Re: [libstdc++/65033] Give alignment info to libatomic

[Jonathan Wakely <jwakely@redhat.com>]

On 05/04/15 21:07 -0400, Hans-Peter Nilsson wrote:
>On Fri, 3 Apr 2015, Jonathan Wakely wrote:
>
>> On 03/04/15 05:24 -0400, Hans-Peter Nilsson wrote:
>> > On Thu, 2 Apr 2015, Hans-Peter Nilsson wrote:
>> > > Why then use __alignof(_M_i) (the object-alignment)
>> > > instead of _S_alignment (the deduced alas insufficiently
>> > > increased type-alignment)?
>>
>> Isn't the object aligned to _S_alignment?
>
>We did specify that with the alignas.  Is the alignof always
>exactly the same as an alignas, if one is specified?  (And will
>that not change in a future amendment, standard and/or
>implementation?)  Either way, is there a test-case to guard all
>this?

The language guarantees that's what alignas() does, if the argument is
a valid alignment (which it must be if we derive it from some other
type's alignment).

>Those questions wouldn't even be asked if we used _S_alignment
>for the fake-pointer too, just as a matter of defensive
>programming.
>
>> Instead of changing every case in the condition to include sizeof why
>> not just do it afterwards using sizeof(_Tp), in the _S_alignment
>> calculation?
>
>Doh.
>
>> We know sizeof(_Tp) == sizeof(corresponding integer type) because
>> that's the whole point of the conditionals! See attachment.
>>
>> > @@ -216,7 +216,7 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION
>> >       is_lock_free() const noexcept
>> >       {
>> > 	// Produce a fake, minimally aligned pointer.
>> > -	void *__a = reinterpret_cast<void *>(-__alignof(_M_i));
>> > +	void *__a = reinterpret_cast<void *>(-_S_alignment);
>> > 	return __atomic_is_lock_free(sizeof(_M_i), __a);
>> >       }
>>
>> If _M_i is aligned to _S_alignment then what difference does the
>> change above make?
>>
>> It doesn't matter if the value is per-object if we've forced all such
>> objects to have the same alignment, does it?
>>
>> Or is it different if a std::atomic<T> is included in some other
>> struct and the user forces a different alignment on it? I don't think
>> we really need to support that, users shouldn't be doing that.
>
>Why do we even need to ask those questions, when the patch takes
>care of the per-type business without doubt?

Well if we know the object is guaranteed to be correctly aligned we
might not even need a fake, minimally aligned pointer. We could go
back to passing &_M_i or just a null pointer to __atomic_is_lock_free.

The whole point of alignas() is to fix the alignment to a known value.