Re: [TAPSETS] Linux Kernel Event Trace Tool

["Guanglei Li" <guanglei at cn dot ibm dot com>]

> If the person is interested into tracing the whole system (process 
> events,
> softirqs, interrupts and NMI) to see the root of a problem, then a 
> "event lost"
> will happen as soon as an NMI tries to reenter into the logging 
> code. That will
> leave an incomplete trace.
>
>  Mathieu Desnoyers <compudj@krystal.dyndns.org>

> FChE
> Not exactly.  Those functions could still in theory be instrumented,
> but an actual run-time reentry into the probing system would be
> detected, the probe short-circuited, and a "missed probe" would be
> recorded.

So how about using semaphore in probe handler or spin lock in 
_stp_sprintf. So that no event will be lost. The current 
implementation will just abandon the processing of a probe handler if 
there's already another probe handler running, even on a different 
CPU.

> Mathieu Desnoyers
> This will happend on systems where preemption is enabled if you do 
> not disable
> preemption. You could save the cpu_id() once for the current logging, 
> and then
> keep the same until the end. The side-effect is that, sometimes, a 
> CPU will
> write in another cpu's buffer.

But I have no way to pass cpuid into _stp_printf. So for my tapsets, 
is there a good way to solve multiple calls to _stp_printf besides 
print all stuffs using a single call of _stp_printf?