[Bug target/128310] [14/15/16/17 Regression] wrong code at -O2 with _mm_crc32_u16 and _mm_crc32_u32 on the same constant

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

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=128310

Drea Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |14.5
   Last reconfirmed|                            |2026-10-29
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |NEW

--- Comment #1 from Drea Pinski <pinskia at gcc dot gnu.org> ---
(insn 7 6 8 (set (reg:SI 105)
        (unspec:SI [
                (reg/v:SI 104 [ cD.44427 ])
                (reg:HI 106)
            ] UNSPEC_CRC32))
"/cefs/ff/ffdc7d3937d618b8789d4ffb_gcc-trunk-20261029/lib/gcc/x86_64-linux-gnu/17.0.0/include/smmintrin.h":830:10
-1
     (nil))


vs:
(insn 10 9 11 (set (reg:SI 107)
        (unspec:SI [
                (reg/v:SI 104 [ cD.44427 ])
                (reg:SI 108)
            ] UNSPEC_CRC32))
"/cefs/ff/ffdc7d3937d618b8789d4ffb_gcc-trunk-20261029/lib/gcc/x86_64-linux-gnu/17.0.0/include/smmintrin.h":836:10
-1
     (nil))

But 106/108 are constants which don't carry a mode.

So this gets CSE'd together.