tracing: relax trace_event_eval_update() execution with cond_resched()
authorClément Léger <cleger@rivosinc.com>
Fri, 29 Sep 2023 19:16:37 +0000 (21:16 +0200)
committerSteven Rostedt (Google) <rostedt@goodmis.org>
Sat, 30 Sep 2023 20:24:55 +0000 (16:24 -0400)
commit23cce5f25491968b23fb9c399bbfb25f13870cd9
tree38c3459a5628bd3eed67f9bfd2f061e5503ba7fc
parent1e0cb399c7653462d9dadf8ab9425337c355d358
tracing: relax trace_event_eval_update() execution with cond_resched()

When kernel is compiled without preemption, the eval_map_work_func()
(which calls trace_event_eval_update()) will not be preempted up to its
complete execution. This can actually cause a problem since if another
CPU call stop_machine(), the call will have to wait for the
eval_map_work_func() function to finish executing in the workqueue
before being able to be scheduled. This problem was observe on a SMP
system at boot time, when the CPU calling the initcalls executed
clocksource_done_booting() which in the end calls stop_machine(). We
observed a 1 second delay because one CPU was executing
eval_map_work_func() and was not preempted by the stop_machine() task.

Adding a call to cond_resched() in trace_event_eval_update() allows
other tasks to be executed and thus continue working asynchronously
like before without blocking any pending task at boot time.

Link: https://lore.kernel.org/linux-trace-kernel/20230929191637.416931-1-cleger@rivosinc.com
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Signed-off-by: Clément Léger <cleger@rivosinc.com>
Tested-by: Atish Patra <atishp@rivosinc.com>
Reviewed-by: Atish Patra <atishp@rivosinc.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
kernel/trace/trace_events.c