Testbench add process components support by singalsu · Pull Request #2510 · thesofproject/sof

singalsu · 2020-03-09T15:24:10Z

Here's a cleaned up and more complete version of testbench PR. It consists of these commits:

Testbench: Add processing component load
Testbench: Fix possible uninitialized use of variable ret
Testbench: Build topology parser with debug symbols
Tools: Test: Cleanup component run scripts
Topology: Fix in EQ only pipelines with macro PIPELINE_SINK

tools/topology/sof/pipe-eq-iir-capture.m4

lgirdwood

General, components can support > controls (use different names). Also we need to de-duplicate. Please also add more inline comments.

lgirdwood · 2020-03-09T21:03:26Z

tools/testbench/topology.c

This looks like we are starting to duplicate a lot of component loading code. Can we make this all generic.

I simplified the controls load / ignore from previous version but otherwise I didn't see similar flows. There's always some component specifics.

singalsu · 2020-03-16T17:55:48Z

General, components can support > controls (use different names). Also we need to de-duplicate. Please also add more inline comments.

I didn't yet see such cases so I could not test. EQs and volume had only one control. An error is reported if there's more though it's not what you recommended.

I tried to seek for de-duplication but it didn't look straightforward. The topology handler and loader in kernel sure is nicer than this. A bigger repair could reuse more from there.

I'd like to pursue more important things like building a simplified testbench (statically linked, possibly remove fseeks) for xt-run/dbg environment.

sebcarlucci · 2020-03-17T00:46:33Z

I tested this with the dc blocker (PR #2462) as well. It works fine! Once this is merged, I will open a new PR with the changes to make this work with DCB.

fyi @cujomalainey

singalsu · 2020-03-17T09:34:32Z

@sebcarlucci Wow, that's great news! Thanks for testing!

I've so far run some components with the SRC test script, e.g. "src_test(24, 24, 48000, 48000)" with src_run.sh edited to pass "eq-iir" instead of "src" to comp_run.sh. Many of the measurements done for SRC can be applied for any processing. I will upstream dedicated EQ test scripts later to check that the measured response matches theoretical response defined by the blob.

lgirdwood · 2020-03-17T17:02:26Z

SOFCI TEST

lgirdwood · 2020-03-17T17:03:06Z

@singalsu can you follow up here with the CI results. It seems merge is blocked on CI atm.

This patch adds to testbench capability to run pipelines with EQ_IIR and EQ_FIR components. The component is configured with the topology embedded configuration blob when instantiated with new(). The load of volume component has been enhanced to parse volume tokens and compute the min and max volume parameters. The topology parsing and component library load has been changed to support processing component load and extracting of control private data. The testbench build has been changed to create libraries for EQ components. The testbench command line parsing has been improved to return error for invalid library request. Signed-off-by: Seppo Ingalsuo <seppo.ingalsuo@linux.intel.com>

This patch fixes the read_samples_32() function in file component. Signed-off-by: Seppo Ingalsuo <seppo.ingalsuo@linux.intel.com>

This patch adds compilation with -g. Optimization -O was added to get optimization that preserves sane debugging with gdb. This component is not critical for speed in simulation. Signed-off-by: Seppo Ingalsuo <seppo.ingalsuo@linux.intel.com>

The src_run.sh is converted into a generic component run script that is called by new simple component specific scripts asrc_run.sh, eqfir_run.sh, eqiir_run.sh, src_run.sh, and volume_run.sh. Note: The testbench does not yet have fully working capture support so the direction is not exposed yet. Signed-off-by: Seppo Ingalsuo <seppo.ingalsuo@linux.intel.com>

This patch fixes the buffer to define as PIPELINE_SINK. The upper level topologies do not use these topologies so there has not been related errors. However test pipelines were impacted. Also the PIPELINE_PCM defined name is made more generic. It can be whatever and not just high-pass type. Signed-off-by: Seppo Ingalsuo <seppo.ingalsuo@linux.intel.com>

singalsu · 2020-03-19T15:55:47Z

The patch check reports some things those are not relevant for testbench

Does not understand "FILE *handle";
There's no memcpy_s() in C library
Two long strings lines in shell scripts for usage
One long shell script line define the topology file name to use. The test topology file names are rather long.

The sof-ci/jenkins/pr-device-test is not yet completed

lgirdwood · 2020-03-19T22:28:14Z

SOFCI TEST

singalsu · 2020-03-20T12:27:54Z

@lgirdwood Should I push this PR again to get jenkins test executed? I think only the CI checkpatch report is valid.

lgirdwood · 2020-03-20T13:28:53Z

jenkins CI - known issues.
internal CI - unrelated test.

lgirdwood · 2020-03-20T13:33:14Z

@wwittbrx fyi I've very occasionally seen this error. I'm pretty sure it cant be related to this ?

2020-03-19 23:30:17=> Configure SOF
2020-03-19 23:30:17     => Environment initialization
2020-03-19 23:30:17     => Loading base firmware
2020-03-19 23:30:18=> Enabling DMA trace buffer
2020-03-19 23:30:18ITERATION TestPtpComplexStreamOperations24000Hz24b32b4ch.test === 1584657018.4064174
2020-03-19 23:30:18=> TestPtpComplexStreamOperations24000Hz24b32b4ch.test
2020-03-19 23:30:19     => Verify SSP ports:
2020-03-19 23:30:19     => Generate playback file
2020-03-19 23:30:20     => Topology iteration 0
2020-03-19 23:30:20                 => Create pipelines:
2020-03-19 23:30:20                 => Build topology
2020-03-19 23:30:20                             => Build and connect components of Playback Pipeline (id: 1)
2020-03-19 23:30:20                             => Build and connect components of Capture Pipeline (id: 2)
2020-03-19 23:30:20                             => Build Playback Pipeline (id: 1)
2020-03-19 23:30:20                             => Build Capture Pipeline (id: 2)
2020-03-19 23:30:20                             => Connect pipelines
2020-03-19 23:30:20                             => Complete Playback Pipeline (id: 1)
2020-03-19 23:30:20                             => Complete Capture Pipeline (id: 2)
2020-03-19 23:30:20                 => Configure audio interfaces
2020-03-19 23:30:20                             => Interface type: SSP, port index: 0 
2020-03-19 23:30:20                 => Playback iteration 0
2020-03-19 23:30:20                             => Create playback streams (0)
2020-03-19 23:30:20                             => Create capture streams (0)
2020-03-19 23:30:20                             => Start playback and capture streams (0)
2020-03-19 23:30:26                             => Stop playback streams (0)
2020-03-19 23:30:26                 => Free components
2020-03-19 23:30:26     => Verify captured stream 0
2020-03-19 23:30:26                 Expression check failed:
                                    Iter 0, checking if time of inter recorded in interval 2, on channel 1.
                                    Expected:	1.02
                                    Got:		0.00
2020-03-19 23:30:26

singalsu requested review from jajanusz, lgirdwood and ranj063 March 9, 2020 15:24

singalsu requested review from dbaluta, lbetlej, mmaka1, plbossart and tlauda as code owners March 9, 2020 15:24

singalsu commented Mar 9, 2020

View reviewed changes

tools/topology/sof/pipe-eq-iir-capture.m4 Outdated Show resolved Hide resolved

lgirdwood reviewed Mar 9, 2020

View reviewed changes

singalsu force-pushed the testbench_load_process branch from a934b18 to a1006c3 Compare March 16, 2020 17:38

singalsu changed the title ~~[RFC] Testbench add process components support~~ Testbench add process components support Mar 16, 2020

lgirdwood approved these changes Mar 16, 2020

View reviewed changes

singalsu added 5 commits March 19, 2020 13:52

Testbench: Fix possible uninitialized use of variable ret

784f1bb

This patch fixes the read_samples_32() function in file component. Signed-off-by: Seppo Ingalsuo <seppo.ingalsuo@linux.intel.com>

singalsu force-pushed the testbench_load_process branch from a1006c3 to 426b313 Compare March 19, 2020 11:52

lgirdwood merged commit 26ea50e into thesofproject:master Mar 20, 2020

cujomalainey mentioned this pull request Mar 20, 2020

[RFC] audio: Add DC Blocking Filter Component #2462

Merged

singalsu deleted the testbench_load_process branch September 15, 2022 13:14

Conversation

singalsu commented Mar 9, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

lgirdwood left a comment

Choose a reason for hiding this comment

Uh oh!

lgirdwood Mar 9, 2020

Choose a reason for hiding this comment

Uh oh!

singalsu Mar 16, 2020

Choose a reason for hiding this comment

Uh oh!

singalsu commented Mar 16, 2020

Uh oh!

sebcarlucci commented Mar 17, 2020

Uh oh!

singalsu commented Mar 17, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

lgirdwood commented Mar 17, 2020

Uh oh!

lgirdwood commented Mar 17, 2020

Uh oh!

singalsu commented Mar 19, 2020

Uh oh!

lgirdwood commented Mar 19, 2020

Uh oh!

singalsu commented Mar 20, 2020

Uh oh!

lgirdwood commented Mar 20, 2020

Uh oh!

lgirdwood commented Mar 20, 2020

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

singalsu commented Mar 9, 2020 •

edited

Loading

singalsu commented Mar 17, 2020 •

edited

Loading