Re: [PATCH] rudimentary tracing for Classic RCU

["Paul E. McKenney" <paulmck@linux.vnet.ibm.com>]

On Thu, Oct 09, 2008 at 12:23:13PM +0200, Frédéric Weisbecker wrote:
> 2008/10/9 Paul E. McKenney <paulmck@linux.vnet.ibm.com>:
> > Hello!
> >
> > This is a tracing patch for Classic RCU, which creates an "rcu/rcucb"
> > file in debugfs.  This patch can be handy when you need to work out
> > why RCU is refusing to end the current grace period.
> >
> > Reading from the file results in something like the following:
> >
> >        rcu: cur=1129  completed=1128  np=0  s=0
> >                0,3,7
> >        rcu_bh: cur=-287  completed=-287  np=0  s=0
> >
> >        online: 0-7
> 
> 
> Hi Paul,
> 
> Why don't you use the ring-buffer tracing engine?
> You will really make your life better by putting it as a tracer in
> kernel/trace and by using the relevant API.
> That will avoid you to manage the debugfs things, the memory
> allocation, the buffer managment.....

Hello, Frédéric,

Well, one reason is that I didn't know about it.  ;-)

Does it allow the user to trigger a one-shot trace?  Right now, what
one does is:

	cat /debug/rcu/rcucb

whenever one wants to see what RCU is up to.  You really don't want to
see every new value, as that would generate hundreds of trace records
per second -- per CPU.  What does the user do with the ring-buffer
tracing enging?

							Thanx, Paul