[Bug c++/16951] New: ICE with recursion in C++ constructor

[Bug c++/16951] New: ICE with recursion in C++ constructor

[rjpeters at klab dot caltech dot edu]

This looks like a 3.5 regression; not sure if I should put "[3.5 regression]" in
the subject line myself, or if I just let someone else do that.

Got an internal compiler error with the latest 3.5 snapshot with 'g++ -O1'; the
code compiles cleanly with 'g++-3.3.3 -O1', also with 'g++-3.4.1 -O1', and also
with recent 3.4 snapshot 'g++-3.4-20040730 -O1'.

[sideswipe 02:33 PM 1]$ cat fail.cc

struct foo;

void dummy_use(foo*);

struct foo
{
  foo()
  {
    dummy_use(this);
    foo r;
  }
};

void crash()
{
  foo r;
}

[sideswipe 02:33 PM 2]$ \
/home/rjpeters/local/gcc-3.5-20040808/bin/g++-3.5-20040808 -v -O1 -dHa gdb-c \
fail.cc

Reading specs from
/home/rjpeters/local/gcc-3.5-20040808/lib/gcc/i686-pc-linux-gnu/3.5.0/specs
Configured with: /home/rjpeters/build/gcc-3.5-20040808/configure
--prefix=/home/rjpeters/local/gcc-3.5-20040808 --program-suffix=-3.5-20040808
--enable-shared -enable-threads=posix --enable-checking --with-system-zlib
--enable-__cxa_atexit --enable-languages=c,c++
Thread model: posix
gcc version 3.5.0 20040808 (experimental)
 /home/rjpeters/local/gcc-3.5-20040808/libexec/gcc/i686-pc-linux-gnu/3.5.0/cc1plus
-quiet -v -D_GNU_SOURCE fail.cc -quiet -dumpbase fail.cc -mtune=pentiumpro
-auxbase fail -O1 -version -o /tmp/ccM74Apb.s
ignoring nonexistent directory "/usr/local/include"
ignoring nonexistent directory
"/home/rjpeters/local/gcc-3.5-20040808/lib/gcc/i686-pc-linux-gnu/3.5.0/../../../../i686-pc-linux-gnu/include"
#include "..." search starts here:
#include <...> search starts here:
 /home/rjpeters/local/gcc-3.5-20040808/lib/gcc/i686-pc-linux-gnu/3.5.0/../../../../include/c++/3.5.0
 /home/rjpeters/local/gcc-3.5-20040808/lib/gcc/i686-pc-linux-gnu/3.5.0/../../../../include/c++/3.5.0/i686-pc-linux-gnu
 /home/rjpeters/local/gcc-3.5-20040808/lib/gcc/i686-pc-linux-gnu/3.5.0/../../../../include/c++/3.5.0/backward
 /home/rjpeters/local/gcc-3.5-20040808/include
 /home/rjpeters/local/gcc-3.5-20040808/lib/gcc/i686-pc-linux-gnu/3.5.0/include
 /usr/include
End of search list.
GNU C++ version 3.5.0 20040808 (experimental) (i686-pc-linux-gnu)
	compiled by GNU C version 3.5.0 20040808 (experimental).
GGC heuristics: --param ggc-min-expand=30 --param ggc-min-heapsize=4096

fail.cc: In constructor `foo::foo()':
fail.cc:9: internal compiler error: in make_decl_rtl, at varasm.c:768
g++-3.5-20040808: Internal error: Aborted (program cc1plus)
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://gcc.gnu.org/bugs.html> for instructions.

[sideswipe 02:35 PM 3]$ gdb \
/home/rjpeters/local/gcc-3.5-20040808/libexec/gcc/i686-pc-linux-gnu/3.5.0/cc1plus\
core.15430

GNU gdb Red Hat Linux (6.0post-0.20040223.19rh)
Copyright 2004 Free Software Foundation, Inc.
GDB is free software, covered by the GNU General Public License, and you are
welcome to change it and/or distribute copies of it under certain conditions.
Type "show copying" to see the conditions.
There is absolutely no warranty for GDB.  Type "show warranty" for details.
This GDB was configured as "i386-redhat-linux-gnu"...Using host libthread_db
library "/lib/tls/libthread_db.so.1".

Core was generated by
`/home/rjpeters/local/gcc-3.5-20040808/libexec/gcc/i686-pc-linux-gnu/3.5.0/cc1pl'.
Program terminated with signal 6, Aborted.

warning: svr4_current_sos: Can't read pathname for load map: Input/output error

Error while mapping shared library sections:
: Success.
Error while reading shared library symbols:
: No such file or directory.
Reading symbols from /lib/tls/libc.so.6...done.
Loaded symbols for /lib/tls/libc.so.6
Reading symbols from /lib/ld-linux.so.2...done.
Loaded symbols for /lib/ld-linux.so.2
#0  0x00a41402 in ?? ()
(gdb) bt
#0  0x00a41402 in ?? ()
#1  0x0016ee49 in raise () from /lib/tls/libc.so.6
#2  0x00170872 in abort () from /lib/tls/libc.so.6
#3  0x0826f03b in real_abort ()
    at /home/rjpeters/build/gcc-3.5-20040808/gcc/diagnostic.c:604
#4  0x0826f1d5 in diagnostic_report_diagnostic (context=0x860f080, 
    diagnostic=0x25effc)
    at /home/rjpeters/build/gcc-3.5-20040808/gcc/diagnostic.c:291
#5  0x0826f283 in internal_error (msgid=0x0)
    at /home/rjpeters/build/gcc-3.5-20040808/gcc/diagnostic.c:548
#6  0x0826f2bf in fancy_abort (file=0x0, line=6, function=0x0)
    at /home/rjpeters/build/gcc-3.5-20040808/gcc/diagnostic.c:594
#7  0x0849726e in make_decl_rtl (decl=0xf6dd5f68)
    at /home/rjpeters/build/gcc-3.5-20040808/gcc/varasm.c:768
#8  0x082b9cb8 in expand_expr_real_1 (exp=0xf6dd5f68, target=0xf6dd7078, 
    tmode=VOIDmode, modifier=EXPAND_CONST_ADDRESS, alt_rtl=0x0)
    at /home/rjpeters/build/gcc-3.5-20040808/gcc/expr.c:6344
#9  0x082c28fc in expand_expr_real (exp=0xf6dd5f68, target=0x0, 
    tmode=VOIDmode, modifier=EXPAND_CONST_ADDRESS, alt_rtl=0x0)
    at /home/rjpeters/build/gcc-3.5-20040808/gcc/expr.c:6177
#10 0x082b413e in expand_expr_real_1 (exp=0xf6dd6160, target=Variable "target"
is not available.
) at expr.h:497
#11 0x082c28fc in expand_expr_real (exp=0xf6dd6160, target=0xf6dd9534, 
    tmode=VOIDmode, modifier=EXPAND_STACK_PARM, alt_rtl=0x0)
    at /home/rjpeters/build/gcc-3.5-20040808/gcc/expr.c:6177
#12 0x0823bc55 in store_one_arg (arg=0xfefa0f20, argblock=0xf6e86070, 
    flags=0, variable_size=0, reg_parm_stack_space=0) at expr.h:497
#13 0x0823dfd7 in expand_call (exp=0xf6e13960, target=0x0, ignore=1)
    at /home/rjpeters/build/gcc-3.5-20040808/gcc/calls.c:2606
#14 0x082b3aa6 in expand_expr_real_1 (exp=0xf6e13960, target=Variable "target"
is not available.
)
    at /home/rjpeters/build/gcc-3.5-20040808/gcc/expr.c:7107
#15 0x082c2a32 in expand_expr_real (exp=0xf6e13960, target=0xf6e85210, 
    tmode=VOIDmode, modifier=EXPAND_NORMAL, alt_rtl=0x0)
    at /home/rjpeters/build/gcc-3.5-20040808/gcc/expr.c:6171
#16 0x0846b515 in expand_expr_stmt (exp=0xf6e13960) at expr.h:497
#17 0x084abc8b in tree_expand_cfg ()
    at /home/rjpeters/build/gcc-3.5-20040808/gcc/cfgexpand.c:295
#18 0x081b3ab2 in execute_pass_list (pass=0x860a7c0)
    at /home/rjpeters/build/gcc-3.5-20040808/gcc/tree-optimize.c:453
#19 0x081b3d21 in tree_rest_of_compilation (fndecl=0xf6dd4074, 
    nested_p=0 '\0')
    at /home/rjpeters/build/gcc-3.5-20040808/gcc/tree-optimize.c:556
#20 0x08131d71 in expand_body (fn=0xf6dd4074)
    at /home/rjpeters/build/gcc-3.5-20040808/gcc/cp/semantics.c:2882
#21 0x084cc277 in cgraph_expand_function (node=0xf6dd4d98)
    at /home/rjpeters/build/gcc-3.5-20040808/gcc/cgraphunit.c:801
#22 0x084cdb3f in cgraph_optimize ()
    at /home/rjpeters/build/gcc-3.5-20040808/gcc/cgraphunit.c:1673
#23 0x080d96be in finish_file ()
    at /home/rjpeters/build/gcc-3.5-20040808/gcc/cp/decl2.c:3056
#24 0x081894da in c_common_parse_file (set_yydebug=0)
    at /home/rjpeters/build/gcc-3.5-20040808/gcc/c-opts.c:1093
#25 0x08478d55 in toplev_main (argc=0, argv=0xfefa1674)
    at /home/rjpeters/build/gcc-3.5-20040808/gcc/toplev.c:976
#26 0x0015cad4 in __libc_start_main () from /lib/tls/libc.so.6
#27 0x08049b31 in _start ()


[sideswipe 02:40 PM 5]$ ls fail.cc.*
fail.cc.00.cgraph

I'll attach 'fail.cc' and 'fail.cc.00.cgraph'.

Cheers,
Rob

-- 
           Summary: ICE with recursion in C++ constructor
           Product: gcc
           Version: 3.5.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: c++
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: rjpeters at klab dot caltech dot edu
                CC: gcc-bugs at gcc dot gnu dot org
 GCC build triplet: i686-pc-linux-gnu
  GCC host triplet: i686-pc-linux-gnu
GCC target triplet: i686-pc-linux-gnu


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16951

[Bug c++/16951] ICE with recursion in C++ constructor

[rjpeters at klab dot caltech dot edu]

------- Additional Comments From rjpeters at klab dot caltech dot edu  2004-08-09 21:47 -------
Created an attachment (id=6908)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=6908&action=view)
gives an ICE with 'g++-3.5-20040808 -O1'


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16951

[Bug c++/16951] ICE with recursion in C++ constructor

[rjpeters at klab dot caltech dot edu]

------- Additional Comments From rjpeters at klab dot caltech dot edu  2004-08-09 21:48 -------
Created an attachment (id=6909)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=6909&action=view)
debug dump from aforementioned compilation


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16951

[Bug c++/16951] [3.5 regression] ICE in make_decl_rtl with recursion in C++ constructor

[bangerth at dealii dot org]

------- Additional Comments From bangerth at dealii dot org  2004-08-09 22:39 -------
Confirmed, a regression on mainline. 
W. 

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|                            |1
           Keywords|                            |ice-on-valid-code
      Known to fail|                            |3.5.0
      Known to work|                            |3.4.1
   Last reconfirmed|0000-00-00 00:00:00         |2004-08-09 22:39:53
               date|                            |
            Summary|ICE with recursion in C++   |[3.5 regression] ICE in
                   |constructor                 |make_decl_rtl with recursion
                   |                            |in C++ constructor
   Target Milestone|---                         |3.5.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16951

[Bug tree-optimization/16951] [3.5 regression] ICE in make_decl_rtl with recursion in C++ constructor (-O1)

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-08-10 00:16 -------
: Search converges between 2004-07-29-trunk (#498) and 2004-07-30-trunk (#499).

Hmm, I did a -fdump-tree-vars to find out why this was ICE and I noticed that before, it was inlining  
the constructor once and after it was inlining it a couple of times.  Also after it removed some variables 
which it should not be removed from scope.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|normal                      |critical
          Component|c++                         |tree-optimization
            Summary|[3.5 regression] ICE in     |[3.5 regression] ICE in
                   |make_decl_rtl with recursion|make_decl_rtl with recursion
                   |in C++ constructor          |in C++ constructor (-O1)


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16951

[Bug tree-optimization/16951] [3.5 regression] ICE in make_decl_rtl with recursion in C++ constructor (-O1)

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-08-10 01:35 -------
Note with -O3, I get a different dates:
: Search converges between 2004-07-06-trunk (#480) and 2004-07-07-trunk (#481).

Which means that the dates for -O1 is really just a side effect of more inlining (or maybe not).

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16951

[Bug tree-optimization/16951] [4.0 regression] ICE in make_decl_rtl with recursion in C++ constructor (-O1)

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-09-21 13:48 -------
Hmm, this looks like an inlining bug, the problem is that we are not adding the correct variables in the 
first place.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
      Known to fail|4.0                         |4.0.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16951

[Bug tree-optimization/16951] [4.0 regression] ICE in make_decl_rtl with recursion in C++ constructor (-O1)

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-12-02 18:37 -------
*** Bug 18790 has been marked as a duplicate of this bug. ***

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |tsarkov at cs dot man dot ac
                   |                            |dot uk


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16951

[Bug tree-optimization/16951] [4.0 regression] ICE in make_decl_rtl with recursion in C++ constructor (-O1)

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-12-09 22:09 -------
*** Bug 18914 has been marked as a duplicate of this bug. ***

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |hans dot buchmann at fhso
                   |                            |dot ch


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16951

[Bug tree-optimization/16951] [4.0 regression] ICE in make_decl_rtl with recursion in C++ constructor (-O1)

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-12-11 22:20 -------
Patch here: <http://gcc.gnu.org/ml/gcc-patches/2004-12/msg00807.html>.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |patch


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16951

[Bug tree-optimization/16951] [4.0 regression] ICE in make_decl_rtl with recursion in C++ constructor (-O1)

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2004-12-13 21:03 -------
Subject: Bug 16951

CVSROOT:	/cvs/gcc
Module name:	gcc
Changes by:	aoliva@gcc.gnu.org	2004-12-13 21:03:47

Modified files:
	gcc            : ChangeLog tree-inline.c 
	gcc/testsuite  : ChangeLog 
Added files:
	gcc/testsuite/gcc.c-torture/compile: 20041211-1.c 

Log message:
	gcc/ChangeLog:
	PR tree-opt/16951
	* tree-inline.c (setup_one_parameter): Don't directly map a
	parameter to the address of another variable of the same
	function.
	gcc/testsuite/ChangeLog:
	PR tree-opt/16951
	* gcc.c-torture/compile/20041211-1.c: New.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.diff?cvsroot=gcc&r1=2.6815&r2=2.6816
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/tree-inline.c.diff?cvsroot=gcc&r1=1.156&r2=1.157
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/ChangeLog.diff?cvsroot=gcc&r1=1.4743&r2=1.4744
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/gcc.c-torture/compile/20041211-1.c.diff?cvsroot=gcc&r1=NONE&r2=1.1



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16951

[Bug tree-optimization/16951] [4.0 regression] ICE in make_decl_rtl with recursion in C++ constructor (-O1)

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-12-13 21:12 -------
Fixed, and I doubled checked to make sure that all of the bugs which were marked as a dup as fixed 
too.

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16951

[Bug tree-optimization/16951] [4.0 regression] ICE in make_decl_rtl with recursion in C++ constructor (-O1)

[rjpeters at klab dot caltech dot edu]

------- Additional Comments From rjpeters at klab dot caltech dot edu  2004-12-13 21:18 -------
Thanks for the bugfix.
Cheers,
Rob

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16951

[Bug tree-optimization/16951] [4.0 regression] ICE in make_decl_rtl with recursion in C++ constructor (-O1)

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2004-12-23 19:37 -------
Subject: Bug 16951

CVSROOT:	/cvs/gcc
Module name:	gcc
Branch: 	apple-ppc-branch
Changes by:	dalej@gcc.gnu.org	2004-12-23 19:36:58

Modified files:
	gcc            : ChangeLog.apple-ppc tree.c tree-inline.c 

Log message:
	2004-12-23  Dale Johannesen  <dalej@apple.com>
	
	Radar 3926600
	Following patches brought in from mainline
	* tree.c (iterative_hash_expr):  Canonicalize builtins.
	PR tree-opt/16951
	* tree-inline.c (setup_one_parameter): Don't directly map a
	parameter to the address of another variable of the same
	function.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.apple-ppc.diff?cvsroot=gcc&only_with_tag=apple-ppc-branch&r1=1.1.2.248&r2=1.1.2.249
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/tree.c.diff?cvsroot=gcc&only_with_tag=apple-ppc-branch&r1=1.263.2.84.2.24&r2=1.263.2.84.2.25
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/tree-inline.c.diff?cvsroot=gcc&only_with_tag=apple-ppc-branch&r1=1.26.2.87.2.9&r2=1.26.2.87.2.10



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16951