Hi Vinod, Hi Russell, On 11 December 2014 at 11:42, Jassi Brar <jaswinder.singh@xxxxxxxxxx> wrote: > On 11 December 2014 at 10:17, Vinod Koul <vinod.koul@xxxxxxxxx> wrote: >> On Tue, Dec 09, 2014 at 08:48:04PM +0530, Jassi Brar wrote: >>> As Russell pointed out, that ain't the case either. >>> So we are yet to figure out benefits of having explicit >>> issue_pending() after tx_submit(). >> callback ? >> > The callback is set after prep() and before tx_submit(), but here we > talk after tx_submit(). Perhaps the idea dates back to async-only days, when client drivers would prepare and queue descriptors in the controller driver rather than having to manage the dependency queues themselves (?). Today ~95% clients are slave and I am yet to find one that really can't work with submit and issue_pending tied together. Not to forget the 100% of the controller drivers have to manage 'submitted' and 'active' queues -- only to have arguably negative side-effects. If we agree that clubbing submit and issue_pending is the right thing to do, I can start converting the ~90 client drivers. Please let me know either way. Cheers! -- To unsubscribe from this list: send the line "unsubscribe dmaengine" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
