[Bug target/125320] [16/17 Regression] riscv: broken assembly with -march=rv64gc_xtheadmemidx -fstack-protector-strong -O2

[xry111 at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=125320

Xi Ruoyao <xry111 at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|riscv: broken assembly with |[16/17 Regression] riscv:
                   |-march=rv64gc_xtheadmemidx  |broken assembly with
                   |-fstack-protector-strong    |-march=rv64gc_xtheadmemidx
                   |-O2                         |-fstack-protector-strong
                   |                            |-O2
   Target Milestone|---                         |16.2

--- Comment #5 from Xi Ruoyao <xry111 at gcc dot gnu.org> ---
(In reply to Xi Ruoyao from comment #4)

> Indeed.  And is it really intended for "m" to accept all the memory
> load/store insns in the extensions?

It seems so (via riscv_legitimate_address_p => riscv_classify_address =>
th_classify_address).  But I've no idea why it didn't manifest in GCC 14 and 15
then...