Re: [Xen-devel] [PATCH 21/27 v11] xen/arm: vpl011: Add support for multiple consoles in xenconsole

[Boris Ostrovsky <boris.ostrovsky@xxxxxxxxxx>]

On 10/04/2017 10:29 AM, Wei Liu wrote:
> On Wed, Oct 04, 2017 at 10:28:26AM -0400, Boris Ostrovsky wrote:
>> On 10/04/2017 10:06 AM, Wei Liu wrote:
>>> On Wed, Oct 04, 2017 at 10:03:05AM -0400, Boris Ostrovsky wrote:
>>>>>  
>>>>> @@ -1223,13 +1316,13 @@ void handle_io(void)
>>>>>           /* Re-calculate any event counter allowances & unblock
>>>>>              domains with new allowance */
>>>>>           for (d = dom_head; d; d = d->next) {
>>>>> -                 struct console *con = &d->console;
>>>>>  
>>>>> -                 console_evtchn_unmask(con, (void *)now);
>>>>> +                 console_iter_void_arg2(d, console_evtchn_unmask, (void 
>>>>> *)now);
>>>> This (together with patch 15's "long long now = (long long)data" cast)
>>>> generates fatal warning when building for 32-bit.
>>>>
>>>> The warning can probably be eliminated by casting to uintptr_t (or some
>>>> such) but my question is whether 'now' should be a 'long long' --- isn't
>>>> time_t a 4-byte type on 32-bit? And if it's not then casting will lose
>>>> the upper word.
>>>>
>>> It would be better to just pass the pointer instead.
>> Or we can do that.
>>
>> I just noticed Bhupinder's message so I assume he is going to take care
>> of this.
> I sent a patch just now.
>
>    xenconsole: fix 32bit compilation
>
> Feel free to test it.

Builds fine but I haven't had a chance to test it.

-boris

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel