summaryrefslogtreecommitdiffstats
path: root/kernel/power (follow)
Commit message (Collapse)AuthorAgeFilesLines
* PM / sleep: Add support for read-only sysfs attributesRafael J. Wysocki2016-01-042-15/+11
| | | | | | | | | Some sysfs attributes in /sys/power/ should really be read-only, so add support for that, convert those attributes to read-only and drop the stub .show() routines from them. Original-by: Sergey Senozhatsky <sergey.senozhatsky.work@gmail.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
* mm, page_alloc: rename __GFP_WAIT to __GFP_RECLAIMMel Gorman2015-11-071-8/+8
| | | | | | | | | | | | | | | | | | | | | | | | __GFP_WAIT was used to signal that the caller was in atomic context and could not sleep. Now it is possible to distinguish between true atomic context and callers that are not willing to sleep. The latter should clear __GFP_DIRECT_RECLAIM so kswapd will still wake. As clearing __GFP_WAIT behaves differently, there is a risk that people will clear the wrong flags. This patch renames __GFP_WAIT to __GFP_RECLAIM to clearly indicate what it does -- setting it allows all reclaim activity, clearing them prevents it. [akpm@linux-foundation.org: fix build] [akpm@linux-foundation.org: coding-style fixes] Signed-off-by: Mel Gorman <mgorman@techsingularity.net> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Vlastimil Babka <vbabka@suse.cz> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Christoph Lameter <cl@linux.com> Acked-by: David Rientjes <rientjes@google.com> Cc: Vitaly Wool <vitalywool@gmail.com> Cc: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* mm, page_alloc: distinguish between being unable to sleep, unwilling to ↵Mel Gorman2015-11-071-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | sleep and avoiding waking kswapd __GFP_WAIT has been used to identify atomic context in callers that hold spinlocks or are in interrupts. They are expected to be high priority and have access one of two watermarks lower than "min" which can be referred to as the "atomic reserve". __GFP_HIGH users get access to the first lower watermark and can be called the "high priority reserve". Over time, callers had a requirement to not block when fallback options were available. Some have abused __GFP_WAIT leading to a situation where an optimisitic allocation with a fallback option can access atomic reserves. This patch uses __GFP_ATOMIC to identify callers that are truely atomic, cannot sleep and have no alternative. High priority users continue to use __GFP_HIGH. __GFP_DIRECT_RECLAIM identifies callers that can sleep and are willing to enter direct reclaim. __GFP_KSWAPD_RECLAIM to identify callers that want to wake kswapd for background reclaim. __GFP_WAIT is redefined as a caller that is willing to enter direct reclaim and wake kswapd for background reclaim. This patch then converts a number of sites o __GFP_ATOMIC is used by callers that are high priority and have memory pools for those requests. GFP_ATOMIC uses this flag. o Callers that have a limited mempool to guarantee forward progress clear __GFP_DIRECT_RECLAIM but keep __GFP_KSWAPD_RECLAIM. bio allocations fall into this category where kswapd will still be woken but atomic reserves are not used as there is a one-entry mempool to guarantee progress. o Callers that are checking if they are non-blocking should use the helper gfpflags_allow_blocking() where possible. This is because checking for __GFP_WAIT as was done historically now can trigger false positives. Some exceptions like dm-crypt.c exist where the code intent is clearer if __GFP_DIRECT_RECLAIM is used instead of the helper due to flag manipulations. o Callers that built their own GFP flags instead of starting with GFP_KERNEL and friends now also need to specify __GFP_KSWAPD_RECLAIM. The first key hazard to watch out for is callers that removed __GFP_WAIT and was depending on access to atomic reserves for inconspicuous reasons. In some cases it may be appropriate for them to use __GFP_HIGH. The second key hazard is callers that assembled their own combination of GFP flags instead of starting with something like GFP_KERNEL. They may now wish to specify __GFP_KSWAPD_RECLAIM. It's almost certainly harmless if it's missed in most cases as other activity will wake kswapd. Signed-off-by: Mel Gorman <mgorman@techsingularity.net> Acked-by: Vlastimil Babka <vbabka@suse.cz> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Vitaly Wool <vitalywool@gmail.com> Cc: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* PM / hibernate: fix a comment typoGeliang Tang2015-10-141-1/+1
| | | | | | | | Just fix a typo in a function name in kerneldoc comments. Signed-off-by: Geliang Tang <geliangtang@163.com> Acked-by: Pavel Machek <pavel@ucw.cz> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
* PM / sleep: Add flags to indicate platform firmware involvementRafael J. Wysocki2015-10-141-0/+4
| | | | | | | | | | | | | | | | | There are quite a few cases in which device drivers, bus types or even the PM core itself may benefit from knowing whether or not the platform firmware will be involved in the upcoming system power transition (during system suspend) or whether or not it was involved in it (during system resume). For this reason, introduce global system suspend flags that can be used by the platform code to expose that information for the benefit of the other parts of the kernel and make the ACPI core set them as appropriate. Users of the new flags will be added later. Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
* PM / sleep: Report interrupt that caused system wakeupAlexandra Yates2015-09-161-0/+17
| | | | | | | | | | | | | | Add a sysfs attribute, /sys/power/pm_wakeup_irq, reporting the IRQ number of the first wakeup interrupt (that is, the first interrupt from an IRQ line armed for system wakeup) seen by the kernel during the most recent system suspend/resume cycle. This feature will be useful for system wakeup diagnostics of spurious wakeup interrupts. Signed-off-by: Alexandra Yates <alexandra.yates@linux.intel.com> [ rjw: Fixed up pm_wakeup_irq definition ] Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
* Merge branch 'for-4.3/core' of git://git.kernel.dk/linux-blockLinus Torvalds2015-09-021-8/+4
|\ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Pull core block updates from Jens Axboe: "This first core part of the block IO changes contains: - Cleanup of the bio IO error signaling from Christoph. We used to rely on the uptodate bit and passing around of an error, now we store the error in the bio itself. - Improvement of the above from myself, by shrinking the bio size down again to fit in two cachelines on x86-64. - Revert of the max_hw_sectors cap removal from a revision again, from Jeff Moyer. This caused performance regressions in various tests. Reinstate the limit, bump it to a more reasonable size instead. - Make /sys/block/<dev>/queue/discard_max_bytes writeable, by me. Most devices have huge trim limits, which can cause nasty latencies when deleting files. Enable the admin to configure the size down. We will look into having a more sane default instead of UINT_MAX sectors. - Improvement of the SGP gaps logic from Keith Busch. - Enable the block core to handle arbitrarily sized bios, which enables a nice simplification of bio_add_page() (which is an IO hot path). From Kent. - Improvements to the partition io stats accounting, making it faster. From Ming Lei. - Also from Ming Lei, a basic fixup for overflow of the sysfs pending file in blk-mq, as well as a fix for a blk-mq timeout race condition. - Ming Lin has been carrying Kents above mentioned patches forward for a while, and testing them. Ming also did a few fixes around that. - Sasha Levin found and fixed a use-after-free problem introduced by the bio->bi_error changes from Christoph. - Small blk cgroup cleanup from Viresh Kumar" * 'for-4.3/core' of git://git.kernel.dk/linux-block: (26 commits) blk: Fix bio_io_vec index when checking bvec gaps block: Replace SG_GAPS with new queue limits mask block: bump BLK_DEF_MAX_SECTORS to 2560 Revert "block: remove artifical max_hw_sectors cap" blk-mq: fix race between timeout and freeing request blk-mq: fix buffer overflow when reading sysfs file of 'pending' Documentation: update notes in biovecs about arbitrarily sized bios block: remove bio_get_nr_vecs() fs: use helper bio_add_page() instead of open coding on bi_io_vec block: kill merge_bvec_fn() completely md/raid5: get rid of bio_fits_rdev() md/raid5: split bio for chunk_aligned_read block: remove split code in blkdev_issue_{discard,write_same} btrfs: remove bio splitting and merge_bvec_fn() calls bcache: remove driver private bio splitting code block: simplify bio_add_page() block: make generic_make_request handle arbitrarily sized bios blk-cgroup: Drop unlikely before IS_ERR(_OR_NULL) block: don't access bio->bi_error after bio_put() block: shrink struct bio down to 2 cache lines again ...
| * block: add a bi_error field to struct bioChristoph Hellwig2015-07-291-8/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Currently we have two different ways to signal an I/O error on a BIO: (1) by clearing the BIO_UPTODATE flag (2) by returning a Linux errno value to the bi_end_io callback The first one has the drawback of only communicating a single possible error (-EIO), and the second one has the drawback of not beeing persistent when bios are queued up, and are not passed along from child to parent bio in the ever more popular chaining scenario. Having both mechanisms available has the additional drawback of utterly confusing driver authors and introducing bugs where various I/O submitters only deal with one of them, and the others have to add boilerplate code to deal with both kinds of error returns. So add a new bi_error field to store an errno value directly in struct bio and remove the existing mechanisms to clean all this up. Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Hannes Reinecke <hare@suse.de> Reviewed-by: NeilBrown <neilb@suse.com> Signed-off-by: Jens Axboe <axboe@fb.com>
* | PM / suspend: make sync() on suspend-to-RAM build-time optionalLen Brown2015-07-312-0/+12
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The Linux kernel suspend path has traditionally invoked sys_sync() before freezing user threads. But sys_sync() can be expensive, and some user-space OS's do not want the kernel to pay the cost of sys_sync() on every suspend -- preferring invoke sync() from user-space if/when they want it. So make sys_sync on suspend build-time optional. The default is unchanged. Signed-off-by: Len Brown <len.brown@intel.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
* | PM / autosleep: Use workqueue for user space wakeup sources garbage collectorSungEun Kim2015-07-141-3/+15
|/ | | | | | | | | | | | | | | The synchronous synchronize_rcu() in wakeup_source_remove() makes user process which writes to /sys/kernel/wake_unlock blocked sometimes. For example, when android eventhub tries to release a wakelock, this blocking process can occur, and eventhub can't get input events for a while. Using a work item instead of direct function call at pm_wake_unlock() can prevent this unnecessary delay from happening. Signed-off-by: SungEun Kim <cleaneye.kim@lge.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
* Merge tag 'pm+acpi-4.2-rc1-2' of ↵Linus Torvalds2015-07-012-2/+4
|\ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm Pull power management and ACPI fixes from Rafael Wysocki: "These are fixes that didn't make it to the previous PM+ACPI pull request or are fixing issues introduced by it. Specifics: - Fix a recently added memory leak in an error path in the ACPI resources management code (Dan Carpenter) - Fix a build warning triggered by an ACPI video header function that should be static inline (Borislav Petkov) - Change names of helper function converting struct fwnode_handle pointers to either struct device_node or struct acpi_device pointers so they don't conflict with local variable names (Alexander Sverdlin) - Make the hibernate core re-enable nonboot CPUs on failures to disable them as expected (Vitaly Kuznetsov) - Increase the default timeout of the device suspend watchdog to prevent it from triggering too early on some systems (Takashi Iwai) - Prevent the cpuidle powernv driver from registering idle states with CPUIDLE_FLAG_TIMER_STOP set if CONFIG_TICK_ONESHOT is unset which leads to boot hangs (Preeti U Murthy)" * tag 'pm+acpi-4.2-rc1-2' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm: tick/idle/powerpc: Do not register idle states with CPUIDLE_FLAG_TIMER_STOP set in periodic mode PM / sleep: Increase default DPM watchdog timeout to 60 PM / hibernate: re-enable nonboot cpus on disable_nonboot_cpus() failure ACPI / OF: Rename of_node() and acpi_node() to to_of_node() and to_acpi_node() ACPI / video: Inline acpi_video_set_dmi_backlight_type ACPI / resources: free memory on error in add_region_before()
| * PM / sleep: Increase default DPM watchdog timeout to 60Takashi Iwai2015-06-251-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Many harddisks (mostly WD ones) have firmware problems and take too long, more than 10 seconds, to resume from suspend. And this often exceeds the default DPM watchdog timeout (12 seconds), resulting in a kernel panic out of sudden. Since most distros just take the default as is, we should give a bit more safer value. This patch increases the default value from 12 seconds to one minute, which has been confirmed to be long enough for such problematic disks. Link: https://bugzilla.kernel.org/show_bug.cgi?id=91921 Fixes: 70fea60d888d (PM / Sleep: Detect device suspend/resume lockup and log event) Cc: 3.13+ <stable@vger.kernel.org> # 3.13+ Signed-off-by: Takashi Iwai <tiwai@suse.de> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
| * PM / hibernate: re-enable nonboot cpus on disable_nonboot_cpus() failureVitaly Kuznetsov2015-06-241-1/+3
| | | | | | | | | | | | | | | | | | | | | | | | | | | | When disable_nonboot_cpus() fails on some cpu it doesn't bring back all cpus it managed to offline, a consequent call to enable_nonboot_cpus() is expected. In hibernation_platform_enter() we don't call enable_nonboot_cpus() on error so cpus stay offlined. create_image() and resume_target_kernel() functions handle disable_nonboot_cpus() faults correctly, hibernation_platform_enter() is the only one which is doing it wrong. Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
* | Merge branch 'for-4.2/core' of git://git.kernel.dk/linux-blockLinus Torvalds2015-06-254-153/+121
|\ \ | |/ |/| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Pull core block IO update from Jens Axboe: "Nothing really major in here, mostly a collection of smaller optimizations and cleanups, mixed with various fixes. In more detail, this contains: - Addition of policy specific data to blkcg for block cgroups. From Arianna Avanzini. - Various cleanups around command types from Christoph. - Cleanup of the suspend block I/O path from Christoph. - Plugging updates from Shaohua and Jeff Moyer, for blk-mq. - Eliminating atomic inc/dec of both remaining IO count and reference count in a bio. From me. - Fixes for SG gap and chunk size support for data-less (discards) IO, so we can merge these better. From me. - Small restructuring of blk-mq shared tag support, freeing drivers from iterating hardware queues. From Keith Busch. - A few cfq-iosched tweaks, from Tahsin Erdogan and me. Makes the IOPS mode the default for non-rotational storage" * 'for-4.2/core' of git://git.kernel.dk/linux-block: (35 commits) cfq-iosched: fix other locations where blkcg_to_cfqgd() can return NULL cfq-iosched: fix sysfs oops when attempting to read unconfigured weights cfq-iosched: move group scheduling functions under ifdef cfq-iosched: fix the setting of IOPS mode on SSDs blktrace: Add blktrace.c to BLOCK LAYER in MAINTAINERS file block, cgroup: implement policy-specific per-blkcg data block: Make CFQ default to IOPS mode on SSDs block: add blk_set_queue_dying() to blkdev.h blk-mq: Shared tag enhancements block: don't honor chunk sizes for data-less IO block: only honor SG gap prevention for merges that contain data block: fix returnvar.cocci warnings block, dm: don't copy bios for request clones block: remove management of bi_remaining when restoring original bi_end_io block: replace trylock with mutex_lock in blkdev_reread_part() block: export blkdev_reread_part() and __blkdev_reread_part() suspend: simplify block I/O handling block: collapse bio bit space block: remove unused BIO_RW_BLOCK and BIO_EOF flags block: remove BIO_EOPNOTSUPP ...
| * suspend: simplify block I/O handlingChristoph Hellwig2015-05-194-153/+121
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Stop abusing struct page functionality and the swap end_io handler, and instead add a modified version of the blk-lib.c bio_batch helpers. Also move the block I/O code into swap.c as they are directly tied into each other. Signed-off-by: Christoph Hellwig <hch@lst.de> Tested-by: Pavel Machek <pavel@ucw.cz> Tested-by: Ming Lin <mlin@kernel.org> Acked-by: Pavel Machek <pavel@ucw.cz> Acked-by: Rafael J. Wysocki <rjw@rjwysocki.net> Signed-off-by: Jens Axboe <axboe@fb.com>
* | PM / sleep: Return -EBUSY from suspend_enter() on wakeup detectionRuchi Kandoi2015-05-191-0/+2
| | | | | | | | | | | | | | | | | | | | | | | | | | | | If a wakeup source is found to be pending in the last stage of suspend after syscore suspend, then the machine won't suspend, but suspend_enter() will return 0. That is confusing, as wakeup detection elsewhere causes -EBUSY to be returned from suspend_enter(). To avoid the confusion, make suspend_enter() return -EBUSY in that case too. Signed-off-by: Ruchi Kandoi <kandoiruchi@google.com> [ rjw: Subject and changelog ] Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
* | PM / sleep: Fix symbol name in a comment in kernel/power/main.cRafael J. Wysocki2015-05-131-1/+1
| | | | | | | | Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
* | PM / sleep: Refine diagnostic messages in enter_state()Rafael J. Wysocki2015-05-121-3/+3
|/ | | | | | | | | | | | Some of the system suspend diagnostic messages related to suspend-to-idle refer to it as "freeze sleep" or "freeze state" while the others say "suspend-to-idle". To reduce the possible confusion that may result from that, refine the former either to say "suspend to idle" too or to make it clearer that what is printed is a state string written to /sys/power/state ("mem", "standby", or "freeze"). Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
* Merge back earlier suspend/hibernate material for v4.1.Rafael J. Wysocki2015-04-102-3/+12
|\
| * PM / sleep: add pm-trace support for suspending phaseZhonghui Fu2015-03-181-1/+1
| | | | | | | | | | | | | | | | | | | | | | Occasionally, the system can't come back up after suspend/resume due to problems of device suspending phase. This patch make PM_TRACE infrastructure cover device suspending phase of suspend/resume process, and the information in RTC can tell developers which device suspending function make system hang. Signed-off-by: Zhonghui Fu <zhonghui.fu@linux.intel.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
| * PM / sleep: add configurable delay for pm_testBrian Norris2015-02-261-2/+11
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | When CONFIG_PM_DEBUG=y, we provide a sysfs file (/sys/power/pm_test) for selecting one of a few suspend test modes, where rather than entering a full suspend state, the kernel will perform some subset of suspend steps, wait 5 seconds, and then resume back to normal operation. This mode is useful for (among other things) observing the state of the system just before entering a sleep mode, for debugging or analysis purposes. However, a constant 5 second wait is not sufficient for some sorts of analysis; for example, on an SoC, one might want to use external tools to probe the power states of various on-chip controllers or clocks. This patch turns this 5 second delay into a configurable module parameter, so users can determine how long to wait in this pseudo-suspend state before resuming the system. Example (wait 30 seconds); # echo 30 > /sys/module/suspend/parameters/pm_test_delay # echo core > /sys/power/pm_test # time echo mem > /sys/power/state ... [ 17.583625] suspend debug: Waiting for 30 second(s). ... real 0m30.381s user 0m0.017s sys 0m0.080s Signed-off-by: Brian Norris <computersforpeace@gmail.com> Acked-by: Pavel Machek <pavel@ucw.cz> Reviewed-by: Kevin Cernekee <cernekee@chromium.org> Acked-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
* | Revert "PM / hibernate: avoid unsafe pages in e820 reserved regions"Rafael J. Wysocki2015-04-071-20/+1
|/ | | | | | | | | | | | | | Commit 84c91b7ae07c (PM / hibernate: avoid unsafe pages in e820 reserved regions) is reported to make resume from hibernation on Lenovo x230 unreliable, so revert it. We will revisit the issue the commit in question was supposed to fix in the future. Link: https://bugzilla.kernel.org/show_bug.cgi?id=96111 Reported-by: rhn <kebuac.rhn@porcupinefactory.org> Cc: 3.17+ <stable@vger.kernel.org> # 3.17+ Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
* PM / sleep: Re-implement suspend-to-idle handlingRafael J. Wysocki2015-02-131-7/+36
| | | | | | | | | | | | | | | | | | | | | | | | | | | | In preparation for adding support for quiescing timers in the final stage of suspend-to-idle transitions, rework the freeze_enter() function making the system wait on a wakeup event, the freeze_wake() function terminating the suspend-to-idle loop and the mechanism by which deep idle states are entered during suspend-to-idle. First of all, introduce a simple state machine for suspend-to-idle and make the code in question use it. Second, prevent freeze_enter() from losing wakeup events due to race conditions and ensure that the number of online CPUs won't change while it is being executed. In addition to that, make it force all of the CPUs re-enter the idle loop in case they are in idle states already (so they can enter deeper idle states if possible). Next, drop cpuidle_use_deepest_state() and replace use_deepest_state checks in cpuidle_select() and cpuidle_reflect() with a single suspend-to-idle state check in cpuidle_idle_call(). Finally, introduce cpuidle_enter_freeze() that will simply find the deepest idle state available to the given CPU and enter it using cpuidle_enter(). Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
* oom, PM: make OOM detection in the freezer path racelessMichal Hocko2015-02-121-41/+9
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Commit 5695be142e20 ("OOM, PM: OOM killed task shouldn't escape PM suspend") has left a race window when OOM killer manages to note_oom_kill after freeze_processes checks the counter. The race window is quite small and really unlikely and partial solution deemed sufficient at the time of submission. Tejun wasn't happy about this partial solution though and insisted on a full solution. That requires the full OOM and freezer's task freezing exclusion, though. This is done by this patch which introduces oom_sem RW lock and turns oom_killer_disable() into a full OOM barrier. oom_killer_disabled check is moved from the allocation path to the OOM level and we take oom_sem for reading for both the check and the whole OOM invocation. oom_killer_disable() takes oom_sem for writing so it waits for all currently running OOM killer invocations. Then it disable all the further OOMs by setting oom_killer_disabled and checks for any oom victims. Victims are counted via mark_tsk_oom_victim resp. unmark_oom_victim. The last victim wakes up all waiters enqueued by oom_killer_disable(). Therefore this function acts as the full OOM barrier. The page fault path is covered now as well although it was assumed to be safe before. As per Tejun, "We used to have freezing points deep in file system code which may be reacheable from page fault." so it would be better and more robust to not rely on freezing points here. Same applies to the memcg OOM killer. out_of_memory tells the caller whether the OOM was allowed to trigger and the callers are supposed to handle the situation. The page allocation path simply fails the allocation same as before. The page fault path will retry the fault (more on that later) and Sysrq OOM trigger will simply complain to the log. Normally there wouldn't be any unfrozen user tasks after try_to_freeze_tasks so the function will not block. But if there was an OOM killer racing with try_to_freeze_tasks and the OOM victim didn't finish yet then we have to wait for it. This should complete in a finite time, though, because - the victim cannot loop in the page fault handler (it would die on the way out from the exception) - it cannot loop in the page allocator because all the further allocation would fail and __GFP_NOFAIL allocations are not acceptable at this stage - it shouldn't be blocked on any locks held by frozen tasks (try_to_freeze expects lockless context) and kernel threads and work queues are not frozen yet Signed-off-by: Michal Hocko <mhocko@suse.cz> Suggested-by: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Cong Wang <xiyou.wangcong@gmail.com> Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* PM: convert printk to pr_* equivalentMichal Hocko2015-02-121-14/+15
| | | | | | | | | | | | | | | While touching this area let's convert printk to pr_*. This also makes the printing of continuation lines done properly. Signed-off-by: Michal Hocko <mhocko@suse.cz> Acked-by: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Cong Wang <xiyou.wangcong@gmail.com> Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* Merge tag 'pm+acpi-3.20-rc1' of ↵Linus Torvalds2015-02-112-7/+95
|\ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm Pull ACPI and power management updates from Rafael Wysocki: "We have a few new features this time, including a new SFI-based cpufreq driver, a new devfreq driver for Tegra Activity Monitor, a new devfreq class for providing its governors with raw utilization data and a new ACPI driver for AMD SoCs. Still, the majority of changes here are reworks of existing code to make it more straightforward or to prepare it for implementing new features on top of it. The primary example is the rework of ACPI resources handling from Jiang Liu, Thomas Gleixner and Lv Zheng with support for IOAPIC hotplug implemented on top of it, but there is quite a number of changes of this kind in the cpufreq core, ACPICA, ACPI EC driver, ACPI processor driver and the generic power domains core code too. The most active developer is Viresh Kumar with his cpufreq changes. Specifics: - Rework of the core ACPI resources parsing code to fix issues in it and make using resource offsets more convenient and consolidation of some resource-handing code in a couple of places that have grown analagous data structures and code to cover the the same gap in the core (Jiang Liu, Thomas Gleixner, Lv Zheng). - ACPI-based IOAPIC hotplug support on top of the resources handling rework (Jiang Liu, Yinghai Lu). - ACPICA update to upstream release 20150204 including an interrupt handling rework that allows drivers to install raw handlers for ACPI GPEs which then become entirely responsible for the given GPE and the ACPICA core code won't touch it (Lv Zheng, David E Box, Octavian Purdila). - ACPI EC driver rework to fix several concurrency issues and other problems related to events handling on top of the ACPICA's new support for raw GPE handlers (Lv Zheng). - New ACPI driver for AMD SoCs analogous to the LPSS (Low-Power Subsystem) driver for Intel chips (Ken Xue). - Two minor fixes of the ACPI LPSS driver (Heikki Krogerus, Jarkko Nikula). - Two new blacklist entries for machines (Samsung 730U3E/740U3E and 510R) where the native backlight interface doesn't work correctly while the ACPI one does (Hans de Goede). - Rework of the ACPI processor driver's handling of idle states to make the code more straightforward and less bloated overall (Rafael J Wysocki). - Assorted minor fixes related to ACPI and SFI (Andreas Ruprecht, Andy Shevchenko, Hanjun Guo, Jan Beulich, Rafael J Wysocki, Yaowei Bai). - PCI core power management modification to avoid resuming (some) runtime-suspended devices during system suspend if they are in the right states already (Rafael J Wysocki). - New SFI-based cpufreq driver for Intel platforms using SFI (Srinidhi Kasagar). - cpufreq core fixes, cleanups and simplifications (Viresh Kumar, Doug Anderson, Wolfram Sang). - SkyLake CPU support and other updates for the intel_pstate driver (Kristen Carlson Accardi, Srinivas Pandruvada). - cpufreq-dt driver cleanup (Markus Elfring). - Init fix for the ARM big.LITTLE cpuidle driver (Sudeep Holla). - Generic power domains core code fixes and cleanups (Ulf Hansson). - Operating Performance Points (OPP) core code cleanups and kernel documentation update (Nishanth Menon). - New dabugfs interface to make the list of PM QoS constraints available to user space (Nishanth Menon). - New devfreq driver for Tegra Activity Monitor (Tomeu Vizoso). - New devfreq class (devfreq_event) to provide raw utilization data to devfreq governors (Chanwoo Choi). - Assorted minor fixes and cleanups related to power management (Andreas Ruprecht, Krzysztof Kozlowski, Rickard Strandqvist, Pavel Machek, Todd E Brandt, Wonhong Kwon). - turbostat updates (Len Brown) and cpupower Makefile improvement (Sriram Raghunathan)" * tag 'pm+acpi-3.20-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm: (151 commits) tools/power turbostat: relax dependency on APERF_MSR tools/power turbostat: relax dependency on invariant TSC Merge branch 'pci/host-generic' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci into acpi-resources tools/power turbostat: decode MSR_*_PERF_LIMIT_REASONS tools/power turbostat: relax dependency on root permission ACPI / video: Add disable_native_backlight quirk for Samsung 510R ACPI / PM: Remove unneeded nested #ifdef USB / PM: Remove unneeded #ifdef and associated dead code intel_pstate: provide option to only use intel_pstate with HWP ACPI / EC: Add GPE reference counting debugging messages ACPI / EC: Add query flushing support ACPI / EC: Refine command storm prevention support ACPI / EC: Add command flushing support. ACPI / EC: Introduce STARTED/STOPPED flags to replace BLOCKED flag ACPI: add AMD ACPI2Platform device support for x86 system ACPI / table: remove duplicate NULL check for the handler of acpi_table_parse() ACPI / EC: Update revision due to raw handler mode. ACPI / EC: Reduce ec_poll() by referencing the last register access timestamp. ACPI / EC: Fix several GPE handling issues by deploying ACPI_GPE_DISPATCH_RAW_HANDLER mode. ACPICA: Events: Enable APIs to allow interrupt/polling adaptive request based GPE handling model ...
| *-. Merge branches 'pm-sleep' and 'pm-runtime'Rafael J. Wysocki2015-02-101-5/+6
| |\ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | * pm-sleep: PM / hibernate: exclude freed pages from allocated pages printout PM / sleep: export suspend_resume trace event PM / sleep: Mention async suspend in PM_TRACE documentation PM / hibernate: Remove unused function * pm-runtime: ACPI / PM: Remove unneeded nested #ifdef USB / PM: Remove unneeded #ifdef and associated dead code
| | * | PM / hibernate: exclude freed pages from allocated pages printoutWonhong Kwon2015-02-031-3/+6
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | hibernate_preallocate_memory() prints out that how many pages are allocated, but it doesn't take into consideration the pages freed by free_unnecessary_pages(). Therefore, it always shows the count more than actually allocated. Signed-off-by: Wonhong Kwon <wonhong.kwon@lge.com> [ rjw: Subject ] Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
| | * | PM / hibernate: Remove unused functionRickard Strandqvist2015-01-231-2/+0
| | |/ | | | | | | | | | | | | | | | | | | | | | | | | | | | Remove the function get_safe_write_buffer() that is not used anywhere. This was partially found by using a static code analysis program called cppcheck. Signed-off-by: Rickard Strandqvist <rickard_strandqvist@spectrumdigital.se> Acked-by: Pavel Machek <pavel@ucw.cz> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
| * / PM / QoS: Add debugfs support to view the list of constraintsNishanth Menon2015-01-231-2/+89
| |/ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | PM QoS requests are notoriously hard to debug and made even more so due to their highly dynamic nature. Having visibility into the internal data representation per constraint allows us to have much better appreciation of potential issues or bad usage by drivers in the system. So introduce for all classes of PM QoS, an entry in /sys/kernel/debug/pm_qos that shall show all the current requests as well as the snapshot of the value these requests boil down to. For example: ==> /sys/kernel/debug/pm_qos/cpu_dma_latency <== 1: 4444: Active 2: 2000000000: Default 3: 2000000000: Default 4: 2000000000: Default Type=Minimum, Value=4444, Requests: active=1 / total=4 ==> /sys/kernel/debug/pm_qos/memory_bandwidth <== Empty! ... The actual value listed will have their meaning based on the QoS it is on, the 'Type' indicates what logic it would use to collate the information - Minimum, Maximum, or Sum. Value is the collation of all requests. This interface also compares the values with the defaults for the QoS class and marks the ones that are currently active. Signed-off-by: Nishanth Menon <nm@ti.com> Signed-off-by: Dave Gerlach <d-gerlach@ti.com> Acked-by: Kevin Hilman <khilman@linaro.org> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
* / rcu: Make SRCU optional by using CONFIG_SRCUPranith Kumar2015-01-061-0/+1
|/ | | | | | | | | | | | | | | | | | | | | | | | | | | | | SRCU is not necessary to be compiled by default in all cases. For tinification efforts not compiling SRCU unless necessary is desirable. The current patch tries to make compiling SRCU optional by introducing a new Kconfig option CONFIG_SRCU which is selected when any of the components making use of SRCU are selected. If we do not select CONFIG_SRCU, srcu.o will not be compiled at all. text data bss dec hex filename 2007 0 0 2007 7d7 kernel/rcu/srcu.o Size of arch/powerpc/boot/zImage changes from text data bss dec hex filename 831552 64180 23944 919676 e087c arch/powerpc/boot/zImage : before 829504 64180 23952 917636 e0084 arch/powerpc/boot/zImage : after so the savings are about ~2000 bytes. Signed-off-by: Pranith Kumar <bobby.prani@gmail.com> CC: Paul E. McKenney <paulmck@linux.vnet.ibm.com> CC: Josh Triplett <josh@joshtriplett.org> CC: Lai Jiangshan <laijs@cn.fujitsu.com> Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com> [ paulmck: resolve conflict due to removal of arch/ia64/kvm/Kconfig. ]
* PM: Eliminate CONFIG_PM_RUNTIMERafael J. Wysocki2014-12-191-10/+6
| | | | | | | | | | Having switched over all of the users of CONFIG_PM_RUNTIME to use CONFIG_PM directly, turn the latter into a user-selectable option and drop the former entirely from the tree. Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org> Acked-by: Kevin Hilman <khilman@linaro.org>
* Merge branch 'pm-runtime'Rafael J. Wysocki2014-12-081-5/+1
|\ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | * pm-runtime: (25 commits) i2c-omap / PM: Drop CONFIG_PM_RUNTIME from i2c-omap.c dmaengine / PM: Replace CONFIG_PM_RUNTIME with CONFIG_PM drivers: sh / PM: Replace CONFIG_PM_RUNTIME with CONFIG_PM e1000e / igb / PM: Eliminate CONFIG_PM_RUNTIME MMC / PM: Replace CONFIG_PM_RUNTIME with CONFIG_PM MFD / PM: Replace CONFIG_PM_RUNTIME with CONFIG_PM misc / PM: Replace CONFIG_PM_RUNTIME with CONFIG_PM media / PM: Replace CONFIG_PM_RUNTIME with CONFIG_PM input / PM: Replace CONFIG_PM_RUNTIME with CONFIG_PM iio / PM: Replace CONFIG_PM_RUNTIME with CONFIG_PM hsi / OMAP / PM: Replace CONFIG_PM_RUNTIME with CONFIG_PM i2c-hid / PM: Replace CONFIG_PM_RUNTIME with CONFIG_PM drm / exynos / PM: Replace CONFIG_PM_RUNTIME with CONFIG_PM gpio / PM: Replace CONFIG_PM_RUNTIME with CONFIG_PM hwrandom / exynos / PM: Use CONFIG_PM in #ifdef block / PM: Replace CONFIG_PM_RUNTIME with CONFIG_PM USB / PM: Drop CONFIG_PM_RUNTIME from the USB core PM: Merge the SET*_RUNTIME_PM_OPS() macros PM / Kconfig: Do not select PM directly from Kconfig files PCI / PM: Drop CONFIG_PM_RUNTIME from the PCI core ...
| * PM: Drop CONFIG_PM_RUNTIME from the driver coreRafael J. Wysocki2014-12-041-4/+0
| | | | | | | | | | | | | | | | | | | | | | | | After commit b2b49ccbdd54 (PM: Kconfig: Set PM_RUNTIME if PM_SLEEP is selected) PM_RUNTIME is always set if PM is set, so quite a few depend on CONFIG_PM or even may be dropped entirely in some cases. Replace CONFIG_PM_RUNTIME with CONFIG_PM in the PM core code. Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org> Acked-by: Kevin Hilman <khilman@linaro.org> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
| * Merge back earlier 'pm-runtime' material for 3.19-rc1.Rafael J. Wysocki2014-11-271-1/+1
| |\
| | * PM: Kconfig: Set PM_RUNTIME if PM_SLEEP is selectedRafael J. Wysocki2014-11-181-0/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The number of and dependencies between high-level power management Kconfig options make life much harder than necessary. Several conbinations of them have to be tested and supported, even though some of those combinations are very rarely used in practice (if they are used in practice at all). Moreover, the fact that we have separate independent Kconfig options for runtime PM and system suspend is a serious obstacle for integration between the two frameworks. To overcome these difficulties, always select PM_RUNTIME if PM_SLEEP is set. Among other things, this will allow system suspend callbacks provided by bus types and device drivers to rely on the runtime PM framework regardless of the kernel configuration. Enthusiastically-acked-by: Kevin Hilman <khilman@linaro.org> Tested-by: Geert Uytterhoeven <geert+renesas@glider.be> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
| | * PM / Runtime: Kconfig: move ia64 dependency to arch/ia64/KconfigKevin Hilman2014-11-151-1/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The IA64_HP_SIM dependency on PM_RUNTIME should be done in the arch Kconfig instead of in the PM core. Move it accordingly. NOTE: arch/ia64/Kconfig currently does a 'select PM', which since commit 1eb208aea317 (PM: Make CONFIG_PM depend on (CONFIG_PM_SLEEP || CONFIG_PM_RUNTIME)) is effectively a noop unless PM_SLEEP or PM_RUNTIME are set elsewhere. Signed-off-by: Kevin Hilman <khilman@linaro.org> Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
| | |
| \ \
*-. \ \ Merge branches 'pm-domains', 'pm-sleep' and 'pm-tools'Rafael J. Wysocki2014-12-084-35/+34
|\ \ \ \ | | |/ / | |/| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | * pm-domains: ARM: shmobile: Convert to genpd flags for PM clocks for R-mobile ARM: shmobile: Convert to genpd flags for PM clocks for r8a7779 PM / Domains: Initial PM clock support for genpd PM / Domains: Power on the PM domain right after attach completes PM / Domains: Move struct pm_domain_data to pm_domain.h PM / Domains: Extract code to power off/on a PM domain PM / Domains: Make genpd parameter of pm_genpd_present() const * pm-sleep: PM / hibernate: Deletion of an unnecessary check before the function call "vfree" PM / Hibernate: Migrate to ktime_t * pm-tools: tools: cpupower: fix return checks for sysfs_get_idlestate_count()
| | * | PM / hibernate: Deletion of an unnecessary check before the function call ↵Markus Elfring2014-11-181-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | "vfree" The vfree() function performs also input parameter validation. Thus the test around the call is not needed. This issue was detected by using the Coccinelle software. Signed-off-by: Markus Elfring <elfring@users.sourceforge.net> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
| | * | Merge back 'pm-sleep' material for 3.19-rc1.Rafael J. Wysocki2014-11-184-34/+33
| |/| |
| | * | PM / Hibernate: Migrate to ktime_tTina Ruchandani2014-11-034-34/+33
| | |/ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This patch migrates swsusp_show_speed and its callers to using ktime_t instead of 'struct timeval' which suffers from the y2038 problem. Changes to swsusp_show_speed: - use ktime_t for start and stop times - pass start and stop times by value Calling functions affected: - load_image - load_image_lzo - save_image - save_image_lzo - hibernate_preallocate_memory Design decisions: - use ktime_t to preserve same granularity of reporting as before - use centisecs logic as before to avoid 'div by zero' issues caused by using seconds and nanoseconds directly - use monotonic time (ktime_get()) since we only care about elapsed time. Signed-off-by: Tina Ruchandani <ruchandani.tina@gmail.com> Suggested-by: Arnd Bergmann <arnd@arndb.de> Reviewed-by: Arnd Bergmann <arnd@arndb.de> Acked-by: Pavel Machek <pavel@ucw.cz> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
* | / PM: Kconfig: fix unmet dependency for CPU_PMPankaj Dubey2014-11-181-1/+0
|/ / | | | | | | | | | | | | | | | | | | | | | | | | | | | | If BL_SWITCHER is enabled but SUSPEND and CPU_IDLE is not enabled we are getting following config warning. warning: (BL_SWITCHER) selects CPU_PM which has unmet direct dependencies (SUSPEND || CPU_IDLE) It has been noticed that CPU_PM dependencies in this file are not really required so let's remove these dependencies from CPU_PM. Signed-off-by: Pankaj Dubey <pankaj.dubey@samsung.com> Acked-by: Nicolas Pitre <nico@linaro.org> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
* / PM / sleep: Fix entering suspend-to-IDLE if no freeze_oops is setDmitry Eremin-Solenikov2014-11-081-2/+2
|/ | | | | | | | | | If no freeze_ops is set, trying to enter suspend-to-IDLE will cause a nice oops in platform_suspend_prepare_late(). Add respective checks to platform_suspend_prepare_late() and platform_resume_early() functions. Fixes: a8d46b9e4e48 (ACPI / sleep: Rework the handling of ACPI GPE wakeup ...) Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
* PM / Sleep: fix recovery during resuming from hibernationImre Deak2014-10-271-1/+7
| | | | | | | | | | | If a device's dev_pm_ops::freeze callback fails during the QUIESCE phase, we don't rollback things correctly calling the thaw and complete callbacks. This could leave some devices in a suspended state in case of an error during resuming from hibernation. Signed-off-by: Imre Deak <imre.deak@intel.com> Cc: All applicable <stable@vger.kernel.org> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
* Merge branch 'freezer'Rafael J. Wysocki2014-10-231-9/+48
|\ | | | | | | | | | | | | | | | | * freezer: PM / freezer: Clean up code after recent fixes PM: convert do_each_thread to for_each_process_thread OOM, PM: OOM killed task shouldn't escape PM suspend freezer: remove obsolete comments in __thaw_task() freezer: Do not freeze tasks killed by OOM killer
| * PM / freezer: Clean up code after recent fixesRafael J. Wysocki2014-10-221-15/+16
| | | | | | | | | | | | | | Clean up the code in process.c after recent changes to get rid of unnecessary labels and goto statements. Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
| * PM: convert do_each_thread to for_each_process_threadMichal Hocko2014-10-211-8/+8
| | | | | | | | | | | | | | | | | | | | | | | | as per 0c740d0afc3b (introduce for_each_thread() to replace the buggy while_each_thread()) get rid of do_each_thread { } while_each_thread() construct and replace it by a more error prone for_each_thread. This patch doesn't introduce any user visible change. Suggested-by: Oleg Nesterov <oleg@redhat.com> Signed-off-by: Michal Hocko <mhocko@suse.cz> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
| * OOM, PM: OOM killed task shouldn't escape PM suspendMichal Hocko2014-10-211-1/+39
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | PM freezer relies on having all tasks frozen by the time devices are getting frozen so that no task will touch them while they are getting frozen. But OOM killer is allowed to kill an already frozen task in order to handle OOM situtation. In order to protect from late wake ups OOM killer is disabled after all tasks are frozen. This, however, still keeps a window open when a killed task didn't manage to die by the time freeze_processes finishes. Reduce the race window by checking all tasks after OOM killer has been disabled. This is still not race free completely unfortunately because oom_killer_disable cannot stop an already ongoing OOM killer so a task might still wake up from the fridge and get killed without freeze_processes noticing. Full synchronization of OOM and freezer is, however, too heavy weight for this highly unlikely case. Introduce and check oom_kills counter which gets incremented early when the allocator enters __alloc_pages_may_oom path and only check all the tasks if the counter changes during the freezing attempt. The counter is updated so early to reduce the race window since allocator checked oom_killer_disabled which is set by PM-freezing code. A false positive will push the PM-freezer into a slow path but that is not a big deal. Changes since v1 - push the re-check loop out of freeze_processes into check_frozen_processes and invert the condition to make the code more readable as per Rafael Fixes: f660daac474c6f (oom: thaw threads if oom killed thread is frozen before deferring) Cc: 3.2+ <stable@vger.kernel.org> # 3.2+ Signed-off-by: Michal Hocko <mhocko@suse.cz> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
* | Merge branch 'pm-qos'Rafael J. Wysocki2014-10-231-1/+26
|\ \ | |/ |/| | | | | * pm-qos: PM / QoS: Add PM_QOS_MEMORY_BANDWIDTH class
| * PM / QoS: Add PM_QOS_MEMORY_BANDWIDTH classTomeu Vizoso2014-09-251-1/+26
| | | | | | | | | | | | | | | | | | | | | | Also adds a class type PM_QOS_SUM that aggregates the values by summing them. It can be used by memory controllers to calculate the optimum clock frequency based on the bandwidth needs of the different memory clients. Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com> Acked-by: Pavel Machek <pavel@ucw.cz> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>