| Age | Commit message (Collapse) | Author |
|---|
|
This fixes an issue where old load on a cfs_rq is not properly decayed,
resulting in strange behavior where fairness can decrease drastically.
Real workloads with equally weighted control groups have ended up
getting a respective 99% and 1%(!!) of cpu time.
When an idle task is attached to a cfs_rq by attaching a pid to a cgroup,
the old load of the task is attached to the new cfs_rq and sched_entity by
attach_entity_cfs_rq. If the task is then moved to another cpu (and
therefore cfs_rq) before being enqueued/woken up, the load will be moved
to cfs_rq->removed from the sched_entity. Such a move will happen when
enforcing a cpuset on the task (eg. via a cgroup) that force it to move.
The load will however not be removed from the task_group itself, making
it look like there is a constant load on that cfs_rq. This causes the
vruntime of tasks on other sibling cfs_rq's to increase faster than they
are supposed to; causing severe fairness issues. If no other task is
started on the given cfs_rq, and due to the cpuset it would not happen,
this load would never be properly unloaded. With this patch the load
will be properly removed inside update_blocked_averages. This also
applies to tasks moved to the fair scheduling class and moved to another
cpu, and this path will also fix that. For fork, the entity is queued
right away, so this problem does not affect that.
This applies to cases where the new process is the first in the cfs_rq,
issue introduced 3d30544f0212 ("sched/fair: Apply more PELT fixes"), and
when there has previously been load on the cgroup but the cgroup was
removed from the leaflist due to having null PELT load, indroduced
in 039ae8bcf7a5 ("sched/fair: Fix O(nr_cgroups) in the load balancing
path").
For a simple cgroup hierarchy (as seen below) with two equally weighted
groups, that in theory should get 50/50 of cpu time each, it often leads
to a load of 60/40 or 70/30.
parent/
cg-1/
cpu.weight: 100
cpuset.cpus: 1
cg-2/
cpu.weight: 100
cpuset.cpus: 1
If the hierarchy is deeper (as seen below), while keeping cg-1 and cg-2
equally weighted, they should still get a 50/50 balance of cpu time.
This however sometimes results in a balance of 10/90 or 1/99(!!) between
the task groups.
$ ps u -C stress
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
root 18568 1.1 0.0 3684 100 pts/12 R+ 13:36 0:00 stress --cpu 1
root 18580 99.3 0.0 3684 100 pts/12 R+ 13:36 0:09 stress --cpu 1
parent/
cg-1/
cpu.weight: 100
sub-group/
cpu.weight: 1
cpuset.cpus: 1
cg-2/
cpu.weight: 100
sub-group/
cpu.weight: 10000
cpuset.cpus: 1
This can be reproduced by attaching an idle process to a cgroup and
moving it to a given cpuset before it wakes up. The issue is evident in
many (if not most) container runtimes, and has been reproduced
with both crun and runc (and therefore docker and all its "derivatives"),
and with both cgroup v1 and v2.
Fixes: 3d30544f0212 ("sched/fair: Apply more PELT fixes")
Fixes: 039ae8bcf7a5 ("sched/fair: Fix O(nr_cgroups) in the load balancing path")
Signed-off-by: Odin Ugedal <odin@uged.al>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
Link: https://lkml.kernel.org/r/20210501141950.23622-2-odin@uged.al
|
|
Util-clamp places tasks in different buckets based on their clamp values
for performance reasons. However, the size of buckets is currently
computed using a rounding division, which can lead to an off-by-one
error in some configurations.
For instance, with 20 buckets, the bucket size will be 1024/20=51. A
task with a clamp of 1024 will be mapped to bucket id 1024/51=20. Sadly,
correct indexes are in range [0,19], hence leading to an out of bound
memory access.
Clamp the bucket id to fix the issue.
Fixes: 69842cba9ace ("sched/uclamp: Add CPU's clamp buckets refcounting")
Suggested-by: Qais Yousef <qais.yousef@arm.com>
Signed-off-by: Quentin Perret <qperret@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
Reviewed-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Link: https://lkml.kernel.org/r/20210430151412.160913-1-qperret@google.com
|
|
4117cebf1a9f ("psi: Optimize task switch inside shared cgroups")
introduced a race condition that corrupts internal psi state. This
manifests as kernel warnings, sometimes followed by bogusly high IO
pressure:
psi: task underflow! cpu=1 t=2 tasks=[0 0 0 0] clear=c set=0
(schedule() decreasing RUNNING and ONCPU, both of which are 0)
psi: incosistent task state! task=2412744:systemd cpu=17 psi_flags=e clear=3 set=0
(cgroup_move_task() clearing MEMSTALL and IOWAIT, but task is MEMSTALL | RUNNING | ONCPU)
What the offending commit does is batch the two psi callbacks in
schedule() to reduce the number of cgroup tree updates. When prev is
deactivated and removed from the runqueue, nothing is done in psi at
first; when the task switch completes, TSK_RUNNING and TSK_IOWAIT are
updated along with TSK_ONCPU.
However, the deactivation and the task switch inside schedule() aren't
atomic: pick_next_task() may drop the rq lock for load balancing. When
this happens, cgroup_move_task() can run after the task has been
physically dequeued, but the psi updates are still pending. Since it
looks at the task's scheduler state, it doesn't move everything to the
new cgroup that the task switch that follows is about to clear from
it. cgroup_move_task() will leak the TSK_RUNNING count in the old
cgroup, and psi_sched_switch() will underflow it in the new cgroup.
A similar thing can happen for iowait. TSK_IOWAIT is usually set when
a p->in_iowait task is dequeued, but again this update is deferred to
the switch. cgroup_move_task() can see an unqueued p->in_iowait task
and move a non-existent TSK_IOWAIT. This results in the inconsistent
task state warning, as well as a counter underflow that will result in
permanent IO ghost pressure being reported.
Fix this bug by making cgroup_move_task() use task->psi_flags instead
of looking at the potentially mismatching scheduler state.
[ We used the scheduler state historically in order to not rely on
task->psi_flags for anything but debugging. But that ship has sailed
anyway, and this is simpler and more robust.
We previously already batched TSK_ONCPU clearing with the
TSK_RUNNING update inside the deactivation call from schedule(). But
that ordering was safe and didn't result in TSK_ONCPU corruption:
unlike most places in the scheduler, cgroup_move_task() only checked
task_current() and handled TSK_ONCPU if the task was still queued. ]
Fixes: 4117cebf1a9f ("psi: Optimize task switch inside shared cgroups")
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20210503174917.38579-1-hannes@cmpxchg.org
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull scheduler updates from Ingo Molnar:
- Clean up SCHED_DEBUG: move the decades old mess of sysctl, procfs and
debugfs interfaces to a unified debugfs interface.
- Signals: Allow caching one sigqueue object per task, to improve
performance & latencies.
- Improve newidle_balance() irq-off latencies on systems with a large
number of CPU cgroups.
- Improve energy-aware scheduling
- Improve the PELT metrics for certain workloads
- Reintroduce select_idle_smt() to improve load-balancing locality -
but without the previous regressions
- Add 'scheduler latency debugging': warn after long periods of pending
need_resched. This is an opt-in feature that requires the enabling of
the LATENCY_WARN scheduler feature, or the use of the
resched_latency_warn_ms=xx boot parameter.
- CPU hotplug fixes for HP-rollback, and for the 'fail' interface. Fix
remaining balance_push() vs. hotplug holes/races
- PSI fixes, plus allow /proc/pressure/ files to be written by
CAP_SYS_RESOURCE tasks as well
- Fix/improve various load-balancing corner cases vs. capacity margins
- Fix sched topology on systems with NUMA diameter of 3 or above
- Fix PF_KTHREAD vs to_kthread() race
- Minor rseq optimizations
- Misc cleanups, optimizations, fixes and smaller updates
* tag 'sched-core-2021-04-28' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (61 commits)
cpumask/hotplug: Fix cpu_dying() state tracking
kthread: Fix PF_KTHREAD vs to_kthread() race
sched/debug: Fix cgroup_path[] serialization
sched,psi: Handle potential task count underflow bugs more gracefully
sched: Warn on long periods of pending need_resched
sched/fair: Move update_nohz_stats() to the CONFIG_NO_HZ_COMMON block to simplify the code & fix an unused function warning
sched/debug: Rename the sched_debug parameter to sched_verbose
sched,fair: Alternative sched_slice()
sched: Move /proc/sched_debug to debugfs
sched,debug: Convert sysctl sched_domains to debugfs
debugfs: Implement debugfs_create_str()
sched,preempt: Move preempt_dynamic to debug.c
sched: Move SCHED_DEBUG sysctl to debugfs
sched: Don't make LATENCYTOP select SCHED_DEBUG
sched: Remove sched_schedstats sysctl out from under SCHED_DEBUG
sched/numa: Allow runtime enabling/disabling of NUMA balance without SCHED_DEBUG
sched: Use cpu_dying() to fix balance_push vs hotplug-rollback
cpumask: Introduce DYING mask
cpumask: Make cpu_{online,possible,present,active}() inline
rseq: Optimise rseq_get_rseq_cs() and clear_rseq_cs()
...
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull locking updates from Ingo Molnar:
- rtmutex cleanup & spring cleaning pass that removes ~400 lines of
code
- Futex simplifications & cleanups
- Add debugging to the CSD code, to help track down a tenacious race
(or hw problem)
- Add lockdep_assert_not_held(), to allow code to require a lock to not
be held, and propagate this into the ath10k driver
- Misc LKMM documentation updates
- Misc KCSAN updates: cleanups & documentation updates
- Misc fixes and cleanups
- Fix locktorture bugs with ww_mutexes
* tag 'locking-core-2021-04-28' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (44 commits)
kcsan: Fix printk format string
static_call: Relax static_call_update() function argument type
static_call: Fix unused variable warn w/o MODULE
locking/rtmutex: Clean up signal handling in __rt_mutex_slowlock()
locking/rtmutex: Restrict the trylock WARN_ON() to debug
locking/rtmutex: Fix misleading comment in rt_mutex_postunlock()
locking/rtmutex: Consolidate the fast/slowpath invocation
locking/rtmutex: Make text section and inlining consistent
locking/rtmutex: Move debug functions as inlines into common header
locking/rtmutex: Decrapify __rt_mutex_init()
locking/rtmutex: Remove pointless CONFIG_RT_MUTEXES=n stubs
locking/rtmutex: Inline chainwalk depth check
locking/rtmutex: Move rt_mutex_debug_task_free() to rtmutex.c
locking/rtmutex: Remove empty and unused debug stubs
locking/rtmutex: Consolidate rt_mutex_init()
locking/rtmutex: Remove output from deadlock detector
locking/rtmutex: Remove rtmutex deadlock tester leftovers
locking/rtmutex: Remove rt_mutex_timed_lock()
MAINTAINERS: Add myself as futex reviewer
locking/mutex: Remove repeated declaration
...
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm
Pull power management updates from Rafael Wysocki:
"These add some new hardware support (for example, IceLake-D idle
states in intel_idle), fix some issues (for example, the handling of
negative "sleep length" values in cpuidle governors), add new
functionality to the existing drivers (for example, scale-invariance
support in the ACPI CPPC cpufreq driver) and clean up code all over.
Specifics:
- Add idle states table for IceLake-D to the intel_idle driver and
update IceLake-X C6 data in it (Artem Bityutskiy).
- Fix the C7 idle state on Tegra114 in the tegra cpuidle driver and
drop the unused do_idle() firmware call from it (Dmitry Osipenko).
- Fix cpuidle-qcom-spm Kconfig entry (He Ying).
- Fix handling of possible negative tick_nohz_get_next_hrtimer()
return values of in cpuidle governors (Rafael Wysocki).
- Add support for frequency-invariance to the ACPI CPPC cpufreq
driver and update the frequency-invariance engine (FIE) to use it
as needed (Viresh Kumar).
- Simplify the default delay_us setting in the ACPI CPPC cpufreq
driver (Tom Saeger).
- Clean up frequency-related computations in the intel_pstate cpufreq
driver (Rafael Wysocki).
- Fix TBG parent setting for load levels in the armada-37xx cpufreq
driver and drop the CPU PM clock .set_parent method for armada-37xx
(Marek Behún).
- Fix multiple issues in the armada-37xx cpufreq driver (Pali Rohár).
- Fix handling of dev_pm_opp_of_cpumask_add_table() return values in
cpufreq-dt to take the -EPROBE_DEFER one into acconut as
appropriate (Quanyang Wang).
- Fix format string in ia64-acpi-cpufreq (Sergei Trofimovich).
- Drop the unused for_each_policy() macro from cpufreq (Shaokun
Zhang).
- Simplify computations in the schedutil cpufreq governor to avoid
unnecessary overhead (Yue Hu).
- Fix typos in the s5pv210 cpufreq driver (Bhaskar Chowdhury).
- Fix cpufreq documentation links in Kconfig (Alexander Monakov).
- Fix PCI device power state handling in pci_enable_device_flags() to
avoid issuse in some cases when the device depends on an ACPI power
resource (Rafael Wysocki).
- Add missing documentation of pm_runtime_resume_and_get() (Alan
Stern).
- Add missing static inline stub for pm_runtime_has_no_callbacks() to
pm_runtime.h and drop the unused try_to_freeze_nowarn() definition
(YueHaibing).
- Drop duplicate struct device declaration from pm.h and fix a
structure type declaration in intel_rapl.h (Wan Jiabing).
- Use dev_set_name() instead of an open-coded equivalent of it in the
wakeup sources code and drop a redundant local variable
initialization from it (Andy Shevchenko, Colin Ian King).
- Use crc32 instead of md5 for e820 memory map integrity check during
resume from hibernation on x86 (Chris von Recklinghausen).
- Fix typos in comments in the system-wide and hibernation support
code (Lu Jialin).
- Modify the generic power domains (genpd) code to avoid resuming
devices in the "prepare" phase of system-wide suspend and
hibernation (Ulf Hansson).
- Add Hygon Fam18h RAPL support to the intel_rapl power capping
driver (Pu Wen).
- Add MAINTAINERS entry for the dynamic thermal power management
(DTPM) code (Daniel Lezcano).
- Add devm variants of operating performance points (OPP) API
functions and switch over some users of the OPP framework to the
new resource-managed API (Yangtao Li and Dmitry Osipenko).
- Update devfreq core:
* Register devfreq devices as cooling devices on demand (Daniel
Lezcano).
* Add missing unlock opeation in devfreq_add_device() (Lukasz
Luba).
* Use the next frequency as resume_freq instead of the previous
frequency when using the opp-suspend property (Dong Aisheng).
* Check get_dev_status in devfreq_update_stats() (Dong Aisheng).
* Fix set_freq path for the userspace governor in Kconfig (Dong
Aisheng).
* Remove invalid description of get_target_freq() (Dong Aisheng).
- Update devfreq drivers:
* imx8m-ddrc: Remove imx8m_ddrc_get_dev_status() and unneeded
of_match_ptr() (Dong Aisheng, Fabio Estevam).
* rk3399_dmc: dt-bindings: Add rockchip,pmu phandle and drop
references to undefined symbols (Enric Balletbo i Serra, Gaël
PORTAY).
* rk3399_dmc: Use dev_err_probe() to simplify the code (Krzysztof
Kozlowski).
* imx-bus: Remove unneeded of_match_ptr() (Fabio Estevam).
- Fix kernel-doc warnings in three places (Pierre-Louis Bossart).
- Fix typo in the pm-graph utility code (Ricardo Ribalda)"
* tag 'pm-5.13-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm: (74 commits)
PM: wakeup: remove redundant assignment to variable retval
PM: hibernate: x86: Use crc32 instead of md5 for hibernation e820 integrity check
cpufreq: Kconfig: fix documentation links
PM: wakeup: use dev_set_name() directly
PM: runtime: Add documentation for pm_runtime_resume_and_get()
cpufreq: intel_pstate: Simplify intel_pstate_update_perf_limits()
cpufreq: armada-37xx: Fix module unloading
cpufreq: armada-37xx: Remove cur_frequency variable
cpufreq: armada-37xx: Fix determining base CPU frequency
cpufreq: armada-37xx: Fix driver cleanup when registration failed
clk: mvebu: armada-37xx-periph: Fix workaround for switching from L1 to L0
clk: mvebu: armada-37xx-periph: Fix switching CPU freq from 250 Mhz to 1 GHz
cpufreq: armada-37xx: Fix the AVS value for load L1
clk: mvebu: armada-37xx-periph: remove .set_parent method for CPU PM clock
cpufreq: armada-37xx: Fix setting TBG parent for load levels
cpuidle: Fix ARM_QCOM_SPM_CPUIDLE configuration
cpuidle: tegra: Remove do_idle firmware call
cpuidle: tegra: Fix C7 idling state on Tegra114
PM: sleep: fix typos in comments
cpufreq: Remove unused for_each_policy macro
...
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull irq updates from Thomas Gleixner:
"The usual updates from the irq departement:
Core changes:
- Provide IRQF_NO_AUTOEN as a flag for request*_irq() so drivers can
be cleaned up which either use a seperate mechanism to prevent
auto-enable at request time or have a racy mechanism which disables
the interrupt right after request.
- Get rid of the last usage of irq_create_identity_mapping() and
remove the interface.
- An overhaul of tasklet_disable().
Most usage sites of tasklet_disable() are in task context and
usually in cleanup, teardown code pathes. tasklet_disable()
spinwaits for a tasklet which is currently executed. That's not
only a problem for PREEMPT_RT where this can lead to a live lock
when the disabling task preempts the softirq thread. It's also
problematic in context of virtualization when the vCPU which runs
the tasklet is scheduled out and the disabling code has to spin
wait until it's scheduled back in.
There are a few code pathes which invoke tasklet_disable() from
non-sleepable context. For these a new disable variant which still
spinwaits is provided which allows to switch tasklet_disable() to a
sleep wait mechanism. For the atomic use cases this does not solve
the live lock issue on PREEMPT_RT. That is mitigated by blocking on
the RT specific softirq lock.
- The PREEMPT_RT specific implementation of softirq processing and
local_bh_disable/enable().
On RT enabled kernels soft interrupt processing happens always in
task context and all interrupt handlers, which are not explicitly
marked to be invoked in hard interrupt context are forced into task
context as well. This allows to protect against softirq processing
with a per CPU lock, which in turn allows to make BH disabled
regions preemptible.
Most of the softirq handling code is still shared. The RT/non-RT
specific differences are addressed with a set of inline functions
which provide the context specific functionality. The
local_bh_disable() / local_bh_enable() mechanism are obviously
seperate.
- The usual set of small improvements and cleanups
Driver changes:
- New drivers for Nuvoton WPCM450 and DT 79rc3243x interrupt
controllers
- Extended functionality for MStar, STM32 and SC7280 irq chips
- Enhanced robustness for ARM GICv3/4.1 drivers
- The usual set of cleanups and improvements all over the place"
* tag 'irq-core-2021-04-26' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (53 commits)
irqchip/xilinx: Expose Kconfig option for Zynq/ZynqMP
irqchip/gic-v3: Do not enable irqs when handling spurious interrups
dt-bindings: interrupt-controller: Add IDT 79RC3243x Interrupt Controller
irqchip: Add support for IDT 79rc3243x interrupt controller
irqdomain: Drop references to recusive irqdomain setup
irqdomain: Get rid of irq_create_strict_mappings()
irqchip/jcore-aic: Kill use of irq_create_strict_mappings()
ARM: PXA: Kill use of irq_create_strict_mappings()
irqchip/gic-v4.1: Disable vSGI upon (GIC CPUIF < v4.1) detection
irqchip/tb10x: Use 'fallthrough' to eliminate a warning
genirq: Reduce irqdebug cacheline bouncing
kernel: Initialize cpumask before parsing
irqchip/wpcm450: Drop COMPILE_TEST
irqchip/irq-mst: Support polarity configuration
irqchip: Add driver for WPCM450 interrupt controller
dt-bindings: interrupt-controller: Add nuvoton, wpcm450-aic
dt-bindings: qcom,pdc: Add compatible for sc7280
irqchip/stm32: Add usart instances exti direct event support
irqchip/gic-v3: Fix OF_BAD_ADDR error handling
irqchip/sifive-plic: Mark two global variables __ro_after_init
...
|
|
The kthread_is_per_cpu() construct relies on only being called on
PF_KTHREAD tasks (per the WARN in to_kthread). This gives rise to the
following usage pattern:
if ((p->flags & PF_KTHREAD) && kthread_is_per_cpu(p))
However, as reported by syzcaller, this is broken. The scenario is:
CPU0 CPU1 (running p)
(p->flags & PF_KTHREAD) // true
begin_new_exec()
me->flags &= ~(PF_KTHREAD|...);
kthread_is_per_cpu(p)
to_kthread(p)
WARN(!(p->flags & PF_KTHREAD) <-- *SPLAT*
Introduce __to_kthread() that omits the WARN and is sure to check both
values.
Use this to remove the problematic pattern for kthread_is_per_cpu()
and fix a number of other kthread_*() functions that have similar
issues but are currently not used in ways that would expose the
problem.
Notably kthread_func() is only ever called on 'current', while
kthread_probe_data() is only used for PF_WQ_WORKER, which implies the
task is from kthread_create*().
Fixes: ac687e6e8c26 ("kthread: Extract KTHREAD_IS_PER_CPU")
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Valentin Schneider <Valentin.Schneider@arm.com>
Link: https://lkml.kernel.org/r/YH6WJc825C4P0FCK@hirez.programming.kicks-ass.net
|
|
The handling of sysrq key can be activated by echoing the key to
/proc/sysrq-trigger or via the magic key sequence typed into a terminal
that is connected to the system in some way (serial, USB or other mean).
In the former case, the handling is done in a user context. In the
latter case, it is likely to be in an interrupt context.
Currently in print_cpu() of kernel/sched/debug.c, sched_debug_lock is
taken with interrupt disabled for the whole duration of the calls to
print_*_stats() and print_rq() which could last for the quite some time
if the information dump happens on the serial console.
If the system has many cpus and the sched_debug_lock is somehow busy
(e.g. parallel sysrq-t), the system may hit a hard lockup panic
depending on the actually serial console implementation of the
system.
The purpose of sched_debug_lock is to serialize the use of the global
cgroup_path[] buffer in print_cpu(). The rests of the printk calls don't
need serialization from sched_debug_lock.
Calling printk() with interrupt disabled can still be problematic if
multiple instances are running. Allocating a stack buffer of PATH_MAX
bytes is not feasible because of the limited size of the kernel stack.
The solution implemented in this patch is to allow only one caller at a
time to use the full size group_path[], while other simultaneous callers
will have to use shorter stack buffers with the possibility of path
name truncation. A "..." suffix will be printed if truncation may have
happened. The cgroup path name is provided for informational purpose
only, so occasional path name truncation should not be a big problem.
Fixes: efe25c2c7b3a ("sched: Reinstate group names in /proc/sched_debug")
Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Waiman Long <longman@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20210415195426.6677-1-longman@redhat.com
|
|
psi_group_cpu->tasks, represented by the unsigned int, stores the
number of tasks that could be stalled on a psi resource(io/mem/cpu).
Decrementing these counters at zero leads to wrapping which further
leads to the psi_group_cpu->state_mask is being set with the
respective pressure state. This could result into the unnecessary time
sampling for the pressure state thus cause the spurious psi events.
This can further lead to wrong actions being taken at the user land
based on these psi events.
Though psi_bug is set under these conditions but that just for debug
purpose. Fix it by decrementing the ->tasks count only when it is
non-zero.
Signed-off-by: Charan Teja Reddy <charante@codeaurora.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Link: https://lkml.kernel.org/r/1618585336-37219-1-git-send-email-charante@codeaurora.org
|
|
CPU scheduler marks need_resched flag to signal a schedule() on a
particular CPU. But, schedule() may not happen immediately in cases
where the current task is executing in the kernel mode (no
preemption state) for extended periods of time.
This patch adds a warn_on if need_resched is pending for more than the
time specified in sysctl resched_latency_warn_ms. If it goes off, it is
likely that there is a missing cond_resched() somewhere. Monitoring is
done via the tick and the accuracy is hence limited to jiffy scale. This
also means that we won't trigger the warning if the tick is disabled.
This feature (LATENCY_WARN) is default disabled.
Signed-off-by: Paul Turner <pjt@google.com>
Signed-off-by: Josh Don <joshdon@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20210416212936.390566-1-joshdon@google.com
|
|
simplify the code & fix an unused function warning
When !CONFIG_NO_HZ_COMMON we get this new GCC warning:
kernel/sched/fair.c:8398:13: warning: ‘update_nohz_stats’ defined but not used [-Wunused-function]
Move update_nohz_stats() to an already existing CONFIG_NO_HZ_COMMON #ifdef
block.
Beyond fixing the GCC warning, this also simplifies the update_nohz_stats() function.
[ mingo: Rewrote the changelog. ]
Fixes: 0826530de3cb ("sched/fair: Remove update of blocked load from newidle_balance")
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
Link: https://lore.kernel.org/r/20210329144029.29200-1-yuehaibing@huawei.com
|
|
CONFIG_SCHED_DEBUG is the build-time Kconfig knob, the boot param
sched_debug and the /debug/sched/debug_enabled knobs control the
sched_debug_enabled variable, but what they really do is make
SCHED_DEBUG more verbose, so rename the lot.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
|
|
The current sched_slice() seems to have issues; there's two possible
things that could be improved:
- the 'nr_running' used for __sched_period() is daft when cgroups are
considered. Using the RQ wide h_nr_running seems like a much more
consistent number.
- (esp) cgroups can slice it real fine, which makes for easy
over-scheduling, ensure min_gran is what the name says.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Valentin Schneider <valentin.schneider@arm.com>
Link: https://lkml.kernel.org/r/20210412102001.611897312@infradead.org
|
|
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Tested-by: Valentin Schneider <valentin.schneider@arm.com>
Link: https://lkml.kernel.org/r/20210412102001.548833671@infradead.org
|
|
Stop polluting sysctl, move to debugfs for SCHED_DEBUG stuff.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Reviewed-by: Valentin Schneider <valentin.schneider@arm.com>
Tested-by: Valentin Schneider <valentin.schneider@arm.com>
Link: https://lkml.kernel.org/r/YHgB/s4KCBQ1ifdm@hirez.programming.kicks-ass.net
|
|
Move the #ifdef SCHED_DEBUG bits to kernel/sched/debug.c in order to
collect all the debugfs bits.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Valentin Schneider <valentin.schneider@arm.com>
Link: https://lkml.kernel.org/r/20210412102001.353833279@infradead.org
|
|
Stop polluting sysctl with undocumented knobs that really are debug
only, move them all to /debug/sched/ along with the existing
/debug/sched_* files that already exist.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Tested-by: Valentin Schneider <valentin.schneider@arm.com>
Link: https://lkml.kernel.org/r/20210412102001.287610138@infradead.org
|
|
Use the new cpu_dying() state to simplify and fix the balance_push()
vs CPU hotplug rollback state.
Specifically, we currently rely on notifiers sched_cpu_dying() /
sched_cpu_activate() to terminate balance_push, however if the
cpu_down() fails when we're past sched_cpu_deactivate(), it should
terminate balance_push at that point and not wait until we hit
sched_cpu_activate().
Similarly, when cpu_up() fails and we're going back down, balance_push
should be active, where it currently is not.
So instead, make sure balance_push is enabled below SCHED_AP_ACTIVE
(when !cpu_active()), and gate it's utility with cpu_dying().
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Valentin Schneider <valentin.schneider@arm.com>
Link: https://lkml.kernel.org/r/YHgAYef83VQhKdC2@hirez.programming.kicks-ass.net
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/vireshk/pm
Pull ARM cpufreq updates for v5.13 from Viresh Kumar:
"- Fix typos in s5pv210 cpufreq driver (Bhaskar Chowdhury).
- Armada 37xx: Fix cpufreq changing base CPU speed to 800 MHz from
1000 MHz (Pali Rohár and Marek Behún).
- cpufreq-dt: Return -EPROBE_DEFER on failure to add table (Quanyang
Wang).
- Minor cleanup in cppc driver (Tom Saeger).
- Add frequency invariance support for CPPC driver and generalize
freq invariance support arch-topology driver (Viresh Kumar)."
* 'cpufreq/arm/linux-next' of git://git.kernel.org/pub/scm/linux/kernel/git/vireshk/pm:
cpufreq: armada-37xx: Fix module unloading
cpufreq: armada-37xx: Remove cur_frequency variable
cpufreq: armada-37xx: Fix determining base CPU frequency
cpufreq: armada-37xx: Fix driver cleanup when registration failed
clk: mvebu: armada-37xx-periph: Fix workaround for switching from L1 to L0
clk: mvebu: armada-37xx-periph: Fix switching CPU freq from 250 Mhz to 1 GHz
cpufreq: armada-37xx: Fix the AVS value for load L1
clk: mvebu: armada-37xx-periph: remove .set_parent method for CPU PM clock
cpufreq: armada-37xx: Fix setting TBG parent for load levels
cpufreq: dt: dev_pm_opp_of_cpumask_add_table() may return -EPROBE_DEFER
cpufreq: cppc: simplify default delay_us setting
cpufreq: Rudimentary typos fix in the file s5pv210-cpufreq.c
cpufreq: CPPC: Add support for frequency invariance
arch_topology: Export arch_freq_scale and helpers
arch_topology: Allow multiple entities to provide sched_freq_tick() callback
arch_topology: Rename freq_scale as arch_freq_scale
|
|
During load-balance, groups classified as group_misfit_task are filtered
out if they do not pass
group_smaller_max_cpu_capacity(<candidate group>, <local group>);
which itself employs fits_capacity() to compare the sgc->max_capacity of
both groups.
Due to the underlying margin, fits_capacity(X, 1024) will return false for
any X > 819. Tough luck, the capacity_orig's on e.g. the Pixel 4 are
{261, 871, 1024}. If a CPU-bound task ends up on one of those "medium"
CPUs, misfit migration will never intentionally upmigrate it to a CPU of
higher capacity due to the aforementioned margin.
One may argue the 20% margin of fits_capacity() is excessive in the advent
of counter-enhanced load tracking (APERF/MPERF, AMUs), but one point here
is that fits_capacity() is meant to compare a utilization value to a
capacity value, whereas here it is being used to compare two capacity
values. As CPU capacity and task utilization have different dynamics, a
sensible approach here would be to add a new helper dedicated to comparing
CPU capacities.
Also note that comparing capacity extrema of local and source sched_group's
doesn't make much sense when at the day of the day the imbalance will be
pulled by a known env->dst_cpu, whose capacity can be anywhere within the
local group's capacity extrema.
While at it, replace group_smaller_{min, max}_cpu_capacity() with
comparisons of the source group's min/max capacity and the destination
CPU's capacity.
Signed-off-by: Valentin Schneider <valentin.schneider@arm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Reviewed-by: Qais Yousef <qais.yousef@arm.com>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
Tested-by: Lingutla Chandrasekhar <clingutla@codeaurora.org>
Link: https://lkml.kernel.org/r/20210407220628.3798191-4-valentin.schneider@arm.com
|
|
When triggering an active load balance, sd->nr_balance_failed is set to
such a value that any further can_migrate_task() using said sd will ignore
the output of task_hot().
This behaviour makes sense, as active load balance intentionally preempts a
rq's running task to migrate it right away, but this asynchronous write is
a bit shoddy, as the stopper thread might run active_load_balance_cpu_stop
before the sd->nr_balance_failed write either becomes visible to the
stopper's CPU or even happens on the CPU that appended the stopper work.
Add a struct lb_env flag to denote active balancing, and use it in
can_migrate_task(). Remove the sd->nr_balance_failed write that served the
same purpose. Cleanup the LBF_DST_PINNED active balance special case.
Signed-off-by: Valentin Schneider <valentin.schneider@arm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
Link: https://lkml.kernel.org/r/20210407220628.3798191-3-valentin.schneider@arm.com
|
|
During load balance, LBF_SOME_PINNED will be set if any candidate task
cannot be detached due to CPU affinity constraints. This can result in
setting env->sd->parent->sgc->group_imbalance, which can lead to a group
being classified as group_imbalanced (rather than any of the other, lower
group_type) when balancing at a higher level.
In workloads involving a single task per CPU, LBF_SOME_PINNED can often be
set due to per-CPU kthreads being the only other runnable tasks on any
given rq. This results in changing the group classification during
load-balance at higher levels when in reality there is nothing that can be
done for this affinity constraint: per-CPU kthreads, as the name implies,
don't get to move around (modulo hotplug shenanigans).
It's not as clear for userspace tasks - a task could be in an N-CPU cpuset
with N-1 offline CPUs, making it an "accidental" per-CPU task rather than
an intended one. KTHREAD_IS_PER_CPU gives us an indisputable signal which
we can leverage here to not set LBF_SOME_PINNED.
Note that the aforementioned classification to group_imbalance (when
nothing can be done) is especially problematic on big.LITTLE systems, which
have a topology the likes of:
DIE [ ]
MC [ ][ ]
0 1 2 3
L L B B
arch_scale_cpu_capacity(L) < arch_scale_cpu_capacity(B)
Here, setting LBF_SOME_PINNED due to a per-CPU kthread when balancing at MC
level on CPUs [0-1] will subsequently prevent CPUs [2-3] from classifying
the [0-1] group as group_misfit_task when balancing at DIE level. Thus, if
CPUs [0-1] are running CPU-bound (misfit) tasks, ill-timed per-CPU kthreads
can significantly delay the upgmigration of said misfit tasks. Systems
relying on ASYM_PACKING are likely to face similar issues.
Signed-off-by: Lingutla Chandrasekhar <clingutla@codeaurora.org>
[Use kthread_is_per_cpu() rather than p->nr_cpus_allowed]
[Reword changelog]
Signed-off-by: Valentin Schneider <valentin.schneider@arm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
Link: https://lkml.kernel.org/r/20210407220628.3798191-2-valentin.schneider@arm.com
|
|
Mel Gorman did some nice work in 9fe1f127b913 ("sched/fair: Merge
select_idle_core/cpu()"), resulting in the kernel being more efficient
at finding an idle CPU, and in tasks spending less time waiting to be
run, both according to the schedstats run_delay numbers, and according
to measured application latencies. Yay.
The flip side of this is that we see more task migrations (about 30%
more), higher cache misses, higher memory bandwidth utilization, and
higher CPU use, for the same number of requests/second.
This is most pronounced on a memcache type workload, which saw a
consistent 1-3% increase in total CPU use on the system, due to those
increased task migrations leading to higher L2 cache miss numbers, and
higher memory utilization. The exclusive L3 cache on Skylake does us
no favors there.
On our web serving workload, that effect is usually negligible.
It appears that the increased number of CPU migrations is generally a
good thing, since it leads to lower cpu_delay numbers, reflecting the
fact that tasks get to run faster. However, the reduced locality and
the corresponding increase in L2 cache misses hurts a little.
The patch below appears to fix the regression, while keeping the
benefit of the lower cpu_delay numbers, by reintroducing
select_idle_smt with a twist: when a socket has no idle cores, check
to see if the sibling of "prev" is idle, before searching all the
other CPUs.
This fixes both the occasional 9% regression on the web serving
workload, and the continuous 2% CPU use regression on the memcache
type workload.
With Mel's patches and this patch together, task migrations are still
high, but L2 cache misses, memory bandwidth, and CPU time used are
back down to what they were before. The p95 and p99 response times for
the memcache type application improve by about 10% over what they were
before Mel's patches got merged.
Signed-off-by: Rik van Riel <riel@surriel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Vincent Guittot <vincent.guittot@linaro.org>
Link: https://lkml.kernel.org/r/20210326151932.2c187840@imladris.surriel.com
|
|
static_call_update() had stronger type requirements than regular C,
relax them to match. Instead of requiring the @func argument has the
exact matching type, allow any type which C is willing to promote to the
right (function) pointer type. Specifically this allows (void *)
arguments.
This cleans up a bunch of static_call_update() callers for
PREEMPT_DYNAMIC and should get around silly GCC11 warnings for free.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/YFoN7nCl8OfGtpeh@hirez.programming.kicks-ass.net
|
|
Currently only root can write files under /proc/pressure. Relax this to
allow tasks running as unprivileged users with CAP_SYS_RESOURCE to be
able to write to these files.
Signed-off-by: Josh Hunt <johunt@akamai.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Link: https://lkml.kernel.org/r/20210402025833.27599-1-johunt@akamai.com
|
|
mask is built in build_balance_mask() by for_each_cpu(i, sg_span), so
it must be a subset of sched_group_span(sg).
So the cpumask_and() call is redundant - remove it.
[ mingo: Adjusted the changelog a bit. ]
Signed-off-by: Barry Song <song.bao.hua@hisilicon.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Valentin Schneider <Valentin.Schneider@arm.com>
Link: https://lore.kernel.org/r/20210325023140.23456-1-song.bao.hua@hisilicon.com
|
|
-1 is -EPERM which is a somewhat odd error to return from
sched_dynamic_write(). No other callers care about which negative
value is used.
Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: https://lore.kernel.org/r/20210325004515.531631-2-linux@rasmusvillemoes.dk
|
|
Use the enum names which are also what is used in the switch() in
sched_dynamic_update().
Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: https://lore.kernel.org/r/20210325004515.531631-1-linux@rasmusvillemoes.dk
|
|
A long-tail load balance cost is observed on the newly idle path,
this is caused by a race window between the first nr_running check
of the busiest runqueue and its nr_running recheck in detach_tasks.
Before the busiest runqueue is locked, the tasks on the busiest
runqueue could be pulled by other CPUs and nr_running of the busiest
runqueu becomes 1 or even 0 if the running task becomes idle, this
causes detach_tasks breaks with LBF_ALL_PINNED flag set, and triggers
load_balance redo at the same sched_domain level.
In order to find the new busiest sched_group and CPU, load balance will
recompute and update the various load statistics, which eventually leads
to the long-tail load balance cost.
This patch clears LBF_ALL_PINNED flag for this race condition, and hence
reduces the long-tail cost of newly idle balance.
Signed-off-by: Aubrey Li <aubrey.li@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
Link: https://lkml.kernel.org/r/1614154549-116078-1-git-send-email-aubrey.li@intel.com
|
|
update_idle_core() is only done for the case of sched_smt_present.
but test_idle_cores() is done for all machines even those without
SMT.
This can contribute to up 8%+ hackbench performance loss on a
machine like kunpeng 920 which has no SMT. This patch removes the
redundant test_idle_cores() for !SMT machines.
Hackbench is ran with -g {2..14}, for each g it is ran 10 times to get
an average.
$ numactl -N 0 hackbench -p -T -l 20000 -g $1
The below is the result of hackbench w/ and w/o this patch:
g= 2 4 6 8 10 12 14
w/o: 1.8151 3.8499 5.5142 7.2491 9.0340 10.7345 12.0929
w/ : 1.8428 3.7436 5.4501 6.9522 8.2882 9.9535 11.3367
+4.1% +8.3% +7.3% +6.3%
Signed-off-by: Barry Song <song.bao.hua@hisilicon.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
Acked-by: Mel Gorman <mgorman@suse.de>
Link: https://lkml.kernel.org/r/20210320221432.924-1-song.bao.hua@hisilicon.com
|
|
We noticed that the cost of psi increases with the increase in the
levels of the cgroups. Particularly the cost of cpu_clock() sticks out
as the kernel calls it multiple times as it traverses up the cgroup
tree. This patch reduces the calls to cpu_clock().
Performed perf bench on Intel Broadwell with 3 levels of cgroup.
Before the patch:
$ perf bench sched all
# Running sched/messaging benchmark...
# 20 sender and receiver processes per group
# 10 groups == 400 processes run
Total time: 0.747 [sec]
# Running sched/pipe benchmark...
# Executed 1000000 pipe operations between two processes
Total time: 3.516 [sec]
3.516689 usecs/op
284358 ops/sec
After the patch:
$ perf bench sched all
# Running sched/messaging benchmark...
# 20 sender and receiver processes per group
# 10 groups == 400 processes run
Total time: 0.640 [sec]
# Running sched/pipe benchmark...
# Executed 1000000 pipe operations between two processes
Total time: 3.329 [sec]
3.329820 usecs/op
300316 ops/sec
Signed-off-by: Shakeel Butt <shakeelb@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Link: https://lkml.kernel.org/r/20210321205156.4186483-1-shakeelb@google.com
|
|
The Frequency Invariance Engine (FIE) is providing a frequency scaling
correction factor that helps achieve more accurate load-tracking.
Normally, this scaling factor can be obtained directly with the help of
the cpufreq drivers as they know the exact frequency the hardware is
running at. But that isn't the case for CPPC cpufreq driver.
Another way of obtaining that is using the arch specific counter
support, which is already present in kernel, but that hardware is
optional for platforms.
This patch updates the CPPC driver to register itself with the topology
core to provide its own implementation (cppc_scale_freq_tick()) of
topology_scale_freq_tick() which gets called by the scheduler on every
tick. Note that the arch specific counters have higher priority than
CPPC counters, if available, though the CPPC driver doesn't need to have
any special handling for that.
On an invocation of cppc_scale_freq_tick(), we schedule an irq work
(since we reach here from hard-irq context), which then schedules a
normal work item and cppc_scale_freq_workfn() updates the per_cpu
arch_freq_scale variable based on the counter updates since the last
tick.
To allow platforms to disable this CPPC counter-based frequency
invariance support, this is all done under CONFIG_ACPI_CPPC_CPUFREQ_FIE,
which is enabled by default.
This also exports sched_setattr_nocheck() as the CPPC driver can be
built as a module.
Cc: linux-acpi@vger.kernel.org
Reviewed-by: Ionela Voinescu <ionela.voinescu@arm.com>
Tested-by: Ionela Voinescu <ionela.voinescu@arm.com>
Tested-by: Vincent Guittot <vincent.guittot@linaro.org>
Acked-by: Rafael J. Wysocki <rafael@kernel.org>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
|
|
Fix ~42 single-word typos in scheduler code comments.
We have accumulated a few fun ones over the years. :-)
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Juri Lelli <juri.lelli@redhat.com>
Cc: Vincent Guittot <vincent.guittot@linaro.org>
Cc: Dietmar Eggemann <dietmar.eggemann@arm.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Ben Segall <bsegall@google.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: linux-kernel@vger.kernel.org
|
|
fast_switch_enabled
Note that sugov_update_next_freq() may return false, that means the
caller sugov_fast_switch() will do nothing except fast switch check.
Similarly, sugov_deferred_update() also has unnecessary operations
of raw_spin_{lock,unlock} in sugov_update_single_freq() for that case.
So, let's call sugov_update_next_freq() before the fast switch check
to avoid unnecessary behaviors above. Accordingly, update interface
definition to sugov_deferred_update() and remove sugov_fast_switch()
since we will call cpufreq_driver_fast_switch() directly instead.
Signed-off-by: Yue Hu <huyue2@yulong.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
vtime_account_irq and irqtime_account_irq() base checks on preempt_count()
which fails on RT because preempt_count() does not contain the softirq
accounting which is seperate on RT.
These checks do not need the full preempt count as they only operate on the
hard and softirq sections.
Use irq_count() instead which provides the correct value on both RT and non
RT kernels. The compiler is clever enough to fold the masking for !RT:
99b: 65 8b 05 00 00 00 00 mov %gs:0x0(%rip),%eax
- 9a2: 25 ff ff ff 7f and $0x7fffffff,%eax
+ 9a2: 25 00 ff ff 00 and $0xffff00,%eax
Reported-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Tested-by: Paul E. McKenney <paulmck@kernel.org>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20210309085727.153926793@linutronix.de
|
|
Since 565790d28b1 (sched: Fix balance_callback(), 2020-05-11), there
is no longer a need to reuse the result value of the call to finish_task_switch()
inside schedule_tail(), therefore the variable used to hold that value
(rq) is no longer needed.
Signed-off-by: Edmundo Carmona Antoranz <eantoranz@gmail.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20210306210739.1370486-1-eantoranz@gmail.com
|
|
A significant portion of __calc_delta() time is spent in the loop
shifting a u64 by 32 bits. Use `fls` instead of iterating.
This is ~7x faster on benchmarks.
The generic `fls` implementation (`generic_fls`) is still ~4x faster
than the loop.
Architectures that have a better implementation will make use of it. For
example, on x86 we get an additional factor 2 in speed without dedicated
implementation.
On GCC, the asm versions of `fls` are about the same speed as the
builtin. On Clang, the versions that use fls are more than twice as
slow as the builtin. This is because the way the `fls` function is
written, clang puts the value in memory:
https://godbolt.org/z/EfMbYe. This bug is filed at
https://bugs.llvm.org/show_bug.cgi?idI406.
```
name cpu/op
BM_Calc<__calc_delta_loop> 9.57ms Â=B112%
BM_Calc<__calc_delta_generic_fls> 2.36ms Â=B113%
BM_Calc<__calc_delta_asm_fls> 2.45ms Â=B113%
BM_Calc<__calc_delta_asm_fls_nomem> 1.66ms Â=B112%
BM_Calc<__calc_delta_asm_fls64> 2.46ms Â=B113%
BM_Calc<__calc_delta_asm_fls64_nomem> 1.34ms Â=B115%
BM_Calc<__calc_delta_builtin> 1.32ms Â=B111%
```
Signed-off-by: Clement Courbet <courbet@google.com>
Signed-off-by: Josh Don <joshdon@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20210303224653.2579656-1-joshdon@google.com
|
|
The commit 36b238d57172 ("psi: Optimize switching tasks inside shared
cgroups") only update cgroups whose state actually changes during a
task switch only in task preempt case, not in task sleep case.
We actually don't need to clear and set TSK_ONCPU state for common cgroups
of next and prev task in sleep case, that can save many psi_group_change
especially when most activity comes from one leaf cgroup.
sleep before:
psi_dequeue()
while ((group = iterate_groups(prev))) # all ancestors
psi_group_change(prev, .clear=TSK_RUNNING|TSK_ONCPU)
psi_task_switch()
while ((group = iterate_groups(next))) # all ancestors
psi_group_change(next, .set=TSK_ONCPU)
sleep after:
psi_dequeue()
nop
psi_task_switch()
while ((group = iterate_groups(next))) # until (prev & next)
psi_group_change(next, .set=TSK_ONCPU)
while ((group = iterate_groups(prev))) # all ancestors
psi_group_change(prev, .clear=common?TSK_RUNNING:TSK_RUNNING|TSK_ONCPU)
When a voluntary sleep switches to another task, we remove one call of
psi_group_change() for every common cgroup ancestor of the two tasks.
Co-developed-by: Muchun Song <songmuchun@bytedance.com>
Signed-off-by: Muchun Song <songmuchun@bytedance.com>
Signed-off-by: Chengming Zhou <zhouchengming@bytedance.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Link: https://lkml.kernel.org/r/20210303034659.91735-5-zhouchengming@bytedance.com
|
|
Move the unlikely branches out of line. This eliminates undesirable
jumps during wakeup and sleeps for workloads that aren't under any
sort of resource pressure.
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Chengming Zhou <zhouchengming@bytedance.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lkml.kernel.org/r/20210303034659.91735-4-zhouchengming@bytedance.com
|
|
Move the reclaim detection from the timer tick to the task state
tracking machinery using the recently added ONCPU state. And we
also add task psi_flags changes checking in the psi_task_switch()
optimization to update the parents properly.
In terms of performance and cost, this ONCPU task state tracking
is not cheaper than previous timer tick in aggregate. But the code is
simpler and shorter this way, so it's a maintainability win. And
Johannes did some testing with perf bench, the performace and cost
changes would be acceptable for real workloads.
Thanks to Johannes Weiner for pointing out the psi_task_switch()
optimization things and the clearer changelog.
Co-developed-by: Muchun Song <songmuchun@bytedance.com>
Signed-off-by: Muchun Song <songmuchun@bytedance.com>
Signed-off-by: Chengming Zhou <zhouchengming@bytedance.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Link: https://lkml.kernel.org/r/20210303034659.91735-3-zhouchengming@bytedance.com
|
|
The FULL state doesn't exist for the CPU resource at the system level,
but exist at the cgroup level, means all non-idle tasks in a cgroup are
delayed on the CPU resource which used by others outside of the cgroup
or throttled by the cgroup cpu.max configuration.
Co-developed-by: Muchun Song <songmuchun@bytedance.com>
Signed-off-by: Muchun Song <songmuchun@bytedance.com>
Signed-off-by: Chengming Zhou <zhouchengming@bytedance.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Link: https://lkml.kernel.org/r/20210303034659.91735-2-zhouchengming@bytedance.com
|
|
diameter > 2
As long as NUMA diameter > 2, building sched_domain by sibling's child
domain will definitely create a sched_domain with sched_group which will
span out of the sched_domain:
+------+ +------+ +-------+ +------+
| node | 12 |node | 20 | node | 12 |node |
| 0 +---------+1 +--------+ 2 +-------+3 |
+------+ +------+ +-------+ +------+
domain0 node0 node1 node2 node3
domain1 node0+1 node0+1 node2+3 node2+3
+
domain2 node0+1+2 |
group: node0+1 |
group:node2+3 <-------------------+
when node2 is added into the domain2 of node0, kernel is using the child
domain of node2's domain2, which is domain1(node2+3). Node 3 is outside
the span of the domain including node0+1+2.
This will make load_balance() run based on screwed avg_load and group_type
in the sched_group spanning out of the sched_domain, and it also makes
select_task_rq_fair() pick an idle CPU outside the sched_domain.
Real servers which suffer from this problem include Kunpeng920 and 8-node
Sun Fire X4600-M2, at least.
Here we move to use the *child* domain of the *child* domain of node2's
domain2 as the new added sched_group. At the same, we re-use the lower
level sgc directly.
+------+ +------+ +-------+ +------+
| node | 12 |node | 20 | node | 12 |node |
| 0 +---------+1 +--------+ 2 +-------+3 |
+------+ +------+ +-------+ +------+
domain0 node0 node1 +- node2 node3
|
domain1 node0+1 node0+1 | node2+3 node2+3
|
domain2 node0+1+2 |
group: node0+1 |
group:node2 <-------------------+
While the lower level sgc is re-used, this patch only changes the remote
sched_groups for those sched_domains playing grandchild trick, therefore,
sgc->next_update is still safe since it's only touched by CPUs that have
the group span as local group. And sgc->imbalance is also safe because
sd_parent remains the same in load_balance and LB only tries other CPUs
from the local group.
Moreover, since local groups are not touched, they are still getting
roughly equal size in a TL. And should_we_balance() only matters with
local groups, so the pull probability of those groups are still roughly
equal.
Tested by the below topology:
qemu-system-aarch64 -M virt -nographic \
-smp cpus=8 \
-numa node,cpus=0-1,nodeid=0 \
-numa node,cpus=2-3,nodeid=1 \
-numa node,cpus=4-5,nodeid=2 \
-numa node,cpus=6-7,nodeid=3 \
-numa dist,src=0,dst=1,val=12 \
-numa dist,src=0,dst=2,val=20 \
-numa dist,src=0,dst=3,val=22 \
-numa dist,src=1,dst=2,val=22 \
-numa dist,src=2,dst=3,val=12 \
-numa dist,src=1,dst=3,val=24 \
-m 4G -cpu cortex-a57 -kernel arch/arm64/boot/Image
w/o patch, we get lots of "groups don't span domain->span":
[ 0.802139] CPU0 attaching sched-domain(s):
[ 0.802193] domain-0: span=0-1 level=MC
[ 0.802443] groups: 0:{ span=0 cap=1013 }, 1:{ span=1 cap=979 }
[ 0.802693] domain-1: span=0-3 level=NUMA
[ 0.802731] groups: 0:{ span=0-1 cap=1992 }, 2:{ span=2-3 cap=1943 }
[ 0.802811] domain-2: span=0-5 level=NUMA
[ 0.802829] groups: 0:{ span=0-3 cap=3935 }, 4:{ span=4-7 cap=3937 }
[ 0.802881] ERROR: groups don't span domain->span
[ 0.803058] domain-3: span=0-7 level=NUMA
[ 0.803080] groups: 0:{ span=0-5 mask=0-1 cap=5843 }, 6:{ span=4-7 mask=6-7 cap=4077 }
[ 0.804055] CPU1 attaching sched-domain(s):
[ 0.804072] domain-0: span=0-1 level=MC
[ 0.804096] groups: 1:{ span=1 cap=979 }, 0:{ span=0 cap=1013 }
[ 0.804152] domain-1: span=0-3 level=NUMA
[ 0.804170] groups: 0:{ span=0-1 cap=1992 }, 2:{ span=2-3 cap=1943 }
[ 0.804219] domain-2: span=0-5 level=NUMA
[ 0.804236] groups: 0:{ span=0-3 cap=3935 }, 4:{ span=4-7 cap=3937 }
[ 0.804302] ERROR: groups don't span domain->span
[ 0.804520] domain-3: span=0-7 level=NUMA
[ 0.804546] groups: 0:{ span=0-5 mask=0-1 cap=5843 }, 6:{ span=4-7 mask=6-7 cap=4077 }
[ 0.804677] CPU2 attaching sched-domain(s):
[ 0.804687] domain-0: span=2-3 level=MC
[ 0.804705] groups: 2:{ span=2 cap=934 }, 3:{ span=3 cap=1009 }
[ 0.804754] domain-1: span=0-3 level=NUMA
[ 0.804772] groups: 2:{ span=2-3 cap=1943 }, 0:{ span=0-1 cap=1992 }
[ 0.804820] domain-2: span=0-5 level=NUMA
[ 0.804836] groups: 2:{ span=0-3 mask=2-3 cap=3991 }, 4:{ span=0-1,4-7 mask=4-5 cap=5985 }
[ 0.804944] ERROR: groups don't span domain->span
[ 0.805108] domain-3: span=0-7 level=NUMA
[ 0.805134] groups: 2:{ span=0-5 mask=2-3 cap=5899 }, 6:{ span=0-1,4-7 mask=6-7 cap=6125 }
[ 0.805223] CPU3 attaching sched-domain(s):
[ 0.805232] domain-0: span=2-3 level=MC
[ 0.805249] groups: 3:{ span=3 cap=1009 }, 2:{ span=2 cap=934 }
[ 0.805319] domain-1: span=0-3 level=NUMA
[ 0.805336] groups: 2:{ span=2-3 cap=1943 }, 0:{ span=0-1 cap=1992 }
[ 0.805383] domain-2: span=0-5 level=NUMA
[ 0.805399] groups: 2:{ span=0-3 mask=2-3 cap=3991 }, 4:{ span=0-1,4-7 mask=4-5 cap=5985 }
[ 0.805458] ERROR: groups don't span domain->span
[ 0.805605] domain-3: span=0-7 level=NUMA
[ 0.805626] groups: 2:{ span=0-5 mask=2-3 cap=5899 }, 6:{ span=0-1,4-7 mask=6-7 cap=6125 }
[ 0.805712] CPU4 attaching sched-domain(s):
[ 0.805721] domain-0: span=4-5 level=MC
[ 0.805738] groups: 4:{ span=4 cap=984 }, 5:{ span=5 cap=924 }
[ 0.805787] domain-1: span=4-7 level=NUMA
[ 0.805803] groups: 4:{ span=4-5 cap=1908 }, 6:{ span=6-7 cap=2029 }
[ 0.805851] domain-2: span=0-1,4-7 level=NUMA
[ 0.805867] groups: 4:{ span=4-7 cap=3937 }, 0:{ span=0-3 cap=3935 }
[ 0.805915] ERROR: groups don't span domain->span
[ 0.806108] domain-3: span=0-7 level=NUMA
[ 0.806130] groups: 4:{ span=0-1,4-7 mask=4-5 cap=5985 }, 2:{ span=0-3 mask=2-3 cap=3991 }
[ 0.806214] CPU5 attaching sched-domain(s):
[ 0.806222] domain-0: span=4-5 level=MC
[ 0.806240] groups: 5:{ span=5 cap=924 }, 4:{ span=4 cap=984 }
[ 0.806841] domain-1: span=4-7 level=NUMA
[ 0.806866] groups: 4:{ span=4-5 cap=1908 }, 6:{ span=6-7 cap=2029 }
[ 0.806934] domain-2: span=0-1,4-7 level=NUMA
[ 0.806953] groups: 4:{ span=4-7 cap=3937 }, 0:{ span=0-3 cap=3935 }
[ 0.807004] ERROR: groups don't span domain->span
[ 0.807312] domain-3: span=0-7 level=NUMA
[ 0.807386] groups: 4:{ span=0-1,4-7 mask=4-5 cap=5985 }, 2:{ span=0-3 mask=2-3 cap=3991 }
[ 0.807686] CPU6 attaching sched-domain(s):
[ 0.807710] domain-0: span=6-7 level=MC
[ 0.807750] groups: 6:{ span=6 cap=1017 }, 7:{ span=7 cap=1012 }
[ 0.807840] domain-1: span=4-7 level=NUMA
[ 0.807870] groups: 6:{ span=6-7 cap=2029 }, 4:{ span=4-5 cap=1908 }
[ 0.807952] domain-2: span=0-1,4-7 level=NUMA
[ 0.807985] groups: 6:{ span=4-7 mask=6-7 cap=4077 }, 0:{ span=0-5 mask=0-1 cap=5843 }
[ 0.808045] ERROR: groups don't span domain->span
[ 0.808257] domain-3: span=0-7 level=NUMA
[ 0.808571] groups: 6:{ span=0-1,4-7 mask=6-7 cap=6125 }, 2:{ span=0-5 mask=2-3 cap=5899 }
[ 0.808848] CPU7 attaching sched-domain(s):
[ 0.808860] domain-0: span=6-7 level=MC
[ 0.808880] groups: 7:{ span=7 cap=1012 }, 6:{ span=6 cap=1017 }
[ 0.808953] domain-1: span=4-7 level=NUMA
[ 0.808974] groups: 6:{ span=6-7 cap=2029 }, 4:{ span=4-5 cap=1908 }
[ 0.809034] domain-2: span=0-1,4-7 level=NUMA
[ 0.809055] groups: 6:{ span=4-7 mask=6-7 cap=4077 }, 0:{ span=0-5 mask=0-1 cap=5843 }
[ 0.809128] ERROR: groups don't span domain->span
[ 0.810361] domain-3: span=0-7 level=NUMA
[ 0.810400] groups: 6:{ span=0-1,4-7 mask=6-7 cap=5961 }, 2:{ span=0-5 mask=2-3 cap=5903 }
w/ patch, we don't get "groups don't span domain->span" any more:
[ 1.486271] CPU0 attaching sched-domain(s):
[ 1.486820] domain-0: span=0-1 level=MC
[ 1.500924] groups: 0:{ span=0 cap=980 }, 1:{ span=1 cap=994 }
[ 1.515717] domain-1: span=0-3 level=NUMA
[ 1.515903] groups: 0:{ span=0-1 cap=1974 }, 2:{ span=2-3 cap=1989 }
[ 1.516989] domain-2: span=0-5 level=NUMA
[ 1.517124] groups: 0:{ span=0-3 cap=3963 }, 4:{ span=4-5 cap=1949 }
[ 1.517369] domain-3: span=0-7 level=NUMA
[ 1.517423] groups: 0:{ span=0-5 mask=0-1 cap=5912 }, 6:{ span=4-7 mask=6-7 cap=4054 }
[ 1.520027] CPU1 attaching sched-domain(s):
[ 1.520097] domain-0: span=0-1 level=MC
[ 1.520184] groups: 1:{ span=1 cap=994 }, 0:{ span=0 cap=980 }
[ 1.520429] domain-1: span=0-3 level=NUMA
[ 1.520487] groups: 0:{ span=0-1 cap=1974 }, 2:{ span=2-3 cap=1989 }
[ 1.520687] domain-2: span=0-5 level=NUMA
[ 1.520744] groups: 0:{ span=0-3 cap=3963 }, 4:{ span=4-5 cap=1949 }
[ 1.520948] domain-3: span=0-7 level=NUMA
[ 1.521038] groups: 0:{ span=0-5 mask=0-1 cap=5912 }, 6:{ span=4-7 mask=6-7 cap=4054 }
[ 1.522068] CPU2 attaching sched-domain(s):
[ 1.522348] domain-0: span=2-3 level=MC
[ 1.522606] groups: 2:{ span=2 cap=1003 }, 3:{ span=3 cap=986 }
[ 1.522832] domain-1: span=0-3 level=NUMA
[ 1.522885] groups: 2:{ span=2-3 cap=1989 }, 0:{ span=0-1 cap=1974 }
[ 1.523043] domain-2: span=0-5 level=NUMA
[ 1.523092] groups: 2:{ span=0-3 mask=2-3 cap=4037 }, 4:{ span=4-5 cap=1949 }
[ 1.523302] domain-3: span=0-7 level=NUMA
[ 1.523352] groups: 2:{ span=0-5 mask=2-3 cap=5986 }, 6:{ span=0-1,4-7 mask=6-7 cap=6102 }
[ 1.523748] CPU3 attaching sched-domain(s):
[ 1.523774] domain-0: span=2-3 level=MC
[ 1.523825] groups: 3:{ span=3 cap=986 }, 2:{ span=2 cap=1003 }
[ 1.524009] domain-1: span=0-3 level=NUMA
[ 1.524086] groups: 2:{ span=2-3 cap=1989 }, 0:{ span=0-1 cap=1974 }
[ 1.524281] domain-2: span=0-5 level=NUMA
[ 1.524331] groups: 2:{ span=0-3 mask=2-3 cap=4037 }, 4:{ span=4-5 cap=1949 }
[ 1.524534] domain-3: span=0-7 level=NUMA
[ 1.524586] groups: 2:{ span=0-5 mask=2-3 cap=5986 }, 6:{ span=0-1,4-7 mask=6-7 cap=6102 }
[ 1.524847] CPU4 attaching sched-domain(s):
[ 1.524873] domain-0: span=4-5 level=MC
[ 1.524954] groups: 4:{ span=4 cap=958 }, 5:{ span=5 cap=991 }
[ 1.525105] domain-1: span=4-7 level=NUMA
[ 1.525153] groups: 4:{ span=4-5 cap=1949 }, 6:{ span=6-7 cap=2006 }
[ 1.525368] domain-2: span=0-1,4-7 level=NUMA
[ 1.525428] groups: 4:{ span=4-7 cap=3955 }, 0:{ span=0-1 cap=1974 }
[ 1.532726] domain-3: span=0-7 level=NUMA
[ 1.532811] groups: 4:{ span=0-1,4-7 mask=4-5 cap=6003 }, 2:{ span=0-3 mask=2-3 cap=4037 }
[ 1.534125] CPU5 attaching sched-domain(s):
[ 1.534159] domain-0: span=4-5 level=MC
[ 1.534303] groups: 5:{ span=5 cap=991 }, 4:{ span=4 cap=958 }
[ 1.534490] domain-1: span=4-7 level=NUMA
[ 1.534572] groups: 4:{ span=4-5 cap=1949 }, 6:{ span=6-7 cap=2006 }
[ 1.534734] domain-2: span=0-1,4-7 level=NUMA
[ 1.534783] groups: 4:{ span=4-7 cap=3955 }, 0:{ span=0-1 cap=1974 }
[ 1.536057] domain-3: span=0-7 level=NUMA
[ 1.536430] groups: 4:{ span=0-1,4-7 mask=4-5 cap=6003 }, 2:{ span=0-3 mask=2-3 cap=3896 }
[ 1.536815] CPU6 attaching sched-domain(s):
[ 1.536846] domain-0: span=6-7 level=MC
[ 1.536934] groups: 6:{ span=6 cap=1005 }, 7:{ span=7 cap=1001 }
[ 1.537144] domain-1: span=4-7 level=NUMA
[ 1.537262] groups: 6:{ span=6-7 cap=2006 }, 4:{ span=4-5 cap=1949 }
[ 1.537553] domain-2: span=0-1,4-7 level=NUMA
[ 1.537613] groups: 6:{ span=4-7 mask=6-7 cap=4054 }, 0:{ span=0-1 cap=1805 }
[ 1.537872] domain-3: span=0-7 level=NUMA
[ 1.537998] groups: 6:{ span=0-1,4-7 mask=6-7 cap=6102 }, 2:{ span=0-5 mask=2-3 cap=5845 }
[ 1.538448] CPU7 attaching sched-domain(s):
[ 1.538505] domain-0: span=6-7 level=MC
[ 1.538586] groups: 7:{ span=7 cap=1001 }, 6:{ span=6 cap=1005 }
[ 1.538746] domain-1: span=4-7 level=NUMA
[ 1.538798] groups: 6:{ span=6-7 cap=2006 }, 4:{ span=4-5 cap=1949 }
[ 1.539048] domain-2: span=0-1,4-7 level=NUMA
[ 1.539111] groups: 6:{ span=4-7 mask=6-7 cap=4054 }, 0:{ span=0-1 cap=1805 }
[ 1.539571] domain-3: span=0-7 level=NUMA
[ 1.539610] groups: 6:{ span=0-1,4-7 mask=6-7 cap=6102 }, 2:{ span=0-5 mask=2-3 cap=5845 }
Signed-off-by: Barry Song <song.bao.hua@hisilicon.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Valentin Schneider <valentin.schneider@arm.com>
Tested-by: Meelis Roos <mroos@linux.ee>
Link: https://lkml.kernel.org/r/20210224030944.15232-1-song.bao.hua@hisilicon.com
|
|
Being called for each dequeue, util_est reduces the number of its updates
by filtering out when the EWMA signal is different from the task util_avg
by less than 1%. It is a problem for a sudden util_avg ramp-up. Due to the
decay from a previous high util_avg, EWMA might now be close enough to
the new util_avg. No update would then happen while it would leave
ue.enqueued with an out-of-date value.
Taking into consideration the two util_est members, EWMA and enqueued for
the filtering, ensures, for both, an up-to-date value.
This is for now an issue only for the trace probe that might return the
stale value. Functional-wise, it isn't a problem, as the value is always
accessed through max(enqueued, ewma).
This problem has been observed using LISA's UtilConvergence:test_means on
the sd845c board.
No regression observed with Hackbench on sd845c and Perf-bench sched pipe
on hikey/hikey960.
Signed-off-by: Vincent Donnefort <vincent.donnefort@arm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
Link: https://lkml.kernel.org/r/20210225165820.1377125-1-vincent.donnefort@arm.com
|
|
Syzbot reported a handful of occurrences where an sd->nr_balance_failed can
grow to much higher values than one would expect.
A successful load_balance() resets it to 0; a failed one increments
it. Once it gets to sd->cache_nice_tries + 3, this *should* trigger an
active balance, which will either set it to sd->cache_nice_tries+1 or reset
it to 0. However, in case the to-be-active-balanced task is not allowed to
run on env->dst_cpu, then the increment is done without any further
modification.
This could then be repeated ad nauseam, and would explain the absurdly high
values reported by syzbot (86, 149). VincentG noted there is value in
letting sd->cache_nice_tries grow, so the shift itself should be
fixed. That means preventing:
"""
If the value of the right operand is negative or is greater than or equal
to the width of the promoted left operand, the behavior is undefined.
"""
Thus we need to cap the shift exponent to
BITS_PER_TYPE(typeof(lefthand)) - 1.
I had a look around for other similar cases via coccinelle:
@expr@
position pos;
expression E1;
expression E2;
@@
(
E1 >> E2@pos
|
E1 >> E2@pos
)
@cst depends on expr@
position pos;
expression expr.E1;
constant cst;
@@
(
E1 >> cst@pos
|
E1 << cst@pos
)
@script:python depends on !cst@
pos << expr.pos;
exp << expr.E2;
@@
# Dirty hack to ignore constexpr
if exp.upper() != exp:
coccilib.report.print_report(pos[0], "Possible UB shift here")
The only other match in kernel/sched is rq_clock_thermal() which employs
sched_thermal_decay_shift, and that exponent is already capped to 10, so
that one is fine.
Fixes: 5a7f55590467 ("sched/fair: Relax constraint on task's load during load balance")
Reported-by: syzbot+d7581744d5fd27c9fbe1@syzkaller.appspotmail.com
Signed-off-by: Valentin Schneider <valentin.schneider@arm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: http://lore.kernel.org/r/000000000000ffac1205b9a2112f@google.com
|
|
The sub_positive local version is saving an explicit load-store and is
enough for the cpu_util_next() usage.
Signed-off-by: Vincent Donnefort <vincent.donnefort@arm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Quentin Perret <qperret@google.com>
Reviewed-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Link: https://lkml.kernel.org/r/20210225083612.1113823-3-vincent.donnefort@arm.com
|
|
find_energy_efficient_cpu() (feec()) computes for each perf_domain (pd) an
energy delta as follows:
feec(task)
for_each_pd
base_energy = compute_energy(task, -1, pd)
-> for_each_cpu(pd)
-> cpu_util_next(cpu, task, -1)
energy_delta = compute_energy(task, dst_cpu, pd)
-> for_each_cpu(pd)
-> cpu_util_next(cpu, task, dst_cpu)
energy_delta -= base_energy
Then it picks the best CPU as being the one that minimizes energy_delta.
cpu_util_next() estimates the CPU utilization that would happen if the
task was placed on dst_cpu as follows:
max(cpu_util + task_util, cpu_util_est + _task_util_est)
The task contribution to the energy delta can then be either:
(1) _task_util_est, on a mostly idle CPU, where cpu_util is close to 0
and _task_util_est > cpu_util.
(2) task_util, on a mostly busy CPU, where cpu_util > _task_util_est.
(cpu_util_est doesn't appear here. It is 0 when a CPU is idle and
otherwise must be small enough so that feec() takes the CPU as a
potential target for the task placement)
This is problematic for feec(), as cpu_util_next() might give an unfair
advantage to a CPU which is mostly busy (2) compared to one which is
mostly idle (1). _task_util_est being always bigger than task_util in
feec() (as the task is waking up), the task contribution to the energy
might look smaller on certain CPUs (2) and this breaks the energy
comparison.
This issue is, moreover, not sporadic. By starving idle CPUs, it keeps
their cpu_util < _task_util_est (1) while others will maintain cpu_util >
_task_util_est (2).
Fix this problem by always using max(task_util, _task_util_est) as a task
contribution to the energy (ENERGY_UTIL). The new estimated CPU
utilization for the energy would then be:
max(cpu_util, cpu_util_est) + max(task_util, _task_util_est)
compute_energy() still needs to know which OPP would be selected if the
task would be migrated in the perf_domain (FREQUENCY_UTIL). Hence,
cpu_util_next() is still used to estimate the maximum util within the pd.
Signed-off-by: Vincent Donnefort <vincent.donnefort@arm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Quentin Perret <qperret@google.com>
Reviewed-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Link: https://lkml.kernel.org/r/20210225083612.1113823-2-vincent.donnefort@arm.com
|
|
Start to update last_blocked_load_update_tick to reduce the possibility
of another cpu starting the update one more time
Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Valentin Schneider <valentin.schneider@arm.com>
Link: https://lkml.kernel.org/r/20210224133007.28644-8-vincent.guittot@linaro.org
|
|
Instead of waking up a random and already idle CPU, we can take advantage
of this_cpu being about to enter idle to run the ILB and update the
blocked load.
Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Valentin Schneider <valentin.schneider@arm.com>
Link: https://lkml.kernel.org/r/20210224133007.28644-7-vincent.guittot@linaro.org
|
|
The function sync_runqueues_membarrier_state() should copy the
membarrier state from the @mm received as parameter to each runqueue
currently running tasks using that mm.
However, the use of smp_call_function_many() skips the current runqueue,
which is unintended. Replace by a call to on_each_cpu_mask().
Fixes: 227a4aadc75b ("sched/membarrier: Fix p->mm->membarrier_state racy load")
Reported-by: Nadav Amit <nadav.amit@gmail.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: stable@vger.kernel.org # 5.4.x+
Link: https://lore.kernel.org/r/74F1E842-4A84-47BF-B6C2-5407DFDD4A4A@gmail.com
|
