[Bug c++/13983] New: packed attribute does not take effect when using typedef

[Bug c++/13983] New: packed attribute does not take effect when using typedef

[bkramer at ciprico dot com]

Reading specs from
/devel/new_compiler/engsrc/tools/linux/lib/gcc-lib/powerpc-wrs-vxworks/3.4.0/specs
Configured with: ../../src/gcc/configure --target=powerpc-wrs-vxworks
--prefix=/devel/gcc-vxworks --enable-languages=c,c++ --enable-thread=vxworks
--with-headers=/devel/Tornado/target/h --with-gnu-ld --with-gnu-as
--enable-altivec --enable-long-long
Thread model: vxworks
gcc version 3.4.0 20040130 (prerelease)
 /devel/new_compiler/engsrc/tools/linux/bin/../libexec/gcc/powerpc-wrs-vxworks/3.4.0/cc1plus -E -quiet -nostdinc -v -I/devel/new_compiler/engsrc/tools/Tornado/target/config/all -I/devel/new_compiler/engsrc/tools/Tornado/target/src/config -I/devel/new_compiler/engsrc/tools/Tornado/target/src/drv -I/devel/new_compiler/engsrc/tools/Tornado/target/h -I/devel/new_compiler/engsrc/tools/Tornado/target/h/c++/3.4.0-O2 -iprefix /devel/new_compiler/engsrc/tools/linux/bin/../lib/gcc/powerpc-wrs-vxworks/3.4.0/ -isystem /devel/new_compiler/engsrc/tools/linux/lib/gcc-lib/include -isystem /devel/new_compiler/engsrc/tools/linux/powerpc-wrs-vxworks/bin/include -DCPU_FAMILY=PPC -D__ppc -D__EABI__ -DCPU=PPC603 -D__hardfp -DVXWORKS -DROM_SIZE=0x00080000 -DRAM_LOW_ADRS=0x00010000 -DRAM_HIGH_ADRS=0x01000000 Test.cpp -mcpu=603 -mstrict-align -mstrict-align -Wall -fno-exceptions -fno-builtin -fno-exceptions -fworking-directory -o Test.ii
ignoring nonexistent directory
"/devel/new_compiler/engsrc/tools/linux/lib/gcc-lib/include"
ignoring nonexistent directory
"/devel/new_compiler/engsrc/tools/linux/powerpc-wrs-vxworks/bin/include"
ignoring nonexistent directory
"/devel/new_compiler/engsrc/tools/Tornado/target/h/c++/3.4.0-O2"
#include "..." search starts here:
#include <...> search starts here:
 /devel/new_compiler/engsrc/tools/Tornado/target/config/all
 /devel/new_compiler/engsrc/tools/Tornado/target/src/config
 /devel/new_compiler/engsrc/tools/Tornado/target/src/drv
 /devel/new_compiler/engsrc/tools/Tornado/target/h
End of search list.
 /devel/new_compiler/engsrc/tools/linux/bin/../libexec/gcc/powerpc-wrs-vxworks/3.4.0/cc1plus -fpreprocessed Test.ii -mcpu=603 -mstrict-align -quiet -dumpbase Test.cpp -mstrict-align -auxbase Test -g -gdwarf-2 -Wall -version -fno-exceptions -fno-builtin -fno-exceptions -o Test.s
GNU C++ version 3.4.0 20040130 (prerelease) (powerpc-wrs-vxworks)
        compiled by GNU C version 3.2.2 20030222 (Red Hat Linux 3.2.2-5).
GGC heuristics: --param ggc-min-expand=63 --param ggc-min-heapsize=63140
Test.cpp: In function `void test()':
Test.cpp:26: warning: ignoring packed attribute on unpacked non-POD field
`TempStructType test()::<anonymous struct>::tmp3'
 /devel/new_compiler/engsrc/tools/linux/powerpc-wrs-vxworks/bin/as -mppc -V -Qy
-mbig -o Test.o Test.s
GNU assembler version 2.13.1 (powerpc-wrs-vxworks) using BFD version 2.13.1

-- 
           Summary: packed attribute does not take effect when using typedef
           Product: gcc
           Version: 3.4.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: c++
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: bkramer at ciprico dot com
                CC: gcc-bugs at gcc dot gnu dot org
 GCC build triplet: i686-pc-linux-gnu
  GCC host triplet: i686-pc-linux-gnu
GCC target triplet: powerpc-wrs-vxworks


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13983

[Bug c++/13983] packed attribute does not take effect when using typedef

[bkramer at ciprico dot com]

------- Additional Comments From bkramer at ciprico dot com  2004-02-02 16:56 -------
Created an attachment (id=5651)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=5651&action=view)
Preprocessed file


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13983

[Bug c++/13983] no warning on some non-POD struct with packed attribute

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-02-02 17:02 -------
Actually the warning is correct, the problem is that the warning is not on the other structs for some 
reason when it should be as non-POD's ignore the packed attribute.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|normal                      |enhancement
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|                            |1
           Keywords|                            |diagnostic
   Last reconfirmed|0000-00-00 00:00:00         |2004-02-02 17:02:03
               date|                            |
            Summary|packed attribute does not   |no warning on some non-POD
                   |take effect when using      |struct with packed attribute
                   |typedef                     |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13983

[Bug c++/13983] no warning on some non-POD struct with packed attribute

[mdorey at bluearc dot com]

------- Additional Comments From mdorey at bluearc dot com  2004-03-27 00:08 -------
(In reply to comment #2)
> non-POD's ignore the packed attribute.

We have a packed non-POD which abstracts byte-ordering which we use in lots of
packed structs.  Although the language spec doesn't talk about packing, being
able to do this efficiently is one of the key benefits of C++ for us.  Is it
really the case that *all* non-PODs ignore the packed attribute?  Nathan seemed
to contradict that in gcc/testsuite/g++.dg/ext/packed2.C, where he wrote:

  // July 2003
  // packing of non-pods is now only allowed if the non-pod is itself
  // packed. Also only such pods can be reference bound to non-consts

Which would let us do what we want to do just fine.  To the same end, in the
original(?) PR for packed2.C, Jason said:

    The problem is that the #pragma pack means that _GEIdx::addr
    might not have the usual alignment of an ftn_addr, so we
    can't just take the address of current.addr in order to bind
    it to a reference.  expand_expr knows this, and tries to
    make a bitwise copy and take the address of that.  But since
    ftn_addr has a copy constructor, we can't make bitwise
    copies, so we abort.

This is because "ftn_addr" isn't packed but it's being used to define a field in
a packed structure.  If they're both packed, then the alignment of "ftn_addr" is
always the same (1) and the need to copy won't arise.

If that's agreed, then can we go back to saying that this is a problem with
typedef?  This code compiles OK with 3.4.0 20040324:

template <class T>
struct A {
  A();
} __attribute__ ((packed));

//typedef A<int> Ai;

struct B {
  A<int> a;
} __attribute__ ((packed));

typedef A<int> Ai;

struct C {
  Ai a;
} __attribute__ ((packed));

But uncommenting the first typedef for Ai prevents it from compiling.

Perhaps it's to do with when the typedef is first seen?  If it's first seen
after the non-POD has already been used in a packed structure, the typedef seems
to be packed.  If before, then unpacked - even though the non-POD it's
typedef()d to is packed.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13983

[Bug c++/13983] no warning on some non-POD struct with packed attribute

[mdorey at bluearc dot com]

------- Additional Comments From mdorey at bluearc dot com  2004-03-29 21:26 -------
At the risk of annoying particularly Andrew Pinski, who does such a sterling job
with triaging all my idiot bugs, I wonder if I can set the severity back to
normal, as I'm claiming - in comment #3 - that this isn't a missing warning but
an incorrect warning, as was originally submitted?

...

Darn, it seems I can't.  Perhaps this bug is doomed to languish unread.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13983

[Bug c++/13983] no warning on some non-POD struct with packed attribute

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-03-29 22:16 -------
Nathan could you comment on this bug.

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13983

[Bug c++/13983] no warning on some non-POD struct with packed attribute

[mdorey at bluearc dot com]

------- Additional Comments From mdorey at bluearc dot com  2005-04-26 19:53 -------
(In reply to comment #5)
> Nathan could you comment on this bug.

This bug may be related to http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17519

There's an interesting discussion in that bug, including comments from Nathan.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13983

[Bug c++/13983] no warning on some non-POD struct with packed attribute

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-09-18 02:06 -------
*** Bug 23833 has been marked as a duplicate of this bug. ***

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |Dmitry dot Chepel at acronis
                   |                            |dot com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13983