[Bug tree-optimization/46183] New: [4.6 Regression] ICE: in calc_dfs_tree, at dominance.c:396 with -O -fno-dse -fgcse -ftree-pre

[Bug tree-optimization/46183] New: [4.6 Regression] ICE: in calc_dfs_tree, at dominance.c:396 with -O -fno-dse -fgcse -ftree-pre

[zsojka at seznam dot cz]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46183

           Summary: [4.6 Regression] ICE: in calc_dfs_tree, at
                    dominance.c:396 with -O -fno-dse -fgcse -ftree-pre
           Product: gcc
           Version: 4.6.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: tree-optimization
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: zsojka@seznam.cz


Created attachment 22159
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=22159
reduced testcase

Compiler output:
$ gcc -O -fno-dse -fgcse -ftree-pre pr46183.c 
pr46183.c: In function 'foo':
pr46183.c:16:1: internal compiler error: in calc_dfs_tree, at dominance.c:396
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.

Tested revisions:
r165915 - crash
r161659 - crash
r159696 - OK
4.5 r165781 - OK

[Bug tree-optimization/46183] [4.6 Regression] ICE: in calc_dfs_tree, at dominance.c:396 with -O -fno-dse -fgcse -ftree-pre

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46183

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.6.0

[Bug tree-optimization/46183] [4.6 Regression] ICE: in calc_dfs_tree, at dominance.c:396 with -O -fno-dse -fgcse -ftree-pre

[hjl.tools at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46183

H.J. Lu <hjl.tools at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2010.10.26 15:03:00
                 CC|                            |rguenth at gcc dot gnu.org
     Ever Confirmed|0                           |1

--- Comment #1 from H.J. Lu <hjl.tools at gmail dot com> 2010-10-26 15:03:00 UTC ---
It is caused by revision 161655:

http://gcc.gnu.org/ml/gcc-cvs/2010-07/msg00006.html

[Bug tree-optimization/46183] [4.6 Regression] ICE: in calc_dfs_tree, at dominance.c:396 with -O -fno-dse -fgcse -ftree-pre

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46183

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P1
                 CC|                            |jakub at gcc dot gnu.org

[Bug tree-optimization/46183] [4.6 Regression] ICE: in calc_dfs_tree, at dominance.c:396 with -O -fno-dse -fgcse -ftree-pre

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46183

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
         AssignedTo|unassigned at gcc dot       |rguenth at gcc dot gnu.org
                   |gnu.org                     |

--- Comment #2 from Richard Guenther <rguenth at gcc dot gnu.org> 2010-10-29 12:52:07 UTC ---
Mine.

[Bug tree-optimization/46183] [4.6 Regression] ICE: in calc_dfs_tree, at dominance.c:396 with -O -fno-dse -fgcse -ftree-pre

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46183

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |NEW
                 CC|                            |stevenb.gcc at gmail dot
                   |                            |com
         AssignedTo|rguenth at gcc dot gnu.org  |unassigned at gcc dot
                   |                            |gnu.org

--- Comment #3 from Richard Guenther <rguenth at gcc dot gnu.org> 2010-11-02 13:58:05 UTC ---
RTL cprop leaves us with a not cleaned up CFG (an unreachable block), that
causes dominance computation to ICE.

;; Start of basic block () -> 16
;; bb 16 artificial_defs: { }
;; bb 16 artificial_uses: { u109(6){ }u110(7){ }u111(16){ }u112(20){ }}
;; lr  in        6 [bp] 7 [sp] 16 [argp] 20 [frame] 74
;; lr  use       6 [bp] 7 [sp] 16 [argp] 20 [frame]
;; lr  def       17 [flags]

(code_label 90 87 91 16 14 "" [0 uses])

(note 91 90 92 16 [bb 16] NOTE_INSN_BASIC_BLOCK)

(insn 92 91 93 16 (parallel [
            (set (mem/s/f/j/c:DI (plus:DI (reg/f:DI 20 frame)
                        (const_int -16 [0xfffffffffffffff0])) [0
ap.overflow_arg_area+0 S8 A64])
                (plus:DI (mem/s/f/j/c:DI (plus:DI (reg/f:DI 20 frame)
                            (const_int -16 [0xfffffffffffffff0])) [0
ap.overflow_arg_area+0 S8 A64])
                    (const_int 8 [0x8])))
            (clobber (reg:CC 17 flags))
        ]) t.c:13 253 {*adddi_1}
     (expr_list:REG_UNUSED (reg:CC 17 flags)
        (nil)))
;; End of basic block 16 -> ( 17)
;; lr  out       6 [bp] 7 [sp] 16 [argp] 20 [frame] 74


;; Succ edge  17 [100.0%]  (fallthru)


The solution may be as simple as making sure to remove unreachable blocks
after cprop, but already the exiting

  if (local_cprop_pass ())
    {
      delete_unreachable_blocks ();
      df_analyze ();
    }

has some weird comment in front of it.  Maybe steven remembers what's going
on here.

Not my bug, really only exposed by me.

[Bug tree-optimization/46183] [4.6 Regression] ICE: in calc_dfs_tree, at dominance.c:396 with -O -fno-dse -fgcse -ftree-pre

[steven at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46183

Steven Bosscher <steven at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
                 CC|                            |steven at gcc dot gnu.org

--- Comment #4 from Steven Bosscher <steven at gcc dot gnu.org> 2010-11-02 14:37:21 UTC ---
There is nothing weird about the comment before the quoted code. It just
mentions that the simple solver for global CPROP does not work if there are
unreachable blocks in the CFG (and that the DF solver is insensitive to this).

But oh well, I'll have a look...

[Bug tree-optimization/46183] [4.6 Regression] ICE: in calc_dfs_tree, at dominance.c:396 with -O -fno-dse -fgcse -ftree-pre

[steven at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46183

--- Comment #5 from Steven Bosscher <steven at gcc dot gnu.org> 2010-11-02 14:48:28 UTC ---
For the record:

Program received signal SIGSEGV, Segmentation fault.
0x000000000057f64c in flow_loops_find (loops=0x7ffff72516e0) at
../../trunk/gcc/cfgloop.c:462
462               if (!TEST_BIT (headers, rc_order[b]))
(gdb) bt
#0  0x000000000057f64c in flow_loops_find (loops=0x7ffff72516e0) at
../../trunk/gcc/cfgloop.c:462
#1  0x0000000000703e36 in loop_optimizer_init (flags=0) at
../../trunk/gcc/loop-init.c:52
#2  0x0000000000cc7c2b in fwprop_init () at ../../trunk/gcc/fwprop.c:1381
#3  0x0000000000cc7ec9 in fwprop_addr () at ../../trunk/gcc/fwprop.c:1464
#4  0x000000000074ab8f in execute_one_pass (pass=0x1380200) at
../../trunk/gcc/passes.c:1560
#5  0x000000000074ae2c in execute_pass_list (pass=0x1380200) at
../../trunk/gcc/passes.c:1615
#6  0x000000000074ae4a in execute_pass_list (pass=0x137a620) at
../../trunk/gcc/passes.c:1616
#7  0x0000000000875b7e in tree_rest_of_compilation (fndecl=0x7ffff71f0000) at
../../trunk/gcc/tree-optimize.c:422
#8  0x0000000000a13ec9 in cgraph_expand_function (node=0x7ffff71f3000) at
../../trunk/gcc/cgraphunit.c:1494
#9  0x0000000000a13ff0 in cgraph_expand_all_functions () at
../../trunk/gcc/cgraphunit.c:1553
#10 0x0000000000a14582 in cgraph_optimize () at
../../trunk/gcc/cgraphunit.c:1809
#11 0x0000000000a12a8f in cgraph_finalize_compilation_unit () at
../../trunk/gcc/cgraphunit.c:1016
#12 0x00000000004a69df in c_write_global_declarations () at
../../trunk/gcc/c-decl.c:9827
#13 0x000000000081c8cf in compile_file () at ../../trunk/gcc/toplev.c:931
#14 0x000000000081ec3e in do_compile () at ../../trunk/gcc/toplev.c:2359
#15 0x000000000081ed4a in toplev_main (argc=6, argv=0x7fffffffe2a8) at
../../trunk/gcc/toplev.c:2419
#16 0x0000000000525643 in main (argc=6, argv=0x7fffffffe2a8) at
../../trunk/gcc/main.c:36
(gdb) up 5
#5  0x000000000074ae2c in execute_pass_list (pass=0x1380200) at
../../trunk/gcc/passes.c:1615
1615          if (execute_one_pass (pass) && pass->sub)
(gdb) p *pass
$1 = {type = RTL_PASS, name = 0x1024d89 "fwprop2", gate = 0xcc7cba
<gate_fwprop>, execute = 0xcc7ebc <fwprop_addr>, sub = 0x0,
  next = 0x137fb00, static_pass_number = 180, tv_id = TV_FWPROP,
properties_required = 0, properties_provided = 0,
  properties_destroyed = 0, todo_flags_start = 524288, todo_flags_finish =
132097}
(gdb)

[Bug tree-optimization/46183] [4.6 Regression] ICE: in calc_dfs_tree, at dominance.c:396 with -O -fno-dse -fgcse -ftree-pre

[steven at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46183

Steven Bosscher <steven at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |NEW
         AssignedTo|steven at gcc dot gnu.org   |unassigned at gcc dot
                   |                            |gnu.org

--- Comment #6 from Steven Bosscher <steven at gcc dot gnu.org> 2010-11-02 15:02:39 UTC ---
Fix would look something like this. No time to finish this, though.

Index: gcse.c
===================================================================
--- gcse.c      (revision 166183)
+++ gcse.c      (working copy)
@@ -5246,11 +5246,13 @@ gate_rtl_cprop (void)
 static unsigned int
 execute_rtl_cprop (void)
 {
+  int changed;
   delete_unreachable_blocks ();
   df_set_flags (DF_LR_RUN_DCE);
   df_analyze ();
-  flag_rerun_cse_after_global_opts |= one_cprop_pass ();
-  return 0;
+  changed = one_cprop_pass ();
+  flag_rerun_cse_after_global_opts |= changed;
+  return changed ? TODO_cleanup_cfg : 0;
 }

 static bool
@@ -5265,10 +5267,12 @@ gate_rtl_pre (void)
 static unsigned int
 execute_rtl_pre (void)
 {
+  int changed;
   delete_unreachable_blocks ();
   df_analyze ();
-  flag_rerun_cse_after_global_opts |= one_pre_gcse_pass ();
-  return 0;
+  changed = one_pre_gcse_pass ();
+  flag_rerun_cse_after_global_opts |= changed;
+  return changed ? TODO_cleanup_cfg : 0;
 }

 static bool
@@ -5286,10 +5290,12 @@ gate_rtl_hoist (void)
 static unsigned int
 execute_rtl_hoist (void)
 {
+  int changed;
   delete_unreachable_blocks ();
   df_analyze ();
-  flag_rerun_cse_after_global_opts |= one_code_hoisting_pass ();
-  return 0;
+  changed = one_code_hoisting_pass ();
+  flag_rerun_cse_after_global_opts |= changed;
+  return changed ? TODO_cleanup_cfg : 0;
 }

 struct rtl_opt_pass pass_rtl_cprop =

[Bug tree-optimization/46183] [4.6 Regression] ICE: in calc_dfs_tree, at dominance.c:396 with -O -fno-dse -fgcse -ftree-pre

[ebotcazou at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46183

Eric Botcazou <ebotcazou at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |ebotcazou at gcc dot
                   |                            |gnu.org

--- Comment #7 from Eric Botcazou <ebotcazou at gcc dot gnu.org> 2010-11-02 16:04:07 UTC ---
TODO_cleanup_cfg doesn't work for RTL though.

[Bug tree-optimization/46183] [4.6 Regression] ICE: in calc_dfs_tree, at dominance.c:396 with -O -fno-dse -fgcse -ftree-pre

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46183

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
         AssignedTo|unassigned at gcc dot       |rguenth at gcc dot gnu.org
                   |gnu.org                     |

--- Comment #8 from Richard Guenther <rguenth at gcc dot gnu.org> 2010-11-04 14:02:01 UTC ---
Testing cleanup_cfg (0) instead.

[Bug tree-optimization/46183] [4.6 Regression] ICE: in calc_dfs_tree, at dominance.c:396 with -O -fno-dse -fgcse -ftree-pre

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46183

--- Comment #9 from Richard Guenther <rguenth at gcc dot gnu.org> 2010-11-04 15:19:24 UTC ---
Author: rguenth
Date: Thu Nov  4 15:19:19 2010
New Revision: 166316

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=166316
Log:
2010-11-04  Richard Guenther  <rguenther@suse.de>

    PR rtl-optimization/46183
    * gcse.c (execute_rtl_cprop): Cleanup the CFG if something changed.
    (execute_rtl_pre): Likewise.
    (execute_rtl_hoist): Likewise.

    * gcc.dg/torture/pr46183.c: New testcase.

Added:
    trunk/gcc/testsuite/gcc.dg/torture/pr46183.c
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/gcse.c
    trunk/gcc/testsuite/ChangeLog

[Bug tree-optimization/46183] [4.6 Regression] ICE: in calc_dfs_tree, at dominance.c:396 with -O -fno-dse -fgcse -ftree-pre

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46183

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

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

--- Comment #10 from Richard Guenther <rguenth at gcc dot gnu.org> 2010-11-04 15:21:00 UTC ---
Fixed.