There was an inconsistency in the API as z_nrf_rtc_timer_chan_alloc
returned int but other function were using uint32_t for channel
argument. Updated api to use int32_t everywhere.
Update nrf_802154 driver which was using this api to use int32_t.
Signed-off-by: Krzysztof Chruscinski <krzysztof.chruscinski@nordicsemi.no>
Add initial support for the Cortex-M55 Core which is an implementation
of the Armv8.1-M mainline architecture and includes support for the
M‑profile Vector Extension (MVE).
The support is based on the Cortex-M33 support that already exists in
Zephyr.
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
This feature predated the tickless kernel and has been in legacy mode
for a while. We now have no drivers or systems that do not support
tickless, so remove this option and cleanup the code to only use
tickless.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
This is another API that is being used in all timer drivers and is not
internal to the clock subsystem. Remove the leading z_ and make promote
it to a cross-subsystem API.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
The clock/timer APIs are not application facing APIs, however, similar
to arch_ and a few other APIs they are available to implement drivers
and add support for new hardware and are documented and available to be
used outside of the clock/kernel subsystems.
Remove the leading z_ and provide them as clock_* APIs for someone
writing a new timer driver to use.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
There was a bunch of dead historical cruft floating around in the
arch/xtensa tree, left over from older code versions. It's time to do
a cleanup pass. This is entirely refactoring and size optimization,
no behavior changes on any in-tree devices should be present.
Among the more notable changes:
+ xtensa_context.h offered an elaborate API to deal with a stack frame
and context layout that we no longer use.
+ xtensa_rtos.h was entirely dead code
+ xtensa_timer.h was a parallel abstraction layer implementing in the
architecture layer what we're already doing in our timer driver.
+ The architecture thread structs (_callee_saved and _thread_arch)
aren't used by current code, and had dead fields that were removed.
Unfortunately for standards compliance and C++ compatibility it's
not possible to leave an empty struct here, so they have a single
byte field.
+ xtensa_api.h was really just some interrupt management inlines used
by irq.h, so fold that code into the outer header.
+ Remove the stale assembly offsets. This architecture doesn't use
that facility.
All told, more than a thousand lines have been removed. Not bad.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
If next_timeout() returns INT_MAX and pass it to
z_clock_set_timeout(), and machine goes to freeze or slow down.
Bad scenario as follows:
- If an argument int32_t ticks is set large value 0xffffffff,
ticks = MAX(MIN(ticks - 1, (int32_t)MAX_TICKS), 0);
replaces it into MAX_TICKS.
- uint32_t cyc will be set near by 0xffffffff
(this is 0xfffd7280 in 100 ticks per second).
- Add adjustment to cyc, adjustment max value is MAX_CYC.
(cyc = 0xffff14fd)
- Over 0x80000000 value of uint32_t is considered as negative
value of int32_t.
if ((int32_t)(cyc + last_count - now) < MIN_DELAY)
This condition is always true.
- Because cyc += CYC_PER_TICK will get overflow, driver sets mtimecmp
near value of current mtime.
(cyc = 0x00007fc0)
- Next timer interrupt will happen soon after return from interrupt
handler.
- By repeating these events, machine cannot go to next instruction,
and it's going to freeze or slow down.
Signed-off-by: Katsuhiro Suzuki <katsuhiro@katsuster.net>
Use timeout mechanism instead of unbounded loop during enabling ITIM32
module which source clock is LFCLK (32KHz).
Signed-off-by: Mulin Chao <mlchao@nuvoton.com>
Qemu when running more than one processor has a known synchronization
bug where counter values read from the HPET (notionally a single
global device) can be seen going "backwards" when read from different
CPUs.
There was a pre-existing workaround in the ISR that knew about this,
but the problem can crop up anywhere the counter value is used. In
particular I caught it aliasing with the "max_ticks" computation in
z_clock_set_timeout(), where it would cause a rollover and the
resulting negative comparator value would result in no end of
hilarity.
Wrap all access to the counter register with a counter() inline that
(when the workaround is enabled) forces the result to be monotonic by
clamping it to a minimum of one more than the previously read value.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
This CL introduces power management driver that improves the efficiency
of ec operation by adjusting the chip’s power consumption to the level
of activity required by the application in npcx series.
The following list summarizes the main properties of the various chip
power states. Please refer the power.c file for more detail.
Main power states in npcx series include:
- Active: Core, RAM and modules operate at the clocks generated by PLL.
- Idle: Enter this state when the Core executes WFI or WFE instruction.
- Sleep: clock is stopped for most of modules but PLL is enabled.
- Deep Sleep: As Sleep mode but PLL is disabled.
- Standby: All power rails are turned off besides standby and battery
power rails.
And this CL implements one power state, PM_STATE_SUSPEND_TO_IDLE, with
two sub-states for Zephyr power management system.
Sub-state 0 - "Deep Sleep" mode with “Instant” wake-up if residency
time is greater or equal to 1 ms
Sub-state 1 - "Deep Sleep" mode with "Standard" wake-up if residency
time is greater or equal to 201 ms
Signed-off-by: Mulin Chao <mlchao@nuvoton.com>
This CL introduces a kernel device driver implemented by the internal
64/32-bit timers in Nuvoton NPCX series. Via these two kinds of timer,
the driver provides an standard "system clock driver" interface.
It includes:
- A system timer based on an ITIM64 (Internal 64-bit timer) instance,
clocked by APB2 which freq is CONFIG_SYS_CLOCK_HW_CYCLES_PER_SEC.
- Its prescaler is set to 1 and provide the kernel cycles reading
without handling overflow mechanism.
- A event timer based on an ITIM32 (Internal 32-bit timer) instance,
clocked by LCLK which frequency is 32KHz and still activated when ec
entered "idle/deep idle" power state for better power consumption.
- Its prescaler is set to 1 and provide timeout event mechansim.
- Compensate system timer which clock is gating for better power
consumption after ec left"idle/deep idle" power state.
This CL passed starve, timer_api, and timer_monotonic test suites.
Signed-off-by: Mulin Chao <mlchao@nuvoton.com>
This include make possible to use the arm_arch_timer on
platform such as Cortex-A9 or Cortex-R7 which has support for
ARM Global Timer.
The global timer is a 64 bit incrementing counter, memory-mapped
in the private memory region.
Signed-off-by: Julien Massot <julien.massot@iot.bzh>
The count register is 64 bits, but we're a 32 bit CPU that can only
read four bytes at a time, so a bit of care is needed to prevent
racing against a wraparound of the low word. Wrap the low read
between two reads of the high word and make sure it didn't change.
Fixes#31599
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
idle is only considered in other timer implementations if ticks ==
K_TICKS_FOREVER but in arm_arch_timer. Just fix it.
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
TICKLESS_CAPABLE is now selectable only and without prompt, so remove it
from _defconfig files and select it directly by the timer.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
z_timer_idle_enter is declared only when CONFIG_TICKLESS_IDLE is
selected. This function is not implemented anywhere, but the only
driver including this header is not TICKLESS_CAPABLE. So, no undefined
reference will happen.
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
Due to clock discrepancy, busy waiting for 15us was not covering for
half tick in certain cases. Busy wait runs from HF clock source.
Increased to 19us to cover it. Anyway, this case is hit very rarely,
only when there was aborted, not-cancelled compare value that was
about to expire. Because of that, increase shall not impact the
performance.
Signed-off-by: Krzysztof Chruscinski <krzysztof.chruscinski@nordicsemi.no>
For a while now, we've had two APIC drivers. The older was preserved
initially as the new (much smaller, "new style") code didn't have
support for Quark interrupt handling. But that's long dead now. Just
remove it.
Note that this migrates the one board using this driver (acrn) to
CONFIG_APIC_TIMER instead.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
As timer interrupt is level triggered, we need to mask it before leaving
ISR or it will be delivered again.
Also, Xen automatically masks timer interrupt when it injects IRQ to
a guest, so we need to unmask it again, when setting new timeout.
Signed-off-by: Volodymyr Babchuk <volodymyr_babchuk@epam.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
This commit is about the it8xxx2 timer driver.
We use the timer 5 as system timer for count time,
so the timer interrupt is trigged by it.
Signed-off-by: Cheryl Su <cheryl.su@ite.com.tw>
Add a new test for k_busy_wait and cpu_hold
Signed-off-by: Alberto Escolar Piedras <alpi@oticon.com>
Signed-off-by: Wolfgang Puffitsch <wopu@demant.com>
In native_posix and nrf52_bsim add the cpu_hold() function,
which can be used to emulate the time it takes for code
to execute.
It is very similar to arch_busy_wait(), but while
arch_busy_wait() returns when the requested time has passed,
cpu_hold() ensures that the time passes in the callers
context independently of how much time may pass in some
other context.
Signed-off-by: Alberto Escolar Piedras <alpi@oticon.com>
When tickless mode was disable, sys clock timeout handler was calling
public API function for setting new compare value. Public API function
asserts when chan 0 is used which is reserved for system clock.
Signed-off-by: Krzysztof Chruscinski <krzysztof.chruscinski@nordicsemi.no>
Added clearing of CC event which may occure due to previous
CC value which was closed to current counter value.
Fixed int_mask initialization.
Signed-off-by: Krzysztof Chruscinski <krzysztof.chruscinski@nordicsemi.no>
Fix TIMER0 and RTC0 being selectable when using out-of-tree Bluetooth
controller.
Generalize the Kconfig to have the features that use the HW peripheral
select them as reserved to make the dependencies more manageable.
Signed-off-by: Joakim Andersson <joakim.andersson@nordicsemi.no>
- Remove SYS_ prefix
- shorten POWER_MANAGEMENT to just PM
- DEVICE_POWER_MANAGEMENT -> PM_DEVICE
and use PM_ as the prefix for all PM related Kconfigs
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Extended nrf_rtc_timer driver to expose API for using RTC for
other purposes. System timer is using one compare channels,
other channels may be used through this API.
Signed-off-by: Krzysztof Chruscinski <krzysztof.chruscinski@nordicsemi.no>
This adds support for GRLIB GPTIMER general purpose timer used in
LEON3/4/5 systems.
One of the GPTIMER subtimers is used to generate periodic interrutps
for announcing ticks. Another subtimer is used as upcounter for the
cycle_get_32() service.
Signed-off-by: Martin Åberg <martin.aberg@gaisler.com>
Replaces all existing variants of value clamping with the MIN and MAX
macros with the CLAMP macro.
Signed-off-by: Trond Einar Snekvik <Trond.Einar.Snekvik@nordicsemi.no>
The HDA wall clock timer is a 64 bit timer with 64 bit compare
registers, but it's being used from a 32 bit CPU. Writing the
comparator piecewise with a 64 bit C assignment will write the low
dword first, opening the possibility that the hardware will see time
go "backwards" and trigger an interrupt incorrectly.
Disable the enable bit while setting the comparator.
Found by inspection. In practice this will be very rare, and spurious
timer interrupts are supposed to be benign anyway (though they can
result in timeout expirations being misaligned to ticks, which might
be surprising to applications). Best to get it right.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
