https://bugzilla.suse.com/show_bug.cgi?id=1220119 https://bugzilla.suse.com/show_bug.cgi?id=1220119#c28 Jiri Wiesner <jwiesner@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Attachment #872947|0 |1 is obsolete| | --- Comment #28 from Jiri Wiesner <jwiesner@suse.com> --- Created attachment 873455 --> https://bugzilla.suse.com/attachment.cgi?id=873455&action=edit test 02 debug scripts, scheduling latency with ftrace The full context switch trace showed that I made a mistake in my approach. The scheduling latency reported by the test01 scripts is not real. Measuring scheduling latency between the sched/sched_waking and raw_syscalls/sys_exit tracepoints is a simplification that cannot be applied to qemu processes that spend a long time in kernel mode (running the code executed by the vCPU) after being worked up. I have started the test02 scripts on both openqaworker21 and openqaworker24. -- You are receiving this mail because: You are on the CC list for the bug.