select_idle_siblings() is a known pain point for a number of
workloads; it either does too much or not enough and sometimes just
does plain wrong.
This rewrite attempts to address a number of issues (but sadly not
all).
The current code does an unconditional sched_domain iteration; with
the intent of finding an idle core (on SMT hardware). The problems
which this patch tries to address are:
- its pointless to look for idle cores if the machine is real busy;
at which point you're just wasting cycles.
- it's behaviour is inconsistent between SMT and !SMT hardware in
that !SMT hardware ends up doing a scan for any idle CPU in the LLC
domain, while SMT hardware does a scan for idle cores and if that
fails, falls back to a scan for idle threads on the 'target' core.
The new code replaces the sched_domain scan with 3 explicit scans:
1) search for an idle core in the LLC
2) search for an idle CPU in the LLC
3) search for an idle thread in the 'target' core
where 1 and 3 are conditional on SMT support and 1 and 2 have runtime
heuristics to skip the step.
Step 1) is conditional on sd_llc_shared->has_idle_cores; when a cpu
goes idle and sd_llc_shared->has_idle_cores is false, we scan all SMT
siblings of the CPU going idle. Similarly, we clear
sd_llc_shared->has_idle_cores when we fail to find an idle core.
Step 2) tracks the average cost of the scan and compares this to the
average idle time guestimate for the CPU doing the wakeup. There is a
significant fudge factor involved to deal with the variability of the
averages. Esp. hackbench was sensitive to this.
Step 3) is unconditional; we assume (also per step 1) that scanning
all SMT siblings in a core is 'cheap'.
With this; SMT systems gain step 2, which cures a few benchmarks --
notably one from Facebook.
One 'feature' of the sched_domain iteration, which we preserve in the
new code, is that it would start scanning from the 'target' CPU,
instead of scanning the cpumask in cpu id order. This avoids multiple
CPUs in the LLC scanning for idle to gang up and find the same CPU
quite as much. The down side is that tasks can end up hopping across
the LLC for no apparent reason.
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>
Since struct sched_domain is strictly per cpu; introduce a structure
that is shared between all 'identical' sched_domains.
Limit to SD_SHARE_PKG_RESOURCES domains for now, as we'll only use it
for shared cache state; if another use comes up later we can easily
relax this.
While the sched_group's are normally shared between CPUs, these are
not natural to use when we need some shared state on a domain level --
since that would require the domain to have a parent, which is not a
given.
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>
There is no point in doing a call_rcu() for each domain, only do a
callback for the root sched domain and clean up the entire set in one
go.
Also make the entire call chain be called destroy_sched_domain*() to
remove confusion with the free_sched_domains() call, which does an
entirely different thing.
Both cpu_attach_domain() callers of destroy_sched_domain() can live
without the call_rcu() because at those points the sched_domain hasn't
been published yet.
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>
Otherwise this logic only works if mode is "compatible" with another
exclusive waiter.
If some wq has both TASK_INTERRUPTIBLE and TASK_UNINTERRUPTIBLE waiters,
abort_exclusive_wait() won't wait an uninterruptible waiter.
The main user is __wait_on_bit_lock() and currently it is fine but only
because TASK_KILLABLE includes TASK_UNINTERRUPTIBLE and we do not have
lock_page_interruptible() yet.
Just use TASK_NORMAL and remove the "mode" arg from abort_exclusive_wait().
Yes, this means that (say) wake_up_interruptible() can wake up the non-
interruptible waiter(s), but I think this is fine. And in fact I think
that abort_exclusive_wait() must die, see the next change.
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Cc: Bart Van Assche <bvanassche@acm.org>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Neil Brown <neilb@suse.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/20160906140047.GA6157@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull cgroup fixes from Tejun Heo:
"Three late fixes for cgroup: Two cpuset ones, one trivial and the
other pretty obscure, and a cgroup core fix for a bug which impacts
cgroup v2 namespace users"
* 'for-4.8-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup:
cgroup: fix invalid controller enable rejections with cgroup namespace
cpuset: fix non static symbol warning
cpuset: handle race between CPU hotplug and cpuset_hotplug_work
Pull tracefs fixes from Steven Rostedt:
"Al Viro has been looking at the tracefs code, and has pointed out some
issues. This contains one fix by me and one by Al. I'm sure that
he'll come up with more but for now I tested these patches and they
don't appear to have any negative impact on tracing"
* tag 'trace-v4.8-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace:
fix memory leaks in tracing_buffers_splice_read()
tracing: Move mutex to protect against resetting of seq data
The iter->seq can be reset outside the protection of the mutex. So can
reading of user data. Move the mutex up to the beginning of the function.
Fixes: d7350c3f45 ("tracing/core: make the read callbacks reentrants")
Cc: stable@vger.kernel.org # 2.6.30+
Reported-by: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Pull perf fixes from Thomas Gleixner:
"Three fixlets for perf:
- add a missing NULL pointer check in the intel BTS driver
- make BTS an exclusive PMU because BTS can only handle one event at
a time
- ensure that exclusive events are limited to one PMU so that several
exclusive events can be scheduled on different PMU instances"
* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
perf/core: Limit matching exclusive events to one PMU
perf/x86/intel/bts: Make it an exclusive PMU
perf/x86/intel/bts: Make sure debug store is valid
Pull irq fixes from Thomas Gleixner:
"Three fixes for irq core and irq chip drivers:
- Do not set the irq type if type is NONE. Fixes a boot regression
on various SoCs
- Use the proper cpu for setting up the GIC target list. Discovered
by the cpumask debugging code.
- A rather large fix for the MIPS-GIC so per cpu local interrupts
work again. This was discovered late because the code falls back
to slower timers which use normal device interrupts"
* 'irq-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
irqchip/mips-gic: Fix local interrupts
irqchip/gicv3: Silence noisy DEBUG_PER_CPU_MAPS warning
genirq: Skip chained interrupt trigger setup if type is IRQ_TYPE_NONE
On the v2 hierarchy, "cgroup.subtree_control" rejects controller
enables if the cgroup has processes in it. The enforcement of this
logic assumes that the cgroup wouldn't have any css_sets associated
with it if there are no tasks in the cgroup, which is no longer true
since a79a908fd2 ("cgroup: introduce cgroup namespaces").
When a cgroup namespace is created, it pins the css_set of the
creating task to use it as the root css_set of the namespace. This
extra reference stays as long as the namespace is around and makes
"cgroup.subtree_control" think that the namespace root cgroup is not
empty even when it is and thus reject controller enables.
Fix it by making cgroup_subtree_control() walk and test emptiness of
each css_set instead of testing whether the list_head is empty.
While at it, update the comment of cgroup_task_count() to indicate
that the returned value may be higher than the number of tasks, which
has always been true due to temporary references and doesn't break
anything.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Evgeny Vereshchagin <evvers@ya.ru>
Cc: Serge E. Hallyn <serge.hallyn@ubuntu.com>
Cc: Aditya Kali <adityakali@google.com>
Cc: Eric W. Biederman <ebiederm@xmission.com>
Cc: stable@vger.kernel.org # v4.6+
Fixes: a79a908fd2 ("cgroup: introduce cgroup namespaces")
Link: https://github.com/systemd/systemd/pull/3589#issuecomment-249089541
SCHED_HRTICK feature is useful to preempt SCHED_FAIR tasks on-the-dot
(just when they would have exceeded their ideal_runtime).
It makes use of a per-CPU hrtimer resource and hence arming that
hrtimer should be based on total SCHED_FAIR tasks a CPU has across its
various cfs_rqs, rather than being based on number of tasks in a
particular cfs_rq (as implemented currently).
As a result, with current code, its possible for a running task (which
is the sole task in its cfs_rq) to be preempted much after its
ideal_runtime has elapsed, resulting in increased latency for tasks in
other cfs_rq on same CPU.
Fix this by arming sched hrtimer based on total number of SCHED_FAIR
tasks a CPU has across its various cfs_rqs.
Signed-off-by: Srivatsa Vaddagiri <vatsa@codeaurora.org>
Signed-off-by: Joonwoo Park <joonwoop@codeaurora.org>
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>
Link: http://lkml.kernel.org/r/1474075731-11550-1-git-send-email-joonwoop@codeaurora.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Fixes the following sparse warning:
kernel/cpuset.c:2088:6: warning:
symbol 'cpuset_fork' was not declared. Should it be static?
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
Pull scheduler fix from Ingo Molnar:
"A try_to_wake_up() memory ordering race fix causing a busy-loop in
ttwu()"
* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
sched/core: Fix a race between try_to_wake_up() and a woken up task
Pull perf fixes from Ingo Molnar:
"This contains:
- a set of fixes found by directed-random perf fuzzing efforts by
Vince Weaver, Alexander Shishkin and Peter Zijlstra
- a cqm driver crash fix
- an AMD uncore driver use after free fix"
* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
perf/x86/intel: Fix PEBSv3 record drain
perf/x86/intel/bts: Kill a silly warning
perf/x86/intel/bts: Fix BTS PMI detection
perf/x86/intel/bts: Fix confused ordering of PMU callbacks
perf/core: Fix aux_mmap_count vs aux_refcount order
perf/core: Fix a race between mmap_close() and set_output() of AUX events
perf/x86/amd/uncore: Prevent use after free
perf/x86/intel/cqm: Check cqm/mbm enabled state in event init
perf/core: Remove WARN from perf_event_read()
A discrepancy between cpu_online_mask and cpuset's effective_cpus
mask is inevitable during hotplug since cpuset defers updating of
effective_cpus mask using a workqueue, during which time nothing
prevents the system from more hotplug operations. For that reason
guarantee_online_cpus() walks up the cpuset hierarchy until it finds
an intersection under the assumption that top cpuset's effective_cpus
mask intersects with cpu_online_mask even with such a race occurring.
However a sequence of CPU hotplugs can open a time window, during which
none of the effective CPUs in the top cpuset intersect with
cpu_online_mask.
For example when there are 4 possible CPUs 0-3 and only CPU0 is online:
======================== ===========================
cpu_online_mask top_cpuset.effective_cpus
======================== ===========================
echo 1 > cpu2/online.
CPU hotplug notifier woke up hotplug work but not yet scheduled.
[0,2] [0]
echo 0 > cpu0/online.
The workqueue is still runnable.
[2] [0]
======================== ===========================
Now there is no intersection between cpu_online_mask and
top_cpuset.effective_cpus. Thus invoking sys_sched_setaffinity() at
this moment can cause following:
Unable to handle kernel NULL pointer dereference at virtual address 000000d0
------------[ cut here ]------------
Kernel BUG at ffffffc0001389b0 [verbose debug info unavailable]
Internal error: Oops - BUG: 96000005 [#1] PREEMPT SMP
Modules linked in:
CPU: 2 PID: 1420 Comm: taskset Tainted: G W 4.4.8+ #98
task: ffffffc06a5c4880 ti: ffffffc06e124000 task.ti: ffffffc06e124000
PC is at guarantee_online_cpus+0x2c/0x58
LR is at cpuset_cpus_allowed+0x4c/0x6c
<snip>
Process taskset (pid: 1420, stack limit = 0xffffffc06e124020)
Call trace:
[<ffffffc0001389b0>] guarantee_online_cpus+0x2c/0x58
[<ffffffc00013b208>] cpuset_cpus_allowed+0x4c/0x6c
[<ffffffc0000d61f0>] sched_setaffinity+0xc0/0x1ac
[<ffffffc0000d6374>] SyS_sched_setaffinity+0x98/0xac
[<ffffffc000085cb0>] el0_svc_naked+0x24/0x28
The top cpuset's effective_cpus are guaranteed to be identical to
cpu_online_mask eventually. Hence fall back to cpu_online_mask when
there is no intersection between top cpuset's effective_cpus and
cpu_online_mask.
Signed-off-by: Joonwoo Park <joonwoop@codeaurora.org>
Acked-by: Li Zefan <lizefan@huawei.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: cgroups@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Cc: <stable@vger.kernel.org> # 3.17+
Signed-off-by: Tejun Heo <tj@kernel.org>
Pull libnvdimm fixes from Dan Williams:
"nvdimm fixes for v4.8, two of them are tagged for -stable:
- Fix devm_memremap_pages() to use track_pfn_insert(). Otherwise,
DAX pmd mappings end up with an uncached pgprot, and unusable
performance for the device-dax interface. The device-dax interface
appeared in 4.7 so this is tagged for -stable.
- Fix a couple VM_BUG_ON() checks in the show_smaps() path to
understand DAX pmd entries. This fix is tagged for -stable.
- Fix a mis-merge of the nfit machine-check handler to flip the
polarity of an if() to match the final version of the patch that
Vishal sent for 4.8-rc1. Without this the nfit machine check
handler never detects / inserts new 'badblocks' entries which
applications use to identify lost portions of files.
- For test purposes, fix the nvdimm_clear_poison() path to operate on
legacy / simulated nvdimm memory ranges. Without this fix a test
can set badblocks, but never clear them on these ranges.
- Fix the range checking done by dax_dev_pmd_fault(). This is not
tagged for -stable since this problem is mitigated by specifying
aligned resources at device-dax setup time.
These patches have appeared in a next release over the past week. The
recent rebase you can see in the timestamps was to drop an invalid fix
as identified by the updated device-dax unit tests [1]. The -mm
touches have an ack from Andrew"
[1]: "[ndctl PATCH 0/3] device-dax test for recent kernel bugs"
https://lists.01.org/pipermail/linux-nvdimm/2016-September/006855.html
* 'libnvdimm-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm/nvdimm:
libnvdimm: allow legacy (e820) pmem region to clear bad blocks
nfit, mce: Fix SPA matching logic in MCE handler
mm: fix cache mode of dax pmd mappings
mm: fix show_smap() for zone_device-pmd ranges
dax: fix mapping size check
The order of accesses to ring buffer's aux_mmap_count and aux_refcount
has to be preserved across the users, namely perf_mmap_close() and
perf_aux_output_begin(), otherwise the inversion can result in the latter
holding the last reference to the aux buffer and subsequently free'ing
it in atomic context, triggering a warning.
> ------------[ cut here ]------------
> WARNING: CPU: 0 PID: 257 at kernel/events/ring_buffer.c:541 __rb_free_aux+0x11a/0x130
> CPU: 0 PID: 257 Comm: stopbug Not tainted 4.8.0-rc1+ #2596
> Call Trace:
> [<ffffffff810f3e0b>] __warn+0xcb/0xf0
> [<ffffffff810f3f3d>] warn_slowpath_null+0x1d/0x20
> [<ffffffff8121182a>] __rb_free_aux+0x11a/0x130
> [<ffffffff812127a8>] rb_free_aux+0x18/0x20
> [<ffffffff81212913>] perf_aux_output_begin+0x163/0x1e0
> [<ffffffff8100c33a>] bts_event_start+0x3a/0xd0
> [<ffffffff8100c42d>] bts_event_add+0x5d/0x80
> [<ffffffff81203646>] event_sched_in.isra.104+0xf6/0x2f0
> [<ffffffff8120652e>] group_sched_in+0x6e/0x190
> [<ffffffff8120694e>] ctx_sched_in+0x2fe/0x5f0
> [<ffffffff81206ca0>] perf_event_sched_in+0x60/0x80
> [<ffffffff81206d1b>] ctx_resched+0x5b/0x90
> [<ffffffff81207281>] __perf_event_enable+0x1e1/0x240
> [<ffffffff81200639>] event_function+0xa9/0x180
> [<ffffffff81202000>] ? perf_cgroup_attach+0x70/0x70
> [<ffffffff8120203f>] remote_function+0x3f/0x50
> [<ffffffff811971f3>] flush_smp_call_function_queue+0x83/0x150
> [<ffffffff81197bd3>] generic_smp_call_function_single_interrupt+0x13/0x60
> [<ffffffff810a6477>] smp_call_function_single_interrupt+0x27/0x40
> [<ffffffff81a26ea9>] call_function_single_interrupt+0x89/0x90
> [<ffffffff81120056>] finish_task_switch+0xa6/0x210
> [<ffffffff81120017>] ? finish_task_switch+0x67/0x210
> [<ffffffff81a1e83d>] __schedule+0x3dd/0xb50
> [<ffffffff81a1efe5>] schedule+0x35/0x80
> [<ffffffff81128031>] sys_sched_yield+0x61/0x70
> [<ffffffff81a25be5>] entry_SYSCALL_64_fastpath+0x18/0xa8
> ---[ end trace 6235f556f5ea83a9 ]---
This patch puts the checks in perf_aux_output_begin() in the same order
as that of perf_mmap_close().
Reported-by: Vince Weaver <vincent.weaver@maine.edu>
Signed-off-by: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Arnaldo Carvalho de Melo <acme@infradead.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: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: vince@deater.net
Link: http://lkml.kernel.org/r/20160906132353.19887-3-alexander.shishkin@linux.intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
In the mmap_close() path we need to stop all the AUX events that are
writing data to the AUX area that we are unmapping, before we can
safely free the pages. To determine if an event needs to be stopped,
we're comparing its ->rb against the one that's getting unmapped.
However, a SET_OUTPUT ioctl may turn up inside an AUX transaction
and swizzle event::rb to some other ring buffer, but the transaction
will keep writing data to the old ring buffer until the event gets
scheduled out. At this point, mmap_close() will skip over such an
event and will proceed to free the AUX area, while it's still being
used by this event, which will set off a warning in the mmap_close()
path and cause a memory corruption.
To avoid this, always stop an AUX event before its ->rb is updated;
this will release the (potentially) last reference on the AUX area
of the buffer. If the event gets restarted, its new ring buffer will
be used. If another SET_OUTPUT comes and switches it back to the
old ring buffer that's getting unmapped, it's also fine: this
ring buffer's aux_mmap_count will be zero and AUX transactions won't
start any more.
Reported-by: Vince Weaver <vincent.weaver@maine.edu>
Signed-off-by: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Arnaldo Carvalho de Melo <acme@infradead.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: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: vince@deater.net
Link: http://lkml.kernel.org/r/20160906132353.19887-2-alexander.shishkin@linux.intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
track_pfn_insert() in vmf_insert_pfn_pmd() is marking dax mappings as
uncacheable rendering them impractical for application usage. DAX-pte
mappings are cached and the goal of establishing DAX-pmd mappings is to
attain more performance, not dramatically less (3 orders of magnitude).
track_pfn_insert() relies on a previous call to reserve_memtype() to
establish the expected page_cache_mode for the range. While memremap()
arranges for reserve_memtype() to be called, devm_memremap_pages() does
not. So, teach track_pfn_insert() and untrack_pfn() how to handle
tracking without a vma, and arrange for devm_memremap_pages() to
establish the write-back-cache reservation in the memtype tree.
Cc: <stable@vger.kernel.org>
Cc: Matthew Wilcox <mawilcox@microsoft.com>
Cc: Ross Zwisler <ross.zwisler@linux.intel.com>
Cc: Nilesh Choudhury <nilesh.choudhury@oracle.com>
Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Reported-by: Toshi Kani <toshi.kani@hpe.com>
Reported-by: Kai Zhang <kai.ka.zhang@oracle.com>
Acked-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
of_clk_init() ends up calling into pm_qos_update_request() very early
during boot where irq is expected to stay disabled.
pm_qos_update_request() uses cancel_delayed_work_sync() which
correctly assumes that irq is enabled on invocation and
unconditionally disables and re-enables it.
Gate cancel_delayed_work_sync() invocation with kevented_up() to avoid
enabling irq unexpectedly during early boot.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-and-tested-by: Qiao Zhou <qiaozhou@asrmicro.com>
Link: http://lkml.kernel.org/r/d2501c4c-8e7b-bea3-1b01-000b36b5dfe9@asrmicro.com
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The dl task will be replenished after dl task timer fire and start a
new period. It will be enqueued and to re-evaluate its dependency on
the tick in order to restart it. However, if the CPU is hot-unplugged,
irq_work_queue will splash since the target CPU is offline.
As a result we get:
WARNING: CPU: 2 PID: 0 at kernel/irq_work.c:69 irq_work_queue_on+0xad/0xe0
Call Trace:
dump_stack+0x99/0xd0
__warn+0xd1/0xf0
warn_slowpath_null+0x1d/0x20
irq_work_queue_on+0xad/0xe0
tick_nohz_full_kick_cpu+0x44/0x50
tick_nohz_dep_set_cpu+0x74/0xb0
enqueue_task_dl+0x226/0x480
activate_task+0x5c/0xa0
dl_task_timer+0x19b/0x2c0
? push_dl_task.part.31+0x190/0x190
This can be triggered by hot-unplugging the full dynticks CPU which dl
task is running on.
We enqueue the dl task on the offline CPU, because we need to do
replenish for start_dl_timer(). So, as Juri pointed out, we would
need to do is calling replenish_dl_entity() directly, instead of
enqueue_task_dl(). pi_se shouldn't be a problem as the task shouldn't
be boosted if it was throttled.
This patch fixes it by avoiding the whole enqueue+dequeue+enqueue story, by
first migrating (set_task_cpu()) and then doing 1 enqueue.
Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Wanpeng Li <wanpeng.li@hotmail.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Juri Lelli <juri.lelli@arm.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Luca Abeni <luca.abeni@unitn.it>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/1472639264-3932-1-git-send-email-wanpeng.li@hotmail.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
