[Bug c++/18744] New: C++ ABI is incomplete for ILP64

[Bug c++/18744] New: C++ ABI is incomplete for ILP64

[markus at oberhumer dot com]

The C++ ABI is missing fallbacks for exact 16- and 32-bit types when
using programming models like ILP64 (e.g. MIPS -mint64 option below):


/opt/cross/mips-linux-uclibc-gcc341/bin/mips-linux-uclibc-g++ -c -mint64 bug01.cpp
bug01.cpp:19: internal compiler error: in write_builtin_type, at cp/mangle.c:1686
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://gcc.gnu.org/bugs.html> for instructions.


typedef int int16_exact_t __attribute__((__mode__(__HI__)));
typedef int int32_exact_t __attribute__((__mode__(__SI__)));

long getbit(int16_exact_t v) { return v & 1; }
long getbit(int32_exact_t v) { return v & 1; }

-- 
           Summary: C++ ABI is incomplete for ILP64
           Product: gcc
           Version: 3.4.3
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: c++
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: markus at oberhumer dot com
                CC: gcc-bugs at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18744

[Bug c++/18744] C++ ABI is incomplete for ILP64

[markus at oberhumer dot com]

------- Additional Comments From markus at oberhumer dot com  2004-11-30 20:39 -------
Created an attachment (id=7642)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=7642&action=view)
Testcase to reproduce the ICE.

Testcase to reproduce the ICE.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18744

[Bug c++/18744] C++ ABI is incomplete for ILP64

[markus at oberhumer dot com]

------- Additional Comments From markus at oberhumer dot com  2004-11-30 20:42 -------
Created an attachment (id=7643)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=7643&action=view)
Quick and dirty patch for gcc/cp/mangle.c .

This is just a quick patch, but I think this issue should be addressed in the
C++ ABI.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18744

[Bug c++/18744] C++ ABI is incomplete for ILP64

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-11-30 20:46 -------
Note we reject this code right now on the mainline, see PR 18442.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
  BugsThisDependsOn|                            |18442
           Keywords|                            |ABI, ice-on-valid-code


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18744

[Bug c++/18744] C++ ABI is incomplete for ILP64

[markus at oberhumer dot com]

------- Additional Comments From markus at oberhumer dot com  2004-11-30 20:56 -------
I'm not sure if this is really related to bug18442 - actually I've ported
current mainline to generate ILP64 code on AMD64, and __mode__(__SI__) works
without problems there.

But it's the incomplete C++ ABI that is a showstopper for C++.



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18744

[Bug c++/18744] C++ ABI is incomplete for ILP64

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-11-30 20:59 -------
using the mainline with MIPS and -mint64 cause the code to be rejected which is what I was trying to 
say.  For the x86_64 problem you might want to write an email to gcc@ asking about the ABI, they will 
ask you write to the IA64 C++ ABI emailing list most likely.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18744

[Bug c++/18744] C++ ABI is incomplete for ILP64

[echristo at redhat dot com]

-- 
Bug 18744 depends on bug 18442, which changed state.

Bug 18442 Summary: [4.0 Regression] Rejects attribute((mode(SI))) when using -mint64
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18442

           What    |Old Value                   |New Value
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18744

[Bug c++/18744] C++ ABI is incomplete for ILP64

[markus at oberhumer dot com]

------- Additional Comments From markus at oberhumer dot com  2005-01-21 02:14 -------
Just a short update that the MIPS -mint64 option works again in mainline, so the
ICE reported here is back for both the 3.4 and 4.0 branches.



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18744

[Bug c++/18744] C++ ABI is incomplete for ILP64

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-04-22 05:13 -------
Confirmed.

Waiting on copy right assignment:
<http://gcc.gnu.org/ml/gcc-patches/2005-04/msg01886.html>.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|                            |1
   Last reconfirmed|0000-00-00 00:00:00         |2005-04-22 05:13:50
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18744

[Bug c++/18744] C++ ABI is incomplete for ILP64

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-04-24 17:38 -------
Fixed by:
2005-04-12  Markus F.X.J. Oberhumer  <markus@oberhumer.com>

        * mangle.c (write_builtin_type): Handle integer types which are
        not one of the shared integer type nodes and emit a "vendor
        extended builtin type" with an encoding in the form of "u5int96".

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED
   Target Milestone|---                         |4.1.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18744