Re: [GIT_PULL] SOC: TI Keystone Ring Accelerator driver for v5.6

From: Olof Johansson
Date: Fri Jan 17 2020 - 13:11:57 EST


On Thu, Jan 16, 2020 at 9:05 PM <santosh.shilimkar@xxxxxxxxxx> wrote:
>
> On 1/16/20 4:03 PM, Olof Johansson wrote:
> > Hi,
> >
> > On Thu, Jan 16, 2020 at 12:07:39PM -0800, Santosh Shilimkar wrote:
> >> Its bit late for pull request, but if possible, please pull it to
> >> soc drivers tree.
> >>
> >> The following changes since commit e42617b825f8073569da76dc4510bfa019b1c35a:
> >>
> >> Linux 5.5-rc1 (2019-12-08 14:57:55 -0800)
> >>
> >> are available in the git repository at:
> >>
> >> git://git.kernel.org/pub/scm/linux/kernel/git/ssantosh/linux-keystone.git tags/drivers_soc_for_5.6
> >>
> >> for you to fetch changes up to 3277e8aa2504d97e022ecb9777d784ac1a439d36:
> >>
> >> soc: ti: k3: add navss ringacc driver (2020-01-15 10:07:27 -0800)
> >>
> >> ----------------------------------------------------------------
> >> SOC: TI Keystone Ring Accelerator driver
> >>
> >> The Ring Accelerator (RINGACC or RA) provides hardware acceleration to
> >> enable straightforward passing of work between a producer and a consumer.
> >> There is one RINGACC module per NAVSS on TI AM65x SoCs.
> >
> > This driver doesn't seem to have exported symbols, and no in-kernel
> > users. So how will it be used?
> >
> > Usually we ask to hold off until the consuming side/drivers are also ready.
> >
> The other patches getting merged via Vinod's tree. The combined series
> is split into couple of series. Vinod is going to pull this branch
> and apply rest of the patchset. And then couple of additional consumer
> drivers will get posted.

Ok -- might have been useful to get that in the tag description for
context. Something to consider next time.

> > Also, is there a reason this is under drivers/soc/ instead of somewhere more
> > suitable in the drivers subsystem? It's not "soc glue code" in the same way as
> > drivers/soc was intended originally.
> >
> These kind of SOC IP drivers, we put into drivers/soc/ because of lack
> of specific subsystem where they fit in. Navigator was also similar example.

Hmm. At some point we'll have to push the brakes on this, since
drivers/soc can't become a catch-all for random stuff like the old
mach directories were. But it's tricky to tell just when -- sometimes
you have to let the mess show up too.

I'll merge this when I do the next pass (today, likely).

-Olof