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 c/47337] [4.6 Regression] memory unsafety bug leading to wrong code generation


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47337

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P1
                 CC|                            |jakub at gcc dot gnu.org
   Target Milestone|---                         |4.6.0
            Summary|memory unsafety bug leading |[4.6 Regression] memory
                   |to wrong code generation    |unsafety bug leading to
                   |                            |wrong code generation

--- Comment #1 from Jakub Jelinek <jakub at gcc dot gnu.org> 2011-01-18 09:04:46 UTC ---
Confirmed, goes away with -fno-partial-inlining, happens only with -m32.
>From brief look at it it looks like a RTL DCE bug, it removes a __divdi3 call,
but takes away with it also a needed unrelated store:
(insn 203 105 204 6 (set (mem:SI (plus:SI (reg/f:SI 7 sp)
                (const_int 8 [0x8])) [0 S4 A32])
        (const_int 10 [0xa])) pr47337.c:14 64 {*movsi_internal}
     (nil))

(insn 204 203 205 6 (set (mem:SI (plus:SI (reg/f:SI 7 sp)
                (const_int 12 [0xc])) [0 S4 A32])
        (const_int 0 [0])) pr47337.c:14 64 {*movsi_internal}
     (nil))

(insn 205 204 206 6 (set (mem:SI (reg/f:SI 7 sp) [0 S4 A32])
        (const_int 0 [0])) pr47337.c:14 64 {*movsi_internal}
     (nil))

(insn 206 205 190 6 (set (mem:SI (plus:SI (reg/f:SI 7 sp)
                (const_int 4 [0x4])) [0 S4 A32])
        (const_int 0 [0])) pr47337.c:14 64 {*movsi_internal}
     (nil))

(insn 190 206 110 6 (set (mem/c:SI (plus:SI (reg/f:SI 7 sp)
                (const_int 28 [0x1c])) [5 S4 A32])
        (reg:SI 2 cx)) pr47337.c:14 64 {*movsi_internal}
     (expr_list:REG_DEAD (reg:SI 2 cx)
        (nil)))

(call_insn/u 110 190 179 6 (set (reg:DI 0 ax)
        (call (mem:QI (symbol_ref:SI ("__divdi3") [flags 0x41]) [0 S1 A8])
            (const_int 16 [0x10]))) pr47337.c:14 927 {*call_value_0}
     (expr_list:REG_UNUSED (reg:DI 0 ax)
        (expr_list:REG_EH_REGION (const_int -2147483648 [0xffffffff80000000])
            (nil)))
    (expr_list:REG_DEP_TRUE (use (mem:DI (reg/f:SI 7 sp) [0 S8 A8]))
        (expr_list:REG_DEP_TRUE (use (mem:DI (plus:SI (reg/f:SI 7 sp)
                        (const_int 8 [0x8])) [0 S8 A8]))
            (nil))))

The sp+0x1c store shouldn't be deleted, but it is.


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