linux-toradex.git/drivers/cpufreq, branch v3.10.3

cpufreq: Revert commit 2f7021a8 to fix CPU hotplug regression

2013-07-25T21:07:23+00:00

commit e8d05276f236ee6435e78411f62be9714e0b9377 upstream.

commit 2f7021a8 "cpufreq: protect 'policy->cpus' from offlining
during __gov_queue_work()" caused a regression in CPU hotplug,
because it lead to a deadlock between cpufreq governor worker thread
and the CPU hotplug writer task.

Lockdep splat corresponding to this deadlock is shown below:

[   60.277396] ======================================================
[   60.277400] [ INFO: possible circular locking dependency detected ]
[   60.277407] 3.10.0-rc7-dbg-01385-g241fd04-dirty #1744 Not tainted
[   60.277411] -------------------------------------------------------
[   60.277417] bash/2225 is trying to acquire lock:
[   60.277422]  ((&(&j_cdbs->work)->work)){+.+...}, at: [] flush_work+0x5/0x280
[   60.277444] but task is already holding lock:
[   60.277449]  (cpu_hotplug.lock){+.+.+.}, at: [] cpu_hotplug_begin+0x2b/0x60
[   60.277465] which lock already depends on the new lock.

[   60.277472] the existing dependency chain (in reverse order) is:
[   60.277477] -> #2 (cpu_hotplug.lock){+.+.+.}:
[   60.277490]        [] lock_acquire+0xa4/0x200
[   60.277503]        [] mutex_lock_nested+0x67/0x410
[   60.277514]        [] get_online_cpus+0x3c/0x60
[   60.277522]        [] gov_queue_work+0x2a/0xb0
[   60.277532]        [] cs_dbs_timer+0xc1/0xe0
[   60.277543]        [] process_one_work+0x1cd/0x6a0
[   60.277552]        [] worker_thread+0x121/0x3a0
[   60.277560]        [] kthread+0xdb/0xe0
[   60.277569]        [] ret_from_fork+0x7c/0xb0
[   60.277580] -> #1 (&j_cdbs->timer_mutex){+.+...}:
[   60.277592]        [] lock_acquire+0xa4/0x200
[   60.277600]        [] mutex_lock_nested+0x67/0x410
[   60.277608]        [] cs_dbs_timer+0x8d/0xe0
[   60.277616]        [] process_one_work+0x1cd/0x6a0
[   60.277624]        [] worker_thread+0x121/0x3a0
[   60.277633]        [] kthread+0xdb/0xe0
[   60.277640]        [] ret_from_fork+0x7c/0xb0
[   60.277649] -> #0 ((&(&j_cdbs->work)->work)){+.+...}:
[   60.277661]        [] __lock_acquire+0x1766/0x1d30
[   60.277669]        [] lock_acquire+0xa4/0x200
[   60.277677]        [] flush_work+0x3d/0x280
[   60.277685]        [] __cancel_work_timer+0x8a/0x120
[   60.277693]        [] cancel_delayed_work_sync+0x13/0x20
[   60.277701]        [] cpufreq_governor_dbs+0x529/0x6f0
[   60.277709]        [] cs_cpufreq_governor_dbs+0x17/0x20
[   60.277719]        [] __cpufreq_governor+0x48/0x100
[   60.277728]        [] __cpufreq_remove_dev.isra.14+0x80/0x3c0
[   60.277737]        [] cpufreq_cpu_callback+0x38/0x4c
[   60.277747]        [] notifier_call_chain+0x5d/0x110
[   60.277759]        [] __raw_notifier_call_chain+0xe/0x10
[   60.277768]        [] _cpu_down+0x88/0x330
[   60.277779]        [] cpu_down+0x36/0x50
[   60.277788]        [] store_online+0x98/0xd0
[   60.277796]        [] dev_attr_store+0x18/0x30
[   60.277806]        [] sysfs_write_file+0xdb/0x150
[   60.277818]        [] vfs_write+0xbd/0x1f0
[   60.277826]        [] SyS_write+0x4c/0xa0
[   60.277834]        [] tracesys+0xd0/0xd5
[   60.277842] other info that might help us debug this:

[   60.277848] Chain exists of:
  (&(&j_cdbs->work)->work) --> &j_cdbs->timer_mutex --> cpu_hotplug.lock

[   60.277864]  Possible unsafe locking scenario:

[   60.277869]        CPU0                    CPU1
[   60.277873]        ----                    ----
[   60.277877]   lock(cpu_hotplug.lock);
[   60.277885]                                lock(&j_cdbs->timer_mutex);
[   60.277892]                                lock(cpu_hotplug.lock);
[   60.277900]   lock((&(&j_cdbs->work)->work));
[   60.277907]  *** DEADLOCK ***

[   60.277915] 6 locks held by bash/2225:
[   60.277919]  #0:  (sb_writers#6){.+.+.+}, at: [] vfs_write+0x1c3/0x1f0
[   60.277937]  #1:  (&buffer->mutex){+.+.+.}, at: [] sysfs_write_file+0x3c/0x150
[   60.277954]  #2:  (s_active#61){.+.+.+}, at: [] sysfs_write_file+0xc3/0x150
[   60.277972]  #3:  (x86_cpu_hotplug_driver_mutex){+.+...}, at: [] cpu_hotplug_driver_lock+0x17/0x20
[   60.277990]  #4:  (cpu_add_remove_lock){+.+.+.}, at: [] cpu_down+0x22/0x50
[   60.278007]  #5:  (cpu_hotplug.lock){+.+.+.}, at: [] cpu_hotplug_begin+0x2b/0x60
[   60.278023] stack backtrace:
[   60.278031] CPU: 3 PID: 2225 Comm: bash Not tainted 3.10.0-rc7-dbg-01385-g241fd04-dirty #1744
[   60.278037] Hardware name: Acer             Aspire 5741G    /Aspire 5741G    , BIOS V1.20 02/08/2011
[   60.278042]  ffffffff8204e110 ffff88014df6b9f8 ffffffff815b3d90 ffff88014df6ba38
[   60.278055]  ffffffff815b0a8d ffff880150ed3f60 ffff880150ed4770 3871c4002c8980b2
[   60.278068]  ffff880150ed4748 ffff880150ed4770 ffff880150ed3f60 ffff88014df6bb00
[   60.278081] Call Trace:
[   60.278091]  [] dump_stack+0x19/0x1b
[   60.278101]  [] print_circular_bug+0x2b6/0x2c5
[   60.278111]  [] __lock_acquire+0x1766/0x1d30
[   60.278123]  [] ? __kernel_text_address+0x58/0x80
[   60.278134]  [] lock_acquire+0xa4/0x200
[   60.278142]  [] ? flush_work+0x5/0x280
[   60.278151]  [] flush_work+0x3d/0x280
[   60.278159]  [] ? flush_work+0x5/0x280
[   60.278169]  [] ? mark_held_locks+0x94/0x140
[   60.278178]  [] ? __cancel_work_timer+0x77/0x120
[   60.278188]  [] ? trace_hardirqs_on_caller+0xfd/0x1c0
[   60.278196]  [] __cancel_work_timer+0x8a/0x120
[   60.278206]  [] cancel_delayed_work_sync+0x13/0x20
[   60.278214]  [] cpufreq_governor_dbs+0x529/0x6f0
[   60.278225]  [] cs_cpufreq_governor_dbs+0x17/0x20
[   60.278234]  [] __cpufreq_governor+0x48/0x100
[   60.278244]  [] __cpufreq_remove_dev.isra.14+0x80/0x3c0
[   60.278255]  [] cpufreq_cpu_callback+0x38/0x4c
[   60.278265]  [] notifier_call_chain+0x5d/0x110
[   60.278275]  [] __raw_notifier_call_chain+0xe/0x10
[   60.278284]  [] _cpu_down+0x88/0x330
[   60.278292]  [] ? cpu_hotplug_driver_lock+0x17/0x20
[   60.278302]  [] cpu_down+0x36/0x50
[   60.278311]  [] store_online+0x98/0xd0
[   60.278320]  [] dev_attr_store+0x18/0x30
[   60.278329]  [] sysfs_write_file+0xdb/0x150
[   60.278337]  [] vfs_write+0xbd/0x1f0
[   60.278347]  [] ? fget_light+0x320/0x4b0
[   60.278355]  [] SyS_write+0x4c/0xa0
[   60.278364]  [] tracesys+0xd0/0xd5
[   60.280582] smpboot: CPU 1 is now offline

The intention of that commit was to avoid warnings during CPU
hotplug, which indicated that offline CPUs were getting IPIs from the
cpufreq governor's work items.  But the real root-cause of that
problem was commit a66b2e5 (cpufreq: Preserve sysfs files across
suspend/resume) because it totally skipped all the cpufreq callbacks
during CPU hotplug in the suspend/resume path, and hence it never
actually shut down the cpufreq governor's worker threads during CPU
offline in the suspend/resume path.

Reflecting back, the reason why we never suspected that commit as the
root-cause earlier, was that the original issue was reported with
just the halt command and nobody had brought in suspend/resume to the
equation.

The reason for _that_ in turn, as it turns out, is that earlier
halt/shutdown was being done by disabling non-boot CPUs while tasks
were frozen, just like suspend/resume....  but commit cf7df378a
(reboot: migrate shutdown/reboot to boot cpu) which came somewhere
along that very same time changed that logic: shutdown/halt no longer
takes CPUs offline.  Thus, the test-cases for reproducing the bug
were vastly different and thus we went totally off the trail.

Overall, it was one hell of a confusion with so many commits
affecting each other and also affecting the symptoms of the problems
in subtle ways.  Finally, now since the original problematic commit
(a66b2e5) has been completely reverted, revert this intermediate fix
too (2f7021a8), to fix the CPU hotplug deadlock.  Phew!

Reported-by: Sergey Senozhatsky 
Reported-by: Bartlomiej Zolnierkiewicz 
Signed-off-by: Srivatsa S. Bhat 
Tested-by: Peter Wu 
Signed-off-by: Rafael J. Wysocki 
Signed-off-by: Greg Kroah-Hartman

cpufreq: Revert commit a66b2e to fix suspend/resume regression

2013-07-25T21:07:23+00:00

commit aae760ed21cd690fe8a6db9f3a177ad55d7e12ab upstream.

commit a66b2e (cpufreq: Preserve sysfs files across suspend/resume)
has unfortunately caused several things in the cpufreq subsystem to
break subtly after a suspend/resume cycle.

The intention of that patch was to retain the file permissions of the
cpufreq related sysfs files across suspend/resume.  To achieve that,
the commit completely removed the calls to cpufreq_add_dev() and
__cpufreq_remove_dev() during suspend/resume transitions.  But the
problem is that those functions do 2 kinds of things:
  1. Low-level initialization/tear-down that are critical to the
     correct functioning of cpufreq-core.
  2. Kobject and sysfs related initialization/teardown.

Ideally we should have reorganized the code to cleanly separate these
two responsibilities, and skipped only the sysfs related parts during
suspend/resume.  Since we skipped the entire callbacks instead (which
also included some CPU and cpufreq-specific critical components),
cpufreq subsystem started behaving erratically after suspend/resume.

So revert the commit to fix the regression.  We'll revisit and address
the original goal of that commit separately, since it involves quite a
bit of careful code reorganization and appears to be non-trivial.

(While reverting the commit, note that another commit f51e1eb
 (cpufreq: Fix cpufreq regression after suspend/resume) already
 reverted part of the original set of changes.  So revert only the
 remaining ones).

Signed-off-by: Srivatsa S. Bhat 
Acked-by: Viresh Kumar 
Tested-by: Paul Bolle 
Signed-off-by: Rafael J. Wysocki 
Signed-off-by: Greg Kroah-Hartman

cpufreq: Fix cpufreq regression after suspend/resume

2013-07-13T18:42:27+00:00

commit f51e1eb63d9c28cec188337ee656a13be6980cfd upstream.

Toralf Förster reported that the cpufreq ondemand governor behaves erratically
(doesn't scale well) after a suspend/resume cycle. The problem was that the
cpufreq subsystem's idea of the cpu frequencies differed from the actual
frequencies set in the hardware after a suspend/resume cycle. Toralf bisected
the problem to commit a66b2e5 (cpufreq: Preserve sysfs files across
suspend/resume).

Among other (harmless) things, that commit skipped the call to
cpufreq_update_policy() in the resume path. But cpufreq_update_policy() plays
an important role during resume, because it is responsible for checking if
the BIOS changed the cpu frequencies behind our back and resynchronize the
cpufreq subsystem's knowledge of the cpu frequencies, and update them
accordingly.

So, restore the call to cpufreq_update_policy() in the resume path to fix
the cpufreq regression.

Reported-and-tested-by: Toralf Förster 
Signed-off-by: Srivatsa S. Bhat 
Signed-off-by: Rafael J. Wysocki 
Signed-off-by: Greg Kroah-Hartman

cpufreq: fix NULL pointer deference at od_set_powersave_bias()

2013-06-25T20:42:37+00:00

When initializing the default powersave_bias value, we need to first
make sure that this policy is running the ondemand governor.

Reported-and-tested-by: Tim Gardner 
Signed-off-by: Jacob Shin 
Signed-off-by: Rafael J. Wysocki

cpufreq: cpufreq-cpu0: use the exact frequency for clk_set_rate()

2013-06-05T11:51:29+00:00

clk_set_rate() isn't supposed to accept approximate frequencies, instead
a supported frequency should be obtained from clk_round_rate() and then
used to set the clock.

Signed-off-by: Guennadi Liakhovetski 
Acked-by: Shawn Guo 
Acked-by: Viresh Kumar 
Signed-off-by: Rafael J. Wysocki

cpufreq: protect 'policy->cpus' from offlining during __gov_queue_work()

2013-06-05T11:46:54+00:00

Jiri Kosina  and Borislav Petkov 
reported the warning:

[   51.616759] ------------[ cut here ]------------
[   51.621460] WARNING: at arch/x86/kernel/smp.c:123 native_smp_send_reschedule+0x58/0x60()
[   51.629638] Modules linked in: ext2 vfat fat loop snd_hda_codec_hdmi usbhid snd_hda_codec_realtek coretemp kvm_intel kvm snd_hda_intel snd_hda_codec crc32_pclmul crc32c_intel ghash_clmulni_intel snd_hwdep snd_pcm aesni_intel sb_edac aes_x86_64 ehci_pci snd_page_alloc glue_helper snd_timer xhci_hcd snd iTCO_wdt iTCO_vendor_support ehci_hcd edac_core lpc_ich acpi_cpufreq lrw gf128mul ablk_helper cryptd mperf usbcore usb_common soundcore mfd_core dcdbas evdev pcspkr processor i2c_i801 button microcode
[   51.675581] CPU: 0 PID: 244 Comm: kworker/1:1 Tainted: G        W    3.10.0-rc1+ #10
[   51.683407] Hardware name: Dell Inc. Precision T3600/0PTTT9, BIOS A08 01/24/2013
[   51.690901] Workqueue: events od_dbs_timer
[   51.695069]  0000000000000009 ffff88043a2f5b68 ffffffff8161441c ffff88043a2f5ba8
[   51.702602]  ffffffff8103e540 0000000000000033 0000000000000001 ffff88043d5f8000
[   51.710136]  00000000ffff0ce1 0000000000000001 ffff88044fc4fc08 ffff88043a2f5bb8
[   51.717691] Call Trace:
[   51.720191]  [] dump_stack+0x19/0x1b
[   51.725396]  [] warn_slowpath_common+0x70/0xa0
[   51.731473]  [] warn_slowpath_null+0x1a/0x20
[   51.737378]  [] native_smp_send_reschedule+0x58/0x60
[   51.744013]  [] wake_up_nohz_cpu+0x2d/0xa0
[   51.749745]  [] add_timer_on+0x8f/0x110
[   51.755214]  [] __queue_delayed_work+0x16e/0x1a0
[   51.761470]  [] ? try_to_grab_pending+0xd1/0x1a0
[   51.767724]  [] mod_delayed_work_on+0x5a/0xa0
[   51.773719]  [] gov_queue_work+0x4d/0xc0
[   51.779271]  [] od_dbs_timer+0xcb/0x170
[   51.784734]  [] process_one_work+0x1fd/0x540
[   51.790634]  [] ? process_one_work+0x192/0x540
[   51.796711]  [] worker_thread+0x122/0x380
[   51.802350]  [] ? rescuer_thread+0x320/0x320
[   51.808264]  [] kthread+0xea/0xf0
[   51.813200]  [] ? flush_kthread_worker+0x150/0x150
[   51.819644]  [] ret_from_fork+0x7c/0xb0
[   51.918165] nouveau E[     DRM] GPU lockup - switching to software fbcon
[   51.930505]  [] ? flush_kthread_worker+0x150/0x150
[   51.936994] ---[ end trace f419538ada83b5c5 ]---

It was caused by the policy->cpus changed during the process of
__gov_queue_work(), in other word, cpu offline happened.

Use get/put_online_cpus() to prevent the offline from happening while
__gov_queue_work() is running.

[rjw: The problem has been present since recent commit 031299b
(cpufreq: governors: Avoid unnecessary per cpu timer interrupts)]

References: https://lkml.org/lkml/2013/6/5/88
Reported-by: Borislav Petkov 
Reported-and-tested-by: Jiri Kosina 
Signed-off-by: Michael Wang 
Acked-by: Viresh Kumar 
Signed-off-by: Rafael J. Wysocki

acpi-cpufreq: set current frequency based on target P-State

2013-06-05T11:10:57+00:00

Commit 4b31e774 (Always set P-state on initialization) fixed bug
#4634 and caused the driver to always set the target P-State at
least once since the initial P-State may not be the desired one.
Commit 5a1c0228 (cpufreq: Avoid calling cpufreq driver's target()
routine if target_freq == policy->cur) caused a regression in
this behavior.

This fixes the regression by setting policy->cur based on the CPU's
target frequency rather than the CPU's current reported frequency
(which may be different).  This means that the P-State will be set
initially if the CPU's target frequency is different from the
governor's target frequency.

This fixes an issue where setting the default governor to
performance wouldn't correctly enable turbo mode on all cores.

Signed-off-by: Ross Lagerwall 
Reviewed-by: Len Brown 
Acked-by: Viresh Kumar 
Cc: 3.8+ 
Signed-off-by: Rafael J. Wysocki

Merge tag 'pm+acpi-3.10-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm

2013-05-26T03:32:00+00:00

Pull power management and ACPI fixes from Rafael Wysocki:

 - Additional CPU ID for the intel_pstate driver from Dirk Brandewie.

 - More cpufreq fixes related to ARM big.LITTLE support and locking from
   Viresh Kumar.

 - VIA C7 cpufreq build fix from Rafał Bilski.

 - ACPI power management fix making it possible to use device power
   states regardless of the CONFIG_PM setting from Rafael J Wysocki.

 - New ACPI video blacklist item from Bastian Triller.

* tag 'pm+acpi-3.10-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  ACPI / video: Add "Asus UL30A" to ACPI video detect blacklist
  cpufreq: arm_big_little_dt: Instantiate as platform_driver
  cpufreq: arm_big_little_dt: Register driver only if DT has valid data
  cpufreq / e_powersaver: Fix linker error when ACPI processor is a module
  cpufreq / intel_pstate: Add additional supported CPU ID
  cpufreq: Drop rwsem lock around CPUFREQ_GOV_POLICY_EXIT
  ACPI / PM: Allow device power states to be used for CONFIG_PM unset

cpufreq: arm_big_little_dt: Instantiate as platform_driver

2013-05-22T10:43:33+00:00

As multiplatform build is being adopted by more and more ARM platforms, initcall
function should be used very carefully. For example, when both arm_big_little_dt
and cpufreq-cpu0 drivers are compiled in, arm_big_little_dt driver may try to
register even if we had platform device for cpufreq-cpu0 registered.

To eliminate this undesired the effect, the patch changes arm_big_little_dt
driver to have it instantiated as a platform_driver. Then it will only run on
platforms that create the platform_device "arm-bL-cpufreq-dt".

Reported-and-tested-by: Rob Herring 
Signed-off-by: Viresh Kumar 
Signed-off-by: Rafael J. Wysocki

cpufreq: arm_big_little_dt: Register driver only if DT has valid data

2013-05-22T10:42:34+00:00

If arm_big_little_dt driver is enabled, then it will always try to register with
big LITTLE cpufreq core driver. In case DT doesn't have relevant data for cpu
nodes, i.e. operating points aren't present, then we should exit early and
shouldn't register with big LITTLE cpufreq core driver. Otherwise we will fail
continuously from the driver->init() routine.

This patch fixes this issue.

Reported-and-tested-by: Jon Medhurst 
Reviewed-by: Jon Medhurst 
Signed-off-by: Viresh Kumar 
Signed-off-by: Rafael J. Wysocki