[RFC 09/11] coresight: etm-perf: Disable the path before capturing the trace data

[Suzuki K Poulose]

On 11/10/20 12:45 PM, Anshuman Khandual wrote:
> perf handle structure needs to be shared with the TRBE IRQ handler for
> capturing trace data and restarting the handle. There is a probability
> of an undefined reference based crash when etm event is being stopped
> while a TRBE IRQ also getting processed. This happens due the release
> of perf handle via perf_aux_output_end(). This stops the sinks via the
> link before releasing the handle, which will ensure that a simultaneous
> TRBE IRQ could not happen.
> 
> Signed-off-by: Anshuman Khandual <anshuman.khandual at arm.com>
> ---
> This might cause problem with traditional sink devices which can be
> operated in both sysfs and perf mode. This needs to be addressed
> correctly. One option would be to move the update_buffer callback
> into the respective sink devices. e.g, disable().
> 
>   drivers/hwtracing/coresight/coresight-etm-perf.c | 2 ++
>   1 file changed, 2 insertions(+)
> 
> diff --git a/drivers/hwtracing/coresight/coresight-etm-perf.c b/drivers/hwtracing/coresight/coresight-etm-perf.c
> index 534e205..1a37991 100644
> --- a/drivers/hwtracing/coresight/coresight-etm-perf.c
> +++ b/drivers/hwtracing/coresight/coresight-etm-perf.c
> @@ -429,7 +429,9 @@ static void etm_event_stop(struct perf_event *event, int mode)
>   
>   		size = sink_ops(sink)->update_buffer(sink, handle,
>   					      event_data->snk_config);
> +		coresight_disable_path(path);
>   		perf_aux_output_end(handle, size);
> +		return;
>   	}

As you mentioned, this is not ideal where another session could be triggered on
the sink from a different ETM (not for per-CPU sink) in a different mode before
you collect the buffer. I believe the best option is to leave the
update_buffer() to disable_hw. This would need to pass on the "handle" to the
disable_path.

That way the races can be handled inside the sinks. Also, this aligns the
perf mode of the sinks with that of the sysfs mode.

Suzuki