Hi, On Tue, Mar 29, 2016 at 06:28:07PM +0200, Gregory CLEMENT wrote: > >> So I checked and for this XOR engine there is a real minimum transfer > >> size, whereas there is no constraint on the alignment. (Actually it is > >> better to be aligned on the burst size, but it is not mandatory at all) > >> > >> So I think it would make sens to add this information, as for the > >> alignment the modules using the dma engine still can ignore this but it > >> could be an interesting information. > > > > If it's really a minimum transfer size, no one should ignore it, and > > the framework should reject any transfer size lower than that. > > Currently it is the driver who rejects it. What I thought is if the user > can have this information it is possible to optimize the > transfer. Instead of using a fallback if the size is too small, then the > user can either chose to directly not using the DMA or to use a bigger > buffer. Yes, having it available to the users is something that we definitely need, but you'll also have to check that the users cooperate. And duplicating that code over and over again in the drivers seems counter-productive. Maxime -- Maxime Ripard, Free Electrons Embedded Linux, Kernel and Android engineering http://free-electrons.com
Attachment:
signature.asc
Description: Digital signature