Re: [Xen-devel] [PATCH 1/2] page-alloc/x86: don't restrict DMA heap to node 0

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 11.08.16 at 11:53, <julien.grall@xxxxxxx> wrote:
> Hi Jan,
> 
> On 10/08/2016 11:23, Jan Beulich wrote:
>> --- a/xen/include/asm-arm/numa.h
>> +++ b/xen/include/asm-arm/numa.h
>> @@ -17,6 +17,11 @@ static inline __attribute__((pure)) node
>>  #define node_start_pfn(nid) (pdx_to_pfn(frametable_base_pdx))
>>  #define __node_distance(a, b) (20)
>>
>> +static inline unsigned int arch_get_dma_bitsize(void)
>> +{
>> +    return 32;
>> +}
>> +
> 
> I am not sure why we return 32 here for ARM. Anyway, as it was already 
> the case before this patch:

In fact I think we could get away without that inline function for the
time being: Since NODES_SHIFT is undefined on ARM, the call out of
page_alloc.c should get elided by the compiler anyway, so a possible
alternative would be to just have a declaration of the function on
ARM without actual implementation.

> Acked-by: Julien Grall <julien.grall@xxxxxxx>

Thanks.

Jan


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel