On Tue, Sep 6, 2016 at 11:11 AM, Florian Weimer <fwei...@redhat.com> wrote:
> On 09/05/2016 07:06 PM, Joseph Myers wrote:
>
>> Such an increase is of course an ABI change, but a reasonably safe
>> one, in that max_align_t doesn't tend to appear in library interfaces
>> (rather, it's something to use when writing allocators and similar
>> code; most matches found on codesearch.debian.net look like copies of
>> the gnulib stddef.h module rather than anything actually using
>> max_align_t at all).
>
>
> I've thought some more about this. I'll try to rephrase my reservations in
> a less diplomatic way, hopefully making my thoughts clearer.
>
> I think this change is only safe because nothing relies on it. max_align_t
> is a committee invention with no actual users. As such, you can change it
> in any way you see fit and it won't make a difference in any way.
>
> Why aren't there any users? The standard isn't very clear what the value of
> _Alignof (max_align_t) actually means. Does the phrase “all contexts”
> include pointers returned malloc? Even if the requested size is smaller
> than the fundamental alignment? Did those who wrote the standard expect
> there to be *any* relationship between malloc and max_align_t?
>
> The existing situation is that most mallocs to do not provide _Alignof
> (max_align_t) alignment unconditionally. But they can provide arbitrarily
> large alignment with aligned_alloc/memalign-style interfaces.
>
> For object alignment declarations, I think GCC can satisfy most requests,
> perhaps subject to binutils/dynamic linker limitations for global variables
> on some targets.
>
> The question then is, how can we make max_align_t more useful? If it is
> supposed to reflect a malloc property, it cannot be a compile-time constant
> because we don't know at compile time which malloc is going to be used
> (malloc has to remain interposable). If there is no relationship to malloc,
> GCC essentially supports unbounded alignment on many targets. How is it
> possible to have a meaningful definition of max_align_t under such
> circumstances?
I thought max_align_t was to replace uses like
union {
long long x;
double y;
...
char buf[N];
};
to get statically allocated "max" aligned memory. That is, you now know
_which_ type you need to put into the union.
Richard.
> Florian
