https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80266
Wilco <wilco at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |NEW
Last reconfirmed| |2017-07-05
CC| |wilco at gcc dot gnu.org
Ever confirmed|0 |1
--- Comment #1 from Wilco <wilco at gcc dot gnu.org> ---
Confirmed. This is the same issue as:
https://gcc.gnu.org/ml/gcc-patches/2017-06/msg02422.html
Ada somehow generates SImode addresses for local variables which isn't correct.
This happens early on in Expand, eg for
testsuite/gnat.dg/abstract_with_anonymous_result.adb:
(insn 9 8 10 2 (set (reg:DI 403)
(plus:DI (reg/f:DI 68 virtual-stack-vars)
(const_int -368 [0xfffffffffffffe90])))
(nil))
(insn 10 9 11 2 (set (reg:SI 404)
(plus:SI (subreg:SI (reg:DI 403) 0)
(const_int 56 [0x38])))
(nil))
(insn 11 10 12 2 (set (reg:DI 405)
(plus:DI (reg/f:DI 68 virtual-stack-vars)
(const_int -368 [0xfffffffffffffe90])))
(nil))
(insn 12 11 13 2 (set (reg:SI 406)
(subreg:SI (reg:DI 405) 0))
(nil))
(insn 13 12 14 2 (set (mem/c:SI (reg:SI 404) [0 S4 A8])
(reg:SI 406))
(nil))
