Re: rcu_process_callbacks irqsoff latency caused by taking spinlock with irqs disabled

2018-04-17 Thread Paul E. McKenney
On Sun, Apr 08, 2018 at 02:06:18PM -0700, Paul E. McKenney wrote: > On Sat, Apr 07, 2018 at 07:40:42AM +1000, Nicholas Piggin wrote: > > On Thu, 5 Apr 2018 08:53:20 -0700 > > "Paul E. McKenney" wrote: [ . . . ] > > > > Note that rcu doesn't show up consistently at the top, this was > > > > just

Re: rcu_process_callbacks irqsoff latency caused by taking spinlock with irqs disabled

2018-04-08 Thread Paul E. McKenney
On Sat, Apr 07, 2018 at 07:40:42AM +1000, Nicholas Piggin wrote: > On Thu, 5 Apr 2018 08:53:20 -0700 > "Paul E. McKenney" wrote: > > > On Thu, Apr 05, 2018 at 10:45:12AM +1000, Nicholas Piggin wrote: > > > On Wed, 4 Apr 2018 17:13:58 -0700 > > > "Paul E. McKenney" wrote: > > > > > > > On Thu,

Re: rcu_process_callbacks irqsoff latency caused by taking spinlock with irqs disabled

2018-04-06 Thread Nicholas Piggin
On Thu, 5 Apr 2018 08:53:20 -0700 "Paul E. McKenney" wrote: > On Thu, Apr 05, 2018 at 10:45:12AM +1000, Nicholas Piggin wrote: > > On Wed, 4 Apr 2018 17:13:58 -0700 > > "Paul E. McKenney" wrote: > > > > > On Thu, Apr 05, 2018 at 09:34:14AM +1000, Nicholas Piggin wrote: > > > > Hi Paul, > >

Re: rcu_process_callbacks irqsoff latency caused by taking spinlock with irqs disabled

2018-04-05 Thread Paul E. McKenney
On Thu, Apr 05, 2018 at 10:45:12AM +1000, Nicholas Piggin wrote: > On Wed, 4 Apr 2018 17:13:58 -0700 > "Paul E. McKenney" wrote: > > > On Thu, Apr 05, 2018 at 09:34:14AM +1000, Nicholas Piggin wrote: > > > Hi Paul, > > > > > > Just looking at latencies, and RCU showed up as one of the maximums.

Re: rcu_process_callbacks irqsoff latency caused by taking spinlock with irqs disabled

2018-04-04 Thread Nicholas Piggin
On Wed, 4 Apr 2018 17:13:58 -0700 "Paul E. McKenney" wrote: > On Thu, Apr 05, 2018 at 09:34:14AM +1000, Nicholas Piggin wrote: > > Hi Paul, > > > > Just looking at latencies, and RCU showed up as one of the maximums. > > This is a 2 socket system with (176 CPU threads). Just doing a > > `make -j

Re: rcu_process_callbacks irqsoff latency caused by taking spinlock with irqs disabled

2018-04-04 Thread Paul E. McKenney
On Thu, Apr 05, 2018 at 09:34:14AM +1000, Nicholas Piggin wrote: > Hi Paul, > > Just looking at latencies, and RCU showed up as one of the maximums. > This is a 2 socket system with (176 CPU threads). Just doing a > `make -j 352` kernel build. Got a max latency of 3ms. I don't think > that's anyth

rcu_process_callbacks irqsoff latency caused by taking spinlock with irqs disabled

2018-04-04 Thread Nicholas Piggin
Hi Paul, Just looking at latencies, and RCU showed up as one of the maximums. This is a 2 socket system with (176 CPU threads). Just doing a `make -j 352` kernel build. Got a max latency of 3ms. I don't think that's anything to worry about really, but I wanted to check the cause. # tracer: irqsof