From: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
To: "Rafael J. Wysocki" <rafael@kernel.org>,
Pavel Machek <pavel@kernel.org>, Len Brown <lenb@kernel.org>,
Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
Danilo Krummrich <dakr@kernel.org>, Ulf Hansson <ulfh@kernel.org>
Cc: Devarsh Thakkar <devarsht@ti.com>,
linux-pm@vger.kernel.org, driver-core@lists.linux.dev,
linux-kernel@vger.kernel.org
Subject: Re: [PATCH] pmdomain: core: Add genpd_dev_pm_attach_by_node()
Date: Tue, 10 Nov 2026 15:58:37 +0200 [thread overview]
Message-ID: <af65c273-70c5-4926-96e7-ab1089503286@ideasonboard.com> (raw)
In-Reply-To: <20261023-pm_attach_by_node-v1-1-4dbaccdf8b0e@ideasonboard.com>
Hi,
On 23/10/2026 21:16, Tomi Valkeinen wrote:
> genpd_dev_pm_attach_by_id() and genpd_dev_pm_attach_by_name() attach a
> virtual device to a PM domain listed in the OF node of the passed
> device.
>
> Some hardware blocks are described in a child node of a main device.
> They have no device of their own, but they do have a PM domain of their
> own. Neither function can attach that domain, as both parse only the OF
> node of the device.
>
> An example case is TI's display subsystem and its OLDI (LVDS)
> transmitters:
>
> dss@30200000 {
> power-domains = <...>;
>
> oldi-transmitters {
> oldi@0 {
> power-domains = <...>;
> };
>
> oldi@1 {
> power-domains = <...>;
> };
> };
> };
>
> The OLDI TX blocks are not part of the DSS IP, but they are not quite
> real devices either as they don't have registers of their own, and are
> instead controlled via DSS registers.
>
> Add genpd_dev_pm_attach_by_node(), which is similar to
> genpd_dev_pm_attach_by_id() but instead of looking for the power-domains
> from the passed device's DT node, the DT node is passed explicitly. Also
> add a matching wrapper function dev_pm_domain_attach_by_node().
Ping, any thoughts on this?
Tomi
> Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
> ---
> This is a new try for managing the OLDI power domains. Previously I
> tried using auxiliary device/driver for OLDIs, and handling the PDs the
> normal way:
>
> https://lore.kernel.org/all/20260529-beagley-ai-display-v3-12-7fefdc5d1adf%40ideasonboard.com/
>
> Then I realized an issue, mentioned in
> Documentation/driver-api/driver-model/driver.rst: "-EPROBE_DEFER must
> not be returned if probe() has already created child devices".
>
> The PD virtual devices used in this patch's approach do not hit the
> issue as they don't get a driver, so they are not probed. The tidss
> driver's OLDI part can then use normal pm_runtime_get/put functions to
> manage the OLDI PDs.
> ---
> drivers/base/power/common.c | 27 +++++++++++++
> drivers/pmdomain/core.c | 97 +++++++++++++++++++++++++++++++++------------
> include/linux/pm_domain.h | 19 +++++++++
> 3 files changed, 117 insertions(+), 26 deletions(-)
>
> diff --git a/drivers/base/power/common.c b/drivers/base/power/common.c
> index 9bef9248a705..a873541f9350 100644
> --- a/drivers/base/power/common.c
> +++ b/drivers/base/power/common.c
> @@ -171,6 +171,33 @@ struct device *dev_pm_domain_attach_by_name(struct device *dev,
> }
> EXPORT_SYMBOL_GPL(dev_pm_domain_attach_by_name);
>
> +/**
> + * dev_pm_domain_attach_by_node - Associate a device with a PM domain of a DT
> + * node.
> + * @dev: The device that uses the PM domain.
> + * @np: The DT node with the PM domain.
> + * @index: The index of the PM domain in @np.
> + *
> + * Like dev_pm_domain_attach_by_id(), but the PM domain is looked up
> + * in @np instead of the OF node of @dev.
> + *
> + * Unlike dev_pm_domain_attach_by_id(), @dev may already be attached to a PM
> + * domain of its own. The driver controls the power to the PM domain through
> + * runtime PM of the returned virtual device, or through a device link to it.
> + *
> + * Returns the virtual created device when successfully attached to its PM
> + * domain, NULL in case @np has no PM domain at @index, else an ERR_PTR().
> + * Note that, to detach the returned virtual device, the driver shall call
> + * dev_pm_domain_detach() on it, typically during the remove phase.
> + */
> +struct device *dev_pm_domain_attach_by_node(struct device *dev,
> + struct device_node *np,
> + unsigned int index)
> +{
> + return genpd_dev_pm_attach_by_node(dev, np, index);
> +}
> +EXPORT_SYMBOL_GPL(dev_pm_domain_attach_by_node);
> +
> /**
> * dev_pm_domain_attach_list - Associate a device with its PM domains.
> * @dev: The device used to lookup the PM domains for.
> diff --git a/drivers/pmdomain/core.c b/drivers/pmdomain/core.c
> index 052ec478dfcc..327fbd40c854 100644
> --- a/drivers/pmdomain/core.c
> +++ b/drivers/pmdomain/core.c
> @@ -3462,34 +3462,16 @@ int genpd_dev_pm_attach(struct device *dev)
> }
> EXPORT_SYMBOL_GPL(genpd_dev_pm_attach);
>
> -/**
> - * genpd_dev_pm_attach_by_id - Associate a device with one of its PM domains.
> - * @dev: The device used to lookup the PM domain.
> - * @index: The index of the PM domain.
> - *
> - * Parse device's OF node to find a PM domain specifier at the provided @index.
> - * If such is found, creates a virtual device and attaches it to the retrieved
> - * pm_domain ops. To deal with detaching of the virtual device, the ->detach()
> - * callback in the struct dev_pm_domain are assigned to genpd_dev_pm_detach().
> - *
> - * Returns the created virtual device if successfully attached PM domain, NULL
> - * when the device don't need a PM domain, else an ERR_PTR() in case of
> - * failures. If a power-domain exists for the device, but cannot be found or
> - * turned on, then ERR_PTR(-EPROBE_DEFER) is returned to ensure that the device
> - * is not probed and to re-try again later.
> - */
> -struct device *genpd_dev_pm_attach_by_id(struct device *dev,
> - unsigned int index)
> +static struct device *genpd_dev_pm_attach_virt(struct device *dev,
> + struct device_node *np,
> + unsigned int index)
> {
> struct device *virt_dev;
> int num_domains;
> int ret;
>
> - if (!dev->of_node)
> - return NULL;
> -
> /* Verify that the index is within a valid range. */
> - num_domains = of_count_phandle_with_args(dev->of_node, "power-domains",
> + num_domains = of_count_phandle_with_args(np, "power-domains",
> "#power-domain-cells");
> if (num_domains < 0 || index >= num_domains)
> return NULL;
> @@ -3502,10 +3484,14 @@ struct device *genpd_dev_pm_attach_by_id(struct device *dev,
> if (!virt_dev)
> return ERR_PTR(-ENOMEM);
>
> - dev_set_name(virt_dev, "genpd:%u:%s", index, dev_name(dev));
> + if (np == dev->of_node)
> + dev_set_name(virt_dev, "genpd:%u:%s", index, dev_name(dev));
> + else
> + dev_set_name(virt_dev, "genpd:%u:%s:%pOFP", index,
> + dev_name(dev), np);
> virt_dev->bus = &genpd_bus_type;
> virt_dev->release = genpd_release_dev;
> - virt_dev->of_node = of_node_get(dev->of_node);
> + virt_dev->of_node = of_node_get(np);
>
> ret = device_register(virt_dev);
> if (ret) {
> @@ -3513,8 +3499,13 @@ struct device *genpd_dev_pm_attach_by_id(struct device *dev,
> return ERR_PTR(ret);
> }
>
> - /* Try to attach the device to the PM domain at the specified index. */
> - ret = __genpd_dev_pm_attach(virt_dev, dev, index, num_domains, false);
> + /*
> + * Try to attach the device to the PM domain at the specified index.
> + * Pass 0 as the num_domains when np != dev->of_node, so that the dev's
> + * required OPPs are not used for the PD.
> + */
> + ret = __genpd_dev_pm_attach(virt_dev, dev, index,
> + np == dev->of_node ? num_domains : 0, false);
> if (ret < 1) {
> device_unregister(virt_dev);
> return ret ? ERR_PTR(ret) : NULL;
> @@ -3525,8 +3516,62 @@ struct device *genpd_dev_pm_attach_by_id(struct device *dev,
>
> return virt_dev;
> }
> +
> +/**
> + * genpd_dev_pm_attach_by_id - Associate a device with one of its PM domains.
> + * @dev: The device used to lookup the PM domain.
> + * @index: The index of the PM domain.
> + *
> + * Parse device's OF node to find a PM domain specifier at the provided @index.
> + * If such is found, creates a virtual device and attaches it to the retrieved
> + * pm_domain ops. To deal with detaching of the virtual device, the ->detach()
> + * callback in the struct dev_pm_domain are assigned to genpd_dev_pm_detach().
> + *
> + * Returns the created virtual device if successfully attached PM domain, NULL
> + * when the device don't need a PM domain, else an ERR_PTR() in case of
> + * failures. If a power-domain exists for the device, but cannot be found or
> + * turned on, then ERR_PTR(-EPROBE_DEFER) is returned to ensure that the device
> + * is not probed and to re-try again later.
> + */
> +struct device *genpd_dev_pm_attach_by_id(struct device *dev,
> + unsigned int index)
> +{
> + if (!dev->of_node)
> + return NULL;
> +
> + return genpd_dev_pm_attach_virt(dev, dev->of_node, index);
> +}
> EXPORT_SYMBOL_GPL(genpd_dev_pm_attach_by_id);
>
> +/**
> + * genpd_dev_pm_attach_by_node - Associate a device with a PM domain of a DT
> + * node.
> + * @dev: The device that uses the PM domain.
> + * @np: The DT node with the PM domain.
> + * @index: The index of the PM domain in @np.
> + *
> + * Like genpd_dev_pm_attach_by_id(), but the PM domain is looked up in
> + * @np instead of the OF node of @dev.
> + *
> + * The virtual device is named after @dev and @np. The required OPPs of @dev are
> + * not linked to the PM domain.
> + *
> + * Returns the created virtual device if successfully attached PM domain, NULL
> + * when @np has no PM domain at @index, else an ERR_PTR() in case of failures.
> + * If the PM domain exists, but cannot be found or turned on, then
> + * ERR_PTR(-EPROBE_DEFER) is returned.
> + */
> +struct device *genpd_dev_pm_attach_by_node(struct device *dev,
> + struct device_node *np,
> + unsigned int index)
> +{
> + if (!np)
> + return NULL;
> +
> + return genpd_dev_pm_attach_virt(dev, np, index);
> +}
> +EXPORT_SYMBOL_GPL(genpd_dev_pm_attach_by_node);
> +
> /**
> * genpd_dev_pm_attach_by_name - Associate a device with one of its PM domains.
> * @dev: The device used to lookup the PM domain.
> diff --git a/include/linux/pm_domain.h b/include/linux/pm_domain.h
> index f925614aebdb..c38c90d63c02 100644
> --- a/include/linux/pm_domain.h
> +++ b/include/linux/pm_domain.h
> @@ -477,6 +477,9 @@ struct device *genpd_dev_pm_attach_by_id(struct device *dev,
> unsigned int index);
> struct device *genpd_dev_pm_attach_by_name(struct device *dev,
> const char *name);
> +struct device *genpd_dev_pm_attach_by_node(struct device *dev,
> + struct device_node *np,
> + unsigned int index);
> #else /* !CONFIG_PM_GENERIC_DOMAINS_OF */
> static inline int of_genpd_add_provider_simple(struct device_node *np,
> struct generic_pm_domain *genpd)
> @@ -535,6 +538,13 @@ static inline struct device *genpd_dev_pm_attach_by_name(struct device *dev,
> return NULL;
> }
>
> +static inline struct device *genpd_dev_pm_attach_by_node(struct device *dev,
> + struct device_node *np,
> + unsigned int index)
> +{
> + return NULL;
> +}
> +
> static inline
> struct generic_pm_domain *of_genpd_remove_last(struct device_node *np)
> {
> @@ -560,6 +570,9 @@ struct device *dev_pm_domain_attach_by_id(struct device *dev,
> unsigned int index);
> struct device *dev_pm_domain_attach_by_name(struct device *dev,
> const char *name);
> +struct device *dev_pm_domain_attach_by_node(struct device *dev,
> + struct device_node *np,
> + unsigned int index);
> int dev_pm_domain_attach_list(struct device *dev,
> const struct dev_pm_domain_attach_data *data,
> struct dev_pm_domain_list **list);
> @@ -586,6 +599,12 @@ static inline struct device *dev_pm_domain_attach_by_name(struct device *dev,
> {
> return NULL;
> }
> +static inline struct device *dev_pm_domain_attach_by_node(struct device *dev,
> + struct device_node *np,
> + unsigned int index)
> +{
> + return NULL;
> +}
> static inline int dev_pm_domain_attach_list(struct device *dev,
> const struct dev_pm_domain_attach_data *data,
> struct dev_pm_domain_list **list)
>
> ---
> base-commit: c03c1a14507dfaf8b856c2608f9ac216021fc222
> change-id: 20261023-pm_attach_by_node-13788662f3a1
>
> Best regards,
> --
> Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
>
next prev parent reply other threads:[~2026-11-10 13:58 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-10-23 18:16 Tomi Valkeinen
2026-11-10 13:58 ` Tomi Valkeinen [this message]
2026-11-16 17:08 ` Tomi Valkeinen
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=af65c273-70c5-4926-96e7-ab1089503286@ideasonboard.com \
--to=tomi.valkeinen@ideasonboard.com \
--cc=dakr@kernel.org \
--cc=devarsht@ti.com \
--cc=driver-core@lists.linux.dev \
--cc=gregkh@linuxfoundation.org \
--cc=lenb@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-pm@vger.kernel.org \
--cc=pavel@kernel.org \
--cc=rafael@kernel.org \
--cc=ulfh@kernel.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
all inboxes | Powered by JetHome®