[Bug other/52443] New: ICE: verify_gimple failed: invalid types in nop conversion

[Bug other/52443] New: ICE: verify_gimple failed: invalid types in nop conversion

[gjl at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52443

             Bug #: 52443
           Summary: ICE: verify_gimple failed: invalid types in nop
                    conversion
    Classification: Unclassified
           Product: gcc
           Version: 4.7.0
            Status: UNCONFIRMED
          Keywords: addr-space, ice-on-valid-code
          Severity: normal
          Priority: P3
         Component: other
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: gjl@gcc.gnu.org
            Target: avr


extern void bar (const __memx void*);

void func (char a)
{
    bar (a);
}

const __memx void* func2 (char a)
{
    return a;
}

>> avr-gcc ice-memx.c -S -w

ice-memx.c: In function 'func2':
ice-memx.c:8:20: error: invalid types in nop conversion
const <address-space-7> void *
long int
D.1326 = (const <address-space-7> void *) D.1327;

ice-memx.c:8:20: internal compiler error: verify_gimple failed
Please submit a full bug report,
with preprocessed source if appropriate.


Target: avr
Configured with: ../gcc-trunk/configure --with-gmp=/usr/local
--prefix=/usr/local/avr --target=avr --enable-languages=c,c++ --with-dwarf2
--disable-shared --disable-libada --disable-libssp --disable-nls
Thread model: single
gcc version 4.7.0 20120217 (experimental) (GCC) 
GNU C (GCC) version 4.7.0 20120217 (experimental) (avr)
    compiled by GNU C version 4.6.2, GMP version 5.0.4, MPFR version 3.1.0, MPC
version 0.9


I can also see the error on a build from 2012-02-28.

[Bug other/52443] ICE: verify_gimple failed: invalid types in nop conversion

[gjl at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52443

--- Comment #1 from Georg-Johann Lay <gjl at gcc dot gnu.org> 2012-02-29 22:11:16 UTC ---
Created attachment 26793
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=26793
ice-memx.c: C source

[Bug other/52443] ICE: verify_gimple failed: invalid types in nop conversion

[gjl at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52443

Georg-Johann Lay <gjl at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2012-02-29
                 CC|                            |eric.weddington at atmel
                   |                            |dot com
     Ever Confirmed|0                           |1

[Bug other/52443] ICE: verify_gimple failed: invalid types in nop conversion

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52443

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

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

--- Comment #2 from Richard Guenther <rguenth at gcc dot gnu.org> 2012-03-01 09:50:40 UTC ---
        /* Allow conversions between integral types and pointers only if
           there is no sign or zero extension involved.
           For targets were the precision of ptrofftype doesn't match that
           of pointers we need to allow arbitrary conversions from and
           to ptrofftype.  */
        if ((POINTER_TYPE_P (lhs_type)
             && INTEGRAL_TYPE_P (rhs1_type)
             && (TYPE_PRECISION (lhs_type) >= TYPE_PRECISION (rhs1_type)
                 || ptrofftype_p (rhs1_type)))
            || (POINTER_TYPE_P (rhs1_type)
                && INTEGRAL_TYPE_P (lhs_type)
                && (TYPE_PRECISION (rhs1_type) >= TYPE_PRECISION (lhs_type)
                    || ptrofftype_p (sizetype))))
          return false;

Looks like the above should be

        if ((POINTER_TYPE_P (lhs_type)
             && INTEGRAL_TYPE_P (rhs1_type)
             && (TYPE_PRECISION (lhs_type) <= TYPE_PRECISION (rhs1_type)
                 || ptrofftype_p (rhs1_type)))
            || (POINTER_TYPE_P (rhs1_type)
                && INTEGRAL_TYPE_P (lhs_type)
                && (TYPE_PRECISION (rhs1_type) >= TYPE_PRECISION (lhs_type)
                    || ptrofftype_p (sizetype))))
          return false;

does that help?

[Bug other/52443] ICE: verify_gimple failed: invalid types in nop conversion

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52443

--- Comment #3 from Richard Guenther <rguenth at gcc dot gnu.org> 2012-03-01 10:20:20 UTC ---
Or rather

        /* Allow conversions pointer type to integral type only if
           there is no sign or zero extension involved.
           For targets were the precision of ptrofftype doesn't match that
           of pointers we need to allow arbitrary conversions to ptrofftype. 
*/
        if (POINTER_TYPE_P (rhs1_type)
            && INTEGRAL_TYPE_P (lhs_type)
            && (TYPE_PRECISION (rhs1_type) >= TYPE_PRECISION (lhs_type)
                || ptrofftype_p (sizetype)))
          return false;

as restricting conversion from integral to pointers does not seem necessary
(only pointers do not have a "sign").

[Bug other/52443] ICE: verify_gimple failed: invalid types in nop conversion

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52443

--- Comment #4 from Richard Guenther <rguenth at gcc dot gnu.org> 2012-03-01 10:23:34 UTC ---
Err ...

        /* Allow conversions pointer type to integral type only if
           there is no sign or zero extension involved.
           For targets were the precision of ptrofftype doesn't match that
           of pointers we need to allow arbitrary conversions to ptrofftype. 
*/
        if ((POINTER_TYPE_P (lhs_type)
             && INTEGRAL_TYPE_P (rhs1_type))
            || (POINTER_TYPE_P (rhs1_type)
                && INTEGRAL_TYPE_P (lhs_type)
                && (TYPE_PRECISION (rhs1_type) >= TYPE_PRECISION (lhs_type)
                    || ptrofftype_p (sizetype))))
          return false;

tricky business ;)

[Bug other/52443] ICE: verify_gimple failed: invalid types in nop conversion

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52443

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 #5 from Richard Guenther <rguenth at gcc dot gnu.org> 2012-03-01 10:24:39 UTC ---
Mine, in case that helps (it's in tree-cfg.c)

[Bug other/52443] ICE: verify_gimple failed: invalid types in nop conversion

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52443

--- Comment #6 from Richard Guenther <rguenth at gcc dot gnu.org> 2012-03-01 12:26:25 UTC ---
Author: rguenth
Date: Thu Mar  1 12:26:20 2012
New Revision: 184739

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=184739
Log:
2012-03-01  Richard Guenther  <rguenther@suse.de>

    PR middle-end/52443
    * tree-cfg.c (verify_gimple_assign_unary): Allow any
    conversions from integral types to pointer types.

Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/tree-cfg.c

[Bug other/52443] ICE: verify_gimple failed: invalid types in nop conversion

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52443

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

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

--- Comment #7 from Richard Guenther <rguenth at gcc dot gnu.org> 2012-03-01 12:27:43 UTC ---
Fixed.

[Bug other/52443] ICE: verify_gimple failed: invalid types in nop conversion

[gjl at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52443

--- Comment #8 from Georg-Johann Lay <gjl at gcc dot gnu.org> 2012-03-01 13:36:08 UTC ---
Yes, thanks. It works now :-)