[Bug target/118594] New: insn does not satisfy its constraints with _Float16 in some cases

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

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=118594

            Bug ID: 118594
           Summary: insn does not satisfy its constraints with _Float16 in
                    some cases
           Product: gcc
           Version: 15.0
            Status: UNCONFIRMED
          Keywords: ice-on-valid-code
          Severity: normal
          Priority: P3
         Component: target
          Assignee: unassigned at gcc dot gnu.org
          Reporter: pinskia at gcc dot gnu.org
  Target Milestone: ---
            Target: aarch64

Created attachment 60229
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=60229&action=edit
testcase

Split off from PR 116507.
`aarch64-unknown-linux-gnu-gcc -O2 -ffast-math`

Since this uses _BitInt, I can't tell if it is a regression or not.
```
<source>:36:1: error: insn does not satisfy its constraints:
   36 | }
      | ^
(insn 47 46 48 2 (set (reg:HI 1 x1 [172])
        (mem/v/j/c:HI (plus:DI (reg/f:DI 4 x4 [96])
                (const_int -257 [0xfffffffffffffeff])) [1 +0 S2 A8]))
"<source>":21:17 68 {*movhi_aarch64}
     (nil))
during RTL pass: reload
```