aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorLen Brown <len.brown@intel.com>2007-10-03 18:58:00 -0400
committerLen Brown <len.brown@intel.com>2007-10-10 00:12:41 -0400
commit4f86d3a8e297205780cca027e974fd5f81064780 (patch)
tree3c89624dea48a9de756256c935660ff7b24d5376
parentbbf25010f1a6b761914430f5fca081ec8c7accd1 (diff)
cpuidle: consolidate 2.6.22 cpuidle branch into one patch
commit e5a16b1f9eec0af7cfa0830304b41c1c0833cf9f Author: Len Brown <len.brown@intel.com> Date: Tue Oct 2 23:44:44 2007 -0400 cpuidle: shrink diff processor_idle.c | 440 +++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 429 insertions(+), 11 deletions(-) Signed-off-by: Len Brown <len.brown@intel.com> commit dfbb9d5aedfb18848a3e0d6f6e3e4969febb209c Author: Len Brown <len.brown@intel.com> Date: Wed Sep 26 02:17:55 2007 -0400 cpuidle: reduce diff size Reduces the cpuidle processor_idle.c diff vs 2.6.22 from this processor_idle.c | 2006 ++++++++++++++++++++++++++----------------- 1 file changed, 1219 insertions(+), 787 deletions(-) to this: processor_idle.c | 502 +++++++++++++++++++++++++++++++++++++++---- 1 file changed, 458 insertions(+), 44 deletions(-) ...for the purpose of making the cpuilde patch less invasive and easier to review. no functional changes. build tested only. Signed-off-by: Len Brown <len.brown@intel.com> commit 889172fc915f5a7fe20f35b133cbd205ce69bf6c Author: Venki Pallipadi <venkatesh.pallipadi@intel.com> Date: Thu Sep 13 13:40:05 2007 -0700 cpuidle: Retain old ACPI policy for !CONFIG_CPU_IDLE Retain the old policy in processor_idle, so that when CPU_IDLE is not configured, old C-state policy will still be used. This provides a clean gradual migration path from old ACPI policy to new cpuidle based policy. Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Signed-off-by: Len Brown <len.brown@intel.com> commit 9544a8181edc7ecc33b3bfd69271571f98ed08bc Author: Venki Pallipadi <venkatesh.pallipadi@intel.com> Date: Thu Sep 13 13:39:17 2007 -0700 cpuidle: Configure governors by default Quoting Len "Do not give an option to users to shoot themselves in the foot". Remove the configurability of ladder and menu governors as they are needed for default policy of cpuidle. That way users will not be able to have cpuidle without any policy loosing all C-state power savings. Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Signed-off-by: Len Brown <len.brown@intel.com> commit 8975059a2c1e56cfe83d1bcf031bcf4cb39be743 Author: Adam Belay <abelay@novell.com> Date: Tue Aug 21 18:27:07 2007 -0400 CPUIDLE: load ACPI properly when CPUIDLE is disabled Change the registration return codes for when CPUIDLE support is not compiled into the kernel. As a result, the ACPI processor driver will load properly even if CPUIDLE is unavailable. However, it may be possible to cleanup the ACPI processor driver further and eliminate some dead code paths. Signed-off-by: Adam Belay <abelay@novell.com> Acked-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Signed-off-by: Len Brown <len.brown@intel.com> commit e0322e2b58dd1b12ec669bf84693efe0dc2414a8 Author: Adam Belay <abelay@novell.com> Date: Tue Aug 21 18:26:06 2007 -0400 CPUIDLE: remove cpuidle_get_bm_activity() Remove cpuidle_get_bm_activity() and updates governors accordingly. Signed-off-by: Adam Belay <abelay@novell.com> Acked-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Signed-off-by: Len Brown <len.brown@intel.com> commit 18a6e770d5c82ba26653e53d240caa617e09e9ab Author: Adam Belay <abelay@novell.com> Date: Tue Aug 21 18:25:58 2007 -0400 CPUIDLE: max_cstate fix Currently max_cstate is limited to 0, resulting in no idle processor power management on ACPI platforms. This patch restores the value to the array size. Signed-off-by: Adam Belay <abelay@novell.com> Acked-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Signed-off-by: Len Brown <len.brown@intel.com> commit 1fdc0887286179b40ce24bcdbde663172e205ef0 Author: Adam Belay <abelay@novell.com> Date: Tue Aug 21 18:25:40 2007 -0400 CPUIDLE: handle BM detection inside the ACPI Processor driver Update the ACPI processor driver to detect BM activity and limit state entry depth internally, rather than exposing such requirements to CPUIDLE. As a result, CPUIDLE can drop this ACPI-specific interface and become more platform independent. BM activity is now handled much more aggressively than it was in the original implementation, so some testing coverage may be needed to verify that this doesn't introduce any DMA buffer under-run issues. Signed-off-by: Adam Belay <abelay@novell.com> Acked-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Signed-off-by: Len Brown <len.brown@intel.com> commit 0ef38840db666f48e3cdd2b769da676c57228dd9 Author: Adam Belay <abelay@novell.com> Date: Tue Aug 21 18:25:14 2007 -0400 CPUIDLE: menu governor updates Tweak the menu governor to more effectively handle non-timer break events. Non-timer break events are detected by comparing the actual sleep time to the expected sleep time. In future revisions, it may be more reliable to use the timer data structures directly. Signed-off-by: Adam Belay <abelay@novell.com> Acked-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Signed-off-by: Len Brown <len.brown@intel.com> commit bb4d74fca63fa96cf3ace644b15ae0f12b7df5a1 Author: Adam Belay <abelay@novell.com> Date: Tue Aug 21 18:24:40 2007 -0400 CPUIDLE: fix 'current_governor' sysfs entry Allow the "current_governor" sysfs entry to properly handle input terminated with '\n'. Signed-off-by: Adam Belay <abelay@novell.com> Acked-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Signed-off-by: Len Brown <len.brown@intel.com> commit df3c71559bb69b125f1a48971bf0d17f78bbdf47 Author: Len Brown <len.brown@intel.com> Date: Sun Aug 12 02:00:45 2007 -0400 cpuidle: fix IA64 build (again) Signed-off-by: Len Brown <len.brown@intel.com> commit a02064579e3f9530fd31baae16b1fc46b5a7bca8 Author: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Date: Sun Aug 12 01:39:27 2007 -0400 cpuidle: Remove support for runtime changing of max_cstate Remove support for runtime changeability of max_cstate. Drivers can use use latency APIs. max_cstate can still be used as a boot time option and dmi override. Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Signed-off-by: Len Brown <len.brown@intel.com> commit 0912a44b13adf22f5e3f607d263aed23b4910d7e Author: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Date: Sun Aug 12 01:39:16 2007 -0400 cpuidle: Remove ACPI cstate_limit calls from ipw2100 ipw2100 already has code to use accetable_latency interfaces to limit the C-state. Remove the calls to acpi_set_cstate_limit and acpi_get_cstate_limit as they are redundant. Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Signed-off-by: Len Brown <len.brown@intel.com> commit c649a76e76be6bff1fd770d0a775798813a3f6e0 Author: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Date: Sun Aug 12 01:35:39 2007 -0400 cpuidle: compile fix for pause and resume functions Fix the compilation failure when cpuidle is not compiled in. Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Acked-by: Adam Belay <adam.belay@novell.com> Signed-off-by: Len Brown <len.brown@intel.com> commit 2305a5920fb8ee6ccec1c62ade05aa8351091d71 Author: Adam Belay <abelay@novell.com> Date: Thu Jul 19 00:49:00 2007 -0400 cpuidle: re-write Some portions have been rewritten to make the code cleaner and lighter weight. The following is a list of changes: 1.) the state name is now included in the sysfs interface 2.) detection, hotplug, and available state modifications are handled by CPUIDLE drivers directly 3.) the CPUIDLE idle handler is only ever installed when at least one cpuidle_device is enabled and ready 4.) the menu governor BM code no longer overflows 5.) the sysfs attributes are now printed as unsigned integers, avoiding negative values 6.) a variety of other small cleanups Also, Idle drivers are no longer swappable during runtime through the CPUIDLE sysfs inteface. On i386 and x86_64 most idle handlers (e.g. poll, mwait, halt, etc.) don't benefit from an infrastructure that supports multiple states, so I think using a more general case idle handler selection mechanism would be cleaner. Signed-off-by: Adam Belay <abelay@novell.com> Acked-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Acked-by: Shaohua Li <shaohua.li@intel.com> Signed-off-by: Len Brown <len.brown@intel.com> commit df25b6b56955714e6e24b574d88d1fd11f0c3ee5 Author: Len Brown <len.brown@intel.com> Date: Tue Jul 24 17:08:21 2007 -0400 cpuidle: fix IA64 buid Signed-off-by: Len Brown <len.brown@intel.com> commit fd6ada4c14488755ff7068860078c437431fbccd Author: Adrian Bunk <bunk@stusta.de> Date: Mon Jul 9 11:33:13 2007 -0700 cpuidle: static make cpuidle_replace_governor() static Signed-off-by: Adrian Bunk <bunk@stusta.de> Cc: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Len Brown <len.brown@intel.com> commit c1d4a2cebcadf2429c0c72e1d29aa2a9684c32e0 Author: Adrian Bunk <bunk@stusta.de> Date: Tue Jul 3 00:54:40 2007 -0400 cpuidle: static This patch makes the needlessly global struct menu_governor static. Signed-off-by: Adrian Bunk <bunk@stusta.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Len Brown <len.brown@intel.com> commit dbf8780c6e8d572c2c273da97ed1cca7608fd999 Author: Andrew Morton <akpm@linux-foundation.org> Date: Tue Jul 3 00:49:14 2007 -0400 export symbol tick_nohz_get_sleep_length ERROR: "tick_nohz_get_sleep_length" [drivers/cpuidle/governors/menu.ko] undefined! ERROR: "tick_nohz_get_idle_jiffies" [drivers/cpuidle/governors/menu.ko] undefined! And please be sure to get your changes to core kernel suitably reviewed. Cc: Adam Belay <abelay@novell.com> Cc: Venki Pallipadi <venkatesh.pallipadi@intel.com> Cc: Ingo Molnar <mingo@elte.hu> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: john stultz <johnstul@us.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Len Brown <len.brown@intel.com> commit 29f0e248e7017be15f99febf9143a2cef00b2961 Author: Andrew Morton <akpm@linux-foundation.org> Date: Tue Jul 3 00:43:04 2007 -0400 tick.h needs hrtimer.h It uses hrtimers. Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Len Brown <len.brown@intel.com> commit e40cede7d63a029e92712a3fe02faee60cc38fb4 Author: Venki Pallipadi <venkatesh.pallipadi@intel.com> Date: Tue Jul 3 00:40:34 2007 -0400 cpuidle: first round of documentation updates Documentation changes based on Pavel's feedback. Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Len Brown <len.brown@intel.com> commit 83b42be2efece386976507555c29e7773a0dfcd1 Author: Venki Pallipadi <venkatesh.pallipadi@intel.com> Date: Tue Jul 3 00:39:25 2007 -0400 cpuidle: add rating to the governors and pick the one with highest rating by default Introduce a governor rating scheme to pick the right governor by default. Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Len Brown <len.brown@intel.com> commit d2a74b8c5e8f22def4709330d4bfc4a29209b71c Author: Venki Pallipadi <venkatesh.pallipadi@intel.com> Date: Tue Jul 3 00:38:08 2007 -0400 cpuidle: make cpuidle sysfs driver governor switch off by default Make default cpuidle sysfs to show current_governor and current_driver in read-only mode. More elaborate available_governors and available_drivers with writeable current_governor and current_driver interface only appear with "cpuidle_sysfs_switch" boot parameter. Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Len Brown <len.brown@intel.com> commit 1f60a0e80bf83cf6b55c8845bbe5596ed8f6307b Author: Venki Pallipadi <venkatesh.pallipadi@intel.com> Date: Tue Jul 3 00:37:00 2007 -0400 cpuidle: menu governor: change the early break condition Change the C-state early break out algorithm in menu governor. We only look at early breakouts that result in wakeups shorter than idle state's target_residency. If such a breakout is frequent enough, eliminate the particular idle state upto a timeout period. Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Len Brown <len.brown@intel.com> commit 45a42095cf64b003b4a69be3ce7f434f97d7af51 Author: Venki Pallipadi <venkatesh.pallipadi@intel.com> Date: Tue Jul 3 00:35:38 2007 -0400 cpuidle: fix uninitialized variable in sysfs routine Fix the uninitialized usage of ret. Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Len Brown <len.brown@intel.com> commit 80dca7cdba3e6ee13eae277660873ab9584eb3be Author: Venki Pallipadi <venkatesh.pallipadi@intel.com> Date: Tue Jul 3 00:34:16 2007 -0400 cpuidle: reenable /proc/acpi//power interface for the time being Keep /proc/acpi/processor/CPU*/power around for a while as powertop depends on it. It will be marked deprecated and removed in future. powertop can use cpuidle interfaces instead. Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Len Brown <len.brown@intel.com> commit 589c37c2646c5e3813a51255a5ee1159cb4c33fc Author: Venki Pallipadi <venkatesh.pallipadi@intel.com> Date: Tue Jul 3 00:32:37 2007 -0400 cpuidle: menu governor and hrtimer compile fix Compile fix for menu governor. Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Len Brown <len.brown@intel.com> commit 0ba80bd9ab3ed304cb4f19b722e4cc6740588b5e Author: Len Brown <len.brown@intel.com> Date: Thu May 31 22:51:43 2007 -0400 cpuidle: build fix - cpuidle vs ipw2100 module ERROR: "acpi_set_cstate_limit" [drivers/net/wireless/ipw2100.ko] undefined! Signed-off-by: Len Brown <len.brown@intel.com> commit d7d8fa7f96a7f7682be7c6cc0cc53fa7a18c3b58 Author: Adam Belay <abelay@novell.com> Date: Sat Mar 24 03:47:07 2007 -0400 cpuidle: add the 'menu' governor Here is my first take at implementing an idle PM governor that takes full advantage of NO_HZ. I call it the 'menu' governor because it considers the full list of idle states before each entry. I've kept the implementation fairly simple. It attempts to guess the next residency time and then chooses a state that would meet at least the break-even point between power savings and entry cost. To this end, it selects the deepest idle state that satisfies the following constraints: 1. If the idle time elapsed since bus master activity was detected is below a threshold (currently 20 ms), then limit the selection to C2-type or above. 2. Do not choose a state with a break-even residency that exceeds the expected time remaining until the next timer interrupt. 3. Do not choose a state with a break-even residency that exceeds the elapsed time between the last pair of break events, excluding timer interrupts. This governor has an advantage over "ladder" governor because it proactively checks how much time remains until the next timer interrupt using the tick infrastructure. Also, it handles device interrupt activity more intelligently by not including timer interrupts in break event calculations. Finally, it doesn't make policy decisions using the number of state entries, which can have variable residency times (NO_HZ makes these potentially very large), and instead only considers sleep time deltas. The menu governor can be selected during runtime using the cpuidle sysfs interface like so: "echo "menu" > /sys/devices/system/cpu/cpuidle/current_governor" Signed-off-by: Adam Belay <abelay@novell.com> Signed-off-by: Len Brown <len.brown@intel.com> commit a4bec7e65aa3b7488b879d971651cc99a6c410fe Author: Adam Belay <abelay@novell.com> Date: Sat Mar 24 03:47:03 2007 -0400 cpuidle: export time until next timer interrupt using NO_HZ Expose information about the time remaining until the next timer interrupt expires by utilizing the dynticks infrastructure. Also modify the main idle loop to allow dynticks to handle non-interrupt break events (e.g. DMA). Finally, expose sleep ticks information to external code. Thomas Gleixner is responsible for much of the code in this patch. However, I've made some additional changes, so I'm probably responsible if there are any bugs or oversights :) Signed-off-by: Adam Belay <abelay@novell.com> Signed-off-by: Len Brown <len.brown@intel.com> commit 2929d8996fbc77f41a5ff86bb67cdde3ca7d2d72 Author: Adam Belay <abelay@novell.com> Date: Sat Mar 24 03:46:58 2007 -0400 cpuidle: governor API changes This patch prepares cpuidle for the menu governor. It adds an optional stage after idle state entry to give the governor an opportunity to check why the state was exited. Also it makes sure the idle loop returns after each state entry, allowing the appropriate dynticks code to run. Signed-off-by: Adam Belay <abelay@novell.com> Signed-off-by: Len Brown <len.brown@intel.com> commit 3a7fd42f9825c3b03e364ca59baa751bb350775f Author: Venki Pallipadi <venkatesh.pallipadi@intel.com> Date: Thu Apr 26 00:03:59 2007 -0700 cpuidle: hang fix Prevent hang on x86-64, when ACPI processor driver is added as a module on a system that does not support C-states. x86-64 expects all idle handlers to enable interrupts before returning from idle handler. This is due to enter_idle(), exit_idle() races. Make cpuidle_idle_call() confirm to this when there is no pm_idle_old. Also, cpuidle look at the return values of attch_driver() and set current_driver to NULL if attach fails on all CPUs. Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Len Brown <len.brown@intel.com> commit 4893339a142afbd5b7c01ffadfd53d14746e858e Author: Shaohua Li <shaohua.li@intel.com> Date: Thu Apr 26 10:40:09 2007 +0800 cpuidle: add support for max_cstate limit With CPUIDLE framework, the max_cstate (to limit max cpu c-state) parameter is ingored. Some systems require it to ignore C2/C3 and some drivers like ipw require it too. Signed-off-by: Shaohua Li <shaohua.li@intel.com> Signed-off-by: Len Brown <len.brown@intel.com> commit 43bbbbe1cb998cbd2df656f55bb3bfe30f30e7d1 Author: Shaohua Li <shaohua.li@intel.com> Date: Thu Apr 26 10:40:13 2007 +0800 cpuidle: add cpuidle_fore_redetect_devices API add cpuidle_force_redetect_devices API, which forces all CPU redetect idle states. Next patch will use it. Signed-off-by: Shaohua Li <shaohua.li@intel.com> Signed-off-by: Len Brown <len.brown@intel.com> commit d1edadd608f24836def5ec483d2edccfb37b1d19 Author: Shaohua Li <shaohua.li@intel.com> Date: Thu Apr 26 10:40:01 2007 +0800 cpuidle: fix sysfs related issue Fix the cpuidle sysfs issue. a. make kobject dynamicaly allocated b. fixed sysfs init issue to avoid suspend/resume issue Signed-off-by: Shaohua Li <shaohua.li@intel.com> Signed-off-by: Len Brown <len.brown@intel.com> commit 7169a5cc0d67b263978859672e86c13c23a5570d Author: Randy Dunlap <randy.dunlap@oracle.com> Date: Wed Mar 28 22:52:53 2007 -0400 cpuidle: 1-bit field must be unsigned A 1-bit bitfield has no room for a sign bit. drivers/cpuidle/governors/ladder.c:54:16: error: dubious bitfield without explicit `signed' or `unsigned' Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com> Cc: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Len Brown <len.brown@intel.com> commit 4658620158dc2fbd9e4bcb213c5b6fb5d05ba7d4 Author: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Date: Wed Mar 28 22:52:41 2007 -0400 cpuidle: fix boot hang Patch for cpuidle boot hang reported by Larry Finger here. http://www.ussg.iu.edu/hypermail/linux/kernel/0703.2/2025.html Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Cc: Larry Finger <larry.finger@lwfinger.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Len Brown <len.brown@intel.com> commit c17e168aa6e5fe3851baaae8df2fbc1cf11443a9 Author: Len Brown <len.brown@intel.com> Date: Wed Mar 7 04:37:53 2007 -0500 cpuidle: ladder does not depend on ACPI build fix for CONFIG_ACPI=n In file included from drivers/cpuidle/governors/ladder.c:21: include/acpi/processor.h:88: error: expected specifier-qualifier-list before ‘acpi_integer’ include/acpi/processor.h:106: error: expected specifier-qualifier-list before ‘acpi_integer’ include/acpi/processor.h:168: error: expected specifier-qualifier-list before ‘acpi_handle’ Signed-off-by: Len Brown <len.brown@intel.com> commit 8c91d958246bde68db0c3f0c57b535962ce861cb Author: Adrian Bunk <bunk@stusta.de> Date: Tue Mar 6 02:29:40 2007 -0800 cpuidle: make code static This patch makes the following needlessly global code static: - driver.c: __cpuidle_find_driver() - governor.c: __cpuidle_find_governor() - ladder.c: struct ladder_governor Signed-off-by: Adrian Bunk <bunk@stusta.de> Cc: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Cc: Adam Belay <abelay@novell.com> Cc: Shaohua Li <shaohua.li@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Len Brown <len.brown@intel.com> commit 0c39dc3187094c72c33ab65a64d2017b21f372d2 Author: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Date: Wed Mar 7 02:38:22 2007 -0500 cpu_idle: fix build break This patch fixes a build breakage with !CONFIG_HOTPLUG_CPU and CONFIG_CPU_IDLE. Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Signed-off-by: Adrian Bunk <bunk@stusta.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Len Brown <len.brown@intel.com> commit 8112e3b115659b07df340ef170515799c0105f82 Author: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Date: Tue Mar 6 02:29:39 2007 -0800 cpuidle: build fix for !CPU_IDLE Fix the compile issues when CPU_IDLE is not configured. Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Cc: Adam Belay <abelay@novell.com> Cc: Shaohua Li <shaohua.li@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Len Brown <len.brown@intel.com> commit 1eb4431e9599cd25e0d9872f3c2c8986821839dd Author: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Date: Thu Feb 22 13:54:57 2007 -0800 cpuidle take2: Basic documentation for cpuidle Documentation for cpuidle infrastructure Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Signed-off-by: Adam Belay <abelay@novell.com> Signed-off-by: Shaohua Li <shaohua.li@intel.com> Signed-off-by: Len Brown <len.brown@intel.com> commit ef5f15a8b79123a047285ec2e3899108661df779 Author: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Date: Thu Feb 22 13:54:03 2007 -0800 cpuidle take2: Hookup ACPI C-states driver with cpuidle Hookup ACPI C-states onto generic cpuidle infrastructure. drivers/acpi/procesor_idle.c is now a ACPI C-states driver that registers as a driver in cpuidle infrastructure and the policy part is removed from drivers/acpi/processor_idle.c. We use governor in cpuidle instead. Signed-off-by: Shaohua Li <shaohua.li@intel.com> Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Signed-off-by: Adam Belay <abelay@novell.com> Signed-off-by: Len Brown <len.brown@intel.com> commit 987196fa82d4db52c407e8c9d5dec884ba602183 Author: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Date: Thu Feb 22 13:52:57 2007 -0800 cpuidle take2: Core cpuidle infrastructure Announcing 'cpuidle', a new CPU power management infrastructure to manage idle CPUs in a clean and efficient manner. cpuidle separates out the drivers that can provide support for multiple types of idle states and policy governors that decide on what idle state to use at run time. A cpuidle driver can support multiple idle states based on parameters like varying power consumption, wakeup latency, etc (ACPI C-states for example). A cpuidle governor can be usage model specific (laptop, server, laptop on battery etc). Main advantage of the infrastructure being, it allows independent development of drivers and governors and allows for better CPU power management. A huge thanks to Adam Belay and Shaohua Li who were part of this mini-project since its beginning and are greatly responsible for this patchset. This patch: Core cpuidle infrastructure. Introduces a new abstraction layer for cpuidle: * which manages drivers that can support multiple idles states. Drivers can be generic or particular to specific hardware/platform * allows pluging in multiple policy governors that can take idle state policy decision * The core also has a set of sysfs interfaces with which administrato can know about supported drivers and governors and switch them at run time. Signed-off-by: Adam Belay <abelay@novell.com> Signed-off-by: Shaohua Li <shaohua.li@intel.com> Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Signed-off-by: Len Brown <len.brown@intel.com> Signed-off-by: Len Brown <len.brown@intel.com>
Diffstat
-rw-r--r--arch/i386/Kconfig2
-rw-r--r--arch/x86_64/Kconfig2
-rw-r--r--drivers/Makefile1
-rw-r--r--drivers/acpi/osl.c8
-rw-r--r--drivers/acpi/processor_core.c22
-rw-r--r--drivers/acpi/processor_idle.c482
-rw-r--r--drivers/cpuidle/Kconfig20
-rw-r--r--drivers/cpuidle/Makefile5
-rw-r--r--drivers/cpuidle/cpuidle.c295
-rw-r--r--drivers/cpuidle/cpuidle.h33
-rw-r--r--drivers/cpuidle/driver.c56
-rw-r--r--drivers/cpuidle/governor.c141
-rw-r--r--drivers/cpuidle/governors/Makefile6
-rw-r--r--drivers/cpuidle/governors/ladder.c166
-rw-r--r--drivers/cpuidle/governors/menu.c137
-rw-r--r--drivers/cpuidle/sysfs.c361
-rw-r--r--drivers/net/wireless/ipw2100.c29
-rw-r--r--drivers/net/wireless/ipw2100.h2
-rw-r--r--include/acpi/processor.h5
-rw-r--r--include/linux/acpi.h26
-rw-r--r--include/linux/cpuidle.h180
-rw-r--r--include/linux/tick.h9
-rw-r--r--kernel/time/tick-sched.c16
23 files changed, 1895 insertions, 109 deletions
diff --git a/arch/i386/Kconfig b/arch/i386/Kconfig
index 97b64d7d6bf6..3328e0ab5428 100644
--- a/arch/i386/Kconfig
+++ b/arch/i386/Kconfig
@@ -1069,6 +1069,8 @@ endif # APM
1069 1069
1070source "arch/i386/kernel/cpu/cpufreq/Kconfig" 1070source "arch/i386/kernel/cpu/cpufreq/Kconfig"
1071 1071
1072source "drivers/cpuidle/Kconfig"
1073
1072endmenu 1074endmenu
1073 1075
1074menu "Bus options (PCI, PCMCIA, EISA, MCA, ISA)" 1076menu "Bus options (PCI, PCMCIA, EISA, MCA, ISA)"
diff --git a/arch/x86_64/Kconfig b/arch/x86_64/Kconfig
index b4d9089a6a06..e8dbb37a29cf 100644
--- a/arch/x86_64/Kconfig
+++ b/arch/x86_64/Kconfig
@@ -706,6 +706,8 @@ source "drivers/acpi/Kconfig"
706 706
707source "arch/x86_64/kernel/cpufreq/Kconfig" 707source "arch/x86_64/kernel/cpufreq/Kconfig"
708 708
709source "drivers/cpuidle/Kconfig"
710
709endmenu 711endmenu
710 712
711menu "Bus options (PCI etc.)" 713menu "Bus options (PCI etc.)"
diff --git a/drivers/Makefile b/drivers/Makefile
index f0878b2ec55e..10a9c52c9100 100644
--- a/drivers/Makefile
+++ b/drivers/Makefile
@@ -76,6 +76,7 @@ obj-$(CONFIG_MCA) += mca/
76obj-$(CONFIG_EISA) += eisa/ 76obj-$(CONFIG_EISA) += eisa/
77obj-$(CONFIG_LGUEST_GUEST) += lguest/ 77obj-$(CONFIG_LGUEST_GUEST) += lguest/
78obj-$(CONFIG_CPU_FREQ) += cpufreq/ 78obj-$(CONFIG_CPU_FREQ) += cpufreq/
79obj-$(CONFIG_CPU_IDLE) += cpuidle/
79obj-$(CONFIG_MMC) += mmc/ 80obj-$(CONFIG_MMC) += mmc/
80obj-$(CONFIG_NEW_LEDS) += leds/ 81obj-$(CONFIG_NEW_LEDS) += leds/
81obj-$(CONFIG_INFINIBAND) += infiniband/ 82obj-$(CONFIG_INFINIBAND) += infiniband/
diff --git a/drivers/acpi/osl.c b/drivers/acpi/osl.c
index 12c09fafce9a..5d14d4f10b12 100644
--- a/drivers/acpi/osl.c
+++ b/drivers/acpi/osl.c
@@ -1043,14 +1043,6 @@ static int __init acpi_wake_gpes_always_on_setup(char *str)
1043__setup("acpi_wake_gpes_always_on", acpi_wake_gpes_always_on_setup); 1043__setup("acpi_wake_gpes_always_on", acpi_wake_gpes_always_on_setup);
1044 1044
1045/* 1045/*
1046 * max_cstate is defined in the base kernel so modules can
1047 * change it w/o depending on the state of the processor module.
1048 */
1049unsigned int max_cstate = ACPI_PROCESSOR_MAX_POWER;
1050
1051EXPORT_SYMBOL(max_cstate);
1052
1053/*
1054 * Acquire a spinlock. 1046 * Acquire a spinlock.
1055 * 1047 *
1056 * handle is a pointer to the spinlock_t. 1048 * handle is a pointer to the spinlock_t.
diff --git a/drivers/acpi/processor_core.c b/drivers/acpi/processor_core.c
index 9f11dc296cdd..a7351084f2f8 100644
--- a/drivers/acpi/processor_core.c
+++ b/drivers/acpi/processor_core.c
@@ -44,6 +44,7 @@
44#include <linux/seq_file.h> 44#include <linux/seq_file.h>
45#include <linux/dmi.h> 45#include <linux/dmi.h>
46#include <linux/moduleparam.h> 46#include <linux/moduleparam.h>
47#include <linux/cpuidle.h>
47 48
48#include <asm/io.h> 49#include <asm/io.h>
49#include <asm/system.h> 50#include <asm/system.h>
@@ -1049,11 +1050,13 @@ static int __init acpi_processor_init(void)
1049 return -ENOMEM; 1050 return -ENOMEM;
1050 acpi_processor_dir->owner = THIS_MODULE; 1051 acpi_processor_dir->owner = THIS_MODULE;
1051 1052
1053 result = cpuidle_register_driver(&acpi_idle_driver);
1054 if (result < 0)
1055 goto out_proc;
1056
1052 result = acpi_bus_register_driver(&acpi_processor_driver); 1057 result = acpi_bus_register_driver(&acpi_processor_driver);
1053 if (result < 0) { 1058 if (result < 0)
1054 remove_proc_entry(ACPI_PROCESSOR_CLASS, acpi_root_dir); 1059 goto out_cpuidle;
1055 return result;
1056 }
1057 1060
1058 acpi_processor_install_hotplug_notify(); 1061 acpi_processor_install_hotplug_notify();
1059 1062
@@ -1062,11 +1065,18 @@ static int __init acpi_processor_init(void)
1062 acpi_processor_ppc_init(); 1065 acpi_processor_ppc_init();
1063 1066
1064 return 0; 1067 return 0;
1068
1069out_cpuidle:
1070 cpuidle_unregister_driver(&acpi_idle_driver);
1071
1072out_proc:
1073 remove_proc_entry(ACPI_PROCESSOR_CLASS, acpi_root_dir);
1074
1075 return result;
1065} 1076}
1066 1077
1067static void __exit acpi_processor_exit(void) 1078static void __exit acpi_processor_exit(void)
1068{ 1079{
1069
1070 acpi_processor_ppc_exit(); 1080 acpi_processor_ppc_exit();
1071 1081
1072 acpi_thermal_cpufreq_exit(); 1082 acpi_thermal_cpufreq_exit();
@@ -1075,6 +1085,8 @@ static void __exit acpi_processor_exit(void)
1075 1085
1076 acpi_bus_unregister_driver(&acpi_processor_driver); 1086 acpi_bus_unregister_driver(&acpi_processor_driver);
1077 1087
1088 cpuidle_unregister_driver(&acpi_idle_driver);
1089
1078 remove_proc_entry(ACPI_PROCESSOR_CLASS, acpi_root_dir); 1090 remove_proc_entry(ACPI_PROCESSOR_CLASS, acpi_root_dir);
1079 1091
1080 return; 1092 return;
diff --git a/drivers/acpi/processor_idle.c b/drivers/acpi/processor_idle.c
index f18261368e76..99da6a790857 100644
--- a/drivers/acpi/processor_idle.c
+++ b/drivers/acpi/processor_idle.c
@@ -40,6 +40,7 @@
40#include <linux/sched.h> /* need_resched() */ 40#include <linux/sched.h> /* need_resched() */
41#include <linux/latency.h> 41#include <linux/latency.h>
42#include <linux/clockchips.h> 42#include <linux/clockchips.h>
43#include <linux/cpuidle.h>
43 44
44/* 45/*
45 * Include the apic definitions for x86 to have the APIC timer related defines 46 * Include the apic definitions for x86 to have the APIC timer related defines
@@ -64,14 +65,22 @@ ACPI_MODULE_NAME("processor_idle");
64#define ACPI_PROCESSOR_FILE_POWER "power" 65#define ACPI_PROCESSOR_FILE_POWER "power"
65#define US_TO_PM_TIMER_TICKS(t) ((t * (PM_TIMER_FREQUENCY/1000)) / 1000) 66#define US_TO_PM_TIMER_TICKS(t) ((t * (PM_TIMER_FREQUENCY/1000)) / 1000)
66#define PM_TIMER_TICK_NS (1000000000ULL/PM_TIMER_FREQUENCY) 67#define PM_TIMER_TICK_NS (1000000000ULL/PM_TIMER_FREQUENCY)
68#ifndef CONFIG_CPU_IDLE
67#define C2_OVERHEAD 4 /* 1us (3.579 ticks per us) */ 69#define C2_OVERHEAD 4 /* 1us (3.579 ticks per us) */
68#define C3_OVERHEAD 4 /* 1us (3.579 ticks per us) */ 70#define C3_OVERHEAD 4 /* 1us (3.579 ticks per us) */
69static void (*pm_idle_save) (void) __read_mostly; 71static void (*pm_idle_save) (void) __read_mostly;
70module_param(max_cstate, uint, 0644); 72#else
73#define C2_OVERHEAD 1 /* 1us */
74#define C3_OVERHEAD 1 /* 1us */
75#endif
76#define PM_TIMER_TICKS_TO_US(p) (((p) * 1000)/(PM_TIMER_FREQUENCY/1000))
71 77
78static unsigned int max_cstate __read_mostly = ACPI_PROCESSOR_MAX_POWER;
79module_param(max_cstate, uint, 0000);
72static unsigned int nocst __read_mostly; 80static unsigned int nocst __read_mostly;
73module_param(nocst, uint, 0000); 81module_param(nocst, uint, 0000);
74 82
83#ifndef CONFIG_CPU_IDLE
75/* 84/*
76 * bm_history -- bit-mask with a bit per jiffy of bus-master activity 85 * bm_history -- bit-mask with a bit per jiffy of bus-master activity
77 * 1000 HZ: 0xFFFFFFFF: 32 jiffies = 32ms 86 * 1000 HZ: 0xFFFFFFFF: 32 jiffies = 32ms
@@ -82,9 +91,10 @@ module_param(nocst, uint, 0000);
82static unsigned int bm_history __read_mostly = 91static unsigned int bm_history __read_mostly =
83 (HZ >= 800 ? 0xFFFFFFFF : ((1U << (HZ / 25)) - 1)); 92 (HZ >= 800 ? 0xFFFFFFFF : ((1U << (HZ / 25)) - 1));
84module_param(bm_history, uint, 0644); 93module_param(bm_history, uint, 0644);
85/* -------------------------------------------------------------------------- 94
86 Power Management 95static int acpi_processor_set_power_policy(struct acpi_processor *pr);
87 -------------------------------------------------------------------------- */ 96
97#endif
88 98
89/* 99/*
90 * IBM ThinkPad R40e crashes mysteriously when going into C2 or C3. 100 * IBM ThinkPad R40e crashes mysteriously when going into C2 or C3.
@@ -177,6 +187,18 @@ static inline u32 ticks_elapsed(u32 t1, u32 t2)
177 return ((0xFFFFFFFF - t1) + t2); 187 return ((0xFFFFFFFF - t1) + t2);
178} 188}
179 189
190static inline u32 ticks_elapsed_in_us(u32 t1, u32 t2)
191{
192 if (t2 >= t1)
193 return PM_TIMER_TICKS_TO_US(t2 - t1);
194 else if (!(acpi_gbl_FADT.flags & ACPI_FADT_32BIT_TIMER))
195 return PM_TIMER_TICKS_TO_US(((0x00FFFFFF - t1) + t2) & 0x00FFFFFF);
196 else
197 return PM_TIMER_TICKS_TO_US((0xFFFFFFFF - t1) + t2);
198}
199
200#ifndef CONFIG_CPU_IDLE
201
180static void 202static void
181acpi_processor_power_activate(struct acpi_processor *pr, 203acpi_processor_power_activate(struct acpi_processor *pr,
182 struct acpi_processor_cx *new) 204 struct acpi_processor_cx *new)
@@ -248,6 +270,7 @@ static void acpi_cstate_enter(struct acpi_processor_cx *cstate)
248 unused = inl(acpi_gbl_FADT.xpm_timer_block.address); 270 unused = inl(acpi_gbl_FADT.xpm_timer_block.address);
249 } 271 }
250} 272}
273#endif /* !CONFIG_CPU_IDLE */
251 274
252#ifdef ARCH_APICTIMER_STOPS_ON_C3 275#ifdef ARCH_APICTIMER_STOPS_ON_C3
253 276
@@ -342,6 +365,7 @@ int acpi_processor_resume(struct acpi_device * device)
342 return 0; 365 return 0;
343} 366}
344 367
368#ifndef CONFIG_CPU_IDLE
345static void acpi_processor_idle(void) 369static void acpi_processor_idle(void)
346{ 370{
347 struct acpi_processor *pr = NULL; 371 struct acpi_processor *pr = NULL;
@@ -439,7 +463,7 @@ static void acpi_processor_idle(void)
439 * an SMP system. We do it here instead of doing it at _CST/P_LVL 463 * an SMP system. We do it here instead of doing it at _CST/P_LVL
440 * detection phase, to work cleanly with logical CPU hotplug. 464 * detection phase, to work cleanly with logical CPU hotplug.
441 */ 465 */
442 if ((cx->type != ACPI_STATE_C1) && (num_online_cpus() > 1) && 466 if ((cx->type != ACPI_STATE_C1) && (num_online_cpus() > 1) &&
443 !pr->flags.has_cst && !(acpi_gbl_FADT.flags & ACPI_FADT_C2_MP_SUPPORTED)) 467 !pr->flags.has_cst && !(acpi_gbl_FADT.flags & ACPI_FADT_C2_MP_SUPPORTED))
444 cx = &pr->power.states[ACPI_STATE_C1]; 468 cx = &pr->power.states[ACPI_STATE_C1];
445#endif 469#endif
@@ -739,6 +763,7 @@ static int acpi_processor_set_power_policy(struct acpi_processor *pr)
739 763
740 return 0; 764 return 0;
741} 765}
766#endif /* !CONFIG_CPU_IDLE */
742 767
743static int acpi_processor_get_power_info_fadt(struct acpi_processor *pr) 768static int acpi_processor_get_power_info_fadt(struct acpi_processor *pr)
744{ 769{
@@ -756,7 +781,7 @@ static int acpi_processor_get_power_info_fadt(struct acpi_processor *pr)
756#ifndef CONFIG_HOTPLUG_CPU 781#ifndef CONFIG_HOTPLUG_CPU
757 /* 782 /*
758 * Check for P_LVL2_UP flag before entering C2 and above on 783 * Check for P_LVL2_UP flag before entering C2 and above on
759 * an SMP system. 784 * an SMP system.
760 */ 785 */
761 if ((num_online_cpus() > 1) && 786 if ((num_online_cpus() > 1) &&
762 !(acpi_gbl_FADT.flags & ACPI_FADT_C2_MP_SUPPORTED)) 787 !(acpi_gbl_FADT.flags & ACPI_FADT_C2_MP_SUPPORTED))
@@ -957,7 +982,12 @@ static void acpi_processor_power_verify_c2(struct acpi_processor_cx *cx)
957 * Normalize the C2 latency to expidite policy 982 * Normalize the C2 latency to expidite policy
958 */ 983 */
959 cx->valid = 1; 984 cx->valid = 1;
985
986#ifndef CONFIG_CPU_IDLE
960 cx->latency_ticks = US_TO_PM_TIMER_TICKS(cx->latency); 987 cx->latency_ticks = US_TO_PM_TIMER_TICKS(cx->latency);
988#else
989 cx->latency_ticks = cx->latency;
990#endif
961 991
962 return; 992 return;
963} 993}
@@ -1037,7 +1067,12 @@ static void acpi_processor_power_verify_c3(struct acpi_processor *pr,
1037 * use this in our C3 policy 1067 * use this in our C3 policy
1038 */ 1068 */
1039 cx->valid = 1; 1069 cx->valid = 1;
1070
1071#ifndef CONFIG_CPU_IDLE
1040 cx->latency_ticks = US_TO_PM_TIMER_TICKS(cx->latency); 1072 cx->latency_ticks = US_TO_PM_TIMER_TICKS(cx->latency);
1073#else
1074 cx->latency_ticks = cx->latency;
1075#endif
1041 1076
1042 return; 1077 return;
1043} 1078}
@@ -1102,6 +1137,7 @@ static int acpi_processor_get_power_info(struct acpi_processor *pr)
1102 1137
1103 pr->power.count = acpi_processor_power_verify(pr); 1138 pr->power.count = acpi_processor_power_verify(pr);
1104 1139
1140#ifndef CONFIG_CPU_IDLE
1105 /* 1141 /*
1106 * Set Default Policy 1142 * Set Default Policy
1107 * ------------------ 1143 * ------------------
@@ -1113,6 +1149,7 @@ static int acpi_processor_get_power_info(struct acpi_processor *pr)
1113 result = acpi_processor_set_power_policy(pr); 1149 result = acpi_processor_set_power_policy(pr);
1114 if (result) 1150 if (result)
1115 return result; 1151 return result;
1152#endif
1116 1153
1117 /* 1154 /*
1118 * if one state of type C2 or C3 is available, mark this 1155 * if one state of type C2 or C3 is available, mark this
@@ -1129,35 +1166,6 @@ static int acpi_processor_get_power_info(struct acpi_processor *pr)
1129 return 0; 1166 return 0;
1130} 1167}
1131 1168
1132int acpi_processor_cst_has_changed(struct acpi_processor *pr)
1133{
1134 int result = 0;
1135
1136
1137 if (!pr)
1138 return -EINVAL;
1139
1140 if (nocst) {
1141 return -ENODEV;
1142 }
1143
1144 if (!pr->flags.power_setup_done)
1145 return -ENODEV;
1146
1147 /* Fall back to the default idle loop */
1148 pm_idle = pm_idle_save;
1149 synchronize_sched(); /* Relies on interrupts forcing exit from idle. */
1150
1151 pr->flags.power = 0;
1152 result = acpi_processor_get_power_info(pr);
1153 if ((pr->flags.power == 1) && (pr->flags.power_setup_done))
1154 pm_idle = acpi_processor_idle;
1155
1156 return result;
1157}
1158
1159/* proc interface */
1160
1161static int acpi_processor_power_seq_show(struct seq_file *seq, void *offset) 1169static int acpi_processor_power_seq_show(struct seq_file *seq, void *offset)
1162{ 1170{
1163 struct acpi_processor *pr = seq->private; 1171 struct acpi_processor *pr = seq->private;
@@ -1239,6 +1247,35 @@ static const struct file_operations acpi_processor_power_fops = {
1239 .release = single_release, 1247 .release = single_release,
1240}; 1248};
1241 1249
1250#ifndef CONFIG_CPU_IDLE
1251
1252int acpi_processor_cst_has_changed(struct acpi_processor *pr)
1253{
1254 int result = 0;
1255
1256
1257 if (!pr)
1258 return -EINVAL;
1259
1260 if (nocst) {
1261 return -ENODEV;
1262 }
1263
1264 if (!pr->flags.power_setup_done)
1265 return -ENODEV;
1266
1267 /* Fall back to the default idle loop */
1268 pm_idle = pm_idle_save;
1269 synchronize_sched(); /* Relies on interrupts forcing exit from idle. */
1270
1271 pr->flags.power = 0;
1272 result = acpi_processor_get_power_info(pr);
1273 if ((pr->flags.power == 1) && (pr->flags.power_setup_done))
1274 pm_idle = acpi_processor_idle;
1275
1276 return result;
1277}
1278
1242#ifdef CONFIG_SMP 1279#ifdef CONFIG_SMP
1243static void smp_callback(void *v) 1280static void smp_callback(void *v)
1244{ 1281{
@@ -1261,7 +1298,360 @@ static int acpi_processor_latency_notify(struct notifier_block *b,
1261static struct notifier_block acpi_processor_latency_notifier = { 1298static struct notifier_block acpi_processor_latency_notifier = {
1262 .notifier_call = acpi_processor_latency_notify, 1299 .notifier_call = acpi_processor_latency_notify,
1263}; 1300};
1301
1302#endif
1303
1304#else /* CONFIG_CPU_IDLE */
1305
1306/**
1307 * acpi_idle_bm_check - checks if bus master activity was detected
1308 */
1309static int acpi_idle_bm_check(void)
1310{
1311 u32 bm_status = 0;
1312
1313 acpi_get_register(ACPI_BITREG_BUS_MASTER_STATUS, &bm_status);
1314 if (bm_status)
1315 acpi_set_register(ACPI_BITREG_BUS_MASTER_STATUS, 1);
1316 /*
1317 * PIIX4 Erratum #18: Note that BM_STS doesn't always reflect
1318 * the true state of bus mastering activity; forcing us to
1319 * manually check the BMIDEA bit of each IDE channel.
1320 */
1321 else if (errata.piix4.bmisx) {
1322 if ((inb_p(errata.piix4.bmisx + 0x02) & 0x01)
1323 || (inb_p(errata.piix4.bmisx + 0x0A) & 0x01))
1324 bm_status = 1;
1325 }
1326 return bm_status;
1327}
1328
1329/**
1330 * acpi_idle_update_bm_rld - updates the BM_RLD bit depending on target state
1331 * @pr: the processor
1332 * @target: the new target state
1333 */
1334static inline void acpi_idle_update_bm_rld(struct acpi_processor *pr,
1335 struct acpi_processor_cx *target)
1336{
1337 if (pr->flags.bm_rld_set && target->type != ACPI_STATE_C3) {
1338 acpi_set_register(ACPI_BITREG_BUS_MASTER_RLD, 0);
1339 pr->flags.bm_rld_set = 0;
1340 }
1341
1342 if (!pr->flags.bm_rld_set && target->type == ACPI_STATE_C3) {
1343 acpi_set_register(ACPI_BITREG_BUS_MASTER_RLD, 1);
1344 pr->flags.bm_rld_set = 1;
1345 }
1346}
1347
1348/**
1349 * acpi_idle_do_entry - a helper function that does C2 and C3 type entry
1350 * @cx: cstate data
1351 */
1352static inline void acpi_idle_do_entry(struct acpi_processor_cx *cx)
1353{
1354 if (cx->space_id == ACPI_CSTATE_FFH) {
1355 /* Call into architectural FFH based C-state */
1356 acpi_processor_ffh_cstate_enter(cx);
1357 } else {
1358 int unused;
1359 /* IO port based C-state */
1360 inb(cx->address);
1361 /* Dummy wait op - must do something useless after P_LVL2 read
1362 because chipsets cannot guarantee that STPCLK# signal
1363 gets asserted in time to freeze execution properly. */
1364 unused = inl(acpi_gbl_FADT.xpm_timer_block.address);
1365 }
1366}
1367
1368/**
1369 * acpi_idle_enter_c1 - enters an ACPI C1 state-type
1370 * @dev: the target CPU
1371 * @state: the state data
1372 *
1373 * This is equivalent to the HALT instruction.
1374 */
1375static int acpi_idle_enter_c1(struct cpuidle_device *dev,
1376 struct cpuidle_state *state)
1377{
1378 struct acpi_processor *pr;
1379 struct acpi_processor_cx *cx = cpuidle_get_statedata(state);
1380 pr = processors[smp_processor_id()];
1381
1382 if (unlikely(!pr))
1383 return 0;
1384
1385 if (pr->flags.bm_check)
1386 acpi_idle_update_bm_rld(pr, cx);
1387
1388 current_thread_info()->status &= ~TS_POLLING;
1389 /*
1390 * TS_POLLING-cleared state must be visible before we test
1391 * NEED_RESCHED:
1392 */
1393 smp_mb();
1394 if (!need_resched())
1395 safe_halt();
1396 current_thread_info()->status |= TS_POLLING;
1397
1398 cx->usage++;
1399
1400 return 0;
1401}
1402
1403/**
1404 * acpi_idle_enter_simple - enters an ACPI state without BM handling
1405 * @dev: the target CPU
1406 * @state: the state data
1407 */
1408static int acpi_idle_enter_simple(struct cpuidle_device *dev,
1409 struct cpuidle_state *state)
1410{
1411 struct acpi_processor *pr;
1412 struct acpi_processor_cx *cx = cpuidle_get_statedata(state);
1413 u32 t1, t2;
1414 pr = processors[smp_processor_id()];
1415
1416 if (unlikely(!pr))
1417 return 0;
1418
1419 if (pr->flags.bm_check)
1420 acpi_idle_update_bm_rld(pr, cx);
1421
1422 local_irq_disable();
1423 current_thread_info()->status &= ~TS_POLLING;
1424 /*
1425 * TS_POLLING-cleared state must be visible before we test
1426 * NEED_RESCHED:
1427 */
1428 smp_mb();
1429
1430 if (unlikely(need_resched())) {
1431 current_thread_info()->status |= TS_POLLING;
1432 local_irq_enable();
1433 return 0;
1434 }
1435
1436 if (cx->type == ACPI_STATE_C3)
1437 ACPI_FLUSH_CPU_CACHE();
1438
1439 t1 = inl(acpi_gbl_FADT.xpm_timer_block.address);
1440 acpi_state_timer_broadcast(pr, cx, 1);
1441 acpi_idle_do_entry(cx);
1442 t2 = inl(acpi_gbl_FADT.xpm_timer_block.address);
1443
1444#if defined (CONFIG_GENERIC_TIME) && defined (CONFIG_X86_TSC)
1445 /* TSC could halt in idle, so notify users */
1446 mark_tsc_unstable("TSC halts in idle");;
1447#endif
1448
1449 local_irq_enable();
1450 current_thread_info()->status |= TS_POLLING;
1451
1452 cx->usage++;
1453
1454 acpi_state_timer_broadcast(pr, cx, 0);
1455 cx->time += ticks_elapsed(t1, t2);
1456 return ticks_elapsed_in_us(t1, t2);
1457}
1458
1459static int c3_cpu_count;
1460static DEFINE_SPINLOCK(c3_lock);
1461
1462/**
1463 * acpi_idle_enter_bm - enters C3 with proper BM handling
1464 * @dev: the target CPU
1465 * @state: the state data
1466 *
1467 * If BM is detected, the deepest non-C3 idle state is entered instead.
1468 */
1469static int acpi_idle_enter_bm(struct cpuidle_device *dev,
1470 struct cpuidle_state *state)
1471{
1472 struct acpi_processor *pr;
1473 struct acpi_processor_cx *cx = cpuidle_get_statedata(state);
1474 u32 t1, t2;
1475 pr = processors[smp_processor_id()];
1476
1477 if (unlikely(!pr))
1478 return 0;
1479
1480 local_irq_disable();
1481 current_thread_info()->status &= ~TS_POLLING;
1482 /*
1483 * TS_POLLING-cleared state must be visible before we test
1484 * NEED_RESCHED:
1485 */
1486 smp_mb();
1487
1488 if (unlikely(need_resched())) {
1489 current_thread_info()->status |= TS_POLLING;
1490 local_irq_enable();
1491 return 0;
1492 }
1493
1494 /*
1495 * Must be done before busmaster disable as we might need to
1496 * access HPET !
1497 */
1498 acpi_state_timer_broadcast(pr, cx, 1);
1499
1500 if (acpi_idle_bm_check()) {
1501 cx = pr->power.bm_state;
1502
1503 acpi_idle_update_bm_rld(pr, cx);
1504
1505 t1 = inl(acpi_gbl_FADT.xpm_timer_block.address);
1506 acpi_idle_do_entry(cx);
1507 t2 = inl(acpi_gbl_FADT.xpm_timer_block.address);
1508 } else {
1509 acpi_idle_update_bm_rld(pr, cx);
1510
1511 spin_lock(&c3_lock);
1512 c3_cpu_count++;
1513 /* Disable bus master arbitration when all CPUs are in C3 */
1514 if (c3_cpu_count == num_online_cpus())
1515 acpi_set_register(ACPI_BITREG_ARB_DISABLE, 1);
1516 spin_unlock(&c3_lock);
1517
1518 t1 = inl(acpi_gbl_FADT.xpm_timer_block.address);
1519 acpi_idle_do_entry(cx);
1520 t2 = inl(acpi_gbl_FADT.xpm_timer_block.address);
1521
1522 spin_lock(&c3_lock);
1523 /* Re-enable bus master arbitration */
1524 if (c3_cpu_count == num_online_cpus())
1525 acpi_set_register(ACPI_BITREG_ARB_DISABLE, 0);
1526 c3_cpu_count--;
1527 spin_unlock(&c3_lock);
1528 }
1529
1530#if defined (CONFIG_GENERIC_TIME) && defined (CONFIG_X86_TSC)
1531 /* TSC could halt in idle, so notify users */
1532 mark_tsc_unstable("TSC halts in idle");
1533#endif
1534
1535 local_irq_enable();
1536 current_thread_info()->status |= TS_POLLING;
1537
1538 cx->usage++;
1539
1540 acpi_state_timer_broadcast(pr, cx, 0);
1541 cx->time += ticks_elapsed(t1, t2);
1542 return ticks_elapsed_in_us(t1, t2);
1543}
1544
1545struct cpuidle_driver acpi_idle_driver = {
1546 .name = "acpi_idle",
1547 .owner = THIS_MODULE,
1548};
1549
1550/**
1551 * acpi_processor_setup_cpuidle - prepares and configures CPUIDLE
1552 * @pr: the ACPI processor
1553 */
1554static int acpi_processor_setup_cpuidle(struct acpi_processor *pr)
1555{
1556 int i, count = 0;
1557 struct acpi_processor_cx *cx;
1558 struct cpuidle_state *state;
1559 struct cpuidle_device *dev = &pr->power.dev;
1560
1561 if (!pr->flags.power_setup_done)
1562 return -EINVAL;
1563
1564 if (pr->flags.power == 0) {
1565 return -EINVAL;
1566 }
1567
1568 for (i = 1; i < ACPI_PROCESSOR_MAX_POWER && i <= max_cstate; i++) {
1569 cx = &pr->power.states[i];
1570 state = &dev->states[count];
1571
1572 if (!cx->valid)
1573 continue;
1574
1575#ifdef CONFIG_HOTPLUG_CPU
1576 if ((cx->type != ACPI_STATE_C1) && (num_online_cpus() > 1) &&
1577 !pr->flags.has_cst &&
1578 !(acpi_gbl_FADT.flags & ACPI_FADT_C2_MP_SUPPORTED))
1579 continue;
1264#endif 1580#endif
1581 cpuidle_set_statedata(state, cx);
1582
1583 snprintf(state->name, CPUIDLE_NAME_LEN, "C%d", i);
1584 state->exit_latency = cx->latency;
1585 state->target_residency = cx->latency * 6;
1586 state->power_usage = cx->power;
1587
1588 state->flags = 0;
1589 switch (cx->type) {
1590 case ACPI_STATE_C1:
1591 state->flags |= CPUIDLE_FLAG_SHALLOW;
1592 state->enter = acpi_idle_enter_c1;
1593 break;
1594
1595 case ACPI_STATE_C2:
1596 state->flags |= CPUIDLE_FLAG_BALANCED;
1597 state->flags |= CPUIDLE_FLAG_TIME_VALID;
1598 state->enter = acpi_idle_enter_simple;
1599 break;
1600
1601 case ACPI_STATE_C3:
1602 state->flags |= CPUIDLE_FLAG_DEEP;
1603 state->flags |= CPUIDLE_FLAG_TIME_VALID;
1604 state->flags |= CPUIDLE_FLAG_CHECK_BM;
1605 state->enter = pr->flags.bm_check ?
1606 acpi_idle_enter_bm :
1607 acpi_idle_enter_simple;
1608 break;
1609 }
1610
1611 count++;
1612 }
1613
1614 dev->state_count = count;
1615
1616 if (!count)
1617 return -EINVAL;
1618
1619 /* find the deepest state that can handle active BM */
1620 if (pr->flags.bm_check) {
1621 for (i = 1; i < ACPI_PROCESSOR_MAX_POWER && i <= max_cstate; i++)
1622 if (pr->power.states[i].type == ACPI_STATE_C3)
1623 break;
1624 pr->power.bm_state = &pr->power.states[i-1];
1625 }
1626
1627 return 0;
1628}
1629
1630int acpi_processor_cst_has_changed(struct acpi_processor *pr)
1631{
1632 int ret;
1633
1634 if (!pr)
1635 return -EINVAL;
1636
1637 if (nocst) {
1638 return -ENODEV;
1639 }
1640
1641 if (!pr->flags.power_setup_done)
1642 return -ENODEV;
1643
1644 cpuidle_pause_and_lock();
1645 cpuidle_disable_device(&pr->power.dev);
1646 acpi_processor_get_power_info(pr);
1647 acpi_processor_setup_cpuidle(pr);
1648 ret = cpuidle_enable_device(&pr->power.dev);
1649 cpuidle_resume_and_unlock();
1650
1651 return ret;
1652}
1653
1654#endif /* CONFIG_CPU_IDLE */
1265 1655
1266int __cpuinit acpi_processor_power_init(struct acpi_processor *pr, 1656int __cpuinit acpi_processor_power_init(struct acpi_processor *pr,
1267 struct acpi_device *device) 1657 struct acpi_device *device)
@@ -1279,7 +1669,7 @@ int __cpuinit acpi_processor_power_init(struct acpi_processor *pr,
1279 "ACPI: processor limited to max C-state %d\n", 1669 "ACPI: processor limited to max C-state %d\n",
1280 max_cstate); 1670 max_cstate);
1281 first_run++; 1671 first_run++;
1282#ifdef CONFIG_SMP 1672#if !defined (CONFIG_CPU_IDLE) && defined (CONFIG_SMP)
1283 register_latency_notifier(&acpi_processor_latency_notifier); 1673 register_latency_notifier(&acpi_processor_latency_notifier);
1284#endif 1674#endif
1285 } 1675 }
@@ -1297,6 +1687,7 @@ int __cpuinit acpi_processor_power_init(struct acpi_processor *pr,
1297 } 1687 }
1298 1688
1299 acpi_processor_get_power_info(pr); 1689 acpi_processor_get_power_info(pr);
1690 pr->flags.power_setup_done = 1;
1300 1691
1301 /* 1692 /*
1302 * Install the idle handler if processor power management is supported. 1693 * Install the idle handler if processor power management is supported.
@@ -1304,6 +1695,13 @@ int __cpuinit acpi_processor_power_init(struct acpi_processor *pr,
1304 * platforms that only support C1. 1695 * platforms that only support C1.
1305 */ 1696 */
1306 if ((pr->flags.power) && (!boot_option_idle_override)) { 1697 if ((pr->flags.power) && (!boot_option_idle_override)) {
1698#ifdef CONFIG_CPU_IDLE
1699 acpi_processor_setup_cpuidle(pr);
1700 pr->power.dev.cpu = pr->id;
1701 if (cpuidle_register_device(&pr->power.dev))
1702 return -EIO;
1703#endif
1704
1307 printk(KERN_INFO PREFIX "CPU%d (power states:", pr->id); 1705 printk(KERN_INFO PREFIX "CPU%d (power states:", pr->id);
1308 for (i = 1; i <= pr->power.count; i++) 1706 for (i = 1; i <= pr->power.count; i++)
1309 if (pr->power.states[i].valid) 1707 if (pr->power.states[i].valid)
@@ -1311,10 +1709,12 @@ int __cpuinit acpi_processor_power_init(struct acpi_processor *pr,
1311 pr->power.states[i].type); 1709 pr->power.states[i].type);
1312 printk(")\n"); 1710 printk(")\n");
1313 1711
1712#ifndef CONFIG_CPU_IDLE
1314 if (pr->id == 0) { 1713 if (pr->id == 0) {
1315 pm_idle_save = pm_idle; 1714 pm_idle_save = pm_idle;
1316 pm_idle = acpi_processor_idle; 1715 pm_idle = acpi_processor_idle;
1317 } 1716 }
1717#endif
1318 } 1718 }
1319 1719
1320 /* 'power' [R] */ 1720 /* 'power' [R] */
@@ -1328,21 +1728,24 @@ int __cpuinit acpi_processor_power_init(struct acpi_processor *pr,
1328 entry->owner = THIS_MODULE; 1728 entry->owner = THIS_MODULE;
1329 } 1729 }
1330 1730
1331 pr->flags.power_setup_done = 1;
1332
1333 return 0; 1731 return 0;
1334} 1732}
1335 1733
1336int acpi_processor_power_exit(struct acpi_processor *pr, 1734int acpi_processor_power_exit(struct acpi_processor *pr,
1337 struct acpi_device *device) 1735 struct acpi_device *device)
1338{ 1736{
1339 1737#ifdef CONFIG_CPU_IDLE
1738 if ((pr->flags.power) && (!boot_option_idle_override))
1739 cpuidle_unregister_device(&pr->power.dev);
1740#endif
1340 pr->flags.power_setup_done = 0; 1741 pr->flags.power_setup_done = 0;
1341 1742
1342 if (acpi_device_dir(device)) 1743 if (acpi_device_dir(device))
1343 remove_proc_entry(ACPI_PROCESSOR_FILE_POWER, 1744 remove_proc_entry(ACPI_PROCESSOR_FILE_POWER,
1344 acpi_device_dir(device)); 1745 acpi_device_dir(device));
1345 1746
1747#ifndef CONFIG_CPU_IDLE
1748
1346 /* Unregister the idle handler when processor #0 is removed. */ 1749 /* Unregister the idle handler when processor #0 is removed. */
1347 if (pr->id == 0) { 1750 if (pr->id == 0) {
1348 pm_idle = pm_idle_save; 1751 pm_idle = pm_idle_save;
@@ -1357,6 +1760,7 @@ int acpi_processor_power_exit(struct acpi_processor *pr,
1357 unregister_latency_notifier(&acpi_processor_latency_notifier); 1760 unregister_latency_notifier(&acpi_processor_latency_notifier);
1358#endif 1761#endif
1359 } 1762 }
1763#endif
1360 1764
1361 return 0; 1765 return 0;
1362} 1766}
diff --git a/drivers/cpuidle/Kconfig b/drivers/cpuidle/Kconfig
new file mode 100644
index 000000000000..3bed4127d4ad
--- /dev/null
+++ b/drivers/cpuidle/Kconfig
@@ -0,0 +1,20 @@
1
2config CPU_IDLE
3 bool "CPU idle PM support"
4 help
5 CPU idle is a generic framework for supporting software-controlled
6 idle processor power management. It includes modular cross-platform
7 governors that can be swapped during runtime.
8
9 If you're using a mobile platform that supports CPU idle PM (e.g.
10 an ACPI-capable notebook), you should say Y here.
11
12config CPU_IDLE_GOV_LADDER
13 bool
14 depends on CPU_IDLE
15 default y
16
17config CPU_IDLE_GOV_MENU
18 bool
19 depends on CPU_IDLE && NO_HZ
20 default y
diff --git a/drivers/cpuidle/Makefile b/drivers/cpuidle/Makefile
new file mode 100644
index 000000000000..5634f88379df
--- /dev/null
+++ b/drivers/cpuidle/Makefile
@@ -0,0 +1,5 @@
1#
2# Makefile for cpuidle.
3#
4
5obj-y += cpuidle.o driver.o governor.o sysfs.o governors/
diff --git a/drivers/cpuidle/cpuidle.c b/drivers/cpuidle/cpuidle.c
new file mode 100644
index 000000000000..fdf4106b817b
--- /dev/null
+++ b/drivers/cpuidle/cpuidle.c
@@ -0,0 +1,295 @@
1/*
2 * cpuidle.c - core cpuidle infrastructure
3 *
4 * (C) 2006-2007 Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
5 * Shaohua Li <shaohua.li@intel.com>
6 * Adam Belay <abelay@novell.com>
7 *
8 * This code is licenced under the GPL.
9 */
10
11#include <linux/kernel.h>
12#include <linux/mutex.h>
13#include <linux/sched.h>
14#include <linux/notifier.h>
15#include <linux/latency.h>
16#include <linux/cpu.h>
17#include <linux/cpuidle.h>
18
19#include "cpuidle.h"
20
21DEFINE_PER_CPU(struct cpuidle_device *, cpuidle_devices);
22EXPORT_PER_CPU_SYMBOL_GPL(cpuidle_devices);
23
24DEFINE_MUTEX(cpuidle_lock);
25LIST_HEAD(cpuidle_detected_devices);
26static void (*pm_idle_old)(void);
27
28static int enabled_devices;
29
30/**
31 * cpuidle_idle_call - the main idle loop
32 *
33 * NOTE: no locks or semaphores should be used here
34 */
35static void cpuidle_idle_call(void)
36{
37 struct cpuidle_device *dev = __get_cpu_var(cpuidle_devices);
38 struct cpuidle_state *target_state;
39 int next_state;
40
41 /* check if the device is ready */
42 if (!dev || !dev->enabled) {
43 if (pm_idle_old)
44 pm_idle_old();
45 else
46 local_irq_enable();
47 return;
48 }
49
50 /* ask the governor for the next state */
51 next_state = cpuidle_curr_governor->select(dev);
52 if (need_resched())
53 return;
54 target_state = &dev->states[next_state];
55
56 /* enter the state and update stats */
57 dev->last_residency = target_state->enter(dev, target_state);
58 dev->last_state = target_state;
59 target_state->time += dev->last_residency;
60 target_state->usage++;
61
62 /* give the governor an opportunity to reflect on the outcome */
63 if (cpuidle_curr_governor->reflect)
64 cpuidle_curr_governor->reflect(dev);
65}
66
67/**
68 * cpuidle_install_idle_handler - installs the cpuidle idle loop handler
69 */
70void cpuidle_install_idle_handler(void)
71{
72 if (enabled_devices && (pm_idle != cpuidle_idle_call)) {
73 /* Make sure all changes finished before we switch to new idle */
74 smp_wmb();
75 pm_idle = cpuidle_idle_call;
76 }
77}
78
79/**
80 * cpuidle_uninstall_idle_handler - uninstalls the cpuidle idle loop handler
81 */
82void cpuidle_uninstall_idle_handler(void)
83{
84 if (enabled_devices && (pm_idle != pm_idle_old)) {
85 pm_idle = pm_idle_old;
86 cpu_idle_wait();
87 }
88}
89
90/**
91 * cpuidle_pause_and_lock - temporarily disables CPUIDLE
92 */
93void cpuidle_pause_and_lock(void)
94{
95 mutex_lock(&cpuidle_lock);
96 cpuidle_uninstall_idle_handler();
97}
98
99EXPORT_SYMBOL_GPL(cpuidle_pause_and_lock);
100
101/**
102 * cpuidle_resume_and_unlock - resumes CPUIDLE operation
103 */
104void cpuidle_resume_and_unlock(void)
105{
106 cpuidle_install_idle_handler();
107 mutex_unlock(&cpuidle_lock);
108}
109
110EXPORT_SYMBOL_GPL(cpuidle_resume_and_unlock);
111
112/**
113 * cpuidle_enable_device - enables idle PM for a CPU
114 * @dev: the CPU
115 *
116 * This function must be called between cpuidle_pause_and_lock and
117 * cpuidle_resume_and_unlock when used externally.
118 */
119int cpuidle_enable_device(struct cpuidle_device *dev)
120{
121 int ret, i;
122
123 if (dev->enabled)
124 return 0;
125 if (!cpuidle_curr_driver || !cpuidle_curr_governor)
126 return -EIO;
127 if (!dev->state_count)
128 return -EINVAL;
129
130 if ((ret = cpuidle_add_state_sysfs(dev)))
131 return ret;
132
133 if (cpuidle_curr_governor->enable &&
134 (ret = cpuidle_curr_governor->enable(dev)))
135 goto fail_sysfs;
136
137 for (i = 0; i < dev->state_count; i++) {
138 dev->states[i].usage = 0;
139 dev->states[i].time = 0;
140 }
141 dev->last_residency = 0;
142 dev->last_state = NULL;
143
144 smp_wmb();
145
146 dev->enabled = 1;
147
148 enabled_devices++;
149 return 0;
150
151fail_sysfs:
152 cpuidle_remove_state_sysfs(dev);
153
154 return ret;
155}
156
157EXPORT_SYMBOL_GPL(cpuidle_enable_device);
158
159/**
160 * cpuidle_disable_device - disables idle PM for a CPU
161 * @dev: the CPU
162 *
163 * This function must be called between cpuidle_pause_and_lock and
164 * cpuidle_resume_and_unlock when used externally.
165 */
166void cpuidle_disable_device(struct cpuidle_device *dev)
167{
168 if (!dev->enabled)
169 return;
170 if (!cpuidle_curr_driver || !cpuidle_curr_governor)
171 return;
172
173 dev->enabled = 0;
174
175 if (cpuidle_curr_governor->disable)
176 cpuidle_curr_governor->disable(dev);
177
178 cpuidle_remove_state_sysfs(dev);
179 enabled_devices--;
180}
181
182EXPORT_SYMBOL_GPL(cpuidle_disable_device);
183
184/**
185 * cpuidle_register_device - registers a CPU's idle PM feature
186 * @dev: the cpu
187 */
188int cpuidle_register_device(struct cpuidle_device *dev)
189{
190 int ret;
191 struct sys_device *sys_dev = get_cpu_sysdev((unsigned long)dev->cpu);
192
193 if (!sys_dev)
194 return -EINVAL;
195 if (!try_module_get(cpuidle_curr_driver->owner))
196 return -EINVAL;
197
198 init_completion(&dev->kobj_unregister);
199
200 mutex_lock(&cpuidle_lock);
201
202 per_cpu(cpuidle_devices, dev->cpu) = dev;
203 list_add(&dev->device_list, &cpuidle_detected_devices);
204 if ((ret = cpuidle_add_sysfs(sys_dev))) {
205 mutex_unlock(&cpuidle_lock);
206 module_put(cpuidle_curr_driver->owner);
207 return ret;
208 }
209
210 cpuidle_enable_device(dev);
211 cpuidle_install_idle_handler();
212
213 mutex_unlock(&cpuidle_lock);
214
215 return 0;
216
217}
218
219EXPORT_SYMBOL_GPL(cpuidle_register_device);
220
221/**
222 * cpuidle_unregister_device - unregisters a CPU's idle PM feature
223 * @dev: the cpu
224 */
225void cpuidle_unregister_device(struct cpuidle_device *dev)
226{
227 struct sys_device *sys_dev = get_cpu_sysdev((unsigned long)dev->cpu);
228
229 cpuidle_pause_and_lock();
230
231 cpuidle_disable_device(dev);
232
233 cpuidle_remove_sysfs(sys_dev);
234 list_del(&dev->device_list);
235 wait_for_completion(&dev->kobj_unregister);
236 per_cpu(cpuidle_devices, dev->cpu) = NULL;
237
238 cpuidle_resume_and_unlock();
239
240 module_put(cpuidle_curr_driver->owner);
241}
242
243EXPORT_SYMBOL_GPL(cpuidle_unregister_device);
244
245#ifdef CONFIG_SMP
246
247static void smp_callback(void *v)
248{
249 /* we already woke the CPU up, nothing more to do */
250}
251
252/*
253 * This function gets called when a part of the kernel has a new latency
254 * requirement. This means we need to get all processors out of their C-state,
255 * and then recalculate a new suitable C-state. Just do a cross-cpu IPI; that
256 * wakes them all right up.
257 */
258static int cpuidle_latency_notify(struct notifier_block *b,
259 unsigned long l, void *v)
260{
261 smp_call_function(smp_callback, NULL, 0, 1);
262 return NOTIFY_OK;
263}
264
265static struct notifier_block cpuidle_latency_notifier = {
266 .notifier_call = cpuidle_latency_notify,
267};
268
269#define latency_notifier_init(x) do { register_latency_notifier(x); } while (0)
270
271#else /* CONFIG_SMP */
272
273#define latency_notifier_init(x) do { } while (0)
274
275#endif /* CONFIG_SMP */
276
277/**
278 * cpuidle_init - core initializer
279 */
280static int __init cpuidle_init(void)
281{
282 int ret;
283
284 pm_idle_old = pm_idle;
285
286 ret = cpuidle_add_class_sysfs(&cpu_sysdev_class);
287 if (ret)
288 return ret;
289
290 latency_notifier_init(&cpuidle_latency_notifier);
291
292 return 0;
293}
294
295core_initcall(cpuidle_init);
diff --git a/drivers/cpuidle/cpuidle.h b/drivers/cpuidle/cpuidle.h
new file mode 100644
index 000000000000..9476ba33ee2c
--- /dev/null
+++ b/drivers/cpuidle/cpuidle.h
@@ -0,0 +1,33 @@
1/*
2 * cpuidle.h - The internal header file
3 */
4
5#ifndef __DRIVER_CPUIDLE_H
6#define __DRIVER_CPUIDLE_H
7
8#include <linux/sysdev.h>
9
10/* For internal use only */
11extern struct cpuidle_governor *cpuidle_curr_governor;
12extern struct cpuidle_driver *cpuidle_curr_driver;
13extern struct list_head cpuidle_governors;
14extern struct list_head cpuidle_detected_devices;
15extern struct mutex cpuidle_lock;
16extern spinlock_t cpuidle_driver_lock;
17
18/* idle loop */
19extern void cpuidle_install_idle_handler(void);
20extern void cpuidle_uninstall_idle_handler(void);
21
22/* governors */
23extern int cpuidle_switch_governor(struct cpuidle_governor *gov);
24
25/* sysfs */
26extern int cpuidle_add_class_sysfs(struct sysdev_class *cls);
27extern void cpuidle_remove_class_sysfs(struct sysdev_class *cls);
28extern int cpuidle_add_state_sysfs(struct cpuidle_device *device);
29extern void cpuidle_remove_state_sysfs(struct cpuidle_device *device);
30extern int cpuidle_add_sysfs(struct sys_device *sysdev);
31extern void cpuidle_remove_sysfs(struct sys_device *sysdev);
32
33#endif /* __DRIVER_CPUIDLE_H */
diff --git a/drivers/cpuidle/driver.c b/drivers/cpuidle/driver.c
new file mode 100644
index 000000000000..2257004fe33d
--- /dev/null
+++ b/drivers/cpuidle/driver.c
@@ -0,0 +1,56 @@
1/*
2 * driver.c - driver support
3 *
4 * (C) 2006-2007 Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
5 * Shaohua Li <shaohua.li@intel.com>
6 * Adam Belay <abelay@novell.com>
7 *
8 * This code is licenced under the GPL.
9 */
10
11#include <linux/mutex.h>
12#include <linux/module.h>
13#include <linux/cpuidle.h>
14
15#include "cpuidle.h"
16
17struct cpuidle_driver *cpuidle_curr_driver;
18DEFINE_SPINLOCK(cpuidle_driver_lock);
19
20/**
21 * cpuidle_register_driver - registers a driver
22 * @drv: the driver
23 */
24int cpuidle_register_driver(struct cpuidle_driver *drv)
25{
26 if (!drv)
27 return -EINVAL;
28
29 spin_lock(&cpuidle_driver_lock);
30 if (cpuidle_curr_driver) {
31 spin_unlock(&cpuidle_driver_lock);
32 return -EBUSY;
33 }
34 cpuidle_curr_driver = drv;
35 spin_unlock(&cpuidle_driver_lock);
36
37 return 0;
38}
39
40EXPORT_SYMBOL_GPL(cpuidle_register_driver);
41
42/**
43 * cpuidle_unregister_driver - unregisters a driver
44 * @drv: the driver
45 */
46void cpuidle_unregister_driver(struct cpuidle_driver *drv)
47{
48 if (!drv)
49 return;
50
51 spin_lock(&cpuidle_driver_lock);
52 cpuidle_curr_driver = NULL;
53 spin_unlock(&cpuidle_driver_lock);
54}
55
56EXPORT_SYMBOL_GPL(cpuidle_unregister_driver);
diff --git a/drivers/cpuidle/governor.c b/drivers/cpuidle/governor.c
new file mode 100644
index 000000000000..bb699cb2dc5a
--- /dev/null
+++ b/drivers/cpuidle/governor.c
@@ -0,0 +1,141 @@
1/*
2 * governor.c - governor support
3 *
4 * (C) 2006-2007 Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
5 * Shaohua Li <shaohua.li@intel.com>
6 * Adam Belay <abelay@novell.com>
7 *
8 * This code is licenced under the GPL.
9 */
10
11#include <linux/mutex.h>
12#include <linux/module.h>
13#include <linux/cpuidle.h>
14
15#include "cpuidle.h"
16
17LIST_HEAD(cpuidle_governors);
18struct cpuidle_governor *cpuidle_curr_governor;
19
20/**
21 * __cpuidle_find_governor - finds a governor of the specified name
22 * @str: the name
23 *
24 * Must be called with cpuidle_lock aquired.
25 */
26static struct cpuidle_governor * __cpuidle_find_governor(const char *str)
27{
28 struct cpuidle_governor *gov;
29
30 list_for_each_entry(gov, &cpuidle_governors, governor_list)
31 if (!strnicmp(str, gov->name, CPUIDLE_NAME_LEN))
32 return gov;
33
34 return NULL;
35}
36
37/**
38 * cpuidle_switch_governor - changes the governor
39 * @gov: the new target governor
40 *
41 * NOTE: "gov" can be NULL to specify disabled
42 * Must be called with cpuidle_lock aquired.
43 */
44int cpuidle_switch_governor(struct cpuidle_governor *gov)
45{
46 struct cpuidle_device *dev;
47
48 if (gov == cpuidle_curr_governor)
49 return 0;
50
51 cpuidle_uninstall_idle_handler();
52
53 if (cpuidle_curr_governor) {
54 list_for_each_entry(dev, &cpuidle_detected_devices, device_list)
55 cpuidle_disable_device(dev);
56 module_put(cpuidle_curr_governor->owner);
57 }
58
59 cpuidle_curr_governor = gov;
60
61 if (gov) {
62 if (!try_module_get(cpuidle_curr_governor->owner))
63 return -EINVAL;
64 list_for_each_entry(dev, &cpuidle_detected_devices, device_list)
65 cpuidle_enable_device(dev);
66 cpuidle_install_idle_handler();
67 printk(KERN_INFO "cpuidle: using governor %s\n", gov->name);
68 }
69
70 return 0;
71}
72
73/**
74 * cpuidle_register_governor - registers a governor
75 * @gov: the governor
76 */
77int cpuidle_register_governor(struct cpuidle_governor *gov)
78{
79 int ret = -EEXIST;
80
81 if (!gov || !gov->select)
82 return -EINVAL;
83
84 mutex_lock(&cpuidle_lock);
85 if (__cpuidle_find_governor(gov->name) == NULL) {
86 ret = 0;
87 list_add_tail(&gov->governor_list, &cpuidle_governors);
88 if (!cpuidle_curr_governor ||
89 cpuidle_curr_governor->rating < gov->rating)
90 cpuidle_switch_governor(gov);
91 }
92 mutex_unlock(&cpuidle_lock);
93
94 return ret;
95}
96
97EXPORT_SYMBOL_GPL(cpuidle_register_governor);
98
99/**
100 * cpuidle_replace_governor - find a replacement governor
101 * @exclude_rating: the rating that will be skipped while looking for
102 * new governor.
103 */
104static struct cpuidle_governor *cpuidle_replace_governor(int exclude_rating)
105{
106 struct cpuidle_governor *gov;
107 struct cpuidle_governor *ret_gov = NULL;
108 unsigned int max_rating = 0;
109
110 list_for_each_entry(gov, &cpuidle_governors, governor_list) {
111 if (gov->rating == exclude_rating)
112 continue;
113 if (gov->rating > max_rating) {
114 max_rating = gov->rating;
115 ret_gov = gov;
116 }
117 }
118
119 return ret_gov;
120}
121
122/**
123 * cpuidle_unregister_governor - unregisters a governor
124 * @gov: the governor
125 */
126void cpuidle_unregister_governor(struct cpuidle_governor *gov)
127{
128 if (!gov)
129 return;
130
131 mutex_lock(&cpuidle_lock);
132 if (gov == cpuidle_curr_governor) {
133 struct cpuidle_governor *new_gov;
134 new_gov = cpuidle_replace_governor(gov->rating);
135 cpuidle_switch_governor(new_gov);
136 }
137 list_del(&gov->governor_list);
138 mutex_unlock(&cpuidle_lock);
139}
140
141EXPORT_SYMBOL_GPL(cpuidle_unregister_governor);
diff --git a/drivers/cpuidle/governors/Makefile b/drivers/cpuidle/governors/Makefile
new file mode 100644
index 000000000000..1b512722689f
--- /dev/null
+++ b/drivers/cpuidle/governors/Makefile
@@ -0,0 +1,6 @@
1#
2# Makefile for cpuidle governors.
3#
4
5obj-$(CONFIG_CPU_IDLE_GOV_LADDER) += ladder.o
6obj-$(CONFIG_CPU_IDLE_GOV_MENU) += menu.o
diff --git a/drivers/cpuidle/governors/ladder.c b/drivers/cpuidle/governors/ladder.c
new file mode 100644
index 000000000000..eb666ecae7c9
--- /dev/null
+++ b/drivers/cpuidle/governors/ladder.c
@@ -0,0 +1,166 @@
1/*
2 * ladder.c - the residency ladder algorithm
3 *
4 * Copyright (C) 2001, 2002 Andy Grover <andrew.grover@intel.com>
5 * Copyright (C) 2001, 2002 Paul Diefenbaugh <paul.s.diefenbaugh@intel.com>
6 * Copyright (C) 2004, 2005 Dominik Brodowski <linux@brodo.de>
7 *
8 * (C) 2006-2007 Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
9 * Shaohua Li <shaohua.li@intel.com>
10 * Adam Belay <abelay@novell.com>
11 *
12 * This code is licenced under the GPL.
13 */
14
15#include <linux/kernel.h>
16#include <linux/cpuidle.h>
17#include <linux/latency.h>
18#include <linux/moduleparam.h>
19#include <linux/jiffies.h>
20
21#include <asm/io.h>
22#include <asm/uaccess.h>
23
24#define PROMOTION_COUNT 4
25#define DEMOTION_COUNT 1
26
27struct ladder_device_state {
28 struct {
29 u32 promotion_count;
30 u32 demotion_count;
31 u32 promotion_time;
32 u32 demotion_time;
33 } threshold;
34 struct {
35 int promotion_count;
36 int demotion_count;
37 } stats;
38};
39
40struct ladder_device {
41 struct ladder_device_state states[CPUIDLE_STATE_MAX];
42 int last_state_idx;
43};
44
45static DEFINE_PER_CPU(struct ladder_device, ladder_devices);
46
47/**
48 * ladder_do_selection - prepares private data for a state change
49 * @ldev: the ladder device
50 * @old_idx: the current state index
51 * @new_idx: the new target state index
52 */
53static inline void ladder_do_selection(struct ladder_device *ldev,
54 int old_idx, int new_idx)
55{
56 ldev->states[old_idx].stats.promotion_count = 0;
57 ldev->states[old_idx].stats.demotion_count = 0;
58 ldev->last_state_idx = new_idx;
59}
60
61/**
62 * ladder_select_state - selects the next state to enter
63 * @dev: the CPU
64 */
65static int ladder_select_state(struct cpuidle_device *dev)
66{
67 struct ladder_device *ldev = &__get_cpu_var(ladder_devices);
68 struct ladder_device_state *last_state;
69 int last_residency, last_idx = ldev->last_state_idx;
70
71 if (unlikely(!ldev))
72 return 0;
73
74 last_state = &ldev->states[last_idx];
75
76 if (dev->states[last_idx].flags & CPUIDLE_FLAG_TIME_VALID)
77 last_residency = cpuidle_get_last_residency(dev) - dev->states[last_idx].exit_latency;
78 else
79 last_residency = last_state->threshold.promotion_time + 1;
80
81 /* consider promotion */
82 if (last_idx < dev->state_count - 1 &&
83 last_residency > last_state->threshold.promotion_time &&
84 dev->states[last_idx + 1].exit_latency <= system_latency_constraint()) {
85 last_state->stats.promotion_count++;
86 last_state->stats.demotion_count = 0;
87 if (last_state->stats.promotion_count >= last_state->threshold.promotion_count) {
88 ladder_do_selection(ldev, last_idx, last_idx + 1);
89 return last_idx + 1;
90 }
91 }
92
93 /* consider demotion */
94 if (last_idx > 0 &&
95 last_residency < last_state->threshold.demotion_time) {
96 last_state->stats.demotion_count++;
97 last_state->stats.promotion_count = 0;
98 if (last_state->stats.demotion_count >= last_state->threshold.demotion_count) {
99 ladder_do_selection(ldev, last_idx, last_idx - 1);
100 return last_idx - 1;
101 }
102 }
103
104 /* otherwise remain at the current state */
105 return last_idx;
106}
107
108/**
109 * ladder_enable_device - setup for the governor
110 * @dev: the CPU
111 */
112static int ladder_enable_device(struct cpuidle_device *dev)
113{
114 int i;
115 struct ladder_device *ldev = &per_cpu(ladder_devices, dev->cpu);
116 struct ladder_device_state *lstate;
117 struct cpuidle_state *state;
118
119 ldev->last_state_idx = 0;
120
121 for (i = 0; i < dev->state_count; i++) {
122 state = &dev->states[i];
123 lstate = &ldev->states[i];
124
125 lstate->stats.promotion_count = 0;
126 lstate->stats.demotion_count = 0;
127
128 lstate->threshold.promotion_count = PROMOTION_COUNT;
129 lstate->threshold.demotion_count = DEMOTION_COUNT;
130
131 if (i < dev->state_count - 1)
132 lstate->threshold.promotion_time = state->exit_latency;
133 if (i > 0)
134 lstate->threshold.demotion_time = state->exit_latency;
135 }
136
137 return 0;
138}
139
140static struct cpuidle_governor ladder_governor = {
141 .name = "ladder",
142 .rating = 10,
143 .enable = ladder_enable_device,
144 .select = ladder_select_state,
145 .owner = THIS_MODULE,
146};
147
148/**
149 * init_ladder - initializes the governor
150 */
151static int __init init_ladder(void)
152{
153 return cpuidle_register_governor(&ladder_governor);
154}
155
156/**
157 * exit_ladder - exits the governor
158 */
159static void __exit exit_ladder(void)
160{
161 cpuidle_unregister_governor(&ladder_governor);
162}
163
164MODULE_LICENSE("GPL");
165module_init(init_ladder);
166module_exit(exit_ladder);
diff --git a/drivers/cpuidle/governors/menu.c b/drivers/cpuidle/governors/menu.c
new file mode 100644
index 000000000000..299d45c3bdd2
--- /dev/null
+++ b/drivers/cpuidle/governors/menu.c
@@ -0,0 +1,137 @@
1/*
2 * menu.c - the menu idle governor
3 *
4 * Copyright (C) 2006-2007 Adam Belay <abelay@novell.com>
5 *
6 * This code is licenced under the GPL.
7 */
8
9#include <linux/kernel.h>
10#include <linux/cpuidle.h>
11#include <linux/latency.h>
12#include <linux/time.h>
13#include <linux/ktime.h>
14#include <linux/hrtimer.h>
15#include <linux/tick.h>
16
17#define BREAK_FUZZ 4 /* 4 us */
18
19struct menu_device {
20 int last_state_idx;
21
22 unsigned int expected_us;
23 unsigned int predicted_us;
24 unsigned int last_measured_us;
25 unsigned int elapsed_us;
26};
27
28static DEFINE_PER_CPU(struct menu_device, menu_devices);
29
30/**
31 * menu_select - selects the next idle state to enter
32 * @dev: the CPU
33 */
34static int menu_select(struct cpuidle_device *dev)
35{
36 struct menu_device *data = &__get_cpu_var(menu_devices);
37 int i;
38
39 /* determine the expected residency time */
40 data->expected_us =
41 (u32) ktime_to_ns(tick_nohz_get_sleep_length()) / 1000;
42
43 /* find the deepest idle state that satisfies our constraints */
44 for (i = 1; i < dev->state_count; i++) {
45 struct cpuidle_state *s = &dev->states[i];
46
47 if (s->target_residency > data->expected_us)
48 break;
49 if (s->target_residency > data->predicted_us)
50 break;
51 if (s->exit_latency > system_latency_constraint())
52 break;
53 }
54
55 data->last_state_idx = i - 1;
56 return i - 1;
57}
58
59/**
60 * menu_reflect - attempts to guess what happened after entry
61 * @dev: the CPU
62 *
63 * NOTE: it's important to be fast here because this operation will add to
64 * the overall exit latency.
65 */
66static void menu_reflect(struct cpuidle_device *dev)
67{
68 struct menu_device *data = &__get_cpu_var(menu_devices);
69 int last_idx = data->last_state_idx;
70 unsigned int measured_us =
71 cpuidle_get_last_residency(dev) + data->elapsed_us;
72 struct cpuidle_state *target = &dev->states[last_idx];
73
74 /*
75 * Ugh, this idle state doesn't support residency measurements, so we
76 * are basically lost in the dark. As a compromise, assume we slept
77 * for one full standard timer tick. However, be aware that this
78 * could potentially result in a suboptimal state transition.
79 */
80 if (!(target->flags & CPUIDLE_FLAG_TIME_VALID))
81 measured_us = USEC_PER_SEC / HZ;
82
83 /* Predict time remaining until next break event */
84 if (measured_us + BREAK_FUZZ < data->expected_us - target->exit_latency) {
85 data->predicted_us = max(measured_us, data->last_measured_us);
86 data->last_measured_us = measured_us;
87 data->elapsed_us = 0;
88 } else {
89 if (data->elapsed_us < data->elapsed_us + measured_us)
90 data->elapsed_us = measured_us;
91 else
92 data->elapsed_us = -1;
93 data->predicted_us = max(measured_us, data->last_measured_us);
94 }
95}
96
97/**
98 * menu_enable_device - scans a CPU's states and does setup
99 * @dev: the CPU
100 */
101static int menu_enable_device(struct cpuidle_device *dev)
102{
103 struct menu_device *data = &per_cpu(menu_devices, dev->cpu);
104
105 memset(data, 0, sizeof(struct menu_device));
106
107 return 0;
108}
109
110static struct cpuidle_governor menu_governor = {
111 .name = "menu",
112 .rating = 20,
113 .enable = menu_enable_device,
114 .select = menu_select,
115 .reflect = menu_reflect,
116 .owner = THIS_MODULE,
117};
118
119/**
120 * init_menu - initializes the governor
121 */
122static int __init init_menu(void)
123{
124 return cpuidle_register_governor(&menu_governor);
125}
126
127/**
128 * exit_menu - exits the governor
129 */
130static void __exit exit_menu(void)
131{
132 cpuidle_unregister_governor(&menu_governor);
133}
134
135MODULE_LICENSE("GPL");
136module_init(init_menu);
137module_exit(exit_menu);
diff --git a/drivers/cpuidle/sysfs.c b/drivers/cpuidle/sysfs.c
new file mode 100644
index 000000000000..0f3515e77d4b
--- /dev/null
+++ b/drivers/cpuidle/sysfs.c
@@ -0,0 +1,361 @@
1/*
2 * sysfs.c - sysfs support
3 *
4 * (C) 2006-2007 Shaohua Li <shaohua.li@intel.com>
5 *
6 * This code is licenced under the GPL.
7 */
8
9#include <linux/kernel.h>
10#include <linux/cpuidle.h>
11#include <linux/sysfs.h>
12#include <linux/cpu.h>
13
14#include "cpuidle.h"
15
16static unsigned int sysfs_switch;
17static int __init cpuidle_sysfs_setup(char *unused)
18{
19 sysfs_switch = 1;
20 return 1;
21}
22__setup("cpuidle_sysfs_switch", cpuidle_sysfs_setup);
23
24static ssize_t show_available_governors(struct sys_device *dev, char *buf)
25{
26 ssize_t i = 0;
27 struct cpuidle_governor *tmp;
28
29 mutex_lock(&cpuidle_lock);
30 list_for_each_entry(tmp, &cpuidle_governors, governor_list) {
31 if (i >= (ssize_t) ((PAGE_SIZE/sizeof(char)) - CPUIDLE_NAME_LEN - 2))
32 goto out;
33 i += scnprintf(&buf[i], CPUIDLE_NAME_LEN, "%s ", tmp->name);
34 }
35
36out:
37 i+= sprintf(&buf[i], "\n");
38 mutex_unlock(&cpuidle_lock);
39 return i;
40}
41
42static ssize_t show_current_driver(struct sys_device *dev, char *buf)
43{
44 ssize_t ret;
45
46 spin_lock(&cpuidle_driver_lock);
47 if (cpuidle_curr_driver)
48 ret = sprintf(buf, "%s\n", cpuidle_curr_driver->name);
49 else
50 ret = sprintf(buf, "none\n");
51 spin_unlock(&cpuidle_driver_lock);
52
53 return ret;
54}
55
56static ssize_t show_current_governor(struct sys_device *dev, char *buf)
57{
58 ssize_t ret;
59
60 mutex_lock(&cpuidle_lock);
61 if (cpuidle_curr_governor)
62 ret = sprintf(buf, "%s\n", cpuidle_curr_governor->name);
63 else
64 ret = sprintf(buf, "none\n");
65 mutex_unlock(&cpuidle_lock);
66
67 return ret;
68}
69
70static ssize_t store_current_governor(struct sys_device *dev,
71 const char *buf, size_t count)
72{
73 char gov_name[CPUIDLE_NAME_LEN];
74 int ret = -EINVAL;
75 size_t len = count;
76 struct cpuidle_governor *gov;
77
78 if (!len || len >= sizeof(gov_name))
79 return -EINVAL;
80
81 memcpy(gov_name, buf, len);
82 gov_name[len] = '\0';
83 if (gov_name[len - 1] == '\n')
84 gov_name[--len] = '\0';
85
86 mutex_lock(&cpuidle_lock);
87
88 list_for_each_entry(gov, &cpuidle_governors, governor_list) {
89 if (strlen(gov->name) == len && !strcmp(gov->name, gov_name)) {
90 ret = cpuidle_switch_governor(gov);
91 break;
92 }
93 }
94
95 mutex_unlock(&cpuidle_lock);
96
97 if (ret)
98 return ret;
99 else
100 return count;
101}
102
103static SYSDEV_ATTR(current_driver, 0444, show_current_driver, NULL);
104static SYSDEV_ATTR(current_governor_ro, 0444, show_current_governor, NULL);
105
106static struct attribute *cpuclass_default_attrs[] = {
107 &attr_current_driver.attr,
108 &attr_current_governor_ro.attr,
109 NULL
110};
111
112static SYSDEV_ATTR(available_governors, 0444, show_available_governors, NULL);
113static SYSDEV_ATTR(current_governor, 0644, show_current_governor,
114 store_current_governor);
115
116static struct attribute *cpuclass_switch_attrs[] = {
117 &attr_available_governors.attr,
118 &attr_current_driver.attr,
119 &attr_current_governor.attr,
120 NULL
121};
122
123static struct attribute_group cpuclass_attr_group = {
124 .attrs = cpuclass_default_attrs,
125 .name = "cpuidle",
126};
127
128/**
129 * cpuidle_add_class_sysfs - add CPU global sysfs attributes
130 */
131int cpuidle_add_class_sysfs(struct sysdev_class *cls)
132{
133 if (sysfs_switch)
134 cpuclass_attr_group.attrs = cpuclass_switch_attrs;
135
136 return sysfs_create_group(&cls->kset.kobj, &cpuclass_attr_group);
137}
138
139/**
140 * cpuidle_remove_class_sysfs - remove CPU global sysfs attributes
141 */
142void cpuidle_remove_class_sysfs(struct sysdev_class *cls)
143{
144 sysfs_remove_group(&cls->kset.kobj, &cpuclass_attr_group);
145}
146
147struct cpuidle_attr {
148 struct attribute attr;
149 ssize_t (*show)(struct cpuidle_device *, char *);
150 ssize_t (*store)(struct cpuidle_device *, const char *, size_t count);
151};
152
153#define define_one_ro(_name, show) \
154 static struct cpuidle_attr attr_##_name = __ATTR(_name, 0444, show, NULL)
155#define define_one_rw(_name, show, store) \
156 static struct cpuidle_attr attr_##_name = __ATTR(_name, 0644, show, store)
157
158#define kobj_to_cpuidledev(k) container_of(k, struct cpuidle_device, kobj)
159#define attr_to_cpuidleattr(a) container_of(a, struct cpuidle_attr, attr)
160static ssize_t cpuidle_show(struct kobject * kobj, struct attribute * attr ,char * buf)
161{
162 int ret = -EIO;
163 struct cpuidle_device *dev = kobj_to_cpuidledev(kobj);
164 struct cpuidle_attr * cattr = attr_to_cpuidleattr(attr);
165
166 if (cattr->show) {
167 mutex_lock(&cpuidle_lock);
168 ret = cattr->show(dev, buf);
169 mutex_unlock(&cpuidle_lock);
170 }
171 return ret;
172}
173
174static ssize_t cpuidle_store(struct kobject * kobj, struct attribute * attr,
175 const char * buf, size_t count)
176{
177 int ret = -EIO;
178 struct cpuidle_device *dev = kobj_to_cpuidledev(kobj);
179 struct cpuidle_attr * cattr = attr_to_cpuidleattr(attr);
180
181 if (cattr->store) {
182 mutex_lock(&cpuidle_lock);
183 ret = cattr->store(dev, buf, count);
184 mutex_unlock(&cpuidle_lock);
185 }
186 return ret;
187}
188
189static struct sysfs_ops cpuidle_sysfs_ops = {
190 .show = cpuidle_show,
191 .store = cpuidle_store,
192};
193
194static void cpuidle_sysfs_release(struct kobject *kobj)
195{
196 struct cpuidle_device *dev = kobj_to_cpuidledev(kobj);
197
198 complete(&dev->kobj_unregister);
199}
200
201static struct kobj_type ktype_cpuidle = {
202 .sysfs_ops = &cpuidle_sysfs_ops,
203 .release = cpuidle_sysfs_release,
204};
205
206struct cpuidle_state_attr {
207 struct attribute attr;
208 ssize_t (*show)(struct cpuidle_state *, char *);
209 ssize_t (*store)(struct cpuidle_state *, const char *, size_t);
210};
211
212#define define_one_state_ro(_name, show) \
213static struct cpuidle_state_attr attr_##_name = __ATTR(_name, 0444, show, NULL)
214
215#define define_show_state_function(_name) \
216static ssize_t show_state_##_name(struct cpuidle_state *state, char *buf) \
217{ \
218 return sprintf(buf, "%u\n", state->_name);\
219}
220
221static ssize_t show_state_name(struct cpuidle_state *state, char *buf)
222{
223 return sprintf(buf, "%s\n", state->name);
224}
225
226define_show_state_function(exit_latency)
227define_show_state_function(power_usage)
228define_show_state_function(usage)
229define_show_state_function(time)
230define_one_state_ro(name, show_state_name);
231define_one_state_ro(latency, show_state_exit_latency);
232define_one_state_ro(power, show_state_power_usage);
233define_one_state_ro(usage, show_state_usage);
234define_one_state_ro(time, show_state_time);
235
236static struct attribute *cpuidle_state_default_attrs[] = {
237 &attr_name.attr,
238 &attr_latency.attr,
239 &attr_power.attr,
240 &attr_usage.attr,
241 &attr_time.attr,
242 NULL
243};
244
245#define kobj_to_state_obj(k) container_of(k, struct cpuidle_state_kobj, kobj)
246#define kobj_to_state(k) (kobj_to_state_obj(k)->state)
247#define attr_to_stateattr(a) container_of(a, struct cpuidle_state_attr, attr)
248static ssize_t cpuidle_state_show(struct kobject * kobj,
249 struct attribute * attr ,char * buf)
250{
251 int ret = -EIO;
252 struct cpuidle_state *state = kobj_to_state(kobj);
253 struct cpuidle_state_attr * cattr = attr_to_stateattr(attr);
254
255 if (cattr->show)
256 ret = cattr->show(state, buf);
257
258 return ret;
259}
260
261static struct sysfs_ops cpuidle_state_sysfs_ops = {
262 .show = cpuidle_state_show,
263};
264
265static void cpuidle_state_sysfs_release(struct kobject *kobj)
266{
267 struct cpuidle_state_kobj *state_obj = kobj_to_state_obj(kobj);
268
269 complete(&state_obj->kobj_unregister);
270}
271
272static struct kobj_type ktype_state_cpuidle = {
273 .sysfs_ops = &cpuidle_state_sysfs_ops,
274 .default_attrs = cpuidle_state_default_attrs,
275 .release = cpuidle_state_sysfs_release,
276};
277
278static void inline cpuidle_free_state_kobj(struct cpuidle_device *device, int i)
279{
280 kobject_unregister(&device->kobjs[i]->kobj);
281 wait_for_completion(&device->kobjs[i]->kobj_unregister);
282 kfree(device->kobjs[i]);
283 device->kobjs[i] = NULL;
284}
285
286/**
287 * cpuidle_add_driver_sysfs - adds driver-specific sysfs attributes
288 * @device: the target device
289 */
290int cpuidle_add_state_sysfs(struct cpuidle_device *device)
291{
292 int i, ret = -ENOMEM;
293 struct cpuidle_state_kobj *kobj;
294
295 /* state statistics */
296 for (i = 0; i < device->state_count; i++) {
297 kobj = kzalloc(sizeof(struct cpuidle_state_kobj), GFP_KERNEL);
298 if (!kobj)
299 goto error_state;
300 kobj->state = &device->states[i];
301 init_completion(&kobj->kobj_unregister);
302
303 kobj->kobj.parent = &device->kobj;
304 kobj->kobj.ktype = &ktype_state_cpuidle;
305 kobject_set_name(&kobj->kobj, "state%d", i);
306 ret = kobject_register(&kobj->kobj);
307 if (ret) {
308 kfree(kobj);
309 goto error_state;
310 }
311 device->kobjs[i] = kobj;
312 }
313
314 return 0;
315
316error_state:
317 for (i = i - 1; i >= 0; i--)
318 cpuidle_free_state_kobj(device, i);
319 return ret;
320}
321
322/**
323 * cpuidle_remove_driver_sysfs - removes driver-specific sysfs attributes
324 * @device: the target device
325 */
326void cpuidle_remove_state_sysfs(struct cpuidle_device *device)
327{
328 int i;
329
330 for (i = 0; i < device->state_count; i++)
331 cpuidle_free_state_kobj(device, i);
332}
333
334/**
335 * cpuidle_add_sysfs - creates a sysfs instance for the target device
336 * @sysdev: the target device
337 */
338int cpuidle_add_sysfs(struct sys_device *sysdev)
339{
340 int cpu = sysdev->id;
341 struct cpuidle_device *dev;
342
343 dev = per_cpu(cpuidle_devices, cpu);
344 dev->kobj.parent = &sysdev->kobj;
345 dev->kobj.ktype = &ktype_cpuidle;
346 kobject_set_name(&dev->kobj, "%s", "cpuidle");
347 return kobject_register(&dev->kobj);
348}
349
350/**
351 * cpuidle_remove_sysfs - deletes a sysfs instance on the target device
352 * @sysdev: the target device
353 */
354void cpuidle_remove_sysfs(struct sys_device *sysdev)
355{
356 int cpu = sysdev->id;
357 struct cpuidle_device *dev;
358
359 dev = per_cpu(cpuidle_devices, cpu);
360 kobject_unregister(&dev->kobj);
361}
diff --git a/drivers/net/wireless/ipw2100.c b/drivers/net/wireless/ipw2100.c
index 8990585bd228..e8a94b746295 100644
--- a/drivers/net/wireless/ipw2100.c
+++ b/drivers/net/wireless/ipw2100.c
@@ -1858,14 +1858,6 @@ static void ipw2100_down(struct ipw2100_priv *priv)
1858 1858
1859 modify_acceptable_latency("ipw2100", INFINITE_LATENCY); 1859 modify_acceptable_latency("ipw2100", INFINITE_LATENCY);
1860 1860
1861#ifdef ACPI_CSTATE_LIMIT_DEFINED
1862 if (priv->config & CFG_C3_DISABLED) {
1863 IPW_DEBUG_INFO(": Resetting C3 transitions.\n");
1864 acpi_set_cstate_limit(priv->cstate_limit);
1865 priv->config &= ~CFG_C3_DISABLED;
1866 }
1867#endif
1868
1869 /* We have to signal any supplicant if we are disassociating */ 1861 /* We have to signal any supplicant if we are disassociating */
1870 if (associated) 1862 if (associated)
1871 wireless_send_event(priv->net_dev, SIOCGIWAP, &wrqu, NULL); 1863 wireless_send_event(priv->net_dev, SIOCGIWAP, &wrqu, NULL);
@@ -2088,14 +2080,6 @@ static void isr_indicate_rf_kill(struct ipw2100_priv *priv, u32 status)
2088 /* RF_KILL is now enabled (else we wouldn't be here) */ 2080 /* RF_KILL is now enabled (else we wouldn't be here) */
2089 priv->status |= STATUS_RF_KILL_HW; 2081 priv->status |= STATUS_RF_KILL_HW;
2090 2082
2091#ifdef ACPI_CSTATE_LIMIT_DEFINED
2092 if (priv->config & CFG_C3_DISABLED) {
2093 IPW_DEBUG_INFO(": Resetting C3 transitions.\n");
2094 acpi_set_cstate_limit(priv->cstate_limit);
2095 priv->config &= ~CFG_C3_DISABLED;
2096 }
2097#endif
2098
2099 /* Make sure the RF Kill check timer is running */ 2083 /* Make sure the RF Kill check timer is running */
2100 priv->stop_rf_kill = 0; 2084 priv->stop_rf_kill = 0;
2101 cancel_delayed_work(&priv->rf_kill); 2085 cancel_delayed_work(&priv->rf_kill);
@@ -2326,23 +2310,10 @@ static void ipw2100_corruption_detected(struct ipw2100_priv *priv, int i)
2326 u32 match, reg; 2310 u32 match, reg;
2327 int j; 2311 int j;
2328#endif 2312#endif
2329#ifdef ACPI_CSTATE_LIMIT_DEFINED
2330 int limit;
2331#endif
2332 2313
2333 IPW_DEBUG_INFO(": PCI latency error detected at 0x%04zX.\n", 2314 IPW_DEBUG_INFO(": PCI latency error detected at 0x%04zX.\n",
2334 i * sizeof(struct ipw2100_status)); 2315 i * sizeof(struct ipw2100_status));
2335 2316
2336#ifdef ACPI_CSTATE_LIMIT_DEFINED
2337 IPW_DEBUG_INFO(": Disabling C3 transitions.\n");
2338 limit = acpi_get_cstate_limit();
2339 if (limit > 2) {
2340 priv->cstate_limit = limit;
2341 acpi_set_cstate_limit(2);
2342 priv->config |= CFG_C3_DISABLED;
2343 }
2344#endif
2345
2346#ifdef IPW2100_DEBUG_C3 2317#ifdef IPW2100_DEBUG_C3
2347 /* Halt the fimrware so we can get a good image */ 2318 /* Halt the fimrware so we can get a good image */
2348 write_register(priv->net_dev, IPW_REG_RESET_REG, 2319 write_register(priv->net_dev, IPW_REG_RESET_REG,
diff --git a/drivers/net/wireless/ipw2100.h b/drivers/net/wireless/ipw2100.h
index de7d384d38af..2b8be2418fa7 100644
--- a/drivers/net/wireless/ipw2100.h
+++ b/drivers/net/wireless/ipw2100.h
@@ -479,7 +479,6 @@ enum {
479#define CFG_ASSOCIATE (1<<6) 479#define CFG_ASSOCIATE (1<<6)
480#define CFG_FIXED_RATE (1<<7) 480#define CFG_FIXED_RATE (1<<7)
481#define CFG_ADHOC_CREATE (1<<8) 481#define CFG_ADHOC_CREATE (1<<8)
482#define CFG_C3_DISABLED (1<<9)
483#define CFG_PASSIVE_SCAN (1<<10) 482#define CFG_PASSIVE_SCAN (1<<10)
484#ifdef CONFIG_IPW2100_MONITOR 483#ifdef CONFIG_IPW2100_MONITOR
485#define CFG_CRC_CHECK (1<<11) 484#define CFG_CRC_CHECK (1<<11)
@@ -508,7 +507,6 @@ struct ipw2100_priv {
508 u8 bssid[ETH_ALEN]; 507 u8 bssid[ETH_ALEN];
509 u8 channel; 508 u8 channel;
510 int last_mode; 509 int last_mode;
511 int cstate_limit;
512 510
513 unsigned long connect_start; 511 unsigned long connect_start;
514 unsigned long last_reset; 512 unsigned long last_reset;
diff --git a/include/acpi/processor.h b/include/acpi/processor.h
index 99934a999e66..26d79f6db8a0 100644
--- a/include/acpi/processor.h
+++ b/include/acpi/processor.h
@@ -3,6 +3,7 @@
3 3
4#include <linux/kernel.h> 4#include <linux/kernel.h>
5#include <linux/cpu.h> 5#include <linux/cpu.h>
6#include <linux/cpuidle.h>
6 7
7#include <asm/acpi.h> 8#include <asm/acpi.h>
8 9
@@ -75,7 +76,9 @@ struct acpi_processor_cx {
75}; 76};
76 77
77struct acpi_processor_power { 78struct acpi_processor_power {
79 struct cpuidle_device dev;
78 struct acpi_processor_cx *state; 80 struct acpi_processor_cx *state;
81 struct acpi_processor_cx *bm_state;
79 unsigned long bm_check_timestamp; 82 unsigned long bm_check_timestamp;
80 u32 default_state; 83 u32 default_state;
81 u32 bm_activity; 84 u32 bm_activity;
@@ -199,6 +202,7 @@ struct acpi_processor_flags {
199 u8 bm_check:1; 202 u8 bm_check:1;
200 u8 has_cst:1; 203 u8 has_cst:1;
201 u8 power_setup_done:1; 204 u8 power_setup_done:1;
205 u8 bm_rld_set:1;
202}; 206};
203 207
204struct acpi_processor { 208struct acpi_processor {
@@ -322,6 +326,7 @@ int acpi_processor_power_exit(struct acpi_processor *pr,
322 struct acpi_device *device); 326 struct acpi_device *device);
323int acpi_processor_suspend(struct acpi_device * device, pm_message_t state); 327int acpi_processor_suspend(struct acpi_device * device, pm_message_t state);
324int acpi_processor_resume(struct acpi_device * device); 328int acpi_processor_resume(struct acpi_device * device);
329extern struct cpuidle_driver acpi_idle_driver;
325 330
326/* in processor_thermal.c */ 331/* in processor_thermal.c */
327int acpi_processor_get_limit_info(struct acpi_processor *pr); 332int acpi_processor_get_limit_info(struct acpi_processor *pr);
diff --git a/include/linux/acpi.h b/include/linux/acpi.h
index bf5e0009de75..8ccedf7a0a5a 100644
--- a/include/linux/acpi.h
+++ b/include/linux/acpi.h
@@ -189,32 +189,6 @@ extern int ec_transaction(u8 command,
189extern int acpi_blacklisted(void); 189extern int acpi_blacklisted(void);
190extern void acpi_bios_year(char *s); 190extern void acpi_bios_year(char *s);
191 191
192#define ACPI_CSTATE_LIMIT_DEFINED /* for driver builds */
193#ifdef CONFIG_ACPI
194
195/*
196 * Set highest legal C-state
197 * 0: C0 okay, but not C1
198 * 1: C1 okay, but not C2
199 * 2: C2 okay, but not C3 etc.
200 */
201
202extern unsigned int max_cstate;
203
204static inline unsigned int acpi_get_cstate_limit(void)
205{
206 return max_cstate;
207}
208static inline void acpi_set_cstate_limit(unsigned int new_limit)
209{
210 max_cstate = new_limit;
211 return;
212}
213#else
214static inline unsigned int acpi_get_cstate_limit(void) { return 0; }
215static inline void acpi_set_cstate_limit(unsigned int new_limit) { return; }
216#endif
217
218#ifdef CONFIG_ACPI_NUMA 192#ifdef CONFIG_ACPI_NUMA
219int acpi_get_pxm(acpi_handle handle); 193int acpi_get_pxm(acpi_handle handle);
220int acpi_get_node(acpi_handle *handle); 194int acpi_get_node(acpi_handle *handle);
diff --git a/include/linux/cpuidle.h b/include/linux/cpuidle.h
new file mode 100644
index 000000000000..16a51546db44
--- /dev/null
+++ b/include/linux/cpuidle.h
@@ -0,0 +1,180 @@
1/*
2 * cpuidle.h - a generic framework for CPU idle power management
3 *
4 * (C) 2007 Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
5 * Shaohua Li <shaohua.li@intel.com>
6 * Adam Belay <abelay@novell.com>
7 *
8 * This code is licenced under the GPL.
9 */
10
11#ifndef _LINUX_CPUIDLE_H
12#define _LINUX_CPUIDLE_H
13
14#include <linux/percpu.h>
15#include <linux/list.h>
16#include <linux/module.h>
17#include <linux/kobject.h>
18#include <linux/completion.h>
19
20#define CPUIDLE_STATE_MAX 8
21#define CPUIDLE_NAME_LEN 16
22
23struct cpuidle_device;
24
25
26/****************************
27 * CPUIDLE DEVICE INTERFACE *
28 ****************************/
29
30struct cpuidle_state {
31 char name[CPUIDLE_NAME_LEN];
32 void *driver_data;
33
34 unsigned int flags;
35 unsigned int exit_latency; /* in US */
36 unsigned int power_usage; /* in mW */
37 unsigned int target_residency; /* in US */
38
39 unsigned int usage;
40 unsigned int time; /* in US */
41
42 int (*enter) (struct cpuidle_device *dev,
43 struct cpuidle_state *state);
44};
45
46/* Idle State Flags */
47#define CPUIDLE_FLAG_TIME_VALID (0x01) /* is residency time measurable? */
48#define CPUIDLE_FLAG_CHECK_BM (0x02) /* BM activity will exit state */
49#define CPUIDLE_FLAG_SHALLOW (0x10) /* low latency, minimal savings */
50#define CPUIDLE_FLAG_BALANCED (0x20) /* medium latency, moderate savings */
51#define CPUIDLE_FLAG_DEEP (0x40) /* high latency, large savings */
52
53#define CPUIDLE_DRIVER_FLAGS_MASK (0xFFFF0000)
54
55/**
56 * cpuidle_get_statedata - retrieves private driver state data
57 * @state: the state
58 */
59static inline void * cpuidle_get_statedata(struct cpuidle_state *state)
60{
61 return state->driver_data;
62}
63
64/**
65 * cpuidle_set_statedata - stores private driver state data
66 * @state: the state
67 * @data: the private data
68 */
69static inline void
70cpuidle_set_statedata(struct cpuidle_state *state, void *data)
71{
72 state->driver_data = data;
73}
74
75struct cpuidle_state_kobj {
76 struct cpuidle_state *state;
77 struct completion kobj_unregister;
78 struct kobject kobj;
79};
80
81struct cpuidle_device {
82 int enabled:1;
83 unsigned int cpu;
84
85 int last_residency;
86 int state_count;
87 struct cpuidle_state states[CPUIDLE_STATE_MAX];
88 struct cpuidle_state_kobj *kobjs[CPUIDLE_STATE_MAX];
89 struct cpuidle_state *last_state;
90
91 struct list_head device_list;
92 struct kobject kobj;
93 struct completion kobj_unregister;
94 void *governor_data;
95};
96
97DECLARE_PER_CPU(struct cpuidle_device *, cpuidle_devices);
98
99/**
100 * cpuidle_get_last_residency - retrieves the last state's residency time
101 * @dev: the target CPU
102 *
103 * NOTE: this value is invalid if CPUIDLE_FLAG_TIME_VALID isn't set
104 */
105static inline int cpuidle_get_last_residency(struct cpuidle_device *dev)
106{
107 return dev->last_residency;
108}
109
110
111/****************************
112 * CPUIDLE DRIVER INTERFACE *
113 ****************************/
114
115struct cpuidle_driver {
116 char name[CPUIDLE_NAME_LEN];
117 struct module *owner;
118};
119
120#ifdef CONFIG_CPU_IDLE
121
122extern int cpuidle_register_driver(struct cpuidle_driver *drv);
123extern void cpuidle_unregister_driver(struct cpuidle_driver *drv);
124extern int cpuidle_register_device(struct cpuidle_device *dev);
125extern void cpuidle_unregister_device(struct cpuidle_device *dev);
126
127extern void cpuidle_pause_and_lock(void);
128extern void cpuidle_resume_and_unlock(void);
129extern int cpuidle_enable_device(struct cpuidle_device *dev);
130extern void cpuidle_disable_device(struct cpuidle_device *dev);
131
132#else
133
134static inline int cpuidle_register_driver(struct cpuidle_driver *drv)
135{return 0;}
136static inline void cpuidle_unregister_driver(struct cpuidle_driver *drv) { }
137static inline int cpuidle_register_device(struct cpuidle_device *dev)
138{return 0;}
139static inline void cpuidle_unregister_device(struct cpuidle_device *dev) { }
140
141static inline void cpuidle_pause_and_lock(void) { }
142static inline void cpuidle_resume_and_unlock(void) { }
143static inline int cpuidle_enable_device(struct cpuidle_device *dev)
144{return 0;}
145static inline void cpuidle_disable_device(struct cpuidle_device *dev) { }
146
147#endif
148
149/******************************
150 * CPUIDLE GOVERNOR INTERFACE *
151 ******************************/
152
153struct cpuidle_governor {
154 char name[CPUIDLE_NAME_LEN];
155 struct list_head governor_list;
156 unsigned int rating;
157
158 int (*enable) (struct cpuidle_device *dev);
159 void (*disable) (struct cpuidle_device *dev);
160
161 int (*select) (struct cpuidle_device *dev);
162 void (*reflect) (struct cpuidle_device *dev);
163
164 struct module *owner;
165};
166
167#ifdef CONFIG_CPU_IDLE
168
169extern int cpuidle_register_governor(struct cpuidle_governor *gov);
170extern void cpuidle_unregister_governor(struct cpuidle_governor *gov);
171
172#else
173
174static inline int cpuidle_register_governor(struct cpuidle_governor *gov)
175{return 0;}
176static inline void cpuidle_unregister_governor(struct cpuidle_governor *gov) { }
177
178#endif
179
180#endif /* _LINUX_CPUIDLE_H */
diff --git a/include/linux/tick.h b/include/linux/tick.h
index 9a7252e089b9..f4a1395e05ff 100644
--- a/include/linux/tick.h
+++ b/include/linux/tick.h
@@ -40,6 +40,7 @@ enum tick_nohz_mode {
40 * @idle_sleeps: Number of idle calls, where the sched tick was stopped 40 * @idle_sleeps: Number of idle calls, where the sched tick was stopped
41 * @idle_entrytime: Time when the idle call was entered 41 * @idle_entrytime: Time when the idle call was entered
42 * @idle_sleeptime: Sum of the time slept in idle with sched tick stopped 42 * @idle_sleeptime: Sum of the time slept in idle with sched tick stopped
43 * @sleep_length: Duration of the current idle sleep
43 */ 44 */
44struct tick_sched { 45struct tick_sched {
45 struct hrtimer sched_timer; 46 struct hrtimer sched_timer;
@@ -52,6 +53,7 @@ struct tick_sched {
52 unsigned long idle_sleeps; 53 unsigned long idle_sleeps;
53 ktime_t idle_entrytime; 54 ktime_t idle_entrytime;
54 ktime_t idle_sleeptime; 55 ktime_t idle_sleeptime;
56 ktime_t sleep_length;
55 unsigned long last_jiffies; 57 unsigned long last_jiffies;
56 unsigned long next_jiffies; 58 unsigned long next_jiffies;
57 ktime_t idle_expires; 59 ktime_t idle_expires;
@@ -100,10 +102,17 @@ static inline int tick_check_oneshot_change(int allow_nohz) { return 0; }
100extern void tick_nohz_stop_sched_tick(void); 102extern void tick_nohz_stop_sched_tick(void);
101extern void tick_nohz_restart_sched_tick(void); 103extern void tick_nohz_restart_sched_tick(void);
102extern void tick_nohz_update_jiffies(void); 104extern void tick_nohz_update_jiffies(void);
105extern ktime_t tick_nohz_get_sleep_length(void);
103# else 106# else
104static inline void tick_nohz_stop_sched_tick(void) { } 107static inline void tick_nohz_stop_sched_tick(void) { }
105static inline void tick_nohz_restart_sched_tick(void) { } 108static inline void tick_nohz_restart_sched_tick(void) { }
106static inline void tick_nohz_update_jiffies(void) { } 109static inline void tick_nohz_update_jiffies(void) { }
110static inline ktime_t tick_nohz_get_sleep_length(void)
111{
112 ktime_t len = { .tv64 = NSEC_PER_SEC/HZ };
113
114 return len;
115}
107# endif /* !NO_HZ */ 116# endif /* !NO_HZ */
108 117
109#endif 118#endif
diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c
index 8c3fef1db09c..637519af6151 100644
--- a/kernel/time/tick-sched.c
+++ b/kernel/time/tick-sched.c
@@ -153,6 +153,7 @@ void tick_nohz_stop_sched_tick(void)
153 unsigned long seq, last_jiffies, next_jiffies, delta_jiffies, flags; 153 unsigned long seq, last_jiffies, next_jiffies, delta_jiffies, flags;
154 struct tick_sched *ts; 154 struct tick_sched *ts;
155 ktime_t last_update, expires, now, delta; 155 ktime_t last_update, expires, now, delta;
156 struct clock_event_device *dev = __get_cpu_var(tick_cpu_device).evtdev;
156 int cpu; 157 int cpu;
157 158
158 local_irq_save(flags); 159 local_irq_save(flags);
@@ -302,11 +303,26 @@ void tick_nohz_stop_sched_tick(void)
302out: 303out:
303 ts->next_jiffies = next_jiffies; 304 ts->next_jiffies = next_jiffies;
304 ts->last_jiffies = last_jiffies; 305 ts->last_jiffies = last_jiffies;
306 ts->sleep_length = ktime_sub(dev->next_event, now);
305end: 307end:
306 local_irq_restore(flags); 308 local_irq_restore(flags);
307} 309}
308 310
309/** 311/**
312 * tick_nohz_get_sleep_length - return the length of the current sleep
313 *
314 * Called from power state control code with interrupts disabled
315 */
316ktime_t tick_nohz_get_sleep_length(void)
317{
318 struct tick_sched *ts = &__get_cpu_var(tick_cpu_sched);
319
320 return ts->sleep_length;
321}
322
323EXPORT_SYMBOL_GPL(tick_nohz_get_sleep_length);
324
325/**
310 * nohz_restart_sched_tick - restart the idle tick from the idle task 326 * nohz_restart_sched_tick - restart the idle tick from the idle task
311 * 327 *
312 * Restart the idle tick when the CPU is woken up from idle 328 * Restart the idle tick when the CPU is woken up from idle
generated by cgit v1.2.2 (git 2.25.0) at 2024-11-08 10:16:27 -0500