ASoC: SOF: reduce verbosity of IPC logs

[plbossart]

We currently log the initiation of an IPC as well at its success.

[ 3906.106987] kernel: sof-audio-pci-intel-tgl 0000:00:1f.3: ipc tx: 0x80010000: GLB_DAI_MSG: CONFIG
[ 3906.107189] kernel: sof-audio-pci-intel-tgl 0000:00:1f.3: ipc tx succeeded: 0x80010000: GLB_DAI_MSG: CONFIG

This is overkill, we already have a message thrown in case of errors
and have tracepoints enabled to check for IPC duration.

In addition, the DMA_POSITION_UPDATE for traces brings absolutely zero
information but pollutes the logs for no good reason.

[ 3906.322256] kernel: sof-audio-pci-intel-tgl 0000:00:1f.3: ipc rx: 0x90020000: GLB_TRACE_MSG: DMA_POSITION
[ 3906.322308] kernel: sof-audio-pci-intel-tgl 0000:00:1f.3: ipc rx done: 0x90020000: GLB_TRACE_MSG: DMA_POSITION
[ 3906.822261] kernel: sof-audio-pci-intel-tgl 0000:00:1f.3: ipc rx: 0x90020000: GLB_TRACE_MSG: DMA_POSITION
[ 3906.822319] kernel: sof-audio-pci-intel-tgl 0000:00:1f.3: ipc rx done: 0x90020000: GLB_TRACE_MSG: DMA_POSITION
[ 3907.822261] kernel: sof-audio-pci-intel-tgl 0000:00:1f.3: ipc rx: 0x90020000: GLB_TRACE_MSG: DMA_POSITION
[ 3907.822319] kernel: sof-audio-pci-intel-tgl 0000:00:1f.3: ipc rx done: 0x90020000: GLB_TRACE_MSG: DMA_POSITION
[ 3908.822251] kernel: sof-audio-pci-intel-tgl 0000:00:1f.3: ipc rx: 0x90020000: GLB_TRACE_MSG: DMA_POSITION
[ 3908.822309] kernel: sof-audio-pci-intel-tgl 0000:00:1f.3: ipc rx done: 0x90020000: GLB_TRACE_MSG: DMA_POSITION

Add a flag to squelch all DMA position update logs and make it the
default when the trace is enabled.

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

[marc-hb]

I just had a quick look at the mere volume of kernel logs in the test results for this PR at https://sof-ci.01.org/linuxpr/PR3625/build84/devicetest/ , taking a random, recent PR as a reference point https://sof-ci.01.org/sofpr/PR5757/build198/devicetest/

Looking at du -shc */*/dmesg.html, this PR decreased the total to 41M down from 55M. grep GLB_TRACE_MSG.*DMA_POSITION */*/dmesg.html | wc -l goes down to 40k lines from 64k lines.
These are averages; the gain is very different depending on the test and/or platform.

This may help with "hung task" issue #3387? This won't make a big results storage difference though because when adding firmware and user space logs the total is about 250M but for sure it won't hurt either.

[marc-hb]

Superseded by #3631