Re: [GIT PULL rcu/urgent] Fixes for RCU regressions in 3.8

2013-01-24 Thread Ingo Molnar
* Paul E. McKenney wrote: > Hello, Ingo, > > Paul Gortmaker found some problems with no-CBs-CPU support in 3.8 > mainline, and kindly supplied fixes. The first problem is soft-lockup > complaints that can occur on systems running with no-CBs CPUs that have > long periods of time where RCU is c

[GIT PULL rcu/urgent] Fixes for RCU regressions in 3.8

2013-01-11 Thread Paul E. McKenney
Hello, Ingo, Paul Gortmaker found some problems with no-CBs-CPU support in 3.8 mainline, and kindly supplied fixes. The first problem is soft-lockup complaints that can occur on systems running with no-CBs CPUs that have long periods of time where RCU is completely idle, which can happen on some

Re: [GIT PULL rcu/urgent] Fixes for RCU regressions

2012-10-12 Thread Ingo Molnar
* Paul E. McKenney wrote: > Hello, Ingo, > > This series fixes some RCU regressions in 3.7. > > Kirill A. Shutemov found a case where synchronize_rcu() is invoked > from a CPU-hotplug notifier, which, with RCU's shiny new kthread-based > grace-period implementation, results in deadlock. The f

[GIT PULL rcu/urgent] Fixes for RCU regressions

2012-10-11 Thread Paul E. McKenney
Hello, Ingo, This series fixes some RCU regressions in 3.7. Kirill A. Shutemov found a case where synchronize_rcu() is invoked from a CPU-hotplug notifier, which, with RCU's shiny new kthread-based grace-period implementation, results in deadlock. The first commit in this series fixes this deadl