[Bug tree-optimization/61289] New: Bad jump threading generates infinite loop

[Bug tree-optimization/61289] New: Bad jump threading generates infinite loop

[xinliangli at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61289

            Bug ID: 61289
           Summary: Bad jump threading generates infinite loop
           Product: gcc
           Version: unknown
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: tree-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: xinliangli at gmail dot com

Build the following program with -fno-exceptions -O2, the program aborts at
runtime.

Adding -fno-tree-dominator-opts, the program runs fine.

To trigger the problem, the first iteration needs to return list == new_list
--> the code will then get into the cloned path which is actually an infinite
loop.

(the toy program does something nonsense, but it should finish without
aborting).

jpthread.h
---------

struct MyList {
 inline int LastHit () const { return last_; }
 // inline int FirstHit () const { return last_-1; }

 MyList* PushBack(int);
// MyList* PushFront(int);
 void Destroy();

 static MyList* Create(int);
 MyList(int i): last_(i) {}
 void PopBack();
 int last_;
};

jpthread.cc
------------

#include "jpthread.h"

void test  ()
{

  MyList *list = MyList::Create(20);

  for (int i = 0; i < 2; ) {
    MyList* new_list = list->PushBack(
        list->LastHit() + 10);
    if (new_list != list) {
      list->Destroy();
      list = new_list;
      ++i;
    }
  }

  list->Destroy();
}
int main()
{
 test();
}


m.cc
-----
#include <stdlib.h>
#include <stdio.h>
#include "jpthread.h"


int cnt = 0;
MyList *prev = 0;

MyList* MyList::PushBack(int i)
{
  cnt ++;

  if (cnt == 1)
     return prev;

  if (cnt > 3) {
     fprintf (stderr, "Loop not ending, aborting ..\n");
     abort ();
  }

  prev = this;
  return new MyList(i+last_);
}

MyList* MyList::Create(int i)
{
  prev = new MyList(i);

  return prev;
}

void MyList::Destroy() {
   delete this;
}

[Bug tree-optimization/61289] Bad jump threading generates infinite loop

[ppluzhnikov at google dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61289

Paul Pluzhnikov <ppluzhnikov at google dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |ppluzhnikov at google dot com

--- Comment #1 from Paul Pluzhnikov <ppluzhnikov at google dot com> ---
Google ref: b/15146357

[Bug tree-optimization/61289] [4.9/4.10 Regression] Bad jump threading generates infinite loop

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61289

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2014-05-23
                 CC|                            |jakub at gcc dot gnu.org
            Version|unknown                     |4.9.0
   Target Milestone|---                         |4.9.1
            Summary|Bad jump threading          |[4.9/4.10 Regression] Bad
                   |generates infinite loop     |jump threading generates
                   |                            |infinite loop
     Ever confirmed|0                           |1

--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
struct S
{
  inline int fn1 () const { return s; }
  __attribute__ ((noinline, noclone)) S *fn2 (int);
  __attribute__ ((noinline, noclone)) void fn3 ();
  __attribute__ ((noinline, noclone)) static S *fn4 (int);
  S (int i) : s (i) {}
  int s;
};

int a = 0;
S *b = 0;

S *
S::fn2 (int i)
{
  a++;
  if (a == 1)
    return b;
  if (a > 3)
    __builtin_abort ();
  b = this;
  return new S (i + s);
}

S *
S::fn4 (int i)
{
  b = new S (i);
  return b;
}

void
S::fn3 ()
{
  delete this;
}

void
foo ()
{
  S *c = S::fn4 (20);
  for (int i = 0; i < 2;)
    {
      S *d = c->fn2 (c->fn1 () + 10);
      if (d != c)
{
  c->fn3 ();
  c = d;
  ++i;
}
    }
  c->fn3 ();
}

int
main ()
{
  foo ();
}

Started with r205279.

[Bug tree-optimization/61289] [4.9/4.10 Regression] Bad jump threading generates infinite loop

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61289

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |wrong-code
           Priority|P3                          |P2

[Bug tree-optimization/61289] [4.9/4.10 Regression] Bad jump threading generates infinite loop

[law at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61289

--- Comment #3 from Jeffrey A. Law <law at gcc dot gnu.org> ---
Author: law
Date: Thu Jun  5 18:25:02 2014
New Revision: 211287

URL: http://gcc.gnu.org/viewcvs?rev=211287&root=gcc&view=rev
Log:
    PR tree-optimization/61289
    * tree-ssa-threadedge.c (invalidate_equivalences): Remove SRC_MAP and
    DST_MAP parameters.   Invalidate by walking all the SSA_NAME_VALUES
    looking for those which match LHS.  All callers changed.
    (record_temporary_equivalences_from_phis): Remove SRC_MAP and DST_MAP
    parameters and code which manipulated them.  All callers changed.
    (record_temporary_equivalences_from_stmts_at_dest): Remove SRC_MAP
    and DST_MAP parameters.  Simplify invalidation code by just calling
    invalidate_equivalences.  All callers changed.
    (thread_across_edge): Simplify now that we don't need to maintain
    the map of equivalences to invalidate.

        PR tree-optimization/61289
    * g++.dg/pr61289.C: New test.
    * g++.dg/pr61289-2.C: New test.

Added:
    trunk/gcc/testsuite/g++.dg/pr61289-2.c
    trunk/gcc/testsuite/g++.dg/pr61289.C
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/tree-ssa-threadedge.c

[Bug tree-optimization/61289] [4.9 Regression] Bad jump threading generates infinite loop

[ppluzhnikov at google dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61289

--- Comment #4 from Paul Pluzhnikov <ppluzhnikov at google dot com> ---
Back-port to gcc-4_9-branch?

Thanks,

[Bug tree-optimization/61289] [4.9 Regression] Bad jump threading generates infinite loop

[law at redhat dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61289

--- Comment #5 from Jeffrey A. Law <law at redhat dot com> ---
yes, I'll backport to the 4.9 branch.  Waiting to see if there's any
follow-ups.

[Bug tree-optimization/61289] [4.9 Regression] Bad jump threading generates infinite loop

[law at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61289

--- Comment #7 from Jeffrey A. Law <law at gcc dot gnu.org> ---
Author: law
Date: Thu Jun 12 18:38:20 2014
New Revision: 211588

URL: http://gcc.gnu.org/viewcvs?rev=211588&root=gcc&view=rev
Log:
    Backports from mainline:

    PR tree-optimization/61289
    * tree-ssa-threadedge.c (invalidate_equivalences): Remove SRC_MAP and
    DST_MAP parameters.   Invalidate by walking all the SSA_NAME_VALUES
    looking for those which match LHS.  All callers changed.
    (record_temporary_equivalences_from_phis): Remove SRC_MAP and DST_MAP
    parameters and code which manipulated them.  All callers changed.
    (record_temporary_equivalences_from_stmts_at_dest): Remove SRC_MAP
    and DST_MAP parameters.  Simplify invalidation code by just calling
    invalidate_equivalences.  All callers changed.
    (thread_across_edge): Simplify now that we don't need to maintain
    the map of equivalences to invalidate.

        PR tree-optimization/61289
    * g++.dg/pr61289.C: New test.
    * g++.dg/pr61289-2.C: New test.

Added:
    branches/gcc-4_9-branch/gcc/testsuite/g++.dg/pr61289-2.c
    branches/gcc-4_9-branch/gcc/testsuite/g++.dg/pr61289.C
Modified:
    branches/gcc-4_9-branch/gcc/ChangeLog
    branches/gcc-4_9-branch/gcc/testsuite/ChangeLog
    branches/gcc-4_9-branch/gcc/tree-ssa-threadedge.c

[Bug tree-optimization/61289] [4.9 Regression] Bad jump threading generates infinite loop

[law at redhat dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61289

Jeffrey A. Law <law at redhat dot com> changed:

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

--- Comment #6 from Jeffrey A. Law <law at redhat dot com> ---
Backported to 4.9 branch as well.