Re: [PATCH 1/5] tracing: Prevent kernel oops with corrupted buffer

[Steven Rostedt <rostedt@goodmis.org>]

On Tue, 2010-01-26 at 23:39 +0100, Frederic Weisbecker wrote:
> On Tue, Jan 26, 2010 at 02:32:23PM -0800, Andrew Morton wrote:
> > > +	if (WARN_ON_ONCE(pid < 0)) {
> > > +		strcpy(comm, "<XXX>");
> > > +		return;
> > > +	}
> > > +
> > >  	if (pid > PID_MAX_DEFAULT) {
> > >  		strcpy(comm, "<...>");
> > >  		return;
> > 
> > But why is it WARN_ON_ONCE()?  That will only fix the problem a single
> > time.  On the second occurrence, it will oops again.
> 
> 
> 
> The warning will be produced only once, but after that,
> the condition is still checked like a simple if:
> 
> #define WARN_ON_ONCE(condition)	({				\
> 	static bool __warned;					\
> 	int __ret_warn_once = !!(condition);			\
> 								\
> 	if (unlikely(__ret_warn_once))				\
> 		if (WARN_ON(!__warned)) 			\
> 			__warned = true;			\
> 	unlikely(__ret_warn_once);				\
> })
> 
> 
> And since this function can be called anytime we have a trace
> to print to the user, we don't want to encumber with thousands
> of warnings.

Exactly!

-- Steve