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/80693] New: [6/7/8 Regression] wrong code with -O -fno-tree-coalesce-vars


https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80693

            Bug ID: 80693
           Summary: [6/7/8 Regression] wrong code with -O
                    -fno-tree-coalesce-vars
           Product: gcc
           Version: 8.0
            Status: UNCONFIRMED
          Keywords: wrong-code
          Severity: normal
          Priority: P3
         Component: rtl-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: zsojka at seznam dot cz
  Target Milestone: ---

Created attachment 41335
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=41335&action=edit
reduced testcase

Output:
$ x86_64-pc-linux-gnu-gcc -O -fno-tree-coalesce-vars testcase.c
$ ./a.out
Aborted

The .combine dump shows:
...
Trying 18, 19 -> 20:
Failed to match this instruction:
(set (reg:HI 116)
    (and:HI (not:HI (reg/v:HI 112 [ u16_1D.1800 ]))
        (const_int 1 [0x1])))
Successfully matched this instruction:
(set (reg:HI 117)
    (not:HI (reg/v:HI 112 [ u16_1D.1800 ])))
Successfully matched this instruction:
(set (reg:HI 116)
    (and:HI (reg:HI 117)
        (const_int 1 [0x1])))
allowing combination of insns 18, 19 and 20
original costs 0 + 4 + 4 = 0
replacement costs 4 + 4 = 8
deferring deletion of insn with uid = 18.
modifying insn i2    19: r117:HI=!r112:HI
deferring rescan insn with uid = 19.
modifying insn i3    20: {r116:HI=r117:HI&0x1;clobber flags:CC;}
      REG_UNUSED flags:CC
      REG_DEAD r117:HI
deferring rescan insn with uid = 20.
...
Trying 19 -> 20:
Successfully matched this instruction:
(parallel [
        (set (reg:HI 116)
            (xor:HI (reg/v:HI 112 [ u16_1D.1800 ])
                (const_int 1 [0x1])))
        (clobber (reg:CC 17 flags))
    ])
allowing combination of insns 19 and 20
original costs 4 + 4 = 8
replacement cost 4
deferring deletion of insn with uid = 19.
modifying insn i3    20: {r116:HI=r112:HI^0x1;clobber flags:CC;}
      REG_UNUSED flags:CC
deferring rescan insn with uid = 20.
...


Which looks wrong, since it changes "r116:HI=!r112:HI&1" (can be 0 or 1) to
"r116:HI=r112:HI^0x1" (can be anything).

Also, the costs dump in the first replacement looks wrong:
original costs 0 + 4 + 4 = 0

$ x86_64-pc-linux-gnu-gcc
COLLECT_LTO_WRAPPER=/repo/gcc-trunk/binary-trunk-247635-checking-yes-rtl-df-extra-nographite-amd64/bin/../libexec/gcc/x86_64-pc-linux-gnu/8.0.0/lto-wrapper
Target: x86_64-pc-linux-gnu
Configured with: /repo/gcc-trunk//configure --enable-languages=c,c++
--enable-valgrind-annotations --disable-nls --enable-checking=yes,rtl,df,extra
--without-cloog --without-ppl --without-isl --build=x86_64-pc-linux-gnu
--host=x86_64-pc-linux-gnu --target=x86_64-pc-linux-gnu
--with-ld=/usr/bin/x86_64-pc-linux-gnu-ld
--with-as=/usr/bin/x86_64-pc-linux-gnu-as --disable-libstdcxx-pch
--prefix=/repo/gcc-trunk//binary-trunk-247635-checking-yes-rtl-df-extra-nographite-amd64
Thread model: posix
gcc version 8.0.0 20170505 (experimental) (GCC) 


Tested revisions:
trunk r247635 - FAIL
7-branch r247545 - FAIL
6-branch r247359 - FAIL
5-branch r247456 - OK

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