On 09/30/2016 11:53 AM, Daniel Vetter wrote: > On Thu, Sep 29, 2016 at 11:44 PM, Marek Vasut <marex@xxxxxxx> wrote: >> I have the following right now, I think that's more descriptive as this >> function is not preparing the FB in any way. >> >> /** >> * drm_fb_cma_extract_and_attach_fence() - Extract fence from plane and >> attach to planestate >> * @plane: Which plane >> * @state: Plane state attach fence to >> * >> * If the plane fb has an dma-buf attached, fish out the exclusive >> * fence and attach it to plane state for the atomic helper to wait >> * on. >> */ > > That choice of color is ok with me too, but then you need to have a > pile of text to explain where it should be used (i.e. directly as the > prepare_fb hook). There can be more stuff in the prepare_fb hook though. > And it is a bit inconsistent with all the other > helpers that can be put into hooks directly. Also, > extract_and_attach_fence _is_ officially part of what a driver should > do in their ->prepare_fb hook. Hence I'm still leaning towards that > (and it's shorter!). And with cma there's nothing else to do (dma > memory is always pinned, which is the other thing prepare_fb should > do), so really it's not even a lie: Your function fully prapares a cma > fb for display as expected by th atomic helpers. OK, I have no good argument not to go for this one. Unless someone has, I will redo it. > -Daniel > -- Best regards, Marek Vasut _______________________________________________ dri-devel mailing list dri-devel@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/dri-devel
