[Bug c++/19448] New: Different value representation for bitfield width exceeding its type size.

[Bug c++/19448] New: Different value representation for bitfield width exceeding its type size.

[yanliu at ca dot ibm dot com]

The gcc compiler information: 
Reading specs from /usr/lib/gcc/ppc64-redhat-linux/3.4.3/specs
Configured with: ../configure --prefix=/usr --mandir=/usr/share/man --
infodir=/usr/share/info --enable-shared --enable-threads=posix --disable-
checking --with-system-zlib --enable-__cxa_atexit --disable-libunwind-
exceptions --enable-languages=c,c++,objc,java,f77 --enable-java-awt=gtk --
host=ppc64-redhat-linux --build=ppc64-redhat-linux --target=ppc64-redhat-linux -
-with-cpu=default32
Thread model: posix
gcc version 3.4.3 20041125 (Red Hat 3.4.3-6.EL4)

In the attached testcase, the bitfield width is bigger than its type size, and 
the bitfield is initialized with a value greater than 127.  

When c1.m1 is assigned with different value, the result is different using 
g++3.4.3 and g++ 4.0.0 compilers:

c1.m1=128, g++ 3.4.3's result: 128, g++ 4.0.0's result: -128.
c1.m1=400, g++ 3.4.3's result: 144, g++ 4.0.0's result: -112.
c1.m1=500, g++ 3.4.3's result: 244, g++ 4.0.0's result: -12.

Could you tell me which compiler version gives the correct result, and why?

-- 
           Summary: Different value representation for bitfield width
                    exceeding its type size.
           Product: gcc
           Version: 3.4.3
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: c++
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: yanliu at ca dot ibm dot com
                CC: gcc-bugs at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19448

[Bug c++/19448] Different value representation for bitfield width exceeding its type size.

[yanliu at ca dot ibm dot com]

------- Additional Comments From yanliu at ca dot ibm dot com  2005-01-14 18:21 -------
Created an attachment (id=7960)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=7960&action=view)
testcase


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19448

[Bug c++/19448] Different value representation for bitfield width exceeding its type size.

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-01-14 19:01 -------
Hmm, are you sure that you are not comparing against gcc for ppc-darwin and gcc for ppc-linux 
because the default sign-ness of char is different between them.

On ppc-darwin, the char is signed by default and for ppc-linux it is unsigned.  


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19448

[Bug c++/19448] Different value representation for bitfield width exceeding its type size.

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-01-14 19:06 -------
I will just have a note here from what the standard say about this issue:
" in such cases the extra bits are used for padding and do not participate in the value representation of 
the bit-field" (9.6P1)

"It is implmentation-defined whether a plain (neitther explicitly signed nor unsigned) char, short, int or 
long bit-field is signed or unsigned."

So both are right if we changed the definedness (but I still think you are comparing ppc-darwin vs ppc-
linux).

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19448

[Bug c++/19448] Different value representation for bitfield width exceeding its type size.

[yanliu at ca dot ibm dot com]

------- Additional Comments From yanliu at ca dot ibm dot com  2005-01-14 20:10 -------
You are right, I was using linux and darwin two different compilers. In thse 
two situations, the gcc results are both right. 

I have another question for you for the same testcase: if initializing the 
bitfield with 1, and the bitwidth=17.  I got two different layouts for the 
bitfield member with different g++ compilers:
Both compilers are ppc-rhat. One is 3.4.3, the other is 3.2.3.

Here is the first compiler's info:
Reading specs from /usr/lib/gcc/ppc64-redhat-linux/3.4.3/specs
Configured with: ../configure --prefix=/usr --mandir=/usr/share/man --
infodir=/usr/share/info --enable-shared --enable-threads=posix --disable-
checking --with-system-zlib --enable-__cxa_atexit --disable-libunwind-
exceptions --enable-languages=c,c++,objc,java,f77 --enable-java-awt=gtk --
host=ppc64-redhat-linux --build=ppc64-redhat-linux --target=ppc64-redhat-linux -
-with-cpu=default32
Thread model: posix
gcc version 3.4.3 20041125 (Red Hat 3.4.3-6.EL4)

Running the testcase, here is what I got:
with g++3.2.3, the result is:
Dumping array with size of 4
Character 0 is 0
Character 1 is 1
Character 2 is 0
Character 3 is 0
 c1.m1=1,

With 3.4.3 g++ compiler, the result is:
Dumping array with size of 4
Character 0 is 1
Character 1 is 0
Character 2 is 0
Character 3 is 0
 c1.m1=1,

The Itanium C++ ABI didn't specify how to layout bitfield member if their width 
exceeds its type. Did you put your least-important bit at the left hand side? 
Could you tell me your algorithm on the bitfield member layout? Thanks. 



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19448

[Bug c++/19448] Different value representation for bitfield width exceeding its type size.

[janis187 at us dot ibm dot com]

------- Additional Comments From janis187 at us dot ibm dot com  2005-01-15 01:24 -------
Created an attachment (id=7963)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=7963&action=view)
example binary compatibility testcase


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19448

[Bug c++/19448] Different value representation for bitfield width exceeding its type size.

[janis187 at us dot ibm dot com]

------- Additional Comments From janis187 at us dot ibm dot com  2005-01-15 01:27 -------
Darn, I did my comments wrong.

Yan, the testcase you attached doesn't match the output you show.  For all
compilers I have available on a RHEL3 system I get:

Dumping array with size of 2
Character 0 is 1
Character 1 is 0
 c1.m1=1,

Please provide some examples of classes with bitfields that are not compatible
between releases of GCC.  I've attached an example testcase for binary
compatibility tests in the format used by the GCC testsuite.  Our binary
compatibility tests are not at all complete and we need more tests to cover
this kind of functionality; that's on my todo list already.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19448

[Bug c++/19448] Different value representation for bitfield width exceeding its type size.

[janis187 at us dot ibm dot com]

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19448

[Bug c++/19448] Different value representation for bitfield width exceeding its type size.

[yanliu at ca dot ibm dot com]

------- Additional Comments From yanliu at ca dot ibm dot com  2005-01-17 15:20 -------
Did you change the bitfield width to 17 in the testcase? 


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19448

[Bug c++/19448] Different value representation for bitfield width exceeding its type size.

[janis187 at us dot ibm dot com]

------- Additional Comments From janis187 at us dot ibm dot com  2005-01-17 18:14 -------
I had not changed the size of the bitfield to 17 in my test case.  When I do
that I can see the binary compatibility breakage.

I'll look into this to find out why this change was introduced.  It appears to
be a general C++ change not specific to powerpc-linux.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19448

[Bug c++/19448] Different value representation for bitfield width exceeding its type size.

[janis187 at us dot ibm dot com]

------- Additional Comments From janis187 at us dot ibm dot com  2005-01-19 02:01 -------
There are two changes that affect this binary incompatibility: one
changes the layout of the class, and the other changes how the bitfield
is accessed.  Mark, did this change on purpose and is it covered by the
C++ ABI?
                                                                                
The first change is:
                                                                                
--- gcc/gcc/cp/ChangeLog ---
                                                                                
2003-04-29  Mark Mitchell  <mark@codesourcery.com>
                                                                                
        PR c++/10549
        * class.c (layout_class_type): Mark overlong bitfields as having
        the maximum size permitted by their type, after layout.
                                                                                
The second change is:
                                                                                
--- gcc/gcc/ChangeLog ---
                                                                                
2003-12-23  Mark Mitchell  <mark@codesourcery.com>
                                                                                
        * c-common.c (flag_abi_version): Default to 2.
        * c-cppbuiltin.c (c_cpp_builtins): Define __GXX_ABI_VERSION
        uniformly for versions above 2.
        * doc/invoke.texi: Update documentation for -fabi-version.
                                                                                
--- gcc/gcc/cp/ChangeLog ---
                                                                                
2003-12-23  Mark Mitchell  <mark@codesourcery.com>
                                                                                
        * cp-lang.c (cp_expr_size): Return zero for empty classes.
                                                                                
        * cp-tree.h (warn_if_uknown_interface): Remove unused function.
        * decl2.c (warn_if_unknown_interface): Likewise.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |mark at codesourcery dot com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19448

[Bug c++/19448] Different value representation for bitfield width exceeding its type size.

[janis187 at us dot ibm dot com]

------- Additional Comments From janis187 at us dot ibm dot com  2005-01-19 02:05 -------
Created an attachment (id=7988)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=7988&action=view)
test case with script

I get the following output from running the attached script using compilers
I built before and after each of the patches listed in the previous comment.

-------------------------------------------------------------------
define with 3.2.3
use with    2003-04-29 21:28 UTC
									       

PASSED
									       

define with 2003-04-29 21:28 UTC
use with    3.2.3
									       

PASSED
-------------------------------------------------------------------
define with 2003-04-29 21:28 UTC
use with    2003-04-29 21:28 UTC
									       

PASSED
-------------------------------------------------------------------
define with 2003-04-29 21:28 UTC
use with    2003-04-29 21:30 UTC
									       

./pr19448-2.sh: line 116: 26808 Aborted 		a.out
									       

define with 2003-04-29 21:30 UTC
use with    2003-04-29 21:28 UTC
									       

PASSED
-------------------------------------------------------------------
define with 2003-04-29 21:30 UTC
use with    2003-04-29 21:30 UTC
									       

./pr19448-2.sh: line 118: 26832 Aborted 		a.out
-------------------------------------------------------------------
define with 2003-12-23 16:53 UTC
use with    2003-12-23 16:53 UTC
									       

./pr19448-2.sh: line 120: 26852 Aborted 		a.out
-------------------------------------------------------------------
define with 2003-12-23 16:53 UTC
use with    2003-12-23 16:55 UTC
									       

./pr19448-2.sh: line 122: 26872 Aborted 		a.out
									       

define with 2003-12-23 16:55 UTC
use with    2003-12-23 16:53 UTC
									       

PASSED
-------------------------------------------------------------------
define with 2003-12-23 16:55 UTC
use with    2003-12-23 16:55 UTC
									       

PASSED
-------------------------------------------------------------------
define with 2003-12-23 16:55 UTC
use with    3.4.3
									       

PASSED
									       

define with 3.4.3
use with    2003-12-23 16:55 UTC
									       

PASSED
-------------------------------------------------------------------
define with 3.2.3
use with    3.4.3
									       

./pr19448-2.sh: line 128: 26940 Aborted 		a.out
									       

define with 3.4.3
use with    3.2.3
									       

./pr19448-2.sh: line 128: 26944 Aborted 		a.out
-------------------------------------------------------------------
In file included from pr19448_define.C:1:
bc.h:3: warning: width of `bc::m1' exceeds its type
bc.h:1: warning: the offset of `bc::m1' may not be ABI-compliant and may change

in a future version of GCC
In file included from pr19448_use.C:1:
bc.h:3: warning: width of `bc::m1' exceeds its type
bc.h:1: warning: the offset of `bc::m1' may not be ABI-compliant and may change

in a future version of GCC
In file included from pr19448_define.C:1:
bc.h:3: warning: width of `bc::m1' exceeds its type
In file included from pr19448_use.C:1:
bc.h:3: warning: width of `bc::m1' exceeds its type
define with 3.2.3 with -Wabi
use with    3.4.3 with -fabi-version=1
									       

./pr19448-2.sh: line 131: 26964 Aborted 		a.out
									       

define with 3.4.3 with -fabi-version=1
use with    3.2.3 with -Wabi
									       

PASSED
-------------------------------------------------------------------

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
Attachment #7963 is|0                           |1
           obsolete|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19448

[Bug c++/19448] Different value representation for bitfield width exceeding its type size.

[mmitchel at gcc dot gnu dot org]

------- Additional Comments From mmitchel at gcc dot gnu dot org  2005-01-19 02:11 -------
G++ 4.0 is correct.  This declaration gives only a signed 9-bit type,
independent of the ABI.  Since +128 is not representable in 9 signed bits, it
wraps to -128.  The same issue explains the other tests.

AFAICT, G++ 4.0 is following the ABI.  The bump of the default ABI version (to
2) is the cause because we fixed this for version 2 of the ABI.  The same goes
for the empty class change.

Therefore, I'm marking this as INVALID.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|                            |INVALID


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19448

[Bug c++/19448] Different value representation for bitfield width exceeding its type size.

[janis187 at us dot ibm dot com]

------- Additional Comments From janis187 at us dot ibm dot com  2005-01-19 17:04 -------
Mark, your response addresses the original message but not the later ones, and
not either of the attached test cases.  In those the class is:

class bc {
public:
  char m1 :17;
};

m1 is assigned a value of 1, which certainly fits.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19448

[Bug c++/19448] Different value representation for bitfield width exceeding its type size.

[janis at gcc dot gnu dot org]

------- Additional Comments From janis at gcc dot gnu dot org  2005-01-19 23:08 -------
Mark closed this based on early comments discussing invalid code, but later
comments discuss a problem with valid code.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |UNCONFIRMED
         Resolution|INVALID                     |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19448

[Bug c++/19448] Different value representation for bitfield width exceeding its type size.

[janis at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |janis at gcc dot gnu dot org
                   |dot org                     |
             Status|UNCONFIRMED                 |ASSIGNED
     Ever Confirmed|                            |1
   Last reconfirmed|0000-00-00 00:00:00         |2005-01-28 01:35:58
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19448

[Bug c++/19448] Different value representation for bitfield width exceeding its type size.

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19448

Paolo Carlini <paolo.carlini at oracle dot com> changed:

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

--- Comment #22 from Paolo Carlini <paolo.carlini at oracle dot com> ---
Thanks. Let's close this then.

[Bug c++/19448] Different value representation for bitfield width exceeding its type size.

[janis at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19448

--- Comment #21 from Janis Johnson <janis at gcc dot gnu.org> ---
I'm definitely not working on the bug anymore, and would have to do a lot of
work (better left to experts) to figure out if the test is valid.  Please
assign it to someone else, or at least unassign it from me (or tell me how to
do that).  If the functionality really mattered to the submitter he would have
spoken up again.

[Bug c++/19448] Different value representation for bitfield width exceeding its type size.

[dje at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19448

David Edelsohn <dje at gcc dot gnu.org> changed:

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

--- Comment #20 from David Edelsohn <dje at gcc dot gnu.org> ---
I don't think that Janis is working on this PR any more.  I'm not sure if the
examples still are valid C++.

unsigned char m1:17;

warning: width of 'bc::m1' exceeds its type [enabled by default]
   unsigned char m1 :17; 
                     ^

The testcase do not fail with GCC 4.7.2 and GCC 4.8.1.

[Bug c++/19448] Different value representation for bitfield width exceeding its type size.

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19448

Paolo Carlini <paolo.carlini at oracle dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |dje.gcc at gmail dot com

--- Comment #19 from Paolo Carlini <paolo.carlini at oracle dot com> ---
Adding David as powerpc expert.

[Bug c++/19448] Different value representation for bitfield width exceeding its type size.

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19448

--- Comment #18 from Paolo Carlini <paolo.carlini at oracle dot com> ---
So, are you still actively working on it? (I'm asking because the bug is
assigned to you) Do you think it's still an issue today?

[Bug c++/19448] Different value representation for bitfield width exceeding its type size.

[janis at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19448

--- Comment #17 from Janis Johnson <janis at gcc dot gnu.org> ---
Paolo, I don't remember, but assume I didn't uncover anything else that was
interesting.