Re: [RFC] tracing: Improve design of preemptirq tracepoints and its users

[Steven Rostedt]

On Mon, 5 Feb 2018 23:15:32 -0800
Joel Fernandes <joelaf@xxxxxxxxxx> wrote:

> On Mon, Feb 5, 2018 at 6:39 PM, Joel Fernandes <joelaf@xxxxxxxxxx> wrote:
> > This patch detaches the preemptirq tracepoints from the tracers and
> > keeps it separate. With this, several ifdefs are cleaner, and lockdep
> > and other users can use the preemptirq tracepoints by registering probes
> > onto them. This makes it much cleaner, but not just that: PROVE_LOCKING
> > and CONFIG_IRQSOFF_TRACER can be enabled now at the same time and we
> > get rid of all the horrific ifdeferry around PROVE_LOCKING.
> >  
> 
> Sorry for the misstatement, irqsoff tracer and PROVE_LOCKING can still
> be used at the same time (without this patch). I will correct the
> commit message in the next rev. However this patch avoids the
> complication that arises when enabling the 2 together, by eliminating
> time_hardirqs_on/off.
> 

Hi Joel,

This looks nice, but I currently don't have time to look at it. Please
ping me again if you don't hear from me by beginning of next week.

-- Steve