This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [Patch, Fortran] PR56907 - do not 'pack' arrays passed to C_LOC

>>> Well, it does: As it doesn't know whether the array is contiguous or not
>>> - it packs the array. That's what one usually wants. However, for C_LOC one
>>> knows that the array is contiguous - the user promises this to the compiler
>>> - and, hence, no packing is needed.
>> Ok, I see. Then the patch is certainly ok. (The fact that
>> conv_isocbinding_function is used also for C_ASSOCIATED and C_FUNLOC
>> is no problem either, I guess).
> Well, the code in question is under:
>   if (expr->value.function.isym->id == GFC_ISYM_C_LOC)

Ah, sure, I missed that.

Btw, wouldn't it make more sense to split up
'conv_isocbinding_function' into three separate functions, since there
isn't really any common code and one could directly call them in:

    case GFC_ISYM_C_LOC:
      conv_isocbinding_function (se, expr);

Anyway, the patch is ok as is (or with this additional cleanup, if you prefer).

>> However, if calling C_LOC on a non-contiguous array is invalid,
>> shouldn't one add a check for cases like
>>    integer, dimension(1:5,1:5), target :: zzz
>>    type(c_ptr) :: ptr
>>    ptr = c_loc (zzz(4:,4:))
>> where the compiler can easily tell that the argument is not contiguous ...
>> ?
> Definitely. I think there is also a PR about adding a
> gfc_is_simply_noncontiguous() or something like that. It has several uses:
> C_LOC, pointer-assignment to a contiguous pointer, removing some "if"s
> related to packing (as one knows that internal_pack will pack). And for
> compile-time simplification of the (unimplemented) Fortran 2008 function
> "is_contiguous".



Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]