Re: CONFIG_PREEMPT_RCU in next/mmotm

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

On Sun, Aug 09, 2009 at 04:32:28AM +0100, Hugh Dickins wrote:
> On Sat, 8 Aug 2009, Paul E. McKenney wrote:
> > On Sat, Aug 08, 2009 at 08:34:03PM +0100, Hugh Dickins wrote:
> > 
> > > CONFIG_PREEMPT_TREE=y seems okay on PowerPC; and when I briefly tried
> > 
> > CONFIG_TREE_RCU=y, you mean, right?
> 
> Whoops, yes indeed.
> 
> > > My guess is that there's some other issue which is triggering
> > > the RCU disgrace, but that is just a guess.
> > 
> > The one other issue I know of is one that apparently happens only on
> > one of Ingo's machines, which suddenly decides that it need not inform
> > RCU when onlining a CPU (at boot time).
> > 
> > But you seem to be making it past boot, and I would guess that you are
> > not doing any hotplug CPU operations, right?
> 
> That's right, I'm well past boot, and don't explicitly get into CPU
> hotplug, nor using suspend/hibernate on that machine; and looking at
> the config, see I don't even have CONFIG_HOTPLUG_CPU set.
> 
> In the morning I'll check that /proc/cpuinfo really shows all four
> cpus before I start, and is still showing four cpus when it goes bad,
> in case there's nevertheless an issue with that.
> 
> Is there some particular attribute of a cpu, most relevant to its RCU
> grace status, that I can check on with a printk?

The debugfs files enabled by CONFIG_RCU_TRACE, except rcu/rcugp, which,
since it will attempt to wait for a grace period, will probably wait
forever.

> > > Config attached.  Any suggestions?
> 
> [ Of course I forgot the config, but sent it just to Paul later. ]
> 
> > 
> > /me scratches head...  Will try some more CONFIG_PREEMPT_RCU testing
> > locally.
> 
> Thanks: I'm kind of glad that it's interesting,
> though it may sink some of our time.

Well, it certainly does have me a bit confused at the moment!

							Thanx, Paul