Re: ftrace performance (sched events): cyclictest shows 25% more latency

2018-07-09 Thread Steven Rostedt
On Fri, 6 Jul 2018 08:22:01 +0200 Claudio wrote: > Hello all, > > I have been experimenting with the idea of leaving ftrace enabled, with sched > events, > on production systems. > > The main concern that I am having at the moment is about the impact on the > system. > Enabling the sched

Re: ftrace performance (sched events): cyclictest shows 25% more latency

2018-07-09 Thread Steven Rostedt
On Fri, 6 Jul 2018 08:22:01 +0200 Claudio wrote: > Hello all, > > I have been experimenting with the idea of leaving ftrace enabled, with sched > events, > on production systems. > > The main concern that I am having at the moment is about the impact on the > system. > Enabling the sched

Re: ftrace performance (sched events): cyclictest shows 25% more latency

2018-07-09 Thread Steven Rostedt
On Mon, 9 Jul 2018 16:53:52 +0200 Claudio wrote: > > One additional data point, based on brute force again: > > I applied this change, in order to understand if it was the > > trace_event_raw_event_* (I suppose primarily trace_event_raw_event_switch) > > that contained the latency

Re: ftrace performance (sched events): cyclictest shows 25% more latency

2018-07-09 Thread Steven Rostedt
On Mon, 9 Jul 2018 16:53:52 +0200 Claudio wrote: > > One additional data point, based on brute force again: > > I applied this change, in order to understand if it was the > > trace_event_raw_event_* (I suppose primarily trace_event_raw_event_switch) > > that contained the latency

Re: ftrace performance (sched events): cyclictest shows 25% more latency

2018-07-09 Thread Claudio
On 07/09/2018 12:06 PM, Claudio wrote: > Hello Steve, > > thank you for your response, > > On 07/06/2018 11:24 PM, Steven Rostedt wrote: >> On Fri, 6 Jul 2018 08:22:01 +0200 >> Claudio wrote: >> >>> Hello all, >>> >>> I have been experimenting with the idea of leaving ftrace enabled, with

Re: ftrace performance (sched events): cyclictest shows 25% more latency

2018-07-09 Thread Claudio
On 07/09/2018 12:06 PM, Claudio wrote: > Hello Steve, > > thank you for your response, > > On 07/06/2018 11:24 PM, Steven Rostedt wrote: >> On Fri, 6 Jul 2018 08:22:01 +0200 >> Claudio wrote: >> >>> Hello all, >>> >>> I have been experimenting with the idea of leaving ftrace enabled, with

Re: ftrace performance (sched events): cyclictest shows 25% more latency

2018-07-09 Thread Claudio
Hello Steve, thank you for your response, On 07/06/2018 11:24 PM, Steven Rostedt wrote: > On Fri, 6 Jul 2018 08:22:01 +0200 > Claudio wrote: > >> Hello all, >> >> I have been experimenting with the idea of leaving ftrace enabled, with >> sched events, >> on production systems. >> >> The main

Re: ftrace performance (sched events): cyclictest shows 25% more latency

2018-07-09 Thread Claudio
Hello Steve, thank you for your response, On 07/06/2018 11:24 PM, Steven Rostedt wrote: > On Fri, 6 Jul 2018 08:22:01 +0200 > Claudio wrote: > >> Hello all, >> >> I have been experimenting with the idea of leaving ftrace enabled, with >> sched events, >> on production systems. >> >> The main

Re: ftrace performance (sched events): cyclictest shows 25% more latency

2018-07-06 Thread Steven Rostedt
On Fri, 6 Jul 2018 17:39:22 -0400 Steven Rostedt wrote: > On Fri, 6 Jul 2018 17:24:28 -0400 > Steven Rostedt wrote: > > > I'll investigate further. > > Note, I enabled function tracing to trace the scheduler function: > > # trace-cmd start -p function -l schedule > > And then ran

Re: ftrace performance (sched events): cyclictest shows 25% more latency

2018-07-06 Thread Steven Rostedt
On Fri, 6 Jul 2018 17:39:22 -0400 Steven Rostedt wrote: > On Fri, 6 Jul 2018 17:24:28 -0400 > Steven Rostedt wrote: > > > I'll investigate further. > > Note, I enabled function tracing to trace the scheduler function: > > # trace-cmd start -p function -l schedule > > And then ran

Re: ftrace performance (sched events): cyclictest shows 25% more latency

2018-07-06 Thread Steven Rostedt
On Fri, 6 Jul 2018 17:24:28 -0400 Steven Rostedt wrote: > I'll investigate further. Note, I enabled function tracing to trace the scheduler function: # trace-cmd start -p function -l schedule And then ran cyclictest. That does not cause any noticeable increase in latency (try it). So there

Re: ftrace performance (sched events): cyclictest shows 25% more latency

2018-07-06 Thread Steven Rostedt
On Fri, 6 Jul 2018 17:24:28 -0400 Steven Rostedt wrote: > I'll investigate further. Note, I enabled function tracing to trace the scheduler function: # trace-cmd start -p function -l schedule And then ran cyclictest. That does not cause any noticeable increase in latency (try it). So there

Re: ftrace performance (sched events): cyclictest shows 25% more latency

2018-07-06 Thread Steven Rostedt
On Fri, 6 Jul 2018 08:22:01 +0200 Claudio wrote: > Hello all, > > I have been experimenting with the idea of leaving ftrace enabled, with sched > events, > on production systems. > > The main concern that I am having at the moment is about the impact on the > system. > Enabling the sched

Re: ftrace performance (sched events): cyclictest shows 25% more latency

2018-07-06 Thread Steven Rostedt
On Fri, 6 Jul 2018 08:22:01 +0200 Claudio wrote: > Hello all, > > I have been experimenting with the idea of leaving ftrace enabled, with sched > events, > on production systems. > > The main concern that I am having at the moment is about the impact on the > system. > Enabling the sched

ftrace performance (sched events): cyclictest shows 25% more latency

2018-07-06 Thread Claudio
Hello all, I have been experimenting with the idea of leaving ftrace enabled, with sched events, on production systems. The main concern that I am having at the moment is about the impact on the system. Enabling the sched events that I currently need for the tracing application seems to slow

ftrace performance (sched events): cyclictest shows 25% more latency

2018-07-06 Thread Claudio
Hello all, I have been experimenting with the idea of leaving ftrace enabled, with sched events, on production systems. The main concern that I am having at the moment is about the impact on the system. Enabling the sched events that I currently need for the tracing application seems to slow