> -----Original Message----- > From: Alan Stern [mailto:st...@rowland.harvard.edu] > Sent: Wednesday, September 9, 2015 6:30 PM > To: Laurent Pinchart > Cc: Hans de Goede; Kaukab, Yousaf; linux-media@vger.kernel.org; > mche...@osg.samsung.com; linux-...@vger.kernel.org > Subject: Re: [PATCH v1] media: uvcvideo: handle urb completion in a work > queue > > On Wed, 9 Sep 2015, Laurent Pinchart wrote: > > > > > Instead of fixing the issue in the uvcvideo driver, would it then > > > > make more sense to fix it in the remaining hcd drivers ? > > > > > > Unfortunately that's not so easy. It involves some subtle changes > > > related to the way isochronous endpoints are handled. I wouldn't > > > know what to change in any of the HCDs, except the ones that I maintain. > > > > I'm not saying it would be easy, but I'm wondering whether it makes > > change to move individual USB device drivers to work queues when the > > long term goal is to use tasklets for URB completion anyway. > > I'm not sure that this is a long-term goal for every HCD. For instance, there > probably isn't much incentive to convert a driver if its host controllers can > only > run at low speed or full speed. >
I can convert this patch to use tasklets instead and only schedule the tasklet if urb->complete is called in interrupt context. This way, hcd using tasklet or not, uvc driver behavior will be almost same. Only difference is that local interrupts will still be enabled when tasklet scheduled from uvc driver is executing the completion function. This patch can be dropped once all hcd's start using tasklets for the completion callback (if that ever happens). BR, Yousaf -- To unsubscribe from this list: send the line "unsubscribe linux-media" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
