Re: [PATCH 3/4] xen/events: initialize local per-cpu mask for all possible events

[David Vrabel <david.vrabel@citrix.com>]

On 15/08/13 14:06, Jan Beulich wrote:
>>>> On 15.08.13 at 14:49, David Vrabel <david.vrabel@citrix.com> wrote:
>> On 15/08/13 13:44, Jan Beulich wrote:
>>>>>> On 15.08.13 at 14:21, David Vrabel <david.vrabel@citrix.com> wrote:
>>>> --- a/drivers/xen/events.c
>>>> +++ b/drivers/xen/events.c
>>>> @@ -348,7 +348,7 @@ static void init_evtchn_cpu_bindings(void)
>>>>  
>>>>  	for_each_possible_cpu(i)
>>>>  		memset(per_cpu(cpu_evtchn_mask, i),
>>>> -		       (i == 0) ? ~0 : 0, sizeof(*per_cpu(cpu_evtchn_mask, i)));
>>>> +		       (i == 0) ? ~0 : 0, NR_EVENT_CHANNELS/8);
>>>
>>> I think simply dropping the bogus * would have been the better fix.
>>
>> I disagree.  Using NR_EVENT_CHANNELS/8 makes it clearer that this is
>> initializing all bits.
> 
> My preference of using sizeof() wherever possible is mainly attributed
> to the fact that if someone changes the array size, (s)he'd have to go
> through and identify and change all the corresponding loop or
> whatever boundaries.

Normally I would agree, but in this case a) but this array won't change
its size and b) someone was already confused by the original code and
copied it across incorrectly.

David