pipeline: set scheduling comp during complete

[ranj063]

Move the code that sets the scheduling comp for
a pipeline to ipc_pipeline_complete(). This removes the
restriction that the scheduling comp must be set up
before the pipeline widget and provides the
flexibility in the kernel to set up the widgets
in any order while parsing topology.

Signed-off-by: Ranjani Sridharan ranjani.sridharan@linux.intel.com

[plbossart]

@ranj063 is this related to a specific issue or evolution? Also wondering if this means a specific kernel would need a firmware with a minimal version, which could be problematic in terms of backwards compatibility.

[ranj063]

@ranj063 is this related to a specific issue or evolution? Also wondering if this means a specific kernel would need a firmware with a minimal version, which could be problematic in terms of backwards compatibility.

@plbossart this is mainly required for IPC4 support with dynamic pipelines. A newer kernel that tries to set up the pipeline widget willl require this patch. So I suppose this calls for an ABI bump

[plbossart]

@ranj063 is this related to a specific issue or evolution? Also wondering if this means a specific kernel would need a firmware with a minimal version, which could be problematic in terms of backwards compatibility.

@plbossart this is mainly required for IPC4 support with dynamic pipelines. A newer kernel that tries to set up the pipeline widget willl require this patch. So I suppose this calls for an ABI bump

I guessed this was for IPC4, but you modified common code and helper-ipc3.c, so not sure how this might work for current platforms.

[ranj063]

I guessed this was for IPC4, but you modified common code and helper-ipc3.c, so not sure how this might work for current platforms.

yes, I suppose this could be in the common helper if things work the same way for IPC3 and IPC4.

[marc-hb]

The https://github.com/thesofproject/sof/pull/4113/checks?check_run_id=2461033508 regression was added by #4111 and is unrelated to this PR

[lgirdwood]

@ranj063 UT's build failing - best to locally check with ./scripts/run-mocks.sh

[ 76%] Built target find_min_int16
cc1: warnings being treated as errors
/quickbuild/workspace1/24733/SOF_FW/test/cmocka/src/audio/pipeline/pipeline_new.c: In function ‘test_audio_pipeline_pipeline_new_creation’:
/quickbuild/workspace1/24733/SOF_FW/test/cmocka/src/audio/pipeline/pipeline_new.c:50: warning: passing argument 1 of ‘pipeline_new’ makes integer from pointer without a cast
/quickbuild/workspace1/24733/SOF_FW/test/cmocka/src/audio/pipeline/pipeline_new.c:50: error: too many arguments to function ‘pipeline_new’
[ 77%] Building C object test/cmocka/src/audio/pipeline/CMakeFiles/pipeline_free.dir/pipeline_connection_mocks.c.o
test/cmocka/src/audio/pipeline/CMakeFiles/pipeline_new.dir/build.make:63: recipe for target 'test/cmocka/src/audio/pipeline/CMakeFiles/pipeline_new.dir/pipeline_new.c.o' failed
make[2]: *** [test/cmocka/src/audio/pipeline/CMakeFiles/pipeline_new.dir/pipeline_new.c.o] Error 2
make[2]: Leaving directory '/quickbuild/workspace1/24733/SOF_FW/build_ut'
cd /quickbuild/workspace1/24733/SOF_FW/build_ut/test/cmocka/src/audio/pipeline && /localdisk/tools/xtensa/RG-2017.8-linux/XtensaTools/bin/xt-xcc -DUNIT_TEST -D_UINTPTR_T_DEFINED -DRELATIVE_FILE=\"test/cmocka/src/audio/pipeline/pipeline_connection_mocks.c\" -I/quickbuild/workspace1/24733/SOF_FW/test/cmocka/include -I/localdisk/tools/cmocka/build_cavs2x_LX6HiFi3_2017_8/install/include -I/quickbuild/workspace1/24733/SOF_FW/src/arch/xtensa/include -I/quickbuild/workspace1/24733/SOF_FW/src/arch/xtensa/xtos -I/quickbuild/workspace1/24733/SOF_FW/src/platform/tigerlake/include -I/localdisk/tools/xtensa/RG-2017.8-linux/cavs2x_LX6HiFi3_2017_8/xtensa-elf/arch/include -I/localdisk/tools/xtensa/RG-2017.8-linux/cavs2x_LX6HiFi3_2017_8/xtensa-elf/include -I/quickbuild/workspace1/24733/SOF_FW/src/platform/intel/cavs/include -I/quickbuild/workspace1/24733/SOF_FW/src/include -I/quickbuild/workspace1/24733/SOF_FW/rimage/src/include -I/quickbuild/workspace1/24733/SOF_FW/build_ut/generated/include -fdata-sections -ffunction-sections -fno-inline-functions -mlongcalls -O2 -g -Wall -Werror -Wl,-EL -Wmissing-prototypes -Wpointer-arith -mtext-section-literals -imacros/quickbuild/workspace1/24733/SOF_FW/build_ut/generated/include/autoconfig.h -o CMakeFiles/pipeline_free.dir/pipeline_connection_mocks.c.o -c /quickbuild/workspace1/24733/SOF_FW/test/cmocka/src/audio/pipeline/pipeline_connection_mocks.c
CMakeFiles/Makefile2:1646: recipe for target 'test/cmocka/src/audio/pipeline/CMakeFiles/pipeline_new.dir/all' failed
make[1]: *** [test/cmocka/src/audio/pipeline/CMakeFiles/pipeline_new.dir/all] Error 2
make[1]: *** Waiting for unfinished jobs....

[lgirdwood]

@ranj063 I assume you also need a kernel path (perhaps we can use the MINOR ABI) to set the order of the pipeline new IPC in the driver ?

[lgirdwood]

@juimonen fyi - you will need this PR for pipeline allocator.

[ranj063]

@ranj063 I assume you also need a kernel path (perhaps we can use the MINOR ABI) to set the order of the pipeline new IPC in the driver ?

@lgirdwood do I really need an ABI bump here? This change wont impact the static pipeline topologies that we have today. The hda-generic pipelines are the only ones using dynamic pipelines and Im not sure if they are used in any product. So if and when we switch to dynamic pipeliine in tplg, we can be assured that the FW will already have this change right?

[lgirdwood]

@ranj063 I assume you also need a kernel path (perhaps we can use the MINOR ABI) to set the order of the pipeline new IPC in the driver ?

@lgirdwood do I really need an ABI bump here? This change wont impact the static pipeline topologies that we have today. The hda-generic pipelines are the only ones using dynamic pipelines and Im not sure if they are used in any product. So if and when we switch to dynamic pipeliine in tplg, we can be assured that the FW will already have this change right?

ok, sounds fine. Btw, lets get this in for v1.8-rc1, just needs the UT fixed. See above comment on running the UTs locally.

[ranj063]

ok, sounds fine. Btw, lets get this in for v1.8-rc1, just needs the UT fixed. See above comment on running the UTs locally.

@lgirdwood should be fixed now

[ranj063]

this patch should be merged after #4143 though.

[lgirdwood]

@ranj063 looks like a pipeline error happening from time to time in QB logs.

[ 1745240.026042] ( 6.380208) c0 ipc src/ipc/helper-ipc3.c:411 ERROR ipc_pipeline_complete(): icd->core (0) != p->core (0) for pipeline scheduling component icd->id 0

[lgirdwood]

@ranj063 I'm assuming this is a fix for v1.8 ? if so, we can get into rc2

[ranj063]

@zrombel could you please help me understand why the "17_03_TestMultiCoreMultiKernelHda48000hz24b32b2ch" test is failing with this PR?