Message-ID: <d6b158ff-382c-c659-04c8-930b2681e8d6@redhat.com>
Date:   Mon, 13 Feb 2023 13:59:06 +0100
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.6.0
Subject: Re: [PATCH 2/4] drm/vc4: hdmi: Enable power domain before setting
 minimum
Content-Language: en-US
To:     Maxime Ripard <maxime@cerno.tech>, Emma Anholt <emma@anholt.net>,
        Maxime Ripard <mripard@kernel.org>,
        David Airlie <airlied@gmail.com>,
        Daniel Vetter <daniel@ffwll.ch>
Cc:     dri-devel@lists.freedesktop.org,
        Thomas Zimmermann <tzimmermann@suse.de>,
        linux-kernel@vger.kernel.org
References: <20230126-rpi-display-fw-clk-cleanup-v1-0-d646ff6fb842@cerno.tech>
 <20230126-rpi-display-fw-clk-cleanup-v1-2-d646ff6fb842@cerno.tech>
From:   Javier Martinez Canillas <javierm@redhat.com>
In-Reply-To: <20230126-rpi-display-fw-clk-cleanup-v1-2-d646ff6fb842@cerno.tech>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 1/26/23 18:05, Maxime Ripard wrote:
> On the RaspberryPi0-3, the HSM clock was provided by the clk-bcm2835
> driver, but on the Pi4 it was provided by the firmware through the
> clk-raspberrypi driver.
> 
> The clk-bcm2835 driver registers the HSM clock using the
> CLK_SET_RATE_GATE flag that prevents any modification to the rate while
> the clock is active.
> 
> This meant that we needed to call clk_set_min_rate() before our call to
> pm_runtime_resume_and_get() since our runtime_resume implementation
> needs to enable the HSM clock for the HDMI controller registers to be
> functional.
> 
> However, the HSM clock is part of the HDMI power domain which might not
> be powered prior to the pm_runtime_resume_and_get() call, so we could
> end up changing the rate of the HSM clock while its power domain was
> disabled.
> 
> We recently changed the backing driver for the RaspberryPi0-3 to
> clk-raspberrypi though, which doesn't have such restrictions. We can
> thus move the clk_set_min_rate() after our call to runtime_resume and
> avoid the access while the power domain is disabled.
> 
> Signed-off-by: Maxime Ripard <maxime@cerno.tech>
> ---

I'm not familiar with the RPi clock hierarchy but the commit message explains
why a clk_set_min_rate() was needed before the pm_runtime_resume_and_get(),
and why that isn't needed anymore after the switch to clk-raspberrypi driver.

And certainly, the correct thing to do is to enable the power domain that a
controller is part of, before attempting to change the rate for one of its
clocks. So the patch looks good to me.

Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>

-- 
Best regards,

Javier Martinez Canillas
Core Platforms
Red Hat