Hello Mauro, On 12/15/2016 09:37 AM, Mauro Carvalho Chehab wrote: [snip] > > What happens is that omap3isp driver calls media_device_unregister() > too early. Right now, it is called at omap3isp_video_device_release(), > with happens when a driver unbind is ordered by userspace, and not after > the last usage of all /dev/video?? devices. > > There are two possible fixes: > > 1) at omap3isp_video_device_release(), streamoff all streams and mark > that the media device will be gone. > > 2) instead of using video_device_release_empty for the video->video.release, > create a omap3isp_video_device_release() that will call > media_device_unregister() when destroying the last /dev/video?? devnode. > There's also option (3), to have a proper refcounting to make sure that the media device node is not freed until all references to it are gone. I understand that's what Sakari's RFC patches do. I'll try to make some time tomorrow to test and review his patches. Best regards, -- Javier Martinez Canillas Open Source Group Samsung Research America -- To unsubscribe from this list: send the line "unsubscribe linux-media" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html