Re: [Xen-devel] [PATCH v7 1/6] x86/domain: remove the 'oos_off' flag

[Jan Beulich <jbeulich@xxxxxxxx>]

On 02.09.2019 15:59, Paul Durrant wrote:
>> -----Original Message-----
>> From: Jan Beulich <jbeulich@xxxxxxxx>
>> Sent: 02 September 2019 14:46
>> To: Paul Durrant <Paul.Durrant@xxxxxxxxxx>
>> Cc: Andrew Cooper <Andrew.Cooper3@xxxxxxxxxx>; George Dunlap 
>> <George.Dunlap@xxxxxxxxxx>; Roger Pau
>> Monne <roger.pau@xxxxxxxxxx>; xen-devel@xxxxxxxxxxxxxxxxxxxx; Tim (Xen.org) 
>> <tim@xxxxxxx>; WeiLiu
>> <wl@xxxxxxx>
>> Subject: Re: [PATCH v7 1/6] x86/domain: remove the 'oos_off' flag
>>
>> On 02.09.2019 15:06, Paul Durrant wrote:
>>>> From: Jan Beulich <jbeulich@xxxxxxxx>
>>>> Sent: 02 September 2019 13:34
>>>>
>>>> On 30.08.2019 10:29, Paul Durrant wrote:
>>>>> --- a/xen/common/domain.c
>>>>> +++ b/xen/common/domain.c
>>>>> @@ -313,11 +313,19 @@ static int sanitise_domain_config(struct 
>>>>> xen_domctl_createdomain *config)
>>>>>          return -EINVAL;
>>>>>      }
>>>>>
>>>>> -    if ( !(config->flags & XEN_DOMCTL_CDF_hvm_guest) &&
>>>>> -         (config->flags & XEN_DOMCTL_CDF_hap) )
>>>>> +    if ( !(config->flags & XEN_DOMCTL_CDF_hvm_guest) )
>>>>>      {
>>>>> -        dprintk(XENLOG_INFO, "HAP requested for non-HVM guest\n");
>>>>> -        return -EINVAL;
>>>>> +        if ( config->flags & XEN_DOMCTL_CDF_hap )
>>>>> +        {
>>>>> +            dprintk(XENLOG_INFO, "HAP requested for non-HVM guest\n");
>>>>> +            return -EINVAL;
>>>>> +        }
>>>>> +
>>>>> +        /*
>>>>> +         * It is only meaningful for XEN_DOMCTL_CDF_oos_off to be clear
>>>>> +         * for HVM guests.
>>>>> +         */
>>>>> +        config->flags |= XEN_DOMCTL_CDF_oos_off;
>>>>
>>>> ... I wonder whether this last part wouldn't better belong into
>>>> x86's arch_sanitise_domain_config(). Arm, to the contrary, should
>>>> force/require the bit to be uniformly off.
>>>>
>>>
>>> I'm sure I had a reason for doing it like this but it's sufficiently long
>>> ago now that I've forgotten what it was*. Would it be ok to take the code
>>> as-is and I'll investigate whether this can be tidied up?
>>
>> Well, with this pending question I'm less inclined to stop waiting for
>> the outstanding acks.
>>
>>> [ * I suspect it was concern over breaking existing tool-stacks by
>>> requiring them to set a flag that previously they did not need to, but
>>> I'm not sure that was the only reason ]
>>
>> Seems rather unlikely to me, as there wouldn't be any difference (from
>> tool stack perspective) if the adjustment was done by per-arch code.
> 
> Ok, if you feel strongly about it I'll move the hunk.

Well, wait - not the hunk. The HAP part should remain in common code
imo. The OOS part wants doing differently in x86 and Arm code.

Jan

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel