The clocksource watchdog code has over time reached the state of an
impenetrable maze of duct tape and staples. The original design, which was
made in the context of systems far smaller than today, is based on the
assumption that the to be monitored clocksource (TSC) can be trivially
compared against a known to be stable clocksource (HPET/ACPI-PM timer).
Over the years it turned out that this approach has major flaws:
- Long delays between watchdog invocations can result in wrap arounds
of the reference clocksource
- Scalability of the reference clocksource readout can degrade on large
multi-socket systems due to interconnect congestion
This was addressed with various heuristics which degraded the accuracy of
the watchdog to the point that it fails to detect actual TSC problems on
older hardware which exposes slow inter CPU drifts due to firmware
manipulating the TSC to hide SMI time.
To address this and bring back sanity to the watchdog, rewrite the code
completely with a different approach:
1) Restrict the validation against a reference clocksource to the boot
CPU, which is usually the CPU/Socket closest to the legacy block which
contains the reference source (HPET/ACPI-PM timer). Validate that the
reference readout is within a bound latency so that the actual
comparison against the TSC stays within 500ppm as long as the clocks
are stable.
2) Compare the TSCs of the other CPUs in a round robin fashion against
the boot CPU in the same way the TSC synchronization on CPU hotplug
works. This still can suffer from delayed reaction of the remote CPU
to the SMP function call and the latency of the control variable cache
line. But this latency is not affecting correctness. It only affects
the accuracy. With low contention the readout latency is in the low
nanoseconds range, which detects even slight skews between CPUs. Under
high contention this becomes obviously less accurate, but still
detects slow skews reliably as it solely relies on subsequent readouts
being monotonically increasing. It just can take slightly longer to
detect the issue.
3) Rewrite the watchdog test so it tests the various mechanisms one by
one and validating the result against the expectation.
Signed-off-by: Thomas Gleixner <tglx@kernel.org>
Tested-by: Borislav Petkov (AMD) <bp@alien8.de>
Tested-by: Daniel J Blueman <daniel@quora.org>
Reviewed-by: Jiri Wiesner <jwiesner@suse.de>
Reviewed-by: Daniel J Blueman <daniel@quora.org>
Link: https://patch.msgid.link/20260123231521.926490888@kernel.org
Link: https://patch.msgid.link/87h5qeomm5.ffs@tglx
Currently hrtimer_interrupt() runs expired timers, which can re-arm
themselves, after which it computes the next expiration time and
re-programs the hardware.
However, things like HRTICK, a highres timer driving preemption, cannot
re-arm itself at the point of running, since the next task has not been
determined yet. The schedule() in the interrupt return path will switch to
the next task, which then causes a new hrtimer to be programmed.
This then results in reprogramming the hardware at least twice, once after
running the timers, and once upon selecting the new task.
Notably, *both* events happen in the interrupt.
By pushing the hrtimer reprogram all the way into the interrupt return
path, it runs after schedule() picks the new task and the double reprogram
can be avoided.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@kernel.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://patch.msgid.link/20260224163431.273488269@kernel.org
The hrtimer interrupt expires timers and at the end of the interrupt it
rearms the clockevent device for the next expiring timer.
That's obviously correct, but in the case that a expired timer set
NEED_RESCHED the return from interrupt ends up in schedule(). If HRTICK is
enabled then schedule() will modify the hrtick timer, which causes another
reprogramming of the hardware.
That can be avoided by deferring the rearming to the return from interrupt
path and if the return results in a immediate schedule() invocation then it
can be deferred until the end of schedule().
To make this correct the affected code parts need to be made aware of this.
Provide empty stubs for the deferred rearming mechanism, so that the
relevant code changes for entry, softirq and scheduler can be split up into
separate changes independent of the actual enablement in the hrtimer code.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@kernel.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://patch.msgid.link/20260224163431.000891171@kernel.org
Some clockevent devices are coupled to the system clocksource by
implementing a less than or equal comparator which compares the programmed
absolute expiry time against the underlying time counter.
The timekeeping core provides a function to convert and absolute
CLOCK_MONOTONIC based expiry time to a absolute clock cycles time which can
be directly fed into the comparator. That spares two time reads in the next
event progamming path, one to convert the absolute nanoseconds time to a
delta value and the other to convert the delta value back to a absolute
time value suitable for the comparator.
Provide a new clocksource callback which takes the absolute cycle value and
wire it up in clockevents_program_event(). Similar to clocksources allow
architectures to inline the rearm operation.
Signed-off-by: Thomas Gleixner <tglx@kernel.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://patch.msgid.link/20260224163430.010425428@kernel.org
Some architectures have clockevent devices which are coupled to the system
clocksource by implementing a less than or equal comparator which compares
the programmed absolute expiry time against the underlying time
counter. Well known examples are TSC/TSC deadline timer and the S390 TOD
clocksource/comparator.
While the concept is nice it has some downsides:
1) The clockevents core code is strictly based on relative expiry times
as that's the most common case for clockevent device hardware. That
requires to convert the absolute expiry time provided by the caller
(hrtimers, NOHZ code) to a relative expiry time by reading and
substracting the current time.
The clockevent::set_next_event() callback must then read the counter
again to convert the relative expiry back into a absolute one.
2) The conversion factors from nanoseconds to counter clock cycles are
set up when the clockevent is registered. When NTP applies corrections
then the clockevent conversion factors can deviate from the
clocksource conversion substantially which either results in timers
firing late or in the worst case early. The early expiry then needs to
do a reprogam with a short delta.
In most cases this is papered over by the fact that the read in the
set_next_event() callback happens after the read which is used to
calculate the delta. So the tendency is that timers expire mostly
late.
All of this can be avoided by providing support for these devices in the
core code:
1) The timekeeping core keeps track of the last update to the clocksource
by storing the base nanoseconds and the corresponding clocksource
counter value. That's used to keep the conversion math for reading the
time within 64-bit in the common case.
This information can be used to avoid both reads of the underlying
clocksource in the clockevents reprogramming path:
delta = expiry - base_ns;
cycles = base_cycles + ((delta * clockevent::mult) >> clockevent::shift);
The resulting cycles value can be directly used to program the
comparator.
2) As #1 does not longer provide the "compensation" through the second
read the deviation of the clocksource and clockevent conversions
caused by NTP become more prominent.
This can be cured by letting the timekeeping core compute and store
the reverse conversion factors when the clocksource cycles to
nanoseconds factors are modified by NTP:
CS::MULT (1 << NS_TO_CYC_SHIFT)
--------------- = ----------------------
(1 << CS:SHIFT) NS_TO_CYC_MULT
Ergo: NS_TO_CYC_MULT = (1 << (CS::SHIFT + NS_TO_CYC_SHIFT)) / CS::MULT
The NS_TO_CYC_SHIFT value is calculated when the clocksource is
installed so that it aims for a one hour maximum sleep time.
Signed-off-by: Thomas Gleixner <tglx@kernel.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://patch.msgid.link/20260224163429.944763521@kernel.org
On some architectures clocksource::read() boils down to a single
instruction, so the indirect function call is just a massive overhead
especially with speculative execution mitigations in effect.
Allow architectures to enable conditional inlining of that read to avoid
that by:
- providing a static branch to switch to the inlined variant
- disabling the branch before clocksource changes
- enabling the branch after a clocksource change, when the clocksource
indicates in a feature flag that it is the one which provides the
inlined variant
This is intentionally not a static call as that would only remove the
indirect call, but not the rest of the overhead.
Signed-off-by: Thomas Gleixner <tglx@kernel.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://patch.msgid.link/20260224163429.675151545@kernel.org
To support auxiliary timekeeping and the related user space interfaces,
it's required to define a clock ID range for them.
Reserve 8 auxiliary clock IDs after the regular timekeeping clock ID space.
This is the maximum number of auxiliary clocks the kernel can support. The actual
number of supported clocks depends obviously on the presence of related devices
and might be constraint by the available VDSO space.
Add the corresponding timekeeper IDs as well.
Signed-off-by: Anna-Maria Behnsen <anna-maria@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: John Stultz <jstultz@google.com>
Link: https://lore.kernel.org/all/20250519083025.905800695@linutronix.de
clocksource_delta() has two variants. One with a check for negative motion,
which is only selected by x86. This is a historic leftover as this function
was previously used in the time getter hot paths.
Since 135225a363 timekeeping_cycles_to_ns() has unconditional protection
against this as a by-product of the protection against 64bit math overflow.
clocksource_delta() is only used in the clocksource watchdog and in
timekeeping_advance(). The extra conditional there is not hurting anyone.
Remove the config option and unconditionally prevent negative motion of the
readout.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: John Stultz <jstultz@google.com>
Link: https://lore.kernel.org/all/20241031120328.599430157@linutronix.de
When doing
make menuconfig
and searching for the CLOCKSOURCE_WATCHDOG_MAX_SKEW_US config item, the
help says:
│ Symbol: CLOCKSOURCE_WATCHDOG_MAX_SKEW_US [=125]
│ Type : integer
│ Range : [50 1000]
│ Defined at kernel/time/Kconfig:204
│ Prompt: Clocksource watchdog maximum allowable skew (in s)
^^^
│ Depends on: GENERIC_CLOCKEVENTS [=y] && CLOCKSOURCE_WATCHDOG [=y]
because on some terminals, it cannot display the 'μ' char, unicode
number 0x3bc.
So simply write it out so that there's no trouble.
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Paul E. McKenney <paulmck@kernel.org>
Acked-by: Randy Dunlap <rdunlap@infradead.org>
Link: https://lore.kernel.org/r/20240428102143.26764-1-bp@kernel.org
The x86 architecture has an idle routine for AMD CPUs which are affected
by erratum 400. On the affected CPUs the local APIC timer stops in the
C1E halt state.
It therefore requires tick broadcasting. The invocation of
tick_broadcast_enter()/exit() from this function violates the RCU
constraints because it can end up in lockdep or tracing, which
rightfully triggers a warning.
tick_broadcast_enter()/exit() must be invoked before ct_cpuidle_enter()
and after ct_cpuidle_exit() in default_idle_call().
Add a static branch conditional invocation of tick_broadcast_enter()/exit()
into this function to allow X86 to replace the AMD specific idle code. It's
guarded by a config switch which will be selected by x86. Otherwise it's
a NOOP.
Reported-by: Borislav Petkov <bp@alien8.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Link: https://lore.kernel.org/r/20240229142248.266708822@linutronix.de
Currently, MAX_SKEW_USEC is set to 100 microseconds, which has worked
reasonably well. However, NTP is willing to tolerate 500 microseconds
of skew per second, and a clocksource that is good enough for NTP should
be good enough for the clocksource watchdog. The watchdog's skew is
controlled by MAX_SKEW_USEC and the CLOCKSOURCE_WATCHDOG_MAX_SKEW_US
Kconfig option. However, these values are doubled before being associated
with a clocksource's ->uncertainty_margin, and the ->uncertainty_margin
values of the pair of clocksource's being compared are summed before
checking against the skew.
Therefore, set both MAX_SKEW_USEC and the default for the
CLOCKSOURCE_WATCHDOG_MAX_SKEW_US Kconfig option to 125 microseconds of
skew per second, resulting in 500 microseconds of skew per second in
the clocksource watchdog's skew comparison.
Suggested-by Rik van Riel <riel@surriel.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
A watchdog maximum skew of 100us may still be too small for
some systems or archs. It may also be too small when some kernel
debug config options are enabled. So add a new Kconfig option
CLOCKSOURCE_WATCHDOG_MAX_SKEW_US to allow kernel builders to have more
control on the threshold for marking clocksource as unstable.
Signed-off-by: Waiman Long <longman@redhat.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Pull timer updates from Thomas Gleixner:
"Time and clocksource/clockevent related updates:
Core changes:
- Infrastructure to support per CPU "broadcast" devices for per CPU
clockevent devices which stop in deep idle states. This allows us
to utilize the more efficient architected timer on certain ARM SoCs
for normal operation instead of permanentely using the slow to
access SoC specific clockevent device.
- Print the name of the broadcast/wakeup device in /proc/timer_list
- Make the clocksource watchdog more robust against delays between
reading the current active clocksource and the watchdog
clocksource. Such delays can be caused by NMIs, SMIs and vCPU
preemption.
Handle this by reading the watchdog clocksource twice, i.e. before
and after reading the current active clocksource. In case that the
two watchdog reads shows an excessive time delta, the read sequence
is repeated up to 3 times.
- Improve the debug output and add a test module for the watchdog
mechanism.
- Reimplementation of the venerable time64_to_tm() function with a
faster and significantly smaller version. Straight from the source,
i.e. the author of the related research paper contributed this!
Driver changes:
- No new drivers, not even new device tree bindings!
- Fixes, improvements and cleanups and all over the place"
* tag 'timers-core-2021-06-29' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (30 commits)
time/kunit: Add missing MODULE_LICENSE()
time: Improve performance of time64_to_tm()
clockevents: Use list_move() instead of list_del()/list_add()
clocksource: Print deviation in nanoseconds when a clocksource becomes unstable
clocksource: Provide kernel module to test clocksource watchdog
clocksource: Reduce clocksource-skew threshold
clocksource: Limit number of CPUs checked for clock synchronization
clocksource: Check per-CPU clock synchronization when marked unstable
clocksource: Retry clock read if long delays detected
clockevents: Add missing parameter documentation
clocksource/drivers/timer-ti-dm: Drop unnecessary restore
clocksource/arm_arch_timer: Improve Allwinner A64 timer workaround
clocksource/drivers/arm_global_timer: Remove duplicated argument in arm_global_timer
clocksource/drivers/arm_global_timer: Make symbol 'gt_clk_rate_change_nb' static
arm: zynq: don't disable CONFIG_ARM_GLOBAL_TIMER due to CONFIG_CPU_FREQ anymore
clocksource/drivers/arm_global_timer: Implement rate compensation whenever source clock changes
clocksource/drivers/ingenic: Rename unreasonable array names
clocksource/drivers/timer-ti-dm: Save and restore timer TIOCP_CFG
clocksource/drivers/mediatek: Ack and disable interrupts on suspend
clocksource/drivers/samsung_pwm: Constify source IO memory
...
The current implementation of time64_to_tm() contains unnecessary loops,
branches and look-up tables. The new one uses an arithmetic-based algorithm
appeared in [1] and is approximately 3x faster (YMMV).
The drawback is that the new code isn't intuitive and contains many 'magic
numbers' (not unusual for this type of algorithm). However, [1] justifies
all those numbers and, given this function's history, the code is unlikely
to need much maintenance, if any at all.
Add a KUnit test for it which checks every day in a 160,000 years interval
centered at 1970-01-01 against the expected result.
[1] Neri, Schneider, "Euclidean Affine Functions and Applications to
Calendar Algorithms". https://arxiv.org/abs/2102.06959
Signed-off-by: Cassio Neri <cassio.neri@gmail.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20210622213616.313046-1-cassio.neri@gmail.com
Pull timer fixes from Ingo Molnar:
"Update/fix two CPU sanity checks in the hotplug and the boot code, and
fix a typo in the Kconfig help text.
[ Context: the first two commits are the result of an ongoing
annotation+review work of (intentional) tick_do_timer_cpu() data
races reported by KCSAN, but the annotations aren't fully cooked
yet ]"
* tag 'timers-urgent-2020-12-27' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
timekeeping: Fix spelling mistake in Kconfig "fullfill" -> "fulfill"
tick/sched: Remove bogus boot "safety" check
tick: Remove pointless cpu valid check in hotplug code
Almost all machines use GENERIC_CLOCKEVENTS, so it feels wrong to
require each one to select that symbol manually.
Instead, enable it whenever CONFIG_LEGACY_TIMER_TICK is disabled as
a simplification. It should be possible to select both
GENERIC_CLOCKEVENTS and LEGACY_TIMER_TICK from an architecture now
and decide at runtime between the two.
For the clockevents arch-support.txt file, this means that additional
architectures are marked as TODO when they have at least one machine
that still uses LEGACY_TIMER_TICK, rather than being marked 'ok' when
at least one machine has been converted. This means that both m68k and
arm (for riscpc) revert to TODO.
At this point, we could just always enable CONFIG_GENERIC_CLOCKEVENTS
rather than leaving it off when not needed. I built an m68k
defconfig kernel (using gcc-10.1.0) and found that this would add
around 5.5KB in kernel image size:
text data bss dec hex filename
3861936 1092236 196656 5150828 4e986c obj-m68k/vmlinux-no-clockevent
3866201 1093832 196184 5156217 4ead79 obj-m68k/vmlinux-clockevent
On Arm (MACH_RPC), that difference appears to be twice as large,
around 11KB on top of an 6MB vmlinux.
Reviewed-by: Geert Uytterhoeven <geert@linux-m68k.org>
Acked-by: Geert Uytterhoeven <geert@linux-m68k.org>
Tested-by: Geert Uytterhoeven <geert@linux-m68k.org>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
All platforms that currently do not use generic clockevents roughly call
the same set of functions in their timer interrupts: xtime_update(),
update_process_times() and profile_tick(), sometimes in a different
sequence.
Add a helper function that performs all three of them, to make the
callers more uniform and simplify the interface.
Reviewed-by: Geert Uytterhoeven <geert@linux-m68k.org>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
With Arm EBSA110 gone, nothing uses it any more, so the corresponding
code and the Kconfig option can be removed.
Acked-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Running posix CPU timers in hard interrupt context has a few downsides:
- For PREEMPT_RT it cannot work as the expiry code needs to take
sighand lock, which is a 'sleeping spinlock' in RT. The original RT
approach of offloading the posix CPU timer handling into a high
priority thread was clumsy and provided no real benefit in general.
- For fine grained accounting it's just wrong to run this in context of
the timer interrupt because that way a process specific CPU time is
accounted to the timer interrupt.
- Long running timer interrupts caused by a large amount of expiring
timers which can be created and armed by unpriviledged user space.
There is no hard requirement to expire them in interrupt context.
If the signal is targeted at the task itself then it won't be delivered
before the task returns to user space anyway. If the signal is targeted at
a supervisor process then it might be slightly delayed, but posix CPU
timers are inaccurate anyway due to the fact that they are tied to the
tick.
Provide infrastructure to schedule task work which allows splitting the
posix CPU timer code into a quick check in interrupt context and a thread
context expiry and signal delivery function. This has to be enabled by
architectures as it requires that the architecture specific KVM
implementation handles pending task work before exiting to guest mode.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Oleg Nesterov <oleg@redhat.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20200730102337.783470146@linutronix.de
Add SPDX license identifiers to all Make/Kconfig files which:
- Have no license information of any form
These files fall under the project license, GPL v2 only. The resulting SPDX
license identifier is:
GPL-2.0-only
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Both CONTEXT_TRACKING and CONTEXT_TRACKING_FORCE are currently defined
in kernel/rcu/kconfig, which might have made sense at some point, but
no longer does given that RCU refers to neither of these Kconfig options.
Therefore move them to kernel/time/Kconfig, where the rest of the
NO_HZ_FULL Kconfig options live.
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Link: https://lkml.kernel.org/r/20181220170525.GA12579@linux.ibm.com
Commit 6f1982fedd ("sched/isolation: Handle the nohz_full= parameter")
broke CONFIG_NO_HZ_FULL_ALL=y kernels. This breakage is due to the code
under CONFIG_NO_HZ_FULL_ALL failing to invoke the shiny new housekeeping
functions. This means that rcutorture scenario TREE04 now emits RCU CPU
stall warnings due to the RCU grace-period kthreads not being awakened
at a time of their choosing, or perhaps even not at all:
[ 27.731422] rcu_bh kthread starved for 21001 jiffies! g18446744073709551369 c18446744073709551368 f0x0 RCU_GP_WAIT_FQS(3) ->state=0x402 ->cpu=3
[ 27.731423] rcu_bh I14936 9 2 0x80080000
[ 27.731435] Call Trace:
[ 27.731440] __schedule+0x31a/0x6d0
[ 27.731442] schedule+0x31/0x80
[ 27.731446] schedule_timeout+0x15a/0x320
[ 27.731453] ? call_timer_fn+0x130/0x130
[ 27.731457] rcu_gp_kthread+0x66c/0xea0
[ 27.731458] ? rcu_gp_kthread+0x66c/0xea0
Because no one has complained about CONFIG_NO_HZ_FULL_ALL=y being broken,
I hypothesize that no one is in fact using it, other than rcutorture.
This commit therefore eliminates CONFIG_NO_HZ_FULL_ALL and updates
rcutorture's config files to instead use the nohz_full= kernel parameter
to put the desired CPUs into nohz_full mode.
Fixes: 6f1982fedd ("sched/isolation: Handle the nohz_full= parameter")
Reported-by: kernel test robot <xiaolong.ye@intel.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Frederic Weisbecker <frederic@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Chris Metcalf <cmetcalf@mellanox.com>
Cc: Christoph Lameter <cl@linux.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Luiz Capitulino <lcapitulino@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: Wanpeng Li <kernellwp@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Jonathan Corbet <corbet@lwn.net>
CONFIG_NO_HZ_FULL doesn't make sense without CONFIG_CPU_ISOLATION. In
fact enabling the first without the second is a regression as nohz_full=
boot parameter gets silently ignored.
Besides this unnatural combination hangs RCU gp kthread when running
rcutorture for reasons that are not yet fully understood:
rcu_preempt kthread starved for 9974 jiffies! g4294967208
+c4294967207 f0x0 RCU_GP_WAIT_FQS(3) ->state=0x402 ->cpu=0
rcu_preempt I 7464 8 2 0x80000000
Call Trace:
__schedule+0x493/0x620
schedule+0x24/0x40
schedule_timeout+0x330/0x3b0
? preempt_count_sub+0xea/0x140
? collect_expired_timers+0xb0/0xb0
rcu_gp_kthread+0x6bf/0xef0
This commit therefore makes NO_HZ_FULL select CPU_ISOLATION, which
prevents all these bad behaviours.
Reported-by: kernel test robot <xiaolong.ye@intel.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Cc: Chris Metcalf <cmetcalf@mellanox.com>
Cc: Christoph Lameter <cl@linux.com>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Luiz Capitulino <lcapitulino@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Wanpeng Li <kernellwp@gmail.com>
Fixes: 5c4991e24c ("sched/isolation: Split out new CONFIG_CPU_ISOLATION=y config from CONFIG_NO_HZ_FULL")
Link: http://lkml.kernel.org/r/1513275507-29200-2-git-send-email-frederic@kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The NO_HZ_FULL_SYSIDLE full-system-idle capability was added in 2013
by commit 0edd1b1784 ("nohz_full: Add full-system-idle state machine"),
but has not been used. This commit therefore removes it.
If it turns out to be needed later, this commit can always be reverted.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Ingo Molnar <mingo@kernel.org>
Acked-by: Linus Torvalds <torvalds@linux-foundation.org>
The RCU_USER_QS Kconfig parameter is now just a synonym for NO_HZ_FULL,
so this commit eliminates RCU_USER_QS, replacing all uses with NO_HZ_FULL.
Reported-by: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Acked-by: Frederic Weisbecker <fweisbec@gmail.com>
The only user of the cycle_last validation is the x86 TSC. In order to
provide NMI safe accessor functions for clock monotonic and
monotonic_raw we need to do that in the core.
We can't do the TSC specific
if (now < cycle_last)
now = cycle_last;
for the other wrapping around clocksources, but TSC has
CLOCKSOURCE_MASK(64) which actually does not mask out anything so if
now is less than cycle_last the subtraction will give a negative
result. So we can check for that in clocksource_delta() and return 0
for that case.
Implement and enable it for x86
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: John Stultz <john.stultz@linaro.org>
The non-scalar ktime_t implementation is basically a timespec
which has to be changed to support dates past 2038 on 32bit
systems.
This patch removes the non-scalar ktime_t implementation, forcing
the scalar s64 nanosecond version on all architectures.
This may have additional performance overhead on some 32bit
systems when converting between ktime_t and timespec structures,
however the majority of 32bit systems (arm and i386) were already
using scalar ktime_t, so no performance regressions will be seen
on those platforms.
On affected platforms, I'm open to finding optimizations, including
avoiding converting to timespecs where possible.
[ tglx: We can now cleanup the ktime_t.tv64 mess, but thats a
different issue and we can throw a coccinelle script at it ]
Signed-off-by: John Stultz <john.stultz@linaro.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: John Stultz <john.stultz@linaro.org>
With VIRT_CPU_ACCOUNTING_GEN, cputime_t becomes 64-bit. In order
to use that feature, arch code should be audited to ensure there are no
races in concurrent read/write of cputime_t. For example,
reading/writing 64-bit cputime_t on some 32-bit arches may require
multiple accesses for low and high value parts, so proper locking
is needed to protect against concurrent accesses.
Therefore, add CONFIG_HAVE_VIRT_CPU_ACCOUNTING_GEN which arches can
enable after they've been audited for potential races.
This option is automatically enabled on 64-bit platforms.
Feature requested by Frederic Weisbecker.
Signed-off-by: Kevin Hilman <khilman@linaro.org>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Russell King <rmk@arm.linux.org.uk>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Arm Linux <linux-arm-kernel@lists.infradead.org>
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Pull timers/nohz changes from Ingo Molnar:
"It mostly contains fixes and full dynticks off-case optimizations, by
Frederic Weisbecker"
* 'timers-nohz-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (24 commits)
nohz: Include local CPU in full dynticks global kick
nohz: Optimize full dynticks's sched hooks with static keys
nohz: Optimize full dynticks state checks with static keys
nohz: Rename a few state variables
vtime: Always debug check snapshot source _before_ updating it
vtime: Always scale generic vtime accounting results
vtime: Optimize full dynticks accounting off case with static keys
vtime: Describe overriden functions in dedicated arch headers
m68k: hardirq_count() only need preempt_mask.h
hardirq: Split preempt count mask definitions
context_tracking: Split low level state headers
vtime: Fix racy cputime delta update
vtime: Remove a few unneeded generic vtime state checks
context_tracking: User/kernel broundary cross trace events
context_tracking: Optimize context switch off case with static keys
context_tracking: Optimize guest APIs off case with static key
context_tracking: Optimize main APIs off case with static key
context_tracking: Ground setup for static key use
context_tracking: Remove full dynticks' hacky dependency on wide context tracking
nohz: Only enable context tracking on full dynticks CPUs
...
This commit adds the state machine that takes the per-CPU idle data
as input and produces a full-system-idle indication as output. This
state machine is driven out of RCU's quiescent-state-forcing
mechanism, which invokes rcu_sysidle_check_cpu() to collect per-CPU
idle state and then rcu_sysidle_report() to drive the state machine.
The full-system-idle state is sampled using rcu_sys_is_idle(), which
also drives the state machine if RCU is idle (and does so by forcing
RCU to become non-idle). This function returns true if all but the
timekeeping CPU (tick_do_timer_cpu) are idle and have been idle long
enough to avoid memory contention on the full_sysidle_state state
variable. The rcu_sysidle_force_exit() may be called externally
to reset the state machine back into non-idle state.
For large systems the state machine is driven out of RCU's
force-quiescent-state logic, which provides good scalability at the price
of millisecond-scale latencies on the transition to full-system-idle
state. This is not so good for battery-powered systems, which are usually
small enough that they don't need to care about scalability, but which
do care deeply about energy efficiency. Small systems therefore drive
the state machine directly out of the idle-entry code. The number of
CPUs in a "small" system is defined by a new NO_HZ_FULL_SYSIDLE_SMALL
Kconfig parameter, which defaults to 8. Note that this is a build-time
definition.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
[ paulmck: Use true and false for boolean constants per Lai Jiangshan. ]
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
[ paulmck: Simplify logic and provide better comments for memory barriers,
based on review comments and questions by Lai Jiangshan. ]
At least one CPU must keep the scheduling-clock tick running for
timekeeping purposes whenever there is a non-idle CPU. However, with
the new nohz_full adaptive-idle machinery, it is difficult to distinguish
between all CPUs really being idle as opposed to all non-idle CPUs being
in adaptive-ticks mode. This commit therefore adds a Kconfig parameter
as a first step towards enabling a scalable detection of full-system
idle state.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
[ paulmck: Update help text per Frederic Weisbecker. ]
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
Now that the full dynticks subsystem only enables the context tracking
on full dynticks CPUs, lets remove the dependency on CONTEXT_TRACKING_FORCE
This dependency was a hack to enable the context tracking widely for the
full dynticks susbsystem until the latter becomes able to enable it in a
more CPU-finegrained fashion.
Now CONTEXT_TRACKING_FORCE only stands for testing on archs that
work on support for the context tracking while full dynticks can't be
used yet due to unmet dependencies. It simulates a system where all CPUs
are full dynticks so that RCU user extended quiescent states and dynticks
cputime accounting can be tested on the given arch.
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Li Zhong <zhong@linux.vnet.ibm.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Kevin Hilman <khilman@linaro.org>
Pull timer fixes from Thomas Gleixner:
- Cure for not using zalloc in the first place, which leads to random
crashes with CPUMASK_OFF_STACK.
- Revert a user space visible change which broke udev
- Add a missing cpu_online early return introduced by the new full
dyntick conversions
- Plug a long standing race in the timer wheel cpu hotplug code.
Sigh...
- Cleanup NOHZ per cpu data on cpu down to prevent stale data on cpu
up.
* 'timers-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
time: Revert ALWAYS_USE_PERSISTENT_CLOCK compile time optimizaitons
timer: Don't reinitialize the cpu base lock during CPU_UP_PREPARE
tick: Don't invoke tick_nohz_stop_sched_tick() if the cpu is offline
tick: Cleanup NOHZ per cpu data on cpu down
tick: Use zalloc_cpumask_var for allocating offstack cpumasks
Kay Sievers noted that the ALWAYS_USE_PERSISTENT_CLOCK config,
which enables some minor compile time optimization to avoid
uncessary code in mostly the suspend/resume path could cause
problems for userland.
In particular, the dependency for RTC_HCTOSYS on
!ALWAYS_USE_PERSISTENT_CLOCK, which avoids setting the time
twice and simplifies suspend/resume, has the side effect
of causing the /sys/class/rtc/rtcN/hctosys flag to always be
zero, and this flag is commonly used by udev to setup the
/dev/rtc symlink to /dev/rtcN, which can cause pain for
older applications.
While the udev rules could use some work to be less fragile,
breaking userland should strongly be avoided. Additionally
the compile time optimizations are fairly minor, and the code
being optimized is likely to be reworked in the future, so
lets revert this change.
Reported-by: Kay Sievers <kay@vrfy.org>
Signed-off-by: John Stultz <john.stultz@linaro.org>
Cc: stable <stable@vger.kernel.org> #3.9
Cc: Feng Tang <feng.tang@intel.com>
Cc: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
Link: http://lkml.kernel.org/r/1366828376-18124-1-git-send-email-john.stultz@linaro.org
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Turn the full dynticks passive dependency on VIRT_CPU_ACCOUNTING_GEN
to an active one.
The full dynticks Kconfig is currently hidden behind the full dynticks
cputime accounting, which is an awkward and counter-intuitive layout:
the user first has to select the dynticks cputime accounting in order
to make the full dynticks feature to be visible.
We definetly want it the other way around. The usual way to perform
this kind of active dependency is use "select" on the depended target.
Now we can't use the Kconfig "select" instruction when the target is
a "choice".
So this patch inspires on how the RCU subsystem Kconfig interact
with its dependencies on SMP and PREEMPT: we make sure that cputime
accounting can't propose another option than VIRT_CPU_ACCOUNTING_GEN
when NO_HZ_FULL is selected by using the right "depends on" instruction
for each cputime accounting choices.
v2: Keep full dynticks cputime accounting available even without
full dynticks, as per Paul McKenney's suggestion.
Reported-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Christoph Lameter <cl@linux.com>
Cc: Hakan Akkan <hakanakkan@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Kevin Hilman <khilman@linaro.org>
Cc: Li Zhong <zhong@linux.vnet.ibm.com>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Paul Gortmaker <paul.gortmaker@windriver.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
