From 2430d12c94ff2bafcfe4f65edf7ee5f300d2d9c6 Mon Sep 17 00:00:00 2001 From: Alan Stern Date: Sun, 13 Jun 2010 00:36:52 +0200 Subject: PM: describe kernel policy regarding wakeup defaults (v. 2) This patch (as1381b) updates a comment describing the kernel's policy toward enabling wakeup by default. It also makes device_set_wakeup_capable() actually do something when CONFIG_PM isn't enabled. It's not clear this is necessary; however if it isn't then device_init_wakeup() and device_can_wakeup() should also be do-nothing routines. Furthermore, I don't expect this change to have any noticeable effect -- but if it does then clearly the old behavior was wrong. Signed-off-by: Alan Stern Signed-off-by: Rafael J. Wysocki --- include/linux/pm_wakeup.h | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-) (limited to 'include') diff --git a/include/linux/pm_wakeup.h b/include/linux/pm_wakeup.h index 22d64c18056c..76aca48722ae 100644 --- a/include/linux/pm_wakeup.h +++ b/include/linux/pm_wakeup.h @@ -29,8 +29,11 @@ #ifdef CONFIG_PM -/* changes to device_may_wakeup take effect on the next pm state change. - * by default, devices should wakeup if they can. +/* Changes to device_may_wakeup take effect on the next pm state change. + * + * By default, most devices should leave wakeup disabled. The exceptions + * are devices that everyone expects to be wakeup sources: keyboards, + * power buttons, possibly network interfaces, etc. */ static inline void device_init_wakeup(struct device *dev, bool val) { @@ -59,7 +62,7 @@ static inline bool device_may_wakeup(struct device *dev) #else /* !CONFIG_PM */ -/* For some reason the next two routines work even without CONFIG_PM */ +/* For some reason the following routines work even without CONFIG_PM */ static inline void device_init_wakeup(struct device *dev, bool val) { dev->power.can_wakeup = val; @@ -67,6 +70,7 @@ static inline void device_init_wakeup(struct device *dev, bool val) static inline void device_set_wakeup_capable(struct device *dev, bool capable) { + dev->power.can_wakeup = capable; } static inline bool device_can_wakeup(struct device *dev) -- cgit v1.2.3 From b14e033e17d0ea0ba12668d0d2f371cd31586994 Mon Sep 17 00:00:00 2001 From: Alan Stern Date: Tue, 29 Jun 2010 22:49:24 +0200 Subject: PNPACPI: Add support for remote wakeup This patch (as1354) adds remote-wakeup support to the pnpacpi driver. The new can_wakeup method also allows other PNP protocol drivers (pnpbios or iaspnp) to add wakeup support, but I don't know enough about how they work to actually do it. Signed-off-by: Alan Stern Reviewed-by: Bjorn Helgaas Signed-off-by: Rafael J. Wysocki --- include/linux/pnp.h | 1 + 1 file changed, 1 insertion(+) (limited to 'include') diff --git a/include/linux/pnp.h b/include/linux/pnp.h index 7c4193eb0072..1bc1338b817b 100644 --- a/include/linux/pnp.h +++ b/include/linux/pnp.h @@ -414,6 +414,7 @@ struct pnp_protocol { int (*disable) (struct pnp_dev *dev); /* protocol specific suspend/resume */ + bool (*can_wakeup) (struct pnp_dev *dev); int (*suspend) (struct pnp_dev * dev, pm_message_t state); int (*resume) (struct pnp_dev * dev); -- cgit v1.2.3 From c125e96f044427f38d106fab7bc5e4a5e6a18262 Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Mon, 5 Jul 2010 22:43:53 +0200 Subject: PM: Make it possible to avoid races between wakeup and system sleep One of the arguments during the suspend blockers discussion was that the mainline kernel didn't contain any mechanisms making it possible to avoid races between wakeup and system suspend. Generally, there are two problems in that area. First, if a wakeup event occurs exactly when /sys/power/state is being written to, it may be delivered to user space right before the freezer kicks in, so the user space consumer of the event may not be able to process it before the system is suspended. Second, if a wakeup event occurs after user space has been frozen, it is not generally guaranteed that the ongoing transition of the system into a sleep state will be aborted. To address these issues introduce a new global sysfs attribute, /sys/power/wakeup_count, associated with a running counter of wakeup events and three helper functions, pm_stay_awake(), pm_relax(), and pm_wakeup_event(), that may be used by kernel subsystems to control the behavior of this attribute and to request the PM core to abort system transitions into a sleep state already in progress. The /sys/power/wakeup_count file may be read from or written to by user space. Reads will always succeed (unless interrupted by a signal) and return the current value of the wakeup events counter. Writes, however, will only succeed if the written number is equal to the current value of the wakeup events counter. If a write is successful, it will cause the kernel to save the current value of the wakeup events counter and to abort the subsequent system transition into a sleep state if any wakeup events are reported after the write has returned. [The assumption is that before writing to /sys/power/state user space will first read from /sys/power/wakeup_count. Next, user space consumers of wakeup events will have a chance to acknowledge or veto the upcoming system transition to a sleep state. Finally, if the transition is allowed to proceed, /sys/power/wakeup_count will be written to and if that succeeds, /sys/power/state will be written to as well. Still, if any wakeup events are reported to the PM core by kernel subsystems after that point, the transition will be aborted.] Additionally, put a wakeup events counter into struct dev_pm_info and make these per-device wakeup event counters available via sysfs, so that it's possible to check the activity of various wakeup event sources within the kernel. To illustrate how subsystems can use pm_wakeup_event(), make the low-level PCI runtime PM wakeup-handling code use it. Signed-off-by: Rafael J. Wysocki Acked-by: Jesse Barnes Acked-by: Greg Kroah-Hartman Acked-by: markgross Reviewed-by: Alan Stern --- include/linux/pm.h | 10 ++++++++++ include/linux/suspend.h | 7 +++++++ 2 files changed, 17 insertions(+) (limited to 'include') diff --git a/include/linux/pm.h b/include/linux/pm.h index 8e258c727971..b417fc46f3fc 100644 --- a/include/linux/pm.h +++ b/include/linux/pm.h @@ -457,6 +457,7 @@ struct dev_pm_info { #ifdef CONFIG_PM_SLEEP struct list_head entry; struct completion completion; + unsigned long wakeup_count; #endif #ifdef CONFIG_PM_RUNTIME struct timer_list suspend_timer; @@ -552,6 +553,11 @@ extern void __suspend_report_result(const char *function, void *fn, int ret); } while (0) extern void device_pm_wait_for_dev(struct device *sub, struct device *dev); + +/* drivers/base/power/wakeup.c */ +extern void pm_wakeup_event(struct device *dev, unsigned int msec); +extern void pm_stay_awake(struct device *dev); +extern void pm_relax(void); #else /* !CONFIG_PM_SLEEP */ #define device_pm_lock() do {} while (0) @@ -565,6 +571,10 @@ static inline int dpm_suspend_start(pm_message_t state) #define suspend_report_result(fn, ret) do {} while (0) static inline void device_pm_wait_for_dev(struct device *a, struct device *b) {} + +static inline void pm_wakeup_event(struct device *dev, unsigned int msec) {} +static inline void pm_stay_awake(struct device *dev) {} +static inline void pm_relax(void) {} #endif /* !CONFIG_PM_SLEEP */ /* How to reorder dpm_list after device_move() */ diff --git a/include/linux/suspend.h b/include/linux/suspend.h index bc7d6bb4cd8e..bf1bab7b059c 100644 --- a/include/linux/suspend.h +++ b/include/linux/suspend.h @@ -286,6 +286,13 @@ extern int unregister_pm_notifier(struct notifier_block *nb); { .notifier_call = fn, .priority = pri }; \ register_pm_notifier(&fn##_nb); \ } + +/* drivers/base/power/wakeup.c */ +extern bool events_check_enabled; + +extern bool pm_check_wakeup_events(void); +extern bool pm_get_wakeup_count(unsigned long *count); +extern bool pm_save_wakeup_count(unsigned long count); #else /* !CONFIG_PM_SLEEP */ static inline int register_pm_notifier(struct notifier_block *nb) -- cgit v1.2.3 From 12e4d0cc2e0a776a526c93bb2fcb9267abc6e0b1 Mon Sep 17 00:00:00 2001 From: James Bottomley Date: Thu, 1 Jul 2010 21:46:36 +0200 Subject: plist: Add plist_last plist is currently used by the scheduler, which only needs to know the highest item in the list. This adds plist_last which allows you to find the lowest. This is necessary for using plists to implement a fast search of dynamic ranges in pm_qos which can have both highest and lowest criteria. Signed-off-by: James Bottomley Signed-off-by: Rafael J. Wysocki --- include/linux/plist.h | 29 +++++++++++++++++++++++++++++ 1 file changed, 29 insertions(+) (limited to 'include') diff --git a/include/linux/plist.h b/include/linux/plist.h index 6898985e7b38..7254eda078e5 100644 --- a/include/linux/plist.h +++ b/include/linux/plist.h @@ -259,6 +259,23 @@ static inline int plist_node_empty(const struct plist_node *node) container_of(plist_first(head), type, member) #endif +/** + * plist_last_entry - get the struct for the last entry + * @head: the &struct plist_head pointer + * @type: the type of the struct this is embedded in + * @member: the name of the list_struct within the struct + */ +#ifdef CONFIG_DEBUG_PI_LIST +# define plist_last_entry(head, type, member) \ +({ \ + WARN_ON(plist_head_empty(head)); \ + container_of(plist_last(head), type, member); \ +}) +#else +# define plist_last_entry(head, type, member) \ + container_of(plist_last(head), type, member) +#endif + /** * plist_first - return the first node (and thus, highest priority) * @head: the &struct plist_head pointer @@ -271,4 +288,16 @@ static inline struct plist_node *plist_first(const struct plist_head *head) struct plist_node, plist.node_list); } +/** + * plist_last - return the last node (and thus, lowest priority) + * @head: the &struct plist_head pointer + * + * Assumes the plist is _not_ empty. + */ +static inline struct plist_node *plist_last(const struct plist_head *head) +{ + return list_entry(head->node_list.prev, + struct plist_node, plist.node_list); +} + #endif -- cgit v1.2.3 From 82f682514a5df89ffb3890627eebf0897b7a84ec Mon Sep 17 00:00:00 2001 From: James Bottomley Date: Mon, 5 Jul 2010 22:53:06 +0200 Subject: pm_qos: Get rid of the allocation in pm_qos_add_request() All current users of pm_qos_add_request() have the ability to supply the memory required by the pm_qos routines, so make them do this and eliminate the kmalloc() with pm_qos_add_request(). This has the double benefit of making the call never fail and allowing it to be called from atomic context. Signed-off-by: James Bottomley Signed-off-by: mark gross Signed-off-by: Rafael J. Wysocki --- include/linux/netdevice.h | 2 +- include/linux/pm_qos_params.h | 13 ++++++++++--- include/sound/pcm.h | 2 +- 3 files changed, 12 insertions(+), 5 deletions(-) (limited to 'include') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index b21e4054c12c..2f22119b4b08 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -779,7 +779,7 @@ struct net_device { */ char name[IFNAMSIZ]; - struct pm_qos_request_list *pm_qos_req; + struct pm_qos_request_list pm_qos_req; /* device name hash chain */ struct hlist_node name_hlist; diff --git a/include/linux/pm_qos_params.h b/include/linux/pm_qos_params.h index 8ba440e5eb7f..77cbddb3784c 100644 --- a/include/linux/pm_qos_params.h +++ b/include/linux/pm_qos_params.h @@ -1,8 +1,10 @@ +#ifndef _LINUX_PM_QOS_PARAMS_H +#define _LINUX_PM_QOS_PARAMS_H /* interface for the pm_qos_power infrastructure of the linux kernel. * * Mark Gross */ -#include +#include #include #include @@ -14,9 +16,12 @@ #define PM_QOS_NUM_CLASSES 4 #define PM_QOS_DEFAULT_VALUE -1 -struct pm_qos_request_list; +struct pm_qos_request_list { + struct plist_node list; + int pm_qos_class; +}; -struct pm_qos_request_list *pm_qos_add_request(int pm_qos_class, s32 value); +void pm_qos_add_request(struct pm_qos_request_list *l, int pm_qos_class, s32 value); void pm_qos_update_request(struct pm_qos_request_list *pm_qos_req, s32 new_value); void pm_qos_remove_request(struct pm_qos_request_list *pm_qos_req); @@ -24,4 +29,6 @@ void pm_qos_remove_request(struct pm_qos_request_list *pm_qos_req); int pm_qos_request(int pm_qos_class); int pm_qos_add_notifier(int pm_qos_class, struct notifier_block *notifier); int pm_qos_remove_notifier(int pm_qos_class, struct notifier_block *notifier); +int pm_qos_request_active(struct pm_qos_request_list *req); +#endif diff --git a/include/sound/pcm.h b/include/sound/pcm.h index dd76cdede64d..6e3a29732dc4 100644 --- a/include/sound/pcm.h +++ b/include/sound/pcm.h @@ -366,7 +366,7 @@ struct snd_pcm_substream { int number; char name[32]; /* substream name */ int stream; /* stream (direction) */ - struct pm_qos_request_list *latency_pm_qos_req; /* pm_qos request */ + struct pm_qos_request_list latency_pm_qos_req; /* pm_qos request */ size_t buffer_bytes_max; /* limit ring buffer size */ struct snd_dma_buffer dma_buffer; unsigned int dma_buf_id; -- cgit v1.2.3 From ce4410116c5debfb0e049f5db4b5cd6211e05b80 Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Wed, 7 Jul 2010 23:43:45 +0200 Subject: PM / Suspend: Fix ordering of calls in suspend error paths The ACPI suspend code calls suspend_nvs_free() at a wrong place, which may lead to a memory leak if there's an error executing acpi_pm_prepare(), because acpi_pm_finish() will not be called in that case. However, the root cause of this problem is the apparently confusing ordering of calls in suspend error paths that needs to be fixed. In addition to that, fix a typo in a label name in suspend.c. Signed-off-by: Rafael J. Wysocki Acked-by: Len Brown --- include/linux/suspend.h | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-) (limited to 'include') diff --git a/include/linux/suspend.h b/include/linux/suspend.h index bf1bab7b059c..4af270ec2204 100644 --- a/include/linux/suspend.h +++ b/include/linux/suspend.h @@ -61,14 +61,15 @@ typedef int __bitwise suspend_state_t; * before device drivers' late suspend callbacks are executed. It returns * 0 on success or a negative error code otherwise, in which case the * system cannot enter the desired sleep state (@prepare_late(), @enter(), - * @wake(), and @finish() will not be called in that case). + * and @wake() will not be called in that case). * * @prepare_late: Finish preparing the platform for entering the system sleep * state indicated by @begin(). * @prepare_late is called before disabling nonboot CPUs and after * device drivers' late suspend callbacks have been executed. It returns * 0 on success or a negative error code otherwise, in which case the - * system cannot enter the desired sleep state (@enter() and @wake()). + * system cannot enter the desired sleep state (@enter() will not be + * executed). * * @enter: Enter the system sleep state indicated by @begin() or represented by * the argument if @begin() is not implemented. @@ -81,14 +82,15 @@ typedef int __bitwise suspend_state_t; * resume callbacks are executed. * This callback is optional, but should be implemented by the platforms * that implement @prepare_late(). If implemented, it is always called - * after @enter(), even if @enter() fails. + * after @prepare_late and @enter(), even if one of them fails. * * @finish: Finish wake-up of the platform. * @finish is called right prior to calling device drivers' regular suspend * callbacks. * This callback is optional, but should be implemented by the platforms * that implement @prepare(). If implemented, it is always called after - * @enter() and @wake(), if implemented, even if any of them fails. + * @enter() and @wake(), even if any of them fails. It is executed after + * a failing @prepare. * * @end: Called by the PM core right after resuming devices, to indicate to * the platform that the system has returned to the working state or -- cgit v1.2.3 From 8d4b9d1bfef117862a2889dec4dac227068544c9 Mon Sep 17 00:00:00 2001 From: Arjan van de Ven Date: Mon, 19 Jul 2010 02:01:06 +0200 Subject: PM / Runtime: Add runtime PM statistics (v3) In order for PowerTOP to be able to report how well the new runtime PM is working for the various drivers, the kernel needs to export some basic statistics in sysfs. This patch adds two sysfs files in the runtime PM domain that expose the total time a device has been active, and the time a device has been suspended. With this PowerTOP can compute the activity percentage Active %age = 100 * (delta active) / (delta active + delta suspended) and present the information to the user. I've written the PowerTOP code (slated for version 1.12) already, and the output looks like this: Runtime Device Power Management statistics Active Device name 10.0% 06:00.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL8101E/RTL8102E PCI Express Fast Ethernet controller [version 2: fix stat update bugs noticed by Alan Stern] [version 3: rebase to -next and move the sysfs declaration] Signed-off-by: Arjan van de Ven Signed-off-by: Rafael J. Wysocki --- include/linux/pm.h | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'include') diff --git a/include/linux/pm.h b/include/linux/pm.h index b417fc46f3fc..52e8c55ff314 100644 --- a/include/linux/pm.h +++ b/include/linux/pm.h @@ -477,9 +477,15 @@ struct dev_pm_info { enum rpm_request request; enum rpm_status runtime_status; int runtime_error; + unsigned long active_jiffies; + unsigned long suspended_jiffies; + unsigned long accounting_timestamp; #endif }; +extern void update_pm_runtime_accounting(struct device *dev); + + /* * The PM_EVENT_ messages are also used by drivers implementing the legacy * suspend framework, based on the ->suspend() and ->resume() callbacks common -- cgit v1.2.3