Re: [PATCH] x86: do away with HAVE_AS_NEGATIVE_TRUE

[Jan Beulich <jbeulich@xxxxxxxx>]

On 19.05.2023 08:15, Jan Beulich wrote:
> On 17.05.2023 19:05, Andrew Cooper wrote:
>> On 17/05/2023 3:22 pm, Jan Beulich wrote:
>>> There's no real need for the associated probing - we can easily convert
>>> to a uniform value without knowing the specific behavior (note also that
>>> the respective comments weren't fully correct and have gone stale). All
>>> we (need to) depend upon is unary ! producing 0 or 1 (and never -1).
>>>
>>> For all present purposes yielding a value with all bits set is more
>>> useful.
>>>
>>> No difference in generated code.
>>>
>>> Signed-off-by: Jan Beulich <jbeulich@xxxxxxxx>
>>> ---
>>> Unlike in C, there's also binary ! in assembly expressions, and even
>>> binary !!. But those don't get in the way here.
>>
>> I had been wanting to do this for a while, but IMO a clearer expression
>> is to take ((x) & 1) to discard the sign.
>>
>> It doesn't change any of the logic to use +1 (I don't think), and it's
>> definitely the more common way for the programmer to think.
> 
> Well, I can certainly switch. It simply seemed to me that with our many
> uses of !! elsewhere, using this here as well would only be consistent.
> (I did in fact consider the ... & 1 alternative.)

Before even starting with this - you do realize that the C macro
(AS_TRUE) expands to just a prefix for the expression to be dealt
with? That would then become "1 & ", which I have to admit I find
a little odd. The alternative of making this a more ordinary macro
(with a parameter) would likely be more intrusive. Plus I assume
you had a reason to do it the way it is right now (and I might end
up figuring that reason the hard way when trying to change things).

Jan