[PATCH] kexec: disable non-boot CPUs

[Stephen Warren]

On 12/20/2012 10:36 AM, Will Deacon wrote:
> On Thu, Dec 20, 2012 at 05:21:56PM +0000, Stephen Warren wrote:
>> On 12/20/2012 03:49 AM, Will Deacon wrote:
>>> If you do manage to get this merged, please can you follow up with a patch
>>> to remove the smp_kill_cpus bits from arch/arm/kernel/smp.c please? It only
>>> exists as a hook to do exactly this and currently nobody is using it afaict.
>>
>> I originally implemented this in
>> arch/arm/kernel/process.c:machine_shutdown(), which currently is:
>>
>> void machine_shutdown(void)
>> {
>> #ifdef CONFIG_SMP
>> 	smp_send_stop();
>> #endif
>> }
>>
>> and I changed it to something like:
>>
>> void machine_shutdown(void)
>> {
>> #ifdef CONFIG_HOTPLUG_CPU
>>  	disable_nonboot_cpus();
>> #elifdef CONFIG_SMP
>> 	smp_send_stop();
>> #endif
>> }
>>
>> ... but then figured that moving it up into the core kexec code would be
>> better, so that everything always worked the same way.
> 
> Hmmm, isn't this racy: requiring the secondaries to hit idle and notice
> they're offline and call cpu_die before the primary has replace the kernel
> image?

Isn't disable_nonboot_cpus() synchronous? If not, I imagine my original
patch wasn't any better in this respect, except that the hotunplug
happened earlier, and hence reduced the likelihood of actually seeing
any such issues.

>> Anyway, the change above addresses Eric's concern about isolating the
>> change to ARM. Does that seem like a reasonable thing for the ARM code
>> to do?
> 
> I think you're better off using what we currently have and hanging your code
> off platform_cpu_kill.

OK, I'll look into that. Joseph Lo just posted patches to implement
cpu_kill() on Tegra, which was needed to fix some issues in our hotplug
code anyway. Perhaps that will remove the need for any other changes...