[plbossart]

We have multiple bugs reporting that the DMA trace helps avoid underflows.
Let's put that to the test. of course sof-logger will not work but let's see what happens.

[plbossart]

well that was a spectacular failure across the board. Looks like sof-test is unable to deal with the case where the trace is disabled.

I am not sure however if the kernel does the right thing and enables the e-trace even when the DMA trace is not allowed. @ranj063 @kv2019i feedback welcome.

[marc-hb]

Looks like sof-test is unable to deal with the case where the trace is disabled.

By default this is considered a failure per thesofproject/sof-test#297

There is code to disable these checks but it requires some small sof-test or configuration change: thesofproject/sof-test@325e57244e2a

I am not sure however if the kernel does the right thing and enables the e-trace even when the DMA trace is not allowed.

It does, I just verified this: /sys/kernel/debug/sof/etrace is here and works even when /sys/kernel/debug/sof/trace is not.

BTW I just submitted sof_insert.sh: add support for module parameters thesofproject/sof-test#734