[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20150108114444.GM1987@ulmo.nvidia.com>
Date: Thu, 8 Jan 2015 12:44:46 +0100
From: Thierry Reding <thierry.reding@...il.com>
To: Peter De Schrijver <pdeschrijver@...dia.com>
Cc: Vince Hsu <vinceh@...dia.com>, Lucas Stach <dev@...xeye.de>,
swarren@...dotorg.org, gnurou@...il.com, bskeggs@...hat.com,
martin.peres@...e.fr, seven@...rod-online.com,
samuel.pitoiset@...il.com, nouveau@...ts.freedesktop.org,
linux-tegra@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 1/11] ARM: tegra: add function to control the GPU rail
clamp
On Thu, Jan 08, 2015 at 11:39:57AM +0200, Peter De Schrijver wrote:
> > > And specify the dependencies between domains in DT?
> >
> > I think the dependencies could be in the driver. Of course the power
> > domains are per-SoC data, so really shouldn't be in the DTS either (the
> > data is all implied by the compatible value) but there's no good way to
>
> The clock references could also be retrieved via clk_get_sys(). We could add
> some more clkdev entries. If we use the domain name as the dev_id and the
> module names as the con_id's, the domain code could then retrieve the
> clocks by iterating over the module names and performing a
> clk_get_sys(domain_name, module_name) for each module. Unfortunately no such
> mechanism exists for resets.
I don't think having both clock and reset references in the device tree
is all that bad. We could possibly add a lookup mechanism for reset
controls that doesn't rely on DT, but I'm not sure it's really worth it.
Thierry
Content of type "application/pgp-signature" skipped
Powered by blists - more mailing lists