The symbol _end is used to indicate the start of heap in
the common libc malloc code. On ACE, heap is in uncached
area. However, _end was in the cached area while end of
heap is in uncached area. This resulted in incorrect
calculation of heap size. So move _end into uncached
area so correct heap size can be calculated.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
PRE_KERNEL_1 is more suited for dealing with devices, so move out of
EARLY.
Verified on hardware and things seem to behave the same. Something was
changed since this was first introduced as this was not possible for
some reason.
Fixes#62627
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Flash segments require VMA to proper work. Executing from LMA
is not possible. Current implementation did not take into account
runtime iterable rom sections that any application could implement.
In the above cenario and as reported in the issue below, ESP32 won't run
when those ROM sections are created in application level.
This change make sure all flash segments are properly mapped
accordingly.
Fixes#61834
Signed-off-by: Sylvio Alves <sylvio.alves@espressif.com>
Selecting Low Power clock has the same result as selecting High
Performance clock (now IPLL). Therefore, the LP clock will be removed
from the list of available clocks on ACE family platforms.
Signed-off-by: Tomasz Leman <tomasz.m.leman@intel.com>
Updating value of WOVRCO request bit in CLKCTL register according to the
documentation. Previus value was mask used for clock enabling.
Signed-off-by: Tomasz Leman <tomasz.m.leman@intel.com>
Created set of new functions for configure fir coefficients with support
for packed format. This allowed to make the dai_dmic_set_config_nhlt
function simpler.
Signed-off-by: Adrian Warecki <adrian.warecki@intel.com>
All fir filters have an identical set of registers so their definitions
were combined to simplify the code. From the dai_dmic_set_config_nhlt
function, a duplicate piece of code responsible for configuring fir was
separated into a new function.
Signed-off-by: Adrian Warecki <adrian.warecki@intel.com>
All PDM controllers have the same set of registers. Their definitions have
been merged to simplify the code.
Signed-off-by: Adrian Warecki <adrian.warecki@intel.com>
The assert on L3_MEM_BASE_ADDR is incorrect, we need must convert
to uncached before use.
Fixes: ffd2121c65 ("soc: xtensa: intel_adsp: cavs: fix
power_down_cavs() signature")
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
To properly setup L1 exit timing this patch will use buffer interrupt
for HOST DMA and wait for Host HDA to actually start
First interrupt will clear all others.
Signed-off-by: Adrian Bonislawski <adrian.bonislawski@intel.com>
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Adjust the memory organization to avoid overlapping
critical regions from bootloaders (MCUboot and IDF)
Signed-off-by: Almir Okato <almir.okato@espressif.com>
The second argument 'uint32_t *hpsram_pg_mask' must be a cached
pointer and this needs to be reflected in function prototype.
Fixes sparse warning:
/zep_workspace/zephyr/soc/xtensa/intel_adsp/cavs/power.c:106:63:
warning: incorrect type in argument 2 (different address spaces)
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Converting between cached and uncached aliases should have correct
sparse annotations. Fix following sparse warning:
/zep_workspace/zephyr/soc/xtensa/intel_adsp/cavs/power.c:97:53: warning:
incorrect type in argument 1 (different address spaces)
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Use WAIT_FOR to wait for core power changes to be reflected
in status registers. If core power state does not complete in
10ms, k_panic() is raised.
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Use WAIT_FOR to wait for clock source to be enabled. If clock does
not come up in 10ms, raise k_panic().
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
This adds SoC and board configs to support the dc233c core
that is available on QEMU. This core has more features than
sample_controller, such as MMU support.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
With gcc from the zephyr sdk and -Wold-style-declaration is giving this
output:
zephyr/arch/arm/core/aarch32/cortex_a_r/fault.c:101:1: warning:
'inline' is not at beginning of declaration [-Wold-style-declaration]
101 | static void ALWAYS_INLINE
z_arm_fpu_caller_save(struct __fpu_sf *fpu)
| ^~~~~~
I searched to all of the source code to find these further occurances
where inline is not at the beginning of a function declaration.
Signed-off-by: Florian La Roche <Florian.LaRoche@gmail.com>
The existing heap starts at the beginning of SPIRAM which is also occupied
by the WiFi stack if CONFIG_ESP32_WIFI_NET_ALLOC_SPIRAM is enabled. This
results in memory corruption.
Update the linker description to add a _spiram_heap_start symbol and add a
KConfig size option to allow the SPIRAM stack to be correctly located and
sized. Requires matching commit in hal_espressif.
Fixes: 61359
Signed-off-by: Eric Holmberg <eric.holmberg@northriversystems.co.nz>
For some weird reasons, if the sections in linker script are not
in memory address order, there are lots of padding involved in
zephyr_pre0.elf. This moves the .intList section to its memory
ordered location. When building hello_world, this shrinks
zephyr_pre0.elf from 512MB to 339KB.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Implement the sys_poweroff hook for all ESP32 SoCs based on the previous
SOFT_OFF implementation.
Signed-off-by: Gerard Marull-Paretas <gerard@teslabs.com>
In case of HDA Link DMA FIFORDY bit is RO according
to HW specification thus should be managed by HDA controller.
Add a logic to set FIFORDY for HDA Host DMA only
Signed-off-by: Ievgen Ganakov <ievgen.ganakov@intel.com>
File was using stdint.h defined types without including and caused a
build error depending on include ordering.
Signed-off-by: Tom Burdick <thomas.burdick@intel.com>
Add an option to control whether or not hpsram banks should
be switched off during the power down. This is particular useful
when running tests because we don't want to lose the contents
of the memory window before we capture it.
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
Refactor the ESP32 target SOCs together with
all related boards. Most braking changes includes:
- changing the CONFIG_SOC_ESP32* to refer to
the actual soc line (esp32,esp32s2,esp32s3,esp32c3)
- replacing CONFIG_SOC with the CONFIG_SOC_SERIES
- creating CONFIG_SOC_FAMILY_ESP32 to embrace all
the ESP32 across all used architectures
- introducing CONFIG_SOC_PART_NUMBER_* to
provide a SOC model config
- introducing the 'common' folder to hide all
commonly used configs and files.
- updating west.yml to reflect previous changes in hal
Signed-off-by: Marek Matej <marek.matej@espressif.com>
The PM hooks were guarded with CONFIG_PM_POLICY_CUSTOM, however, they
need to be guarded (if file is always compiled) with CONFIG_PM. In fact,
CONFIG_PM_POLICY_CUSTOM requires to implement a custom policy hook,
something this module did not provide.
Signed-off-by: Gerard Marull-Paretas <gerard@teslabs.com>
Remove unnecessary __weak attribute from power management functions.
These functions are now defined once, globally, and mandatory for
systems that support CONFIG_PM.
Signed-off-by: Gerard Marull-Paretas <gerard@teslabs.com>
Add a new Kconfig option that has to be selected by SoCs providing PM
hooks. This option will be now required to enable CONFIG_PM. Before this
change, CONFIG_PM could always be enabled, regardless of SoC providing
any kind of low-power support.
Signed-off-by: Gerard Marull-Paretas <gerard@teslabs.com>
The clock rates for ACE series of Intel Audio DSP have changed.
The values come from the SOF project in their board configs.
CONFIG_XTENSA_CCOUNT_HZ is also set so the arch timing test
can pass.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Align `soc/nxp_adsp` with new `WEST_SIGN_OPTS` option added to
`soc/intel_adsp` by recent commit d98a7c2f8d ("soc: xtensa: cmake: add
new WEST_SIGN_OPTS variable").
This allows per-board rimage customization at the CMake level. Example
in `zephyr/boards/xtensa/nxp_adsp_NEWBOARD/board.cmake`:
set(WEST_SIGN_OPTS -- -c rimage/config/special.toml)
Signed-off-by: Marc Herbert <marc.herbert@intel.com>
Instead of relying on direct memory access via structs to
control CPU power and status, using inline functions instead
to hide the details. This makes reading the common code a bit
cleaner.
The function names are generic and not architecture or
platform specific, in an attempt to ease future arch or
platform additions with code reuse. Or else we would need to
rename these.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Some boards has ESP_SPIRAM enabled by default, which is
causing issues when MCUboot is used as 2nd bootloader.
Signed-off-by: Marek Matej <marek.matej@espressif.com>
This fixes the following build error:
```
zephyr/zephyr.elf', needed by 'zephyr/zephyr.ri', missing and no known
rule to make it
```
This appears when CONFIG_KERNEL_BIN_NAME is used.
Therefore, do not use zephyr.elf since some samples might be called
based on CONFIG_KERNEL_BIN_NAME.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Signed-off-by: Iuliana Prodan <iuliana.prodan@nxp.com>
The Zephyr linker scripts have inconsistent ordering of various chunks of
data which lands in RAM at runtime. This leads to the value of _end not
being consistently defined as the maximum address of static variables used
in the application.
Create a helper linker fragment, zephyr/linker/ram-end.ld, which can be
included after the last possible definition of RAM data, that consistently
sets _image_ram_end, _end and z_mapped_end.
Signed-off-by: Keith Packard <keithp@keithp.com>
After exiting D3 state if IMR context save is enabled, IDC interrupt
must be re-enabled again for all cores.
Signed-off-by: Serhiy Katsyuba <serhiy.katsyuba@intel.com>
