[Bug c/127740] `__atomic_store_n (&v, f + 1, ...)` with a `_Float16` operand into an integer object: ICE in `c_fully_fold_internal` (c-fold.cc:677, `EXCESS_PRECISION_EXPR`)

[redi at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=127740

Jonathan Wakely <redi at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2026-10-06
                 CC|                            |liuhongt at gcc dot gnu.org
     Ever confirmed|0                           |1
           Keywords|                            |ice-on-valid-code
            Summary|[GCC C] `__atomic_store_n   |`__atomic_store_n (&v, f +
                   |(&v, f + 1, ...)` with a    |1, ...)` with a `_Float16`
                   |`_Float16` operand into an  |operand into an integer
                   |integer object: ICE in      |object: ICE in
                   |`c_fully_fold_internal`     |`c_fully_fold_internal`
                   |(c-fold.cc:677,             |(c-fold.cc:677,
                   |`EXCESS_PRECISION_EXPR`)    |`EXCESS_PRECISION_EXPR`)

--- Comment #1 from Jonathan Wakely <redi at gcc dot gnu.org> ---
GCC 11 rejects _Float16 and GCC 12 ICEs, starting with r12-3309 which added
_Float16 for the target:

"Enable _Float16 type for TARGET_SSE2 and above."


N.B. Please don't add made-up "tags" to the summary like "[GCC C]", that's what
the Product=gcc and Component=c fields are for.