[Bug middle-end/44608] New: unspecified signed overflow applied to unsigned int

[Bug middle-end/44608] New: unspecified signed overflow applied to unsigned int

[avr at gjlay dot de]

As discussed in 
http://gcc.gnu.org/ml/gcc-help/2010-06/msg00191.html
I am filing a PR for the following piece of C:

int abssat2 (int x)
{
    unsigned int y = x;

    if (x < 0)
        y = -y;

    if (y >= 0x80000000)
        y--;

    return y;
}

Code looks fine until pass CE1 which introduces ABS rtx for the first
comparison (ifcvt.c:noce_try_abs). 

Pass insn combine then calls combine.c:combine_simplify_rtx() which calls
simplify_rtx.c:simplify_relational_operation (code=LT, mode=SImode,
cmp_mode=SImode, op0=(abs:SI (reg:SI)), op1=(const_int 0))
and the latter removes the second comparison against 0x80000000.

As this problem occurs on a non-standard target, let me know if I con you
provide with any additional information.


-- 
           Summary: unspecified signed overflow applied to unsigned int
           Product: gcc
           Version: 4.4.3
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: middle-end
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: avr at gjlay dot de
 GCC build triplet: pc-linux
  GCC host triplet: pc-linux
GCC target triplet: private


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44608

[Bug middle-end/44608] unspecified signed overflow applied to unsigned int

[avr at gjlay dot de]

------- Comment #1 from avr at gjlay dot de  2010-06-21 11:26 -------
Created an attachment (id=20955)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=20955&action=view)
Dump for .c.123t.optimized

Dump for .c.123t.optimized


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44608

[Bug middle-end/44608] unspecified signed overflow applied to unsigned int

[avr at gjlay dot de]

------- Comment #2 from avr at gjlay dot de  2010-06-21 11:28 -------
Created an attachment (id=20956)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=20956&action=view)
Dump for .c.128r.expand

Dump for .c.128r.expand


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44608

[Bug middle-end/44608] unspecified signed overflow applied to unsigned int

[avr at gjlay dot de]

------- Comment #3 from avr at gjlay dot de  2010-06-21 11:28 -------
Created an attachment (id=20957)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=20957&action=view)
Dump for .c.141r.ce1

Dump for .c.141r.ce1


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44608

[Bug middle-end/44608] unspecified signed overflow applied to unsigned int

[avr at gjlay dot de]

------- Comment #4 from avr at gjlay dot de  2010-06-21 11:30 -------
Created an attachment (id=20958)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=20958&action=view)
Dump for .c.159r.combine

Dump for .c.159r.combine


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44608

[Bug middle-end/44608] unspecified signed overflow applied to unsigned int

[rguenth at gcc dot gnu dot org]

------- Comment #5 from rguenth at gcc dot gnu dot org  2010-06-21 11:38 -------
It is folding from the frontend that changes

    if (y >= 0x80000000)

to

    if ((int) y < 0)

(see code == LT instead of code == GEU)

But the main issue is that y = -y to abs is bogus (but we can't easily tell
that as on RTL operands do not have a signedness).


-- 

rguenth at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|0                           |1
           Keywords|                            |wrong-code
   Last reconfirmed|0000-00-00 00:00:00         |2010-06-21 11:38:28
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44608

[Bug middle-end/44608] unspecified signed overflow applied to unsigned int

[rguenth at gcc dot gnu dot org]

------- Comment #6 from rguenth at gcc dot gnu dot org  2010-06-21 11:50 -------
int __attribute__((noinline))
abssat2 (int x)
{
  unsigned int y = x;
  if (x < 0)
    y = -y;
  if (y >= 0x80000000)
    y--;
  return y;
}
extern void abort (void);
int main()
{
  if (abssat2 (0x80000000) != 0x7fffffff)
    abort ();
  return 0;
}

does not fail on x86_64 or i?86.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44608

[Bug middle-end/44608] unspecified signed overflow applied to unsigned int

[avr at gjlay dot de]

------- Comment #7 from avr at gjlay dot de  2010-06-21 12:27 -------
(In reply to comment #5)
> It is folding from the frontend that changes
> 
>     if (y >= 0x80000000)
> 
> to
> 
>     if ((int) y < 0)
> 
> (see code == LT instead of code == GEU)
> 
> But the main issue is that y = -y to abs is bogus (but we can't easily tell
> that as on RTL operands do not have a signedness).

Both transformation are ok from the target's point of view. The bad thing is
that undefinedness is propagated to RTL which is supposed to be an algebraic
formulation of the target's instruction set. But in this case RTL is treated
like a low level representation of C (or any other language gcc supports).

To reproduce the problem it seems essential that there is an abssi2 insn in the
md. 


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44608

[Bug middle-end/44608] unspecified signed overflow applied to unsigned int

[pinskia at gcc dot gnu dot org]

------- Comment #8 from pinskia at gcc dot gnu dot org  2010-07-21 22:51 -------
For some reason rs6000 has an abssi pattern but it is not produced by ce1.


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
  GCC build triplet|pc-linux                    |
   GCC host triplet|pc-linux                    |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44608