[Bug tree-optimization/48330] New: [4.7 Regression] ICE: in optimize_inline_calls, at tree-inline.c:4201 with -fmudflap -fno-early-inlining

[Bug tree-optimization/48330] New: [4.7 Regression] ICE: in optimize_inline_calls, at tree-inline.c:4201 with -fmudflap -fno-early-inlining

[zsojka at seznam dot cz]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48330

           Summary: [4.7 Regression] ICE: in optimize_inline_calls, at
                    tree-inline.c:4201 with -fmudflap -fno-early-inlining
           Product: gcc
           Version: 4.7.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: tree-optimization
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: zsojka@seznam.cz
                CC: jamborm@gcc.gnu.org
              Host: x86_64-pc-linux-gnu
            Target: x86_64-pc-linux-gnu


It seems any input file is enough to trigger the assertion.

Compiler output:
$ echo '' > testcase.c
$ gcc -O -fmudflap -fno-early-inlining testcase.c 
testcase.c: In function '_GLOBAL__sub_I_00099_0_testcase.c':
testcase.c:1:0: internal compiler error: in optimize_inline_calls, at
tree-inline.c:4201
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.

That assert was added in
http://gcc.gnu.org/viewcvs?view=revision&revision=171602

[Bug tree-optimization/48330] [4.7 Regression] ICE: in optimize_inline_calls, at tree-inline.c:4201 with -fmudflap -fno-early-inlining

[jamborm at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48330

--- Comment #1 from Martin Jambor <jamborm at gcc dot gnu.org> 2011-03-29 12:54:50 UTC ---
It's mudflap adding call graph nodes from outside the pass manager
again:

(gdb) bt
#0  fancy_abort (file=0x89fd0b8 "/home/mjambor/gcc/icln/src/gcc/tree-inline.c",
line=4201, 
    function=0x89fe0a3 "optimize_inline_calls") at
/home/mjambor/gcc/icln/src/gcc/diagnostic.c:893
#1  0x086c4642 in optimize_inline_calls (fn=0xb7649100)
    at /home/mjambor/gcc/icln/src/gcc/tree-inline.c:4201
#2  0x08699622 in cgraph_early_inlining () at
/home/mjambor/gcc/icln/src/gcc/ipa-inline.c:1757
#3  0x083b43f9 in execute_one_pass (pass=0x8b1f300) at
/home/mjambor/gcc/icln/src/gcc/passes.c:1555
#4  0x083b46ad in execute_pass_list (pass=0x8b1f300)
    at /home/mjambor/gcc/icln/src/gcc/passes.c:1610
#5  0x084be2f5 in tree_lowering_passes (fn=0xb7649100)
    at /home/mjambor/gcc/icln/src/gcc/tree-optimize.c:375
#6  0x08688514 in cgraph_lower_function (node=0xb75c63fc)
    at /home/mjambor/gcc/icln/src/gcc/cgraphunit.c:334
#7  0x0868a6c4 in cgraph_analyze_function (node=0xb75c63fc)
    at /home/mjambor/gcc/icln/src/gcc/cgraphunit.c:799
#8  0x086854f7 in cgraph_add_new_function (fndecl=0xb7649100, lowered=0 '\000')
    at /home/mjambor/gcc/icln/src/gcc/cgraph.c:2501
#9  0x086ad21e in cgraph_build_static_cdtor_1 (which=73 'I', body=0xb762d4b0,
priority=99, 
    final=0 '\000') at /home/mjambor/gcc/icln/src/gcc/ipa.c:1593
#10 0x08119465 in mudflap_finish_file () at
/home/mjambor/gcc/icln/src/gcc/tree-mudflap.c:1366
#11 0x0845c559 in compile_file () at
/home/mjambor/gcc/icln/src/gcc/toplev.c:601
#12 do_compile () at /home/mjambor/gcc/icln/src/gcc/toplev.c:1900
#13 toplev_main (argc=20, argv=0xbfffef34) at
/home/mjambor/gcc/icln/src/gcc/toplev.c:1963
#14 0x0816cedb in main (argc=20, argv=0xbfffef34) at
/home/mjambor/gcc/icln/src/gcc/main.c:36

[Bug tree-optimization/48330] [4.7 Regression] ICE: in optimize_inline_calls, at tree-inline.c:4201 with -fmudflap -fno-early-inlining

[jamborm at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48330

Martin Jambor <jamborm at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2011.03.29 13:13:29
                 CC|                            |hubicka at gcc dot gnu.org
     Ever Confirmed|0                           |1

--- Comment #2 from Martin Jambor <jamborm at gcc dot gnu.org> 2011-03-29 13:13:29 UTC ---
The following (untested) patch fixes the issue.  It seems OK to me as
the lowering passes probably should know that the current function is
"analyzed" but I guess we should ask Honza whether it is really
correct:

Index: src/gcc/cgraphunit.c
===================================================================
--- src.orig/gcc/cgraphunit.c
+++ src/gcc/cgraphunit.c
@@ -796,8 +796,8 @@ cgraph_analyze_function (struct cgraph_n
     gimplify_function_tree (decl);
   dump_function (TDI_generic, decl);

-  cgraph_lower_function (node);
   node->analyzed = true;
+  cgraph_lower_function (node);

   pop_cfun ();
   current_function_decl = save;

[Bug tree-optimization/48330] [4.7 Regression] ICE: in optimize_inline_calls, at tree-inline.c:4201 with -fmudflap -fno-early-inlining

[pinskia at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48330

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.7.0

[Bug tree-optimization/48330] [4.7 Regression] ICE: in optimize_inline_calls, at tree-inline.c:4201 with -fmudflap -fno-early-inlining

[jamborm at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48330

Martin Jambor <jamborm at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #3 from Martin Jambor <jamborm at gcc dot gnu.org> 2011-04-18 14:26:13 UTC ---
I wanted to re-visit this but it seems to be fixed already (with trunk
revision 172644 at least) so I'm going to close it as resolved.
Please re-open if I am somehow mistaken, thanks.