2009/1/14 Chris Mason :
> On Wed, 2009-01-14 at 12:18 +0100, Dmitry Adamushko wrote:
>> 2009/1/14 Chris Mason :
>> > On Tue, 2009-01-13 at 18:21 +0100, Peter Zijlstra wrote:
>> >> On Tue, 2009-01-13 at 08:49 -0800, Linus Torvalds wrote:
>> >> >
>> >> > So do a v10, and ask people to test.
>> >>
>>
On Wed, 2009-01-14 at 12:18 +0100, Dmitry Adamushko wrote:
> 2009/1/14 Chris Mason :
> > On Tue, 2009-01-13 at 18:21 +0100, Peter Zijlstra wrote:
> >> On Tue, 2009-01-13 at 08:49 -0800, Linus Torvalds wrote:
> >> >
> >> > So do a v10, and ask people to test.
> >>
> >> ---
> >> Subject: mutex: imple
On Wed, 2009-01-14 at 07:43 -0800, Linus Torvalds wrote:
>
> On Wed, 14 Jan 2009, Ingo Molnar wrote:
>
> >
> > * Chris Mason wrote:
> >
> > > v10 is better that not spinning, but its in the 5-10% range. So, I've
> > > been trying to find ways to close the gap, just to understand exactly
> >
On Wed, 14 Jan 2009, Ingo Molnar wrote:
>
> * Chris Mason wrote:
>
> > v10 is better that not spinning, but its in the 5-10% range. So, I've
> > been trying to find ways to close the gap, just to understand exactly
> > where it is different.
> >
> > If I take out:
> > /*
> > * If
* Chris Mason wrote:
> v10 is better that not spinning, but its in the 5-10% range. So, I've
> been trying to find ways to close the gap, just to understand exactly
> where it is different.
>
> If I take out:
> /*
>* If there are pending waiters, join them.
>*/
>
2009/1/14 Chris Mason :
> On Tue, 2009-01-13 at 18:21 +0100, Peter Zijlstra wrote:
>> On Tue, 2009-01-13 at 08:49 -0800, Linus Torvalds wrote:
>> >
>> > So do a v10, and ask people to test.
>>
>> ---
>> Subject: mutex: implement adaptive spinning
>> From: Peter Zijlstra
>> Date: Mon Jan 12 14:01:4
On Tue, 2009-01-13 at 18:21 +0100, Peter Zijlstra wrote:
> On Tue, 2009-01-13 at 08:49 -0800, Linus Torvalds wrote:
> >
> > So do a v10, and ask people to test.
>
> ---
> Subject: mutex: implement adaptive spinning
> From: Peter Zijlstra
> Date: Mon Jan 12 14:01:47 CET 2009
>
> Change mutex con
* Linus Torvalds wrote:
>
>
> On Tue, 13 Jan 2009, Ingo Molnar wrote:
> >
> > below is the v8 -> v10 delta patch - for all who'd like to review the
> > changes.
>
> Hmm. This does seem to indicate that v10 lost many of the preempt
> changes. Probably because Peter went back to v7 to create
On Tue, 13 Jan 2009, Ingo Molnar wrote:
>
> below is the v8 -> v10 delta patch - for all who'd like to review the
> changes.
Hmm. This does seem to indicate that v10 lost many of the preempt changes.
Probably because Peter went back to v7 to create it.
I also wonder about the timing numbers
* Peter Zijlstra wrote:
> On Tue, 2009-01-13 at 08:49 -0800, Linus Torvalds wrote:
> >
> > So do a v10, and ask people to test.
below is the v8 -> v10 delta patch - for all who'd like to review the
changes.
Ingo
>
>From d154179e2d4d4667bcbf22920eeab563bc042e6a Mon Sep 17
* Linus Torvalds wrote:
> On Tue, 13 Jan 2009, Ingo Molnar wrote:
> >
> > And v8 is rock solid in all my testing - and i'll give v10 a similar
> > workout as well.
>
> The differences between v8 and v10 are very fundamental, since v8 does
> the spinning inside the spinlock'ed loop (the spinni
On Tue, 13 Jan 2009, Ingo Molnar wrote:
>
> And v8 is rock solid in all my testing - and i'll give v10 a similar
> workout as well.
The differences between v8 and v10 are very fundamental, since v8 does the
spinning inside the spinlock'ed loop (the spinning itself is not inside
the spinlock,
* Linus Torvalds wrote:
>
>
> On Tue, 13 Jan 2009, Peter Zijlstra wrote:
> >
> > Ok, tested only 1, but that was the one I remember lockups from -- and
> > that seems to be good with the cmpxchg.
> >
> > Do you fancy me sending v10 or will you make that change locally?
>
> I'd like to get t
On Tue, 2009-01-13 at 08:49 -0800, Linus Torvalds wrote:
>
> So do a v10, and ask people to test.
---
Subject: mutex: implement adaptive spinning
From: Peter Zijlstra
Date: Mon Jan 12 14:01:47 CET 2009
Change mutex contention behaviour such that it will sometimes busy wait on
acquisition - movi
On Tue, 13 Jan 2009, Peter Zijlstra wrote:
>
> Ok, tested only 1, but that was the one I remember lockups from -- and
> that seems to be good with the cmpxchg.
>
> Do you fancy me sending v10 or will you make that change locally?
I'd like to get this in, but I'm not going to apply it in any ca
On Tue, 2009-01-13 at 17:21 +0100, Peter Zijlstra wrote:
> On Tue, 2009-01-13 at 08:16 -0800, Linus Torvalds wrote:
> >
> > On Tue, 13 Jan 2009, Peter Zijlstra wrote:
> > >
> > > Change mutex contention behaviour such that it will sometimes busy wait on
> > > acquisition - moving its behaviour cl
* Peter Zijlstra wrote:
> > Now you're forcing the slow-path on unlock. Maybe it was intentional,
> > maybe it wasn't. Did you perhaps mean
> >
> > if (atomic_cmpxchg(&lock->count, 1, 0) == 1) {
> >
> > here? I thought we agreed it was safe, if only because it should be
> > equivalent to
On Tue, 2009-01-13 at 08:16 -0800, Linus Torvalds wrote:
>
> On Tue, 13 Jan 2009, Peter Zijlstra wrote:
> >
> > Change mutex contention behaviour such that it will sometimes busy wait on
> > acquisition - moving its behaviour closer to that of spinlocks.
>
> Okey, dokey. Looks reasonable, but I
On Tue, 13 Jan 2009, Peter Zijlstra wrote:
>
> Change mutex contention behaviour such that it will sometimes busy wait on
> acquisition - moving its behaviour closer to that of spinlocks.
Okey, dokey. Looks reasonable, but I wonder if this part came from v8 and
wasn't intentional:
> +
Subject: mutex: implement adaptive spinning
From: Peter Zijlstra
Date: Mon Jan 12 14:01:47 CET 2009
Change mutex contention behaviour such that it will sometimes busy wait on
acquisition - moving its behaviour closer to that of spinlocks.
This concept got ported to mainline from the -rt tree, wh
20 matches
Mail list logo
