Re: [PATCH v4, part3 11/15] mm: use a dedicated lock to protect totalram_pages and zone->managed_pages

[Jiang Liu <liuj97@gmail.com>]

On 04/08/2013 09:39 PM, Rik van Riel wrote:
> On 04/06/2013 09:55 AM, Jiang Liu wrote:
> 
>> @@ -5186,6 +5189,22 @@ early_param("movablecore", cmdline_parse_movablecore);
>>
>>   #endif /* CONFIG_HAVE_MEMBLOCK_NODE_MAP */
>>
>> +void adjust_managed_page_count(struct page *page, long count)
>> +{
>> +    bool lock = (system_state != SYSTEM_BOOTING);
>> +
>> +    /* No need to acquire the lock during boot */
>> +    if (lock)
>> +        spin_lock(&managed_page_count_lock);
>> +
>> +    page_zone(page)->managed_pages += count;
>> +    totalram_pages += count;
>> +
>> +    if (lock)
>> +        spin_unlock(&managed_page_count_lock);
>> +}
> 
> While I agree the boot code currently does not need the lock, is
> there any harm to removing that conditional?
> 
> That would simplify the code, and protect against possible future
> cleverness of initializing multiple memory things simultaneously.
> 
Hi Rik,
	Thanks for you comments.
	I'm OK with that. Acquiring/releasing the lock should be lightweight
because there shouldn't be contention during boot. Will remove the logic in
next version.
	Regards!
	Gerry