Hi Marek, On Friday 08 April 2011 15:09:02 Marek Szyprowski wrote: > On Friday, April 08, 2011 2:53 PM Laurent Pinchart wrote: [snip] > > buf_queue is called with a spinlock help, so you can't perform I2C > > communication there. > > In videobuf2 buf_queue() IS NOT called with any spinlock held. buf_queue > can call functions that require sleeping. This makes a lot of sense > especially for drivers that need to perform a lot of operations for > enabling/disabling hardware. Oops, my bad. > I remember we discussed your solution where you wanted to add a spinlock > for calling buf_queue. This case shows one more reason not go that way. :) Hehe. I totally agree with you that we should avoid locking wherever possible. We still have no solution for the disconnection problem though. > AFAIR buf_queue callback in old videobuf was called with spinlock held. That's correct, yes. > I agree that we definitely need more documentation for vb2 and clarification > what is allowed in each callback... -- Regards, Laurent Pinchart -- 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
