[sane-devel] MX375 no recovery from ADF empty when scan started

[Rolf Bensch]

Hi Matthias,

Please try attached patch. It is based on patch1-new.

Again I need a log file created with 'export SANE_DEBUG_PIXMA=11'.

Cheers,
Rolf



Am 21.05.2014 22:33, schrieb Matthias Peter Walther:
> Sure. Zipped, cause they are larger than 100 KB.
> 
> On 21.05.2014 21:44, Rolf Bensch wrote:
>> Hi Matthias,
>>
>> Please create logfiles again with 'export SANE_DEBUG_PIXMA=11'.
>>
>> Many thanks.
>>
>> Cheers,
>> Rolf
>>
>>
>> Am 21.05.2014 20:53, schrieb Matthias Peter Walther:
>>> Hallo Rolf,
>>>
>>> one fixed, one to go.
>>>
>>> An empty ADF still crashes the device. But it starts scanning now, if I
>>> insert paper before the timeout runs out.
>>>
>>> I created logs for both issues.
>>>
>>> Grüße
>>> Matthias
>>>
>>> On 21.05.2014 17:27, Rolf Bensch wrote:
>>>> Hi Matthias,
>>>>
>>>> I created a new patch1. This should fix both issues. I also added a new
>>>> debug output to check paper empty code.
>>>>
>>>> Cheers,
>>>> Rolf
>>>>
>>>>
>>>> Am 21.05.2014 09:51, schrieb Matthias Peter Walther:
>>>>> Hallo Rolf,
>>>>>
>>>>> I can't notice any change of behavior here. To make sure, that I use the
>>>>> patched version, I manually deleted all libsane*.
>>>>>
>>>>> See log attached.
>>>>>
>>>>> And the scan does not start, if I insert paper during the timeout
>>>>> countdown. The scanner beeps, so the firmware detects the inserted
>>>>> paper. But the scan does not start.
>>>>>
>>>>> Grüße
>>>>> Matthias
>>>>>
>>>>> On 20.05.2014 22:44, Rolf Bensch wrote:
>>>>>> Hi Matthias,
>>>>>>
>>>>>> Am 20.05.2014 17:49, schrieb Matthias Peter Walther:
>>>>>>> Hallo Rolf,
>>>>>>>
>>>>>>> sorry to bother you again. But here is another bug with my MX375.
>>>>>> No problem.
>>>>>>
>>>>>>> If I start an ADF-scan while the ADF is empty, the scanner does not
>>>>>>> recover and needs to be restarted by hand.
>>>>>> Attached patch should fix this. The patch is based on recent git code.
>>>>>>
>>>>>>> [...]
>>>>>>>
>>>>>>> I mean scanning ADF without paper in it is stupid, but it should not
>>>>>>> crash the device :)
>>>>>> It's not stupid. Maybe you started the job and forgot to put the paper
>>>>>> into the ADF. Then the job should start scanning if you put the paper
>>>>>> into the ADF *before* timeout counted down to 0 and aborts the job.
>>>>>>
>>>>>> Please test, if scanning starts if you put paper into the ADF during
>>>>>> timeout counts down.
>>>>>>
>>>>>> Cheers,
>>>>>> Rolf
> 
-------------- next part --------------
--- ./pixma_mp150.c	2014-05-23 20:02:28.000000000 +0200
+++ ../sane-backends/backend/pixma_mp150.c	2014-05-23 20:17:41.000000000 +0200
@@ -1361,6 +1361,10 @@
       if (!has_paper (s))
       {
         PDBG (pixma_dbg (4, "*mp150_scan***** no paper in ADF *****\n"));
+        error = abort_session (s);
+        if (error < 0)
+          PDBG (pixma_dbg (1, "WARNING:abort_session() failed %d\n", error));
+
         if (mp->generation == 4 && mp->adf_state == state_idle)
           {
             if (!send_xml_dialog (s, XML_END))