This commit introduces the following changes:
* nrf_rtc_timer is extended with a capability to handle RTC overflow,
allowing it to operate on absolute RTC ticks, rather than relative
ticks.
* overflow handling is ZLI-proof and relies on the sys clock
handler being executed twice every RTC counter's overflow.
* callbacks are given an absolute RTC tick value as a parameter instead
of CC register's value. The absolute RTC tick value is the RTC counter
value set during CC channel configuration extended to 64 bits.
* in case the timer's target time is in the past or is the current tick,
the timer fires as soon as possible, however still from the RTC's ISR
context.
* in case an active timer is set again with the same target time, it is
not scheduled again - only its event data is updated. Otherwise, the
timer is scheduled as usual.
* a scheduled timer can be aborted.
* system clock functions are now using 64 bit values internally.
Signed-off-by: Andrzej Kuroś <andrzej.kuros@nordicsemi.no>
Signed-off-by: Jedrzej Ciupis <jedrzej.ciupis@nordicsemi.no>
Signed-off-by: Krzysztof Chruscinski <krzysztof.chruscinski@nordicsemi.no>
Signed-off-by: Paweł Kwiek <pawel.kwiek@nordicsemi.no>
This commit introduces the following changes:
* nrf_rtc_timer is extended with a capability to handle RTC overflow,
allowing it to operate on absolute RTC ticks, rather than relative
ticks.
* overflow handling is ZLI-proof and relies on the sys clock
handler being executed twice every RTC counter's overflow.
* callbacks are given an absolute RTC tick value as a parameter instead
of CC register's value. The absolute RTC tick value is the RTC counter
value set during CC channel configuration extended to 64 bits.
* in case the timer's target time is in the past or is the current tick,
the timer fires as soon as possible, however still from the RTC's ISR
context.
* in case an active timer is set again with the same target time, it is
not scheduled again - only its event data is updated. Otherwise, the
timer is scheduled as usual.
* a scheduled timer can be aborted.
* system clock functions are now using 64 bit values internally.
Signed-off-by: Andrzej Kuroś <andrzej.kuros@nordicsemi.no>
Signed-off-by: Jedrzej Ciupis <jedrzej.ciupis@nordicsemi.no>
Signed-off-by: Krzysztof Chruscinski <krzysztof.chruscinski@nordicsemi.no>
Signed-off-by: Paweł Kwiek <pawel.kwiek@nordicsemi.no>
The weak symbol sys_clock_driver_init has been removed, therefore moving
the init responsability to the drivers themselves. As a result, the init
function has now been made static on all drivers and moved to the
bottom, following the convention used in other areas.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
RTC interrupt was reading CC value and passing it to the handler.
However, higher priority interrupt could preempt RTC interrupt
and set new CC value. In that case CC value read in the RTC
interrupt context was not the one that triggered the interrupt.
Added fallback to COUNTER value if that case is detected.
Using COUNTER is not as precise as CC because it returns time
when event was handled and not when event occured but it is the
only option since CC value is overwritten.
Signed-off-by: Krzysztof Chruscinski <krzysztof.chruscinski@nordicsemi.no>
As per description of the sys_clock_elapsed() function, "the kernel
will call this with appropriate locking, the driver needs only provide
an instantaneous answer". Remove then the unnecessary locking from the
function, as it only adds an undesirable delay.
Signed-off-by: Andrzej Głąbek <andrzej.glabek@nordicsemi.no>
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>
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>
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>
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>
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>
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>
Added configuration for approach to starting system clock source.
There are 3 options: no wait, wait untill available, wait until
stable.
Added support for those modes in clock control driver which handles
low frequency source clock.
Signed-off-by: Krzysztof Chruscinski <krzysztof.chruscinski@nordicsemi.no>
Now that device_api attribute is unmodified at runtime, as well as all
the other attributes, it is possible to switch all device driver
instance to be constant.
A coccinelle rule is used for this:
@r_const_dev_1
disable optional_qualifier
@
@@
-struct device *
+const struct device *
@r_const_dev_2
disable optional_qualifier
@
@@
-struct device * const
+const struct device *
Fixes#27399
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
Adapted driver to clock_control changes (usage of onoff manager).
Since timer is permanenty requesting the clock, it is using API
dedicated for that: z_nrf_clock_control_lf_on().
Signed-off-by: Krzysztof Chruscinski <krzysztof.chruscinski@nordicsemi.no>
When checking the absolute value of cycles set to the comparator
use the MAX_CYCLES instead of MAX_TICKS.
The commit changes function names and comments to make it clear
where ticks (system ticks) and where RTC cycles are used.
Fixes#26701
Signed-off-by: Pawel Dunaj <pawel.dunaj@nordicsemi.no>
If timeout is being overwrite exactly when previous one is expiring
then hardware event was cleared correctly but interrupt was already
triggered. Interrupt routine was assuming that compare event is set
and proceed with that assumption. However, in that corner case when
compare event was overwritten and event was cleared, that was not the
case.
As the outcome, timeout could be triggered prematurely. Fixed by
clearing pending interrupt after handling previous compare value.
Signed-off-by: Krzysztof Chruscinski <krzysztof.chruscinski@nordicsemi.no>
Some kernel tests use `CONFIG_TICKLESS_KERNEL=n` with
`CONFIG_SYS_CLOCK_TICKS_PER_SEC=1` to detect when a test runs longer
than 1 second. These tests break if a tick is announced every time a
timeout occurs. Only announce if the measured duration since the last
tick is at least the duration of a tick.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
User reported a flaw in the current algorithm which fails when Zero
Latency Interrupts (ZLI) are used. Ported algorithm from
counter_nrfx_rtc.c which covers all cases. Algorithm is lockless so
no distinction for ZLI is needed.
Signed-off-by: Krzysztof Chruscinski <krzysztof.chruscinski@nordicsemi.no>
Add a k_timeout_t type, and use it everywhere that kernel API
functions were accepting a millisecond timeout argument. Instead of
forcing milliseconds everywhere (which are often not integrally
representable as system ticks), do the conversion to ticks at the
point where the timeout is created. This avoids an extra unit
conversion in some application code, and allows us to express the
timeout in units other than milliseconds to achieve greater precision.
The existing K_MSEC() et. al. macros now return initializers for a
k_timeout_t.
The K_NO_WAIT and K_FOREVER constants have now become k_timeout_t
values, which means they cannot be operated on as integers.
Applications which have their own APIs that need to inspect these
vs. user-provided timeouts can now use a K_TIMEOUT_EQ() predicate to
test for equality.
Timer drivers, which receive an integer tick count in ther
z_clock_set_timeout() functions, now use the integer-valued
K_TICKS_FOREVER constant instead of K_FOREVER.
For the initial release, to preserve source compatibility, a
CONFIG_LEGACY_TIMEOUT_API kconfig is provided. When true, the
k_timeout_t will remain a compatible 32 bit value that will work with
any legacy Zephyr application.
Some subsystems present timeout (or timeout-like) values to their own
users as APIs that would re-use the kernel's own constants and
conventions. These will require some minor design work to adapt to
the new scheme (in most cases just using k_timeout_t directly in their
own API), and they have not been changed in this patch, instead
selecting CONFIG_LEGACY_TIMEOUT_API via kconfig. These subsystems
include: CAN Bus, the Microbit display driver, I2S, LoRa modem
drivers, the UART Async API, Video hardware drivers, the console
subsystem, and the network buffer abstraction.
k_sleep() now takes a k_timeout_t argument, with a k_msleep() variant
provided that works identically to the original API.
Most of the changes here are just type/configuration management and
documentation, but there are logic changes in mempool, where a loop
that used a timeout numerically has been reworked using a new
z_timeout_end_calc() predicate. Also in queue.c, a (when POLL was
enabled) a similar loop was needlessly used to try to retry the
k_poll() call after a spurious failure. But k_poll() does not fail
spuriously, so the loop was removed.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
The CONFIG_ prefixes were missing on these.
Found with a work-in-progress scripts/kconfig/lint.py check.
This symbol is defined in kernel/Kconfig.
Signed-off-by: Ulf Magnusson <Ulf.Magnusson@nordicsemi.no>
Low frequency and high frequency clocks had separate devices
while they are actually handled by single peripheral with single
interrupt. The split was done probably because opaque subsys
argument in the API was used for other purposes and there was
no way to pass the information which clock should be controlled.
Implementation changes some time ago and subsys parameter was
no longer used. It now can be used to indicate which clock should
be controlled.
Change become necessary when nrf5340 is taken into account where
there are more clocks and current approach would lead to create
multiple devices - mess.
Signed-off-by: Krzysztof Chruscinski <krzysztof.chruscinski@nordicsemi.no>
When setting a timeout measure the number of accumulated unannounced
ticks. If this value exceeds half the 24-bit cycle counter range
force an announcement so the unannounced cycles are incorporated into
the system tick counter.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
The original code assumed that limiting the tick count to the maximum
cycle value representable without wrapping would guarantee that adding
the resulting cycle offset to last_count would not lap the counter.
This is not true when elapsed time, which is also added to the cycle
offset, exceeds one tick. Cap the maximum offset at the number of
cycles corresponding to the maximum number of ticks without wrapping.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
Fix an inline comment in nrf_rtc_timer.c correcting the
path to the mentioned test.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
Header files of nrfx HALs are not supposed to be included directly
but only with their names prepended with the hal/ directory (so that
an inclusion of an nrfx HAL header clearly differs from an inclusion
of an nrfx driver header).
Signed-off-by: Andrzej Głąbek <andrzej.glabek@nordicsemi.no>
The variable enabling entry to the zero latency interrupt compensation
loop was named generically, and its logic inverted, making the code
difficult to understand. Change the name and initial value to more
clearly indicate its role.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
Update the logic in a corner case, when the target comparator value is
one cycle ahead of the counter value.
Experiments have shown, that `set_comparator(cyc + 1);` might be not
enough in that case, and we still may (rarely) miss the interrupt.
This could happen when the counter incremented its value after the `dt`
variable was set. As we should set the comparator value two cycles
ahead to be on the safe side, increment the target comparator value
by 2 instead of 1.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
On some SoCs the frequency of the system clock is obtained at run time
as the exact configuration of the hardware is not known at compile time.
On such platforms using CONFIG_SYS_CLOCK_HW_CYCLES_PER_SEC define
directly introduces timing errors.
This commit replaces CONFIG_SYS_CLOCK_HW_CYCLES_PER_SEC by the call
to inline function sys_clock_hw_cycles_per_sec() which always returns
correct frequency of the system clock.
Signed-off-by: Piotr Zięcik <piotr.ziecik@nordicsemi.no>
When the tick rate was less than MIN_DELAY, bumping a "too soon"
expiration by just one tick may not be enough and we could
theoretically miss the counter.
Instead, eliminate the MIN_DELAY computation and write to the spec:
NRF guarantees that the RTC will generate an interrupt for a
comparator value two cycles in the future. And further, we can test
at the set point to see if we "just missed" the interrupt (i.e. zero
cycles delay) and flag a synchronous interrupt. So we only need to
miss a requested interrupt now for the special case of exactly one
cycle in the future, and then we're only late by one cycle. That's
optimal.
Also fixes an off-by-one in the next cycle computation. By API
convention, an ticks argument of one or less means "at the next tick"
and not "right now". So we need to add one to the target cycle to
avoid incorrectly triggering a synchronous interrupt. This was a
non-issue when a tick is longer than a hardware cycle but is needed
now.
Also handles the edge case with zero latency interrupts (which are
unmaskable) which might mess up timing. This was always a problem,
but we're more sensitive now and it's comparatively more likely to
occur.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
move clock_control.h to drivers/clock_control.h and
create a shim for backward-compatibility.
No functional changes to the headers.
A warning in the shim can be controlled with CONFIG_COMPAT_INCLUDES.
Related to #16539
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Move internal and architecture specific headers from include/drivers to
subfolder for timer:
include/drivers/timer
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Change code from using now deprecated DT_<COMPAT>_<INSTANCE>_<PROP>
defines to using DT_INST_<INSTANCE>_<COMPAT>_<PROP>.
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
Enabling the RTC event is intended to support peripheral-to-peripheral
interconnects, so introduces a request for HFCLK and PCLK16M when the
event is triggered. This specific event is never used with PPI so
enabling events apparently does nothing but increase power consumption.
Closes#15513
Signed-off-by: Peter A. Bigot <pab@pabigot.com>
Per #13610, recent changes to this driver seem to have introduced
unexpected latency regressions. This patch effectively reverts these
patches which changed the meat of the driver:
ac36886e62 drivers: nrf: timer: add inline qualifier where
inlining is intended
084363a0dc drivers: timer: nrf: refactor for speed and correctness
71882ff8c4 drivers: timer: nrf: drop unnecessary counter mask
4b24e88fa4 drivers: timer: nrf: use irq_lock instead of spinlock
While backporting these seemingly unrelated hygiene patches:
7cbdb6c5c0 drivers/timer: Restore non-tickless tick count behavior
d30c9aeafd drivers: nrf_power_clock: Migrate to DTS.
75f77db432 include: misc: util.h: Rename min/max to MIN/MAX
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Update reserved function names starting with one underscore, replacing
them as follows:
'_k_' with 'z_'
'_K_' with 'Z_'
'_handler_' with 'z_handl_'
'_Cstart' with 'z_cstart'
'_Swap' with 'z_swap'
This renaming is done on both global and those static function names
in kernel/include and include/. Other static function names in kernel/
are renamed by removing the leading underscore. Other function names
not starting with any prefix listed above are renamed starting with
a 'z_' or 'Z_' prefix.
Function names starting with two or three leading underscores are not
automatcally renamed since these names will collide with the variants
with two or three leading underscores.
Various generator scripts have also been updated as well as perf,
linker and usb files. These are
drivers/serial/uart_handlers.c
include/linker/kobject-text.ld
kernel/include/syscall_handler.h
scripts/gen_kobject_list.py
scripts/gen_syscall_header.py
Signed-off-by: Patrik Flykt <patrik.flykt@intel.com>
The newer series of timer drivers will compare counters vs. the last
tick boundary to compute a number of ticks to announce to the kernel.
In the case of CONFIG_TICKLESS=n, this actually represents a change of
behavior from our old scheme where "ticks" always reflected the number
of interrupts received.
The distinction only matters when an interrupt is delayed more than a
full tick, of course. But that actually makes a difference to some
timekeeping code. Restore the old behavior.
This also has the benefit of further reducing code size when !TICKLESS
and improving performance of the ISR by removing the division
(remember Cortex M0 has no hardware divide!).
Fixes#12409
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
