[PATCH 00/17] coresight: next v4.12-rc4

On Wed, 14 Jun 2017 14:22:29 +0800
Leo Yan [off-list ref] wrote:

Hi Kim,

On Tue, Jun 13, 2017 at 04:17:12PM -0500, Kim Phillips wrote:

[...]

quoted

quoted

quoted

Also, do the current juno platforms not have CPU debug modules?  I'd
like to test the new driver.

By all means - Leo pointed out the patch adding the CPU debug entries in
the DT  file for Juno.

Thanks, I applied it, was able to
toggle /sys/kernel/debug/coresight_cpu_debug/enable, but wasn't sure
how to trigger the cpu debug output...

You could trigger panic flow with command, then can see CPU debug output:
echo c > /proc/sysrq-trigger

That worked, thanks.

The coresight CPU debug documentation states:

At the runtime you can disable idle states with below methods:

Set latency request to /dev/cpu_dma_latency to disable all CPUs specific idle
states (if latency = 0uS then disable all idle states):
# echo "what_ever_latency_you_need_in_uS" > /dev/cpu_dma_latency

Disable specific CPU's specific idle state:
# echo 1 > /sys/devices/system/cpu/cpu$cpu/cpuidle/state$state/disable

So I tried two values in /dev/cpu_dma_latency and expected a difference
in the per-cpu latency files, but didn't observe it:

root at juno:~# echo 0 > /dev/cpu_dma_latency
root at juno:~# more /sys/devices/system/cpu/cpu2/cpuidle*/state*/*laten*
::::::::::::::
/sys/devices/system/cpu/cpu2/cpuidle/state0/latency
::::::::::::::
1
::::::::::::::
/sys/devices/system/cpu/cpu2/cpuidle/state1/latency
::::::::::::::
1500
::::::::::::::
/sys/devices/system/cpu/cpu2/cpuidle/state2/latency
::::::::::::::
1600
root at juno:~# echo 1500 > /dev/cpu_dma_latency
root at juno:~# more /sys/devices/system/cpu/cpu2/cpuidle*/state*/*laten*
::::::::::::::
/sys/devices/system/cpu/cpu2/cpuidle/state0/latency
::::::::::::::
1
::::::::::::::
/sys/devices/system/cpu/cpu2/cpuidle/state1/latency
::::::::::::::
1500
::::::::::::::
/sys/devices/system/cpu/cpu2/cpuidle/state2/latency
::::::::::::::
1600
root at juno:~# cat /sys/kernel/debug/coresight_cpu_debug/enable
1

How does one tell whether writes to /dev/cpu_dma_latency are effective?

Thanks,

Kim