Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <59a88e00-0758-32b8-7ce4-8dea84a9b0f7@baylibre.com>
Date:   Tue, 18 Jul 2023 10:37:13 +0200
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.13.0
Subject: Re: [PATCH 2/2] clk: mediatek: mt8195-topckgen: Refactor parents for
 top_dp/edp muxes
To:     AngeloGioacchino Del Regno 
        <angelogioacchino.delregno@collabora.com>,
        Maxime Ripard <mripard@kernel.org>
Cc:     Chen-Yu Tsai <wenst@chromium.org>, sboyd@kernel.org,
        mturquette@baylibre.com, matthias.bgg@gmail.com, msp@baylibre.com,
        yangyingliang@huawei.com, u.kleine-koenig@pengutronix.de,
        miles.chen@mediatek.com, linux-clk@vger.kernel.org,
        linux-kernel@vger.kernel.org, linux-arm-kernel@lists.infradead.org,
        linux-mediatek@lists.infradead.org, kernel@collabora.com
References: <20230713072138.84117-1-angelogioacchino.delregno@collabora.com>
 <20230713072138.84117-3-angelogioacchino.delregno@collabora.com>
 <9a0817c2-4101-5c21-977d-77ac0d83a067@baylibre.com>
 <CAGXv+5E7YYdkG7GtxG90KzdAG8Kke+74Amtbw4mmyVNZgDZHRA@mail.gmail.com>
 <jv6daj2w3pwjtde3m3m26yg4wyxbbio4zqra5yqc4gb32ri5ub@noinbbkjovwm>
 <25724ee3-858a-01eb-352b-3edbfad31c8e@collabora.com>
 <jxgy2pvns4ri2aj5nmdhb4zbluseuzdejbplh2avwz63df2cfx@grrrdm6ujzi4>
 <24d17b07-1e8d-05f6-46b7-9da1ff1bed7a@collabora.com>
Content-Language: en-US
From:   Alexandre Mergnat <amergnat@baylibre.com>
In-Reply-To: <24d17b07-1e8d-05f6-46b7-9da1ff1bed7a@collabora.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk


On 17/07/2023 16:30, AngeloGioacchino Del Regno wrote:
>>>>> However I'm not sure if that works for parents. It should, given the
>>>>> original use case was for the sunxi platforms, which like the MediaTek
>>>>> platform here has 2 PLLs for video related consumers, but I couldn't
>>>>> find code verifying it.
>>>>
>>>> If you want to prevent clocks from ever being reparented, you can use
>>>> the new clk_hw_determine_rate_no_reparent() determine_rate
>>>> implementation.
>>>>
>>>
>>> We want the clocks to be reparented, as we need them to switch 
>>> parents as
>>> explained before... that's more or less how the tree looks:
>>>
>>> TVDPLL(x) -> PLL Divider (fixed) -> MUX -> Gate -> Controller
>>>
>>> Besides, I think that forcing *one* parent to the dp/edp mux would 
>>> produce a
>>> loss of the flexibility that the clock framework provides.
>>>
>>> I again want to emphasize on the fact that TVDPLL1 and TVDPLL2 are 
>>> *identical*
>>> in specs, and on that there will never be a MT8195 SoC that has only 
>>> one of
>>> the two PLLs, for obvious reasons...
>>>
>>> P.S.: If you need more context, I'll be glad to answer to any other 
>>> question!
>>
>> Then I have no idea what the question is :)
>>
>> What are you trying to achieve / fix, and how can I help you ? :)
>>
> 
> Chen-Yu, Alexandre had/have questions about if there was any other 
> solution instead
> of using the solution of *this* commit, so, if there's any other better 
> solution
> than the one that I've sent as this commit.
> 
> I'm the one saying that this commit is the best solution :-P

Hi Angelo,

My solution is based on PLL static allocation, because I missed it could 
be reparented actually. I think I've a better understanding of this 
commit now, thanks to your explanations. Looks fine for me.

Reviewed-by: Alexandre Mergnat <amergnat@baylibre.com>

-- 
Regards,
Alexandre