Laurent, I am not sure if the issue is 10% related to the -84 error, because as you can see on the following log msgs. the interval is kind of high, which made me think that this is only part of the issue. This log was dumped during the execution of luvcview (2007/05/12).
Thx again for your support and in the name of some lost souls like me thx ;). Please let me know if you need me to test or evaluate anything else. A+ [41198.716000] uvcvideo: Non-zero status (-84) in video completion handler, retrying. [41386.168000] uvcvideo: Non-zero status (-84) in video completion handler, retrying. [41537.464000] uvcvideo: Non-zero status (-84) in video completion handler, retrying. [41648.048000] uvcvideo: Non-zero status (-84) in video completion handler, retrying. [41695.720000] uvcvideo: Non-zero status (-84) in video completion handler, retrying. [41744.872000] uvcvideo: Non-zero status (-84) in video completion handler, retrying. [41799.584000] uvcvideo: Non-zero status (-84) in video completion handler, retrying. [41934.364000] uvcvideo: Non-zero status (-84) in video completion handler, retrying. [41978.436000] uvcvideo: Non-zero status (-84) in video completion handler, retrying. [42068.664000] uvcvideo: Non-zero status (-84) in video completion handler, retrying. [42213.120000] uvcvideo: Non-zero status (-84) in video completion handler, retrying. On Fri, Apr 25, 2008 at 12:50 AM, Laurent Pinchart <[EMAIL PROTECTED]> wrote: > Hi Daniel, > > > On Wednesday 23 April 2008, Daniel CABRERA wrote: > > Laurent, > > > > > I just tested your patch with the SVN 205 and its able to recover. I > > tried with luvcview and Skype with the same results. > > Ok. I'll try to come up with a cleaner patch and I'll then ask you to test > it. > > > > I recompiled without the printk ;) but I noticed that as soon as plug > > my webcam and the uvcvideo driver is loaded the performance of my > > laptop is going down down down. In a couple words is running slower. > > Any idea why? > > Maybe because the -84 error keeps getting generated. The UVC driver will then > resubmit URBs without any delay, which will use quite a lot of CPU time. Of > course, without the printk, it's quite difficult to tell. > > I'll try to address that in a new patch, printing a message when too many -84 > errors are generated. > > > > Best regards, > > Laurent Pinchart > _______________________________________________ > Linux-uvc-devel mailing list > [email protected] > https://lists.berlios.de/mailman/listinfo/linux-uvc-devel > _______________________________________________ Linux-uvc-devel mailing list [email protected] https://lists.berlios.de/mailman/listinfo/linux-uvc-devel
