Re: gspca_pac7302: sporatdic problem when plugging the device

["Németh Márton" <nm127@freemail.hu>]

Jean-Francois Moine wrote:
> On Sun, 10 Jan 2010 17:58:58 +0100
> Németh Márton <nm127@freemail.hu> wrote:
> 
>> Then I plugged and unplugged the device 16 times. When I last plugged
>> the device I get the following error in the dmesg:
>>
>> [32393.421313] gspca: probing 093a:2626
>> [32393.426193] gspca: video0 created
>> [32393.426958] gspca: probing 093a:2626
>> [32393.426968] gspca: Interface class 1 not handled here
>> [32394.005917] pac7302: reg_w_page(): Failed to write register to
>> index 0x49, value 0x0, error -71
>> [32394.067799] gspca: set alt 8 err -71
>> [32394.090792] gspca: set alt 8 err -71
>> [32394.118159] gspca: set alt 8 err -71
>>
>> The 17th plug was working correctly again. I executed this test on an
>> EeePC 901.
>>
>> This driver version contains the msleep(4) in the reg_w_buf().
>> However, here the reg_w_page() fails, which does not have msleep()
>> inside. I don't know what is the real problem, but I am afraid that
>> slowing down reg_w_page() would make the time longer when the device
>> can be used starting from the event when it is plugged.
> 
> Hi again,
> 
> I don't understand what you mean by:
>> This driver version contains the msleep(4) in the reg_w_buf().
>> However, here the reg_w_page() fails, which does not have msleep()
>> inside.
I tought that the msleep(4) call introduced recently fixed the plug-in
problem. It seems I misunderstood something.

> Indeed the delay will slow down the webcam start (256 * 4 ms = 1s).
> 
> If having a delay fixes the problem, then, as the error always occurs
> at the same index 0x49 (3 reports), a single delay could be set before
> writing to this index. Do you want I code this for test?

I tested the behaviour a little bit more. Out of 100 plug-ins:

OK: 81 times
"pac7302: reg_w_page(): Failed to write register to index 0x49, value 0x0, error -71": 19 times

Other error message I haven't got, so 19% of the time writing to register
index 0x49 fails in reg_w_page(). So let's try doing fixing the way you
described. If you send me a patch I can test it.

Regards,

	Márton Németh