Re: [v6][PATCH 5/7] hvmloader: introduce hypercall for xc_reserved_device_memory_map

["Chen, Tiejun" <tiejun.chen@intel.com>]

On 2014/9/11 23:45, Jan Beulich wrote:
>>>> On 10.09.14 at 07:49, <tiejun.chen@intel.com> wrote:
>> We will introduce that hypercall xc_reserved_device_memory_map
>> to hvmloader.
>
> Title and text: What does a libxc name do here?

I need to rename this.

>
> Also I don't think this warrants a separate patch - introduce the
> function along with the code needing it.

I will squash the following patch into this.

>
>> Signed-off-by: Tiejun Chen <tiejun.chen@intel.com>
>>
>> diff --git a/tools/firmware/hvmloader/util.c b/tools/firmware/hvmloader/util.c
>> index 80d822f..90dbb6e 100644
>> --- a/tools/firmware/hvmloader/util.c
>> +++ b/tools/firmware/hvmloader/util.c
>> @@ -828,6 +828,28 @@ int hpet_exists(unsigned long hpet_base)
>>       return ((hpet_id >> 16) == 0x8086);
>>   }
>>
>> +int get_reserved_device_memory_map(struct xen_mem_reserved_device_memory entries[],
>> +                                   uint32_t max_entries)
>> +{
>> +    static int map_done = 0;
>> +    struct xen_mem_reserved_device_memory_map memmap = {
>> +        .nr_entries = max_entries
>> +    };
>> +
>> +    if ( map_done )
>> +        return 0;
>> +
>> +    set_xen_guest_handle(memmap.buffer, entries);
>> +
>> +    if ( hypercall_memory_op(XENMEM_reserved_device_memory_map,
>> +                             &memmap) != 0 )
>> +        BUG();
>
> I don't think there's any harm in not BUG()ing here - just return the
> error to the caller (and let it assume there are no entries). In the
> worst case guest startup will subsequently fail (instead of here).

Okay so just return rc.

>
>> +
>> +    map_done = 1;
>
> The map_done variable if completely bogus here, but I think
> someone else already pointed this out.
>

Yes. I will remove this.

Thanks
Tiejun