From: Amit Nischal <anischal@codeaurora.org>
To: Stephen Boyd <sboyd@kernel.org>
Cc: Michael Turquette <mturquette@baylibre.com>,
Andy Gross <andy.gross@linaro.org>,
David Brown <david.brown@linaro.org>,
Rajendra Nayak <rnayak@codeaurora.org>,
Odelu Kukatla <okukatla@codeaurora.org>,
Taniya Das <tdas@codeaurora.org>,
linux-arm-msm@vger.kernel.org, linux-soc@vger.kernel.org,
linux-clk@vger.kernel.org, linux-kernel@vger.kernel.org,
linux-clk-owner@vger.kernel.org
Subject: Re: [PATCH 1/4] clk: qcom: gdsc: Add support to enable/disable the clocks with GDSC
Date: Mon, 30 Jul 2018 16:44:12 +0530 [thread overview]
Message-ID: <e881fd1f970d32d96a9780e8b6812e86@codeaurora.org> (raw)
In-Reply-To: <153250153971.48062.18053635476867512394@swboyd.mtv.corp.google.com>
On 2018-07-25 12:22, Stephen Boyd wrote:
> Quoting Amit Nischal (2018-07-12 05:23:48)
>> On 2018-07-09 11:04, Stephen Boyd wrote:
>> > Quoting Amit Nischal (2018-06-06 04:41:45)
>> >> For some of the GDSCs, there is a requirement to enable/disable the
>> >> few clocks before turning on/off the gdsc power domain. Add support
>> >
>> > Why is there a requirement? Do the clks need to be in hw control mode
>> > or
>> > they can't be turned off when the GDSC is off? It's hard for me to
>> > understand with these vague statements.
>> >
>>
>> This requirement is primarily to turn on the GPU GX GDSC and these
>> clocks
>> do not need to be in HW control mode and clock disable is not related
>> with the GDSC.
>
> Ok that's good to know.
>
>>
>> To turn on the GX GDSC, root clock (GFX3D RCG) needs to be enabled
>> first
>> before writing to SW_COLLAPSE bit of the GDSC. The CLK_ON signal from
>> RCG
>> would power up the GPU GX GDSC.
>
> Can you please put this specific information in the commit text instead
> of making a vague statement about GDSC hardware configurations?
>
> Does anything go wrong if the GDSC is enabled from genpd but doesn't
> actually turn on until the GFX3D RCG root bit is enabled or disabled by
> the clk enable call? I suppose we won't know if the GDSC is enabled or
> not until the clk is enabled? Maybe we should make the clk enable of
> the
> RCG for GPU go check the GDSC status bit as well to make sure it's
> toggling on or off?
As per the current design, before turning on the GDSC(writing to the
GDSCR
register) we are setting the ROOT_EN bit of RCG and GDSC's status will
be
still off without clk_enable call to RCG even though we enable the GDSC.
clk_enable for RCG is CLK_ON signal for the GPU_GX_GDSC.
>
> Also, does the RCG turn on when the GX GDSC is off? I think we may be
> able to rely on the GPU driver to "do the right thing" and enable the
> GPU CX GDSC first, then the RCG and branch for the GFX3D clk, and then
> the GPU GX GDSC for the core GPU logic. Then we don't need to do
> anything special in the GDSC code for this.
Its a GPU_GX_GDSC requirement to enable the RCG first and then GX_GDSC.
We want all of this sequence to be done by the GDSC driver so that
client
only call for clk_apis for the clock branch. For clients, It will be
extra overhead to follow the below sequence.
GPU_CX_GDSC enable -> Enable RCG -> Enable GPU_GX_GDSC -> Enable Branch.
next prev parent reply other threads:[~2018-07-30 11:14 UTC|newest]
Thread overview: 23+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-06-06 11:41 [PATCH 0/4] Add QCOM graphics clock controller driver for SDM845 Amit Nischal
2018-06-06 11:41 ` [PATCH 1/4] clk: qcom: gdsc: Add support to enable/disable the clocks with GDSC Amit Nischal
2018-07-09 5:34 ` Stephen Boyd
2018-07-12 12:23 ` Amit Nischal
2018-07-25 6:52 ` Stephen Boyd
2018-07-30 11:14 ` Amit Nischal [this message]
2018-06-06 11:41 ` [PATCH 2/4] clk: qcom: Add clk_rcg2_gfx3d_ops for SDM845 Amit Nischal
2018-07-09 6:15 ` Stephen Boyd
2018-07-12 12:30 ` Amit Nischal
2018-07-25 6:58 ` Stephen Boyd
2018-07-30 11:28 ` Amit Nischal
2018-08-02 22:44 ` Stephen Boyd
2018-08-06 9:07 ` Amit Nischal
2018-08-06 15:04 ` Jordan Crouse
2018-08-08 5:58 ` Stephen Boyd
2018-08-08 14:51 ` Jordan Crouse
2018-08-13 6:30 ` Amit Nischal
2018-06-06 11:41 ` [PATCH 3/4] dt-bindings: clock: Introduce QCOM Graphics clock bindings Amit Nischal
2018-07-09 5:38 ` Stephen Boyd
2018-07-12 12:32 ` Amit Nischal
2018-06-06 11:41 ` [PATCH 4/4] clk: qcom: Add graphics clock controller driver for SDM845 Amit Nischal
2018-07-09 6:07 ` Stephen Boyd
2018-07-12 12:38 ` Amit Nischal
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=e881fd1f970d32d96a9780e8b6812e86@codeaurora.org \
--to=anischal@codeaurora.org \
--cc=andy.gross@linaro.org \
--cc=david.brown@linaro.org \
--cc=linux-arm-msm@vger.kernel.org \
--cc=linux-clk-owner@vger.kernel.org \
--cc=linux-clk@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-soc@vger.kernel.org \
--cc=mturquette@baylibre.com \
--cc=okukatla@codeaurora.org \
--cc=rnayak@codeaurora.org \
--cc=sboyd@kernel.org \
--cc=tdas@codeaurora.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox