firefly-linux-kernel-4.4.55.git
11 years agoaf_unix: use freezable blocking calls in read
Colin Cross [Mon, 6 May 2013 23:50:21 +0000 (23:50 +0000)]
af_unix: use freezable blocking calls in read

Avoid waking up every thread sleeping in read call on an AF_UNIX
socket during suspend and resume by calling a freezable blocking
call.  Previous patches modified the freezer to avoid sending
wakeups to threads that are blocked in freezable blocking calls.

This call was selected to be converted to a freezable call because
it doesn't hold any locks or release any resources when interrupted
that might be needed by another freezing task or a kernel driver
during suspend, and is a common site where idle userspace tasks are
blocked.

Change-Id: I788246a76780ea892659526e70be018b18f646c4
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Colin Cross <ccross@android.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
11 years agosigtimedwait: use freezable blocking call
Colin Cross [Mon, 6 May 2013 23:50:20 +0000 (23:50 +0000)]
sigtimedwait: use freezable blocking call

Avoid waking up every thread sleeping in a sigtimedwait call during
suspend and resume by calling a freezable blocking call.  Previous
patches modified the freezer to avoid sending wakeups to threads
that are blocked in freezable blocking calls.

This call was selected to be converted to a freezable call because
it doesn't hold any locks or release any resources when interrupted
that might be needed by another freezing task or a kernel driver
during suspend, and is a common site where idle userspace tasks are
blocked.

Change-Id: Ic27469b60a67d50cdc0d0c78975951a99c25adcd
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Colin Cross <ccross@android.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
11 years agonanosleep: use freezable blocking call
Colin Cross [Mon, 6 May 2013 23:50:19 +0000 (23:50 +0000)]
nanosleep: use freezable blocking call

Avoid waking up every thread sleeping in a nanosleep call during
suspend and resume by calling a freezable blocking call.  Previous
patches modified the freezer to avoid sending wakeups to threads
that are blocked in freezable blocking calls.

This call was selected to be converted to a freezable call because
it doesn't hold any locks or release any resources when interrupted
that might be needed by another freezing task or a kernel driver
during suspend, and is a common site where idle userspace tasks are
blocked.

Change-Id: I93383201d4dd62130cd9a9153842d303fc2e2986
Acked-by: Tejun Heo <tj@kernel.org>
Acked-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Colin Cross <ccross@android.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
11 years agofutex: use freezable blocking call
Colin Cross [Mon, 6 May 2013 23:50:18 +0000 (23:50 +0000)]
futex: use freezable blocking call

Avoid waking up every thread sleeping in a futex_wait call during
suspend and resume by calling a freezable blocking call.  Previous
patches modified the freezer to avoid sending wakeups to threads
that are blocked in freezable blocking calls.

This call was selected to be converted to a freezable call because
it doesn't hold any locks or release any resources when interrupted
that might be needed by another freezing task or a kernel driver
during suspend, and is a common site where idle userspace tasks are
blocked.

Change-Id: I9ccab9c2d201adb66c85432801cdcf43fc91e94f
Acked-by: Tejun Heo <tj@kernel.org>
Acked-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Colin Cross <ccross@android.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
11 years agoselect: use freezable blocking call
Colin Cross [Mon, 6 May 2013 23:50:17 +0000 (23:50 +0000)]
select: use freezable blocking call

Avoid waking up every thread sleeping in a select call during
suspend and resume by calling a freezable blocking call.  Previous
patches modified the freezer to avoid sending wakeups to threads
that are blocked in freezable blocking calls.

This call was selected to be converted to a freezable call because
it doesn't hold any locks or release any resources when interrupted
that might be needed by another freezing task or a kernel driver
during suspend, and is a common site where idle userspace tasks are
blocked.

Change-Id: I0d7565ec0b6bc5d44cb55f958589c56e6bd16348
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Colin Cross <ccross@android.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
11 years agoepoll: use freezable blocking call
Colin Cross [Mon, 6 May 2013 23:50:16 +0000 (23:50 +0000)]
epoll: use freezable blocking call

Avoid waking up every thread sleeping in an epoll_wait call during
suspend and resume by calling a freezable blocking call.  Previous
patches modified the freezer to avoid sending wakeups to threads
that are blocked in freezable blocking calls.

This call was selected to be converted to a freezable call because
it doesn't hold any locks or release any resources when interrupted
that might be needed by another freezing task or a kernel driver
during suspend, and is a common site where idle userspace tasks are
blocked.

Change-Id: I848d08d28c89302fd42bbbdfa76489a474ab27bf
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Colin Cross <ccross@android.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
11 years agobinder: use freezable blocking calls
Colin Cross [Mon, 6 May 2013 23:50:15 +0000 (23:50 +0000)]
binder: use freezable blocking calls

Avoid waking up every thread sleeping in a binder call during
suspend and resume by calling a freezable blocking call.  Previous
patches modified the freezer to avoid sending wakeups to threads
that are blocked in freezable blocking calls.

This call was selected to be converted to a freezable call because
it doesn't hold any locks or release any resources when interrupted
that might be needed by another freezing task or a kernel driver
during suspend, and is a common site where idle userspace tasks are
blocked.

Change-Id: Ic4458ae90447f6caa895cc62f08e515caa7790ba
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Colin Cross <ccross@android.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
11 years agofreezer: add new freezable helpers using freezer_do_not_count()
Colin Cross [Mon, 6 May 2013 23:50:14 +0000 (23:50 +0000)]
freezer: add new freezable helpers using freezer_do_not_count()

Freezing tasks will wake up almost every userspace task from
where it is blocking and force it to run until it hits a
call to try_to_sleep(), generally on the exit path from the syscall
it is blocking in.  On resume each task will run again, usually
restarting the syscall and running until it hits the same
blocking call as it was originally blocked in.

To allow tasks to avoid running on every suspend/resume cycle,
this patch adds additional freezable wrappers around blocking calls
that call freezer_do_not_count().  Combined with the previous patch,
these tasks will not run during suspend or resume unless they wake
up for another reason, in which case they will run until they hit
the try_to_freeze() in freezer_count(), and then continue processing
the wakeup after tasks are thawed.

Additional patches will convert the most common locations that
userspace blocks in to use freezable helpers.

Change-Id: Id909760ce460f2532801a4b00d344f0816bfefc9
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Colin Cross <ccross@android.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
11 years agofreezer: convert freezable helpers to static inline where possible
Colin Cross [Mon, 6 May 2013 23:50:13 +0000 (23:50 +0000)]
freezer: convert freezable helpers to static inline where possible

Some of the freezable helpers have to be macros because their
condition argument needs to get evaluated every time through
the wait loop.  Convert the others to static inline to make
future changes easier.

Change-Id: I69d3fc10d26522cb9bf3a616ff4f21245f9c071a
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Colin Cross <ccross@android.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
11 years agofreezer: convert freezable helpers to freezer_do_not_count()
Colin Cross [Mon, 6 May 2013 23:50:12 +0000 (23:50 +0000)]
freezer: convert freezable helpers to freezer_do_not_count()

Freezing tasks will wake up almost every userspace task from
where it is blocking and force it to run until it hits a
call to try_to_sleep(), generally on the exit path from the syscall
it is blocking in.  On resume each task will run again, usually
restarting the syscall and running until it hits the same
blocking call as it was originally blocked in.

Convert the existing wait_event_freezable* wrappers to use
freezer_do_not_count().  Combined with a previous patch,
these tasks will not run during suspend or resume unless they wake
up for another reason, in which case they will run until they hit
the try_to_freeze() in freezer_count(), and then continue processing
the wakeup after tasks are thawed.

This results in a small change in behavior, previously a race
between freezing and a normal wakeup would be won by the wakeup,
now the task will freeze and then handle the wakeup after thawing.

Change-Id: I532e62251f58c1a9ca488b3fb6220c53acf7d33d
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Colin Cross <ccross@android.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
11 years agofreezer: skip waking up tasks with PF_FREEZER_SKIP set
Colin Cross [Mon, 6 May 2013 23:50:11 +0000 (23:50 +0000)]
freezer: skip waking up tasks with PF_FREEZER_SKIP set

Android goes through suspend/resume very often (every few seconds when
on a busy wifi network with the screen off), and a significant portion
of the energy used to go in and out of suspend is spent in the
freezer.  If a task has called freezer_do_not_count(), don't bother
waking it up.  If it happens to wake up later it will call
freezer_count() and immediately enter the refrigerator.

Combined with patches to convert freezable helpers to use
freezer_do_not_count() and convert common sites where idle userspace
tasks are blocked to use the freezable helpers, this reduces the
time and energy required to suspend and resume.

Change-Id: I6ba019d24273619849af757a413271da3261d7db
Acked-by: Tejun Heo <tj@kernel.org>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Colin Cross <ccross@android.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
11 years agofreezer: shorten freezer sleep time using exponential backoff
Colin Cross [Mon, 6 May 2013 23:50:10 +0000 (23:50 +0000)]
freezer: shorten freezer sleep time using exponential backoff

All tasks can easily be frozen in under 10 ms, switch to using
an initial 1 ms sleep followed by exponential backoff until
8 ms.  Also convert the printed time to ms instead of centiseconds.

Change-Id: I7b198b16eefb623c2b0fc45dce50d9bca320afdc
Acked-by: Pavel Machek <pavel@ucw.cz>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Colin Cross <ccross@android.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
11 years agolockdep: check that no locks held at freeze time
Mandeep Singh Baines [Mon, 6 May 2013 23:50:09 +0000 (23:50 +0000)]
lockdep: check that no locks held at freeze time

We shouldn't try_to_freeze if locks are held.  Holding a lock can cause a
deadlock if the lock is later acquired in the suspend or hibernate path
(e.g.  by dpm).  Holding a lock can also cause a deadlock in the case of
cgroup_freezer if a lock is held inside a frozen cgroup that is later
acquired by a process outside that group.

History:
This patch was originally applied as 6aa9707099c and reverted in
dbf520a9d7d4 because NFS was freezing with locks held.  It was
deemed better to keep the bad freeze point in NFS to allow laptops
to suspend consistently.  The previous patch in this series converts
NFS to call _unsafe versions of the freezable helpers so that
lockdep doesn't complain about them until a more correct fix
can be applied.

Change-Id: Ib9d4299fb75a39e611b868be42e413909a994baa
[akpm@linux-foundation.org: export debug_check_no_locks_held]
Signed-off-by: Mandeep Singh Baines <msb@chromium.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Acked-by: Pavel Machek <pavel@ucw.cz>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Colin Cross <ccross@android.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
11 years agolockdep: remove task argument from debug_check_no_locks_held
Colin Cross [Mon, 6 May 2013 23:50:08 +0000 (23:50 +0000)]
lockdep: remove task argument from debug_check_no_locks_held

The only existing caller to debug_check_no_locks_held calls it
with 'current' as the task, and the freezer needs to call
debug_check_no_locks_held but doesn't already have a current
task pointer, so remove the argument.  It is already assuming
that the current task is relevant by dumping the current stack
trace as part of the warning.

This was originally part of 6aa9707099c (lockdep: check that
no locks held at freeze time) which was reverted in
dbf520a9d7d4.

Change-Id: Idbaf1332ce6c80dc49c1d31c324c7fbf210657c5
Original-author: Mandeep Singh Baines <msb@chromium.org>
Acked-by: Pavel Machek <pavel@ucw.cz>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Colin Cross <ccross@android.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
11 years agofreezer: add unsafe versions of freezable helpers for CIFS
Colin Cross [Tue, 7 May 2013 17:52:05 +0000 (17:52 +0000)]
freezer: add unsafe versions of freezable helpers for CIFS

CIFS calls wait_event_freezekillable_unsafe with a VFS lock held,
which is unsafe and will cause lockdep warnings when 6aa9707
"lockdep: check that no locks held at freeze time" is reapplied
(it was reverted in dbf520a).  CIFS shouldn't be doing this, but
it has long-running syscalls that must hold a lock but also
shouldn't block suspend.  Until CIFS freeze handling is rewritten
to use a signal to exit out of the critical section, add a new
wait_event_freezekillable_unsafe helper that will not run the
lockdep test when 6aa9707 is reapplied, and call it from CIFS.

In practice the likley result of holding the lock while freezing
is that a second task blocked on the lock will never freeze,
aborting suspend, but it is possible to manufacture a case using
the cgroup freezer, the lock, and the suspend freezer to create
a deadlock.  Silencing the lockdep warning here will allow
problems to be found in other drivers that may have a more
serious deadlock risk, and prevent new problems from being added.

Change-Id: I420c5392bacf68e58e268293b2b36068ad4df753
Acked-by: Pavel Machek <pavel@ucw.cz>
Acked-by: Tejun Heo <tj@kernel.org>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Colin Cross <ccross@android.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
11 years agofreezer: add unsafe versions of freezable helpers for NFS
Colin Cross [Mon, 6 May 2013 23:50:06 +0000 (23:50 +0000)]
freezer: add unsafe versions of freezable helpers for NFS

NFS calls the freezable helpers with locks held, which is unsafe
and will cause lockdep warnings when 6aa9707 "lockdep: check
that no locks held at freeze time" is reapplied (it was reverted
in dbf520a).  NFS shouldn't be doing this, but it has
long-running syscalls that must hold a lock but also shouldn't
block suspend.  Until NFS freeze handling is rewritten to use a
signal to exit out of the critical section, add new *_unsafe
versions of the helpers that will not run the lockdep test when
6aa9707 is reapplied, and call them from NFS.

In practice the likley result of holding the lock while freezing
is that a second task blocked on the lock will never freeze,
aborting suspend, but it is possible to manufacture a case using
the cgroup freezer, the lock, and the suspend freezer to create
a deadlock.  Silencing the lockdep warning here will allow
problems to be found in other drivers that may have a more
serious deadlock risk, and prevent new problems from being added.

Change-Id: Ia17d32cdd013a6517bdd5759da900970a4427170
Signed-off-by: Colin Cross <ccross@android.com>
Acked-by: Pavel Machek <pavel@ucw.cz>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
11 years agotimerfd: add alarm timers
Todd Poynor [Wed, 8 May 2013 22:50:50 +0000 (15:50 -0700)]
timerfd: add alarm timers

Add support for clocks CLOCK_REALTIME_ALARM and CLOCK_BOOTTIME_ALARM.

Change-Id: Iafc8445d3d7ffb35110c860f1607bf03f1edb895
Signed-off-by: Todd Poynor <toddpoynor@google.com>
11 years agoalarmtimer: add alarm_expires_remaining
Todd Poynor [Wed, 8 May 2013 03:43:29 +0000 (20:43 -0700)]
alarmtimer: add alarm_expires_remaining

Similar to hrtimer_expires_remaining, return the amount of time
remaining until alarm expiry.

Change-Id: I8c57512d619ac66bcdaf2d9ccdf0d7f74af2ff66
Signed-off-by: Todd Poynor <toddpoynor@google.com>
11 years agoalarmtimer: add alarm_start_relative
Todd Poynor [Sat, 11 May 2013 00:41:06 +0000 (17:41 -0700)]
alarmtimer: add alarm_start_relative

Start an alarmtimer with an expires time relative to the current time
of the associated clock.

Change-Id: Ifb5309a15e0d502bb4d0209ca5510a56ee7fa88c
Signed-off-by: Todd Poynor <toddpoynor@google.com>
11 years agoalarmtimer: add alarm_forward_now
Todd Poynor [Sat, 11 May 2013 00:10:42 +0000 (17:10 -0700)]
alarmtimer: add alarm_forward_now

Similar to hrtimer_forward_now, move the expires time forward to an
interval from the current time of the associated clock.

Change-Id: I73fed223321167507b6eddcb7a57d235ffcfc1be
Signed-off-by: Todd Poynor <toddpoynor@google.com>
11 years agoalarmtimer: add alarm_restart
Todd Poynor [Wed, 8 May 2013 22:49:18 +0000 (15:49 -0700)]
alarmtimer: add alarm_restart

Analogous to hrtimer_restart, restart an alarmtimer after the expires
time has already been updated (as with alarm_forward).

Change-Id: Ia2613bbb467404cb2c35c11efa772bc56294963a
Signed-off-by: Todd Poynor <toddpoynor@google.com>
11 years agocpufreq: interactive: avoid underflow on active time calculation
Minsung Kim [Tue, 23 Apr 2013 13:32:01 +0000 (22:32 +0900)]
cpufreq: interactive: avoid underflow on active time calculation

Check for idle time delta less than elapsed time delta, avoid
underflow computing active time.

Change-Id: I3e4c6ef1ad794eec49ed379c0c50fa727fd6ad28
Signed-off-by: Minsung Kim <ms925.kim@samsung.com>
11 years agocpufreq: interactive: reduce chance of zero time delta on load eval
Todd Poynor [Fri, 5 Apr 2013 20:25:21 +0000 (13:25 -0700)]
cpufreq: interactive: reduce chance of zero time delta on load eval

Reschedule load sampling timer after timestamp of sample start taken,
hold spinlock across entire sequence to avoid preemption.  Avoid the
WARN for zero time delta in the load sampling timer function.

Change-Id: Idc10a756f09141decb6df92669521a1ebf0dbc10
Signed-off-by: Todd Poynor <toddpoynor@google.com>
11 years agocpufreq: interactive: handle errors from cpufreq_frequency_table_target
Todd Poynor [Mon, 22 Apr 2013 23:44:58 +0000 (16:44 -0700)]
cpufreq: interactive: handle errors from cpufreq_frequency_table_target

Add checks for error return from cpufreq_frequency_table_target, and be
less noisy on the existing call with an error check.  CPU hotplug and
system shutdown may cause this call to return -EINVAL.

Bug: 8613560
Change-Id: Id78d8829920462c0db1c7e14e717d91740d6cb44
Signed-off-by: Todd Poynor <toddpoynor@google.com>
11 years agoandroid: configs: no MODULES for base, no SIP for recommended
JP Abgrall [Fri, 7 Jun 2013 22:23:31 +0000 (15:23 -0700)]
android: configs: no MODULES for base, no SIP for recommended

We don't like CONFIG_MODULES anymore.
Connection tracker handling of large SIP fails.

Change-Id: Ie3c65aefcc6181752d6656c97e63035e5b5653ff
Signed-off-by: JP Abgrall <jpa@google.com>
11 years agoandroid: configs: Reorder config fragments
JP Abgrall [Fri, 7 Jun 2013 18:56:34 +0000 (11:56 -0700)]
android: configs: Reorder config fragments

Because there is not tool to consistently generate these config
fragments, lets keep the alphabetical instead of random.

Change-Id: I0f098f6be6bdd272544295a3d5a48d04411e4514
Signed-off-by: JP Abgrall <jpa@google.com>
11 years agoashmem: avoid deadlock between read and mmap calls
Todd Poynor [Wed, 5 Jun 2013 00:29:38 +0000 (17:29 -0700)]
ashmem: avoid deadlock between read and mmap calls

Avoid holding ashmem_mutex across code that can page fault.  Page faults
grab the mmap_sem for the process, which are also held by mmap calls
prior to calling ashmem_mmap, which locks ashmem_mutex.  The reversed
order of locking between the two can deadlock.

The calls that can page fault are read() and the ASHMEM_SET_NAME and
ASHMEM_GET_NAME ioctls.  Move the code that accesses userspace pages
outside the ashmem_mutex.

Bug: 9261835
Change-Id: If1322e981d29c889a56cdc9dfcbc6df2729a45e9
Signed-off-by: Todd Poynor <toddpoynor@google.com>
Conflicts:
drivers/staging/android/ashmem.c

11 years agoandroid: configs: Enable KSM support by default
Rom Lemarchand [Fri, 31 May 2013 18:54:35 +0000 (11:54 -0700)]
android: configs: Enable KSM support by default

This will be a noop unless the KSM thread is enabled by userspace

Change-Id: Ia5fde14504cc0da50522e2f875d8d021f9e054ba
Signed-off-by: Rom Lemarchand <romlem@google.com>
11 years agoandroid: configs: Add Logitech unifying receivers to recommended
Michael Wright [Wed, 8 May 2013 21:50:11 +0000 (14:50 -0700)]
android: configs: Add Logitech unifying receivers to recommended

Change-Id: I7647cd7037731df69dfdd513a0808b396d9d5bdd
Signed-off-by: Michael Wright <michaelwr@android.com>
11 years agogpu: ion: Also shrink memory cached in the deferred free list
Rebecca Schultz Zavin [Thu, 23 May 2013 20:37:25 +0000 (13:37 -0700)]
gpu: ion: Also shrink memory cached in the deferred free list

When the system is low on memory, we want to shrink any cached
system memory ion is holding.  Previously we were shrinking memory
in the page pools, but not in the deferred free list.  This patch
makes it possible to shrink both.  It also moves the shrinker
code into the heaps so they can correctly manage any caches they
might contain.

Change-Id: I177f587f999a5220eddbf7af94745aae736cac75
Signed-off-by: Rebecca Schultz Zavin <rebecca@android.com>
11 years agomisc: uidstat: avoid create_stat() race and blockage.
JP Abgrall [Thu, 30 May 2013 22:31:17 +0000 (15:31 -0700)]
misc: uidstat: avoid create_stat() race and blockage.

* create_stat() race would lead to:
  [   58.132324] proc_dir_entry 'uid_stat/10061' already registered

* blocking kmalloc reported by sbranden
 tcp_read_sock()
  uid_stat_tcp_rcv()
    create_stat()
      kmalloc(GFP_KERNEL)

Signed-off-by: JP Abgrall <jpa@google.com>
11 years agommc: Add tracepoints of mmc block operations
Ken Sumrall [Thu, 16 May 2013 03:13:13 +0000 (20:13 -0700)]
mmc: Add tracepoints of mmc block operations

Add tracepoints to record the start and end of each mmc block
operation.  This includes read, write, erase, secure erase,
trim, secure trim1 and secure trim 2, discard and
sanitize commands.

Change-Id: Ic5d1cbdb9adb940d8b1a2a13c73970023575df50
Signed-off-by: Ken Sumrall <ksumrall@android.com>
11 years agolowmemorykiller: make default lowmemorykiller debug message useful
Colin Cross [Fri, 3 May 2013 21:57:29 +0000 (14:57 -0700)]
lowmemorykiller: make default lowmemorykiller debug message useful

lowmemorykiller debug messages are inscrutable and mostly useful
for debugging the lowmemorykiller, not explaining why a process
was killed.  Make the messages more useful by prefixing them
with "lowmemorykiller: " and explaining in more readable terms
what was killed, who it was killed for, and why it was killed.

The messages now look like:
[   76.997631] lowmemorykiller: Killing 'droid.gallery3d' (2172), adj 1000,
[   76.997635]    to free 27436kB on behalf of 'kswapd0' (29) because
[   76.997638]    cache 122624kB is below limit 122880kB for oom_score_adj 1000
[   76.997641]    Free memory is -53356kB above reserved

A negative number for free memory above reserved means some of the
reserved memory has been used and is being regenerated by kswapd,
which is likely what called the shrinkers.

Change-Id: I1fe983381e73e124b90aa5d91cb66e55eaca390f
Signed-off-by: Colin Cross <ccross@android.com>
11 years agoARM: fault: assume no context when IRQs are disabled during data abort.
JP Abgrall [Mon, 29 Apr 2013 23:07:00 +0000 (16:07 -0700)]
ARM: fault: assume no context when IRQs are disabled during data abort.

Bail out early if IRQs are disabled in do_page_fault or else
  [14415.157266] BUG: sleeping function called from invalid context at arch/arm/mm/fault.c:301

Russell King's idea from
  http://comments.gmane.org/gmane.linux.ports.arm.omap/59256

Signed-off-by: JP Abgrall <jpa@google.com>
11 years agocpufreq: interactive: fix uninitialized spinlock
Minsung Kim [Tue, 16 Apr 2013 12:52:50 +0000 (21:52 +0900)]
cpufreq: interactive: fix uninitialized spinlock

Add missing spinlock init

Backtrace:
[<c0011ce4>] (dump_backtrace+0x0/0x10c) from [<c0662a68>] (dump_stack+0x18/0x1c)
 r6:00000032 r5:c0bd09ec r4:e6848000 r3:00000000
[<c0662a50>] (dump_stack+0x0/0x1c) from [<c06670b0>] (spin_dump+0x80/0x94)
[<c0667030>] (spin_dump+0x0/0x94) from [<c06670f0>] (spin_bug+0x2c/0x30)
 r5:c08f91fc r4:c0bd09ec
[<c06670c4>] (spin_bug+0x0/0x30) from [<c0245f74>] (do_raw_spin_unlock+0x88/0xcc)
 r5:e547bac0 r4:c0bd09ec
[<c0245eec>] (do_raw_spin_unlock+0x0/0xcc) from [<c066c9cc>] (_raw_spin_unlock_irqrestore+0x14/0x40)
 r5:e547bac0 r4:60000013
[<c066c9b8>] (_raw_spin_unlock_irqrestore+0x0/0x40) from [<c044b884>] (store_above_hispeed_delay+0x6c/0x80)
 r4:c0b4cf78 r3:00000007
[<c044b818>] (store_above_hispeed_delay+0x0/0x80) from [<c0235d24>] (kobj_attr_store+0x1c/0x28)
 r7:e68ff000 r6:00000032 r5:e58137c0 r4:e61cde80
[<c0235d08>] (kobj_attr_store+0x0/0x28) from [<c0156b78>] (sysfs_write_file+0x104/0x184)
[<c0156a74>] (sysfs_write_file+0x0/0x184) from [<c0100680>] (vfs_write+0xb0/0x140)
[<c01005d0>] (vfs_write+0x0/0x140) from [<c0100900>] (sys_write+0x44/0x70)
 r8:00000000 r7:00000004 r6:00000032 r5:bee43c90 r4:e5600300
[<c01008bc>] (sys_write+0x0/0x70) from [<c000e400>] (ret_fast_syscall+0x0/0x30)
 r9:e6842000 r8:c000e584 r6:00000032 r5:bee43c90 r4:00000009

Change-Id: I80a1e0b3fecb24adba501ff44f568479deeff7fa
Signed-off-by: Minsung Kim <ms925.kim@samsung.com>
11 years agoandroid: configs: Initial commit of Android config fragments
Rom Lemarchand [Wed, 6 Feb 2013 23:49:47 +0000 (15:49 -0800)]
android: configs: Initial commit of Android config fragments

- Add 2 files that contain the minimal and recommended kernel config
options respectively.
- Add a README to explain their purpose and how to use them to
generate a device config compatible with Android.

Change-Id: I3a4883f3b04d2820e90ceb3c4d02390d6458d6ce
Signed-off-by: Rom Lemarchand <romlem@google.com>
11 years agogpu: ion: Fix bug in ion shrinker
Rebecca Schultz Zavin [Wed, 24 Apr 2013 21:33:55 +0000 (14:33 -0700)]
gpu: ion: Fix bug in ion shrinker

The high variable was sometimes used uninitialized

Change-Id: I2f51413fd2d063fdff325047e824dc8c749d9e0a
Signed-off-by: Rebecca Schultz Zavin <rebecca@android.com>
11 years agotrace/events: fix gpu event timestamp formatting
Jamie Gennis [Fri, 19 Apr 2013 03:36:21 +0000 (20:36 -0700)]
trace/events: fix gpu event timestamp formatting

This change fixes the how the gpu_sched_switch timestamp field is formatted.

Signed-off-by: Jamie Gennis <jgennis@google.com>
Change-Id: I273234935254ed15772c9e561c9af20e480004ae

11 years agoARM: convert build of appended dtb zImage to list of dtbs
Colin Cross [Wed, 17 Apr 2013 23:58:36 +0000 (16:58 -0700)]
ARM: convert build of appended dtb zImage to list of dtbs

Allow CONFIG_BUILD_ARM_APPENDED_DTB_IMAGE_NAMES to specify
a space separated list of dtbs to append to the zImage,
and name the resulting file zImage-dtb

Change-Id: I36d9108a2349bdbb373e95076dcb1417d8c7dce6
Signed-off-by: Colin Cross <ccross@android.com>
Conflicts:
arch/arm/boot/Makefile
scripts/Makefile.lib

11 years agogpu: ion: ion_chunk_heap: Zero chunk heap memory at creation time
Rebecca Schultz Zavin [Mon, 1 Apr 2013 23:40:02 +0000 (16:40 -0700)]
gpu: ion: ion_chunk_heap: Zero chunk heap memory at creation time

Allocations from the ion heap need to be zeroed to protect userspace
from seeing memory belonging to other processes.  First allocations
from this heap were not zero'd allowing users to see memory from other
processes on a warm reset.

Change-Id: I524a7b79cb76c390c870fcf8b30d213185fc85a0
Signed-off-by: Rebecca Schultz Zavin <rebecca@android.com>
11 years agogpu: ion: fix kfree/list_del order
JP Abgrall [Mon, 15 Apr 2013 21:23:14 +0000 (14:23 -0700)]
gpu: ion: fix kfree/list_del order

With CONFIG_SLUB_DEBUG_ON it would panic during
ion_alloc()
 ion_buffer_create()
   io_heap_drain_freelist()

Signed-off-by: JP Abgrall <jpa@google.com>
11 years agonetfilter: qtaguid: rate limit some of the printks
JP Abgrall [Mon, 8 Apr 2013 22:09:26 +0000 (15:09 -0700)]
netfilter: qtaguid: rate limit some of the printks

Some of the printks are in the packet handling path.
We now ratelimit the very unlikely errors to avoid
kmsg spamming.

Signed-off-by: JP Abgrall <jpa@google.com>
11 years agocpufreq: interactive: base above_hispeed_delay on target freq, not current
Todd Poynor [Fri, 22 Mar 2013 03:46:00 +0000 (20:46 -0700)]
cpufreq: interactive: base above_hispeed_delay on target freq, not current

Time to wait should be based on the intended target speed, not the
actual speed (which may be held high by another CPU).

Change-Id: Ifc5bb55d06adddb9a02af90af05398a78f282272
Reported-by: Arve Hjønnevåg <arve@android.com>
Signed-off-by: Todd Poynor <toddpoynor@google.com>
11 years agoARM: add config option to build zImage/dtb combo
Erik Gilling [Mon, 25 Mar 2013 22:04:41 +0000 (15:04 -0700)]
ARM: add config option to build zImage/dtb combo

Allows a defconfig to set a default dtb to concatenate with a zImage
to create a zImage-dtb.<dtb name>

Signed-off-by: Erik Gilling <konkers@android.com>
Change-Id: I34b643b1c49228fbae88a56e46c93c478089620d

11 years agocpufreq: interactive: fix crash on error paths in get_tokenized_data
Todd Poynor [Wed, 20 Mar 2013 22:40:46 +0000 (15:40 -0700)]
cpufreq: interactive: fix crash on error paths in get_tokenized_data

Use separate variable for error code, free proper pointer.

Change-Id: Ia83cccb195997789ac6afbf5b8761f7b278196d6
Reported-by: Arve Hjønnevåg <arve@android.com>
Signed-off-by: Todd Poynor <toddpoynor@google.com>
11 years agocpufreq: interactive: add io_is_busy interface
Lianwei Wang [Fri, 22 Feb 2013 03:39:18 +0000 (11:39 +0800)]
cpufreq: interactive: add io_is_busy interface

Previously the idle time returned from get_cpu_idle_time_us included the
iowait time. So the iowait time was always calculated as idle time.

But now the idle time returned from get_cpu_idle_time_us does not include
the iowait time anymore because of below commit which cause the iowait time
always calculated as busy time:
    6beea0c nohz: Fix update_ts_time_stat idle accounting

Add the io_is_busy interface, as does the ondemand governor, and let the user
configure the iowait time as busy or idle through the io_is_busy sysfs
interface.

By default, io_is_busy is disabled.

[toddpoynor@google.com: minor updates]
Change-Id: If7d70ff864c43bc9c8d7fd7cfc66f930d339f9b4
Signed-off-by: Lianwei Wang <lian-wei.wang@motorola.com>
Signed-off-by: Todd Poynor <toddpoynor@google.com>
11 years agocpufreq: interactive: allow arbitrary speed / delay mappings
Minsung Kim [Mon, 25 Feb 2013 14:48:04 +0000 (23:48 +0900)]
cpufreq: interactive: allow arbitrary speed / delay mappings

Accept a string of delays and speeds at which to apply the delay before
raising each step above hispeed. For example, "80000 1300000:200000
1500000:40000" means that the delay at or above 1GHz, until 1.3GHz is 80 msecs,
the delay until 1.5GHz is 200 msecs and the delay at or above 1.5GHz is 40
msecs when hispeed_freq is 1GHz.

[toddpoynor@google.com: add documentation]
Change-Id: Ifeebede8b1acbdd0a53e5c6916bccbf764dc854f
Signed-off-by: Minsung Kim <ms925.kim@samsung.com>
11 years agotrace: add non-hierarchical function_graph option
Jamie Gennis [Thu, 22 Nov 2012 04:14:09 +0000 (20:14 -0800)]
trace: add non-hierarchical function_graph option

Add the 'funcgraph-flat' option to the function_graph tracer to use the default
trace printing format rather than the hierarchical formatting normally used.

Change-Id: If2900bfb86e6f8f51379f56da4f6fabafa630909
Signed-off-by: Jamie Gennis <jgennis@google.com>
11 years agogpu: ion: Make ion_free asynchronous
Rebecca Schultz Zavin [Wed, 13 Feb 2013 22:48:11 +0000 (14:48 -0800)]
gpu: ion: Make ion_free asynchronous

Add the ability for a heap to free buffers asynchrounously.  Freed buffers
are placed on a free list and freed from a low priority background thread.
If allocations from a particular heap fail, the free list is drained.  This
patch also enable asynchronous frees from the chunk heap.

Change-Id: Idfdbc8608b6cbd9e27d2e31ea4fd84fea9f69f7d
Signed-off-by: Rebecca Schultz Zavin <rebecca@android.com>
11 years agotrace: Add an option to show tgids in trace output
Jamie Gennis [Wed, 21 Nov 2012 23:04:25 +0000 (15:04 -0800)]
trace: Add an option to show tgids in trace output

The tgids are tracked along side the saved_cmdlines tracking, and can be
included in trace output by enabling the 'print-tgid' trace option. This is
useful when doing post-processing of the trace data, as it allows events to be
grouped by tgid.

Change-Id: I52ed04c3a8ca7fddbb868b792ce5d21ceb76250e
Signed-off-by: Jamie Gennis <jgennis@google.com>
11 years agoAdd security hooks to binder and implement the hooks for SELinux.
Stephen Smalley [Mon, 5 Nov 2012 13:15:34 +0000 (08:15 -0500)]
Add security hooks to binder and implement the hooks for SELinux.

Add security hooks to the binder and implement the hooks for SELinux.
The security hooks enable security modules such as SELinux to implement
controls over binder IPC.  The security hooks include support for
controlling what process can become the binder context manager
(binder_set_context_mgr), controlling the ability of a process
to invoke a binder transaction/IPC to another process (binder_transaction),
controlling the ability a process to transfer a binder reference to
another process (binder_transfer_binder), and controlling the ability
of a process to transfer an open file to another process (binder_transfer_file).

This support is used by SE Android, http://selinuxproject.org/page/SEAndroid.

Change-Id: I9a64a87825df2e60b9c51400377af4a9cd1c4049
Signed-off-by: Stephen Smalley <sds@tycho.nsa.gov>
11 years agoinput: misc: keychord: log when keychord triggered
JP Abgrall [Tue, 5 Mar 2013 22:25:36 +0000 (14:25 -0800)]
input: misc: keychord: log when keychord triggered

log keychord id at info level just before waking up processes.

Signed-off-by: JP Abgrall <jpa@google.com>
11 years agotrace/events: add gpu trace events
Jamie Gennis [Fri, 22 Feb 2013 01:55:28 +0000 (17:55 -0800)]
trace/events: add gpu trace events

Change-Id: I0607b9c776acf61cb796b8572cf8cfb8b2dc1377
Signed-off-by: Jamie Gennis <jgennis@google.com>
11 years agogpu: ion: Add support for sharing buffers with dma buf kernel handles
Johan Mossberg [Fri, 11 Jan 2013 12:38:13 +0000 (13:38 +0100)]
gpu: ion: Add support for sharing buffers with dma buf kernel handles

Currently ion can only share buffers with dma buf fd's. Fd's can not be
used inside the kernel as they are process specific so support for
sharing buffers with dma buf kernel handles is needed to support kernel
only use cases. An example use case could be a GPU driver using ion
that wants to share its output buffers with a 3d party display
controller driver supporting dma buf.

Change-Id: If1b3753ddbd5b44c5a3e622055d5473e16fc1c48
Signed-off-by: Johan Mossberg <johan.mossberg@stericsson.com>
11 years agonet: bluetooth: Remove the AID_NET_BT* gid numbers
JP Abgrall [Thu, 21 Feb 2013 01:39:53 +0000 (17:39 -0800)]
net: bluetooth: Remove the AID_NET_BT* gid numbers

Removed bluetooth checks for AID_NET_BT and AID_NET_BT_ADMIN
which are not useful anymore.
This is in preparation for getting rid of all the AID_* gids.

Signed-off-by: JP Abgrall <jpa@google.com>
11 years agopower: android-battery: push uevent whenever charge source changes
Todd Poynor [Wed, 13 Feb 2013 04:47:48 +0000 (20:47 -0800)]
power: android-battery: push uevent whenever charge source changes

Ensure userspace reads an up-to-date value for charging status whenever the
charge source is updated.  Avoid races where stale charging state may be
reflected in userspace until the next battery state poll.

Change-Id: Ia744db48584e9e9abf442710c279af9e3a25c079
Signed-off-by: Todd Poynor <toddpoynor@google.com>
11 years agopower: android-battery: remove ac and usb supplies
Todd Poynor [Wed, 9 Jan 2013 02:46:41 +0000 (18:46 -0800)]
power: android-battery: remove ac and usb supplies

Should no longer need to fix up AC/USB online state via these.

Change-Id: I48d2ef0fbefee58cb47eafc11d9a44759920df7a
Signed-off-by: Todd Poynor <toddpoynor@google.com>
11 years agonetfilter: xt_qtaguid: Allow tracking loopback
JP Abgrall [Thu, 7 Feb 2013 01:40:07 +0000 (17:40 -0800)]
netfilter: xt_qtaguid: Allow tracking loopback

In the past it would always ignore interfaces with loopback addresses.
Now we just treat them like any other.
This also helps with writing tests that check for the presence
of the qtaguid module.

Signed-off-by: JP Abgrall <jpa@google.com>
11 years agonetfilter: xt_qtaguid: extend iface stat to report protocols
JP Abgrall [Tue, 29 Jan 2013 00:50:44 +0000 (16:50 -0800)]
netfilter: xt_qtaguid: extend iface stat to report protocols

In the past the iface_stat_fmt would only show global bytes/packets
for the skb-based numbers.
For stall detection in userspace, distinguishing tcp vs other protocols
makes it easier.
Now we report
  ifname total_skb_rx_bytes total_skb_rx_packets total_skb_tx_bytes
  total_skb_tx_packets {rx,tx}_{tcp,udp,ohter}_{bytes,packets}

Bug: 6818637
Signed-off-by: JP Abgrall <jpa@google.com>
11 years agogpu: ion: Only flush buffers in the chunk heap if they were used cached
Rebecca Schultz Zavin [Mon, 14 Jan 2013 23:29:10 +0000 (15:29 -0800)]
gpu: ion: Only flush buffers in the chunk heap if they were used cached

Change-Id: I4ffcf81a6be09e968310bbd882fb017415d61b48
Signed-off-by: Rebecca Schultz Zavin <rebecca@android.com>
11 years agogpu: ion: Refactor the code to zero buffers
Rebecca Schultz Zavin [Wed, 9 Jan 2013 19:26:37 +0000 (11:26 -0800)]
gpu: ion: Refactor the code to zero buffers

Refactor the code in the system heap used to map and zero the buffers
into a seperate utility so it can be called from other heaps.  Use it from
the chunk heap.

Change-Id: I706341ae42b80bc4aae8a8614b4f73435bbf05d9
Signed-off-by: Rebecca Schultz Zavin <rebecca@android.com>
11 years agocpufreq: interactive: fix race on governor start/stop
Lianwei Wang [Mon, 7 Jan 2013 06:15:51 +0000 (14:15 +0800)]
cpufreq: interactive: fix race on governor start/stop

There is race condition when both two cpu do CPUFREQ_GOV_STOP and one cpu
do CPUFREQ_GOV_START soon. The sysfs_remove_group is not done yet on one
cpu, but sysfs_create_group is called on another cpu, which cause governor
start failed and then kernel panic in timer callback because the policy and
cpu mask are all kfree in cpufreq driver.

Replace atomic with mutex to lock the whole START/STOP sequence.

Change-Id: I3762b3d44315ae021b8275aca84f5ea9147cc540
Signed-off-by: Lianwei Wang <a22439@motorola.com>
11 years agonetfilter: xt_qtaguid: remove AID_* dependency for access control
JP Abgrall [Sat, 5 Jan 2013 02:18:36 +0000 (18:18 -0800)]
netfilter: xt_qtaguid: remove AID_* dependency for access control

qtaguid limits what can be done with /ctrl and /stats based on group
membership.
This changes removes AID_NET_BW_STATS and AID_NET_BW_ACCT, and picks
up the groups from the gid of the matching proc entry files.

Signed-off-by: JP Abgrall <jpa@google.com>
Change-Id: I42e477adde78a12ed5eb58fbc0b277cdaadb6f94

11 years agogpu: ion: Modify zeroing code so it only allocates address space once
Rebecca Schultz Zavin [Wed, 19 Dec 2012 06:46:57 +0000 (22:46 -0800)]
gpu: ion: Modify zeroing code so it only allocates address space once

vmap/vunmap spend a significant amount of time allocating the
address space to map into.  Rather than allocating address space
for each page, instead allocate once for the entire allocation
and then just map and unmap each page into that address space.

Change-Id: I4a5c850717c80f75506a36b7ec2bcd55857b8dea
Signed-off-by: Rebecca Schultz Zavin <rschultz@google.com>
11 years agocpufreq: interactive: fix deadlock on spinlock in timer
Todd Poynor [Wed, 2 Jan 2013 21:14:00 +0000 (13:14 -0800)]
cpufreq: interactive: fix deadlock on spinlock in timer

Need to use irqsave/restore spinlock calls to avoid a deadlock in calls
from the timer.

Change-Id: I15b6b590045ba1447e34ca7b5ff342723e53a605
Signed-off-by: Todd Poynor <toddpoynor@google.com>
11 years agocpufreq: interactive: don't handle transition notification if not enabled
Todd Poynor [Sun, 23 Dec 2012 20:28:49 +0000 (12:28 -0800)]
cpufreq: interactive: don't handle transition notification if not enabled

If multiple governors are in use then avoid processing frequency transition
notifications for CPUs on which the interactive governor is not enabled.

Change-Id: Ibd75255b921d887501a64774a8c4f62302f2d4e4
Reported-by: Francisco Franco <francisco.franco@cloudcar.com>
Signed-off-by: Todd Poynor <toddpoynor@google.com>
11 years agocpufreq: interactive: init default values at compile time
Todd Poynor [Fri, 21 Dec 2012 23:32:21 +0000 (15:32 -0800)]
cpufreq: interactive: init default values at compile time

Change-Id: Ia4966e949a6c24c34fdbd4a6e522cd7c37e4108e
Signed-off-by: Todd Poynor <toddpoynor@google.com>
11 years agocpufreq: interactive: default go_hispeed_load 99%, doc updates
Todd Poynor [Fri, 21 Dec 2012 23:13:01 +0000 (15:13 -0800)]
cpufreq: interactive: default go_hispeed_load 99%, doc updates

Update default go_hispeed_load from 85% to 99%.  Recent changes to the
governor now use a default target_load of 90%.  go_hispeed_load should
not be lower than the target load for hispeed_freq, which could lead
to oscillating speed decisions.  Other recent changes reduce the need
to dampen speed jumps on load spikes, while input event boosts from
userspace are the preferred method for anticipating load spikes with
UI impacts.

General update to the documentation to reflect recent changes.

Change-Id: I1b92f3091f42c04b10503cd1169a943b5dfd6faf
Signed-off-by: Todd Poynor <toddpoynor@google.com>
11 years agocpufreq: interactive: fix race on timer restart on governor start
Todd Poynor [Thu, 20 Dec 2012 23:51:00 +0000 (15:51 -0800)]
cpufreq: interactive: fix race on timer restart on governor start

Starting the governor, or restarting on a hotplugged-in CPU, can race
with the timer start in idle, triggering a BUG on timer already pending.
Start the timer before setting the enable flag, and use enable_sem to
protect the sequence (and ensure correct order of the update to the
enable flag).  Delete any existing timer for safety.

Change-Id: Ife77cf9fe099e8fd8543224cbf148c6722c2ffb0
Reported-by: Francisco Franco <francisco.franco@cloudcar.com>
Signed-off-by: Todd Poynor <toddpoynor@google.com>
11 years agocpufreq: interactive: fix racy timer stopping
Todd Poynor [Wed, 19 Dec 2012 01:50:44 +0000 (17:50 -0800)]
cpufreq: interactive: fix racy timer stopping

When stopping the governor, del_timer_sync() can race against an
invocation of the idle notifier callback, which has the potential
to reactivate the timer.

To fix this issue, a read-write semaphore is used. Multiple readers are
allowed as long as pcpu->governor_enabled is true.  However it can be
moved to false only after taking a write semaphore which would wait for
any on-going asynchronous activities to complete and prevent any more of
those activities to be initiated.

[toddpoynor@google.com: cosmetic and commit text changes]
Change-Id: Ib51165a735d73dcf964a06754c48bdc1913e13d0
Signed-off-by: Nicolas Pitre <nicolas.pitre@linaro.org>
11 years agocpufreq: interactive: fix boosting logic
Todd Poynor [Thu, 20 Dec 2012 00:06:48 +0000 (16:06 -0800)]
cpufreq: interactive: fix boosting logic

35a84de cpufreq: interactive: apply above_hispeed_delay to each step above hispeed

caused the speed choice logic to osciallate between boosting and not boosting.
Add back code to ensure speed does not drop below boost frequency while
boosting.

Change-Id: Id420068480fcc7f5c4989ff523e2a8d22e2f4db2
Signed-off-by: Todd Poynor <toddpoynor@google.com>
11 years agocpufreq: interactive: add timer slack to limit idle at speed > min
Todd Poynor [Wed, 19 Dec 2012 01:50:10 +0000 (17:50 -0800)]
cpufreq: interactive: add timer slack to limit idle at speed > min

Always use deferrable timer for load sampling.

Set a non-deferrable timer to an additional slack time to allow prior to
waking up from idle to drop speed when not at minimum speed.  Slack value
-1 avoids wakeups to drop speed.  Default is 80ms.

Remove the governidle module param and its timer management in idle.  For
platforms on which holding speed above mimum in idle costs power, use the
new timer slack to select how long to wait before waking up to drop speed.

Change-Id: I270b3980667e2c70a68e5bff534124b4411dbad5
Signed-off-by: Todd Poynor <toddpoynor@google.com>
11 years agohardlockup: detect hard lockups without NMIs using secondary cpus
Colin Cross [Fri, 11 Jan 2013 21:51:48 +0000 (13:51 -0800)]
hardlockup: detect hard lockups without NMIs using secondary cpus

Emulate NMIs on systems where they are not available by using timer
interrupts on other cpus.  Each cpu will use its softlockup hrtimer
to check that the next cpu is processing hrtimer interrupts by
verifying that a counter is increasing.

This patch is useful on systems where the hardlockup detector is not
available due to a lack of NMIs, for example most ARM SoCs.
Without this patch any cpu stuck with interrupts disabled can
cause a hardware watchdog reset with no debugging information,
but with this patch the kernel can detect the lockup and panic,
which can result in useful debugging info.

Signed-off-by: Colin Cross <ccross@android.com>
11 years agocpufreq: interactive: specify duration of CPU speed boost pulse
Todd Poynor [Sat, 15 Dec 2012 01:31:19 +0000 (17:31 -0800)]
cpufreq: interactive: specify duration of CPU speed boost pulse

Sysfs attribute boostpulse_duration specifies the duration of boosting CPU
speed in response to bootpulse events.  Duration is specified in usecs,
default 80ms.

Change-Id: Ifd41625574891a44f1787a4e85d1e7b4f2afb52b
Signed-off-by: Todd Poynor <toddpoynor@google.com>
11 years agocpufreq: interactive: adjust load for changes in speed
Todd Poynor [Wed, 12 Dec 2012 00:05:03 +0000 (16:05 -0800)]
cpufreq: interactive: adjust load for changes in speed

Add notifier for speed transitions.  Keep a count of CPU active
microseconds times current frequency, converted to a percentage relative
to the current frequency when load is evaluated.

Change-Id: I5c27adb11081c50490219784ca57cc46e97fc28c
Signed-off-by: Todd Poynor <toddpoynor@google.com>
11 years agogpu: ion: Remove heapmask from client
Rebecca Schultz Zavin [Tue, 11 Dec 2012 23:23:14 +0000 (15:23 -0800)]
gpu: ion: Remove heapmask from client

The heapmask in the client generally wasn't being used.  This
patch removes it.

Change-Id: I3526723fbf8f2e81c28c0733deb583ea14bdd837
Signed-off-by: Rebecca Schultz Zavin <rebecca@android.com>
11 years agogpu: ion: Export ion_client_create
Johan Mossberg [Mon, 10 Dec 2012 16:46:16 +0000 (17:46 +0100)]
gpu: ion: Export ion_client_create

Will enable modules to allocate memory with ion.

Signed-off-by: Johan Mossberg <johan.mossberg@stericsson.com>
11 years agogpu: ion: Clarify variable names and comments around heap ids v types
Rebecca Schultz Zavin [Tue, 11 Dec 2012 19:45:59 +0000 (11:45 -0800)]
gpu: ion: Clarify variable names and comments around heap ids v types

There is some confusion between when to use the heap type and when
the id.  This patch clarifies this by using clearer variable names
and describing the intention in the comments.  Also fixes the client
debug code to print heaps by id instead of type.

Change-Id: Ie8b3dadded52e18590fcb2ca94001f6ed46ef07d
Signed-off-by: Rebecca Schultz Zavin <rebecca@android.com>
11 years agogpu: ion: Add chunk heap
Rebecca Schultz Zavin [Thu, 15 Nov 2012 18:52:45 +0000 (10:52 -0800)]
gpu: ion: Add chunk heap

This patch adds support for a chunk heap that allows for buffers that are
made up of a list of fixed size chunks taken from a carveout.  Chunk sizes
are configured when the heaps are created by passing the chunk size in the
priv field of the heap platform data.

Change-Id: Ia9e003f727b553a92804264debe119dcf78b14e0
Signed-off-by: Rebecca Schultz Zavin <rebecca@android.com>
11 years agogpu: ion: Refactor common mapping functions out of system heap
Rebecca Schultz Zavin [Thu, 15 Nov 2012 18:43:46 +0000 (10:43 -0800)]
gpu: ion: Refactor common mapping functions out of system heap

The system heap contained several general purpose functions to map
buffers to the kernel and userspace.  This patch refactors those
into ion_heap.c so they can be used by other heaps.

Change-Id: If64591798bdc2c248bf9064ace2c927909d7adb8
Signed-off-by: Rebecca Schultz Zavin <rebecca@android.com>
11 years agogpu: ion: Switch heap rbtree to a prio list
Rebecca Schultz Zavin [Thu, 15 Nov 2012 18:36:10 +0000 (10:36 -0800)]
gpu: ion: Switch heap rbtree to a prio list

Switches the rbtree tree of heaps for a plist.  This significantly
simplifies the code and the list is small and is modified only at
first boot so the rbtree is unnecessary.  This also switches
the traversal of the heap list to traverse from highest to lowest
id's.  This allows allocations to pass a heap mask that falls
back on the system heap -- typically id 0, which is the common case.

Change-Id: I715be6f4cf020a84ca4f1947c30ee3d2559fb523
Signed-off-by: Rebecca Schultz Zavin <rebecca@android.com>
11 years agogpu: ion: Fix bug where MAP ioctl was no longer supported
Rebecca Schultz Zavin [Thu, 15 Nov 2012 18:34:45 +0000 (10:34 -0800)]
gpu: ion: Fix bug where MAP ioctl was no longer supported

Change-Id: Idbe628ed7dbd6a14469194120b94934d6e99d367
Signed-off-by: Rebecca Schultz Zavin <rebecca@android.com>
11 years agogpu: ion: Modify reserve function for carveouts with no start address
Rebecca Schultz Zavin [Thu, 15 Nov 2012 18:31:02 +0000 (10:31 -0800)]
gpu: ion: Modify reserve function for carveouts with no start address

This patch allows you to specify a heap that requires carveout memory
but that doesn't specify a start address.  Memblock_alloc will be called
to find a location for these heaps.

Change-Id: I9c79b30e3105e796060fc74b058f04093ee5e96e
Signed-off-by: Rebecca Schultz Zavin <rebecca@android.com>
11 years agocpufreq: interactive: remove load since last speed change
Todd Poynor [Sat, 8 Dec 2012 04:08:45 +0000 (20:08 -0800)]
cpufreq: interactive: remove load since last speed change

The longer-term load since last speed change isn't terribly useful,
may delay recognition of dropping load, and would need forthcoming
changes to adjust load for changing CPU speeds.  Drop it.

Change-Id: Ic3cbb0542cc3484617031787e03ed9bdd632dec1
Signed-off-by: Todd Poynor <toddpoynor@google.com>
11 years agocpufreq: interactive: allow arbitrary speed / target load mappings
Todd Poynor [Wed, 14 Nov 2012 19:41:21 +0000 (11:41 -0800)]
cpufreq: interactive: allow arbitrary speed / target load mappings

Accept a string of target loads and speeds at which to apply the
target loads, per the documentation update in this patch.  For example,
"85 1000000:90 1700000:99" targets CPU load 85% below speed 1GHz,  90%
at or above 1GHz, until 1.7GHz and above, at which load 99% is targeted.

Attempt to avoid oscillations by evaluating the current speed
weighted by current load against each new choice of speed, choosing a
higher speed if the current load requires a higher speed.

Change-Id: Ie3300206047c84eca5a26b0b63ea512e5207550e
Signed-off-by: Todd Poynor <toddpoynor@google.com>
11 years agogpu: ion: fix compilation warning
Benjamin Gaignard [Tue, 9 Oct 2012 11:39:25 +0000 (13:39 +0200)]
gpu: ion: fix compilation warning

use atomic_read to get the refcount value to avoid compilation warning

Change-Id: I9bf0a07c787af8f87ac19314f996a78142f72f6b
Signed-off-by: Benjamin Gaignard <benjamin.gaignard@linaro.org>
11 years agogpu: ion: fix carveout ops
Benjamin Gaignard [Wed, 17 Oct 2012 09:32:02 +0000 (11:32 +0200)]
gpu: ion: fix carveout ops

when using carveout heap ion_buffer_create function failed because
map_dma and unmap_dma operations aren't set by carveout heap.

Change-Id: I817bfad742abfab10b01d9b3d2e18bdf812a1307
Signed-off-by: Benjamin Gaignard <benjamin.gaignard@linaro.org>
11 years agogpu: ion: Fix bug in zeroing pages in system heap
Rebecca Schultz Zavin [Mon, 3 Dec 2012 19:43:49 +0000 (11:43 -0800)]
gpu: ion: Fix bug in zeroing pages in system heap

Pages are zeroed for security purposes when returned to the
ion heap.  There was a bug in this code preventing this
from happening.

Bug: 7573871
Change-Id: I79c22ee1da98f306199f3a192eaec4e81d5fd059
Signed-off-by: Rebecca Schultz Zavin <rebecca@android.com>
11 years agocpufreq: interactive: apply above_hispeed_delay to each step above hispeed
Todd Poynor [Thu, 8 Nov 2012 23:06:55 +0000 (15:06 -0800)]
cpufreq: interactive: apply above_hispeed_delay to each step above hispeed

Apply above_hispeed_delay whenever increasing speed to a new speed above
hispeed (not just the first step above hispeed).

Change-Id: Ibb7add7db47f2a4306a9458c4e1ebabb60698636
Signed-off-by: Todd Poynor <toddpoynor@google.com>
11 years agocpufreq: interactive: change speed according to current speed and target load
Todd Poynor [Thu, 29 Nov 2012 01:58:17 +0000 (17:58 -0800)]
cpufreq: interactive: change speed according to current speed and target load

Add a target_load attribute that specifies how aggressively the governor is
to adjust speed to meet the observed load.  New target speed is calculated
as the current actual speed (may be higher than target speed on SMP) times
the CPU load (as a fraction) divided by target load (fraction).

cpufreq_frequency_table_target() call use CPUFREQ_RELATION_L to set
the next higher speed rather than next lower speed.

Change-Id: If432451da82f5fed12e15c9421d7d27792376150
Signed-off-by: Todd Poynor <toddpoynor@google.com>
11 years agocpufreq: interactive: trace actual speed in target speed decisions
Todd Poynor [Thu, 29 Nov 2012 01:56:09 +0000 (17:56 -0800)]
cpufreq: interactive: trace actual speed in target speed decisions

Tracing adds actual speed since this is expected to be key to the
choice of target speed.

Change-Id: Iec936102d0010c4e9dfa143c38a9fd0d551189c3
Signed-off-by: Todd Poynor <toddpoynor@google.com>
11 years agousb: gadget: accessory: Fix section mismatch (again)
Todd Poynor [Wed, 28 Nov 2012 21:03:40 +0000 (13:03 -0800)]
usb: gadget: accessory: Fix section mismatch (again)

create_bulk_endpoints should not be __init since it is called when
accessory is enabled.

Change-Id: Iac6e9f29d53c93760e926efd8e7603432632acb4
Signed-off-by: Todd Poynor <toddpoynor@google.com>
11 years agonetfilter: qtaguid: Don't BUG_ON if create_if_tag_stat fails
Pontus Fuchs [Mon, 19 Nov 2012 19:44:51 +0000 (11:44 -0800)]
netfilter: qtaguid: Don't BUG_ON if create_if_tag_stat fails

If create_if_tag_stat fails to allocate memory (GFP_ATOMIC) the
following will happen:

qtaguid: iface_stat: tag stat alloc failed
...
kernel BUG at xt_qtaguid.c:1482!

Signed-off-by: Pontus Fuchs <pontus.fuchs@gmail.com>
11 years agocpufreq: interactive: kick timer on idle exit past expiry
Todd Poynor [Tue, 9 Oct 2012 03:14:34 +0000 (20:14 -0700)]
cpufreq: interactive: kick timer on idle exit past expiry

The deferrable timer list isn't checked on all idle exits, such as when
hi-res timers expire or ISRs schedule workers.  If the idle loop is
exited and it's past time to run the governor load polling timer,
run it immediately.  This ensures we handle load spikes caused by actvity
that does not run the normal timer list.

Rename the field that timestamps the "time_in_idle" value to be more
accurate.

Change-Id: Ied590ecbefc83c9a9ec5eb9e31903557f6fa1614
Signed-off-by: Todd Poynor <toddpoynor@google.com>
11 years agocpufreq: interactive: use deferrable timer by default
Lianwei Wang [Thu, 1 Nov 2012 01:59:52 +0000 (09:59 +0800)]
cpufreq: interactive: use deferrable timer by default

Avoid wakeups only to handle the governor timer when the system is otherwise
idle.

For platforms where the power cost of remaining in idle at higher CPU
speed may outweigh the cost of a governor wakeup from idle to lower the speed,
set parameter cpufreq_interactive.governidle=1.

Change-Id: Id6c43eb35caecf9b0574fcdd5b769711bc7e6de6
Signed-off-by: LianWei WANG <a22439@motorola.com>
Signed-off-by: Todd Poynor <toddpoynor@google.com>
11 years agocpufreq: interactive: pin timers to associated CPU
Todd Poynor [Mon, 5 Nov 2012 21:09:03 +0000 (13:09 -0800)]
cpufreq: interactive: pin timers to associated CPU

Helps avoid waking up other CPUs to react to activity on the local CPU.

Change-Id: Ife272aaa7916894a437705d44521b1a1693fbe8e
Signed-off-by: Todd Poynor <toddpoynor@google.com>
11 years agoARM: fiq_debugger: fix uninitialised spin_lock.
Mars [Sat, 3 Nov 2012 04:15:38 +0000 (12:15 +0800)]
ARM: fiq_debugger: fix uninitialised spin_lock.

Backtrace:
[<c0045ea0>] (dump_backtrace+0x0/0x10c) from [<c0575c60>] (dump_stack+0x1)
 r6:c07a489c r5:c0c9b9dc r4:00000002 r3:271aed3b
[<c0575c48>] (dump_stack+0x0/0x1c) from [<c00b1b00>] (__lock_acquire+0x93)
[<c00b11c8>] (__lock_acquire+0x0/0xad4) from [<c00b219c>] (lock_acquire+0)
[<c00b210c>] (lock_acquire+0x0/0xa4) from [<c057e544>] (_raw_spin_lock_ir)
[<c057e4f8>] (_raw_spin_lock_irq+0x0/0x5c) from [<c005297c>] (fiq_tty_wri)
 r5:e30f0000 r4:e36f0c00
[<c005293c>] (fiq_tty_write+0x0/0x80) from [<c023168c>] (n_tty_write+0x18)
 r8:e370fc40 r7:e378a000 r6:e3572d1c r5:e36f0c00 r4:00000002
r3:c005293c
[<c0231500>] (n_tty_write+0x0/0x440) from [<c022d4a4>] (tty_write+0x100/0)
[<c022d3a4>] (tty_write+0x0/0x2a8) from [<c0100b80>] (vfs_write+0xa4/0x14)
[<c0100adc>] (vfs_write+0x0/0x148) from [<c0100cdc>] (sys_write+0x40/0x78)
 r8:00000002 r7:4076d2c4 r6:e370fc40 r5:00000000 r4:00000000
[<c0100c9c>] (sys_write+0x0/0x78) from [<c0041740>] (ret_fast_syscall+0x0)
 r8:c0041908 r7:00000004 r6:00000002 r5:00000000 r4:4007cbe0

[ccross: moved spin_lock_init into existing #ifdef]
Change-Id: If400d084eb20433c126ea1dd027a6be7f2ebb1f6
Signed-off-by: Mars <caoziqiang@meizu.com>
Signed-off-by: Colin Cross <ccross@android.com>
11 years agoARM: fiq_debugger: lock between tty and console writes
Colin Cross [Thu, 1 Nov 2012 00:41:39 +0000 (17:41 -0700)]
ARM: fiq_debugger: lock between tty and console writes

debug_console_write calls debug_uart_flush, which will usually wait
until the serial port fifo empties.  If another thread is continuously
calling fiq_tty_write, the fifo will constantly be refilled and
debug_uart_flush might never return.

Add a spinlock that is locked in debug_console_write and fiq_tty_write
to ensure they can't run at the same time.  This has an extra advantage
of preventing lines from the console and tty from being mixed together.

Also reduce the size returned by fiq_tty_write_room to keep the time
spent with the spinlock held to a reasonable value.

In addition, make sure fiq context can't loop forever by never calling
debug_uart_flush when the console is enabled.

Change-Id: I5712b01f740ca0c84f680d2032c9fa16b7656939
Signed-off-by: Colin Cross <ccross@android.com>
11 years agosync: add Documentation/sync.txt
Erik Gilling [Fri, 14 Sep 2012 21:36:34 +0000 (14:36 -0700)]
sync: add Documentation/sync.txt

Change-Id: Ic7f3a6d4622083be607b82ddd8d676609225bb8f
Signed-off-by: Erik Gilling <konkers@android.com>
11 years agogpu: ion: Fix bug in ion_system_heap map_user
Rebecca Schultz Zavin [Tue, 23 Oct 2012 17:47:04 +0000 (10:47 -0700)]
gpu: ion: Fix bug in ion_system_heap map_user

When the requested mmap length was not an integer number of
chunks or the buffer, or if an offset was provided, a bug
would cause extra or incorrect pages of the buffer to be mapped.

Change-Id: I2766763d86048f026eeef0e0388b7de0e25c2093
Signed-off-by: Rebecca Schultz Zavin <rebecca@android.com>