Re: [patch 2/4] introduce the mechanism of disabling cpu hotplug control

["Siddha, Suresh B" <suresh.b.siddha@intel.com>]

On Tue, Nov 07, 2006 at 07:54:30PM -0800, Andrew Morton wrote:
> On Tue, 7 Nov 2006 17:40:25 -0800
> "Siddha, Suresh B" <suresh.b.siddha@intel.com> wrote:
> 
> > Add 'cpu_hotplug_no_control' and when set, the hotplug control file("online")
> > will not be added under /sys/devices/system/cpu/cpuX/
> > 
> > Next patch doing PCI quirks will use this.
> > 
> 
> I don't understand what this (ugly) patch has to do with the overall
> bugfix.  We're fixing the APCI initialisation - what does that have to do
> with presenting cpu-hotplug files in sysfs?

This patch is adding a mechanism which will not export the cpu hotplug control
file.  And the quirk will use this from preventing the users doing cpu hotplug.
On these platforms, we need to have atleast 2 cpus online to workaround the
errata.

In future we can use this mechanism to disable cpu hotplug during bootup.

> But does this variable _have_ to be a negative like this?  The code would
> be simpler if it had the opposite sense and was called, say,
> cpu_hotplug_enable_control_file.

I wanted to add something like disable_cpu_hotplug.. but suspend code
seems to be already using.. Will clean this up.

> Are these patches considered 2.6.19 material?  They look a bit big, ugly
> and scary for that.

Though no one reported failures so far, it would be good to get included
in 2.6.19. If it is too late and sounds scary, we can consider in -stable
after spending sometime in -mm..

thanks,
suresh