Re: [PATCH 4/5] x86,smp: keep spinlock delay values per hashed spinlock address

2013-01-10 Thread Michel Lespinasse
On Thu, Jan 10, 2013 at 5:05 AM, Rik van Riel wrote: > Eric, > > with just patches 1-3, can you still reproduce the > regression on your system? > > In other words, could we get away with dropping the > complexity of patch 4, or do we still need it? To be clear, I must say that I'm not opposing p

Re: [PATCH 4/5] x86,smp: keep spinlock delay values per hashed spinlock address

2013-01-10 Thread Rik van Riel
On 01/10/2013 08:01 AM, Michel Lespinasse wrote: On Tue, Jan 8, 2013 at 2:31 PM, Rik van Riel wrote: From: Eric Dumazet Eric Dumazet found a regression with the first version of the spinlock backoff code, in a workload where multiple spinlocks were contended, each having a different wait time

Re: [PATCH 4/5] x86,smp: keep spinlock delay values per hashed spinlock address

2013-01-10 Thread Michel Lespinasse
On Tue, Jan 8, 2013 at 2:31 PM, Rik van Riel wrote: > From: Eric Dumazet > > Eric Dumazet found a regression with the first version of the spinlock > backoff code, in a workload where multiple spinlocks were contended, > each having a different wait time. > > This patch has multiple delay values

Re: [PATCH 4/5] x86,smp: keep spinlock delay values per hashed spinlock address

2013-01-09 Thread Rafael Aquini
On Tue, Jan 08, 2013 at 05:31:19PM -0500, Rik van Riel wrote: > From: Eric Dumazet > > Eric Dumazet found a regression with the first version of the spinlock > backoff code, in a workload where multiple spinlocks were contended, > each having a different wait time. > > This patch has multiple de

[PATCH 4/5] x86,smp: keep spinlock delay values per hashed spinlock address

2013-01-08 Thread Rik van Riel
From: Eric Dumazet Eric Dumazet found a regression with the first version of the spinlock backoff code, in a workload where multiple spinlocks were contended, each having a different wait time. This patch has multiple delay values per cpu, indexed on a hash of the lock address, to avoid that pro

Re: [RFC PATCH 4/5] x86,smp: keep spinlock delay values per hashed spinlock address

2013-01-03 Thread Eric Dumazet
On Thu, 2013-01-03 at 04:48 -0800, Michel Lespinasse wrote: > On Wed, Jan 2, 2013 at 9:24 PM, Rik van Riel wrote: > > From: Eric Dumazet > > > > Eric Dumazet found a regression with the spinlock backoff code, > > in workloads where multiple spinlocks were contended, each having > > a different wa

Re: [RFC PATCH 4/5] x86,smp: keep spinlock delay values per hashed spinlock address

2013-01-03 Thread Michel Lespinasse
On Wed, Jan 2, 2013 at 9:24 PM, Rik van Riel wrote: > From: Eric Dumazet > > Eric Dumazet found a regression with the spinlock backoff code, > in workloads where multiple spinlocks were contended, each having > a different wait time. I think you should really clarify that the regression was obse

[RFC PATCH 4/5] x86,smp: keep spinlock delay values per hashed spinlock address

2013-01-02 Thread Rik van Riel
From: Eric Dumazet Eric Dumazet found a regression with the spinlock backoff code, in workloads where multiple spinlocks were contended, each having a different wait time. This patch has multiple delay values per cpu, indexed on a hash of the lock address, to avoid that problem. Eric Dumazet wr