Re: [PATCH v4 1/4] x86/HVM: update the start info structure layout

[Boris Ostrovsky <boris.ostrovsky@oracle.com>]

On 02/17/2016 05:05 AM, Roger Pau Monné wrote:
> El 17/2/16 a les 10:58, Jan Beulich ha escrit:
>>>>> On 16.02.16 at 22:26, <boris.ostrovsky@oracle.com> wrote:
>>> On 02/16/2016 12:37 PM, Roger Pau Monne wrote:
>>>> diff --git a/xen/include/public/xen.h b/xen/include/public/xen.h
>>>> index 7b629b1..6ba060f 100644
>>>> --- a/xen/include/public/xen.h
>>>> +++ b/xen/include/public/xen.h
>>>> @@ -787,25 +787,46 @@ typedef struct start_info start_info_t;
>>>>    /*
>>>>     * Start of day structure passed to PVH guests in %ebx.
>>>>     *
>>>> - * NOTE: nothing will be loaded at physical address 0, so
>>>> - * a 0 value in any of the address fields should be treated
>>>> - * as not present.
>>>> + * NOTE: nothing will be loaded at physical address 0, so a 0 value in any
>>>> + * of the address fields should be treated as not present.
>>>> + *
>>>> + *  0 +----------------+
>>>> + *    | magic          | Contains the magic value HVM_START_MAGIC_VALUE
>>>> + *    |                | ("xEn3" with the 0x80 bit of the "E" set).
>>>> + *  4 +----------------+
>>>> + *    | version        | Version of this structure. Current version is 0. New
>>>> + *    |                | versions are guaranteed to be backwards-compatible.
>>> #define XEN_HVM_START_INFO_VERSION   0
>> What would that buy us? Once it gets bumped to 1, consumers
>> would need to check against literal zero anyway.

Consumers would need to check against what their header file's version 
is, not necessarily zero. And they, for example, may decide not to run 
if the version provided by the structure is smaller than what they support.

-boris

> I agree with Jan that this doesn't seem very useful, the headers don't
> have to be in sync with the underlying hypervisor, so it's better to
> just use literals instead of defines.
>
> Roger.