[Bug rtl-optimization/68194] [6 Regression] wrong code at -O2 and -O3 on x86_64-linux-gnu
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68194 --- Comment #7 from Jeffrey A. Law --- *** Bug 68328 has been marked as a duplicate of this bug. ***
[Bug rtl-optimization/68194] [6 Regression] wrong code at -O2 and -O3 on x86_64-linux-gnu
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68194 --- Comment #8 from Jeffrey A. Law --- *** Bug 68185 has been marked as a duplicate of this bug. ***
[Bug rtl-optimization/68194] [6 Regression] wrong code at -O2 and -O3 on x86_64-linux-gnu
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68194 --- Comment #6 from ktkachov at gcc dot gnu.org --- Proposed patch at: https://gcc.gnu.org/ml/gcc-patches/2015-11/msg01953.html
[Bug rtl-optimization/68194] [6 Regression] wrong code at -O2 and -O3 on x86_64-linux-gnu
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68194
--- Comment #5 from ktkachov at gcc dot gnu.org ---
The different if-conversion behaviour enabled a new cse opportunity
which then produced the RTL that triggered the bad ree behaviour.
The relevant RTL insns before the ree pass look like:
Basic Block A:
(set (reg:HI 0 ax)
(mem:HI (symbol_ref:DI ("f"
...
(set (reg:SI 3 bx)
(if_then_else:SI (eq (reg:CCZ 17 flags)
(const_int 0))
(reg:SI 0 ax)
(reg:SI 3 bx)))
(set (reg:SI 4 si)
(sign_extend:SI (reg:HI 3 bx)))
...
Basic block B (dominated by basic block A):
(set (reg:SI 4 si)
(sign_extend:SI (reg:QI 0 ax))) /* ax contains symbol "f". */
ree changes that into the broken:
Basic block A:
(set (reg:SI 4 si)
(sign_extend:SI (mem:HI (symbol_ref:DI ("f")
(set (reg:SI 3 bx)
(reg:SI 4 si))
...
(set (reg:SI 3 bx)
(if_then_else:SI (eq (reg:CCZ 17 flags)
(const_int 0 [0]))
(reg:SI 0 ax)
(reg:SI 3 bx)))
...
Basic block B:
(set (reg:SI 4 si)
(sign_extend:SI (reg:QI 0 ax))) /* Insn unchanged by ree, but ax now
undefined. */
[Bug rtl-optimization/68194] [6 Regression] wrong code at -O2 and -O3 on x86_64-linux-gnu
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68194 ktkachov at gcc dot gnu.org changed: What|Removed |Added Status|NEW |ASSIGNED Assignee|unassigned at gcc dot gnu.org |ktkachov at gcc dot gnu.org Known to fail||6.0 --- Comment #4 from ktkachov at gcc dot gnu.org --- In the end, I think the "ree" pass is at fault here. I'm testing a patch.
[Bug rtl-optimization/68194] [6 Regression] wrong code at -O2 and -O3 on x86_64-linux-gnu
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68194 --- Comment #3 from ktkachov at gcc dot gnu.org --- Created attachment 36641 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=36641=edit cse2 dump containing the 'bad' RTL attaching the cse2 dump.
[Bug rtl-optimization/68194] [6 Regression] wrong code at -O2 and -O3 on x86_64-linux-gnu
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68194 Richard Biener changed: What|Removed |Added Keywords||wrong-code Target Milestone|--- |6.0 Summary|wrong code at -O2 and -O3 |[6 Regression] wrong code |on x86_64-linux-gnu |at -O2 and -O3 on ||x86_64-linux-gnu
[Bug rtl-optimization/68194] [6 Regression] wrong code at -O2 and -O3 on x86_64-linux-gnu
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68194
--- Comment #2 from ktkachov at gcc dot gnu.org ---
Created attachment 36640
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=36640=edit
cprop dump before the suspected bad transformation
Confirmed the failure.
However I think the cited revision (RTL if-conversion change) exposes another
bug.
I think cse goes wrong and ends up CSEing a memory load from a symbol_ref to
'f':
(insn 67 66 68 13 (set (reg:HI 108 [ f ])
(mem/c:HI (symbol_ref:DI ("f") [flags 0x2] ) [2 f+0 S2 A16])) fail.c:28 88 {*movhi_internal}
(nil))
(insn 68 67 69 13 (set (reg:SI 109 [ f ])
(sign_extend:SI (subreg:QI (reg:HI 108 [ f ]) 0))) fail.c:28 146
{extendqisi2}
(expr_list:REG_DEAD (reg:HI 108 [ f ])
(nil)))
in basic block 13 gets transformed to:
(insn 68 66 69 13 (set (reg:SI 109 [ f ])
(sign_extend:SI (subreg:QI (reg:HI 112 [ f ]) 0))) fail.c:28 146
{extendqisi2}
(expr_list:REG_DEAD (reg:HI 108 [ f ])
(nil)))
because (symbol_ref:DI ("f") gets loaded into reg 112 in another basic block
(bb 8).
This seems bogus to me.
Indeed, adding -fno-rerun-cse-after-loop (the flag controlling cse2) to the
flags makes the testcase pass
[Bug rtl-optimization/68194] [6 Regression] wrong code at -O2 and -O3 on x86_64-linux-gnu
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68194 Marek Polacek changed: What|Removed |Added Status|UNCONFIRMED |NEW Last reconfirmed||2015-11-03 CC||mpolacek at gcc dot gnu.org Ever confirmed|0 |1 --- Comment #1 from Marek Polacek --- Started with r228194.
