Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-23 Thread Ingo Molnar
* Dave Young wrote: > On 01/23/14 at 05:56pm, Peter Zijlstra wrote: > > On Thu, Jan 23, 2014 at 10:10:28AM +0800, Dave Young wrote: > > > Hmm, seems the my physical machine is booting fine with this patch. kvm > > > guest problem still exist, but that kvm thing might be other problem. > > > > D

Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-23 Thread Dave Young
On 01/23/14 at 05:56pm, Peter Zijlstra wrote: > On Thu, Jan 23, 2014 at 10:10:28AM +0800, Dave Young wrote: > > Hmm, seems the my physical machine is booting fine with this patch. kvm > > guest problem still exist, but that kvm thing might be other problem. > > Dave, could you try with tip/master,

Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-23 Thread Peter Zijlstra
On Thu, Jan 23, 2014 at 10:10:28AM +0800, Dave Young wrote: > Hmm, seems the my physical machine is booting fine with this patch. kvm > guest problem still exist, but that kvm thing might be other problem. Dave, could you try with tip/master, Ingo pushed out all the fixes gathered. Lacking that,

Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-23 Thread Josh Boyer
On Thu, Jan 23, 2014 at 4:48 AM, Peter Zijlstra wrote: > On Wed, Jan 22, 2014 at 10:17:40PM +0100, Markus Trippelsdorf wrote: >> Yes. Thanks Peter. >> > > Ah much simpler patch that should have the same effect: This fixes the issue on my baremetal i7 machine as well. josh > --- > Subject: sched

Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-23 Thread Peter Zijlstra
On Thu, Jan 23, 2014 at 11:01:00AM +0100, Markus Trippelsdorf wrote: > On 2014.01.23 at 10:48 +0100, Peter Zijlstra wrote: > > On Wed, Jan 22, 2014 at 10:17:40PM +0100, Markus Trippelsdorf wrote: > > > > Ah much simpler patch that should have the same effect: > > Yes. FWIW it also seems to be fin

Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-23 Thread Markus Trippelsdorf
On 2014.01.23 at 10:48 +0100, Peter Zijlstra wrote: > On Wed, Jan 22, 2014 at 10:17:40PM +0100, Markus Trippelsdorf wrote: > > Ah much simpler patch that should have the same effect: Yes. FWIW it also seems to be fine. -- Markus -- To unsubscribe from this list: send the line "unsubscribe linux

Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-23 Thread Peter Zijlstra
On Wed, Jan 22, 2014 at 10:17:40PM +0100, Markus Trippelsdorf wrote: > Yes. Thanks Peter. > Ah much simpler patch that should have the same effect: --- Subject: sched/x86/tsc: Initialize multiplier to 0 From: Peter Zijlstra Date: Wed, 22 Jan 2014 22:08:14 +0100 Since we keep the clock value li

Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-22 Thread Dave Young
On 01/23/14 at 09:53am, Dave Young wrote: > On 01/22/14 at 10:08pm, Peter Zijlstra wrote: > > > > > > I think its the right region to look through. My current suspect is the > > > linear continuity fit with the initial 'random' multiplier. > > > > > > That initial 'random' multiplier can get us q

Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-22 Thread Dave Young
On 01/22/14 at 10:08pm, Peter Zijlstra wrote: > > > > I think its the right region to look through. My current suspect is the > > linear continuity fit with the initial 'random' multiplier. > > > > That initial 'random' multiplier can get us quite high, and we'll fit > > the function to match tha

Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-22 Thread Dave Young
On 01/22/14 at 12:59pm, Peter Zijlstra wrote: > On Wed, Jan 22, 2014 at 11:45:32AM +0100, Peter Zijlstra wrote: > > Ho humm. > > OK, so I had me a ponder; does the below fix things for you and David? > I've only done a boot test on real proper hardware :-) > > --- > kernel/sched/clock.c | 42 +++

Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-22 Thread Josh Boyer
On Wed, Jan 22, 2014 at 4:08 PM, Peter Zijlstra wrote: >> >> I think its the right region to look through. My current suspect is the >> linear continuity fit with the initial 'random' multiplier. >> >> That initial 'random' multiplier can get us quite high, and we'll fit >> the function to match t

Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-22 Thread Sasha Levin
On 01/22/2014 12:14 PM, Peter Zijlstra wrote: On Tue, Jan 21, 2014 at 05:28:37PM -0500, Sasha Levin wrote: [0.00] Initmem setup node 30 [mem 0x12ee00-0x138dff] [0.00] NODE_DATA [mem 0xcfa42000-0xcfa72fff] [0.00] NODE_DATA(30) on node

Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-22 Thread Markus Trippelsdorf
On 2014.01.22 at 22:08 +0100, Peter Zijlstra wrote: > > > > I think its the right region to look through. My current suspect is the > > linear continuity fit with the initial 'random' multiplier. > > > > That initial 'random' multiplier can get us quite high, and we'll fit > > the function to mat

Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-22 Thread Peter Zijlstra
> > I think its the right region to look through. My current suspect is the > linear continuity fit with the initial 'random' multiplier. > > That initial 'random' multiplier can get us quite high, and we'll fit > the function to match that but continue at a sane rate. > > I'll try and prod a li

Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-22 Thread Peter Zijlstra
On Wed, Jan 22, 2014 at 08:12:54PM +0100, Markus Trippelsdorf wrote: > On 2014.01.22 at 20:09 +0100, Markus Trippelsdorf wrote: > > On 2014.01.22 at 19:42 +0100, Peter Zijlstra wrote: > > > On Wed, Jan 22, 2014 at 07:35:38PM +0100, Markus Trippelsdorf wrote: > > > > > > > >FYI it happens on real h

Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-22 Thread Josh Boyer
On Wed, Jan 22, 2014 at 1:45 PM, Markus Trippelsdorf wrote: > On 2014.01.22 at 19:42 +0100, Peter Zijlstra wrote: >> On Wed, Jan 22, 2014 at 07:35:38PM +0100, Markus Trippelsdorf wrote: >> >> > >FYI it happens on real hardware on my machine: >> >> > >[ 60.375384] process: using AMD E400 aware id

Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-22 Thread Markus Trippelsdorf
On 2014.01.22 at 20:09 +0100, Markus Trippelsdorf wrote: > On 2014.01.22 at 19:42 +0100, Peter Zijlstra wrote: > > On Wed, Jan 22, 2014 at 07:35:38PM +0100, Markus Trippelsdorf wrote: > > > > > >FYI it happens on real hardware on my machine: > > > > > >[ 60.375384] process: using AMD E400 aware

Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-22 Thread Markus Trippelsdorf
On 2014.01.22 at 19:42 +0100, Peter Zijlstra wrote: > On Wed, Jan 22, 2014 at 07:35:38PM +0100, Markus Trippelsdorf wrote: > > > >FYI it happens on real hardware on my machine: > > > >[ 60.375384] process: using AMD E400 aware idle routine > > > But this is a different issue. I've bisected it

Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-22 Thread Markus Trippelsdorf
On 2014.01.22 at 19:42 +0100, Peter Zijlstra wrote: > On Wed, Jan 22, 2014 at 07:35:38PM +0100, Markus Trippelsdorf wrote: > > > >FYI it happens on real hardware on my machine: > > > >[ 60.375384] process: using AMD E400 aware idle routine > > > But this is a different issue. I've bisected it

Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-22 Thread Peter Zijlstra
On Wed, Jan 22, 2014 at 07:35:38PM +0100, Markus Trippelsdorf wrote: > >FYI it happens on real hardware on my machine: > >[ 60.375384] process: using AMD E400 aware idle routine > But this is a different issue. I've bisected it to: > > commit 20d1c86a57762f0a33a78988e3fc8818316badd4 > Author:

Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-22 Thread Markus Trippelsdorf
On 2014.01.22 at 09:26 -0500, Sasha Levin wrote: > On 01/22/2014 08:14 AM, Markus Trippelsdorf wrote: > > On 2014.01.22 at 13:30 +0100, Peter Zijlstra wrote: > >> >On Wed, Jan 22, 2014 at 01:26:09PM +0100, Markus Trippelsdorf wrote: > >>> > >On 2014.01.22 at 13:07 +0100, Peter Zijlstra wrote: > >>>

Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-22 Thread Peter Zijlstra
On Tue, Jan 21, 2014 at 05:28:37PM -0500, Sasha Levin wrote: > [0.00] Initmem setup node 30 [mem 0x12ee00-0x138dff] > [0.00] NODE_DATA [mem 0xcfa42000-0xcfa72fff] > [0.00] NODE_DATA(30) on node 1 > [0.00] Initmem setup node 31 [m

Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-22 Thread Sasha Levin
On 01/22/2014 08:14 AM, Markus Trippelsdorf wrote: On 2014.01.22 at 13:30 +0100, Peter Zijlstra wrote: >On Wed, Jan 22, 2014 at 01:26:09PM +0100, Markus Trippelsdorf wrote: > >On 2014.01.22 at 13:07 +0100, Peter Zijlstra wrote: > > >On Wed, Jan 22, 2014 at 01:00:48PM +0100, Markus Trippelsdorf

Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-22 Thread Markus Trippelsdorf
On 2014.01.22 at 13:30 +0100, Peter Zijlstra wrote: > On Wed, Jan 22, 2014 at 01:26:09PM +0100, Markus Trippelsdorf wrote: > > On 2014.01.22 at 13:07 +0100, Peter Zijlstra wrote: > > > On Wed, Jan 22, 2014 at 01:00:48PM +0100, Markus Trippelsdorf wrote: > > > > FYI it happens on real hardware on my

Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-22 Thread Peter Zijlstra
On Wed, Jan 22, 2014 at 01:26:09PM +0100, Markus Trippelsdorf wrote: > On 2014.01.22 at 13:07 +0100, Peter Zijlstra wrote: > > On Wed, Jan 22, 2014 at 01:00:48PM +0100, Markus Trippelsdorf wrote: > > > FYI it happens on real hardware on my machine: > > > ... > > > [0.00] Hierarchical RCU im

Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-22 Thread Markus Trippelsdorf
On 2014.01.22 at 13:07 +0100, Peter Zijlstra wrote: > On Wed, Jan 22, 2014 at 01:00:48PM +0100, Markus Trippelsdorf wrote: > > FYI it happens on real hardware on my machine: > > ... > > [0.00] Hierarchical RCU implementation. > > [0.00] NR_IRQS:4352 nr_irqs:712 16 > > [0.00]

Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-22 Thread Peter Zijlstra
On Wed, Jan 22, 2014 at 01:07:57PM +0100, Peter Zijlstra wrote: > On Wed, Jan 22, 2014 at 01:00:48PM +0100, Markus Trippelsdorf wrote: > > FYI it happens on real hardware on my machine: > > ... > > [0.00] Hierarchical RCU implementation. > > [0.00] NR_IRQS:4352 nr_irqs:712 16 > > [

Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-22 Thread Peter Zijlstra
On Wed, Jan 22, 2014 at 01:00:48PM +0100, Markus Trippelsdorf wrote: > FYI it happens on real hardware on my machine: > ... > [0.00] Hierarchical RCU implementation. > [0.00] NR_IRQS:4352 nr_irqs:712 16 > [0.00] spurious 8259A interrupt: IRQ7. > [0.00] Console: colou

Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-22 Thread Markus Trippelsdorf
On 2014.01.22 at 11:45 +0100, Peter Zijlstra wrote: > On Tue, Jan 21, 2014 at 05:28:37PM -0500, Sasha Levin wrote: > > On 12/12/2013 09:08 AM, Peter Zijlstra wrote: > > > > This patch seems to be causing an issue with booting a KVM guest. It seems > > that it > > causes the time to go random duri

Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-22 Thread Peter Zijlstra
On Wed, Jan 22, 2014 at 11:45:32AM +0100, Peter Zijlstra wrote: > Ho humm. OK, so I had me a ponder; does the below fix things for you and David? I've only done a boot test on real proper hardware :-) --- kernel/sched/clock.c | 42 +- 1 file changed, 33 in

Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-22 Thread Peter Zijlstra
On Tue, Jan 21, 2014 at 05:28:37PM -0500, Sasha Levin wrote: > On 12/12/2013 09:08 AM, Peter Zijlstra wrote: > > This patch seems to be causing an issue with booting a KVM guest. It seems > that it > causes the time to go random during early boot process: > > [0.00] Initmem setup n

Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable

2014-01-21 Thread Sasha Levin
On 12/12/2013 09:08 AM, Peter Zijlstra wrote: In order to avoid the runtime condition and variable load turn sched_clock_stable into a static_key. Also provide a shorter implementation of local_clock() and cpu_clock(int) when sched_clock_stable==1. MAINLINE PRE

[PATCH 13/15] sched: Use a static_key for sched_clock_stable

2013-12-12 Thread Peter Zijlstra
In order to avoid the runtime condition and variable load turn sched_clock_stable into a static_key. Also provide a shorter implementation of local_clock() and cpu_clock(int) when sched_clock_stable==1. MAINLINE PRE POST sched_clock_stable: 1 1