[PATCH v19 2/7] crash: add generic infrastructure for crash hotplug support

[Eric DeVolder]

On 3/14/23 09:22, Baoquan He wrote:
> On 03/14/23 at 08:28am, Eric DeVolder wrote:
> ......
>>>> +static int crash_memhp_notifier(struct notifier_block *nb, unsigned long val, void *v)
>>>> +{
>>>> +	switch (val) {
>>>> +	case MEM_ONLINE:
>>>> +		crash_handle_hotplug_event(KEXEC_CRASH_HP_ADD_MEMORY,
>>>> +			KEXEC_CRASH_HP_INVALID_CPU);
>>>> +		break;
>>>> +
>>>> +	case MEM_OFFLINE:
>>>> +		crash_handle_hotplug_event(KEXEC_CRASH_HP_REMOVE_MEMORY,
>>>> +			KEXEC_CRASH_HP_INVALID_CPU);
>>>> +		break;
>>>> +	}
>>>> +	return NOTIFY_OK;
>>>> +}
>>>> +
>>>> +static struct notifier_block crash_memhp_nb = {
>>>> +	.notifier_call = crash_memhp_notifier,
>>>> +	.priority = 0
>>>> +};
>>>> +
>>>
>>> Because for_each_possible_cpu() is taken in
>>> crash_prepare_elf64_headers(), x86 doesn't need to respond to cpu
>>> hotplug or doesn't do anything with this patchset. This cpu part in
>>> infrastructure is only for the later powerpc usage, right?
>>
>> That is true, yes.
> 
> Given this patchset is aimed at crash hotplug on x86, while obviously it
> does't need to have the cpu hotplug support on x86 since the
> for_each_possible_cpu() adjustment. People looking into this may be
> confused if they don't follow the discussion thread of v18. Do we need
> to mention this in cover letter or somewhere else? I could miss that
> though it is has been told, please ignore this if yes.
> 
Good point, I'll update the cover letter to reflect this.
eric