| Message ID | 1389469372-17199-5-git-send-email-tomasz.figa@gmail.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
Tomasz Figa <tomasz.figa@gmail.com> writes: > This patch introduces generic code to perform power domain look-up using > device tree and automatically bind devices to their power domains. > Generic device tree binding is introduced to specify power domains of > devices in their device tree nodes. > > Backwards compatibility with legacy Samsung-specific power domain > bindings is provided, but for now the new code is not compiled when > CONFIG_ARCH_EXYNOS is selected to avoid collision with legacy code. This > will change as soon as Exynos power domain code gets converted to use > the generic framework in further patch. > > Signed-off-by: Tomasz Figa <tomasz.figa@gmail.com> I haven't read through this in detail yet, but wanted to make sure that the DT representation can handle nested power domains. At least SH-mobile has a hierarchy of power domains and the genpd code can handle that, so wanted to make sure that the DT representation can handle it as well. Thanks, Kevin
Hi Tomasz, thank you for posting this series. I would like to use the DT bindings for power domains in the bindings for C-states on ARM: http://comments.gmane.org/gmane.linux.power-management.general/41012 and in particular link a given C-state to a given power domain so that the kernel will have a way to actually check what devices are lost upon C-state entry (and for devices I also mean CPU peripheral like PMUs, GIC CPU IF, caches and possibly cpus, all of them already represented with DT nodes). I have a remark: - Can we group device nodes under a single power-domain-parent so that all devices defined under that parent won't have to re-define a power-domain property (a property like interrupt-parent, so to speak) What do you think ? Thanks, Lorenzo On Sat, Jan 11, 2014 at 07:42:46PM +0000, Tomasz Figa wrote: > This patch introduces generic code to perform power domain look-up using > device tree and automatically bind devices to their power domains. > Generic device tree binding is introduced to specify power domains of > devices in their device tree nodes. > > Backwards compatibility with legacy Samsung-specific power domain > bindings is provided, but for now the new code is not compiled when > CONFIG_ARCH_EXYNOS is selected to avoid collision with legacy code. This > will change as soon as Exynos power domain code gets converted to use > the generic framework in further patch. > > Signed-off-by: Tomasz Figa <tomasz.figa@gmail.com> > --- > .../devicetree/bindings/power/power_domain.txt | 51 ++++ > drivers/base/power/domain.c | 339 +++++++++++++++++++++ > include/linux/pm_domain.h | 34 +++ > kernel/power/Kconfig | 4 + > 4 files changed, 428 insertions(+) > create mode 100644 Documentation/devicetree/bindings/power/power_domain.txt > > diff --git a/Documentation/devicetree/bindings/power/power_domain.txt b/Documentation/devicetree/bindings/power/power_domain.txt > new file mode 100644 > index 0000000..93be5d9 > --- /dev/null > +++ b/Documentation/devicetree/bindings/power/power_domain.txt > @@ -0,0 +1,51 @@ > +* Generic power domains > + > +System on chip designs are often divided into multiple power domains that > +can be used for power gating of selected IP blocks for power saving by > +reduced leakage current. > + > +This device tree binding can be used to bind power domain consumer devices > +with their power domains provided by power domain providers. A power domain > +provider can be represented by any node in the device tree and can provide > +one or more power domains. A consumer node can refer to the provider by > +a phandle and a set of phandle arguments (so called power domain specifier) > +of length specified by #power-domain-cells property in the power domain > +provider node. > + > +==Power domain providers== > + > +Required properties: > + - #power-domain-cells : Number of cells in a power domain specifier; > + Typically 0 for nodes representing a single power domain and 1 for nodes > + providing multiple power domains (e.g. power controllers), but can be > + any value as specified by device tree binding documentation of particular > + provider. > + > +Example: > + > + power: power-controller@12340000 { > + compatible = "foo,power-controller"; > + reg = <0x12340000 0x1000>; > + #power-domain-cells = <1>; > + }; > + > +The node above defines a power controller that is a power domain provider > +and expects one cell as its phandle argument. > + > +==Power domain consumers== > + > +Required properties: > + - power-domain : A phandle and power domain specifier as defined by bindings > + of power controller specified by phandle. > + > +Example: > + > + leaky-device@12350000 { > + compatible = "foo,i-leak-current"; > + reg = <0x12350000 0x1000>; > + power-domain = <&power 0>; > + }; > + > +The node above defines a typical power domain consumer device, which is located > +inside power domain with index 0 of power controller represented by node with > +label "power". > diff --git a/drivers/base/power/domain.c b/drivers/base/power/domain.c > index bfb8955..6d47498 100644 > --- a/drivers/base/power/domain.c > +++ b/drivers/base/power/domain.c > @@ -3,12 +3,16 @@ > * > * Copyright (C) 2011 Rafael J. Wysocki <rjw@sisk.pl>, Renesas Electronics Corp. > * > + * Support for Device Tree based power domain providers: > + * Copyright (C) 2014 Tomasz Figa <tomasz.figa@gmail.com> > + * > * This file is released under the GPLv2. > */ > > #include <linux/init.h> > #include <linux/kernel.h> > #include <linux/io.h> > +#include <linux/platform_device.h> > #include <linux/pm_runtime.h> > #include <linux/pm_domain.h> > #include <linux/pm_qos.h> > @@ -2177,3 +2181,338 @@ void pm_genpd_init(struct generic_pm_domain *genpd, > list_add(&genpd->gpd_list_node, &gpd_list); > mutex_unlock(&gpd_list_lock); > } > + > +#ifdef CONFIG_PM_GENERIC_DOMAINS_OF > +/* > + * DEVICE TREE BASED POWER DOMAIN PROVIDERS > + * > + * The code below implements generic device tree based power domain providers > + * that bind device tree nodes with generic power domains registered in the > + * system. > + * > + * Any driver that registers generic power domains and need to support binding > + * of devices to these domains is supposed to register a power domain provider, > + * which maps a power domain specifier retrieved from device tree to a power > + * domain. > + * > + * Two simple mapping functions have been provided for convenience: > + * - of_genpd_xlate_simple() for 1:1 device tree node to domain mapping, > + * - of_genpd_xlate_onecell() for mapping of multiple domains per node > + * by index. > + */ > + > +/** > + * struct of_genpd_provider - Power domain provider registration structure > + * @link: Entry in global list of domain providers > + * @node: Pointer to device tree node of domain provider > + * @xlate: Provider-specific xlate callback mapping a set of specifier cells > + * into a power domain. > + * @data: context pointer to be passed into @xlate callback > + */ > +struct of_genpd_provider { > + struct list_head link; > + > + struct device_node *node; > + genpd_xlate_t xlate; > + void *data; > +}; > + > +/* List of registered power domain providers. */ > +static LIST_HEAD(of_genpd_providers); > +/* Mutex to protect the list above. */ > +static DEFINE_MUTEX(of_genpd_mutex); > + > +/** > + * of_genpd_lock() - Lock access to of_genpd_providers list > + */ > +static void of_genpd_lock(void) > +{ > + mutex_lock(&of_genpd_mutex); > +} > + > +/** > + * of_genpd_unlock() - Unlock access to of_genpd_providers list > + */ > +static void of_genpd_unlock(void) > +{ > + mutex_unlock(&of_genpd_mutex); > +} > + > +/** > + * of_genpd_xlate_simple() - Xlate function for direct node-domain mapping > + * @genpdspec: OF phandle args to map into a power domain > + * @data: xlate function private data - pointer to struct generic_pm_domain > + * > + * This is a generic xlate function that can be used to model power domains > + * that have their own device tree nodes. The private data of xlate function > + * needs to be a valid pointer to struct generic_pm_domain. > + */ > +struct generic_pm_domain *of_genpd_xlate_simple( > + struct of_phandle_args *genpdspec, > + void *data) > +{ > + if (genpdspec->args_count != 0) > + return ERR_PTR(-EINVAL); > + return data; > +} > +EXPORT_SYMBOL_GPL(of_genpd_xlate_simple); > + > +/** > + * of_genpd_xlate_onecell() - Xlate function for providers using single index. > + * @genpdspec: OF phandle args to map into a power domain > + * @data: xlate function private data - pointer to struct genpd_onecell_data > + * > + * This is a generic xlate function that can be used to model simple power > + * domain controllers that have one device tree node and provide multiple > + * power domains. A single cell is used as an index to an array of power > + * domains specified in genpd_onecell_data struct when registering the > + * provider. > + */ > +struct generic_pm_domain *of_genpd_xlate_onecell( > + struct of_phandle_args *genpdspec, > + void *data) > +{ > + struct genpd_onecell_data *genpd_data = data; > + unsigned int idx = genpdspec->args[0]; > + > + if (genpdspec->args_count != 1) > + return ERR_PTR(-EINVAL); > + > + if (idx >= genpd_data->domain_num) { > + pr_err("%s: invalid domain index %d\n", __func__, idx); > + return ERR_PTR(-EINVAL); > + } > + > + return genpd_data->domains[idx]; > +} > +EXPORT_SYMBOL_GPL(of_genpd_xlate_onecell); > + > +/** > + * of_genpd_add_provider() - Register a domain provider for a node > + * @np: Device node pointer associated with domain provider > + * @genpd_src_get: callback for decoding domain > + * @data: context pointer for @genpd_src_get callback. > + */ > +int of_genpd_add_provider(struct device_node *np, genpd_xlate_t xlate, > + void *data) > +{ > + struct of_genpd_provider *cp; > + > + cp = kzalloc(sizeof(struct of_genpd_provider), GFP_KERNEL); > + if (!cp) > + return -ENOMEM; > + > + cp->node = of_node_get(np); > + cp->data = data; > + cp->xlate = xlate; > + > + of_genpd_lock(); > + list_add(&cp->link, &of_genpd_providers); > + of_genpd_unlock(); > + pr_debug("Added domain provider from %s\n", np->full_name); > + > + return 0; > +} > +EXPORT_SYMBOL_GPL(of_genpd_add_provider); > + > +/** > + * of_genpd_del_provider() - Remove a previously registered domain provider > + * @np: Device node pointer associated with domain provider > + */ > +void of_genpd_del_provider(struct device_node *np) > +{ > + struct of_genpd_provider *cp; > + > + of_genpd_lock(); > + list_for_each_entry(cp, &of_genpd_providers, link) { > + if (cp->node == np) { > + list_del(&cp->link); > + of_node_put(cp->node); > + kfree(cp); > + break; > + } > + } > + of_genpd_unlock(); > +} > +EXPORT_SYMBOL_GPL(of_genpd_del_provider); > + > +/* See of_genpd_get_from_provider(). */ > +static struct generic_pm_domain *__of_genpd_get_from_provider( > + struct of_phandle_args *genpdspec) > +{ > + struct of_genpd_provider *provider; > + struct generic_pm_domain *genpd = ERR_PTR(-ENOENT); > + > + /* Check if we have such a provider in our array */ > + list_for_each_entry(provider, &of_genpd_providers, link) { > + if (provider->node == genpdspec->np) > + genpd = provider->xlate(genpdspec, provider->data); > + if (!IS_ERR(genpd)) > + break; > + } > + > + return genpd; > +} > + > +/** > + * of_genpd_get_from_provider() - Look-up power domain > + * @genpdspec: OF phandle args to use for look-up > + * > + * Looks for domain provider under node specified by @genpdspec and if found > + * uses xlate function of the provider to map phandle args to a power domain. > + * > + * Returns a valid pointer to struct generic_pm_domain on success or ERR_PTR() > + * on failure. > + */ > +static struct generic_pm_domain *of_genpd_get_from_provider( > + struct of_phandle_args *genpdspec) > +{ > + struct generic_pm_domain *genpd; > + > + of_genpd_lock(); > + genpd = __of_genpd_get_from_provider(genpdspec); > + of_genpd_unlock(); > + > + return genpd; > +} > + > +/* > + * DEVICE<->DOMAIN BINDING USING DEVICE TREE LOOK-UP > + * > + * The code below registers a notifier for platform bus devices' > + * BUS_NOTIFY_BIND_DRIVER events and tries to attach devices to their power > + * domains by looking them up using Device Tree. > + * > + * Similarly in BUS_NOTIFY_UNBOUND_DRIVER the device is detached from its > + * domain, since it no longer supports runtime PM without any driver bound > + * to it. > + * > + * Both generic and legacy Samsung-specific DT bindings are supported to > + * keep backwards compatibility with existing DTBs. > + */ > + > +/** > + * of_genpd_add_to_domain - Bind device to its power domain using Device Tree. > + * @dev: Device to bind to its power domain. > + * > + * Tries to parse power domain specifier from device's OF node and if succeeds > + * attaches the device to retrieved power domain. > + * > + * Returns 0 on success or negative error code otherwise. > + */ > +static int of_genpd_add_to_domain(struct device *dev) > +{ > + struct of_phandle_args pd_args; > + struct generic_pm_domain *pd; > + int ret; > + > + ret = of_parse_phandle_with_args(dev->of_node, "power-domain", > + "#power-domain-cells", 0, &pd_args); > + if (ret < 0) { > + if (ret != ENOENT) > + return ret; > + > + /* > + * Try legacy Samsung-specific bindings > + * (for backwards compatibility of DT ABI) > + */ > + pd_args.args_count = 0; > + pd_args.np = of_parse_phandle(dev->of_node, > + "samsung,power-domain", 0); > + if (!pd_args.np) > + return 0; > + } > + > + pd = of_genpd_get_from_provider(&pd_args); > + if (IS_ERR(pd)) > + return PTR_ERR(pd); > + > + dev_dbg(dev, "adding to power domain %s\n", pd->name); > + > + while (1) { > + ret = pm_genpd_add_device(pd, dev); > + if (ret != -EAGAIN) > + break; > + cond_resched(); > + } > + > + if (!ret) > + pm_genpd_dev_need_restore(dev, true); > + > + return ret; > +} > + > +/** > + * of_genpd_del_from_domain - Unbind device from its power domain. > + * @dev: Device to unbind from its power domain. > + * > + * Unbinds device from power domain previously bound to it. > + * > + * Returns 0 on success or negative error code otherwise. > + */ > +static int of_genpd_del_from_domain(struct device *dev) > +{ > + struct generic_pm_domain *genpd = dev_to_genpd(dev); > + int ret; > + > + if (IS_ERR(genpd)) > + return 0; > + > + dev_dbg(dev, "removing from power domain %s\n", genpd->name); > + > + while (1) { > + ret = pm_genpd_remove_device(genpd, dev); > + if (ret != -EAGAIN) > + break; > + cond_resched(); > + } > + > + return ret; > +} > + > +/** > + * of_genpd_notifier_call - Receive device driver bind/unbind events > + * @nb: Notifier block which sent the event. > + * @event: Received event. > + * @data: Data attached to received event (struct device *). > + * > + * Registered handler for device driver bind/unbind events that lets the > + * code above perform the magic of adding/removing devices to/from its > + * power domains. > + */ > +static int of_genpd_notifier_call(struct notifier_block *nb, > + unsigned long event, void *data) > +{ > + struct device *dev = data; > + int ret; > + > + if (!dev->of_node) > + return NOTIFY_DONE; > + > + switch (event) { > + case BUS_NOTIFY_BIND_DRIVER: > + ret = of_genpd_add_to_domain(dev); > + break; > + > + case BUS_NOTIFY_UNBOUND_DRIVER: > + ret = of_genpd_del_from_domain(dev); > + break; > + > + default: > + return NOTIFY_DONE; > + } > + > + return notifier_from_errno(ret); > +} > + > +static struct notifier_block of_genpd_notifier_block = { > + .notifier_call = of_genpd_notifier_call, > +}; > + > +static int of_genpd_init(void) > +{ > + return bus_register_notifier(&platform_bus_type, > + &of_genpd_notifier_block); > +} > +core_initcall(of_genpd_init); > +#endif > diff --git a/include/linux/pm_domain.h b/include/linux/pm_domain.h > index 7c1d252..08adac0 100644 > --- a/include/linux/pm_domain.h > +++ b/include/linux/pm_domain.h > @@ -310,4 +310,38 @@ static inline void pm_genpd_syscore_poweron(struct device *dev) > pm_genpd_syscore_switch(dev, false); > } > > +/* OF power domain providers */ > +struct of_device_id; > + > +struct genpd_onecell_data { > + struct generic_pm_domain **domains; > + unsigned int domain_num; > +}; > + > +typedef struct generic_pm_domain *(*genpd_xlate_t)(struct of_phandle_args *args, > + void *data); > + > +#ifdef CONFIG_PM_GENERIC_DOMAINS_OF > +int of_genpd_add_provider(struct device_node *np, genpd_xlate_t xlate, > + void *data); > +void of_genpd_del_provider(struct device_node *np); > + > +struct generic_pm_domain *of_genpd_xlate_simple( > + struct of_phandle_args *genpdspec, > + void *data); > +struct generic_pm_domain *of_genpd_xlate_onecell( > + struct of_phandle_args *genpdspec, > + void *data); > +#else /* !CONFIG_PM_GENERIC_DOMAINS_OF */ > +static inline int of_genpd_add_provider(struct device_node *np, > + genpd_xlate_t xlate, void *data) > +{ > + return 0; > +} > +static inline void of_genpd_del_provider(struct device_node *np) {} > + > +#define of_genpd_xlate_simple NULL > +#define of_genpd_xlate_onecell NULL > +#endif /* CONFIG_PM_GENERIC_DOMAINS_OF */ > + > #endif /* _LINUX_PM_DOMAIN_H */ > diff --git a/kernel/power/Kconfig b/kernel/power/Kconfig > index 2fac9cc..45aa98e 100644 > --- a/kernel/power/Kconfig > +++ b/kernel/power/Kconfig > @@ -306,6 +306,10 @@ config PM_GENERIC_DOMAINS_RUNTIME > def_bool y > depends on PM_RUNTIME && PM_GENERIC_DOMAINS > > +config PM_GENERIC_DOMAINS_OF > + def_bool y > + depends on PM_GENERIC_DOMAINS && OF && !ARCH_EXYNOS > + > config CPU_PM > bool > depends on SUSPEND || CPU_IDLE > -- > 1.8.5.2 > > -- > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html > Please read the FAQ at http://www.tux.org/lkml/ >
Hi Kevin, On 14.01.2014 16:42, Kevin Hilman wrote: > Tomasz Figa <tomasz.figa@gmail.com> writes: > >> This patch introduces generic code to perform power domain look-up using >> device tree and automatically bind devices to their power domains. >> Generic device tree binding is introduced to specify power domains of >> devices in their device tree nodes. >> >> Backwards compatibility with legacy Samsung-specific power domain >> bindings is provided, but for now the new code is not compiled when >> CONFIG_ARCH_EXYNOS is selected to avoid collision with legacy code. This >> will change as soon as Exynos power domain code gets converted to use >> the generic framework in further patch. >> >> Signed-off-by: Tomasz Figa <tomasz.figa@gmail.com> > > I haven't read through this in detail yet, but wanted to make sure that > the DT representation can handle nested power domains. At least > SH-mobile has a hierarchy of power domains and the genpd code can handle > that, so wanted to make sure that the DT representation can handle it as > well. The representation of power domains themselves as implied by this patch is fully platform-specific. The only generic part is the #power-domain-cells property, which defines the number of cells needed to identify the power domain of given provider. You are free to have any platform-specific properties (or even generic ones, added on top of this patch) to let you specify the hierarchy in DT. Best regards, Tomasz
Hi Lorenzo, On 16.01.2014 17:34, Lorenzo Pieralisi wrote: > Hi Tomasz, > > thank you for posting this series. I would like to use the DT bindings > for power domains in the bindings for C-states on ARM: > > http://comments.gmane.org/gmane.linux.power-management.general/41012 > > and in particular link a given C-state to a given power domain so that the > kernel will have a way to actually check what devices are lost upon C-state > entry (and for devices I also mean CPU peripheral like PMUs, GIC CPU IF, > caches and possibly cpus, all of them already represented with DT nodes). > > I have a remark: > > - Can we group device nodes under a single power-domain-parent so that > all devices defined under that parent won't have to re-define a > power-domain property (a property like interrupt-parent, so to speak) > > What do you think ? Hmm, I can see potential benefits of such construct on platforms with clear hierarchy of devices, but to make sure I'm getting it correctly, is the following what you have in mind? soc-domain-x@12340000 { compatible = "..."; reg = <...>; power-domain-parent = <&power_domains DOMAIN_X>; device@1000 { compatible = "..."; // inherits power-domain = <&power_domains DOMAIN_X> }; device@2000 { compatible = "..."; // inherits power-domain = <&power_domains DOMAIN_X> }; }; Best regards, Tomasz
On Mon, Jan 20, 2014 at 05:32:53PM +0000, Tomasz Figa wrote: > Hi Lorenzo, > > On 16.01.2014 17:34, Lorenzo Pieralisi wrote: > > Hi Tomasz, > > > > thank you for posting this series. I would like to use the DT bindings > > for power domains in the bindings for C-states on ARM: > > > > http://comments.gmane.org/gmane.linux.power-management.general/41012 > > > > and in particular link a given C-state to a given power domain so that the > > kernel will have a way to actually check what devices are lost upon C-state > > entry (and for devices I also mean CPU peripheral like PMUs, GIC CPU IF, > > caches and possibly cpus, all of them already represented with DT nodes). > > > > I have a remark: > > > > - Can we group device nodes under a single power-domain-parent so that > > all devices defined under that parent won't have to re-define a > > power-domain property (a property like interrupt-parent, so to speak) > > > > What do you think ? > > Hmm, I can see potential benefits of such construct on platforms with > clear hierarchy of devices, but to make sure I'm getting it correctly, > is the following what you have in mind? > > soc-domain-x@12340000 { > compatible = "..."; > reg = <...>; > power-domain-parent = <&power_domains DOMAIN_X>; > > device@1000 { > compatible = "..."; > // inherits power-domain = <&power_domains DOMAIN_X> > }; > > device@2000 { > compatible = "..."; > // inherits power-domain = <&power_domains DOMAIN_X> > }; > }; Yes, exactly, it could avoid duplicated data. I still have an issue with nodes that are per-cpu but define just one node (eg PMU), since a CPU might belong in a power-domain on its own (ie one power domain per-CPU) and basically this means that arch-timers, PMU & company should link to multiple power domains, ie one per CPU or we find a way to define a power domain as "banked". I need to think about this a bit more, thanks for your feedback. Lorenzo
On 01/11, Tomasz Figa wrote: > + > +/** > + * of_genpd_lock() - Lock access to of_genpd_providers list > + */ > +static void of_genpd_lock(void) > +{ > + mutex_lock(&of_genpd_mutex); > +} > + > +/** > + * of_genpd_unlock() - Unlock access to of_genpd_providers list > + */ > +static void of_genpd_unlock(void) > +{ > + mutex_unlock(&of_genpd_mutex); > +} Why do we need these functions? Can't we just call mutex_lock/unlock directly? > + > +/** > + * of_genpd_add_provider() - Register a domain provider for a node > + * @np: Device node pointer associated with domain provider > + * @genpd_src_get: callback for decoding domain > + * @data: context pointer for @genpd_src_get callback. These look a little outdated. > + */ > +int of_genpd_add_provider(struct device_node *np, genpd_xlate_t xlate, > + void *data) > +{ > + struct of_genpd_provider *cp; > + > + cp = kzalloc(sizeof(struct of_genpd_provider), GFP_KERNEL); Please use sizeof(*cp) instead. > + if (!cp) > + return -ENOMEM; > + > + cp->node = of_node_get(np); > + cp->data = data; > + cp->xlate = xlate; > + > + of_genpd_lock(); > + list_add(&cp->link, &of_genpd_providers); > + of_genpd_unlock(); > + pr_debug("Added domain provider from %s\n", np->full_name); > + > + return 0; > +} > +EXPORT_SYMBOL_GPL(of_genpd_add_provider); > + [...] > + > +/* See of_genpd_get_from_provider(). */ > +static struct generic_pm_domain *__of_genpd_get_from_provider( > + struct of_phandle_args *genpdspec) > +{ > + struct of_genpd_provider *provider; > + struct generic_pm_domain *genpd = ERR_PTR(-ENOENT); Can this be -EPROBE_DEFER so that we can defer probe until a later time if the power domain provider hasn't registered yet? > + > + /* Check if we have such a provider in our array */ > + list_for_each_entry(provider, &of_genpd_providers, link) { > + if (provider->node == genpdspec->np) > + genpd = provider->xlate(genpdspec, provider->data); > + if (!IS_ERR(genpd)) > + break; > + } > + > + return genpd; > +} > + [...] > +static int of_genpd_notifier_call(struct notifier_block *nb, > + unsigned long event, void *data) > +{ > + struct device *dev = data; > + int ret; > + > + if (!dev->of_node) > + return NOTIFY_DONE; > + > + switch (event) { > + case BUS_NOTIFY_BIND_DRIVER: > + ret = of_genpd_add_to_domain(dev); > + break; > + > + case BUS_NOTIFY_UNBOUND_DRIVER: > + ret = of_genpd_del_from_domain(dev); > + break; > + > + default: > + return NOTIFY_DONE; > + } > + > + return notifier_from_errno(ret); > +} > + > +static struct notifier_block of_genpd_notifier_block = { > + .notifier_call = of_genpd_notifier_call, > +}; > + > +static int of_genpd_init(void) > +{ > + return bus_register_notifier(&platform_bus_type, > + &of_genpd_notifier_block); > +} > +core_initcall(of_genpd_init); Would it be possible to call the of_genpd_add_to_domain() and of_genpd_del_from_domain() functions directly in the driver core, similar to how the pinctrl framework has a hook in there? That way we're not relying on any initcall ordering for this.
Hi Stephen, On 23.01.2014 01:18, Stephen Boyd wrote: > On 01/11, Tomasz Figa wrote: >> + >> +/** >> + * of_genpd_lock() - Lock access to of_genpd_providers list >> + */ >> +static void of_genpd_lock(void) >> +{ >> + mutex_lock(&of_genpd_mutex); >> +} >> + >> +/** >> + * of_genpd_unlock() - Unlock access to of_genpd_providers list >> + */ >> +static void of_genpd_unlock(void) >> +{ >> + mutex_unlock(&of_genpd_mutex); >> +} > > Why do we need these functions? Can't we just call > mutex_lock/unlock directly? That would be fine as well, I guess. Just duplicated the pattern used in CCF, but can remove them in next version if it's found to be better. > >> + >> +/** >> + * of_genpd_add_provider() - Register a domain provider for a node >> + * @np: Device node pointer associated with domain provider >> + * @genpd_src_get: callback for decoding domain >> + * @data: context pointer for @genpd_src_get callback. > > These look a little outdated. Oops, missed this. > >> + */ >> +int of_genpd_add_provider(struct device_node *np, genpd_xlate_t xlate, >> + void *data) >> +{ >> + struct of_genpd_provider *cp; >> + >> + cp = kzalloc(sizeof(struct of_genpd_provider), GFP_KERNEL); > > Please use sizeof(*cp) instead. Right. > >> + if (!cp) >> + return -ENOMEM; >> + >> + cp->node = of_node_get(np); >> + cp->data = data; >> + cp->xlate = xlate; >> + >> + of_genpd_lock(); >> + list_add(&cp->link, &of_genpd_providers); >> + of_genpd_unlock(); >> + pr_debug("Added domain provider from %s\n", np->full_name); >> + >> + return 0; >> +} >> +EXPORT_SYMBOL_GPL(of_genpd_add_provider); >> + > [...] >> + >> +/* See of_genpd_get_from_provider(). */ >> +static struct generic_pm_domain *__of_genpd_get_from_provider( >> + struct of_phandle_args *genpdspec) >> +{ >> + struct of_genpd_provider *provider; >> + struct generic_pm_domain *genpd = ERR_PTR(-ENOENT); > > Can this be -EPROBE_DEFER so that we can defer probe until a > later time if the power domain provider hasn't registered yet? Yes, this could be useful. Makes me wonder why clock code (on which I based this code) doesn't have it done this way. > >> + >> + /* Check if we have such a provider in our array */ >> + list_for_each_entry(provider, &of_genpd_providers, link) { >> + if (provider->node == genpdspec->np) >> + genpd = provider->xlate(genpdspec, provider->data); >> + if (!IS_ERR(genpd)) >> + break; >> + } >> + >> + return genpd; >> +} >> + > [...] >> +static int of_genpd_notifier_call(struct notifier_block *nb, >> + unsigned long event, void *data) >> +{ >> + struct device *dev = data; >> + int ret; >> + >> + if (!dev->of_node) >> + return NOTIFY_DONE; >> + >> + switch (event) { >> + case BUS_NOTIFY_BIND_DRIVER: >> + ret = of_genpd_add_to_domain(dev); >> + break; >> + >> + case BUS_NOTIFY_UNBOUND_DRIVER: >> + ret = of_genpd_del_from_domain(dev); >> + break; >> + >> + default: >> + return NOTIFY_DONE; >> + } >> + >> + return notifier_from_errno(ret); >> +} >> + >> +static struct notifier_block of_genpd_notifier_block = { >> + .notifier_call = of_genpd_notifier_call, >> +}; >> + >> +static int of_genpd_init(void) >> +{ >> + return bus_register_notifier(&platform_bus_type, >> + &of_genpd_notifier_block); >> +} >> +core_initcall(of_genpd_init); > > Would it be possible to call the of_genpd_add_to_domain() and > of_genpd_del_from_domain() functions directly in the driver core, > similar to how the pinctrl framework has a hook in there? That > way we're not relying on any initcall ordering for this. Hmm, the initcall here just registers a notifier, which needs to be done just before any driver registers. So, IMHO, current variant is safe, given an early enough initcall level is used. However, doing it the pinctrl way might still have an advantage of not relying on specific bus type, so this is worth consideration indeed. I'd like to hear Rafael's and Kevin's opinions on this (and other comments above too). Best regards, Tomasz
On 01/20, Tomasz Figa wrote: > Hi Kevin, > > On 14.01.2014 16:42, Kevin Hilman wrote: > >Tomasz Figa <tomasz.figa@gmail.com> writes: > > > >>This patch introduces generic code to perform power domain look-up using > >>device tree and automatically bind devices to their power domains. > >>Generic device tree binding is introduced to specify power domains of > >>devices in their device tree nodes. > >> > >>Backwards compatibility with legacy Samsung-specific power domain > >>bindings is provided, but for now the new code is not compiled when > >>CONFIG_ARCH_EXYNOS is selected to avoid collision with legacy code. This > >>will change as soon as Exynos power domain code gets converted to use > >>the generic framework in further patch. > >> > >>Signed-off-by: Tomasz Figa <tomasz.figa@gmail.com> > > > >I haven't read through this in detail yet, but wanted to make sure that > >the DT representation can handle nested power domains. At least > >SH-mobile has a hierarchy of power domains and the genpd code can handle > >that, so wanted to make sure that the DT representation can handle it as > >well. > > The representation of power domains themselves as implied by this > patch is fully platform-specific. The only generic part is the > #power-domain-cells property, which defines the number of cells > needed to identify the power domain of given provider. You are free > to have any platform-specific properties (or even generic ones, > added on top of this patch) to let you specify the hierarchy in DT. > (Semi-related to this thread, but not really the patchset) I'd like to have a way to say that this power domain is a subdomain of another domain provided by a different power domain provider driver. From what I can tell, the only way to reparent domains as of today is by name or reference and you have to make a function call to do it (pm_genpd_add_subdomain_names() or pm_genpd_add_subdomain()). This is annoying in the case where all the power domains are not regsitered within the same driver because we don't know which driver comes first. It would be great if there was a way to specify this relationship explicitly when initializing a power domain so that the reparenting is done automatically without requiring any explicit function call. Perhaps DT could specify this? Or we could add another field to the generic_power_domain struct like parent_name?
Hi Tomasz, Am Samstag, den 11.01.2014, 20:42 +0100 schrieb Tomasz Figa: > This patch introduces generic code to perform power domain look-up using > device tree and automatically bind devices to their power domains. > Generic device tree binding is introduced to specify power domains of > devices in their device tree nodes. > > Backwards compatibility with legacy Samsung-specific power domain > bindings is provided, but for now the new code is not compiled when > CONFIG_ARCH_EXYNOS is selected to avoid collision with legacy code. This > will change as soon as Exynos power domain code gets converted to use > the generic framework in further patch. > > Signed-off-by: Tomasz Figa <tomasz.figa@gmail.com> > --- > .../devicetree/bindings/power/power_domain.txt | 51 ++++ > drivers/base/power/domain.c | 339 +++++++++++++++++++++ > include/linux/pm_domain.h | 34 +++ > kernel/power/Kconfig | 4 + > 4 files changed, 428 insertions(+) > create mode 100644 Documentation/devicetree/bindings/power/power_domain.txt > > diff --git a/Documentation/devicetree/bindings/power/power_domain.txt b/Documentation/devicetree/bindings/power/power_domain.txt > new file mode 100644 > index 0000000..93be5d9 > --- /dev/null > +++ b/Documentation/devicetree/bindings/power/power_domain.txt > @@ -0,0 +1,51 @@ > +* Generic power domains > + > +System on chip designs are often divided into multiple power domains that > +can be used for power gating of selected IP blocks for power saving by > +reduced leakage current. > + > +This device tree binding can be used to bind power domain consumer devices > +with their power domains provided by power domain providers. A power domain > +provider can be represented by any node in the device tree and can provide > +one or more power domains. A consumer node can refer to the provider by > +a phandle and a set of phandle arguments (so called power domain specifier) > +of length specified by #power-domain-cells property in the power domain > +provider node. > + > +==Power domain providers== > + > +Required properties: > + - #power-domain-cells : Number of cells in a power domain specifier; > + Typically 0 for nodes representing a single power domain and 1 for nodes > + providing multiple power domains (e.g. power controllers), but can be > + any value as specified by device tree binding documentation of particular > + provider. > + > +Example: > + > + power: power-controller@12340000 { > + compatible = "foo,power-controller"; > + reg = <0x12340000 0x1000>; > + #power-domain-cells = <1>; > + }; > + > +The node above defines a power controller that is a power domain provider > +and expects one cell as its phandle argument. > + > +==Power domain consumers== > + > +Required properties: > + - power-domain : A phandle and power domain specifier as defined by bindings > + of power controller specified by phandle. > + > +Example: > + > + leaky-device@12350000 { > + compatible = "foo,i-leak-current"; > + reg = <0x12350000 0x1000>; > + power-domain = <&power 0>; > + }; > + > +The node above defines a typical power domain consumer device, which is located > +inside power domain with index 0 of power controller represented by node with > +label "power". > diff --git a/drivers/base/power/domain.c b/drivers/base/power/domain.c > index bfb8955..6d47498 100644 > --- a/drivers/base/power/domain.c > +++ b/drivers/base/power/domain.c > @@ -3,12 +3,16 @@ > * > * Copyright (C) 2011 Rafael J. Wysocki <rjw@sisk.pl>, Renesas Electronics Corp. > * > + * Support for Device Tree based power domain providers: > + * Copyright (C) 2014 Tomasz Figa <tomasz.figa@gmail.com> > + * > * This file is released under the GPLv2. > */ > > #include <linux/init.h> > #include <linux/kernel.h> > #include <linux/io.h> > +#include <linux/platform_device.h> > #include <linux/pm_runtime.h> > #include <linux/pm_domain.h> > #include <linux/pm_qos.h> > @@ -2177,3 +2181,338 @@ void pm_genpd_init(struct generic_pm_domain *genpd, > list_add(&genpd->gpd_list_node, &gpd_list); > mutex_unlock(&gpd_list_lock); > } > + > +#ifdef CONFIG_PM_GENERIC_DOMAINS_OF > +/* > + * DEVICE TREE BASED POWER DOMAIN PROVIDERS > + * > + * The code below implements generic device tree based power domain providers > + * that bind device tree nodes with generic power domains registered in the > + * system. > + * > + * Any driver that registers generic power domains and need to support binding > + * of devices to these domains is supposed to register a power domain provider, > + * which maps a power domain specifier retrieved from device tree to a power > + * domain. > + * > + * Two simple mapping functions have been provided for convenience: > + * - of_genpd_xlate_simple() for 1:1 device tree node to domain mapping, > + * - of_genpd_xlate_onecell() for mapping of multiple domains per node > + * by index. > + */ > + > +/** > + * struct of_genpd_provider - Power domain provider registration structure > + * @link: Entry in global list of domain providers > + * @node: Pointer to device tree node of domain provider > + * @xlate: Provider-specific xlate callback mapping a set of specifier cells > + * into a power domain. > + * @data: context pointer to be passed into @xlate callback > + */ > +struct of_genpd_provider { > + struct list_head link; > + > + struct device_node *node; > + genpd_xlate_t xlate; > + void *data; > +}; > + > +/* List of registered power domain providers. */ > +static LIST_HEAD(of_genpd_providers); > +/* Mutex to protect the list above. */ > +static DEFINE_MUTEX(of_genpd_mutex); > + > +/** > + * of_genpd_lock() - Lock access to of_genpd_providers list > + */ > +static void of_genpd_lock(void) > +{ > + mutex_lock(&of_genpd_mutex); > +} > + > +/** > + * of_genpd_unlock() - Unlock access to of_genpd_providers list > + */ > +static void of_genpd_unlock(void) > +{ > + mutex_unlock(&of_genpd_mutex); > +} > + > +/** > + * of_genpd_xlate_simple() - Xlate function for direct node-domain mapping > + * @genpdspec: OF phandle args to map into a power domain > + * @data: xlate function private data - pointer to struct generic_pm_domain > + * > + * This is a generic xlate function that can be used to model power domains > + * that have their own device tree nodes. The private data of xlate function > + * needs to be a valid pointer to struct generic_pm_domain. > + */ > +struct generic_pm_domain *of_genpd_xlate_simple( > + struct of_phandle_args *genpdspec, > + void *data) > +{ > + if (genpdspec->args_count != 0) > + return ERR_PTR(-EINVAL); > + return data; > +} > +EXPORT_SYMBOL_GPL(of_genpd_xlate_simple); > + > +/** > + * of_genpd_xlate_onecell() - Xlate function for providers using single index. > + * @genpdspec: OF phandle args to map into a power domain > + * @data: xlate function private data - pointer to struct genpd_onecell_data > + * > + * This is a generic xlate function that can be used to model simple power > + * domain controllers that have one device tree node and provide multiple > + * power domains. A single cell is used as an index to an array of power > + * domains specified in genpd_onecell_data struct when registering the > + * provider. > + */ > +struct generic_pm_domain *of_genpd_xlate_onecell( > + struct of_phandle_args *genpdspec, > + void *data) > +{ > + struct genpd_onecell_data *genpd_data = data; > + unsigned int idx = genpdspec->args[0]; > + > + if (genpdspec->args_count != 1) > + return ERR_PTR(-EINVAL); > + > + if (idx >= genpd_data->domain_num) { > + pr_err("%s: invalid domain index %d\n", __func__, idx); > + return ERR_PTR(-EINVAL); > + } > + > + return genpd_data->domains[idx]; > +} > +EXPORT_SYMBOL_GPL(of_genpd_xlate_onecell); > + > +/** > + * of_genpd_add_provider() - Register a domain provider for a node > + * @np: Device node pointer associated with domain provider > + * @genpd_src_get: callback for decoding domain > + * @data: context pointer for @genpd_src_get callback. > + */ > +int of_genpd_add_provider(struct device_node *np, genpd_xlate_t xlate, > + void *data) > +{ > + struct of_genpd_provider *cp; > + > + cp = kzalloc(sizeof(struct of_genpd_provider), GFP_KERNEL); > + if (!cp) > + return -ENOMEM; > + > + cp->node = of_node_get(np); > + cp->data = data; > + cp->xlate = xlate; > + > + of_genpd_lock(); > + list_add(&cp->link, &of_genpd_providers); > + of_genpd_unlock(); > + pr_debug("Added domain provider from %s\n", np->full_name); > + > + return 0; > +} > +EXPORT_SYMBOL_GPL(of_genpd_add_provider); > + > +/** > + * of_genpd_del_provider() - Remove a previously registered domain provider > + * @np: Device node pointer associated with domain provider > + */ > +void of_genpd_del_provider(struct device_node *np) > +{ > + struct of_genpd_provider *cp; > + > + of_genpd_lock(); > + list_for_each_entry(cp, &of_genpd_providers, link) { > + if (cp->node == np) { > + list_del(&cp->link); > + of_node_put(cp->node); > + kfree(cp); > + break; > + } > + } > + of_genpd_unlock(); > +} > +EXPORT_SYMBOL_GPL(of_genpd_del_provider); > + > +/* See of_genpd_get_from_provider(). */ > +static struct generic_pm_domain *__of_genpd_get_from_provider( > + struct of_phandle_args *genpdspec) > +{ > + struct of_genpd_provider *provider; > + struct generic_pm_domain *genpd = ERR_PTR(-ENOENT); > + > + /* Check if we have such a provider in our array */ > + list_for_each_entry(provider, &of_genpd_providers, link) { > + if (provider->node == genpdspec->np) > + genpd = provider->xlate(genpdspec, provider->data); > + if (!IS_ERR(genpd)) > + break; > + } > + > + return genpd; > +} > + > +/** > + * of_genpd_get_from_provider() - Look-up power domain > + * @genpdspec: OF phandle args to use for look-up > + * > + * Looks for domain provider under node specified by @genpdspec and if found > + * uses xlate function of the provider to map phandle args to a power domain. > + * > + * Returns a valid pointer to struct generic_pm_domain on success or ERR_PTR() > + * on failure. > + */ > +static struct generic_pm_domain *of_genpd_get_from_provider( > + struct of_phandle_args *genpdspec) > +{ > + struct generic_pm_domain *genpd; > + > + of_genpd_lock(); > + genpd = __of_genpd_get_from_provider(genpdspec); > + of_genpd_unlock(); > + > + return genpd; > +} > + > +/* > + * DEVICE<->DOMAIN BINDING USING DEVICE TREE LOOK-UP > + * > + * The code below registers a notifier for platform bus devices' > + * BUS_NOTIFY_BIND_DRIVER events and tries to attach devices to their power > + * domains by looking them up using Device Tree. > + * > + * Similarly in BUS_NOTIFY_UNBOUND_DRIVER the device is detached from its > + * domain, since it no longer supports runtime PM without any driver bound > + * to it. > + * > + * Both generic and legacy Samsung-specific DT bindings are supported to > + * keep backwards compatibility with existing DTBs. > + */ > + > +/** > + * of_genpd_add_to_domain - Bind device to its power domain using Device Tree. > + * @dev: Device to bind to its power domain. > + * > + * Tries to parse power domain specifier from device's OF node and if succeeds > + * attaches the device to retrieved power domain. > + * > + * Returns 0 on success or negative error code otherwise. > + */ > +static int of_genpd_add_to_domain(struct device *dev) > +{ > + struct of_phandle_args pd_args; > + struct generic_pm_domain *pd; > + int ret; > + > + ret = of_parse_phandle_with_args(dev->of_node, "power-domain", > + "#power-domain-cells", 0, &pd_args); > + if (ret < 0) { > + if (ret != ENOENT) > + return ret; > + > + /* > + * Try legacy Samsung-specific bindings > + * (for backwards compatibility of DT ABI) > + */ > + pd_args.args_count = 0; > + pd_args.np = of_parse_phandle(dev->of_node, > + "samsung,power-domain", 0); > + if (!pd_args.np) > + return 0; > + } > + > + pd = of_genpd_get_from_provider(&pd_args); > + if (IS_ERR(pd)) > + return PTR_ERR(pd); > + > + dev_dbg(dev, "adding to power domain %s\n", pd->name); > + > + while (1) { > + ret = pm_genpd_add_device(pd, dev); Since pm_genpd_add_device is used here, no gpd_timing_data can be provided. Do you have a plan to solve this? Should the timing data be provided from the device tree? > + if (ret != -EAGAIN) > + break; > + cond_resched(); > + } > + > + if (!ret) > + pm_genpd_dev_need_restore(dev, true); > + > + return ret; > +} [...] regards Philipp
Hi Philipp, On 19.02.2014 17:53, Philipp Zabel wrote: > Am Samstag, den 11.01.2014, 20:42 +0100 schrieb Tomasz Figa: [snip] >> + pd = of_genpd_get_from_provider(&pd_args); >> + if (IS_ERR(pd)) >> + return PTR_ERR(pd); >> + >> + dev_dbg(dev, "adding to power domain %s\n", pd->name); >> + >> + while (1) { >> + ret = pm_genpd_add_device(pd, dev); > > Since pm_genpd_add_device is used here, no gpd_timing_data can be > provided. Do you have a plan to solve this? Should the timing data be > provided from the device tree? Hmm, a quick grep over kernel sources for genpd_.*_add_device gives just a single user of __pm_genpd_name_add_device(), with custom timing data: > arch/arm/mach-shmobile/pm-rmobile.c-void rmobile_add_device_to_domain_td(const char *domain_name, > arch/arm/mach-shmobile/pm-rmobile.c- struct platform_device *pdev, > arch/arm/mach-shmobile/pm-rmobile.c- struct gpd_timing_data *td) > arch/arm/mach-shmobile/pm-rmobile.c-{ > arch/arm/mach-shmobile/pm-rmobile.c- struct device *dev = &pdev->dev; > arch/arm/mach-shmobile/pm-rmobile.c- > arch/arm/mach-shmobile/pm-rmobile.c: __pm_genpd_name_add_device(domain_name, dev, td); > arch/arm/mach-shmobile/pm-rmobile.c- if (pm_clk_no_clocks(dev)) > arch/arm/mach-shmobile/pm-rmobile.c- pm_clk_add(dev, NULL); > arch/arm/mach-shmobile/pm-rmobile.c-} > arch/arm/mach-shmobile/pm-rmobile.c- > arch/arm/mach-shmobile/pm-rmobile.c-void rmobile_add_devices_to_domains(struct pm_domain_device data[], > arch/arm/mach-shmobile/pm-rmobile.c- int size) > arch/arm/mach-shmobile/pm-rmobile.c-{ > arch/arm/mach-shmobile/pm-rmobile.c- struct gpd_timing_data latencies = { > arch/arm/mach-shmobile/pm-rmobile.c- .stop_latency_ns = DEFAULT_DEV_LATENCY_NS, > arch/arm/mach-shmobile/pm-rmobile.c- .start_latency_ns = DEFAULT_DEV_LATENCY_NS, > arch/arm/mach-shmobile/pm-rmobile.c- .save_state_latency_ns = DEFAULT_DEV_LATENCY_NS, > arch/arm/mach-shmobile/pm-rmobile.c- .restore_state_latency_ns = DEFAULT_DEV_LATENCY_NS, > arch/arm/mach-shmobile/pm-rmobile.c- }; > arch/arm/mach-shmobile/pm-rmobile.c- int j; > arch/arm/mach-shmobile/pm-rmobile.c- > arch/arm/mach-shmobile/pm-rmobile.c- for (j = 0; j < size; j++) > arch/arm/mach-shmobile/pm-rmobile.c- rmobile_add_device_to_domain_td(data[j].domain_name, > arch/arm/mach-shmobile/pm-rmobile.c- data[j].pdev, &latencies); > arch/arm/mach-shmobile/pm-rmobile.c-} Moreover the timings used there are just defaults, which makes me wonder if there is any reason to specify them explicitly. Even more interesting is the fact that genpd code can measure those latencies itself. Do you have a particular use case for those timing data or just wondering? I don't think we need to implement support for them right away, if there is no real need to do so. The code and bindings can be extended later to handle them, if needed. As for whether DT is appropriate place to define them, I'm not quite sure. Stop and start latencies look like hardware parameters, but state save and restore are likely to be driver-specific, as it depends on driver code how much time it takes to save and restore needed state (e.g. driver with register cache will not need to do any state save), if I understand these timing data correctly. Best regards, Tomasz
Hi Tomasz, Am Sonntag, den 23.02.2014, 18:07 +0100 schrieb Tomasz Figa: > Hi Philipp, > > On 19.02.2014 17:53, Philipp Zabel wrote: > > Am Samstag, den 11.01.2014, 20:42 +0100 schrieb Tomasz Figa: > > [snip] > > >> + pd = of_genpd_get_from_provider(&pd_args); > >> + if (IS_ERR(pd)) > >> + return PTR_ERR(pd); > >> + > >> + dev_dbg(dev, "adding to power domain %s\n", pd->name); > >> + > >> + while (1) { > >> + ret = pm_genpd_add_device(pd, dev); > > > > Since pm_genpd_add_device is used here, no gpd_timing_data can be > > provided. Do you have a plan to solve this? Should the timing data be > > provided from the device tree? > > Hmm, a quick grep over kernel sources for genpd_.*_add_device > gives just a single user of __pm_genpd_name_add_device(), with custom > timing data: I had added this to my work progress i.MX patches to silence the noisy "... latency exceeded, new value ..." warnings emitted by the power domain framework: http://patchwork.ozlabs.org/patch/320084/ [...] > Moreover the timings used there are just defaults, which makes me wonder > if there is any reason to specify them explicitly. Even more interesting > is the fact that genpd code can measure those latencies itself. > > Do you have a particular use case for those timing data or just > wondering? I don't think we need to implement support for them right > away, if there is no real need to do so. The code and bindings can be > extended later to handle them, if needed. You are right, this is just superficial. > As for whether DT is appropriate place to define them, I'm not quite > sure. Stop and start latencies look like hardware parameters, but state > save and restore are likely to be driver-specific, as it depends on > driver code how much time it takes to save and restore needed state > (e.g. driver with register cache will not need to do any state save), if > I understand these timing data correctly. I have one more, on i.MX6 I manually need to enable the clocks of devices in the power domain during the power-up sequence so that the reset signals can propagate. So far, I have implemented this by registering the device clocks of devices in the power domain with pm_clk_add and then let the genpd power_on callback temporarily enable them using pm_clk_resume: http://patchwork.ozlabs.org/patch/320085/ Whether this is needed seems to me to be a property of the power domain. Do you think this is something we could add to of_genpd_add_to_domain, depending on some flag set in struct generic_pm_domain? I'd like to avoid having to register my own bus notifier and to deal with ordering issues between that and of_genpd_notifier_call. regards Philipp
On 23 January 2014 01:31, Tomasz Figa <tomasz.figa@gmail.com> wrote: > Hi Stephen, > > > On 23.01.2014 01:18, Stephen Boyd wrote: >> >> On 01/11, Tomasz Figa wrote: >>> >>> + >>> +/** >>> + * of_genpd_lock() - Lock access to of_genpd_providers list >>> + */ >>> +static void of_genpd_lock(void) >>> +{ >>> + mutex_lock(&of_genpd_mutex); >>> +} >>> + >>> +/** >>> + * of_genpd_unlock() - Unlock access to of_genpd_providers list >>> + */ >>> +static void of_genpd_unlock(void) >>> +{ >>> + mutex_unlock(&of_genpd_mutex); >>> +} >> >> >> Why do we need these functions? Can't we just call >> mutex_lock/unlock directly? > > > That would be fine as well, I guess. Just duplicated the pattern used in > CCF, but can remove them in next version if it's found to be better. > > >> >>> + >>> +/** >>> + * of_genpd_add_provider() - Register a domain provider for a node >>> + * @np: Device node pointer associated with domain provider >>> + * @genpd_src_get: callback for decoding domain >>> + * @data: context pointer for @genpd_src_get callback. >> >> >> These look a little outdated. > > > Oops, missed this. > > >> >>> + */ >>> +int of_genpd_add_provider(struct device_node *np, genpd_xlate_t xlate, >>> + void *data) >>> +{ >>> + struct of_genpd_provider *cp; >>> + >>> + cp = kzalloc(sizeof(struct of_genpd_provider), GFP_KERNEL); >> >> >> Please use sizeof(*cp) instead. > > > Right. > > >> >>> + if (!cp) >>> + return -ENOMEM; >>> + >>> + cp->node = of_node_get(np); >>> + cp->data = data; >>> + cp->xlate = xlate; >>> + >>> + of_genpd_lock(); >>> + list_add(&cp->link, &of_genpd_providers); >>> + of_genpd_unlock(); >>> + pr_debug("Added domain provider from %s\n", np->full_name); >>> + >>> + return 0; >>> +} >>> +EXPORT_SYMBOL_GPL(of_genpd_add_provider); >>> + >> >> [...] >>> >>> + >>> +/* See of_genpd_get_from_provider(). */ >>> +static struct generic_pm_domain *__of_genpd_get_from_provider( >>> + struct of_phandle_args >>> *genpdspec) >>> +{ >>> + struct of_genpd_provider *provider; >>> + struct generic_pm_domain *genpd = ERR_PTR(-ENOENT); >> >> >> Can this be -EPROBE_DEFER so that we can defer probe until a >> later time if the power domain provider hasn't registered yet? > > > Yes, this could be useful. Makes me wonder why clock code (on which I based > this code) doesn't have it done this way. > > >> >>> + >>> + /* Check if we have such a provider in our array */ >>> + list_for_each_entry(provider, &of_genpd_providers, link) { >>> + if (provider->node == genpdspec->np) >>> + genpd = provider->xlate(genpdspec, >>> provider->data); >>> + if (!IS_ERR(genpd)) >>> + break; >>> + } >>> + >>> + return genpd; >>> +} >>> + >> >> [...] >>> >>> +static int of_genpd_notifier_call(struct notifier_block *nb, >>> + unsigned long event, void *data) >>> +{ >>> + struct device *dev = data; >>> + int ret; >>> + >>> + if (!dev->of_node) >>> + return NOTIFY_DONE; >>> + >>> + switch (event) { >>> + case BUS_NOTIFY_BIND_DRIVER: >>> + ret = of_genpd_add_to_domain(dev); >>> + break; >>> + >>> + case BUS_NOTIFY_UNBOUND_DRIVER: >>> + ret = of_genpd_del_from_domain(dev); >>> + break; >>> + >>> + default: >>> + return NOTIFY_DONE; >>> + } >>> + >>> + return notifier_from_errno(ret); >>> +} >>> + >>> +static struct notifier_block of_genpd_notifier_block = { >>> + .notifier_call = of_genpd_notifier_call, >>> +}; >>> + >>> +static int of_genpd_init(void) >>> +{ >>> + return bus_register_notifier(&platform_bus_type, >>> + &of_genpd_notifier_block); >>> +} >>> +core_initcall(of_genpd_init); >> >> >> Would it be possible to call the of_genpd_add_to_domain() and >> of_genpd_del_from_domain() functions directly in the driver core, >> similar to how the pinctrl framework has a hook in there? That >> way we're not relying on any initcall ordering for this. > > > Hmm, the initcall here just registers a notifier, which needs to be done > just before any driver registers. So, IMHO, current variant is safe, given > an early enough initcall level is used. > > However, doing it the pinctrl way might still have an advantage of not > relying on specific bus type, so this is worth consideration indeed. I'd > like to hear Rafael's and Kevin's opinions on this (and other comments above > too). As you say; certainly there will be other bus types that we need to support as well. For example the amba bus (drivers/amba/bus.c). Additionally I believe similar reasons, why we added the pinctrl handling to driver core, applies to generic power domains. So I think we should give it a try! Kind regards Ulf Hansson > > Best regards, > Tomasz > > -- > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html > Please read the FAQ at http://www.tux.org/lkml/
diff --git a/Documentation/devicetree/bindings/power/power_domain.txt b/Documentation/devicetree/bindings/power/power_domain.txt new file mode 100644 index 0000000..93be5d9 --- /dev/null +++ b/Documentation/devicetree/bindings/power/power_domain.txt @@ -0,0 +1,51 @@ +* Generic power domains + +System on chip designs are often divided into multiple power domains that +can be used for power gating of selected IP blocks for power saving by +reduced leakage current. + +This device tree binding can be used to bind power domain consumer devices +with their power domains provided by power domain providers. A power domain +provider can be represented by any node in the device tree and can provide +one or more power domains. A consumer node can refer to the provider by +a phandle and a set of phandle arguments (so called power domain specifier) +of length specified by #power-domain-cells property in the power domain +provider node. + +==Power domain providers== + +Required properties: + - #power-domain-cells : Number of cells in a power domain specifier; + Typically 0 for nodes representing a single power domain and 1 for nodes + providing multiple power domains (e.g. power controllers), but can be + any value as specified by device tree binding documentation of particular + provider. + +Example: + + power: power-controller@12340000 { + compatible = "foo,power-controller"; + reg = <0x12340000 0x1000>; + #power-domain-cells = <1>; + }; + +The node above defines a power controller that is a power domain provider +and expects one cell as its phandle argument. + +==Power domain consumers== + +Required properties: + - power-domain : A phandle and power domain specifier as defined by bindings + of power controller specified by phandle. + +Example: + + leaky-device@12350000 { + compatible = "foo,i-leak-current"; + reg = <0x12350000 0x1000>; + power-domain = <&power 0>; + }; + +The node above defines a typical power domain consumer device, which is located +inside power domain with index 0 of power controller represented by node with +label "power". diff --git a/drivers/base/power/domain.c b/drivers/base/power/domain.c index bfb8955..6d47498 100644 --- a/drivers/base/power/domain.c +++ b/drivers/base/power/domain.c @@ -3,12 +3,16 @@ * * Copyright (C) 2011 Rafael J. Wysocki <rjw@sisk.pl>, Renesas Electronics Corp. * + * Support for Device Tree based power domain providers: + * Copyright (C) 2014 Tomasz Figa <tomasz.figa@gmail.com> + * * This file is released under the GPLv2. */ #include <linux/init.h> #include <linux/kernel.h> #include <linux/io.h> +#include <linux/platform_device.h> #include <linux/pm_runtime.h> #include <linux/pm_domain.h> #include <linux/pm_qos.h> @@ -2177,3 +2181,338 @@ void pm_genpd_init(struct generic_pm_domain *genpd, list_add(&genpd->gpd_list_node, &gpd_list); mutex_unlock(&gpd_list_lock); } + +#ifdef CONFIG_PM_GENERIC_DOMAINS_OF +/* + * DEVICE TREE BASED POWER DOMAIN PROVIDERS + * + * The code below implements generic device tree based power domain providers + * that bind device tree nodes with generic power domains registered in the + * system. + * + * Any driver that registers generic power domains and need to support binding + * of devices to these domains is supposed to register a power domain provider, + * which maps a power domain specifier retrieved from device tree to a power + * domain. + * + * Two simple mapping functions have been provided for convenience: + * - of_genpd_xlate_simple() for 1:1 device tree node to domain mapping, + * - of_genpd_xlate_onecell() for mapping of multiple domains per node + * by index. + */ + +/** + * struct of_genpd_provider - Power domain provider registration structure + * @link: Entry in global list of domain providers + * @node: Pointer to device tree node of domain provider + * @xlate: Provider-specific xlate callback mapping a set of specifier cells + * into a power domain. + * @data: context pointer to be passed into @xlate callback + */ +struct of_genpd_provider { + struct list_head link; + + struct device_node *node; + genpd_xlate_t xlate; + void *data; +}; + +/* List of registered power domain providers. */ +static LIST_HEAD(of_genpd_providers); +/* Mutex to protect the list above. */ +static DEFINE_MUTEX(of_genpd_mutex); + +/** + * of_genpd_lock() - Lock access to of_genpd_providers list + */ +static void of_genpd_lock(void) +{ + mutex_lock(&of_genpd_mutex); +} + +/** + * of_genpd_unlock() - Unlock access to of_genpd_providers list + */ +static void of_genpd_unlock(void) +{ + mutex_unlock(&of_genpd_mutex); +} + +/** + * of_genpd_xlate_simple() - Xlate function for direct node-domain mapping + * @genpdspec: OF phandle args to map into a power domain + * @data: xlate function private data - pointer to struct generic_pm_domain + * + * This is a generic xlate function that can be used to model power domains + * that have their own device tree nodes. The private data of xlate function + * needs to be a valid pointer to struct generic_pm_domain. + */ +struct generic_pm_domain *of_genpd_xlate_simple( + struct of_phandle_args *genpdspec, + void *data) +{ + if (genpdspec->args_count != 0) + return ERR_PTR(-EINVAL); + return data; +} +EXPORT_SYMBOL_GPL(of_genpd_xlate_simple); + +/** + * of_genpd_xlate_onecell() - Xlate function for providers using single index. + * @genpdspec: OF phandle args to map into a power domain + * @data: xlate function private data - pointer to struct genpd_onecell_data + * + * This is a generic xlate function that can be used to model simple power + * domain controllers that have one device tree node and provide multiple + * power domains. A single cell is used as an index to an array of power + * domains specified in genpd_onecell_data struct when registering the + * provider. + */ +struct generic_pm_domain *of_genpd_xlate_onecell( + struct of_phandle_args *genpdspec, + void *data) +{ + struct genpd_onecell_data *genpd_data = data; + unsigned int idx = genpdspec->args[0]; + + if (genpdspec->args_count != 1) + return ERR_PTR(-EINVAL); + + if (idx >= genpd_data->domain_num) { + pr_err("%s: invalid domain index %d\n", __func__, idx); + return ERR_PTR(-EINVAL); + } + + return genpd_data->domains[idx]; +} +EXPORT_SYMBOL_GPL(of_genpd_xlate_onecell); + +/** + * of_genpd_add_provider() - Register a domain provider for a node + * @np: Device node pointer associated with domain provider + * @genpd_src_get: callback for decoding domain + * @data: context pointer for @genpd_src_get callback. + */ +int of_genpd_add_provider(struct device_node *np, genpd_xlate_t xlate, + void *data) +{ + struct of_genpd_provider *cp; + + cp = kzalloc(sizeof(struct of_genpd_provider), GFP_KERNEL); + if (!cp) + return -ENOMEM; + + cp->node = of_node_get(np); + cp->data = data; + cp->xlate = xlate; + + of_genpd_lock(); + list_add(&cp->link, &of_genpd_providers); + of_genpd_unlock(); + pr_debug("Added domain provider from %s\n", np->full_name); + + return 0; +} +EXPORT_SYMBOL_GPL(of_genpd_add_provider); + +/** + * of_genpd_del_provider() - Remove a previously registered domain provider + * @np: Device node pointer associated with domain provider + */ +void of_genpd_del_provider(struct device_node *np) +{ + struct of_genpd_provider *cp; + + of_genpd_lock(); + list_for_each_entry(cp, &of_genpd_providers, link) { + if (cp->node == np) { + list_del(&cp->link); + of_node_put(cp->node); + kfree(cp); + break; + } + } + of_genpd_unlock(); +} +EXPORT_SYMBOL_GPL(of_genpd_del_provider); + +/* See of_genpd_get_from_provider(). */ +static struct generic_pm_domain *__of_genpd_get_from_provider( + struct of_phandle_args *genpdspec) +{ + struct of_genpd_provider *provider; + struct generic_pm_domain *genpd = ERR_PTR(-ENOENT); + + /* Check if we have such a provider in our array */ + list_for_each_entry(provider, &of_genpd_providers, link) { + if (provider->node == genpdspec->np) + genpd = provider->xlate(genpdspec, provider->data); + if (!IS_ERR(genpd)) + break; + } + + return genpd; +} + +/** + * of_genpd_get_from_provider() - Look-up power domain + * @genpdspec: OF phandle args to use for look-up + * + * Looks for domain provider under node specified by @genpdspec and if found + * uses xlate function of the provider to map phandle args to a power domain. + * + * Returns a valid pointer to struct generic_pm_domain on success or ERR_PTR() + * on failure. + */ +static struct generic_pm_domain *of_genpd_get_from_provider( + struct of_phandle_args *genpdspec) +{ + struct generic_pm_domain *genpd; + + of_genpd_lock(); + genpd = __of_genpd_get_from_provider(genpdspec); + of_genpd_unlock(); + + return genpd; +} + +/* + * DEVICE<->DOMAIN BINDING USING DEVICE TREE LOOK-UP + * + * The code below registers a notifier for platform bus devices' + * BUS_NOTIFY_BIND_DRIVER events and tries to attach devices to their power + * domains by looking them up using Device Tree. + * + * Similarly in BUS_NOTIFY_UNBOUND_DRIVER the device is detached from its + * domain, since it no longer supports runtime PM without any driver bound + * to it. + * + * Both generic and legacy Samsung-specific DT bindings are supported to + * keep backwards compatibility with existing DTBs. + */ + +/** + * of_genpd_add_to_domain - Bind device to its power domain using Device Tree. + * @dev: Device to bind to its power domain. + * + * Tries to parse power domain specifier from device's OF node and if succeeds + * attaches the device to retrieved power domain. + * + * Returns 0 on success or negative error code otherwise. + */ +static int of_genpd_add_to_domain(struct device *dev) +{ + struct of_phandle_args pd_args; + struct generic_pm_domain *pd; + int ret; + + ret = of_parse_phandle_with_args(dev->of_node, "power-domain", + "#power-domain-cells", 0, &pd_args); + if (ret < 0) { + if (ret != ENOENT) + return ret; + + /* + * Try legacy Samsung-specific bindings + * (for backwards compatibility of DT ABI) + */ + pd_args.args_count = 0; + pd_args.np = of_parse_phandle(dev->of_node, + "samsung,power-domain", 0); + if (!pd_args.np) + return 0; + } + + pd = of_genpd_get_from_provider(&pd_args); + if (IS_ERR(pd)) + return PTR_ERR(pd); + + dev_dbg(dev, "adding to power domain %s\n", pd->name); + + while (1) { + ret = pm_genpd_add_device(pd, dev); + if (ret != -EAGAIN) + break; + cond_resched(); + } + + if (!ret) + pm_genpd_dev_need_restore(dev, true); + + return ret; +} + +/** + * of_genpd_del_from_domain - Unbind device from its power domain. + * @dev: Device to unbind from its power domain. + * + * Unbinds device from power domain previously bound to it. + * + * Returns 0 on success or negative error code otherwise. + */ +static int of_genpd_del_from_domain(struct device *dev) +{ + struct generic_pm_domain *genpd = dev_to_genpd(dev); + int ret; + + if (IS_ERR(genpd)) + return 0; + + dev_dbg(dev, "removing from power domain %s\n", genpd->name); + + while (1) { + ret = pm_genpd_remove_device(genpd, dev); + if (ret != -EAGAIN) + break; + cond_resched(); + } + + return ret; +} + +/** + * of_genpd_notifier_call - Receive device driver bind/unbind events + * @nb: Notifier block which sent the event. + * @event: Received event. + * @data: Data attached to received event (struct device *). + * + * Registered handler for device driver bind/unbind events that lets the + * code above perform the magic of adding/removing devices to/from its + * power domains. + */ +static int of_genpd_notifier_call(struct notifier_block *nb, + unsigned long event, void *data) +{ + struct device *dev = data; + int ret; + + if (!dev->of_node) + return NOTIFY_DONE; + + switch (event) { + case BUS_NOTIFY_BIND_DRIVER: + ret = of_genpd_add_to_domain(dev); + break; + + case BUS_NOTIFY_UNBOUND_DRIVER: + ret = of_genpd_del_from_domain(dev); + break; + + default: + return NOTIFY_DONE; + } + + return notifier_from_errno(ret); +} + +static struct notifier_block of_genpd_notifier_block = { + .notifier_call = of_genpd_notifier_call, +}; + +static int of_genpd_init(void) +{ + return bus_register_notifier(&platform_bus_type, + &of_genpd_notifier_block); +} +core_initcall(of_genpd_init); +#endif diff --git a/include/linux/pm_domain.h b/include/linux/pm_domain.h index 7c1d252..08adac0 100644 --- a/include/linux/pm_domain.h +++ b/include/linux/pm_domain.h @@ -310,4 +310,38 @@ static inline void pm_genpd_syscore_poweron(struct device *dev) pm_genpd_syscore_switch(dev, false); } +/* OF power domain providers */ +struct of_device_id; + +struct genpd_onecell_data { + struct generic_pm_domain **domains; + unsigned int domain_num; +}; + +typedef struct generic_pm_domain *(*genpd_xlate_t)(struct of_phandle_args *args, + void *data); + +#ifdef CONFIG_PM_GENERIC_DOMAINS_OF +int of_genpd_add_provider(struct device_node *np, genpd_xlate_t xlate, + void *data); +void of_genpd_del_provider(struct device_node *np); + +struct generic_pm_domain *of_genpd_xlate_simple( + struct of_phandle_args *genpdspec, + void *data); +struct generic_pm_domain *of_genpd_xlate_onecell( + struct of_phandle_args *genpdspec, + void *data); +#else /* !CONFIG_PM_GENERIC_DOMAINS_OF */ +static inline int of_genpd_add_provider(struct device_node *np, + genpd_xlate_t xlate, void *data) +{ + return 0; +} +static inline void of_genpd_del_provider(struct device_node *np) {} + +#define of_genpd_xlate_simple NULL +#define of_genpd_xlate_onecell NULL +#endif /* CONFIG_PM_GENERIC_DOMAINS_OF */ + #endif /* _LINUX_PM_DOMAIN_H */ diff --git a/kernel/power/Kconfig b/kernel/power/Kconfig index 2fac9cc..45aa98e 100644 --- a/kernel/power/Kconfig +++ b/kernel/power/Kconfig @@ -306,6 +306,10 @@ config PM_GENERIC_DOMAINS_RUNTIME def_bool y depends on PM_RUNTIME && PM_GENERIC_DOMAINS +config PM_GENERIC_DOMAINS_OF + def_bool y + depends on PM_GENERIC_DOMAINS && OF && !ARCH_EXYNOS + config CPU_PM bool depends on SUSPEND || CPU_IDLE
This patch introduces generic code to perform power domain look-up using device tree and automatically bind devices to their power domains. Generic device tree binding is introduced to specify power domains of devices in their device tree nodes. Backwards compatibility with legacy Samsung-specific power domain bindings is provided, but for now the new code is not compiled when CONFIG_ARCH_EXYNOS is selected to avoid collision with legacy code. This will change as soon as Exynos power domain code gets converted to use the generic framework in further patch. Signed-off-by: Tomasz Figa <tomasz.figa@gmail.com> --- .../devicetree/bindings/power/power_domain.txt | 51 ++++ drivers/base/power/domain.c | 339 +++++++++++++++++++++ include/linux/pm_domain.h | 34 +++ kernel/power/Kconfig | 4 + 4 files changed, 428 insertions(+) create mode 100644 Documentation/devicetree/bindings/power/power_domain.txt