On 21.11.2023 01:04, Stefano Stabellini wrote:
> On Mon, 20 Nov 2023, Jan Beulich wrote:
>> On 20.11.2023 14:13, Federico Serafini wrote:
>>> On 20/11/23 10:02, Jan Beulich wrote:
>>>> On 17.11.2023 09:40, Federico Serafini wrote:
>>>>> --- a/xen/include/xen/sort.h
>>>>> +++ b/xen/include/xen/sort.h
>>>>> @@ -23,8 +23,8 @@
>>>>> extern gnu_inline
>>>>> #endif
>>>>> void sort(void *base, size_t num, size_t size,
>>>>> - int (*cmp)(const void *, const void *),
>>>>> - void (*swap)(void *, void *, size_t))
>>>>> + int (*cmp)(const void *key, const void *elem),
>>>>
>>>> Why "key" and "elem" here, but ...
>>>>
>>>>> + void (*swap)(void *a, void *b, size_t size))
>>>>
>>>> ... "a" and "b" here? The first example of users of sort() that I'm
>>>> looking at right now (x86/extable.c) is consistent in its naming.
>>>>
>>>
>>> On the Arm side there are {cmp,swap}_memory_node() and
>>> {cmp,swap}_mmio_handler(): "key"/"elem" are used for the comparison
>>> and "_a"/"_b" for the swap.
>>
>> So - re-raising a question Stefano did raise - is Misra concerned about
>> such discrepancies? If yes, _all_ instances need harmonizing. If not, I
>> see no reason to go with misleading names here.
>
> Federico confirmed that the answer is "no".
>
> I think we can use "key" and "elem" in this patch as they are more
> informative than "a" and "b"
Except that "key" and "elem" are (imo) inapplicable to sort() callbacks
(and inconsistent with the naming in the 2nd callback here); they _may_
be applicable in bsearch() ones. Note also how in the C99 spec these
parameters of callback functions don't have names either.
Jan
