README.linux and scan a small area of
approx. 5mm x 5mm @ 75 dpi to reduce the logfile size.

Cheers,
Rolf


Am 17.01.2013 21:10, schrieb Dan Duvall:
> (Sorry, sent again. Forgot to CC the list.)
> 
> I'm still seeing the same behavior after recompiling without pthread support (through, looking at `configure --help`, it says this is the default under OSes other than OS X).
> 
> One other thing that I forgot to mention is that this first-scan-works-second-scan-hangs behavior is following a power cycle of the scanner. Also, if I power off the scanner while it's timing out on the second scan—which is what people in the office tend to do—it sends scanimage into an infinite loop. I've frequently seen several saned processes spiking the CPU on this server which I attribute to this.
> 
> Could an strace or gdb session help diagnose the problem? I'm willing to really dig in given the right direction.
> 
> Thanks for the help!
> Dan
> 
> On Jan 17, 2013, at 7:33 AM, Rolf Bensch <rolf at bensch-online.de> wrote:
> 
>> Hi Dan,
>>
>> Maybe enabled pthread support is the problem, especially for linux.
>>
>> I would upgrade sane on your lenny system to 1.0.24 with pthread support
>> disabled.
>>
>> Cheers,
>> Rolf
>>
>>
>>
>> Am 17.01.2013 14:32, schrieb Rolf Bensch:
>>>
>>>
>>> Am 16.01.2013 23:35, schrieb Dan Duvall:
>>>> Hey there,
>>>>
>>>> I'm having trouble scanning using the pixma backend with a Canon imageClass MF4150 over USB. The first scan always seem to succeed while subsequent attempts timeout.
>>>>
>>>> These are the commands I'm running.
>>>>
>>>> SANE_DEBUG_PIXMA%6 scanimage 2> pixma_debug-success.log
>>>> SANE_DEBUG_PIXMA%6 scanimage 2> pixma_debug-timeout.log
>>>>
>>>> stderr of each:
>>>> https://gist.github.com/4551560
>>>> https://gist.github.com/4551574
>>>>
>>>> I get this same behavior using sane-backends 1.0.22 on Debian lenny, and 1.0.24 on OS X 10.8.2.
>>>>
>>>> Any suggestions on how to further debug?
>>>>
>>>> Thanks!
>>>> Dan
>>>>
>