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/84614] New: [8 Regression] wrong code with u16->u128 extension at aarch64 -fno-split-wide-types -g3 --param=max-combine-insns=3


https://gcc.gnu.org/bugzilla/show_bug.cgi?id=84614

            Bug ID: 84614
           Summary: [8 Regression] wrong code with u16->u128 extension at
                    aarch64 -fno-split-wide-types -g3
                    --param=max-combine-insns=3
           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: ---
              Host: x86_64-pc-linux-gnu
            Target: aarch64-unknown-linux-gnu
             Build: x86_64-pc-linux-gnu

Created attachment 43528
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=43528&action=edit
reduced testcase

Output:
$ aarch64-unknown-linux-gnu-gcc -Og -fno-split-wide-types
-fno-tree-coalesce-vars -g --param=max-combine-insns=3 testcase.c -static
$ ./a.out 
qemu: uncaught target signal 6 (Aborted) - core dumped
Aborted

The only extra 3-insns combination is:
...
Trying 2, 16 -> 27:
    2: r102:SI=zero_extend(x0:HI)
      REG_DEAD x0:HI
   16: r105:SI=~r102:SI
      REG_DEAD r102:SI
   27: r109:TI#0=zero_extend(r105:SI#0)
      REG_DEAD r105:SI
...
deferring deletion of insn with uid = 2.
modifying insn i2    16: r105:DI=~x0:DI
      REG_DEAD x0:HI
deferring rescan insn with uid = 16.
modifying insn i3    27: r109:TI#0=zero_extend(r105:DI#0)
      REG_DEAD r105:DI
deferring rescan insn with uid = 27.
...

which looks wrong, because the original does:
r109:TI=zero_ext(~r102:SI)=zero_ext(~x0:HI)
"e = 0 - d"
but the new isns do:
r109:TI=zero_ext(~x0:DI)
"e = 0 - (u32)d"
(extends to DI before ~)

-g needs to be present, and there is a -fcompare-debug failure:
$ aarch64-unknown-linux-gnu-gcc -Og -fno-split-wide-types
-fno-tree-coalesce-vars -g --param=max-combine-insns=3 testcase.c -static
-fcompare-debug
aarch64-unknown-linux-gnu-gcc: error: testcase.c: -fcompare-debug failure
(length)
so it seems that debug insns affect this combination.

$ aarch64-unknown-linux-gnu-gcc -v
Using built-in specs.
COLLECT_GCC=/repo/gcc-trunk/binary-latest-aarch64/bin/aarch64-unknown-linux-gnu-gcc
COLLECT_LTO_WRAPPER=/repo/gcc-trunk/binary-trunk-258031-checking-yes-rtl-df-extra-aarch64/bin/../libexec/gcc/aarch64-unknown-linux-gnu/8.0.1/lto-wrapper
Target: aarch64-unknown-linux-gnu
Configured with: /repo/gcc-trunk//configure --enable-languages=c,c++
--enable-valgrind-annotations --disable-nls --enable-checking=yes,rtl,df,extra
--with-cloog --with-ppl --with-isl
--with-sysroot=/usr/aarch64-unknown-linux-gnu --build=x86_64-pc-linux-gnu
--host=x86_64-pc-linux-gnu --target=aarch64-unknown-linux-gnu
--with-ld=/usr/bin/aarch64-unknown-linux-gnu-ld
--with-as=/usr/bin/aarch64-unknown-linux-gnu-as --disable-libstdcxx-pch
--prefix=/repo/gcc-trunk//binary-trunk-258031-checking-yes-rtl-df-extra-aarch64
Thread model: posix
gcc version 8.0.1 20180227 (experimental) (GCC) 

Tested revisions:
r258031 - FAIL
r257139 - FAIL
7-branch r257043 - OK

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