Re: [PATCH] QRCU with lockless fastpath

2007-02-25 Thread Oleg Nesterov
On 02/24, Paul E. McKenney wrote: > > This is an updated version of Oleg Nesterov's QRCU that avoids the > earlier lock acquisition on the synchronize_qrcu() fastpath. This passes > rcutorture on x86 and the weakly ordered POWER. A promela model of the > code passes as noted before for 2

Re: [PATCH] QRCU with lockless fastpath

2007-02-25 Thread Oleg Nesterov
On 02/24, Paul E. McKenney wrote: This is an updated version of Oleg Nesterov's QRCU that avoids the earlier lock acquisition on the synchronize_qrcu() fastpath. This passes rcutorture on x86 and the weakly ordered POWER. A promela model of the code passes as noted before for 2 readers and

[PATCH] QRCU with lockless fastpath

2007-02-24 Thread Paul E. McKenney
Hello! This is an updated version of Oleg Nesterov's QRCU that avoids the earlier lock acquisition on the synchronize_qrcu() fastpath. This passes rcutorture on x86 and the weakly ordered POWER. A promela model of the code passes as noted before for 2 readers and 3 updaters and for 3 readers

[PATCH] QRCU with lockless fastpath

2007-02-24 Thread Paul E. McKenney
Hello! This is an updated version of Oleg Nesterov's QRCU that avoids the earlier lock acquisition on the synchronize_qrcu() fastpath. This passes rcutorture on x86 and the weakly ordered POWER. A promela model of the code passes as noted before for 2 readers and 3 updaters and for 3 readers