[PATCH] dmaengine: Add support for BCM2835.

Andre Heider a.heider at gmail.com
Tue Nov 12 03:58:14 PST 2013


On Mon, Nov 11, 2013 at 01:06:59PM +0100, Martin Sperl wrote:
> Hi Andre!
> >Right now, considering that we have a) but not b), the latter seems like
> >the obvious choice. And since the dma channel mask is consistent over a
> >power cycle I'd think it's a sane choice.
> (or did you mean "former" not "latter"?)

Nope, I meant the latter, read it from the device tree.

> So what does block us from adding the bcm2835_mailbox interface?

I didn't follow the patches too closely, but IIRC there were some issues
with integrating it into remoteproc/rpmsg? If you don't have them in
your inbox, check the archive around september.

> And then embedding the call to mbox in the dmaengine driver would
> make everything easy.
> Also it seems some other parameters from mbox are also needed by other
> drivers as well in the future, so we should get something working.
> (Clocks, enabling HW blocks, sharing memory between GPU and ARM,...
> come to my mind just looking at the API)
> As for working with the foundation on "upstream", has anyone been able to
> contact them about this subject? You might just want to point me to
> a thread instead instead of going to repeat the discussions?
> Thanks,
>         Martin
> P.s: I know that the Foundation and Broadcom can be a bit hard while they
> try avoid certain subjects... See my forum post about "official" Errata in the
> SOC (and I just found another one this morning) - not the user-maintained
> wiki-page on the elinux site.

More information about the linux-rpi-kernel mailing list