... Does this obsolete Ingo's "sched: make cpu_clock() globally synchronous"? > > http://git.kernel.org/?p=linux/ ... . We should probably come up with some clock lib that builds the various clocks: rq->clock and cpu_clock() from sched_clock() depending on ...

linux-kernel - Peter Zijlstra - May 2 2008 - 16:46

... > > OTOH, Andi said he was working on a fastish global sched_clock() thingy, > > Andi got a link to that code? > > ... architectures to signify how much help they need building the various clocks. Sparc, power and s390 would not need any help. -- To ...

linux-kernel - Peter Zijlstra - May 2 2008 - 18:07

... said he was working on a fastish global sched_clock() > > > thingy, Andi got a link to ... based profilers a bit iffy if you're > interested in where the system spends time ... cpu scheduling should use a time based clock, whereas smp load balancing can take the ...

linux-kernel - Peter Zijlstra - May 4 2008 - 08:12

... code on archs like power64, sparc64 and s390x > which have stable clocks. It just seems wrong to do all this on these > ... don't have much choice on x86. All of this scheduler runqueue clock business is probably solving the x86 TSC issue on the wrong ...

linux-kernel - David Miller - May 2 2008 - 18:00

... tick_timestamp's to > one another.. probably sucks without it.. > > OTOH, Andi said he was working on a fastish global sched_clock() thingy, > Andi got a link to that code? While I'm fine with this kind of stuff being added to constantly cope ...

linux-kernel - David Miller - May 2 2008 - 17:48

... > wrote: > Yeah, I was thinking of a way for architectures to signify how much > help they need building the various clocks. Sparc, power and s390 > would not need any help. there's already such a mechanism in sched-devel.git (and has been there ...

linux-kernel - Ingo Molnar - May 3 2008 - 04:28

... : > > > Yeah, I was thinking of a way for architectures to signify how much > > help they need building the various clocks. Sparc, power and s390 > > would not need any help. > > there's already such a mechanism in sched-devel.git (and has been ...

linux-kernel - David Miller - May 3 2008 - 05:05

... only. (recent example: inlining optimizations ;-) And you are totally right about sched_clock() being dead on accurate an globally synchronous on sparc64 - and you are right to find _any_ issue about it a nuisance. I totally envy you that sparc64 ...

linux-kernel - Ingo Molnar - May 3 2008 - 06:10

... a.p.zijlstra@chello.nl> wrote: > Let me try and come up with something slightly less horrible Does this obsolete Ingo's "sched: make cpu_clock() globally synchronous"? http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h= ...

linux-kernel - Guillaume Chazarain - May 2 2008 - 16:38

... > > OTOH, Andi said he was working on a fastish global sched_clock() > > thingy, Andi got a link to that code? > > ... either; this makes cycle based profilers a bit iffy if you're interested in where the system spends time rather than where it spends ...

linux-kernel - Arjan van de Ven - May 2 2008 - 06:09

From: Ingo Molnar Date: Sat, 3 May 2008 12:10:16 +0200 > ( although please note that the growing generalization that goes on > _did_ find a subtle nohz problem on sparc64 early in the merge window, > so it's not like these

linux-kernel - David Miller - May 3 2008 - 15:27

From: Ingo Molnar Date: Sat, 3 May 2008 12:10:16 +0200 > The bus-to-phys complication which is an identity on x86. Etc. etc. And you guys have IOMMUs now, aren't you glad all of this nice DMA API infrastructure exists already and

linux-kernel - David Miller - May 3 2008 - 15:28

* David Miller wrote: > From: Ingo Molnar > Date: Sat, 3 May 2008 12:10:16 +0200 > > > ( although please note that the growing generalization that goes on > > _did_ find a subtle nohz problem on sparc64

linux-kernel - Ingo Molnar - May 3 2008 - 15:37

* David Miller wrote: > From: Ingo Molnar > Date: Sat, 3 May 2008 12:10:16 +0200 > > > The bus-to-phys complication which is an identity on x86. Etc. etc. > > And you guys have IOMMUs now, aren't you glad

linux-kernel - Ingo Molnar - May 3 2008 - 15:39

On Sat, 2008-05-03 at 12:27 -0700, David Miller wrote: > > ( although please note that the growing generalization that goes on > > _did_ find a subtle nohz problem on sparc64 early in the merge > window, > > so it's not like these changes are

linux-kernel - Benjamin Herrenschmidt - May 3 2008 - 18:30

* Benjamin Herrenschmidt wrote: > > On Sat, 2008-05-03 at 12:27 -0700, David Miller wrote: > > > ( although please note that the growing generalization that goes on > > > _did_ find a subtle nohz problem on sparc64

linux-kernel - Ingo Molnar - May 3 2008 - 18:38

From: Ingo Molnar Date: Sun, 4 May 2008 00:38:44 +0200 > the same bug existed on x86 too: that too does a few IPIs without > irq_enter/irq_exit. We now removed the softlockup dependency so it > should not be required to do an

linux-kernel - David Miller - May 3 2008 - 19:04

From: David Miller Date: Sat, 03 May 2008 16:04:08 -0700 (PDT) > From: Ingo Molnar > Date: Sun, 4 May 2008 00:38:44 +0200 > > > the same bug existed on x86 too: that too does a few IPIs without > > irq_enter

linux-kernel - David Miller - May 3 2008 - 19:36

* David Miller wrote: > From: David Miller > Date: Sat, 03 May 2008 16:04:08 -0700 (PDT) > > > From: Ingo Molnar > > Date: Sun, 4 May 2008 00:38:44 +0200 > > > > > the same bug

linux-kernel - Ingo Molnar - May 3 2008 - 19:38

From: Ingo Molnar Date: Sun, 4 May 2008 01:38:25 +0200 > ah, good! It really looks like unnecessary overhead for the "simple > IPIs". But ... i'm wondering ... what about softirq processing? Do these > IPIs process softirqs on the

linux-kernel - David Miller - May 3 2008 - 19:40