a23c05fd76
The timerlat tracer provides an interface for any application to wait for the timerlat's periodic wakeup. Currently, rtla timerlat uses it to dispatch its user-space workload (-u option). But as the tracer interface is generic, rtla timerlat can also be used to monitor any workload that uses it. For example, a user might place their own workload to wait on the tracer interface, and monitor the results with rtla timerlat. Add the -U option to rtla timerlat top and hist. With this option, rtla timerlat will not dispatch its workload but only setting up the system, waiting for a user to dispatch its workload. The sample code in this patch is an example of python application that loops in the timerlat tracer fd. To use it, dispatch: # rtla timerlat -U In a terminal, then run the python program on another terminal, specifying the CPU to run it. For example, setting on CPU 1: #./timerlat_load.py 1 Then rtla timerlat will start printing the statistics of the ./timerlat_load.py app. An interesting point is that the "Ret user Timer Latency" value is the overall response time of the load. The sample load does a memory copy to exemplify that. The stop tracing options on rtla timerlat works in this setup as well, including auto analysis. Link: https://lkml.kernel.org/r/36e6bcf18fe15c7601048fd4c65aeb193c502cc8.1707229706.git.bristot@kernel.org Cc: Jonathan Corbet <corbet@lwn.net> Cc: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Daniel Bristot de Oliveira <bristot@kernel.org>
42 lines
1.5 KiB
ReStructuredText
42 lines
1.5 KiB
ReStructuredText
**-a**, **--auto** *us*
|
|
|
|
Set the automatic trace mode. This mode sets some commonly used options
|
|
while debugging the system. It is equivalent to use **-T** *us* **-s** *us*
|
|
**-t**. By default, *timerlat* tracer uses FIFO:95 for *timerlat* threads,
|
|
thus equilavent to **-P** *f:95*.
|
|
|
|
**-p**, **--period** *us*
|
|
|
|
Set the *timerlat* tracer period in microseconds.
|
|
|
|
**-i**, **--irq** *us*
|
|
|
|
Stop trace if the *IRQ* latency is higher than the argument in us.
|
|
|
|
**-T**, **--thread** *us*
|
|
|
|
Stop trace if the *Thread* latency is higher than the argument in us.
|
|
|
|
**-s**, **--stack** *us*
|
|
|
|
Save the stack trace at the *IRQ* if a *Thread* latency is higher than the
|
|
argument in us.
|
|
|
|
**--dma-latency** *us*
|
|
Set the /dev/cpu_dma_latency to *us*, aiming to bound exit from idle latencies.
|
|
*cyclictest* sets this value to *0* by default, use **--dma-latency** *0* to have
|
|
similar results.
|
|
|
|
**-u**, **--user-threads**
|
|
|
|
Set timerlat to run without a workload, and then dispatches user-space workloads
|
|
to wait on the timerlat_fd. Once the workload is awakes, it goes to sleep again
|
|
adding so the measurement for the kernel-to-user and user-to-kernel to the tracer
|
|
output.
|
|
|
|
**-U**, **--user-load**
|
|
|
|
Set timerlat to run without workload, waiting for the user to dispatch a per-cpu
|
|
task that waits for a new period on the tracing/osnoise/per_cpu/cpu$ID/timerlat_fd.
|
|
See linux/tools/rtla/sample/timerlat_load.py for an example of user-load code.
|