Jon Medhurst [Tue, 8 Apr 2014 15:43:23 +0000 (16:43 +0100)]
Revert "hmp: Use idle pull to perform forced up-migrations"
This reverts commit
aae7721f20f2520d24a149408a74f18e58f56472.
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Jon Medhurst [Tue, 8 Apr 2014 15:43:19 +0000 (16:43 +0100)]
Revert "hmp: dont attempt to pull tasks if affinity doesn't allow it"
This reverts commit
5a570cfc01b06906faa8ac67ad7c0c6f278761c4.
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Chris Redpath [Mon, 24 Mar 2014 13:47:30 +0000 (13:47 +0000)]
hmp: dont attempt to pull tasks if affinity doesn't allow it
When looking for a task to be idle-pulled, don't consider tasks
where the affinity does not allow that task to be placed on the
target CPU. Also ensure that tasks with restricted affinity
do not block selecting other unrestricted busy tasks.
Use the knowledge of target CPU more effectively in idle pull
by passing to hmp_get_heaviest_task when we know it, otherwise
only checking for general affinity matches with any of the CPUs
in the bigger HMP domain.
We still need to explicitly check affinity is allowed in idle pull
since if we find no match in hmp_get_heaviest_task we will return
the current one, which may not be affine to the new CPU despite
having high enough load. In this case, there is nothing to move.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Chris Redpath [Mon, 24 Mar 2014 13:47:29 +0000 (13:47 +0000)]
hmp: Use idle pull to perform forced up-migrations
When a normal forced up-migration takes place we stop the task to
be migrated while the target CPU becomes available. This delay can
range from 80us to 1500us on TC2 if the target CPU is in a deep idle
state.
Instead, interrupt the target CPU and ask it to pull a task.
This lets the current eligible task continue executing on the
original CPU while the target CPU wakes. Use a pinned timer to
prevent the pulling CPU going back into power-down with pending
up-migrations.
If we trigger for a nohz kick, it doesn't matter about triggering
for an idle pull since the idle_pull flag will be set when we
execute the softirq and we'll still do the idle pull.
If the target CPU is busy, we will not pull any tasks.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Chris Redpath [Mon, 24 Mar 2014 13:47:28 +0000 (13:47 +0000)]
sched: hmp: unify active migration code
The HMP active migration code is functionally identical to the CFS
active migration code apart from one flag check. Share the code
and make the flag check optional.
Two wrapper functions allow the flag check to be present or not.
Thanks to tixy@linaro.org for pointing out the build break and a
good solution in an earlier version.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Chris Redpath [Mon, 24 Mar 2014 13:47:27 +0000 (13:47 +0000)]
hmp: sched: Clean up hmp_up_threshold checks into a utility fn
In anticipation of modifying the up_threshold handling, make all
instances use the same utility fn to check if a task is eligible
for up-migration. This also removes the previous difference in
threshold comparison where up-migration used '!<threshold' and
idle pull used '>threshold' to decide up-migration eligibility.
Make them both use '!<threshold' instead for consistency, although
this is unlikely to change any results.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Dietmar Eggemann [Thu, 16 Jan 2014 11:53:14 +0000 (11:53 +0000)]
HMP: Fix rt task allowed cpu mask restriction code on 1x1 system
There is an error scenario where on a 1x1 HMP system (weight of the
hmp_slow_cpu_mask is 1) the short-cut of restricting the allowed cpu mask
of an rt tasks leads to triggering a kernel bug in the rt sched class
set_cpus_allowed function set_cpus_allowed_rt().
In case the task is on the run-queue and the weight of the required cpu mask
is 1 and this is different to the p->nr_cpus_allowed value, this back-end
function interprets this in such a way that a task changed from being
migratable to not migratable anymore and decrements the rt_nr_migratory
counter. There is a BUG_ON(!rq->rt.rt_nr_migratory) check in this code
path which triggers in this situation.
To circumvent this issue, set the number of allowed cpus for a task p to
the weight of the hmp_slow_cpu_mask before calling do_set_cpus_allowed()
in __setscheduler(). It will be set to this value in do_set_cpus_allowed()
after the call to the sched class related backend function any way. By
doing this, set_cpus_allowed_rt() returns without trying to update the
rt_nr_migratory counter.
This patch has been tested with a test device driver requiring a threaded
irq handler on a TC2 system with a reduced cpu mask (1 Cortex A15, 1
Cortex A7).
Signed-off-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Dietmar Eggemann [Thu, 16 Jan 2014 19:44:10 +0000 (19:44 +0000)]
HMP: Restrict irq_default_affinity to hmp_slow_cpu_mask
This patch limits the default affinity mask for all irqs to the cluster of
the little cpus.
This patch has the positive side effect that an irq thread which has its
IRQTF_RUNTHREAD set inside irq_thread() -> irq_wait_for_interrupt() will
not overwrite its struct task_struct->cpus_allowed with a full cpu mask of
desc->irq_data.affinity in irq_thread_check_affinity() essentially reverting
patch "HMP: experimental: Force all rt tasks to start on little domain."
for this irq thread.
Signed-off-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Chris Redpath [Fri, 10 Jan 2014 10:34:08 +0000 (10:34 +0000)]
sched: hmp: Fix potential task_struct memory leak
We use get_task_struct to increment the ref count on a task_struct
so that even if the task dies with a pending migration we are still
able to read the memory without causing a fault.
In the case of non-running tasks, we forgot to decrement the ref
count when we are done with the task.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Chris Redpath [Thu, 9 Jan 2014 10:41:13 +0000 (10:41 +0000)]
config: Make packing present on TC2
The scheduler will default packing to disabled, but this includes the
feature so that we can test it more easily.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Chris Redpath [Thu, 9 Jan 2014 10:40:30 +0000 (10:40 +0000)]
sched: hmp: Change TC2 packing config to disabled default if present
Since TC2 power curves don't really have a utilisation hotspot where
packing makes sense, if it is present for a TC2 system at least make
it default to disabled.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Chris Redpath [Thu, 9 Jan 2014 10:38:54 +0000 (10:38 +0000)]
sched: hmp: Make idle balance behaviour normal when packing disabled
The presence of packing permanently changed the idle balance
behaviour. Do not restrict idle balance on the smallest CPUs when
packing is present but disabled.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Chris Redpath [Thu, 5 Dec 2013 17:57:46 +0000 (17:57 +0000)]
sched: update runqueue clock before migrations away
If we migrate a sleeping task away from a CPU which has the
tick stopped, then both the clock_task and decay_counter will
be out of date for that CPU and we will not decay load correctly
regardless of how often we update the blocked load.
This is only an issue for tasks which are not on a runqueue
(because otherwise that CPU would be awake) and simultaneously
the CPU the task previously ran on has had the tick stopped.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Chris Redpath [Thu, 5 Dec 2013 15:49:32 +0000 (15:49 +0000)]
sched: reset blocked load decay_count during synchronization
If an entity happens to sleep for less than one tick duration
the tracked load associated with that entity can be decayed by an
unexpectedly large amount if it is later migrated to a different
CPU. This can interfere with correct scheduling when entity load
is used for decision making.
The reason for this is that when an entity is dequeued and enqueued
quickly, such that se.avg.decay_count and cfs_rq.decay_counter
do not differ when that entity is enqueued again,
__synchronize_entity_decay skips the calculation step and also skips
clearing the decay_count. At a later time that entity may be
migrated and its load will be decayed incorrectly.
All users of this function expect decay_count to be zero'ed after
use.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Thomas Gleixner [Fri, 25 May 2012 14:59:47 +0000 (16:59 +0200)]
genirq: Add default affinity mask command line option
If we isolate CPUs, then we don't want random device interrupts on
them. Even w/o the user space irq balancer enabled we can end up with
irqs on non boot cpus.
Allow to restrict the default irq affinity mask.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Jon Medhurst [Tue, 21 Jan 2014 09:48:55 +0000 (09:48 +0000)]
config: Add config fragments for big LITTLE MP
This patch adds config fragments used to enable most of the features
used by big LITTLE MP.
This patch is the result of merging the following commits from the
Linaro Stable Kernel (LSK) 3.10 tree...
Commit
313c69010ed52a7bd52095708049e16c380c8e15
Commit
83952e566d1c41b2e0a70321e9d56cb8162f9a6d
Commit
c4a782d985da03016e2b15aaef90bfc5cfe426f6
Commit
2732b0467a356e7cc08bf1e0f33481330335657a
Commit
7c88e2bd50fc093c96c912090b712b682ae72641
Commit
3231c8ce4b005f79ad27309907adc3de319e341c
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Kamalesh Babulal [Tue, 25 Jun 2013 08:03:36 +0000 (13:33 +0530)]
sched/debug: Add load-tracking statistics to task
At present we print per-entity load-tracking statistics for
cfs_rq of cgroups/runqueues. Given that per task statistics
is maintained, it can be used to know the contribution made
by the task to its parenting cfs_rq level.
This patch adds per-task load-tracking statistics to /proc/<PID>/sched.
Signed-off-by: Kamalesh Babulal <kamalesh@linux.vnet.ibm.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/20130625080336.GA20175@linux.vnet.ibm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
(cherry picked from commit
939fd731eb88a0cdd9058d0b0143563172a217d7)
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Chris Redpath [Fri, 22 Nov 2013 13:19:18 +0000 (13:19 +0000)]
sched: hmp: Fix build breakage when not using CONFIG_SCHED_HMP
hmp_variable_scale_convert was used without guards in
__update_entity_runnable_avg. Guard it.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Chris Redpath [Wed, 20 Nov 2013 14:14:44 +0000 (14:14 +0000)]
Documentation: HMP: Small Task Packing explanation
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Chris Redpath [Mon, 11 Nov 2013 16:29:29 +0000 (16:29 +0000)]
sched: hmp: add read-only hmp domain sysfs file
In order to allow userspace to restrict known low-load tasks to
little CPUs, we must export this knowledge from the kernel or
expect userspace to make their own attempts at figuring it out.
Since we now have a userspace requirement for an HMP implementation
to always have at least some sysfs files, change the integration
so that it only depends upon CONFIG_SCHED_HMP rather than
CONFIG_HMP_VARIABLE_SCALE. Fix Kconfig text to match.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Mathieu Poirier [Wed, 20 Nov 2013 14:20:42 +0000 (14:20 +0000)]
HMP: Avoid using the cpu stopper to stop runnable tasks
When migrating a runnable task, we use the CPU stopper on
the source CPU to ensure that the task to be moved is not
currently running. Before this patch, all forced migrations
(up, offload, idle pull) use the stopper for every migration.
Using the CPU stopper is mandatory only when a task is currently
running on a CPU. Otherwise tasks can be moved by locking the
source and destination run queues.
This patch checks to see if the task to be moved are currently
running. If not the task is moved directly without using the
stopper thread.
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Mark Brown [Mon, 14 Oct 2013 12:37:36 +0000 (13:37 +0100)]
arm64: Fix build due to HMP tracepoints
Commit
2353c1f800 (arm: ipi raise/start/end tracing) added tracepoints for
IPIs in the generic GIC driver but only added definitions for them on ARM,
causing build failures on ARM64. Fix this by adding equivalent definitions
for arm64.
Signed-off-by: Mark Brown <broonie@linaro.org>
Acked-by: Liviu Dudau <Liviu.Dudau@arm.com>
Mark Brown [Mon, 14 Oct 2013 11:38:19 +0000 (12:38 +0100)]
smp: Don't use typedef to work around compiler issue with tracepoints
Having the typedef in place for the tracepoints causes compiler crashes
in some situations. Just using void * directly avoids triggering the
issue and should have no effect on the trace.
Signed-off-by: Mark Brown <broonie@linaro.org>
Acked-by: Liviu Dudau <Liviu.Dudau@arm.com>
Chris Redpath [Fri, 11 Oct 2013 10:45:04 +0000 (11:45 +0100)]
HMP: Implement task packing for small tasks in HMP systems
If we wake up a task on a little CPU, fill CPUs rather than
spread. Adds 2 new files to sys/kernel/hmp to control packing
behaviour.
packing_enable: task packing enabled (1) or disabled (0)
packing_limit: Runqueues will be filled up to this load ratio.
This functionality is disabled by default on TC2 as it lacks per-cpu
power gating so packing small tasks there doesn't make sense.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Signed-off-by: Liviu Dudau <Liviu.Dudau@arm.com>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Chris Redpath [Fri, 11 Oct 2013 10:45:03 +0000 (11:45 +0100)]
hmp: Remove potential for task_struct access race
Accessing the task_struct can be racy in certain conditions, so
we need to only acquire the data when needed.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Signed-off-by: Liviu Dudau <Liviu.Dudau@arm.com>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Chris Redpath [Fri, 11 Oct 2013 10:45:02 +0000 (11:45 +0100)]
sched: HMP: fix potential logical errors
The previous API for hmp_up_migration reset the destination
CPU every time, regardless of if a migration was desired. The code
using it assumed that the value would not be changed unless
a migration was required. In one rare circumstance, this could
have lead to a task migrating to a little CPU at the wrong time.
Fixing that lead to a slight logical tweak to make the surrounding
APIs operate a bit more obviously.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Signed-off-by: Robin Randhawa <robin.randhawa@arm.com>
Signed-off-by: Liviu Dudau <Liviu.Dudau@arm.com>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Chris Redpath [Fri, 11 Oct 2013 10:45:01 +0000 (11:45 +0100)]
smp: smp_cross_call function pointer tracing
generic tracing for smp_cross_call function calls
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Signed-off-by: Liviu Dudau <Liviu.Dudau@arm.com>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Chris Redpath [Fri, 11 Oct 2013 10:45:00 +0000 (11:45 +0100)]
arm: ipi raise/start/end tracing
Add tracepoints for IPI raise events, and start and end of the
ipi handler.
Used to inspect the source of CPU wake-ups which are not already
traced - all other reasons for a CPU to wake-up are already
covered.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Signed-off-by: Liviu Dudau <Liviu.Dudau@arm.com>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Chris Redpath [Fri, 11 Oct 2013 10:44:59 +0000 (11:44 +0100)]
sched: HMP: Additional trace points for debugging HMP behaviour
1. Replace magic numbers in code for migration trace.
Trace points still emit a number as force=<n> field:
force=0 : wakeup migration
force=1 : forced migration
force=2 : offload migration
force=3 : idle pull migration
2. Add trace to expose offload decision-making.
Also adds tracing rq->nr_running so that you can
look back to see what state the RQ was in at the time.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Signed-off-by: Liviu Dudau <Liviu.Dudau@arm.com>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Chris Redpath [Fri, 11 Oct 2013 10:44:58 +0000 (11:44 +0100)]
sched: HMP: Change default HMP thresholds
When the up-threshold is at 512 on TC2, behaviour looks OK since
the graphic-related tasks are very heavy due to lack of a GPU.
Increasing the up-threshold does not reduce power consumption.
When a GPU is present, graphic tasks are much less CPU-heavy and
so additional power may be saved by having a higher threshold.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Signed-off-by: Liviu Dudau <Liviu.Dudau@arm.com>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Chris Redpath [Tue, 23 Jul 2013 13:56:45 +0000 (14:56 +0100)]
HMP: Update migration timer when we fork-migrate
Prevents fork-migration adversely interacting with normal
migration (i.e. runqueues containing forked tasks being
selected as migration targets when there is a better
choice available)
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Signed-off-by: Liviu Dudau <liviu.dudau@arm.com>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Chris Redpath [Mon, 22 Jul 2013 14:56:28 +0000 (15:56 +0100)]
HMP: Access runqueue task clocks directly.
Avoids accesses through cfs_rq going bad when the cpu_rq doesn't
have a cfs member.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Signed-off-by: Liviu Dudau <liviu.dudau@arm.com>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Chris Redpath [Thu, 8 Aug 2013 15:41:26 +0000 (16:41 +0100)]
HMP: Implement idle pull for HMP
When an A15 goes idle, we should up-migrate anything which is
above the threshold and running on an A7.
Reuses the HMP force-migration spinlock, but adds its own new
cpu stopper client.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Signed-off-by: Liviu Dudau <liviu.dudau@arm.com>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Chris Redpath [Thu, 8 Aug 2013 15:10:39 +0000 (16:10 +0100)]
sched: HMP change nr_running offload metric
rq->nr_running was better than cfs.nr_running, since it includes
all tasks actually on the CPU. However, it includes RT tasks which
we would rather ignore at this point.
Switching to cfs.h_nr_running includes all the CFS tasks but no
RT tasks.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Signed-off-by: Liviu Dudau <liviu.dudau@arm.com>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Chris Redpath [Mon, 15 Jul 2013 15:06:44 +0000 (16:06 +0100)]
HMP: Explicitly implement all-load-is-max-load policy for HMP targets
Experimentally, one of the best policies for HMP migration CPU
selection is to completely ignore part-loaded CPUs and only look
for idle ones. If there are no idle ones, we will choose the one
which was least-recently-disturbed.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Signed-off-by: Liviu Dudau <liviu.dudau@arm.com>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Chris Redpath [Thu, 8 Aug 2013 15:32:31 +0000 (16:32 +0100)]
HMP: Modify the runqueue stats to add a new child stat
The original intent here was to track unweighted runqueue load
with less resolution so we could use the least-recently-disturbed
runqueue to choose between 'closely related' load levels.
However, after experimenting with the resolution it turns out
that the following algorithm is highly beneficial for mobile
workloads.
In hmp_domain_min_load:
* If any CPU is zero, the overall load is zero
* If no CPUs are idle, the domain is 'fully loaded'
Additionally, the time since last migration count is used to
discriminate between idle CPUs.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Signed-off-by: Liviu Dudau <liviu.dudau@arm.com>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Chris Redpath [Thu, 8 Aug 2013 15:31:07 +0000 (16:31 +0100)]
sched: track per-rq 'last migration time'
Track when migrations were performed to runqueues.
Use this to decide between runqueues as migration targets when run
queues in an hmp domain have equal load.
Intention is to spread migration load amongst CPUs more fairly.
When all CPUs in an hmp domain are fully loaded, the existing code
always selects the last CPU as a migration target - this is unfair
and little better than doing no selection.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Signed-off-by: Liviu Dudau <liviu.dudau@arm.com>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Morten Rasmussen [Tue, 6 Aug 2013 15:14:19 +0000 (16:14 +0100)]
sched: HMP fix traversing the rb-tree from the curr pointer
The hmp_get_{lightest,heaviest}_task() need to use
__pick_first_entity() to get a pointer to a sched_entity on the rq.
The current is not kept on the rq while running, so its rb-tree node
pointers are no longer valid.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Signed-off-by: Liviu Dudau <liviu.dudau@arm.com>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Chris Redpath [Thu, 8 Aug 2013 15:27:34 +0000 (16:27 +0100)]
HMP: select 'best' task for migration rather than 'current'
When we are looking for a task to migrate up, select the heaviest
one in the first 5 runnable on the runqueue.
Likewise, when looking for a task to offload, select the lightest
one in the first 5 runnable on the runqueue.
Ensure task selected is runnable in the target domain.
This change is necessary in order to implement idle pull in a
sensible manner, but here is used in up-migration and offload to
select the correct target task.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Signed-off-by: Liviu Dudau <liviu.dudau@arm.com>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Jon Medhurst (Tixy) [Fri, 2 Aug 2013 17:45:33 +0000 (18:45 +0100)]
HMP: Check the system has little cpus before forcing rt tasks onto them
It is sometimes desirable to run a kernel with HMP scheduling enabled
on a system which is not big.LITTLE, e.g. when building a multi-platform
kernel, or when testing a big.LITTLE system with one cluster disabled.
We should therefore allow for the situation where is no little domain.
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Signed-off-by: Mark Brown <broonie@linaro.org>
Dietmar Eggemann [Fri, 21 Jun 2013 16:50:08 +0000 (17:50 +0100)]
HMP: experimental: Force all rt tasks to start on little domain.
This patch restricts the allowed cpu mask for rt tasks initially started
with a full cpu mask to the little domain.
An rt task is specified as real time in __setscheduler() which is finally
called for all rt tasks (kernel and user land). In this function we
restrict the allowed cpu mask to the little domain.
This also prevents that a rt tasks can later be pushed to the big domain
because the function find_lowest_rq() will only recognize the allowed cpu
mask of a task to find the new cpu the task runs on.
Current kludges of the patch:
* Since we do not have an API to get the cpu mask of the A7 cluster,
hmp_slow_cpu_mask is made global in arm/kernel/topology.c for now.
* The watchdog_enable() function calls sched_setscheduler() before
kthread_bind() for the cpu specific watchdog kernel threads. The order of
these two calls has to be changed to make this patch work.
Signed-off-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Chris Redpath [Mon, 17 Jun 2013 15:20:37 +0000 (16:20 +0100)]
sched: Restrict nohz balance kicks to stay in the HMP domain
There is little point in doing a nohz balance kick on a CPU from a
different HMP domain, since the unset SD_LOAD_BALANCE flag on the CPU
domain level prevents tasks from being balanced across clusters
except through the per-task load driven hmp_migrate/hmp_offload paths.
Further, the nohz balance kick is actively harmful to power usage if
all the tasks fit into the little domain since it causes the big
domain to wake up and do a lot of calculation to determine that
there is nothing to do.
A more generic solution is to walk the sched domain tree and determine
the intersection of potential idle balance cpus with visibility of
tasks on the current CPU, however HMP domains are more easily
accessible.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Chris Redpath [Mon, 17 Jun 2013 15:08:40 +0000 (16:08 +0100)]
HMP: Force new non-kernel tasks onto big CPUs until load stabilises
Initialise the load stats for new tasks so that they do not
see the instability in early task life which makes it so hard to
decide which CPU is appropriate.
Also, change the fork balance algorithm so that the least loaded of
the CPUs in the big cluster is chosen regardless of the bigness of
the parent task.
This is intended to help performance for applications which use
many short-lived tasks. Although best practise is usually to use
a thread pool, apps which do not do this should not be subject to
the randomness of the early stats.
We should ignore real-time threads for forking on big CPUs, but
it is not possible to figure out if a new thread is real-time or
not at the fork stage. Instead, we prevent kernel threads from
getting the initial boost - when they later become real-time they
will only be on big if their compute requirements demand it.
Signed-off-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Chris Redpath [Thu, 9 May 2013 15:21:29 +0000 (16:21 +0100)]
HMP: Avoid multiple calls to hmp_domain_min_load in fast path
When evaluating a migration we make two calls to hmp_domain_min_load.
This is unnecessary if we pass on the target CPU information from the
hmp_up_migration path.
In hmp_down_migration, we don't consider the load of the target CPUS.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Chris Redpath [Thu, 9 May 2013 15:21:15 +0000 (16:21 +0100)]
HMP: Select least-loaded CPU when performing HMP Migrations
The reference patch set always selects the first CPU in an HMP
domain as a migration target. In busy situations, this means that
the migrated thread cannot make immediate use of an idle CPU but
must share a busy one until the load balancer runs across the big
domain.
This patch uses the hmp_domain_min_load function introduced in
global balancing to figure out which of the CPUs is the least busy
and selects that as a migration target - in both directions.
This essentially implements a task-spread strategy and is intended
to maximise performance of migrated threads but is likely
to use more power than the packing strategy previously employed.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Chris Redpath [Mon, 17 Jun 2013 14:48:15 +0000 (15:48 +0100)]
HMP: Use unweighted load for hmp migration decisions
Normal task and runqueue loading is scaled according to priority
to end up with a weighted load, known as the contribution.
We want the CPU time to be allotted according to priority, but
we also want to make big/little decisions based upon raw load.
It is common, for example, for Android apps following the dev
guide to end up with all their long-running or async action
threads as low priority unless they override the AsyncThread
constructor. All these threads are such low priority that they
become invisible to the hmp_offload routine.
Using unweighted load here allows us to maximise CPU usage in busy
situations.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Chris Redpath [Mon, 17 Jun 2013 14:22:58 +0000 (15:22 +0100)]
Revert "sched: Enable HMP priority filter by default"
This reverts commit
68315334e32932739145ddb41a46cc86b8b056b3.
Having the priority filter enabled prevents proper operation
on Android systems where a wider range of priorities are used
by userspace to partition types of tasks. Those tasks should still
be able to benefit from the use of big CPUs when required.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Chris Redpath [Thu, 16 May 2013 16:48:41 +0000 (17:48 +0100)]
sched: cfs.nr_running does not contain the intended metric
rq->nr_running is the actual number of runnable tasks we wish to use
to determine if a task is alone on a CPU.
Change-Id: Icaf3022e02924ecdc94e14d4146c6fadd9580e2b
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Morten Rasmussen [Thu, 29 Nov 2012 15:41:50 +0000 (15:41 +0000)]
sched: Basic global balancing support for HMP
This patch introduces an extra-check at task up-migration to
prevent overloading the cpus in the faster hmp_domain while the
slower hmp_domain is not fully utilized. The patch also introduces
a periodic balance check that can down-migrate tasks if the faster
domain is oversubscribed and the slower is under-utilized.
Signed-off-by: Morten Rasmussen <morten.rasmussen@arm.com>
Chris Redpath [Tue, 20 Nov 2012 05:34:49 +0000 (11:04 +0530)]
ARM: Fix build breakage when big.LITTLE.conf is not used.
Change-Id: I8641f5e930c65b5672130bd4a18d9868bb3ca594
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Signed-off-by: Liviu Dudau <liviu.dudau@arm.com>
Chris Redpath [Fri, 16 Nov 2012 10:03:00 +0000 (10:03 +0000)]
ARM: Experimental Frequency-Invariant Load Scaling Patch
Evaluation Patch to investigate using load as a representation of the
amount of POTENTIAL cpu compute capacity used rather than a representation
of the CURRENT cpu compute capacity.
If CPUFreq is enabled, scales load in accordance with frequency.
Powersave/performance CPUFreq governors are detected and scaling is
disabled while these governors are in use. This is because when a
single-frequency governor is in use, potential CPU capacity is static.
So long as the governors and CPUFreq subsystem correctly report the
frequencies available, the scaling should self tune.
Adds an additional file to sysfs to allow this feature to be disabled
for experimentation.
/sys/kernel/hmp/frequency_invariant_load_scale
write 0 to disable, 1 to enable.
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Olivier Cozette [Wed, 17 Oct 2012 13:30:30 +0000 (14:30 +0100)]
ARM: Change load tracking scale using sysfs
These functions allow to change the load average period used
in the task load average computation through
/sys/kernel/hmp/load_avg_period_ms. This period is the time
in ms to go from 0 to 0.5 load average while running or the
time from 1 to 0.5 while sleeping.
The default one used is 32 and gives the same load_avg_ratio
computation than without this patch. These functions also allow
to change the up and down threshold of HMP using
/sys/kernel/hmp/{up,down}_threshold. Both must be between 0 and
1024. The thresholds are divided by 1024 before being compared
to the load_avg_ratio.
If /sys/kernel/hmp/load_avg_period_ms is 128 and
/sys/kernel/hmp/up_threshold is 512, a task will be migrated
to a bigger cluster after running for 128ms. Because after
load_avg_period_ms the load average is 0.5 and real up_threshold
us 512 / 1024 = 0.5.
Signed-off-by: Olivier Cozette <olivier.cozette@arm.com>
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Chris Redpath [Thu, 16 May 2013 16:48:24 +0000 (17:48 +0100)]
sched: Ignore offline CPUs in HMP migration & load stats
Previously, an offline CPU would always appear to have a zero load
and this would distort the offload functionality used for balancing
big and little domains.
Maintain a mask of online CPUs in each domain and use this instead.
Change-Id: I639b564b2f40cb659af8ceb8bd37f84b8a1fe323
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Chris Redpath [Thu, 16 May 2013 16:48:01 +0000 (17:48 +0100)]
sched: Do not ignore grouped tasks during HMP forced migration.
If the entity is not a task, it is a cfs group rq. Iterate up to
find the task entity.
Change-Id: I7cab7aba0798f6f14e38ad32e566d90e5937ffbc
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Sudeep KarkadaNagesha [Mon, 24 Sep 2012 13:07:20 +0000 (14:07 +0100)]
sched: fix arch_get_fast_and_slow_cpus to get logical cpumask correctly
The patch "sched: Use device-tree to provide fast/slow CPU list for HMP"
depends on the ordering of CPU's in the device tree. It breaks to determine
the logical mask correctly if the logical mask of the CPUs differ from
physical ordering in the device tree.
This patch fix the logic by depending on the mpidr in the device tree
and mapping that mpidr to the logical cpu.
Signed-off-by: Sudeep KarkadaNagesha <sudeep.karkadanagesha@arm.com>
Signed-off-by: Liviu Dudau <Liviu.Dudau@arm.com>
Morten Rasmussen [Fri, 12 Oct 2012 14:25:02 +0000 (15:25 +0100)]
sched: Only down migrate low priority tasks if allowed by affinity mask
Adds an extra check intersection of the task affinity mask and the slower
hmp_domain cpumask before down migrating low priority tasks.
Signed-off-by: Morten Rasmussen <morten.rasmussen@arm.com>
Jon Medhurst [Fri, 12 Oct 2012 12:45:35 +0000 (13:45 +0100)]
ARM: sched: Avoid empty 'slow' HMP domain
On homogeneous (non-heterogeneous) systems all CPUs will be declared
'fast' and the slow cpu list will be empty. In this situation we need to
avoid adding an empty slow HMP domain otherwise the scheduler code will
blow up when it attempts to move a task to the slow domain.
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Morten Rasmussen [Wed, 10 Oct 2012 13:51:25 +0000 (14:51 +0100)]
sched: Enable HMP priority filter by default
This updates the ARM Kconfig to enable the HMP priority filter by default.
Signed-off-by: Morten Rasmussen <morten.rasmussen@arm.com>
Morten Rasmussen [Fri, 14 Sep 2012 13:38:17 +0000 (14:38 +0100)]
sched: SCHED_HMP multi-domain task migration control
We need a way to prevent tasks that are migrating up and down the
hmp_domains from migrating straight on through before the load has
adapted to the new compute capacity of the CPU on the new hmp_domain.
This patch adds a next up/down migration delay that prevents the task
from doing another migration in the same direction until the delay
has expired.
Signed-off-by: Morten Rasmussen <Morten.Rasmussen@arm.com>
Morten Rasmussen [Fri, 14 Sep 2012 13:38:16 +0000 (14:38 +0100)]
sched: Add HMP task migration ftrace event
Adds ftrace event for tracing task migrations using HMP
optimized scheduling.
Signed-off-by: Morten Rasmussen <Morten.Rasmussen@arm.com>
Morten Rasmussen [Fri, 14 Sep 2012 13:38:15 +0000 (14:38 +0100)]
sched: Add ftrace events for entity load-tracking
Adds ftrace events for key variables related to the entity
load-tracking to help debugging scheduler behaviour. Allows tracing
of load contribution and runqueue residency ratio for both entities
and runqueues as well as entity CPU usage ratio.
Signed-off-by: Morten Rasmussen <Morten.Rasmussen@arm.com>
Morten Rasmussen [Fri, 14 Sep 2012 13:38:14 +0000 (14:38 +0100)]
ARM: sched: Setup SCHED_HMP domains
SCHED_HMP requires the different cpu types to be represented by an
ordered list of hmp_domains. Each hmp_domain represents all cpus of
a particular type using a cpumask.
The list is platform specific and therefore must be generated by
platform code by implementing arch_get_hmp_domains().
Signed-off-by: Morten Rasmussen <Morten.Rasmussen@arm.com>
Morten Rasmussen [Fri, 14 Sep 2012 13:38:13 +0000 (14:38 +0100)]
ARM: sched: Use device-tree to provide fast/slow CPU list for HMP
We can't rely on Kconfig options to set the fast and slow CPU lists for
HMP scheduling if we want a single kernel binary to support multiple
devices with different CPU topology. E.g. TC2 (ARM's Test-Chip-2
big.LITTLE system), Fast Models, or even non big.LITTLE devices.
This patch adds the function arch_get_fast_and_slow_cpus() to generate
the lists at run-time by parsing the CPU nodes in device-tree; it
assumes slow cores are A7s and everything else is fast. The function
still supports the old Kconfig options as this is useful for testing the
HMP scheduler on devices without big.LITTLE.
This patch is reuse of a patch by Jon Medhurst <tixy@linaro.org> with a
few bits left out.
Signed-off-by: Morten Rasmussen <Morten.Rasmussen@arm.com>
Morten Rasmussen [Fri, 14 Sep 2012 13:38:12 +0000 (14:38 +0100)]
ARM: Add HMP scheduling support for ARM architecture
Adds Kconfig entries to enable HMP scheduling on ARM platforms.
Currently, it disables CPU level sched_domain load-balacing in order
to simplify things. This needs fixing in a later revision. HMP
scheduling will do the load-balancing at this level instead.
Signed-off-by: Morten Rasmussen <Morten.Rasmussen@arm.com>
Morten Rasmussen [Fri, 14 Sep 2012 13:38:11 +0000 (14:38 +0100)]
sched: Introduce priority-based task migration filter
Introduces a priority threshold which prevents low priority task
from migrating to faster hmp_domains (cpus). This is useful for
user-space software which assigns lower task priority to background
task.
Signed-off-by: Morten Rasmussen <Morten.Rasmussen@arm.com>
Morten Rasmussen [Fri, 14 Sep 2012 13:38:10 +0000 (14:38 +0100)]
sched: Forced task migration on heterogeneous systems
This patch introduces forced task migration for moving suitable
currently running tasks between hmp_domains. Task behaviour is likely
to change over time. Tasks running in a less capable hmp_domain may
change to become more demanding and should therefore be migrated up.
They are unlikely go through the select_task_rq_fair() path anytime
soon and therefore need special attention.
This patch introduces a period check (SCHED_TICK) of the currently
running task on all runqueues and sets up a forced migration using
stop_machine_no_wait() if the task needs to be migrated.
Ideally, this should not be implemented by polling all runqueues.
Signed-off-by: Morten Rasmussen <Morten.Rasmussen@arm.com>
Morten Rasmussen [Fri, 14 Sep 2012 13:38:09 +0000 (14:38 +0100)]
sched: Task placement for heterogeneous systems based on task load-tracking
This patch introduces the basic SCHED_HMP infrastructure. Each class of
cpus is represented by a hmp_domain and tasks will only be moved between
these domains when their load profiles suggest it is beneficial.
SCHED_HMP relies heavily on the task load-tracking introduced in Paul
Turners fair group scheduling patch set:
<https://lkml.org/lkml/2012/8/23/267>
SCHED_HMP requires that the platform implements arch_get_hmp_domains()
which should set up the platform specific list of hmp_domains. It is
also assumed that the platform disables SD_LOAD_BALANCE for the
appropriate sched_domains.
Tasks placement takes place every time a task is to be inserted into
a runqueue based on its load history. The task placement decision is
based on load thresholds.
There are no restrictions on the number of hmp_domains, however,
multiple (>2) has not been tested and the up/down migration policy is
rather simple.
Signed-off-by: Morten Rasmussen <Morten.Rasmussen@arm.com>
Morten Rasmussen [Fri, 14 Sep 2012 13:38:08 +0000 (14:38 +0100)]
sched: entity load-tracking load_avg_ratio
This patch adds load_avg_ratio to each task. The load_avg_ratio is a
variant of load_avg_contrib which is not scaled by the task priority. It
is calculated like this:
runnable_avg_sum * NICE_0_LOAD / (runnable_avg_period + 1).
Signed-off-by: Morten Rasmussen <Morten.Rasmussen@arm.com>
Paul Turner [Fri, 21 Sep 2012 20:27:51 +0000 (13:27 -0700)]
sched: implement usage tracking
With the frame-work for runnable tracking now fully in place. Per-entity usage
tracking is a simple and low-overhead addition.
Signed-off-by: Paul Turner <pjt@google.com>
Reviewed-by: Ben Segall <bsegall@google.com>
Linus Torvalds [Sun, 30 Jun 2013 22:13:29 +0000 (15:13 -0700)]
Linux 3.10
Linus Torvalds [Sun, 30 Jun 2013 22:08:15 +0000 (15:08 -0700)]
Merge branch 'merge' of git://git./linux/kernel/git/benh/powerpc
Pull another powerpc fix from Benjamin Herrenschmidt:
"I mentioned that while we had fixed the kernel crashes, EEH error
recovery didn't always recover... It appears that I had a fix for
that already in powerpc-next (with a stable CC).
I cherry-picked it today and did a few tests and it seems that things
now work quite well. The patch is also pretty simple, so I see no
reason to wait before merging it."
* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc:
powerpc/eeh: Fix fetching bus for single-dev-PE
Linus Torvalds [Sun, 30 Jun 2013 22:06:25 +0000 (15:06 -0700)]
Merge tag 'scsi-fixes' of git://git./linux/kernel/git/jejb/scsi
Pull SCSI fixes from James Bottomley:
"This is a set of seven bug fixes. Several fcoe fixes for locking
problems, initiator issues and a VLAN API change, all of which could
eventually lead to data corruption, one fix for a qla2xxx locking
problem which could lead to multiple completions of the same request
(and subsequent data corruption) and a use after free in the ipr
driver. Plus one minor MAINTAINERS file update"
(only six bugfixes in this pull, since I had already pulled the fcoe API
fix directly from Robert Love)
* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
[SCSI] ipr: Avoid target_destroy accessing memory after it was freed
[SCSI] qla2xxx: Fix for locking issue between driver ISR and mailbox routines
MAINTAINERS: Fix fcoe mailing list
libfc: extend ex_lock to protect all of fc_seq_send
libfc: Correct check for initiator role
libfcoe: Fix Conflicting FCFs issue in the fabric
Gavin Shan [Wed, 5 Jun 2013 07:34:02 +0000 (15:34 +0800)]
powerpc/eeh: Fix fetching bus for single-dev-PE
While running Linux as guest on top of phyp, we possiblly have
PE that includes single PCI device. However, we didn't return
its PCI bus correctly and it leads to failure on recovery from
EEH errors for single-dev-PE. The patch fixes the issue.
Cc: <stable@vger.kernel.org> # v3.7+
Cc: Steve Best <sbest@us.ibm.com>
Signed-off-by: Gavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Linus Torvalds [Sun, 30 Jun 2013 00:02:48 +0000 (17:02 -0700)]
Merge branch 'merge' of git://git./linux/kernel/git/benh/powerpc
Pull powerpc fixes from Ben Herrenschmidt:
"We discovered some breakage in our "EEH" (PCI Error Handling) code
while doing error injection, due to a couple of regressions. One of
them is due to a patch (
37f02195bee9 "powerpc/pci: fix PCI-e devices
rescan issue on powerpc platform") that, in hindsight, I shouldn't
have merged considering that it caused more problems than it solved.
Please pull those two fixes. One for a simple EEH address cache
initialization issue. The other one is a patch from Guenter that I
had originally planned to put in 3.11 but which happens to also fix
that other regression (a kernel oops during EEH error handling and
possibly hotplug).
With those two, the couple of test machines I've hammered with error
injection are remaining up now. EEH appears to still fail to recover
on some devices, so there is another problem that Gavin is looking
into but at least it's no longer crashing the kernel."
* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc:
powerpc/pci: Improve device hotplug initialization
powerpc/eeh: Add eeh_dev to the cache during boot
Olof Johansson [Sat, 29 Jun 2013 23:25:14 +0000 (16:25 -0700)]
ARM: dt: Only print warning, not WARN() on bad cpu map in device tree
Due to recent changes and expecations of proper cpu bindings, there are
now cases for many of the in-tree devicetrees where a WARN() will hit
on boot due to badly formatted /cpus nodes.
Downgrade this to a pr_warn() to be less alarmist, since it's not a
new problem.
Tested on Arndale, Cubox, Seaboard and Panda ES. Panda hits the WARN
without this, the others do not.
Acked-by: Russell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Guenter Roeck [Mon, 10 Jun 2013 17:18:08 +0000 (10:18 -0700)]
powerpc/pci: Improve device hotplug initialization
Commit
37f02195b (powerpc/pci: fix PCI-e devices rescan issue on powerpc
platform) fixes a problem with interrupt and DMA initialization on hot
plugged devices. With this commit, interrupt and DMA initialization for
hot plugged devices is handled in the pci device enable function.
This approach has a couple of drawbacks. First, it creates two code paths
for device initialization, one for hot plugged devices and another for devices
known during the initial PCI scan. Second, the initialization code for hot
plugged devices is only called when the device is enabled, ie typically
in the probe function. Also, the platform specific setup code is called each
time pci_enable_device() is called, not only once during device discovery,
meaning it is actually called multiple times, once for devices discovered
during the initial scan and again each time a driver is re-loaded.
The visible result is that interrupt pins are only assigned to hot plugged
devices when the device driver is loaded. Effectively this changes the PCI
probe API, since pci_dev->irq and the device's dma configuration will now
only be valid after pci_enable() was called at least once. A more subtle
change is that platform specific PCI device setup is moved from device
discovery into the driver's probe function, more specifically into the
pci_enable_device() call.
To fix the inconsistencies, add new function pcibios_add_device.
Call pcibios_setup_device from pcibios_setup_bus_devices if device setup
is not complete, and from pcibios_add_device if bus setup is complete.
With this change, device setup code is moved back into device initialization,
and called exactly once for both static and hot plugged devices.
[ This also fixes a regression introduced by the above patch which
causes dev->irq to be overwritten under some cirumstances after
MSIs have been enabled for the device which leads to crashes due
to the MSI core "hijacking" dev->irq to store the base MSI number
and not the LSI. --BenH
]
Cc: Yuanquan Chen <Yuanquan.Chen@freescale.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Hiroo Matsumoto <matsumoto.hiroo@jp.fujitsu.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Linus Torvalds [Sat, 29 Jun 2013 18:34:18 +0000 (11:34 -0700)]
Merge git://git./linux/kernel/git/herbert/crypto-2.6
Pull crypto fix from Herbert Xu:
"This fixes a crash in the crypto layer exposed by an SCTP test tool"
* git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6:
crypto: algboss - Hold ref count on larval
Linus Torvalds [Sat, 29 Jun 2013 18:32:05 +0000 (11:32 -0700)]
Merge branch 'drm-fixes' of git://people.freedesktop.org/~airlied/linux
Pull drm/qxl fix from Dave Airlie:
"Bad me forgot an access check, possible security issue, but since this
is the first kernel with it, should be fine to just put it in now"
* 'drm-fixes' of git://people.freedesktop.org/~airlied/linux:
drm/qxl: add missing access check for execbuffer ioctl
Mathieu Desnoyers [Fri, 28 Jun 2013 13:49:46 +0000 (09:49 -0400)]
Fix: kernel/ptrace.c: ptrace_peek_siginfo() missing __put_user() validation
This __put_user() could be used by unprivileged processes to write into
kernel memory. The issue here is that even if copy_siginfo_to_user()
fails, the error code is not checked before __put_user() is executed.
Luckily, ptrace_peek_siginfo() has been added within the 3.10-rc cycle,
so it has not hit a stable release yet.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Acked-by: Oleg Nesterov <oleg@redhat.com>
Cc: Andrey Vagin <avagin@openvz.org>
Cc: Roland McGrath <roland@redhat.com>
Cc: Paul McKenney <paulmck@linux.vnet.ibm.com>
Cc: David Howells <dhowells@redhat.com>
Cc: Dave Jones <davej@redhat.com>
Cc: Pavel Emelyanov <xemul@parallels.com>
Cc: Pedro Alves <palves@redhat.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Linus Torvalds [Sat, 29 Jun 2013 17:31:15 +0000 (10:31 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/sage/ceph-client
Pull Ceph fix from Sage Weil:
"This is a recently spotted regression in the snapshot behavior...
It turns out several tests weren't being run in the nightlies so this
took a while to spot"
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client:
rbd: send snapshot context with writes
Linus Torvalds [Sat, 29 Jun 2013 17:30:31 +0000 (10:30 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/viro/vfs
Pull ubifs fixes from Al Viro:
"A couple of ubifs readdir/lseek race fixes. Stable fodder, really
nasty..."
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
UBIFS: fix a horrid bug
UBIFS: prepare to fix a horrid bug
Linus Torvalds [Sat, 29 Jun 2013 17:28:52 +0000 (10:28 -0700)]
Merge tag 'for-linus-
20130628' of git://git./linux/kernel/git/dhowells/linux-mn10300
Pull two MN10300 fixes from David Howells:
"The first fixes a problem with passing arrays rather than pointers to
get_user() where __typeof__ then wants to declare and initialise an
array variable which gcc doesn't like.
The second fixes a problem whereby putting mem=xxx into the kernel
command line causes init=xxx to get an incorrect value."
* tag 'for-linus-
20130628' of git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-mn10300:
mn10300: Use early_param() to parse "mem=" parameter
mn10300: Allow to pass array name to get_user()
Linus Torvalds [Sat, 29 Jun 2013 17:27:19 +0000 (10:27 -0700)]
Merge branch 'timers-urgent-for-linus' of git://git./linux/kernel/git/tip/tip
Pull timer fix from Thomas Gleixner:
"Correct an ordering issue in the tick broadcast code. I really wish
we'd get compensation for pain and suffering for each line of code we
write to work around dysfunctional timer hardware."
* 'timers-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
tick: Fix tick_broadcast_pending_mask not cleared
Linus Torvalds [Sat, 29 Jun 2013 17:26:50 +0000 (10:26 -0700)]
Merge branch 'perf-urgent-for-linus' of git://git./linux/kernel/git/tip/tip
Pull perf fix from Ingo Molnar:
"One more fix for a recently discovered bug"
* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
perf: Disable monitoring on setuid processes for regular users
Artem Bityutskiy [Fri, 28 Jun 2013 11:15:15 +0000 (14:15 +0300)]
UBIFS: fix a horrid bug
Al Viro pointed me to the fact that '->readdir()' and '->llseek()' have no
mutual exclusion, which means the 'ubifs_dir_llseek()' can be run while we are
in the middle of 'ubifs_readdir()'.
This means that 'file->private_data' can be freed while 'ubifs_readdir()' uses
it, and this is a very bad bug: not only 'ubifs_readdir()' can return garbage,
but this may corrupt memory and lead to all kinds of problems like crashes an
security holes.
This patch fixes the problem by using the 'file->f_version' field, which
'->llseek()' always unconditionally sets to zero. We set it to 1 in
'ubifs_readdir()' and whenever we detect that it became 0, we know there was a
seek and it is time to clear the state saved in 'file->private_data'.
I tested this patch by writing a user-space program which runds readdir and
seek in parallell. I could easily crash the kernel without these patches, but
could not crash it with these patches.
Cc: stable@vger.kernel.org
Reported-by: Al Viro <viro@zeniv.linux.org.uk>
Tested-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Artem Bityutskiy [Fri, 28 Jun 2013 11:15:14 +0000 (14:15 +0300)]
UBIFS: prepare to fix a horrid bug
Al Viro pointed me to the fact that '->readdir()' and '->llseek()' have no
mutual exclusion, which means the 'ubifs_dir_llseek()' can be run while we are
in the middle of 'ubifs_readdir()'.
First of all, this means that 'file->private_data' can be freed while
'ubifs_readdir()' uses it. But this particular patch does not fix the problem.
This patch is only a preparation, and the fix will follow next.
In this patch we make 'ubifs_readdir()' stop using 'file->f_pos' directly,
because 'file->f_pos' can be changed by '->llseek()' at any point. This may
lead 'ubifs_readdir()' to returning inconsistent data: directory entry names
may correspond to incorrect file positions.
So here we introduce a local variable 'pos', read 'file->f_pose' once at very
the beginning, and then stick to 'pos'. The result of this is that when
'ubifs_dir_llseek()' changes 'file->f_pos' while we are in the middle of
'ubifs_readdir()', the latter "wins".
Cc: stable@vger.kernel.org
Reported-by: Al Viro <viro@zeniv.linux.org.uk>
Tested-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Akira Takeuchi [Fri, 28 Jun 2013 15:53:03 +0000 (16:53 +0100)]
mn10300: Use early_param() to parse "mem=" parameter
This fixes the problem that "init=" options may not be passed to kernel
correctly.
parse_mem_cmdline() of mn10300 arch gets rid of "mem=" string from
redboot_command_line. Then init_setup() parses the "init=" options from
static_command_line, which is a copy of redboot_command_line, and keeps
the pointer to the init options in execute_command variable.
Since the commit
026cee0 upstream (params: <level>_initcall-like kernel
parameters), static_command_line becomes overwritten by saved_command_line at
do_initcall_level(). Notice that saved_command_line is a command line
which includes "mem=" string.
As a result, execute_command may point to weird string by the length of
"mem=" parameter.
I noticed this problem when using the command line like this:
mem=128M console=ttyS0,115200 init=/bin/sh
Here is the processing flow of command line parameters.
start_kernel()
setup_arch(&command_line)
parse_mem_cmdline(cmdline_p)
* strcpy(boot_command_line, redboot_command_line);
* Remove "mem=xxx" from redboot_command_line.
* *cmdline_p = redboot_command_line;
setup_command_line(command_line) <-- command_line is redboot_command_line
* strcpy(saved_command_line, boot_command_line)
* strcpy(static_command_line, command_line)
parse_early_param()
strlcpy(tmp_cmdline, boot_command_line, COMMAND_LINE_SIZE);
parse_early_options(tmp_cmdline);
parse_args("early options", cmdline, NULL, 0, 0, 0, do_early_param);
parse_args("Booting ..", static_command_line, ...);
init_setup() <-- save the pointer in execute_command
rest_init()
kernel_thread(kernel_init, NULL, CLONE_FS | CLONE_SIGHAND);
At this point, execute_command points to "/bin/sh" string.
kernel_init()
kernel_init_freeable()
do_basic_setup()
do_initcalls()
do_initcall_level()
(*) strcpy(static_command_line, saved_command_line);
Here, execute_command gets to point to "200" string !!
Signed-off-by: David Howells <dhowells@redhat.com>
Akira Takeuchi [Fri, 28 Jun 2013 15:53:01 +0000 (16:53 +0100)]
mn10300: Allow to pass array name to get_user()
This fixes the following compile error:
CC block/scsi_ioctl.o
block/scsi_ioctl.c: In function 'sg_scsi_ioctl':
block/scsi_ioctl.c:449: error: invalid initializer
Signed-off-by: David Howells <dhowells@redhat.com>
Dave Airlie [Fri, 28 Jun 2013 03:27:40 +0000 (13:27 +1000)]
drm/qxl: add missing access check for execbuffer ioctl
Reported-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Thadeu Lima de Souza Cascardo [Thu, 27 Jun 2013 21:00:10 +0000 (18:00 -0300)]
powerpc/eeh: Add eeh_dev to the cache during boot
commit
f8f7d63fd96ead101415a1302035137a866f8998 ("powerpc/eeh: Trace eeh
device from I/O cache") broke EEH on pseries for devices that were
present during boot and have not been hotplugged/DLPARed.
eeh_check_failure will get the eeh_dev from the cache, and will get
NULL. eeh_addr_cache_build adds the addresses to the cache, but eeh_dev
for the giving pci_device is not set yet. Just reordering the call to
eeh_addr_cache_insert_dev works fine. The ordering is similar to the one
in eeh_add_device_late.
Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@linux.vnet.ibm.com>
Acked-by: Gavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Josh Durgin [Wed, 26 Jun 2013 19:56:17 +0000 (12:56 -0700)]
rbd: send snapshot context with writes
Sending the right snapshot context with each write is required for
snapshots to work. Due to the ordering of calls, the snapshot context
is never set for any requests. This causes writes to the current
version of the image to be reflected in all snapshots, which are
supposed to be read-only.
This happens because rbd_osd_req_format_write() sets the snapshot
context based on obj_request->img_request. At this point, however,
obj_request->img_request has not been set yet, to the snapshot context
is set to NULL. Fix this by moving rbd_img_obj_request_add(), which
sets obj_request->img_request, before the osd request formatting
calls.
This resolves:
http://tracker.ceph.com/issues/5465
Reported-by: Karol Jurak <karol.jurak@gmail.com>
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Reviewed-by: Sage Weil <sage@inktank.com>
Reviewed-by: Alex Elder <elder@linaro.org>
James Bottomley [Thu, 27 Jun 2013 06:08:22 +0000 (23:08 -0700)]
Merge tag 'fcoe1' into fixes
This patch fixes a critical bug that was introduced in 3.9
related to VLAN tagging FCoE frames.
James Bottomley [Thu, 27 Jun 2013 06:07:53 +0000 (23:07 -0700)]
Merge tag 'fcoe' into fixes
3.10 fixes
Linus Torvalds [Thu, 27 Jun 2013 05:24:37 +0000 (19:24 -1000)]
Merge git://git./linux/kernel/git/davem/net
Pull networking fixes from David Miller:
1) Found via trinity:
If you connect up an ipv6 socket to an ipv4 mapped address then an
ipv6 one, sendmsg() can croak because ip6_sk_dst_check() assumes the
route cached in the socket is an ipv6 one. In this case there is an
ipv4 route attached, so it gets stomped on.
Reported by Dave Jones and Hannes Frederic Sowa, fixed by Eric
Dumazet.
2) AF_KEY notifications leak some kernel memory to userspace, fix from
Mathias Krause.
3) DLCI calls __dev_get_by_name() without proper locking, and dlci_del
doesn't validate that the device being deleted is actually a DLCI
one. Fixes from Li Zefan.
4) Length check on bluetooth l2cap information responses is wrong, each
response type has a different lenth, so we should make sure it's in
a given range rather than enforce one single valid length. From
Jaganath Kanakkassery.
5) Receive FIFO overflow is really easy to trigger in stress scenerios
in the sh_eth driver, but the event isn't being handled properly at
all. Specifically, the mask of error interrupts doesn't include the
event so we never clear it, resulting in the driver becomming wedged
processing an interrupt that never gets cleared.
Fix from Sergei Shtylyov.
6) qlcnic sleeps while holding a spinlock, use mdelay() instead of
msleep(). From Shahed Shaikh.
7) Missing curly braces causes SIP netfilter NAT module to always drop
packets. Fix from Balazs Peter Odor.
8) ipt_ULOG in netfilter passes the wrong value to timer setup, causing
the timer to dereference crap when it fires. Fix from Gao Feng.
9) Missing RCU protection around txq->axq_acq traversal in
ath_txq_schedule(). Fix from Felix Fietkau.
10) Idle state transition test in ath9k_htc_config() is reversed, fix
from Sujith Manoharan.
11) IPV6 forwarding handles unicast Router Alert packets incorrectly.
It tests the wrong option state. Previously opt->ra being non-zero
indicated a router alert marking in the SKB, but now it's indicated
by a bit in opt->flags. Fix from YOSHIFUJI Hideaki.
12) SKB leak in GRE tunnel GSO handling, from Eric Dumazet.
13) get_user_pages_fast() error handling in TUN and MACVTAP use the same
local variable for the base index and the loop iterator for page
traversal, oops! Fix from Michael S Tsirkin.
14) ipv6_get_lladdr() can fail, and we must therefore check it's return
value in inet6_set_iftoken(). For from Hannes Frederic Sowa.
15) If you change an interface name and meanwhile can sneak in something
that looks up the name (like SO_BINDTODEVICE or SIOCGIFNAME) we can
deadlock with CONFIG_PREEMPT=n. Fix this by providing a helper
function that properly uses raw_seqcount_begin(). From Nicolas
Schichan.
16) Chain noise calibration test is inverted in iwlwifi, fix from
Nikolay Martynov.
17) Properly set TX iwlwifi descriptor flags for back requests. Fix
from Emmanuel Grumbach.
18) We can't assume skb_transport_header() is set in xt_TCPOPTSTRAP
module, fix from Pablo Neira Ayuso.
19) Some crummy APs don't provide the proper High Throughput info in
association response frames. Add a workaround by assume we'll use
whatever is in the beacon/probe. Fix from Johannes Berg.
20) mac80211 call to rate_idx_match_mask() swaps two arguments (mask and
channel width). Fix from Simon Wunderlich.
21) xt_TCPMSS (like xt_TCPOPTSTRAP) must not try to handle fragmented
frames. Fix from Phil Oester.
22) Fix rate control regression causing iwlwifi/iwlegacy chips to use
1Mbit/s on pre-11n networks. From Moshe Benji and Stanslaw Gruszka.
23) Disable brcmsmac power-save functions, they cause regressions. From
Arend van Spriel.
24) Enforce a sane minimum MTU in l2cap_build_cmd() otherwise we can
easily crash. Fix from Anderson Lizardo.
25) If a learning packet arrives during vxlan_stop() we crash, easily
fixed by checking netif_running(). From Stephen Hemminger.
26) Static vxlan FDB entries should not be migrated, also from Stephen.
27) skb_clone() failures not handled in vxlan_xmit(), oops. Also from
Stephen.
28) Add minimal driver for AR816x/AR817x ethernet chips, from Johannes
Berg.
29) Fix regression in userspace VLAN acceleration control, added by the
802.1ad support changes. Fix from Fernando Luis Vazquez Cao.
30) Interval selection for MLD queries in the bridging code was
reversed. Fix from Linus Lüssing.
31) ipv6's ndisc_send_redirect() erroneously writes to the packet we
received not the packet we are building to send out. Fix from
Matthias Schiffer.
32) Don't free netdev before unregistering it, in usb_8dev can driver.
From Marc Kleine-Budde.
33) Fix nl80211 attribute buffer races, from Johannes Berg.
34) Although netlink_diag.h is under uapi/ it isn't present in Kbuild.
From Stephen Hemminger.
35) Wrong address and family passed to MD5 key lookups in TCP, from
Aydin Arik.
36) phy_type attribute created by SFC driver should not be writable.
From Ben Hutchings.
37) Receive/Transmit queue allocations in pxa168_eth and mv643xx_eth
should use kzalloc(). Otherwise if setup fails half-way, we'll
dereference garbage when trying to teardown the rings. From Lubomir
Rintel.
38) Fix double-allocation of dst (resulting in unfreeable net device) in
ipv6's init_loopback(). From Gao Feng.
39) Fix fragmentation handling SKB leak in netfilter conntrack, we were
freeing the wrong skb pointer. From Phil Oester.
40) Don't report "-1" (SPEED_UNKNOWN) in bond_miimon_commit(), from
Nikolay Aleksandrov.
41) davinci_cpdma doesn't check for DMA mapping errors, letting the
device scribble to random addresses. From Sebastian Siewior.
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (69 commits)
dlci: validate the net device in dlci_del()
dlci: acquire rtnl_lock before calling __dev_get_by_name()
af_key: fix info leaks in notify messages
ipv6: ip6_sk_dst_check() must not assume ipv6 dst
net: fix kernel deadlock with interface rename and netdev name retrieval.
net/tg3: Avoid delay during MMIO access
ipv6: check return value of ipv6_get_lladdr
macvtap: fix recovery from gup errors
tun: fix recovery from gup errors
gre: fix a possible skb leak
ipv6: Process unicast packet with Router Alert by checking flag in skb.
ath9k_htc: Handle IDLE state transition properly
ath9k: fix an RCU issue in calling ieee80211_get_tx_rates
netfilter: ipt_ULOG: fix incorrect setting of ulog timer
netfilter: ctnetlink: send event when conntrack label was modified
netfilter: nf_nat_sip: fix mangling
qlcnic: Do not sleep while holding spinlock
drivers: net: cpsw: fix compilation error with cpsw driver
tcp: doc : fix the syncookies default value
sh_eth: fix misreporting of transmit abort
...
Linus Torvalds [Thu, 27 Jun 2013 05:23:15 +0000 (19:23 -1000)]
Merge branch 'drm-fixes' of git://people.freedesktop.org/~airlied/linux
Pull i915 drm fixes from Dave Airlie:
"These should be the last two fixes for i915, one is for a fence leak
killing X on some older GPUs, and one is a late regression partial
revert for an swiotlb/xen/i915 interaction, Konrad has promised to
figure out the proper answer, and this patch is the best thing to do
at this stage to avoid regressing"
* 'drm-fixes' of git://people.freedesktop.org/~airlied/linux:
drm/i915: make compact dma scatter lists creation work with SWIOTLB backend.
drm/i915: Restore fences after resume and GPU resets
Zefan Li [Wed, 26 Jun 2013 07:31:58 +0000 (15:31 +0800)]
dlci: validate the net device in dlci_del()
We triggered an oops while running trinity with 3.4 kernel:
BUG: unable to handle kernel paging request at
0000000100000d07
IP: [<
ffffffffa0109738>] dlci_ioctl+0xd8/0x2d4 [dlci]
PGD
640c0d067 PUD 0
Oops: 0000 [#1] PREEMPT SMP
CPU 3
...
Pid: 7302, comm: trinity-child3 Not tainted 3.4.24.09+ 40 Huawei Technologies Co., Ltd. Tecal RH2285 /BC11BTSA
RIP: 0010:[<
ffffffffa0109738>] [<
ffffffffa0109738>] dlci_ioctl+0xd8/0x2d4 [dlci]
...
Call Trace:
[<
ffffffff8137c5c3>] sock_ioctl+0x153/0x280
[<
ffffffff81195494>] do_vfs_ioctl+0xa4/0x5e0
[<
ffffffff8118354a>] ? fget_light+0x3ea/0x490
[<
ffffffff81195a1f>] sys_ioctl+0x4f/0x80
[<
ffffffff81478b69>] system_call_fastpath+0x16/0x1b
...
It's because the net device is not a dlci device.
Reported-by: Li Jinyue <lijinyue@huawei.com>
Signed-off-by: Li Zefan <lizefan@huawei.com>
Cc: stable@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
Zefan Li [Wed, 26 Jun 2013 07:29:54 +0000 (15:29 +0800)]
dlci: acquire rtnl_lock before calling __dev_get_by_name()
Otherwise the net device returned can be freed at anytime.
Signed-off-by: Li Zefan <lizefan@huawei.com>
Cc: stable@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
Mathias Krause [Wed, 26 Jun 2013 21:52:30 +0000 (23:52 +0200)]
af_key: fix info leaks in notify messages
key_notify_sa_flush() and key_notify_policy_flush() miss to initialize
the sadb_msg_reserved member of the broadcasted message and thereby
leak 2 bytes of heap memory to listeners. Fix that.
Signed-off-by: Mathias Krause <minipli@googlemail.com>
Cc: Steffen Klassert <steffen.klassert@secunet.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Wed, 26 Jun 2013 11:15:07 +0000 (04:15 -0700)]
ipv6: ip6_sk_dst_check() must not assume ipv6 dst
It's possible to use AF_INET6 sockets and to connect to an IPv4
destination. After this, socket dst cache is a pointer to a rtable,
not rt6_info.
ip6_sk_dst_check() should check the socket dst cache is IPv6, or else
various corruptions/crashes can happen.
Dave Jones can reproduce immediate crash with
trinity -q -l off -n -c sendmsg -c connect
With help from Hannes Frederic Sowa
Reported-by: Dave Jones <davej@redhat.com>
Reported-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Nicolas Schichan [Wed, 26 Jun 2013 15:23:42 +0000 (17:23 +0200)]
net: fix kernel deadlock with interface rename and netdev name retrieval.
When the kernel (compiled with CONFIG_PREEMPT=n) is performing the
rename of a network interface, it can end up waiting for a workqueue
to complete. If userland is able to invoke a SIOCGIFNAME ioctl or a
SO_BINDTODEVICE getsockopt in between, the kernel will deadlock due to
the fact that read_secklock_begin() will spin forever waiting for the
writer process (the one doing the interface rename) to update the
devnet_rename_seq sequence.
This patch fixes the problem by adding a helper (netdev_get_name())
and using it in the code handling the SIOCGIFNAME ioctl and
SO_BINDTODEVICE setsockopt.
The netdev_get_name() helper uses raw_seqcount_begin() to avoid
spinning forever, waiting for devnet_rename_seq->sequence to become
even. cond_resched() is used in the contended case, before retrying
the access to give the writer process a chance to finish.
The use of raw_seqcount_begin() will incur some unneeded work in the
reader process in the contended case, but this is better than
deadlocking the system.
Signed-off-by: Nicolas Schichan <nschichan@freebox.fr>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>