On Wed, 7 Jan 2026 11:19:35 -0500
Steven Rostedt <[email protected]> wrote:

> On Wed, 7 Jan 2026 11:17:09 -0500
> Steven Rostedt <[email protected]> wrote:
> 
> > Or we simply change it to:
> > 
> > static inline void  
> 
> Actually, the above should be noinline, as it's in a slower path, and
> should not be adding logic into the cache of the fast path.

However, to be honest, I'm surprized this is considered slow path. My
use case is to record a few selected trace events with "trace-cmd
record", which spends most time polling trace_pipe_raw. Consequently,
there is almost always a pending waiter that requires a wakeup.

In short, irq_work_queue() is the hot path for me.

OTOH I don't mind making it noinline, because on recent Intel and AMD
systems, a function call (noinline) is often cheaper than an increase
in L1 cache footprint (caused by inlining). But I'm confused. I have
always thought most people use tracing same way as I do.

> > rb_irq_work_queue(struct rb_irq_work *irq_work)
> > {
> >     int cpu;
> > 
> >     /* irq_work_queue_on() is not allowed in NMI context */
> >     if (in_nmi()) {
> >             irq_work_queue(&irq_work->work, cpu);
> >             return;
> >     }

Thanks for the idea. There are some downsides. IIUC there is no
fundamental reason IPIs to other CPUs cannot be sent from NMI context.
It's just a limitation of the current Linux kernel code. As such, it
may be lifted in the future, and at that point nobody will remember to
remove this condition.

My current plan is it to keep the patch on hold and have a look why IPI
backends are not NMI-safe. In fact, I'm not even 100% sure the comment
is correct. The issue may have fixed itself e.g. by removing the last
affected architecture. ;-)

Petr T

Reply via email to