Re: [PATCH] tracing: reduce latency and remove percpu trace_seq

[Steven Rostedt]

On Fri, 2010-06-04 at 10:23 +0800, Lai Jiangshan wrote:
> Steven Rostedt wrote:
> 
> > 
> > The "print" in the trace_seq_printf() is the TP_printk() field of the
> > TRACE_EVENT. I believe that there were some events (ext4?) that did
> > things that required preemption disabled.
> > 
> 
> It seems that ext4 has no such events.
> I find jbd2_dev_to_name() in the TP_printk() field of ext4's tracepoints.
> And it seems that jbd2_dev_to_name() requires preemption enabled and sleepable.
> 
> Maybe there were some events did things that required preemption disabled.
> But No document ensure TP_printk() preemption disabled, such events should
> be fixed, I think.

Well, I had a private email conversation with Ted last year where IIRC,
adding the preempt disable would fix things.

I'd like to make sure Ted is OK with removing that before we do so.

Ted?

-- Steve


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/