Re: [PATCH] Remove GP_REPLAY state from rcu_sync

2019-10-07 Thread Oleg Nesterov
On 10/04, Joel Fernandes wrote: > > On Fri, Oct 04, 2019 at 05:41:03PM +0200, Oleg Nesterov wrote: > > On 10/04, Joel Fernandes (Google) wrote: > > > > > > Taking a step back, why did we intend to have > > > to wait for a new GP if another rcu_sync_exit() comes while one is still > > > in progress?

Re: [PATCH] Remove GP_REPLAY state from rcu_sync

2019-10-04 Thread kbuild test robot
Hi "Joel, I love your patch! Perhaps something to improve: [auto build test WARNING on rcu/dev] [cannot apply to v5.4-rc1 next-20191004] [if your patch is applied to the wrong git tree, please drop us a note to help improve the system. BTW, we also suggest to use '--base' option to specify the ba

Re: [PATCH] Remove GP_REPLAY state from rcu_sync

2019-10-04 Thread kbuild test robot
Hi "Joel, I love your patch! Yet something to improve: [auto build test ERROR on rcu/dev] [cannot apply to v5.4-rc1 next-20191004] [if your patch is applied to the wrong git tree, please drop us a note to help improve the system. BTW, we also suggest to use '--base' option to specify the base tre

Re: [PATCH] Remove GP_REPLAY state from rcu_sync

2019-10-04 Thread Joel Fernandes
Hi Oleg, On Fri, Oct 04, 2019 at 05:41:03PM +0200, Oleg Nesterov wrote: > On 10/04, Joel Fernandes (Google) wrote: > > > > But this is not always true if you consider the following events: > > I'm afraid I missed your point, but... > > > --> > > GP num 11

Re: [PATCH] Remove GP_REPLAY state from rcu_sync

2019-10-04 Thread Oleg Nesterov
On 10/04, Joel Fernandes (Google) wrote: > > But this is not always true if you consider the following events: I'm afraid I missed your point, but... > --> > GP num 11 2 > GP state ie px r

Re: [PATCH] Remove GP_REPLAY state from rcu_sync

2019-10-04 Thread Joel Fernandes
On Fri, Oct 04, 2019 at 10:57:41AM -0400, Joel Fernandes (Google) wrote: > From: Joel Fernandes > > Please consider this is an RFC for discussion only. Just want to discuss > why the GP_REPLAY state is needed at all. And I messed up the subject prefix, but this is *really* RFC and for discussion

[PATCH] Remove GP_REPLAY state from rcu_sync

2019-10-04 Thread Joel Fernandes (Google)
From: Joel Fernandes Please consider this is an RFC for discussion only. Just want to discuss why the GP_REPLAY state is needed at all. Here's the intention AFAICS: When rcu_sync_exit() has happened, the gp_state changes to GP_EXIT while we wait for a grace period before transitioning to GP_IDLE