On Tue, 19 Mar 2019 at 04:07, Suzuki K Poulose suzuki.poulose@arm.com wrote:
On 15/03/2019 21:41, Mathieu Poirier wrote:
On Fri, Mar 15, 2019 at 05:24:05PM +0000, Suzuki K Poulose wrote:
Hi Mathieu,
On 06/03/2019 22:57, Mathieu Poirier wrote:
For CPU-wide scenarios trace generated by multiple CPUs can be aggregated by a single sink. This patch paves the way to uses the pid of the process being monitored to allocate and free sink buffer memory along with regimenting access to what source a sink can collect data for.
Signed-off-by: Mathieu Poirier mathieu.poirier@linaro.org
I think it may be a good idea to pass the "event" instead of the pid and the cpu (both of which are deduced from the former), separately to the alloc_buffer. What do you think ?
The problem with passing the event is that each sink needs to call task_pid_nr(), which I thought was redundant.
The problem is we are bloating the arglist of alloc_buffer() already with node, and now with pid, all of which are derived from event. Hence the thought.
I agree... It is a matter of deciding which is the lesser evil. I was ambivalent when I added the cpu and pid in the argument list... I'll just do the change.
Cheers Suzuki