On Thu, Sep 30, 2010 at 10:02:37AM -0700, Colin Cross wrote:

On Wed, Sep 29, 2010 at 4:58 AM, Kukjin Kim [off-list ref] wrote:

quoted

Kukjin Kim wrote:

quoted

From: Changhwan Youn <redacted>

This patch updates CPU hotplug implementation to reduce CPU power
consumption
and will turn off the CPU power when CPU1 is unplugged while previous CPU
hotplug
used CPU idle.
This patch removes init memory freeing code to use CPU boot code when CPU1

is

quoted

plugged-in again and adds vfp_enable() call to allow to access CP10 and

CP11.

quoted

Signed-off-by: Changhwan Youn <redacted>
Signed-off-by: Kukjin Kim <redacted>
Cc: Russell King <redacted>

Cc: Tony Lindgren <tony@atomide.com>
Cc: Colin Cross <redacted>
Cc: Erik Gilling <redacted>
Cc: Olof Johansson <redacted>

snip

quoted

quoted

+#ifndef CONFIG_HOTPLUG_CPU
? ? ? if (!machine_is_integrator() && !machine_is_cintegrator())
? ? ? ? ? ? ? totalram_pages +=

free_area(__phys_to_pfn(__pa(__init_begin)),

quoted

? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? __phys_to_pfn(__pa(__init_end)),
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? "init");
+#endif
?}

?#ifdef CONFIG_BLK_DEV_INITRD

I don't think this is necessary.  Any symbol that is necessary for CPU
hotplug should be marked with __cpuinit instead of __init.  The linker
script will put __cpuinit symbols outside the __init_begin -
__init_end section if CONFIG_HOTPLUG_CPU is enabled.

That's the theory, but not everything required for a complete restart of
a secondary CPU (iow, if it's powered off) is in the right section yet.
I'll have patches hopefully tomorrow to address this.