Re: [PATCH v13 10/14] unwind: Clear unwind_mask on exit back to user space

[Steven Rostedt <rostedt@goodmis.org>]

On Wed, 16 Jul 2025 14:26:09 -0400
Steven Rostedt <rostedt@goodmis.org> wrote:

>   Now before the task gets back to user space, ftrace requests the
>   deferred trace. To do so, it must set the pending bit and its bit,
>   but it must also clear the perf bit as it should not call perf's
>   callback again.

After ftrace clears the bits, it is possible that the first perf
program will request again and this time it will get another callback
with the same cookie. But at least it has a request between the last
callback and the next one.

That is, it would have:

[Task enters kernel]
  request -> add cookie
  request -> add cookie
  [..]
  callback -> add trace + cookie
 [ftrace clears bits]
  request -> add cookie
  callback -> add trace + cookie
[Task exits back to user space]

Which shouldn't be too confusing. But if we just do the fetch_or and it
didn't request a new trace, it would have:

[Task enters kernel]
  request -> add cookie
  request -> add cookie
  [..]
  callback -> add trace + cookie
 [ftrace clears bits]
  callback -> add trace + cookie
[Task exits back to user space]

Where there's two callbacks written to the perf buffer for the same
request.

Maybe this isn't a problem, but I was trying to avoid adding multiple
requests due to other tracers affecting the state.

-- Steve