Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.34 as permitted sender) client-ip=23.128.96.34;
Date:   Tue, 12 Sep 2023 08:44:15 +0200
From:   Janne Grunau <j@jannau.net>
To:     Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Cc:     Thomas Zimmermann <tzimmermann@suse.de>,
        Javier Martinez Canillas <javierm@redhat.com>,
        David Airlie <airlied@gmail.com>,
        Daniel Vetter <daniel@ffwll.ch>,
        dri-devel@lists.freedesktop.org, linux-kernel@vger.kernel.org,
        asahi@lists.linux.dev
Subject: Re: [PATCH] drm/simpledrm: Add support for multiple "power-domains"
Message-ID: <ZQAIv3KchAMsXgIr@jannau.net>
References: <20230910-simpledrm-multiple-power-domains-v1-1-f8718aefc685@jannau.net>
 <0b6ad379-049a-5152-12e5-77c62ffc4541@wanadoo.fr>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <0b6ad379-049a-5152-12e5-77c62ffc4541@wanadoo.fr>
Precedence: bulk

On 2023-09-10 22:16:05 +0200, Christophe JAILLET wrote:
> Le 10/09/2023 ? 18:39, Janne Grunau via B4 Relay a ?crit?:
> > From: Janne Grunau <j@jannau.net>
> > 
> > Multiple power domains need to be handled explicitly in each driver. The
> > driver core can not handle it automatically since it is not aware of
> > power sequencing requirements the hardware might have. This is not a
> > problem for simpledrm since everything is expected to be powered on by
> > the bootloader. simpledrm has just ensure it remains powered on during
> > its lifetime.
> > This is required on Apple silicon M2 and M2 Pro/Max/Ultra desktop
> > systems. The HDMI output initialized by the bootloader requires keeping
> > the display controller and a DP phy power domain on.
> > 
> > Signed-off-by: Janne Grunau <j@jannau.net>
> > ---
> >   drivers/gpu/drm/tiny/simpledrm.c | 106 +++++++++++++++++++++++++++++++++++++++
> >   1 file changed, 106 insertions(+)
> > 
> > diff --git a/drivers/gpu/drm/tiny/simpledrm.c b/drivers/gpu/drm/tiny/simpledrm.c
> > index ff86ba1ae1b8..efedede57d42 100644
> > --- a/drivers/gpu/drm/tiny/simpledrm.c
> > +++ b/drivers/gpu/drm/tiny/simpledrm.c
> > @@ -6,6 +6,7 @@
> >   #include <linux/of_address.h>
> >   #include <linux/platform_data/simplefb.h>
> >   #include <linux/platform_device.h>
> > +#include <linux/pm_domain.h>
> >   #include <linux/regulator/consumer.h>
> >   #include <drm/drm_aperture.h>
> > @@ -227,6 +228,12 @@ struct simpledrm_device {
> >   	unsigned int regulator_count;
> >   	struct regulator **regulators;
> >   #endif
> > +	/* power-domains */
> > +#if defined CONFIG_OF && defined CONFIG_PM_GENERIC_DOMAINS
> > +	int pwr_dom_count;
> > +	struct device **pwr_dom_devs;
> > +	struct device_link **pwr_dom_links;
> > +#endif
> >   	/* simplefb settings */
> >   	struct drm_display_mode mode;
> > @@ -468,6 +475,102 @@ static int simpledrm_device_init_regulators(struct simpledrm_device *sdev)
> >   }
> >   #endif
> > +#if defined CONFIG_OF && defined CONFIG_PM_GENERIC_DOMAINS
> > +/*
> > + * Generic power domain handling code.
> > + *
> > + * Here we handle the power-domains properties of our "simple-framebuffer"
> > + * dt node. This is only necessary if there is more than one power-domain.
> > + * A single power-domains is handled automatically by the driver core. Multiple
> > + * power-domains have to be handled by drivers since the driver core can't know
> > + * the correct power sequencing. Power sequencing is not an issue for simpledrm
> > + * since the bootloader has put the power domains already in the correct state.
> > + * simpledrm has only to ensure they remain active for its lifetime.
> > + *
> > + * When the driver unloads, we detach from the power-domains.
> > + *
> > + * We only complain about errors here, no action is taken as the most likely
> > + * error can only happen due to a mismatch between the bootloader which set
> > + * up the "simple-framebuffer" dt node, and the PM domain providers in the
> > + * device tree. Chances are that there are no adverse effects, and if there are,
> > + * a clean teardown of the fb probe will not help us much either. So just
> > + * complain and carry on, and hope that the user actually gets a working fb at
> > + * the end of things.
> > + */
> > +static void simpledrm_device_detach_genpd(void *res)
> > +{
> > +	int i;
> > +	struct simpledrm_device *sdev = /*(struct simpledrm_device *)*/res;
> > +
> > +
> > +	drm_err(&sdev->dev, "% power-domains count:%d\n", __func__, sdev->pwr_dom_count);
> > +	if (sdev->pwr_dom_count <= 1)
> > +		return;
> > +
> > +	for (i = sdev->pwr_dom_count - 1; i >= 0; i--) {
> > +		if (!sdev->pwr_dom_links[i])
> > +			device_link_del(sdev->pwr_dom_links[i]);
> > +		if (!IS_ERR_OR_NULL(sdev->pwr_dom_devs[i]))
> > +			dev_pm_domain_detach(sdev->pwr_dom_devs[i], true);
> > +	}
> > +}
> > +
> > +static int simpledrm_device_attach_genpd(struct simpledrm_device *sdev)
> > +{
> > +	struct device *dev = sdev->dev.dev;
> > +	int i;
> > +
> > +	sdev->pwr_dom_count = of_count_phandle_with_args(dev->of_node, "power-domains",
> > +							 "#power-domain-cells");
> > +	/*
> > +	 * Single power-domain devices are handled by driver core nothing to do
> > +	 * here. The same for device nodes without "power-domains" property.
> > +	 */
> > +	if (sdev->pwr_dom_count <= 1)
> > +		return 0;
> > +
> > +	sdev->pwr_dom_devs = devm_kcalloc(dev, sdev->pwr_dom_count,
> > +					       sizeof(*sdev->pwr_dom_devs),
> > +					       GFP_KERNEL);
> > +	if (!sdev->pwr_dom_devs)
> > +		return -ENOMEM;
> > +
> > +	sdev->pwr_dom_links = devm_kcalloc(dev, sdev->pwr_dom_count,
> > +						sizeof(*sdev->pwr_dom_links),
> > +						GFP_KERNEL);
> > +	if (!sdev->pwr_dom_links)
> > +		return -ENOMEM;
> > +
> > +	for (i = 0; i < sdev->pwr_dom_count; i++) {
> > +		sdev->pwr_dom_devs[i] = dev_pm_domain_attach_by_id(dev, i);
> > +		if (IS_ERR(sdev->pwr_dom_devs[i])) {
> > +			int ret = PTR_ERR(sdev->pwr_dom_devs[i]);
> > +			if (ret == -EPROBE_DEFER) {
> > +				simpledrm_device_detach_genpd(sdev);
> > +				return PTR_ERR(sdev->pwr_dom_devs[i]);
> > +			}
> > +			drm_err(&sdev->dev,
> > +				"pm_domain_attach_by_id(%u) failed: %d\n", i, ret);
> > +		}
> > +
> 
> sdev->pwr_dom_devs[i] can be an ERR_PTR here.
> Maybe a break or a continue missing after drm_err() above?

yes, a continue is missing, added, locally for v2.

thanks

Janne