Pull timers update for v5.16 from Daniel Lezcano:

- Fix redefined macro in the arc timer ()

- Big cleanup for ARM arch timer clocksource in order to set the scene
  for ARMv8.6 and provide support for higher frequencies with longer
  roll up (Marc Zyngier)

- Make arch dependant the Exynos MCT and Samsung PWM timers (Krzysztof
  Kozlowski)

- Select the TIMER_OF option for the timer TI DM (Kees Cook)

Link: https://lore.kernel.org/r/65693aaf-ab94-c9bb-a97b-a2bb77033a54@linaro.orgSigned-off-by: Borislav Petkov <bp@suse.de>

When building OMAP_DM_TIMER without TIMER_OF, there are orphan sections
due to the use of TIMER_OF_DELCARE() without CONFIG_TIMER_OF. Select
CONFIG_TIMER_OF when enaling OMAP_DM_TIMER:

arm-linux-gnueabi-ld: warning: orphan section `__timer_of_table' from `drivers/clocksource/timer-ti-dm-systimer.o' being placed in section `__timer_of_table'
Reported-by: kernel test robot <lkp@intel.com>
Link: https://lore.kernel.org/lkml/202108282255.tkdt4ani-lkp@intel.com/
Cc: Tony Lindgren <tony@atomide.com>
Cc: Daniel Lezcano <daniel.lezcano@linaro.org>
Cc: Keerthy <j-keerthy@ti.com>
Cc: Sebastian Reichel <sebastian.reichel@collabora.co.uk>
Cc: Ladislav Michl <ladis@linux-mips.org>
Cc: Grygorii Strashko <grygorii.strashko@ti.com>
Cc: linux-omap@vger.kernel.org
Fixes: 52762fbd ("clocksource/drivers/timer-ti-dm: Add clockevent and clocksource support")
Signed-off-by: Kees Cook <keescook@chromium.org>
Acked-by: Tony Lindgren <tony@atomide.com>
Link: https://lore.kernel.org/r/20210828175747.3777891-1-keescook@chromium.orgSigned-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>

The Exynos MCT and Samsung PWM Timer clocksource drivers are not usable
on anything else than Samsung Exynos, S3C or S5P SoC platforms.  These
are integral parts of a SoC.  Even though the drivers are not user
selectable, still document the hardware architecture explicitly with
depends on ARCH_EXYNOS and others.  This also serves a purpose of
documenting use-case, if someone ever wonders whether to select the
driver for his platform.  No functional change, because drivers are
already selected by the platform described in depends. We follow similar
approach also for other SoC-specific drivers.
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20211021063500.39314-1-krzysztof.kozlowski@canonical.comSigned-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>

The branch is a stable branch shared with ARM maintainers for the
first 13th patches of the series:

It is based on v5.14-rc3.

As stated by the changelog:

" [... ] enabling ARMv8.6 support for timer subsystem, and was prompted by a
discussion with Oliver around the fact that an ARMv8.6 implementation
must have a 1GHz counter, which leads to a number of things to break
in the timer code:

- the counter rollover can come pretty quickly as we only advertise a
  56bit counter,
- the maximum timer delta can be remarkably small, as we use the
  countdown interface which is limited to 32bit...

Thankfully, there is a way out: we can compute the minimal width of
the counter based on the guarantees that the architecture gives us,
and we can use the 64bit comparator interface instead of the countdown
to program the timer.

Finally, we start making use of the ARMv8.6 ECV features by switching
accesses to the counters to a self-synchronising register, removing
the need for an ISB. Hopefully, implementations will *not* just stick
an invisible ISB there...

A side effect of the switch to CVAL is that XGene-1 breaks. I have
added a workaround to keep it alive.

I have added Oliver's original patch[0] to the series and tweaked a
couple of things. Blame me if I broke anything.

The whole things has been tested on Juno (sysreg + MMIO timers),
XGene-1 (broken sysreg timers), FVP (FEAT_ECV, CNT*CTSS_EL0).
"

Link: https://lore.kernel.org/r/20211017124225.3018098-1-maz@kernel.orgSigned-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>

We currently handle synchronisation when workarounds are enabled
by having an ISB in the __arch_counter_get_cnt?ct_stable() helpers.

While this works, this prevents us from relaxing this synchronisation.

Instead, move it closer to the point where the synchronisation is
actually needed. Further patches will subsequently relax this.
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20211017124225.3018098-14-maz@kernel.orgSigned-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>

Unfortunately, the architecture provides no means to determine the bit
width of the system counter. However, we do know the following from the
specification:

 - the system counter is at least 56 bits wide
 - Roll-over time of not less than 40 years

To date, the arch timer driver has depended on the first property,
assuming any system counter to be 56 bits wide and masking off the rest.
However, combining a narrow clocksource mask with a high frequency
counter could result in prematurely wrapping the system counter by a
significant margin. For example, a 56 bit wide, 1GHz system counter
would wrap in a mere 2.28 years!

This is a problem for two reasons: v8.6+ implementations are required to
provide a 64 bit, 1GHz system counter. Furthermore, before v8.6,
implementers may select a counter frequency of their choosing.

Fix the issue by deriving a valid clock mask based on the second
property from above. Set the floor at 56 bits, since we know no system
counter is narrower than that.

[maz: fixed width computation not to lose the last bit, added
      max delta generation for the timer]
Suggested-by: Marc Zyngier <maz@kernel.org>
Signed-off-by: Oliver Upton <oupton@google.com>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210807191428.3488948-1-oupton@google.com
Link: https://lore.kernel.org/r/20211017124225.3018098-13-maz@kernel.orgSigned-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>

Switching from TVAL to CVAL has a small drawback: we need an ISB
before reading the counter. We cannot get rid of it, but we can
instead remove the one that comes just after writing to CVAL.

This reduces the number of ISBs from 3 to 2 when programming
the timer.
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20211017124225.3018098-12-maz@kernel.orgSigned-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>

TVAL usage is now long gone, get rid of the leftovers.
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20211017124225.3018098-11-maz@kernel.orgSigned-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>

The Applied Micro XGene-1 SoC has a busted implementation of the
CVAL register: it looks like it is based on TVAL instead of the
other way around. The net effect of this implementation blunder
is that the maximum deadline you can program in the timer is
32bit wide.

Use a MIDR check to notice the broken CPU, and reduce the width
of the timer to 32bit.
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20211017124225.3018098-10-maz@kernel.orgSigned-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>

Proudly tell the code code that we have a timer able to handle
56 bits deltas.
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20211017124225.3018098-9-maz@kernel.orgSigned-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>

Similarily to the sysreg-based timer, move the MMIO over to using
the CVAL registers instead of TVAL. Note that there is no warranty
that the 64bit MMIO access will be atomic, but the timer is always
disabled at the point where we program CVAL.
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20211017124225.3018098-8-maz@kernel.orgSigned-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>

The MMIO timer base address gets published after we have registered
the callbacks and the interrupt handler, which is... a bit dangerous.

Fix this by moving the base address publication to the point where
we register the timer, and expose a pointer to the timer structure
itself rather than a naked value.
Reviewed-by: Oliver Upton <oupton@google.com>
Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20211017124225.3018098-7-maz@kernel.orgSigned-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>

The '_tval' name in the erratum handling function names doesn't
make much sense anymore (and they were using CVAL the first place).

Drop the _tval tag.
Reviewed-by: Oliver Upton <oupton@google.com>
Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Tested-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20211017124225.3018098-6-maz@kernel.orgSigned-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>

In order to cope better with high frequency counters, move the
programming of the timers from the countdown timer (TVAL) over
to the comparator (CVAL).

The programming model is slightly different, as we now need to
read the current counter value to have an absolute deadline
instead of a relative one.

There is a small overhead to this change, which we will address
in the following patches.
Reviewed-by: Oliver Upton <oupton@google.com>
Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Tested-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20211017124225.3018098-5-maz@kernel.orgSigned-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>

The various accessors for the timer sysreg and MMIO registers are
currently hardwired to 32bit. However, we are about to introduce
the use of the CVAL registers, which require a 64bit access.

Upgrade the write side of the accessors to take a 64bit value
(the read side is left untouched as we don't plan to ever read
back any of these registers).

No functional change expected.
Reviewed-by: Oliver Upton <oupton@google.com>
Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Tested-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20211017124225.3018098-4-maz@kernel.orgSigned-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>

The arch timer driver never reads the various TVAL registers, only
writes to them. It is thus pointless to provide accessors
for them and to implement errata workarounds.

Drop these read-side accessors, and add a couple of BUG() statements
for the time being. These statements will be removed further down
the line.
Reviewed-by: Oliver Upton <oupton@google.com>
Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Tested-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20211017124225.3018098-3-maz@kernel.orgSigned-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>

As we are about to change the registers that are used by the driver,
start by adding build-time checks to ensure that we always handle
all registers and access modes.
Suggested-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20211017124225.3018098-2-maz@kernel.orgSigned-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>

In drivers/clocksource/, 3 drivers use "TIMER_CTRL_IE" with 3 different
values.  Two of them (mps2-timer.c and timer-sp804.c/timer-sp.h) are
localized and left unmodifed.

One of them uses a shared header file (<soc/arc/timers.h>), which is
what is causing the "redefined" warnings, so change the macro name in
that driver only. Also change the TIMER_CTRL_NH macro name.
Both macro names are prefixed with "ARC_" to reduce the likelihood
of future name collisions.

In file included from ../drivers/clocksource/timer-sp804.c:24:
../drivers/clocksource/timer-sp.h:25: error: "TIMER_CTRL_IE" redefined [-Werror]
   25 | #define TIMER_CTRL_IE           (1 << 5)        /*   VR */
../include/soc/arc/timers.h:20: note: this is the location of the previous definition
   20 | #define TIMER_CTRL_IE           (1 << 0) /* Interrupt when Count reaches limit */

Fixes: b26c2e38 ("ARC: breakout timer include code into separate header")
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Cc: Vineet Gupta <vgupta@kernel.org>
Cc: linux-snps-arc@lists.infradead.org
Cc: Daniel Lezcano <daniel.lezcano@linaro.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Shahab Vahedi <Shahab.Vahedi@synopsys.com>
Acked-by: Vineet Gupta <vgupta@kernel.org>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20210924020825.20317-1-rdunlap@infradead.org

The clocksource watchdog test sets a local JIFFIES_SHIFT macro and assumes
that HZ is >= 100. For smaller HZ values this shift value is too large and
causes undefined behaviour.

Move the HZ-based definitions of JIFFIES_SHIFT from kernel/time/jiffies.c
to kernel/time/tick-internal.h so the clocksource watchdog test can utilize
them, which makes it work correctly with all HZ values.

[ tglx: Resolved conflicts and massaged changelog ]
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/lkml/20210812000133.GA402890@paulmck-ThinkPad-P17-Gen-1/

Pull timer driver updates from Daniel Lezcano:

 - Prioritize the ARM architected timer on Exynos platform when the
   architecture is ARM64 (Will Deacon)

 - Mark the Exynos timer as a per CPU timer (Will Deacon)

 - DT conversion to yaml for the rockchip platform (Ezequiel Garcia)

 - Fix IRQ setup if there are two channels on the sh_cmt timer (Phong
   Hoang)

 - Use bitfield helper macros in the Ingenic timer (Zhou Yanjie)

 - Clear any pending interrupt to prevent an abort of the suspend on
   the Mediatek platform (Fengquan Chen)

 - Add DT bindings for new Ingenic SoCs (Zhou Yanjie)

Link: https://lore.kernel.org/r/c14ad27a-b1c6-6043-0f5e-71dd984bb4ba@linaro.org

1.Add OST_CLK_EVENT_TIMER for new XBurst®1 SoCs.
2.Add OST_CLK_EVENT_TIMER0 to OST_CLK_EVENT_TIMER15 for new XBurst®2 SoCs.
Signed-off-by: 周琰杰 (Zhou Yanjie) <zhouyanjie@wanyeetech.com>
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/1626370605-120775-1-git-send-email-zhouyanjie@wanyeetech.com

Just pass bool flags from the different initcalls and use the
flags to set the right pointers. This results in less pointers
passed around in init.

Cc: Cédric Le Goater <clg@kaod.org>
Cc: Joel Stanley <joel@jms.id.au>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20210724224424.2085404-1-linus.walleij@linaro.org

mtk_syst_clkevt_shutdown is called after irq disabled in suspend flow,
clear any pending systimer irq when shutdown to avoid suspend aborted
due to timer irq pending

Also as for systimer in mediatek socs, there must be firstly enable
timer before clear systimer irq

Fixes: e3af6776("clocksource/drivers/timer-mediatek: Add support for system timer")
Signed-off-by: Fengquan Chen <fengquan.chen@mediatek.com>
Tested-by: Hsin-Yi Wang <hsinyi@chromium.org>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/1617960162-1988-2-git-send-email-Fengquan.Chen@mediatek.com

Use "FIELD_GET()" and "FIELD_PREP()" to simplify the code.

[dlezcano] : Changed title
Signed-off-by: 周琰杰 (Zhou Yanjie) <zhouyanjie@wanyeetech.com>
Reviewed-by: Paul Cercueil <paul@crapouillou.net>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/1627638188-116163-1-git-send-email-zhouyanjie@wanyeetech.com

If CMT instance has at least two channels, one channel will be used
as a clock source and another one used as a clock event device.
In that case, IRQ is not requested for clock source channel so
sh_cmt_clock_event_program_verify() might work incorrectly.
Besides, when a channel is only used for clock source, don't need to
re-set the next match_value since it should be maximum timeout as
it still is.

On the other hand, due to no IRQ, total_cycles is not counted up
when reaches compare match time (timer counter resets to zero),
so sh_cmt_clocksource_read() returns unexpected value.
Therefore, use 64-bit clocksoure's mask for 32-bit or 16-bit variants
will also lead to wrong delta calculation. Hence, this mask should
correspond to timer counter width, and above function just returns
the raw value of timer counter register.

Fixes: bfa76bb1 ("clocksource: sh_cmt: Request IRQ for clock event device only")
Fixes: 37e7742c ("clocksource/drivers/sh_cmt: Fix clocksource width for 32-bit machines")
Signed-off-by: Phong Hoang <phong.hoang.wz@renesas.com>
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20210422123443.73334-1-niklas.soderlund+renesas@ragnatech.se

Convert Rockchip Timer dt-bindings to YAML.
Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20210506111136.3941-4-ezequiel@collabora.com

The "mct_tick" is a per-cpu clockevents device. Set the
CLOCK_EVT_FEAT_PERCPU feature to prevent e.g. mct_tick0 being unsafely
designated as the global broadcast timer and instead treat the device as
a per-cpu wakeup timer.

Cc: Daniel Lezcano <daniel.lezcano@linaro.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Signed-off-by: Will Deacon <will@kernel.org>
Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Reviewed-by: Chanwoo Choi <cw00.choi@samsung.com>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20210608154341.10794-3-will@kernel.org

All arm64 CPUs feature an architected timer, which offers a relatively
low-latency interface to a per-cpu clocksource and timer. For the most
part, using this interface is a no-brainer, with the exception of SoCs
where it cannot be used to wake up from deep idle state (i.e.
CLOCK_EVT_FEAT_C3STOP is set).

On the contrary, the Exynos MCT is extremely slow to access yet can be
used as a wakeup source. In preparation for using the Exynos MCT as a
potential wakeup timer for the Arm architected timer, reduce its ratings
so that the architected timer is preferred.

This effectively reverts the decision made in 6282edb7
("clocksource/drivers/exynos_mct: Increase priority over ARM arch timer")
for arm64, as the reasoning for the original change was to work around
a 32-bit SoC design.

Cc: Marek Szyprowski <m.szyprowski@samsung.com>
Cc: Krzysztof Kozlowski <krzk@kernel.org>
Cc: Chanwoo Choi <cw00.choi@samsung.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Will Deacon <will@kernel.org>
Tested-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com> # exynos-5422
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20210608154341.10794-2-will@kernel.org

Since the recent consoliation of reprogramming functions,
hrtimer_force_reprogram() is affected by a check whether the new expiry
time is past the current expiry time.

This breaks the NOHZ logic as that relies on the fact that the tick hrtimer
is moved into the future. That means cpu_base->expires_next becomes stale
and subsequent reprogramming attempts fail as well until the situation is
cleaned up by an hrtimer interrupts.

For some yet unknown reason this leads to a complete stall, so for now
partially revert the offending commit to a known working state. The root
cause for the stall is still investigated and will be fixed in a subsequent
commit.

Fixes: b14bca97 ("hrtimer: Consolidate reprogramming code")
Reported-by: Mike Galbraith <efault@gmx.de>
Reported-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Mike Galbraith <efault@gmx.de>
Link: https://lore.kernel.org/r/8735recskh.ffs@tglx

clock_was_set() can be invoked from preemptible context. Use raw_cpu_ptr()
to check whether high resolution mode is active or not. It does not matter
whether the task migrates after acquiring the pointer.

Fixes: e71a4153 ("hrtimer: Force clock_was_set() handling for the HIGHRES=n, NOHZ=y case")
Reported-by: Mike Galbraith <efault@gmx.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/875ywacsmb.ffs@tglx

By unconditionally updating the offsets there are more indicators
whether the SMP function calls on clock_was_set() can be avoided:

  - When the offset update already happened on the remote CPU then the
    remote update attempt will yield the same seqeuence number and no
    IPI is required.

  - When the remote CPU is currently handling hrtimer_interrupt(). In
    that case the remote CPU will reevaluate the timer bases before
    reprogramming anyway, so nothing to do.

  - After updating it can be checked whether the first expiring timer in
    the affected clock bases moves before the first expiring (softirq)
    timer of the CPU. If that's not the case then sending the IPI is not
    required.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20210713135158.887322464@linutronix.de

Setting of clocks triggers an unconditional SMP function call on all online
CPUs to reprogram the clock event device.

However, only some clocks have their offsets updated and therefore
potentially require a reprogram. That's CLOCK_REALTIME and CLOCK_TAI and in
the case of resume (delayed sleep time injection) also CLOCK_BOOTTIME.

Instead of sending an IPI unconditionally, check each per CPU hrtimer base
whether it has active timers in the affected clock bases which are
indicated by the caller in the @bases argument of clock_was_set().

If that's not the case, skip the IPI and update the offsets remotely which
ensures that any subsequently armed timers on the affected clocks are
evaluated with the correct offsets.

[ tglx: Adopted to the new bases argument, removed the softirq_active
  	check, added comment, fixed up stale comment ]
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20210713135158.787536542@linutronix.de

clock_was_set() unconditionaly invokes retrigger_next_event() on all online
CPUs. This was necessary because that mechanism was also used for resume
from suspend to idle which is not longer the case.

The bases arguments allows the callers of clock_was_set() to hand in a mask
which tells clock_was_set() which of the hrtimer clock bases are affected
by the clock setting. This mask will be used in the next step to check
whether a CPU base has timers queued on a clock base affected by the event
and avoid the SMP function call if there are none.

Add a @bases argument, provide defines for the active bases masking and
fixup all callsites.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20210713135158.691083465@linutronix.de

do_adjtimex() might end up scheduling a delayed clock_was_set() via
timekeeping_advance() and then invoke clock_was_set() directly which is
pointless.

Make timekeeping_advance() return whether an invocation of clock_was_set()
is required and handle it at the call sites which allows do_adjtimex() to
issue a single direct call if required.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20210713135158.580966888@linutronix.de

Resuming timekeeping is a clock-was-set event and uses the clock-was-set
notification mechanism. This is in the way of making the clock-was-set
update for hrtimers selective so unnecessary IPIs are avoided when a CPU
base does not have timers queued which are affected by the clock setting.

Distangle it by invoking hrtimer_resume() on each unfreezing CPU and invoke
the new timerfd_resume() function from timekeeping_resume() which is the
only place where this is needed.

Rename hrtimer_resume() to hrtimer_resume_local() to reflect the change.

With this the clock_was_set*() functions are not longer required to IPI all
CPUs unconditionally and can get some smarts to avoid them.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20210713135158.488853478@linutronix.de

Resuming timekeeping is a clock-was-set event and uses the clock-was-set
notification mechanism. This is in the way of making the clock-was-set
update for hrtimers selective so unnecessary IPIs are avoided when a CPU
base does not have timers queued which are affected by the clock setting.

Provide a seperate timerfd_resume() interface so the resume logic and the
clock-was-set mechanism can be distangled in the core code.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20210713135158.395287410@linutronix.de

When CONFIG_HIGH_RES_TIMERS is disabled, but NOHZ is enabled then
clock_was_set() is not doing anything. With HIGHRES=n the kernel relies on
the periodic tick to update the clock offsets, but when NOHZ is enabled and
active then CPUs which are in a deep idle sleep do not have a periodic tick
which means the expiry of timers affected by clock_was_set() can be
arbitrarily delayed up to the point where the CPUs are brought out of idle
again.

Make the clock_was_set() logic unconditionaly available so that idle CPUs
are kicked out of idle to handle the update.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20210713135158.288697903@linutronix.de

If high resolution timers are disabled the timerfd notification about a
clock was set event is not happening for all cases which use
clock_was_set_delayed() because that's a NOP for HIGHRES=n, which is wrong.

Make clock_was_set_delayed() unconditially available to fix that.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20210713135158.196661266@linutronix.de

This code is mostly duplicated. The redudant store in the force reprogram
case does no harm and the in hrtimer interrupt condition cannot be true for
the force reprogram invocations.
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20210713135158.054424875@linutronix.de

If __hrtimer_start_range_ns() is invoked with an already armed hrtimer then
the timer has to be canceled first and then added back. If the timer is the
first expiring timer then on removal the clockevent device is reprogrammed
to the next expiring timer to avoid that the pending expiry fires needlessly.

If the new expiry time ends up to be the first expiry again then the clock
event device has to reprogrammed again.

Avoid this by checking whether the timer is the first to expire and in that
case, keep the timer on the current CPU and delay the reprogramming up to
the point where the timer has been enqueued again.
Reported-by: Lorenzo Colitti <lorenzo@google.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20210713135157.873137732@linutronix.de