[Bug c++/15853] New: temporaries are not destroyed and overwritten later

[Bug c++/15853] New: temporaries are not destroyed and overwritten later

[bugzilla at little-bat dot de]

when initialilizing static data, then temporaries which were required for
automatic type casting are not destroyed. later-on the stack/heap overwrites them. 

example source file can be found at
<http://k1.dyndns.org/Develop/projects/gcc-error>

test.cpp           example source file which reproduces the error
output-BUMMER.txt  output printed by test program when broken
output-OK.txt      output printed by test program when ok
a.out              FC2 i386 binary, hopefully the broken version B-)

used command line:
bash$  gcc -Wall -O2 -lstdc++ test.cpp        // fails
bash$  gcc -Wall     -lstdc++ test.cpp        // ok

see also short summary at start of source file.

the error only shows up under certain circumstances. e.g. in this case no "-O2"
is ok and when destructor and/or creator are inlined it seemed ok too. then the
destructor was called. see output printed by test program.

this bug report form has no file upload section, which i find very confusing.
just wanted to say that too.

           ... kio !

-- 
           Summary: temporaries are not destroyed and overwritten later
           Product: gcc
           Version: 3.3.3
            Status: UNCONFIRMED
          Severity: critical
          Priority: P2
         Component: c++
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: bugzilla at little-bat dot de
                CC: gcc-bugs at gcc dot gnu dot org
 GCC build triplet: gcc-Version 3.3.3 20040412 (Red Hat Linux 3.3.3-7)
  GCC host triplet: i386-redhat-linux
GCC target triplet: i386-redhat-linux


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15853

[Bug rtl-optimization/15853] [3.3 Regression] temporaries are not destroyed and overwritten later

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-06-06 18:54 -------
Confirmed, a regression from 3.0.4.  Note it works in 3.4.0 and 3.5.0.

Also note that -O2 -fno-inline -fno-gcse -fno-strict-aliasing also fails so it has nothing to do with 
inlining.  Also -O1 -fno-inline passes.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
          Component|c++                         |rtl-optimization
     Ever Confirmed|                            |1
           Keywords|                            |wrong-code
      Known to fail|                            |3.3.3
      Known to work|                            |3.4.0 3.5.0 3.0.4
   Last reconfirmed|0000-00-00 00:00:00         |2004-06-06 18:54:47
               date|                            |
            Summary|temporaries are not         |[3.3 Regression] temporaries
                   |destroyed and overwritten   |are not destroyed and
                   |later                       |overwritten later
   Target Milestone|---                         |3.3.4


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15853

[Bug rtl-optimization/15853] [3.3 Regression] temporaries are not destroyed and overwritten later

[bugzilla at little-bat dot de]

------- Additional Comments From bugzilla at little-bat dot de  2004-06-07 19:11 -------
The conclusion that it has nothing to do with inlining is wrong: the error 
occured with not-inlined member functions and adding -fno-inline to the 
command line is just a nop. :-) 
I tested a version of the test proggi where constructor & destructor are 
inlined: 
 
gcc -Wall -O2 -lstdc++             test-inlined.cpp    -> a.out  ok 
gcc -Wall -O2 -lstdc++ -fno-inline test-inlined.cpp    -> a.out  BUMMER 
gcc -Wall -O1 -lstdc++ -fno-inline test-inlined.cpp    -> a.out  ok 
 
see <http://k1.dyndns.org/Develop/projects/gcc-error> 
 
       ... kio ! 

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15853

[Bug rtl-optimization/15853] [3.3 Regression] temporaries are not destroyed and overwritten later

[pinskia at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|3.3.4                       |3.3.5


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15853

[Bug rtl-optimization/15853] [3.3 Regression] temporaries are not destroyed and overwritten later

[gdr at gcc dot gnu dot org]

------- Additional Comments From gdr at gcc dot gnu dot org  2004-09-28 13:33 -------
Adjust milestone

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
      Known to work|3.4.0 4.0 3.0.4             |3.4.0 4.0.0 3.0.4
   Target Milestone|3.3.5                       |3.3.6


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15853

[Bug rtl-optimization/15853] [3.3 Regression] temporaries are not destroyed and overwritten later

[ebotcazou at gcc dot gnu dot org]

------- Additional Comments From ebotcazou at gcc dot gnu dot org  2005-01-23 14:49 -------
A sibcall problem, hence the workaround -O2 -fno-optimize-sibling-calls.


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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15853

[Bug rtl-optimization/15853] [3.3 Regression] temporaries are not destroyed and overwritten later

[ebotcazou at gcc dot gnu dot org]

------- Additional Comments From ebotcazou at gcc dot gnu dot org  2005-01-24 07:53 -------
Fixing.


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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15853

[Bug rtl-optimization/15853] [3.3 Regression] temporaries are not destroyed and overwritten later

[ebotcazou at gcc dot gnu dot org]

------- Additional Comments From ebotcazou at gcc dot gnu dot org  2005-01-24 11:27 -------
Um... first of all, this works on 3.4 branch only by accident, i.e. I think the
underlying problem is still present there.  What happens is that a call has an
argument containing a TARGET_EXPR with cleanups and is eligible for the sibling
call optimization.  The cleanups are expanded during the first pass but, since
the optimization eventually fails, the RTL is thrown away.  Then, during the
second pass, the TARGET_EXPR is expanded again but not the cleanups because they
are not registered (the variable 'cleanups' is NULL at expr.c:9050).

I'm not sure how this is supposed to work.  Richard, do you have any
recollection of this?  In 2000(!), you commited this patch:

Index: calls.c
===================================================================
RCS file: /cvs/gcc/gcc/gcc/calls.c,v
retrieving revision 1.97
retrieving revision 1.98
diff -u -r1.97 -r1.98
--- calls.c     17 Mar 2000 22:40:43 -0000      1.97
+++ calls.c     20 Mar 2000 22:40:50 -0000      1.98
@@ -2020,7 +2020,8 @@
   safe_for_reeval = 0;
   if (optimize >= 2
       && currently_expanding_call == 1
-      && stmt_loop_nest_empty ())
+      && stmt_loop_nest_empty ()
+      && ! any_pending_cleanups (1))
     {
       /* Verify that each argument is safe for re-evaluation.  */
       for (p = actparms; p; p = TREE_CHAIN (p))
@@ -2152,6 +2153,12 @@
              || ! FUNCTION_OK_FOR_SIBCALL (fndecl))
            continue;

+         /* We know at this point that there are not currently any
+            pending cleanups.  If, however, in the process of evaluating
+            the arguments we were to create some, we'll need to be
+            able to get rid of them.  */
+         expand_start_target_temps ();
+
          /* State variables we need to save and restore between
             iterations.  */
          save_pending_stack_adjust = pending_stack_adjust;
@@ -2925,6 +2932,14 @@
        if (args[i].aligned_regs)
          free (args[i].aligned_regs);

+      if (pass == 0)
+       {
+         /* Undo the fake expand_start_target_temps we did earlier.  If
+            there had been any cleanups created, we've already set
+            sibcall_failure.  */
+         expand_end_target_temps ();
+       }
+
       insns = get_insns ();
       end_sequence ();

which is responsible for expanding the cleanups right after the first pass.


-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |rth at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15853

[Bug rtl-optimization/15853] [3.3 Regression] temporaries are not destroyed and overwritten later

[ebotcazou at gcc dot gnu dot org]

------- Additional Comments From ebotcazou at gcc dot gnu dot org  2005-01-25 16:42 -------
Stepping down for now.


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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15853

[Bug rtl-optimization/15853] [3.3 Regression] temporaries are not destroyed and overwritten later

[gdr at gcc dot gnu dot org]

------- Additional Comments From gdr at gcc dot gnu dot org  2005-04-28 08:23 -------
(In reply to comment #6)
> Um... first of all, this works on 3.4 branch only by accident, i.e. I think the
> underlying problem is still present there.  What happens is that a call has an
> argument containing a TARGET_EXPR with cleanups and is eligible for the sibling
> call optimization.  The cleanups are expanded during the first pass but, since
> the optimization eventually fails, the RTL is thrown away.  Then, during the
> second pass, the TARGET_EXPR is expanded again but not the cleanups because they
> are not registered (the variable 'cleanups' is NULL at expr.c:9050).
> 
> I'm not sure how this is supposed to work.  Richard, do you have any
> recollection of this?  In 2000(!), you commited this patch:

RTH --
  I guess you did not have interest in this?  Shall I close it as WONTFIX?

-- Gaby


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15853

[Bug rtl-optimization/15853] [3.3 Regression] temporaries are not destroyed and overwritten later

[gdr at gcc dot gnu dot org]

------- Additional Comments From gdr at gcc dot gnu dot org  2005-04-30 13:44 -------
Chnages to fix this bug would happen in the middle=end which is risky at
this point.  Therefore it won't be fixed for 3.3.6.  It fixed in 3.4.0 though.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED
   Target Milestone|3.3.6                       |3.4.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15853