This reverts commits e0ff1b16d8,
339e5ac217 and
09c69b7607. A different approach will
be used for making SOF a platform-agnostic Zephyr application.
Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
The two components copier and base_fw are
essential to IPC4 support on SOF on Zephyr.
This patch integrates the build of the two
components to SOF build with zephyr RTOS.
Signed-off-by: Chao Song <chao.song@linux.intel.com>
The DF2T IIR core has been moved earlier to library but the
helper functions for configuring it were left to EQ component.
This patch eases use of IIR core from other components. The
file src/audio/eq_iir/iir.c is moved to src/math. There are no
changes except of remove iir.h since the contents specific
for DF2T type were merged to iir_df2t.h
Signed-off-by: Seppo Ingalsuo <seppo.ingalsuo@linux.intel.com>
To align with Zephyr smp.c:smp_init_top(), add the missing call to
smp_timer_init() in secondary_init() of the SOF wrapper. Also
add inline documentation why we are open-coding the function on
SOF side.
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
This reverts commit 414089aa44.
Failures hit with 4core cAVS2.5 configuraions, so this has
to be reverted. We need an implementation that keeps both
Zephyr (and SOF runtime-DSP if still used) in the loop.
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Currently there is code to power up secondary cores both
in Zephyr architecture hooks in wrapper.c, as well as in
SOF runtime-PM driver.
Some of this is necessary overlap until full replacement of
SOF runtime-PM is available. But for secondary core power up,
we can already clean up the flow and hook the Zephyr core
enable code to runtime-PM driver directly. This fixes startup
issues seen on Intel cAvS2.5 platforms.
Core power-down is left intentionally for runtime-PM driver
to handle as the related interface in Zephyr is not ready
for use yet.
BugLink: https://github.com/thesofproject/sof/issues/5018
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
This reverts commit 079ce95b98.
As explained in the comment that it modified, predicted in the #5159
discussion and reproduced in test #5290 /
https://github.com/thesofproject/sof/runs/4988844402, this commit broke
'CONFIG_TRACE=n'. Simple reproduction steps:
$ rm -rf zephyrproject/build-apl
$ echo 'CONFIG_TRACE=n' >> \
boards/xtensa/intel_adsp_cavs15/intel_adsp_cavs15_defconfig
$ sof/scripts/xtensa-build-zephyr.sh [ -p /path/to/zephyrproject ] apl
fw abi main version: 3.20.1
fw abi dbg version: 5.3.0
warning: can't find section .static_log_entries in zephyr/zephyr.elf
error: section .static_log_entries can't be found
error: unable to write dictionaries, -22
Signed-off-by: Marc Herbert <marc.herbert@intel.com>
Zephyr requires CMake 3.21 and SOF requires 3.13, see
commit a0acad3414 ("CMake: require version 3.13")
Fixes commit 4b2edf8eed ("zephyr: wrapper: complete zephyr support for
working audio")
Signed-off-by: Marc Herbert <marc.herbert@intel.com>
_smex_placeholder dowsn't have to be global, make it static.
Also no need to explicitly specify smex_placeholder_f() as inline,
the compiler will decide by itself.
Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
On generic Zephyr platform case, use an OS timer call to get a time
stamp.
Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
This makes it possible to clean sof_versions.h.
Also remove DOT_CONFIG_PATH copy/paste which is unused in
Zephyr.
Signed-off-by: Marc Herbert <marc.herbert@intel.com>
Zephyr also a version.h too so the situation was a buggy #include mess
that (among others) dropped the SOF git version from the SOF banner when
using Zephyr. It's absolutely impossible to fix this mess while keeping
two .h files both named exactly the same and even included with the
exact same '#include version.h' line, I mean no sof/version.h or other
prefix.
As a bonus, this rename also reduces the XTOS-only confusion between
"version.h" and cavs/version.h".
Generating a file in the source tree is a serious bug (to be fixed
later), it means this file is never cleaned. Do not hide this bug with a
.gitignore rule.
Signed-off-by: Marc Herbert <marc.herbert@intel.com>
When built with Zephyr, SOF ideally should be a fully platform-
agnostic application. All platform-specific code should reside in
Zephyr itself, including core architecture and platform
infrastructures and drivers. To prepare for this transition we add a
generic SOF-Zephyr platform. It is based on the current Tiger Lake
architecture and should be usable for cAVS 2.5 and newer systems.
Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
Two directories for cAVS 2.0 and 2.5 in Zephyr CMakeLists.txt don't
exist. They're safely ignored, but remove them to avoid confusion.
Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
Discovered the hard way that the Zephyr build assumes sof-logger is
always up to date when present. It's not clear from cmake's
ExternalProject documentation why that is the case. In any case
BUILD_ALWAYS fixes the issue and takes very little time because it does
not rebuild sof-logger from scratch. BUILD_ALWAYS is already used for
XTOS.
Fixes commit f6c71c21d0 ("zephyr/CMakeLists.txt: build smex and
sof-logger")
Signed-off-by: Marc Herbert <marc.herbert@intel.com>
Zephyr commit 174cb7f9f183 switched 'atomic_t' from 'int' to
'long' and broke SOF compilation as shown below.
Recent SOF commit 970d7d61ec ("zephyr: update print messages for
64-bit atomics") switched all atomic_read() logs to "%ld" instead of
"%d" which fixed compilation with Zephyr after the switch but broke it
with Zephyr _before_ the switch.
This no-op (long) cast makes SOF compatible with Zephyr both before and
after the switch.
Note the sof-logger does not make any difference between %ld and %d and
does not care.
Sample failure:
src/lib/dma.c: In function 'dma_get': sof/src/include/sof/trace/trace.h:290
error: format '%d' expects argument of type 'int', but
argument 5 has type 'atomic_val_t' {aka 'long int'} [-Werror=format=]
290 | printk("%llu " format "\n", platform_timer_get(NULL), \
| ^~~~~~~
sof/src/lib/dma.c:119:2: note: in expansion of macro 'tr_info'
119 | tr_info(&dma_tr, "..., busy channels = %d", atomic_read(...))
Signed-off-by: Marc Herbert <marc.herbert@intel.com>
Primary core during going to D0i3 state disables all another cores
in platform_pg_int_handler() function. Before it secondary cores
should be prepared (disable interrupts, perform writeback) in order
to make proper restore flow after D0i3->D0 transition.
This commit adds cpu_secondary_cores_prepare_d0ix() function in
ipc_pm_gate() handler, which sends idc to secondary cores
with information that they should perform preparation for power down
in the next platform_wait_for_interrupt() invocation. In
platform_wait_for_interrupt() there is cpu_power_down_core()
invocation with CPU_POWER_DOWN_MEMORY_ON (performs writeback/invalidate,
disables interrupts), when proper flag is set.
Signed-off-by: Bartosz Kokoszko <bartoszx.kokoszko@linux.intel.com>
This will add missing secondary cores restore on wake up,
simple pm_runtime_get(PM_RUNTIME_DSP, PLATFORM_PRIMARY_CORE_ID)
in lps is not enough in multicore scenario
Signed-off-by: Adrian Bonislawski <adrian.bonislawski@intel.com>
Instead of hard-coding a guessed heap size, obtain remaining free
SRAM from the linker and use it completely for the heap.
Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
As of July 2021 we support (too) many tracing options and this
duplication is unfortunately the only way I found to support them all
while giving the compiler the opportunity to optimize away as many
strings as possible.
Supported configurations:
- Systems with limited memory and zero space for full strings, must use
SOF dictionary only.
- Systems with enough space for all strings to be in memory.
- Anything in between
- Support to duplicate only important message to both the DMA and the
mailbox (the default)
- CONFIG_TRACEM: supports duplicating ALL messages to both the DMA and
the mailbox
- CONFIG_TRACEV: supports deleting verbose statements at compile time to
save space
- CONFIG_TRACE: support turning off all traces at compile-time
- SOF dict trace de-duplication a.k.a. "adaptive filtering"
- Dynamic log levels per component
- Redirection to Zephyr's shared memory tracing that requires full strings.
Signed-off-by: Marc Herbert <marc.herbert@intel.com>
Recent changes increased the size of various sections, breaking
linkage again. We need to make the linker calculate the heap size
automatically but until that is ready, reduce the heap size further
to fix the current breakage.
Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
As per documentation in include/sof/lib/alloc.h, rmalloc from
MEM_ZONE_SYS will always succeed. In the unlikely case the alloc
fails, trigger a system panic in rmalloc().
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Add caching to more heap zones. With addition of SYS, SYS_RUNTIME
and RUNTIME the mapping is now aligned with XTOS allocator.
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
The XTOS allocator aligns non-cachec allocations to the platform
cacheline size. Test results indicate some SOF application code
relies on this behaviour, so align Zephyr's rmalloc implementation
to match the behaviour.
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Implement the cached SOF heap zones using a single sys_heap
object.
Compared to old implementation which used two separate sys_heaps,
this implementation enables dynamic cached/uncached partitioning
of the heap.
To ensure coherency of the Zephyr heap operation, if cached heap
is enabled, all allocations must be aligned to dcache linesize.
Add a Kconfig option CONFIG_SOF_ZEPHYR_HEAP_CACHED to turn on/off
the cached heap logic, with default set to no for all platforms.
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
The heap code was invalidating blocks on allocation, but that's the
wrong side of the pipe. By definition, new heap memory
will/should/must be written before being used (because the memory is
undefined), so any cached contents are irrelevant as they'll be
overwritten.
But when the user is finished with the block and frees it, there may
still be live dirty cache lines in the region on the current CPU.
Those must be invalidated, otherwise they will be evicted from the
cache at some point in the future, on top of the memory region now
being used for different purposes on another CPU.
Remove the invalidate on allocation. Add it back in free. Leverage a
new Zephyr sys_heap_usable_size() routine to get the size so we don't
have to store it in an extra header.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Building SOF with Zephyr for Apollolake with XCC fails because of
insufficient RAM. Reduce the heap size to fix the problem.
Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
Currently set to obsolete 0.12.4. Better remove it here for everyone
rather than overriding it in Github Actions only.
Also symlinks the toolchains in /opt to /home to fix autodetection
Signed-off-by: Marc Herbert <marc.herbert@intel.com>
We have change the size of runtime_shared zone to be decided at link
stage, here hardcode HEAP_RUNTIME_SHARED_SIZE to unblock the Zephyr
building.
Signed-off-by: Keyon Jie <yang.jie@linux.intel.com>
COMPILER_WORKAROUND_CACHE_ATTR is also enabled on Zephyr,
but the cache_attr.c file is not compiled since is not added in
zephyr_library_sources.
Without this fix, we get:
cache.h:110: undefined reference to `glb_is_cached'
I've added the file for all platforms since cache_attr.c is generic.
Signed-off-by: Iuliana Prodan <iuliana.prodan@nxp.com>
sof_heap variable is used by all platform, with or
without ENABLE_CACHED_HEAP.
Without this fix, on i.MX with Zephyr we get:
error: 'sof_heap' undeclared (first use in this function);
Fixes: fed69b6a34 ("zephyr: wrapper: disable cached heap on Intel platforms")
Signed-off-by: Iuliana Prodan <iuliana.prodan@nxp.com>
Tests with dynamic-pipeline topologies increase the complexity
of alloc/free patterns for heap and these tests have uncovered
new issues with heap corruption. Tests are reporting failed
allocs although heap still has free space. Running same test
with cached heap disabled shows no failures.
The cached heap is already disabled for IMX targets, so build on
this code and add a local ENABLE_CACHED_HEAP define to toggle
the cached heap.
BugLink: https://github.com/thesofproject/sof/issues/4494
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Clean up the ifdefs for ipc_pm_context_save(). By implementing
platform_timer_stop() in the Zephyr wrapper, the original context
save sequence can be used also for Zephyr builds.
By doing this, we can fix issues with IPC timeout on CTX_SAVE.
Disabling all interrupts before the IPC reply is needed also on
Zephyr.
BugLink: https://github.com/thesofproject/sof/issues/4507
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
When the SOF Kconfig is included in other projects, the main menu should
not be that of SOF, rather it should be of the project including SOF.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
In XTOS SOF, ipc_send_queued_msg() is run by task_main_primary_core().
In Zephyr we need to schedule ipc_send_queued_msg() using a notifier
triggered by the periodic ll_scheduler.
This is similar to commit c194125b83 ("zephyr:
add notifier_register(ipc_send_queued_msg) in task_main_start()")
For i.MX we need to use this temporary fix for SOF_SCHEDULE_LL_DMA, also.
Signed-off-by: Iuliana Prodan <iuliana.prodan@nxp.com>
On i.MX the DMA interrupts are routed via IRQ_STEER.
In order for this to work we need to:
- make any second level interrupts handling go
through interrupt-irqsteer.c;
- use first level interrupt handling from
wrapper.c.
TODO: Implement a driver for the IRQ_STEER in Zephyr,
to replace the legacy code (interrupt-irqsteer.c).
Signed-off-by: Iuliana Prodan <iuliana.prodan@nxp.com>
For now, zephyr_ll is limited to timer_domain.
For i.MX we use dma_domain, so keep the ll_schedule
from SOF, until we extend the zephyr_ll for DMA_IRQ.
Signed-off-by: Iuliana Prodan <iuliana.prodan@nxp.com>
Remove zephyr_ll.c from mandatory files for building
SOF with Zephyr and include it where necessary: in CAVS 1.5,
CAVS 1.8, CAVS 2.0 and CAVS 2.5.
While here, add ll_schedule for BROADWELL and BAYTRAIL
when building SOF with Zephyr.
Signed-off-by: Iuliana Prodan <iuliana.prodan@nxp.com>
As of July 2021 we support (too) many tracing options and this
duplication is unfortunately the only way I found to support them all
while giving the compiler the opportunity to optimize away as many
strings as possible.
Supported configurations:
- Systems with limited memory and zero space for full strings, must use
SOF dictionary only.
- Systems with enough space for all strings to be in memory.
- Anything in between
- Support to duplicate only important message to both the DMA and the
mailbox (the default)
- CONFIG_TRACEM: supports duplicating ALL messages to both the DMA and
the mailbox
- CONFIG_TRACEV: supports deleting verbose statements at compile time to
save space
- CONFIG_TRACE: support turning off all traces at compile-time
- SOF dict trace de-duplication a.k.a. "adaptive filtering"
- Dynamic log levels per component
- Redirection to Zephyr's shared memory tracing that requires full strings.
Signed-off-by: Marc Herbert <marc.herbert@intel.com>
Guennadi Liakhovetski