[Bug other/52937] New: [4.8 Regression] ICE in several cpu2006 benchmarks with -fprofile-use (assert in get_loop_body)

[Bug other/52937] New: [4.8 Regression] ICE in several cpu2006 benchmarks with -fprofile-use (assert in get_loop_body)

[wschmidt at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52937

             Bug #: 52937
           Summary: [4.8 Regression] ICE in several cpu2006 benchmarks
                    with -fprofile-use (assert in get_loop_body)
    Classification: Unclassified
           Product: gcc
           Version: 4.8.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: other
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: wschmidt@gcc.gnu.org
                CC: bergner@vnet.ibm.com, rguenther@suse.de
              Host: powerpc64-linux
            Target: powerpc64-linux
             Build: powerpc64-linux


Created attachment 27135
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=27135
SPEC cpu2006 configuration file used for failing build

403.gcc, 445.gobmk, and 454.calculix all fail to build on trunk when applying
profile data (-fprofile-use).  I apologize for not providing a reduced test
case, but I'm not sure how to do that when profile data is involved.  I'm
attaching my cpu2006 .cfg file; hopefully that will help to reproduce.

Richi, copying you since this appears similar to
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52800.

403.gcc
=======
...

/home/wschmidt/gcc/install/gcc-mainline-test/bin/gcc -c -o conflict.o
-DSPEC_CPU -DNDEBUG -I.   -fprofile-generate    -m32 -O3 -mcpu=power7
-fpeel-loops -funroll-loops -ffast-math -fvect-cost-model -mpopcntd
-mrecip=rsqrt -fno-associative-math -fsched-pressure                 conflict.c

...

/home/wschmidt/gcc/install/gcc-mainline-test/bin/gcc  -m32 -O3 -mcpu=power7
-fpeel-loops -funroll-loops -ffast-math -fvect-cost-model -mpopcntd
-mrecip=rsqr
t -fno-associative-math -fsched-pressure        -static
-L/home/wschmidt/gcc/install/gcc-mainline-test/lib -Wl,-q 
-Wl,-rpath=/home/wschmidt/gcc/install/gcc-mainline-test/lib
-Wl,-Map=link.map,--cref -fprofile-generate  alloca.o asprintf.o vasprintf.o
c-parse.o c-lang.o attribs.o c-errors.o c-lex.o c-pragma.o c-decl.o c-typeck.o
c-convert.o c-aux-info.o c-common.o c-format.o c-semantics.o c-objc-common.o
main.o cpplib.o cpplex.o cppmacro.o cppexp.o cppfiles.o cpphash.o cpperror.o
cppinit.o cppdefault.o line-map.o mkdeps.o prefix.o version.o mbchar.o alias.o
bb-reorder.o bitmap.o builtins.o caller-save.o calls.o cfg.o cfganal.o
cfgbuild.o cfgcleanup.o cfglayout.o cfgloop.o cfgrtl.o combine.o conflict.o
convert.o cse.o cselib.o dbxout.o debug.o dependence.o df.o diagnostic.o
doloop.o dominance.o dwarf2asm.o dwarf2out.o dwarfout.o emit-rtl.o except.o
explow.o expmed.o expr.o final.o flow.o fold-const.o function.o gcse.o genrtl.o
ggc-common.o global.o graph.o haifa-sched.o hash.o hashtable.o hooks.o ifcvt.o
insn-attrtab.o insn-emit.o insn-extract.o insn-opinit.o insn-output.o
insn-peep.o insn-recog.o integrate.o intl.o jump.o langhooks.o lcm.o lists.o
local-alloc.o loop.o obstack.o optabs.o params.o predict.o print-rtl.o
print-tree.o profile.o real.o recog.o reg-stack.o regclass.o regmove.o
regrename.o reload.o reload1.o reorg.o resource.o rtl.o rtlanal.o rtl-error.o
sbitmap.o sched-deps.o sched-ebb.o sched-rgn.o sched-vis.o sdbout.o sibcall.o
simplify-rtx.o ssa.o ssa-ccp.o ssa-dce.o stmt.o stor-layout.o stringpool.o
timevar.o toplev.o tree.o tree-dump.o tree-inline.o unroll.o varasm.o varray.o
vmsdbgout.o xcoffout.o ggc-page.o i386.o xmalloc.o xexit.o hashtab.o
safe-ctype.o splay-tree.o xstrdup.o md5.o fibheap.o xstrerror.o concat.o
partition.o hex.o lbasename.o getpwd.o ucbqsort.o             -lm        -o gcc

...

Training 403.gcc with the train workload
Commands to run:
Specinvoke: /home/wschmidt/spec/cpu2006/bin/specinvoke -d
/home/wschmidt/spec/cpu2006/benchspec/CPU2006/403.gcc/build/build_peak_base_32.0000
-e speccmds.err -o speccmds.stdout -f speccmds.cmd -C

...

/home/wschmidt/gcc/install/gcc-mainline-test/bin/gcc -c -o conflict.o
-DSPEC_CPU -DNDEBUG -I.   -fprofile-use -fdump-tree-tracer-details    -m32 -O3
-mcpu=power7 -fpeel-loops -funroll-loops -ffast-math -fvect-cost-model
-mpopcntd -mrecip=rsqrt -fno-associative-math -fsched-pressure                
conflict.c

conflict.c: In function 'conflict_graph_compute':
conflict.c:446:1: internal compiler error: in get_loop_body, at cfgloop.c:831

445.gobmk
=========
Similar build sequence.  Error occurs on:

/home/wschmidt/gcc/install/gcc-mainline-test/bin/gcc -c -o engine/owl.o
-DSPEC_CPU -DNDEBUG -DHAVE_CONFIG_H -I. -I.. -I../include -I./include  
-fprofile-use -fdump-tree-tracer-details    -m32 -O3 -mcpu=power7 -fpeel-loops
-funroll-loops -ffast-math -fvect-cost-model -mpopcntd -mrecip=rsqrt
-fno-associative-math -fsched-pressure                 engine/owl.c

engine/owl.c: In function 'owl_determine_life.isra.2':
engine/owl.c:2352:1: internal compiler error: in get_loop_body, at
cfgloop.c:831

454.calculix
============
Similar build sequence.  Error occurs on:

/home/wschmidt/gcc/install/gcc-mainline-test/bin/gcc -c -o
SPOOLES/GPart/src/GPart_util.o -DSPEC_CPU -DNDEBUG -ISPOOLES   -fprofile-use
-fdump-tree-tracer-details    -m32 -O3 -mcpu=power7 -fpeel-loops -funroll-loops
-ffast-math -fvect-cost-model -mpopcntd -mrecip=rsqrt -fno-associative-math
-fsched-pressure                    SPOOLES/GPart/src/GPart_util.c

SPOOLES/GPart/src/GPart_util.c: In function 'GPart_setCweights':
SPOOLES/GPart/src/GPart_util.c:17:1: internal compiler error: in get_loop_body,
at cfgloop.c:831

[Bug other/52937] [4.8 Regression] ICE in several cpu2006 benchmarks with -fprofile-use (assert in get_loop_body)

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52937

Richard Guenther <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.8.0

--- Comment #1 from Richard Guenther <rguenth at gcc dot gnu.org> 2012-04-12 09:28:53 UTC ---
Hmpf.  Can you isolate the pass that causes this with

Index: gcc/passes.c
===================================================================
*** gcc/passes.c.orig   2012-03-16 15:13:32.000000000 +0100
--- gcc/passes.c        2012-03-16 15:14:17.000000000 +0100
*************** execute_function_todo (void *data)
*** 1732,1737 ****
--- 1732,1741 ----
      verify_gimple_in_cfg (cfun);
    if (flags & TODO_verify_flow)
      verify_flow_info ();
+   if (current_loops
+       /* Between IRA and reload loops are broken but preserved.  */
+       && (cfun->curr_properties & PROP_loops))
+     verify_loop_structure ();
    if (current_loops && loops_state_satisfies_p (LOOP_CLOSED_SSA))
      verify_loop_closed_ssa (false);
    if (flags & TODO_verify_rtl_sharing)

?

[Bug other/52937] [4.8 Regression] ICE in several cpu2006 benchmarks with -fprofile-use (assert in get_loop_body)

[wschmidt at linux dot vnet.ibm.com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52937

--- Comment #2 from wschmidt at linux dot vnet.ibm.com 2012-04-12 11:39:37 UTC ---
Sure, I'll get that information today.

On Thu, 2012-04-12 at 09:28 +0000, rguenth at gcc dot gnu.org wrote:
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52937
> 
> Richard Guenther <rguenth at gcc dot gnu.org> changed:
> 
>            What    |Removed                     |Added
> ----------------------------------------------------------------------------
>    Target Milestone|---                         |4.8.0
> 
> --- Comment #1 from Richard Guenther <rguenth at gcc dot gnu.org> 2012-04-12 09:28:53 UTC ---
> Hmpf.  Can you isolate the pass that causes this with
> 
> Index: gcc/passes.c
> ===================================================================
> *** gcc/passes.c.orig   2012-03-16 15:13:32.000000000 +0100
> --- gcc/passes.c        2012-03-16 15:14:17.000000000 +0100
> *************** execute_function_todo (void *data)
> *** 1732,1737 ****
> --- 1732,1741 ----
>       verify_gimple_in_cfg (cfun);
>     if (flags & TODO_verify_flow)
>       verify_flow_info ();
> +   if (current_loops
> +       /* Between IRA and reload loops are broken but preserved.  */
> +       && (cfun->curr_properties & PROP_loops))
> +     verify_loop_structure ();
>     if (current_loops && loops_state_satisfies_p (LOOP_CLOSED_SSA))
>       verify_loop_closed_ssa (false);
>     if (flags & TODO_verify_rtl_sharing)
> 
> ?
>

[Bug other/52937] [4.8 Regression] ICE in several cpu2006 benchmarks with -fprofile-use (assert in get_loop_body)

[wschmidt at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52937

William J. Schmidt <wschmidt at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|                            |INVALID

--- Comment #3 from William J. Schmidt <wschmidt at gcc dot gnu.org> 2012-04-12 15:06:23 UTC ---
My apologies.  This is occurring in some instrumentation code I added for
testing, and doesn't occur with the unmodified trunk compiler.  I should have
checked that first, but was misled because it looked like similar problems
we've seen recently from the loop changes.

So, take the day off. ;)