Re: [RFC][PATCH 2/9] module: Sanitize RCU usage and locking

2015-03-17 Thread Peter Zijlstra
On Mon, Mar 02, 2015 at 11:37:06AM -0800, Paul E. McKenney wrote: > > - /* Wait for RCU synchronizing before releasing mod->list and buglist. */ > > - synchronize_rcu(); > > + /* Wait for RCU-sched synchronizing before releasing mod->list and > > buglist. */ > > + synchronize_sched(); > >

Re: [RFC][PATCH 2/9] module: Sanitize RCU usage and locking

2015-03-02 Thread Paul E. McKenney
On Sat, Feb 28, 2015 at 10:24:49PM +0100, Peter Zijlstra wrote: > Currently the RCU usage in module is an inconsistent mess of RCU and > RCU-sched, this is broken for CONFIG_PREEMPT where synchronize_rcu() > does not imply synchronize_sched(). > > Convert everything over to RCU-sched. > > Further

Re: [RFC][PATCH 2/9] module: Sanitize RCU usage and locking

2015-03-02 Thread Peter Zijlstra
On Mon, Mar 02, 2015 at 09:46:45PM +1030, Rusty Russell wrote: > Peter Zijlstra writes: > > Currently the RCU usage in module is an inconsistent mess of RCU and > > RCU-sched, this is broken for CONFIG_PREEMPT where synchronize_rcu() > > does not imply synchronize_sched(). > > Huh? It's not "an

Re: [RFC][PATCH 2/9] module: Sanitize RCU usage and locking

2015-03-02 Thread Rusty Russell
Peter Zijlstra writes: > Currently the RCU usage in module is an inconsistent mess of RCU and > RCU-sched, this is broken for CONFIG_PREEMPT where synchronize_rcu() > does not imply synchronize_sched(). Huh? It's not "an inconsistent mess". They're all synchronize_rcu(), except one. That one i

[RFC][PATCH 2/9] module: Sanitize RCU usage and locking

2015-02-28 Thread Peter Zijlstra
Currently the RCU usage in module is an inconsistent mess of RCU and RCU-sched, this is broken for CONFIG_PREEMPT where synchronize_rcu() does not imply synchronize_sched(). Convert everything over to RCU-sched. Furthermore add lockdep asserts to all sites, because its not at all clear to me the