The current timer wheel has some drawbacks:
1) Cascading:
Cascading can be an unbound operation and is completely pointless in most
cases because the vast majority of the timer wheel timers are canceled or
rearmed before expiration. (They are used as timeout safeguards, not as
real timers to measure time.)
2) No fast lookup of the next expiring timer:
In NOHZ scenarios the first timer soft interrupt after a long NOHZ period
must fast forward the base time to the current value of jiffies. As we
have no way to find the next expiring timer fast, the code loops linearly
and increments the base time one by one and checks for expired timers
in each step. This causes unbound overhead spikes exactly in the moment
when we should wake up as fast as possible.
After a thorough analysis of real world data gathered on laptops,
workstations, webservers and other machines (thanks Chris!) I came to the
conclusion that the current 'classic' timer wheel implementation can be
modified to address the above issues.
The vast majority of timer wheel timers is canceled or rearmed before
expiry. Most of them are timeouts for networking and other I/O tasks. The
nature of timeouts is to catch the exception from normal operation (TCP ack
timed out, disk does not respond, etc.). For these kinds of timeouts the
accuracy of the timeout is not really a concern. Timeouts are very often
approximate worst-case values and in case the timeout fires, we already
waited for a long time and performance is down the drain already.
The few timers which actually expire can be split into two categories:
1) Short expiry times which expect halfways accurate expiry
2) Long term expiry times are inaccurate today already due to the
batching which is done for NOHZ automatically and also via the
set_timer_slack() API.
So for long term expiry timers we can avoid the cascading property and just
leave them in the less granular outer wheels until expiry or
cancelation. Timers which are armed with a timeout larger than the wheel
capacity are no longer cascaded. We expire them with the longest possible
timeout (6+ days). We have not observed such timeouts in our data collection,
but at least we handle them, applying the rule of the least surprise.
To avoid extending the wheel levels for HZ=1000 so we can accomodate the
longest observed timeouts (5 days in the network conntrack code) we reduce the
first level granularity on HZ=1000 to 4ms, which effectively is the same as
the HZ=250 behaviour. From our data analysis there is nothing which relies on
that 1ms granularity and as a side effect we get better batching and timer
locality for the networking code as well.
Contrary to the classic wheel the granularity of the next wheel is not the
capacity of the first wheel. The granularities of the wheels are in the
currently chosen setting 8 times the granularity of the previous wheel.
So for HZ=250 we end up with the following granularity levels:
Level Offset Granularity Range
0 0 4 ms 0 ms - 252 ms
1 64 32 ms 256 ms - 2044 ms (256ms - ~2s)
2 128 256 ms 2048 ms - 16380 ms (~2s - ~16s)
3 192 2048 ms (~2s) 16384 ms - 131068 ms (~16s - ~2m)
4 256 16384 ms (~16s) 131072 ms - 1048572 ms (~2m - ~17m)
5 320 131072 ms (~2m) 1048576 ms - 8388604 ms (~17m - ~2h)
6 384 1048576 ms (~17m) 8388608 ms - 67108863 ms (~2h - ~18h)
7 448 8388608 ms (~2h) 67108864 ms - 536870911 ms (~18h - ~6d)
That's a worst case inaccuracy of 12.5% for the timers which are queued at the
beginning of a level.
So the new wheel concept addresses the old issues:
1) Cascading is avoided completely
2) By keeping the timers in the bucket until expiry/cancelation we can track
the buckets which have timers enqueued in a bucket bitmap and therefore can
look up the next expiring timer very fast and O(1).
A further benefit of the concept is that the slack calculation which is done
on every timer start is no longer necessary because the granularity levels
provide natural batching already.
Our extensive testing with various loads did not show any performance
degradation vs. the current wheel implementation.
This patch does not address the 'fast lookup' issue as we wanted to make sure
that there is no regression introduced by the wheel redesign. The
optimizations are in follow up patches.
This patch contains fixes from Anna-Maria Gleixner and Richard Cochran.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Arjan van de Ven <arjan@infradead.org>
Cc: Chris Mason <clm@fb.com>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: George Spelvin <linux@sciencehorizons.net>
Cc: Josh Triplett <josh@joshtriplett.org>
Cc: Len Brown <lenb@kernel.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: rt@linutronix.de
Link: http://lkml.kernel.org/r/20160704094342.108621834@linutronix.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
We want to move the timer migration logic from a 'push' to a 'pull' model.
Under the current 'push' model pinned timers are handled via
a runtime API variant: mod_timer_pinned().
The 'pull' model requires us to store the pinned attribute of a timer
in the timer_list structure itself, as a new TIMER_PINNED bit in
timer->flags.
This flag must be set at initialization time and the timer APIs
recognize the flag.
This patch:
- Implements the new flag and associated new-style initialization
methods
- makes mod_timer() recognize new-style pinned timers,
- and adds some migration helper facility to allow
step by step conversion of old-style to new-style
pinned timers.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Arjan van de Ven <arjan@infradead.org>
Cc: Chris Mason <clm@fb.com>
Cc: Eric Dumazet <edumazet@google.com>
Cc: George Spelvin <linux@sciencehorizons.net>
Cc: Josh Triplett <josh@joshtriplett.org>
Cc: Len Brown <lenb@kernel.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: rt@linutronix.de
Link: http://lkml.kernel.org/r/20160704094341.049338558@linutronix.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The following commit:
66eb579e66 ("perf: allow for PMU-specific event filtering")
added the pmu::filter_match() callback. This was intended to
avoid HW constraints on events from resulting in extremely
pessimistic scheduling.
However, pmu::filter_match() is only called for the leader of each event
group. When the leader is a SW event, we do not filter the groups, and
may fail at pmu::add() time, and when this happens we'll give up on
scheduling any event groups later in the list until they are rotated
ahead of the failing group.
This can result in extremely sub-optimal event scheduling behaviour,
e.g. if running the following on a big.LITTLE platform:
$ taskset -c 0 ./perf stat \
-e 'a57{context-switches,armv8_cortex_a57/config=0x11/}' \
-e 'a53{context-switches,armv8_cortex_a53/config=0x11/}' \
ls
<not counted> context-switches (0.00%)
<not counted> armv8_cortex_a57/config=0x11/ (0.00%)
24 context-switches (37.36%)
57589154 armv8_cortex_a53/config=0x11/ (37.36%)
Here the 'a53' event group was always eligible to be scheduled, but
the 'a57' group never eligible to be scheduled, as the task was always
affine to a Cortex-A53 CPU. The SW (group leader) event in the 'a57'
group was eligible, but the HW event failed at pmu::add() time,
resulting in ctx_flexible_sched_in giving up on scheduling further
groups with HW events.
One way of avoiding this is to check pmu::filter_match() on siblings
as well as the group leader. If any of these fail their
pmu::filter_match() call, we must skip the entire group before
attempting to add any events.
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Will Deacon <will.deacon@arm.com>
Fixes: 66eb579e66 ("perf: allow for PMU-specific event filtering")
Link: http://lkml.kernel.org/r/1465917041-15339-1-git-send-email-mark.rutland@arm.com
[ Small readability edits. ]
Signed-off-by: Ingo Molnar <mingo@kernel.org>
virq is not required to be the same for all msi descs. Use the base irq number
from the desc in the debug printk.
Reported-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Use the affinity hint in the irqdesc allocator. The hint is used to determine
the node for the allocation and to set the affinity of the interrupt.
If multiple interrupts are allocated (multi-MSI) then the allocator iterates
over the cpumask and for each set cpu it allocates on their node and sets the
initial affinity to that cpu.
If a single interrupt is allocated (MSI-X) then the allocator uses the first
cpu in the mask to compute the allocation node and uses the mask for the
initial affinity setting.
Interrupts set up this way are marked with the AFFINITY_MANAGED flag to
prevent userspace from messing with their affinity settings.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Christoph Hellwig <hch@lst.de>
Cc: linux-block@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Cc: linux-nvme@lists.infradead.org
Cc: axboe@fb.com
Cc: agordeev@redhat.com
Link: http://lkml.kernel.org/r/1467621574-8277-5-git-send-email-hch@lst.de
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
While reviewing another patch I noticed that kernel/time/tick-sched.c
had a charmingly (confusingly, annoyingly) rich set of variants for
spelling 'CPU':
cpu
cpus
CPU
CPUs
per CPU
per-CPU
per cpu
... sometimes these were mixed even within the same comment block!
Compress these variants down to a single consistent set of:
CPU
CPUs
per-CPU
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
EXPORT_SYMBOL() get_monotonic_coarse64 for new IIO timestamping clock
selection usage. This provides user apps the ability to request a
particular IIO device to timestamp samples using a monotonic coarse clock
granularity.
Signed-off-by: Gregor Boirie <gregor.boirie@parrot.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
Pull RCU changes from Paul E. McKenney:
- Documentation updates. Just some simple changes, no design-level
additions.
- Miscellaneous fixes.
- Torture-test updates.
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull audit fixes from Paul Moore:
"Two small patches to fix audit problems in 4.7-rcX: the first fixes a
potential kref leak, the second removes some header file noise.
The first is an important bug fix that really should go in before 4.7
is released, the second is not critical, but falls into the very-nice-
to-have category so I'm including in the pull request.
Both patches are straightforward, self-contained, and pass our
testsuite without problem"
* 'stable-4.7' of git://git.infradead.org/users/pcmoore/audit:
audit: move audit_get_tty to reduce scope and kabi changes
audit: move calcs after alloc and check when logging set loginuid
Pull networking fixes from David Miller:
"I've been traveling so this accumulates more than week or so of bug
fixing. It perhaps looks a little worse than it really is.
1) Fix deadlock in ath10k driver, from Ben Greear.
2) Increase scan timeout in iwlwifi, from Luca Coelho.
3) Unbreak STP by properly reinjecting STP packets back into the
stack. Regression fix from Ido Schimmel.
4) Mediatek driver fixes (missing malloc failure checks, leaking of
scratch memory, wrong indexing when mapping TX buffers, etc.) from
John Crispin.
5) Fix endianness bug in icmpv6_err() handler, from Hannes Frederic
Sowa.
6) Fix hashing of flows in UDP in the ruseport case, from Xuemin Su.
7) Fix netlink notifications in ovs for tunnels, delete link messages
are never emitted because of how the device registry state is
handled. From Nicolas Dichtel.
8) Conntrack module leaks kmemcache on unload, from Florian Westphal.
9) Prevent endless jump loops in nft rules, from Liping Zhang and
Pablo Neira Ayuso.
10) Not early enough spinlock initialization in mlx4, from Eric
Dumazet.
11) Bind refcount leak in act_ipt, from Cong WANG.
12) Missing RCU locking in HTB scheduler, from Florian Westphal.
13) Several small MACSEC bug fixes from Sabrina Dubroca (missing RCU
barrier, using heap for SG and IV, and erroneous use of async flag
when allocating AEAD conext.)
14) RCU handling fix in TIPC, from Ying Xue.
15) Pass correct protocol down into ipv4_{update_pmtu,redirect}() in
SIT driver, from Simon Horman.
16) Socket timer deadlock fix in TIPC from Jon Paul Maloy.
17) Fix potential deadlock in team enslave, from Ido Schimmel.
18) Memory leak in KCM procfs handling, from Jiri Slaby.
19) ESN generation fix in ipv4 ESP, from Herbert Xu.
20) Fix GFP_KERNEL allocations with locks held in act_ife, from Cong
WANG.
21) Use after free in netem, from Eric Dumazet.
22) Uninitialized last assert time in multicast router code, from Tom
Goff.
23) Skip raw sockets in sock_diag destruction broadcast, from Willem
de Bruijn.
24) Fix link status reporting in thunderx, from Sunil Goutham.
25) Limit resegmentation of retransmit queue so that we do not
retransmit too large GSO frames. From Eric Dumazet.
26) Delay bpf program release after grace period, from Daniel
Borkmann"
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (141 commits)
openvswitch: fix conntrack netlink event delivery
qed: Protect the doorbell BAR with the write barriers.
neigh: Explicitly declare RCU-bh read side critical section in neigh_xmit()
e1000e: keep VLAN interfaces functional after rxvlan off
cfg80211: fix proto in ieee80211_data_to_8023 for frames without LLC header
qlcnic: use the correct ring in qlcnic_83xx_process_rcv_ring_diag()
bpf, perf: delay release of BPF prog after grace period
net: bridge: fix vlan stats continue counter
tcp: do not send too big packets at retransmit time
ibmvnic: fix to use list_for_each_safe() when delete items
net: thunderx: Fix TL4 configuration for secondary Qsets
net: thunderx: Fix link status reporting
net/mlx5e: Reorganize ethtool statistics
net/mlx5e: Fix number of PFC counters reported to ethtool
net/mlx5e: Prevent adding the same vxlan port
net/mlx5e: Check for BlueFlame capability before allocating SQ uar
net/mlx5e: Change enum to better reflect usage
net/mlx5: Add ConnectX-5 PCIe 4.0 to list of supported devices
net/mlx5: Update command strings
net: marvell: Add separate config ANEG function for Marvell 88E1111
...
Pull cgroup fixes from Tejun Heo:
"Three fix patches. Two are for cgroup / css init failure path. The
last one makes css_set_lock irq-safe as the deadline scheduler ends up
calling put_css_set() from irq context"
* 'for-4.7-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup:
cgroup: Disable IRQs while holding css_set_lock
cgroup: set css->id to -1 during init
cgroup: remove redundant cleanup in css_create
Commit dead9f29dd ("perf: Fix race in BPF program unregister") moved
destruction of BPF program from free_event_rcu() callback to __free_event(),
which is problematic if used with tail calls: if prog A is attached as
trace event directly, but at the same time present in a tail call map used
by another trace event program elsewhere, then we need to delay destruction
via RCU grace period since it can still be in use by the program doing the
tail call (the prog first needs to be dropped from the tail call map, then
trace event with prog A attached destroyed, so we get immediate destruction).
Fixes: dead9f29dd ("perf: Fix race in BPF program unregister")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Cc: Jann Horn <jann@thejh.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
The only users of audit_get_tty and audit_put_tty are internal to
audit, so move it out of include/linux/audit.h to kernel.h and create
a proper function rather than inlining it. This also reduces kABI
changes.
Suggested-by: Paul Moore <pmoore@redhat.com>
Signed-off-by: Richard Guy Briggs <rgb@redhat.com>
[PM: line wrapped description]
Signed-off-by: Paul Moore <paul@paul-moore.com>
Move the calculations of values after the allocation in case the
allocation fails. This avoids wasting effort in the rare case that it
fails, but more importantly saves us extra logic to release the tty
ref.
Signed-off-by: Richard Guy Briggs <rgb@redhat.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
One additional 'rule' for using update_cfs_rq_load_avg() is that one
should call update_tg_load_avg() if it returns true.
Add a bunch of comments to hopefully clarify some of the rules:
o You need to update cfs_rq _before_ any entity attach/detach,
this is important, because while for mathmatical consisency this
isn't strictly needed, it is required for the physical
interpretation of the model, you attach/detach _now_.
o When you modify the cfs_rq avg, you have to then call
update_tg_load_avg() in order to propagate changes upwards.
o (Fair) entities are always attached, switched_{to,from}_fair()
deal with !fair. This directly follows from the definition of the
cfs_rq averages, namely that they are a direct sum of all
(runnable or blocked) entities on that rq.
It is the second rule that this patch enforces, but it adds comments
pertaining to all of them.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Vincent and Yuyang found another few scenarios in which entity
tracking goes wobbly.
The scenarios are basically due to the fact that new tasks are not
immediately attached and thereby differ from the normal situation -- a
task is always attached to a cfs_rq load average (such that it
includes its blocked contribution) and are explicitly
detached/attached on migration to another cfs_rq.
Scenario 1: switch to fair class
p->sched_class = fair_class;
if (queued)
enqueue_task(p);
...
enqueue_entity()
enqueue_entity_load_avg()
migrated = !sa->last_update_time (true)
if (migrated)
attach_entity_load_avg()
check_class_changed()
switched_from() (!fair)
switched_to() (fair)
switched_to_fair()
attach_entity_load_avg()
If @p is a new task that hasn't been fair before, it will have
!last_update_time and, per the above, end up in
attach_entity_load_avg() _twice_.
Scenario 2: change between cgroups
sched_move_group(p)
if (queued)
dequeue_task()
task_move_group_fair()
detach_task_cfs_rq()
detach_entity_load_avg()
set_task_rq()
attach_task_cfs_rq()
attach_entity_load_avg()
if (queued)
enqueue_task();
...
enqueue_entity()
enqueue_entity_load_avg()
migrated = !sa->last_update_time (true)
if (migrated)
attach_entity_load_avg()
Similar as with scenario 1, if @p is a new task, it will have
!load_update_time and we'll end up in attach_entity_load_avg()
_twice_.
Furthermore, notice how we do a detach_entity_load_avg() on something
that wasn't attached to begin with.
As stated above; the problem is that the new task isn't yet attached
to the load tracking and thereby violates the invariant assumption.
This patch remedies this by ensuring a new task is indeed properly
attached to the load tracking on creation, through
post_init_entity_util_avg().
Of course, this isn't entirely as straightforward as one might think,
since the task is hashed before we call wake_up_new_task() and thus
can be poked at. We avoid this by adding TASK_NEW and teaching
cpu_cgroup_can_attach() to refuse such tasks.
Reported-by: Yuyang Du <yuyang.du@intel.com>
Reported-by: Vincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
A new fair task is detached and attached from/to task_group with:
cgroup_post_fork()
ss->fork(child) := cpu_cgroup_fork()
sched_move_task()
task_move_group_fair()
Which is wrong, because at this point in fork() the task isn't fully
initialized and it cannot 'move' to another group, because its not
attached to any group as yet.
In fact, cpu_cgroup_fork() needs a small part of sched_move_task() so we
can just call this small part directly instead sched_move_task(). And
the task doesn't really migrate because it is not yet attached so we
need the following sequence:
do_fork()
sched_fork()
__set_task_cpu()
cgroup_post_fork()
set_task_rq() # set task group and runqueue
wake_up_new_task()
select_task_rq() can select a new cpu
__set_task_cpu
post_init_entity_util_avg
attach_task_cfs_rq()
activate_task
enqueue_task
This patch makes that happen.
Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org>
[ Added TASK_SET_GROUP to set depth properly. ]
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Vincent reported that when a new task is moved into a new cgroup it
gets attached twice to the load tracking:
sched_move_task()
task_move_group_fair()
detach_task_cfs_rq()
set_task_rq()
attach_task_cfs_rq()
attach_entity_load_avg()
se->avg.last_load_update = cfs_rq->avg.last_load_update // == 0
enqueue_entity()
enqueue_entity_load_avg()
update_cfs_rq_load_avg()
now = clock()
__update_load_avg(&cfs_rq->avg)
cfs_rq->avg.last_load_update = now
// ages load/util for: now - 0, load/util -> 0
if (migrated)
attach_entity_load_avg()
se->avg.last_load_update = cfs_rq->avg.last_load_update; // now != 0
The problem is that we don't update cfs_rq load_avg before all
entity attach/detach operations. Only enqueue_task() and migrate_task()
do this.
By fixing this, the above will not happen, because the
sched_move_task() attach will have updated cfs_rq's last_load_update
time before attach, and in turn the attach will have set the entity's
last_load_update stamp.
Note that there is a further problem with sched_move_task() calling
detach on a task that hasn't yet been attached; this will be taken
care of in a subsequent patch.
Reported-by: Vincent Guittot <vincent.guittot@linaro.org>
Tested-by: Vincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Yuyang Du <yuyang.du@intel.com>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The task_fork_fair() callback already calls __set_task_cpu() and takes
rq->lock.
If we move the sched_class::task_fork callback in sched_fork() under
the existing p->pi_lock, right after its set_task_cpu() call, we can
avoid doing two such calls and omit the IRQ disabling on the rq->lock.
Change to __set_task_cpu() to skip the migration bits, this is a new
task, not a migration. Similarly, make wake_up_new_task() use
__set_task_cpu() for the same reason, the task hasn't actually
migrated as it hasn't ever ran.
This cures the problem of calling migrate_task_rq_fair(), which does
remove_entity_from_load_avg() on tasks that have never been added to
the load avg to begin with.
This bug would result in transiently messed up load_avg values, averaged
out after a few dozen milliseconds. This is probably the reason why
this bug was not found for such a long time.
Reported-by: Vincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Commit:
fde7d22e01 ("sched/fair: Fix overly small weight for interactive group entities")
did something non-obvious but also did it buggy yet latent.
The problem was exposed for real by a later commit in the v4.7 merge window:
2159197d66 ("sched/core: Enable increased load resolution on 64-bit kernels")
... after which tg->load_avg and cfs_rq->load.weight had different
units (10 bit fixed point and 20 bit fixed point resp.).
Add a comment to explain the use of cfs_rq->load.weight over the
'natural' cfs_rq->avg.load_avg and add scale_load_down() to correct
for the difference in unit.
Since this is (now, as per a previous commit) the only user of
calc_tg_weight(), collapse it.
The effects of this bug should be randomly inconsistent SMP-balancing
of cgroups workloads.
Reported-by: Jirka Hladky <jhladky@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Fixes: 2159197d66 ("sched/core: Enable increased load resolution on 64-bit kernels")
Fixes: fde7d22e01 ("sched/fair: Fix overly small weight for interactive group entities")
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Starting with the following commit:
fde7d22e01 ("sched/fair: Fix overly small weight for interactive group entities")
calc_tg_weight() doesn't compute the right value as expected by effective_load().
The difference is in the 'correction' term. In order to ensure \Sum
rw_j >= rw_i we cannot use tg->load_avg directly, since that might be
lagging a correction on the current cfs_rq->avg.load_avg value.
Therefore we use tg->load_avg - cfs_rq->tg_load_avg_contrib +
cfs_rq->avg.load_avg.
Now, per the referenced commit, calc_tg_weight() doesn't use
cfs_rq->avg.load_avg, as is later used in @w, but uses
cfs_rq->load.weight instead.
So stop using calc_tg_weight() and do it explicitly.
The effects of this bug are wake_affine() making randomly
poor choices in cgroup-intense workloads.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: <stable@vger.kernel.org> # v4.3+
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Fixes: fde7d22e01 ("sched/fair: Fix overly small weight for interactive group entities")
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull scheduler fixes from Thomas Gleixner:
"A couple of scheduler fixes:
- force watchdog reset while processing sysrq-w
- fix a deadlock when enabling trace events in the scheduler
- fixes to the throttled next buddy logic
- fixes for the average accounting (missing serialization and
underflow handling)
- allow kernel threads for fallback to online but not active cpus"
* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
sched/core: Allow kthreads to fall back to online && !active cpus
sched/fair: Do not announce throttled next buddy in dequeue_task_fair()
sched/fair: Initialize throttle_count for new task-groups lazily
sched/fair: Fix cfs_rq avg tracking underflow
kernel/sysrq, watchdog, sched/core: Reset watchdog on all CPUs while processing sysrq-w
sched/debug: Fix deadlock when enabling sched events
sched/fair: Fix post_init_entity_util_avg() serialization
Pull locking fix from Thomas Gleixner:
"A single fix to address a race in the static key logic"
* 'locking-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
locking/static_key: Fix concurrent static_key_slow_inc()
Commit b235beea9e ("Clarify naming of thread info/stack allocators")
breaks the build on some powerpc configs, where THREAD_SIZE < PAGE_SIZE:
kernel/fork.c:235:2: error: implicit declaration of function 'free_thread_stack'
kernel/fork.c:355:8: error: assignment from incompatible pointer type
stack = alloc_thread_stack_node(tsk, node);
^
Fix it by renaming free_stack() to free_thread_stack(), and updating the
return type of alloc_thread_stack_node().
Fixes: b235beea9e ("Clarify naming of thread info/stack allocators")
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Thomas Gleixner