On Thu, Aug 22, 2013 at 08:08:09PM +0800, Wanpeng Li wrote:
>On Wed, Aug 21, 2013 at 10:19:52PM -0700, Yinghai Lu wrote:
>>On Wed, Aug 21, 2013 at 12:29 AM, Wanpeng Li <liwanp@linux.vnet.ibm.com> wrote:
>>> Hi Yinghai,
>>> On Tue, Aug 20, 2013 at 09:28:29PM -0700, Yinghai Lu wrote:
>>>>On Tue, Aug 20, 2013 at 8:11 PM, Wanpeng Li <liwanp@linux.vnet.ibm.com> wrote:
>>>>> Hi Yinghai,
>>>>> On Tue, Aug 20, 2013 at 05:02:17PM -0700, Yinghai Lu wrote:
>>>>>>>> -     /* ok, last chunk */
>>>>>>>> -     sparse_early_usemaps_alloc_node(usemap_map, pnum_begin, NR_MEM_SECTIONS,
>>>>>>>> -                                      usemap_count, nodeid_begin);
>>>>>>>> +     alloc_usemap_and_memmap(usemap_map, true);
>>>>>>
>>>>>>alloc_usemap_and_memmap() is somehow confusing.
>>>>>>
>>>>>>Please check if you can pass function pointer instead of true/false.
>>>>>>
>>>>>
>>>>> sparse_early_usemaps_alloc_node and sparse_early_mem_maps_alloc_node is
>>>>> similar, however, one has a parameter unsigned long ** and the other has
>>>>> struct page **. function pointer can't help, isn't it? ;-)
>>>>
>>>>you could have one generic function pointer like
>>>>void *alloc_func(void *data);
>>>>
>>>>and in the every alloc function, have own struct data to pass in/out...
>>>>
>

Sorry send you the wrong version, how about this one?