On Wed, Jul 31, 2024, at 10:09, David Laight wrote:
> From: Linus Torvalds
>> Sent: 31 July 2024 00:04
>>
>> On Tue, 30 Jul 2024 at 15:44, Linus Torvalds
>> <torvalds@xxxxxxxxxxxxxxxxxxx> wrote:
>> >
>> > Does this work for you?
>>
>> It seems to at least build cleanly here, but I'm not claiming it's all
>> that great.
>>
>> The nested __cmp() is still rather less than optimal from an expansion
>> standpoint, but at least it expands only those unique temporaries.
>
> That is the main gain, IIRC Arnd did suggest splitting it but that is
> a relatively small gain.
Right, I suggested a split version earlier, but that was without
the __careful_op3, something like
#define min3(x, y, z) ({ __auto_type __xy = min(x, y); min(__xy, z); })
I think Linus' approach with __careful_op3() is better here because
it handles more corner cases with constant arguments, otherwise
the two approaches are equivalent in how they expand the arguments.
Doing both __careful_op3() and another temporary won't add any
advantages that I can see either.
Arnd
