Revert "trace: Kconfig: disable filtering by default"

[plbossart]

This reverts commit 9fadef7.

After multiple trials on a CometLake SoundWire device, this revert to
bring the trace back to what it was seems to be the only solution, the
suggested PR thesofproject/linux#3166 does not
help on this SoundWire device.

We had similar issues with SD offset timeouts and a similar revert
with #4578 at the end of
July, there's something that we are missing on what the trace does and
how it impacts the DMA handling.

BugLink: #4779
Signed-off-by: Pierre-Louis Bossart pierre-louis.bossart@linux.intel.com

[marc-hb]

The DMA trace code is complex and brittle. In the future let's please review and test every change very thoroughly to stop wasting considerable amounts of each other's time. Major test and development disruptions aside, polluting the git history with experiments is also not helping; among others it makes simple bisects impossible.

I broke it in commit 57ee04f , causing regression #4558. Reverted in #4578

Then #4636 caused regressions #4676 and #4699. Reverted in #4760.

Now this important, double default configuration change which was hidden in an unrelated PR: #4548 (comment)

[marc-hb]

The failure signature in https://sof-ci.01.org/sof-pr-viewer/#/build/PR4785/build7329111 seems new, it didn't show in other recent PRs:

22:28:21.372       => Create playback and capture streams
22:28:21.403       => Start playback and capture streams
22:28:22.716   => Exception in test.
22:28:22.716     Exception TimeoutError occurred at 'C:\\work\\tools\\pyexecnetcache\\cavs_scripts-py\\utilities\\hda_streamer.py':594
                 Play stream not running for 1 seconds

All TestHdaLoopback* tests on WHL_HDA failed like this (everything else is green in that run)

Note there have been a few other DMA related changes recently, non exhaustive list:

9fadef7 trace: Kconfig: disable filtering by default (this)
ed5d2fe dma-trace: send ipc msg after update host pointer (mediatek)
509eef9 byt:align DMA mapping with ...
7e042e9 dma: bdw and byt dont have DMIC so remove...

[plbossart]

The failure signature in https://sof-ci.01.org/sof-pr-viewer/#/build/PR4785/build7329111 seems new, it didn't show in other recent PRs:

22:28:21.372       => Create playback and capture streams
22:28:21.403       => Start playback and capture streams
22:28:22.716   => Exception in test.
22:28:22.716     Exception TimeoutError occurred at 'C:\\work\\tools\\pyexecnetcache\\cavs_scripts-py\\utilities\\hda_streamer.py':594
                 Play stream not running for 1 seconds

All TestHdaLoopback* tests on WHL_HDA failed like this (everything else is green in that run)

Note there have been a few other DMA related changes recently, non exhaustive list:

9fadef7 trace: Kconfig: disable filtering by default
ed5d2fe dma-trace: send ipc msg after update host pointer (mediatek)
509eef9 byt:align DMA mapping with ...
7e042e9 dma: bdw and byt dont have DMIC so remove...

we're nowhere near a 1.9 release with such issues...

[marc-hb]

we're nowhere near a 1.9 release with such issues...

Note 1.9 has been branched before all the commits I listed above.

[marc-hb]

The failure signature in https://sof-ci.01.org/sof-pr-viewer/#/build/PR4785/build7329111 seems new, it didn't show in other recent PRs:
All TestHdaLoopback* tests on WHL_HDA failed like this (everything else is green in that run)

I re-submitted this in duplicate TEST PR #4787 just to re-run the tests. In https://sof-ci.01.org/sof-pr-viewer/#/build/PR4787/build7333298 all tests passed on WHL_HDA. Go figure.

I checked that the main branch did not move between this 4785 and my 4787 duplicate. In both cases tests ran on "Merge de2b7d1 into 4de6627 "

EDIT: For some reason 4785 was re-run in https://sof-ci.01.org/sof-pr-viewer/#/build/PR4785/build7333470 and it's all green.

[lgirdwood]

ok, this is simple way to validate the issue. We can re-enable when the successor to thesofproject/linux#3166 is merged.