Re: Re: Re: Re: Re: Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-06 Thread SeongJae Park
On Wed, 6 May 2020 07:41:51 -0700 "Paul E. McKenney" wrote: > On Wed, May 06, 2020 at 02:59:26PM +0200, SeongJae Park wrote: > > TL; DR: It was not kernel's fault, but the benchmark program. > > > > So, the problem is reproducible using the lebench[1] only. I carefully read > > it's code

Re: Re: Re: Re: Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-06 Thread Paul E. McKenney
On Wed, May 06, 2020 at 02:59:26PM +0200, SeongJae Park wrote: > TL; DR: It was not kernel's fault, but the benchmark program. > > So, the problem is reproducible using the lebench[1] only. I carefully read > it's code again. > > Before running the problem occurred "poll big" sub test, lebench

Re: Re: Re: Re: Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-06 Thread Eric Dumazet
On Wed, May 6, 2020 at 5:59 AM SeongJae Park wrote: > > TL; DR: It was not kernel's fault, but the benchmark program. > > So, the problem is reproducible using the lebench[1] only. I carefully read > it's code again. > > Before running the problem occurred "poll big" sub test, lebench executes >

Re: Re: Re: Re: Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-06 Thread SeongJae Park
TL; DR: It was not kernel's fault, but the benchmark program. So, the problem is reproducible using the lebench[1] only. I carefully read it's code again. Before running the problem occurred "poll big" sub test, lebench executes "context switch" sub test. For the test, it sets the cpu

Re: Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-06 Thread SeongJae Park
On Tue, 05 May 2020 12:00:49 -0700 (PDT) David Miller wrote: > From: David Miller > Date: Tue, 05 May 2020 11:48:25 -0700 (PDT) > > > Series applied and queued up for -stable, thanks. > > Nevermind, this doesn't even compile. > > net/smc/af_smc.c: In function ‘smc_switch_to_fallback’: >

Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-05 Thread David Miller
From: David Miller Date: Tue, 05 May 2020 11:48:25 -0700 (PDT) > Series applied and queued up for -stable, thanks. Nevermind, this doesn't even compile. net/smc/af_smc.c: In function ‘smc_switch_to_fallback’: net/smc/af_smc.c:473:19: error: ‘smc->clcsock->wq’ is a pointer; did you mean to use

Re: Re: Re: Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-05 Thread Paul E. McKenney
On Tue, May 05, 2020 at 08:34:02PM +0200, SeongJae Park wrote: > On Tue, 5 May 2020 11:17:07 -0700 "Paul E. McKenney" > wrote: > > > On Tue, May 05, 2020 at 07:56:05PM +0200, SeongJae Park wrote: > > > On Tue, 5 May 2020 10:30:36 -0700 "Paul E. McKenney" > > > wrote: > > > > > > > On Tue,

Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-05 Thread David Miller
From: SeongJae Park Date: Tue, 5 May 2020 10:10:33 +0200 > From: SeongJae Park > > The commit 6d7855c54e1e ("sockfs: switch to ->free_inode()") made the > deallocation of 'socket_alloc' to be done asynchronously using RCU, as > same to 'sock.wq'. And the following commit 333f7909a857

Re: Re: Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-05 Thread Paul E. McKenney
On Tue, May 05, 2020 at 08:40:07PM +0200, SeongJae Park wrote: > On Tue, 5 May 2020 11:27:20 -0700 "Paul E. McKenney" > wrote: > > > On Tue, May 05, 2020 at 07:49:43PM +0200, SeongJae Park wrote: > > > On Tue, 5 May 2020 10:23:58 -0700 "Paul E. McKenney" > > > wrote: > > > > > > > On Tue,

Re: Re: Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-05 Thread SeongJae Park
On Tue, 5 May 2020 11:27:20 -0700 "Paul E. McKenney" wrote: > On Tue, May 05, 2020 at 07:49:43PM +0200, SeongJae Park wrote: > > On Tue, 5 May 2020 10:23:58 -0700 "Paul E. McKenney" > > wrote: > > > > > On Tue, May 05, 2020 at 09:25:06AM -0700, Eric Dumazet wrote: > > > > > > > > > > > > On

Re: Re: Re: Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-05 Thread SeongJae Park
On Tue, 5 May 2020 11:17:07 -0700 "Paul E. McKenney" wrote: > On Tue, May 05, 2020 at 07:56:05PM +0200, SeongJae Park wrote: > > On Tue, 5 May 2020 10:30:36 -0700 "Paul E. McKenney" > > wrote: > > > > > On Tue, May 05, 2020 at 07:05:53PM +0200, SeongJae Park wrote: > > > > On Tue, 5 May 2020

Re: Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-05 Thread Paul E. McKenney
On Tue, May 05, 2020 at 07:49:43PM +0200, SeongJae Park wrote: > On Tue, 5 May 2020 10:23:58 -0700 "Paul E. McKenney" > wrote: > > > On Tue, May 05, 2020 at 09:25:06AM -0700, Eric Dumazet wrote: > > > > > > > > > On 5/5/20 9:13 AM, SeongJae Park wrote: > > > > On Tue, 5 May 2020 09:00:44

Re: Re: Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-05 Thread Paul E. McKenney
On Tue, May 05, 2020 at 07:56:05PM +0200, SeongJae Park wrote: > On Tue, 5 May 2020 10:30:36 -0700 "Paul E. McKenney" > wrote: > > > On Tue, May 05, 2020 at 07:05:53PM +0200, SeongJae Park wrote: > > > On Tue, 5 May 2020 09:37:42 -0700 Eric Dumazet > > > wrote: > > > > > > > > > > > > > >

Re: Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-05 Thread SeongJae Park
On Tue, 5 May 2020 10:28:50 -0700 "Paul E. McKenney" wrote: > On Tue, May 05, 2020 at 09:37:42AM -0700, Eric Dumazet wrote: > > > > > > On 5/5/20 9:31 AM, Eric Dumazet wrote: > > > > > > > > > On 5/5/20 9:25 AM, Eric Dumazet wrote: > > >> > > >> > > >> On 5/5/20 9:13 AM, SeongJae Park wrote:

Re: Re: Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-05 Thread SeongJae Park
On Tue, 5 May 2020 10:30:36 -0700 "Paul E. McKenney" wrote: > On Tue, May 05, 2020 at 07:05:53PM +0200, SeongJae Park wrote: > > On Tue, 5 May 2020 09:37:42 -0700 Eric Dumazet > > wrote: > > > > > > > > > > > On 5/5/20 9:31 AM, Eric Dumazet wrote: > > > > > > > > > > > > On 5/5/20 9:25

Re: Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-05 Thread SeongJae Park
On Tue, 5 May 2020 10:23:58 -0700 "Paul E. McKenney" wrote: > On Tue, May 05, 2020 at 09:25:06AM -0700, Eric Dumazet wrote: > > > > > > On 5/5/20 9:13 AM, SeongJae Park wrote: > > > On Tue, 5 May 2020 09:00:44 -0700 Eric Dumazet > > > wrote: > > > > > >> On Tue, May 5, 2020 at 8:47 AM

Re: Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-05 Thread Paul E. McKenney
On Tue, May 05, 2020 at 07:05:53PM +0200, SeongJae Park wrote: > On Tue, 5 May 2020 09:37:42 -0700 Eric Dumazet wrote: > > > > > > > On 5/5/20 9:31 AM, Eric Dumazet wrote: > > > > > > > > > On 5/5/20 9:25 AM, Eric Dumazet wrote: > > >> > > >> > > >> On 5/5/20 9:13 AM, SeongJae Park wrote: >

Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-05 Thread Paul E. McKenney
On Tue, May 05, 2020 at 09:37:42AM -0700, Eric Dumazet wrote: > > > On 5/5/20 9:31 AM, Eric Dumazet wrote: > > > > > > On 5/5/20 9:25 AM, Eric Dumazet wrote: > >> > >> > >> On 5/5/20 9:13 AM, SeongJae Park wrote: > >>> On Tue, 5 May 2020 09:00:44 -0700 Eric Dumazet > >>> wrote: > >>> >

Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-05 Thread Paul E. McKenney
On Tue, May 05, 2020 at 09:25:06AM -0700, Eric Dumazet wrote: > > > On 5/5/20 9:13 AM, SeongJae Park wrote: > > On Tue, 5 May 2020 09:00:44 -0700 Eric Dumazet wrote: > > > >> On Tue, May 5, 2020 at 8:47 AM SeongJae Park wrote: > >>> > >>> On Tue, 5 May 2020 08:20:50 -0700 Eric Dumazet > >>>

Re: Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-05 Thread SeongJae Park
On Tue, 5 May 2020 09:37:42 -0700 Eric Dumazet wrote: > > > On 5/5/20 9:31 AM, Eric Dumazet wrote: > > > > > > On 5/5/20 9:25 AM, Eric Dumazet wrote: > >> > >> > >> On 5/5/20 9:13 AM, SeongJae Park wrote: > >>> On Tue, 5 May 2020 09:00:44 -0700 Eric Dumazet > >>> wrote: > >>> > On

Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-05 Thread Eric Dumazet
On 5/5/20 9:31 AM, Eric Dumazet wrote: > > > On 5/5/20 9:25 AM, Eric Dumazet wrote: >> >> >> On 5/5/20 9:13 AM, SeongJae Park wrote: >>> On Tue, 5 May 2020 09:00:44 -0700 Eric Dumazet wrote: >>> On Tue, May 5, 2020 at 8:47 AM SeongJae Park wrote: > > On Tue, 5 May 2020 08:20:50

Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-05 Thread Eric Dumazet
On 5/5/20 9:25 AM, Eric Dumazet wrote: > > > On 5/5/20 9:13 AM, SeongJae Park wrote: >> On Tue, 5 May 2020 09:00:44 -0700 Eric Dumazet wrote: >> >>> On Tue, May 5, 2020 at 8:47 AM SeongJae Park wrote: On Tue, 5 May 2020 08:20:50 -0700 Eric Dumazet wrote: > >

Re: Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-05 Thread Al Viro
On Tue, May 05, 2020 at 09:00:44AM -0700, Eric Dumazet wrote: > > Not exactly the 10,000,000, as it is only the possible highest number, but I > > was able to observe clear exponential increase of the number of the objects > > using slabtop. Before the start of the problematic workload, the

Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-05 Thread Eric Dumazet
On 5/5/20 9:13 AM, SeongJae Park wrote: > On Tue, 5 May 2020 09:00:44 -0700 Eric Dumazet wrote: > >> On Tue, May 5, 2020 at 8:47 AM SeongJae Park wrote: >>> >>> On Tue, 5 May 2020 08:20:50 -0700 Eric Dumazet >>> wrote: >>> On 5/5/20 8:07 AM, SeongJae Park wrote: > On

Re: Re: Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-05 Thread SeongJae Park
On Tue, 5 May 2020 09:00:44 -0700 Eric Dumazet wrote: > On Tue, May 5, 2020 at 8:47 AM SeongJae Park wrote: > > > > On Tue, 5 May 2020 08:20:50 -0700 Eric Dumazet > > wrote: > > > > > > > > > > > On 5/5/20 8:07 AM, SeongJae Park wrote: > > > > On Tue, 5 May 2020 07:53:39 -0700 Eric Dumazet

Re: Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-05 Thread Eric Dumazet
On Tue, May 5, 2020 at 8:47 AM SeongJae Park wrote: > > On Tue, 5 May 2020 08:20:50 -0700 Eric Dumazet wrote: > > > > > > > On 5/5/20 8:07 AM, SeongJae Park wrote: > > > On Tue, 5 May 2020 07:53:39 -0700 Eric Dumazet > > > wrote: > > > > > > > >> Why do we have 10,000,000 objects around ?

Re: Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-05 Thread SeongJae Park
On Tue, 5 May 2020 08:20:50 -0700 Eric Dumazet wrote: > > > On 5/5/20 8:07 AM, SeongJae Park wrote: > > On Tue, 5 May 2020 07:53:39 -0700 Eric Dumazet wrote: > > > > >> Why do we have 10,000,000 objects around ? Could this be because of > >> some RCU problem ? > > > > Mainly because of a

Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-05 Thread Eric Dumazet
On 5/5/20 8:07 AM, SeongJae Park wrote: > On Tue, 5 May 2020 07:53:39 -0700 Eric Dumazet wrote: > >> Why do we have 10,000,000 objects around ? Could this be because of >> some RCU problem ? > > Mainly because of a long RCU grace period, as you guess. I have no idea how > the grace period

Re: Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-05 Thread SeongJae Park
On Tue, 5 May 2020 07:53:39 -0700 Eric Dumazet wrote: > On Tue, May 5, 2020 at 4:54 AM SeongJae Park wrote: > > > > CC-ing sta...@vger.kernel.org and adding some more explanations. > > > > On Tue, 5 May 2020 10:10:33 +0200 SeongJae Park wrote: > > > > > From: SeongJae Park > > > > > > The

Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-05 Thread Eric Dumazet
On Tue, May 5, 2020 at 4:54 AM SeongJae Park wrote: > > CC-ing sta...@vger.kernel.org and adding some more explanations. > > On Tue, 5 May 2020 10:10:33 +0200 SeongJae Park wrote: > > > From: SeongJae Park > > > > The commit 6d7855c54e1e ("sockfs: switch to ->free_inode()") made the > >

Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-05 Thread Nuernberger, Stefan
On Tue, 2020-05-05 at 13:54 +0200, SeongJae Park wrote: > CC-ing sta...@vger.kernel.org and adding some more explanations. > > On Tue, 5 May 2020 10:10:33 +0200 SeongJae Park > wrote: > > > > > From: SeongJae Park > > > > The commit 6d7855c54e1e ("sockfs: switch to ->free_inode()") made > >

Re: [PATCH net v2 0/2] Revert the 'socket_alloc' life cycle change

2020-05-05 Thread SeongJae Park
CC-ing sta...@vger.kernel.org and adding some more explanations. On Tue, 5 May 2020 10:10:33 +0200 SeongJae Park wrote: > From: SeongJae Park > > The commit 6d7855c54e1e ("sockfs: switch to ->free_inode()") made the > deallocation of 'socket_alloc' to be done asynchronously using RCU, as >