Created attachment 284613 [details]
Proof of concept patch for raw format of latency printing

We should let the user see the raw flags (the "6dN.20" part) in hex. And then be able to see what the flags are. For example doing:

  trace-cmd report -l -R

Produces:

  <idle>-0       0-0x25-30 333536.173363: sched_wakeup:          comm=ksoftirqd/0 pid=9 prio=120 success=1 target_cpu=0
  rcuc/7-64      7-0x1-20 333536.173365: sched_switch:          prev_comm=rcuc/7 prev_pid=64 prev_prio=98 prev_state=1 next_comm=ksoftirqd/7 next_pid=65 next_prio=120
  <idle>-0       6-0x1-20 333536.173366: sched_switch:          prev_comm=swapper/6 prev_pid=0 prev_prio=120 prev_state=0 next_comm=rcuc/6 next_pid=57 next_prio=98
  <idle>-0       4-0x1-20 333536.173366: sched_switch:          prev_comm=swapper/4 prev_pid=0 prev_prio=120 prev_state=0 next_comm=ksoftirqd/4 next_pid=44 next_prio=120
ksoftirq-65      7-0x0-.2 333536.173366: softirq_entry:         vec=1


Instead of:

  <idle>-0       0dN.30 333536.173363: sched_wakeup:         ksoftirqd/0:9 [120] success=1 CPU:000
  rcuc/7-64      7d..20 333536.173365: sched_switch:         rcuc/7:64 [98] S ==> ksoftirqd/7:65 [120]
  <idle>-0       6d..20 333536.173366: sched_switch:         swapper/6:0 [120] R ==> rcuc/6:57 [98]
  <idle>-0       4d..20 333536.173366: sched_switch:         swapper/4:0 [120] R ==> ksoftirqd/4:44 [120]
ksoftirq-65      7....2 333536.173366: softirq_entry:        vec=1 [action=TIMER]

Attached is a prototype patch.