https://gcc.gnu.org/bugzilla/show_bug.cgi?id=123521
--- Comment #3 from GCC Commits <cvs-commit at gcc dot gnu.org> --- The master branch has been updated by LuluCheng <[email protected]>: https://gcc.gnu.org/g:1a4eb0479ce0827b55a6190ddf13ade330d619e6 commit r16-6815-g1a4eb0479ce0827b55a6190ddf13ade330d619e6 Author: Lulu Cheng <[email protected]> Date: Mon Jan 12 17:23:41 2026 +0800 LoongArch: Fix bug123521. In the vector initialization process, optimization can be performed if it can be determined that all elements are the same, or if the upper or lower halves are identical. However, during this optimization, when the identical element is an immediate value larger than 10 bits, an internal compiler error (ICE) occurs. The reason is that in such cases, the function `simplify_gen_subreg (imode, reg_tmp, GET_MODE (reg_tmp), 0)` is called, where `imode` is `E_DImode`. The mode of `reg_tmp` in `GET_MODE (reg_tmp)` is taken from the immediate value's mode, which is `E_VOIDmode`. This results in a move from `E_VOIDmode` to `E_DImode`, an operation not supported by LoongArch. PR target/123521 gcc/ChangeLog: * config/loongarch/loongarch.cc (loongarch_expand_vector_init_same): Fixed a bug in the vector initialization section.. gcc/testsuite/ChangeLog: * gcc.target/loongarch/vector/lasx/pr123521.c: New test. Suggested-by: Xi Ruoyao <[email protected]>
