On 2026/7/14 19:24, Steven Rostedt wrote:

On Tue, 14 Jul 2026 18:16:26 +0800
Tengda Wu [off-list ref] wrote:

quoted

_Approach 3_: As you suggested earlier, build on v1 but add proper lockdep
assertions to enforce that the parser lock is held by all callers,
regardless of how the parser is used.

Approach 3 seems cleaner and more straightforward.

I thought about having a global lock as writes to tracefs are very much a
slow path, but I think Approach 3 is probably the most "correct" fix.

Yeah, let's go with that.

Thanks,

-- Steve

Thanks, Steve. I'll spin a v2 with the lockdep assertions added.

Tengda