AW: AW: Re: AW: Re: Xen 3.4 strange behaviour as compared to Xen 3.3.1

AW: AW: Re: AW: Re: Xen 3.4 strange behaviour as compared to Xen 3.3.1

[Carsten Schiers]

Kevin, thanks, that fixed the issue of the ondemand govenor with cpufreq=dom0-kernel.

Keir, the other very strange issues (higher load when being the first DomU started, blocking
of /etc/init.d/xendomains when a certain DomU is first), I will try to investigate further 
and post again when I have more information, so that this theat is not getting more complex.

BR,
Carsten.


----- Originalnachricht -----
Von: "Tian, Kevin" <kevin.tian@intel.com>
Gesendet: Mit, 3.6.2009 02:49
An: Carsten Schiers <carsten@schiers.de> ; keir.fraser <keir.fraser@eu.citrix.com> ; xen-devel <xen-devel@lists.xensource.com> ; mark.langsdorf@amd.com
Betreff: RE: AW: Re: AW: Re: [Xen-devel] Xen 3.4 strange behaviour as compared	to Xen 3.3.1

>From: Carsten Schiers
>Sent: 2009年6月3日 0:18
>
>Sorry, I think it could also be that the CPU mask is somehow 
>modified in the
>hypercall itself. Too much code for me to realy understand. 
>
>Just a guess, but does Changeset 18898 take care of 
>cpufreq=dom0-kernel case? It's
>patching the hypercall 52 and the cpufreq_ondemand.c of Xen 
>near a modfication of
>a CPU mask variable, but no change in the cpufreq_ondemand.c 
>of the Dom0 kernel. 
>It also seems to fit into the time window, doesn't it?
>

That's really a silly typo. Could you try whether below fixes for you?

diff -r e8b74e981bfb xen/arch/x86/platform_hypercall.c
--- a/xen/arch/x86/platform_hypercall.c Tue Jun 02 18:58:09 2009 +0800
+++ b/xen/arch/x86/platform_hypercall.c Tue Jun 02 18:59:18 2009 +0800
@@ -313,7 +313,6 @@ ret_t do_platform_op(XEN_GUEST_HANDLE(xe
     {
         uint32_t cpu;
         uint64_t idletime, now = NOW();
-        struct vcpu *v;
         struct xenctl_cpumap ctlmap;
         cpumask_t cpumap;
         XEN_GUEST_HANDLE(uint8) cpumap_bitmap;
@@ -336,7 +335,7 @@ ret_t do_platform_op(XEN_GUEST_HANDLE(xe

         for_each_cpu_mask ( cpu, cpumap )
         {
-            if ( (v = idle_vcpu[cpu]) != NULL )
+            if ( !idle_vcpu[cpu] )
                 cpu_clear(cpu, cpumap);
             idletime = get_cpu_idle_time(cpu);

Thanks
Kevin

Re: AW: AW: Re: AW: Re: Xen 3.4 strange behaviour as compared to Xen 3.3.1

[Keir Fraser]

On 03/06/2009 09:31, "Carsten Schiers" <carsten@schiers.de> wrote:

> Kevin, thanks, that fixed the issue of the ondemand govenor with
> cpufreq=dom0-kernel.
> 
> Keir, the other very strange issues (higher load when being the first DomU
> started, blocking
> of /etc/init.d/xendomains when a certain DomU is first), I will try to
> investigate further
> and post again when I have more information, so that this theat is not getting
> more complex.

Thanks Carsten. And good hunch on the changeset there. I completely missed
the inverted test.

 -- Keir