Add support for INTEL FPGA Nios V/g RISC-V based Processors.
Also amended SOC_NIOSV_M to use ATOMIC_OPERATIONS_BUILTIN.
Signed-off-by: Khor Swee Aun <swee.aun.khor@intel.com>
Amount of memory banks should vary depending on platform, otherwise
power down sequence might result in ipc timeout as the memory used by
the ipc itself is shutdown.
Lift the needed defines from sram.c to adsp_memory.h and shorten the
names to avoid collision with sof code. No functional change to sram.c.
Signed-off-by: Jaska Uimonen <jaska.uimonen@linux.intel.com>
It doesn't make sense to select this option at SoC level. This feature
is meant for subsystems/modules that need device handles to be
modifiable at runtime.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
The field offset is incorrect, L1EXP is at bit 24 and L1ETP at bit 25.
Fixes: cc6e9c094a ("soc/intel_adsp: Low level HDA driver and tests")
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
The Nordic board are selecting CONFIG_ARCH_HAS_CUSTOM_BUSY_WAIT by
default and rely on some HAL code to implement a cycle accurate busy
delay loop.
This is in general fine for real hardware but when QEMU and emulation is
taken into account, a cycle accurate busy wait implementation based on
delay in executing machine code can be misleading.
Let's take for example qemu_cortex_m0 (that is based on the nRF51
chipset) and this code:
uint32_t before, after;
while (1) {
before = k_cycle_get_32();
k_busy_wait(1000 * 1000);
after = k_cycle_get_32();
printk("diff cycles: %d\n", after - before);
}
With CONFIG_SYS_CLOCK_HW_CYCLES_PER_SEC=1000000 this diff cycles should
be always around 1000000, in reality when executed with:
qemu-system-arm -cpu cortex-m0 -machine microbit -nographic
-kernel build/zephyr/zephyr.elf
This results in something like this:
diff cycles: 22285
diff cycles: 24339
diff cycles: 21483
diff cycles: 21063
diff cycles: 21116
diff cycles: 19633
This is possibly due to the fact that the cycle accurate delay busy loop
is too fast in emulation.
When dealing with QEMU let's use the reliable busy loop implementation
based on k_cycle_get_32() instead.
Signed-off-by: Carlo Caione <ccaione@baylibre.com>
As a result of the 11a2107d991b("riscv: timer: driver revamp") commit,
gd32vf103 no longer works properly.
In the
c9c04e491e0f("soc: riscv: Add initial support for GigaDevice GD32V SoC")
that is the first commit of implementation of this SoC,
set the CPU frequency to CONFIG_SYS_CLOCK_HW_CYCLES_PER_SEC,
and riscv_machine_timer divide the clock with the value of
CONFIG_RISCV_MACHINE_TIMER_SYSTEM_CLOCK_DIVIDER.
The CONFIG_SYS_CLOCK_HW_CYCLES_PER_SEC should set the timer's clock,
so I fixed to set the config as 27MHz in this PR. Also, remove the
unnecessary CONFIG_RISCV_MACHINE_TIMER_SYSTEM_CLOCK_DIVIDER setting.
Signed-off-by: TOKITA Hiroshi <tokita.hiroshi@gmail.com>
Updated the flash segment alignment on S3,
so that mapping corresponds with the linkage.
Fixed hard-coded flash size for ESP32.
Signed-off-by: Marek Matej <marek.matej@espressif.com>
For Intel cavs2.5, access from LPGPDMAC to Audio Link Hub
RX/TX registers needs to be explicitly enabled before use.
The logic follow hardware initialization done in
SOF project sof/src/platform/intel/cavs/platform.c
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
1. Add support for the USB Full Speed controller
2. Add a Kconfig to specify if a dedicated USB
RAM is available in the SoC.
Signed-off-by: Mahesh Mahadevan <mahesh.mahadevan@nxp.com>
Add FLEXSPI1 and FLEXSPI2 memory sections for RT5xx SOC. These sections
can be used by the user's application as part of a custom linker script,
if the application needs to relocate a buffer to external SRAM connected
to FLEXSPI1 or FLEXSPI2
Signed-off-by: Daniel DeGrasse <daniel.degrasse@nxp.com>
Adds watchdog reset as a reset source at SoC init. This is achieved by:
1. Writing 0x2 to the RESET_EN bitfield register to indicate watchdog
reset is enabled.
2. Writing 0x1 to the CFG_CDC_SYNC register to commit the change.
3. Polling the CFG_CDC_SYNC register until reading 0 to confirm the
change has been processed.
This patch is part of the OpenTitan watchdog (AON Timer) support patch
series. It is needed to ensure that the watchdog reset functionality
is enabled. Note that the timer itself is not enabled here, only the
reset function.
Signed-off-by: Tyler Ng <tkng@rivosinc.com>
1. Fixes the number of interrupts in OpenTitan by default. This should
be 32 + 185 = 217 IRQs, as there are 185 configurable registers,
including interrupt 0.
2. Adds 2ND_LVL_INTR_00_OFFSET Kconfig, which is needed to generate a
PLIC interrupt on IRQ 11.
Signed-off-by: Tyler Ng <tkng@rivosinc.com>
The SoCs/Boards using those headers were dropped, so remove those
remaining headers that were forgotten.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
This commits selects the `CONFIG_SOC_GECKO_SE` option for the BG27 SoC
series. This is required for BLE support, as it enables the inclusion of
the `sl_protocol_crypto` component from SiLabs HAL.
Signed-off-by: Filip Kokosinski <fkokosinski@antmicro.com>
Internal bootloader will only run application image if valid product
header is present on flash. This means product header is required for
application that are not linked to code partition. Other applications
that are linked to code partition are meant to be run by MCUboot and
should not touch product header, with the exception of MCUboot itself
which is also run by internal bootloader so requires product header.
Default flash load offset for applications not linked to code partition
is set to 0x2400 as this is where internal bootloader looks for an
application image to run based on product header written by flasher.
Flash load offset for MCUboot is set from boot partition.
Valid product header is added by ezFlashCLI when using "flash_image"
command.
Co-authored-by: Andrzej Kaczmarek <andrzej.kaczmarek@codecoup.pl>
Signed-off-by: Andrzej Kaczmarek <andrzej.kaczmarek@codecoup.pl>
Signed-off-by: Niek Ilmer <niek.ilmer.aj@renesas.com>
Applications that are run from MCUboot should reconfigure cache to
ensure they run from cached ared on flash. Initial cache configuration
is done by internal bootloader and is valid for MCUboot only.
Cache is configured once at boot, then new configuration is applied by
triggering SW reset.
Co-authored-by: Andrzej Kaczmarek <andrzej.kaczmarek@codecoup.pl>
Signed-off-by: Andrzej Kaczmarek <andrzej.kaczmarek@codecoup.pl>
Signed-off-by: Niek Ilmer <niek.ilmer.aj@renesas.com>
Flash address is updated to 0x16000000, i.e. actual location instead of
remapped one. FLASH_BASE_ADDRESS is now set via dts.
Co-authored-by: Andrzej Kaczmarek <andrzej.kaczmarek@codecoup.pl>
Signed-off-by: Andrzej Kaczmarek <andrzej.kaczmarek@codecoup.pl>
Signed-off-by: Niek Ilmer <niek.ilmer.aj@renesas.com>
Initialize threadptr with 0 during the boot before it possibly be used
since the architectural reset state is undefined.
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
This make MCUboot build as Zephyr application.
Providing optinal 2nd stage bootloader to the
IDF bootloader, which is used by default.
This provides more flexibility when building
and loading multiple images and aims to
brings better DX to users by using the sysbuild.
MCUboot and applications has now separate
linker scripts.
Signed-off-by: Marek Matej <marek.matej@espressif.com>
sl_power_manager can only be used if HAL services infrastructure
is enabled, and that is controlled by SOC_GECKO_DEV_INIT kconfig option.
Therefore, SOC_GECKO_PM_BACKEND_PMGR may not be enabled without
SOC_GECKO_DEV_INIT.
Signed-off-by: Roman Dobrodii <rdobrodii@antmicro.com>
Using EM2 or deeper sleep states (where HF clocks are off) requires
special care if BLE radio is used, since BLE radio relies on that clock,
and its power/clock requirements need to be taken into account
On SiLabs, radio PM is implemented as part of RAIL blob, which relies
on sl_power_manager HAL service. I've implemented SoC PM
state changes using sl_power_manager instead of emlib, and added
call to RAIL PM initialization in Gecko HCI driver.
Signed-off-by: Roman Dobrodii <rdobrodii@antmicro.com>
This patch modifies intel_adsp_ipc_is_complete function so it don't
report that IPC is completed when we are still waiting for ACK from
HOST.
This change will allow to solve the problem that occurs during the power
state transitions. Occasionally, the Application decides to enter the
power gating state after sending an IPC message, before receiving an ACK
from the HOST. This results in broken IPC communication when coming back
to Idle state.
Signed-off-by: Tomasz Leman <tomasz.m.leman@intel.com>
Driver was based on can_sam. SAMC21 has only 1 interrupt for one
can "output", so can interrupt has to executes two lines of
interrupts.
CAN is configured to use OSC48M clock via GLCK7. GLCK7 is set
by divider configured from dts.
Signed-off-by: Kamil Serwus <kserwus@gmail.com>
This patch enables the PLL clock output and PLL ref clock
for second ethernet module in NXP's i.MxRT106x SoCs
Signed-off-by: Sumit Batra <sumit.batra@nxp.com>
Thanks to PR [1] which introduces support for cache management
operations on xtensa architecture NXP SoCs can now use the
Zephyr native cache management API.
This commit enables some configurations that will allow us
to use the native Zephyr cache management API.
[1]: https://github.com/zephyrproject-rtos/zephyr/pull/50136
Signed-off-by: Laurentiu Mihalcea <laurentiu.mihalcea@nxp.com>
To reduce the interrupt latency of MIWU events, the driver prepares a
dedicated callback function item list for each MIWU group in this PR. We
needn't check the MIWU table and group of the event in ISR. And the
maximum item number of each list is also limited to 8. After applying
this PR, the interrupt latency reduces to ~10us consistently.
Signed-off-by: Mulin Chao <mlchao@nuvoton.com>
The end of the heap should be the same as _heap_sentry. The current end
marker just covers the range of memory that is explicitly put in
.heap_memand not account until the end of L2_SRAM_BASE +
L2_SRAM_SIZE memory.
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
Glitches were observed if a GPIO pin was configured by
ROM to a non-default state and then Zephyr PINCTRL
reconfigured the pin. The fix involves using the correct
PINCTRL YAML output enable and state flags. Reading the
current spin state and reflecting into new pin configuration
if the pin is output and the drive low/high properties are
not present. We also take advantage of GPIO hardware reflecing
the alternate output value in the parallel output bit before
enabling parallel output mode. Interpret boolean flags with
both enable and disable as do not touch if neither flag is
present. We give precedence to enable over disable if both
flags mistakenly appear. Note, PINCTRL always clears the
GPIO control input pad disable bit.
Signed-off-by: Manimaran A <manimaran.a@microchip.com>
1. Enable os_timer as a wakeup-source in the board
dts file.
2. Enable PM_DEVICE when PM is enabled.
Signed-off-by: Mahesh Mahadevan <mahesh.mahadevan@nxp.com>
The macro HW_STATE_CHECK_DELAY represents the number of usec to
busy-wait when waiting for a h/w register state change. This value
has been lowered by 75% to correct a couple of issues related
to Power Management (PM).
1. This resolves a kernel.timer.tickless test failure (see
tests/kernel/timer/timer_api) on the LNL simulator where the test
was taking too long to wake after sleeping for 50 ticks. This
was tracked down to the xtensa SOC power management code where
it was performing a busy wait of 256 usec, which was equivalent
to 2.56 ticks.
2. Furthermore, this delay of 256 usec contradicted the information
in the DTSI (see power-states for d0i3) which states that the
"exit-latency-us" is 100 usec. Reducing this value to 64 helps to
keep that in line.
Signed-off-by: Peter Mitsis <peter.mitsis@intel.com>
This aligns `soc/xtensa/nxp_adsp/` with commit
fad2da39aa ("intel_adsp: move `west sign` from `west flash` to earlier
`west build`")
The --if-tool-available option preserves backwards-compatibility:
nothing happens if rimage is not found.
Signed-off-by: Marc Herbert <marc.herbert@intel.com>
Update the SOC_TOOLCHAIN_NAME to intel_ace15_mtpm so that
we use the correct overlay in Xtensa HAL module. Note that
ace20_lnl will also be using this as well. That will change
once we have a proper toolchain for ace20_lnl.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
The ACE 2.0 LNL platform has 5 HIFI4 cores. Change number
of cores to enable 5th core on the platform.
Signed-off-by: Jaroslaw Stelter <Jaroslaw.Stelter@intel.com>
Added logging declaration to soc.c of stm32wl, which resolves build
errors when using this SOC with DEBUG logging level
Fixes: #57655
Signed-off-by: Oliver King <oliver.king@steadconnect.com>
(resubmitting after it has been reverted by 0f2a352cbd ('Revert
"xtensa: remove ELF section address rewriting"')
Now rimage can handle both cached and uncached addresses correctly,
ELF rewriting isn't needed any more.
Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
Khor Swee Aun