On Wed, Feb 16, 2022 at 01:33:18PM -0500, Steven Rostedt wrote:

On Wed, 16 Feb 2022 10:09:14 -0800
Linus Torvalds [off-list ref] wrote:

quoted

Byungchul, could you fix those two issues? Some of your reports may
well be entirely valid, but the hard-to-read hex offsets and the
knowledge that at least some of them are confused about how
prepare_to_wait -> wait actually works makes the motivation to look at
the details much less..

Hi Byungchul,

I'm not sure what your tool is using to attach to the kernel to analyze the
events (perhaps tracepoints?). But you can have the prepare_to_wait event
just flag the task is about to wait, and then attach to the schedule
(sched_switch) event to denote that it actually waited.

Of course have the finish_wait() clear the flag.

(Sorry for late reply, which was because of my email client issue.)

Thank you for the hint. However, unfortunately, I didn't use tracepoint
for that. However, the key idea is the thing that I should take!

Thanks a lot!

Thanks,
Byungchul

-- Steve