[Bug target/89482] arm aarch32 inline assembly w constraints generate s registers instead of d

[ktkachov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=89482

ktkachov at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|WAITING                     |RESOLVED
         Resolution|---                         |INVALID

--- Comment #4 from ktkachov at gcc dot gnu.org ---
Right. So when dealing with double-precision values you should use the %P
output modifier like so:
    double my_double = 1.5;
    __asm__ (
        "vmov.f64 d0, 1.0;"
        "vadd.f64 %P[my_double], %P[my_double], d0;"
        : [my_double] "+w" (my_double)
        :
        : "d0"
    );

This will print the right D-form of the VFP register allocated.
The documentation for this could, admittedly be improved
(https://gcc.gnu.org/onlinedocs/gcc/Extended-Asm.html). I believe someone is
working on that currently.