On 02/17/15 05:08, Paul Durrant wrote:
>> -----Original Message-----
>> From: xen-devel-boun...@lists.xen.org [mailto:xen-devel-
>> boun...@lists.xen.org] On Behalf Of Don Slutz
>> Sent: 16 February 2015 23:05
>> -static void hvm_unmap_ioreq_page(struct hvm_ioreq_server *s, bool_t
>> buf)
>> +static void hvm_unmap_ioreq_page(struct hvm_ioreq_server *s, int buf)
>> {
>> - struct hvm_ioreq_page *iorp = buf ? &s->bufioreq : &s->ioreq;
>> + struct hvm_ioreq_page *iorp = NULL;
>> +
>> + switch ( buf )
>> + {
>> + case 0:
>> + iorp = &s->ioreq;
>> + break;
>> + case 1:
>> + iorp = &s->bufioreq;
>> + break;
>> + case 2:
>> + iorp = &s->vmport_ioreq;
>> + break;
>
> Now that buf is no longer a bool, could we have #defined values for the types
> rather than magic numbers?
>
Sure.
>> + }
>> + ASSERT(iorp);
>>
>> destroy_ring_for_helper(&iorp->va, iorp->page);
>> }
...
>> + case HVMOP_IO_RANGE_VMWARE_PORT:
>> + type_name = "VMware port";
>> + limit = 1;
>> + break;
>> + case HVMOP_IO_RANGE_TIMEOFFSET:
>> + type_name = "timeoffset";
>> + limit = 1;
>> + break;
>> + case HVMOP_IO_RANGE_INVALIDATE:
>> + type_name = "invalidate";
>> + limit = 1;
>> + break;
>> + default:
>> + break;
>
> Thanks for adding the extra types for timeoffset and
> invalidate, although I don't think the invalidate type is
> needed. This ioreq is actually broadcast, so there is no
> selection required.
>
Ok, will drop the invalidate one.
-Don Slutz
> Paul
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel
