Messages in this thread | | | Date | Mon, 14 Nov 2011 09:37:16 -0500 | From | Konrad Rzeszutek Wilk <> | Subject | Re: [Xen-devel] [PATCH 1/3] cpuidle: If disable_cpuidle() is called, set pm_idle to default_idle. |
| |
On Sun, Nov 13, 2011 at 01:00:15AM -0500, Len Brown wrote: > > And just found that there is a BZ for this as well: > > https://bugzilla.redhat.com/show_bug.cgi?id=739499 > > > > And this patch fixes the Linux kernel to boot under Amazon EC2. > > doesn't matter. > > Working around an Amazon EC2 bug in a newly compiled upstream kernel > isn't going to help with any of the kernels that don't include that workaround.
Prior to 3.1 we did have a solution for this - we would use the default_idle and not pick and/choose one based on the CPUID flags.
It was not really a choice made by "hypervisor wrongly advertises the MWAIT flag", but rather that we want to use the safe_halt, which ends up calling the yield hypercall.
That is the optimal solution irregardless of what version of hypervisor is being used.
> > Amazon EC2 should not advertise MWAIT support that it does not have, > and all kernels should run on it w/o any workaround. > > Len Brown, Intel Open Source Technology Center > > _______________________________________________ > Xen-devel mailing list > Xen-devel@lists.xensource.com > http://lists.xensource.com/xen-devel
| |