On 08.02.2022 13:22, Juergen Gross wrote:
> On 08.02.22 12:55, Andrew Cooper wrote:
>> On 08/02/2022 07:42, Juergen Gross wrote:
>>> --- a/xen/include/public/memory.h
>>> +++ b/xen/include/public/memory.h
>>> @@ -662,6 +662,11 @@ struct xen_mem_acquire_resource {
>>>        * two calls.
>>>        */
>>>       uint32_t nr_frames;
>>> +    /*
>>> +     * Padding field, must be zero on input.
>>> +     * The lowest bit was named XENMEM_rsrc_acq_caller_owned in a previous
>>> +     * version and should not be reused in future.
>>
>> s/should/will/.  This is a statement of how Xen shall behave.
> 
> Okay.
> 
>> I think it's also worth somehow fitting in that it's an output only
>> bit.  It will be important when inevitably we end up changing this back
>> to being a flags field when we need to extend the hypercall.
> 
> Okay.
> 
> In the end the bit only needs to be reserved, if pad _is_ zero on input.
> So the correct way to phrase it would be:
> 
> /*
>   * Padding field, must be zero on input.
>   * In a previous version this was an output field with the lowest
>   * bit named XENMEM_rsrc_acq_caller_owned. Future versions of this
>   * interface will not reuse this bit with the field being zero on
>   * input.
>   */
> 
> Is this fine with you?

FWIW it is at least fine with me this way.

Jan


Reply via email to