This is the mail archive of the gcc-bugs@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[Bug rtl-optimization/60116] [4.8/4.9 Regression] wrong code at -Os on x86_64-linux-gnu in 32-bit mode


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60116

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |ebotcazou at gcc dot gnu.org,
                   |                            |law at gcc dot gnu.org

--- Comment #13 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
So, what happens is that we have:
(insn 39 38 41 5 (set (reg:QI 119)
        (const_int 1 [0x1])) pr60116.c:17 93 {*movqi_internal}
     (expr_list:REG_DEAD (reg:CCZ 17 flags)
        (nil)))

(insn 41 39 83 5 (set (reg:CCZ 17 flags)
        (compare:CCZ (reg:SI 101 [ D.1460 ])
            (const_int 0 [0]))) pr60116.c:18 3 {*cmpsi_ccno_1}
     (nil))

(insn 83 41 42 5 (set (reg:SI 122 [ D.1465 ])
        (zero_extend:SI (reg:QI 119))) 138 {*zero_extendqisi2}
     (expr_list:REG_DEAD (reg:QI 119)
        (nil)))

(the REG_DEAD on flags on the first insn supposedly comes from earlier
combining.
try_combine is then called on i3=insn83 and i2=insn39 and manages to combine
insn 83 into (set (reg:SI 122) (const_int 1 [0x1])).
The problem is when distribute_notes attempts to distribute the REG_DEAD
(reg:CCZ 17 flags) note from elim_i2=insn39 to i3=insn83.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]