Re: [PATCH 0/8] tracing vs rcu vs nmi

[Peter Zijlstra]

On Wed, Feb 12, 2020 at 09:32:34AM -0500, Steven Rostedt wrote:
> I've been trying to get rid of notrace around the kernel, not add more.

As is abundantly clear, I hope, we really need more this time.

You want to trace RCU itself, the concequence is we _MUST_NOT_ trace the
early NMI code until we can distinguish we're *in* NMI code.

There really is no 2 way about it; and I'm not going to wait until some
poor sucker trips over it.