Re: [PATCH V3 11/23] xen/ioreq: Move x86's io_completion/io_req fields to struct vcpu

[Jan Beulich <jbeulich@xxxxxxxx>]

On 08.12.2020 08:52, Paul Durrant wrote:
>> From: Oleksandr <olekstysh@xxxxxxxxx>
>> Sent: 07 December 2020 21:00
>>
>> On 07.12.20 14:32, Jan Beulich wrote:
>>> On 30.11.2020 11:31, Oleksandr Tyshchenko wrote:
>>>> --- a/xen/include/xen/sched.h
>>>> +++ b/xen/include/xen/sched.h
>>>> @@ -145,6 +145,21 @@ void evtchn_destroy_final(struct domain *d); /* from 
>>>> complete_domain_destroy
>> */
>>>>
>>>>   struct waitqueue_vcpu;
>>>>
>>>> +enum io_completion {
>>>> +    IO_no_completion,
>>>> +    IO_mmio_completion,
>>>> +    IO_pio_completion,
>>>> +#ifdef CONFIG_X86
>>>> +    IO_realmode_completion,
>>>> +#endif
>>>> +};
>>> I'm not entirely happy with io_ / IO_ here - they seem a little
>>> too generic. How about ioreq_ / IOREQ_ respectively?
>>
>> I am OK, would like to hear Paul's opinion on both questions.
>>
> 
> No, I think the 'IO_' prefix is better. They relate to a field in the vcpu_io 
> struct. An alternative might be 'VIO_'...
> 
>>
>>>
>>>> +struct vcpu_io {
>>>> +    /* I/O request in flight to device model. */
>>>> +    enum io_completion   completion;
> 
> ... in which case, you could also name the enum 'vio_completion'.

I'd be okay with these - still better than just "io".

Jan