https://gcc.gnu.org/bugzilla/show_bug.cgi?id=121495
Andrew Pinski <pinskia at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |ASSIGNED
Last reconfirmed| |2026-03-10
Ever confirmed|0 |1
Assignee|unassigned at gcc dot gnu.org |pinskia at gcc dot
gnu.org
--- Comment #15 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
So this might be an early ra issue after all:
```
(insn 920 119 120 15 (set (reg:DI 63 v31)
(reg/f:DI 309))
"/usr/lib/gcc/aarch64-unknown-linux-gnu/14/include/g++-v14/bits/unique_ptr.h":1092:63
discrim 1 70 {*movdi_aarch64}
(nil))
(insn 120 920 122 15 (set (mem:DI (post_inc:DI (reg:DI 63 v31)) [16
MEM[(sizetype *)_175]+0 S8 A64])
(reg:DI 310))
"/usr/lib/gcc/aarch64-unknown-linux-gnu/14/include/g++-v14/bits/unique_ptr.h":1092:63
discrim 1 70 {*movdi_aarch64}
(expr_list:REG_DEAD (reg:DI 310)
(expr_list:REG_INC (reg/f:DI 160 [ _176 ])
(nil))))
(insn 122 120 123 15 (set (reg:DI 188 [ ivtmp.25316 ])
(reg:DI 63 v31)) 70 {*movdi_aarch64}
(nil))
```
Then that post_inc gets miscompiled by IRA/LRA.
Let me see if I can reject that in early-ra.
