[Bug java/21070] New: [4.1 Regression]: java compiler generates wrong code on ia64

[Bug java/21070] New: [4.1 Regression]: java compiler generates wrong code on ia64

[hjl at lucon dot org]

The patch

http://gcc.gnu.org/ml/gcc-patches/2005-03/msg00712.html

causes more than 1000 failures:

http://gcc.gnu.org/ml/gcc-testresults/2005-04/msg01089.html

The problem seems uniuqe to ia64. The good ones:

[hjl@gnu-4 libjava]$ readelf --wide  -
sr  ./java/util/logging/.libs/natLogger.o | grep 
ZN4java4util7logging6Logger7getNameEv
0000000000000d17  0000000800000027 R_IA64_DIR64LSB        
0000000000000000 .gnu.linkonce.t._ZN4java4util7logging6Logger7getNameEv + 0
0000000000000d1f  0000000800000027 R_IA64_DIR64LSB        
0000000000000000 .gnu.linkonce.t._ZN4java4util7logging6Logger7getNameEv + 20
0000000000000231  0000000800000027 R_IA64_DIR64LSB        
0000000000000000 .gnu.linkonce.t._ZN4java4util7logging6Logger7getNameEv + 0
0000000000000000  0000000800000027 R_IA64_DIR64LSB        
0000000000000000 .gnu.linkonce.t._ZN4java4util7logging6Logger7getNameEv + 0
0000000000000008  0000000800000027 R_IA64_DIR64LSB        
0000000000000000 .gnu.linkonce.t._ZN4java4util7logging6Logger7getNameEv + 2
0000000000000020  0000000800000027 R_IA64_DIR64LSB        
0000000000000000 .gnu.linkonce.t._ZN4java4util7logging6Logger7getNameEv + 0
    19: 0000000000000000    32 FUNC    WEAK   DEFAULT   11 
_ZN4java4util7logging6Logger7getNameEv
[hjl@gnu-4 libjava]$ readelf --wide  -sr  ./java/util/logging/.libs/Logger.o | 
grep ZN4java4util7logging6Logger7getNameEv
00000000000000e0  000000ef00000081 R_IA64_IPLTLSB         0000000000000df0 
_ZN4java4util7logging6Logger7getNameEv + 0
0000000000000640  000000a000000047 R_IA64_FPTR64LSB       
0000000000000df0 .L_ZN4java4util7logging6Logger7getNameEv13 + 0
   160: 0000000000000df0    32 FUNC    LOCAL  DEFAULT    
1 .L_ZN4java4util7logging6Logger7getNameEv13
   239: 0000000000000df0    32 FUNC    GLOBAL DEFAULT    1 
_ZN4java4util7logging6Logger7getNameEv

Even if ld sees natLogger.o first, the definitions in Logger.o will be used.
The bad ones are

[hjl@gnu-4 libjava]$ readelf --wide  -
sr  ./java/util/logging/.libs/natLogger.o | grep 
ZN4java4util7logging6Logger7getNameEv
0000000000000d17  0000000800000027 R_IA64_DIR64LSB        
0000000000000000 .gnu.linkonce.t._ZN4java4util7logging6Logger7getNameEv + 0
0000000000000d1f  0000000800000027 R_IA64_DIR64LSB        
0000000000000000 .gnu.linkonce.t._ZN4java4util7logging6Logger7getNameEv + 20
0000000000000231  0000000800000027 R_IA64_DIR64LSB        
0000000000000000 .gnu.linkonce.t._ZN4java4util7logging6Logger7getNameEv + 0
0000000000000000  0000000800000027 R_IA64_DIR64LSB        
0000000000000000 .gnu.linkonce.t._ZN4java4util7logging6Logger7getNameEv + 0
0000000000000008  0000000800000027 R_IA64_DIR64LSB        
0000000000000000 .gnu.linkonce.t._ZN4java4util7logging6Logger7getNameEv + 2
0000000000000020  0000000800000027 R_IA64_DIR64LSB        
0000000000000000 .gnu.linkonce.t._ZN4java4util7logging6Logger7getNameEv + 0
    19: 0000000000000000    32 FUNC    WEAK   DEFAULT   11 
_ZN4java4util7logging6Logger7getNameEv
[hjl@gnu-4 libjava]$ readelf --wide  -sr  ./java/util/logging/.libs/Logger.o | 
grep ZN4java4util7logging6Logger7getNameEv
0000000000000df6  0000004100000027 R_IA64_DIR64LSB        
0000000000000000 .gnu.linkonce.t._ZN4java4util7logging6Logger7getNameEv + 0
0000000000000dfe  0000004100000027 R_IA64_DIR64LSB        
0000000000000000 .gnu.linkonce.t._ZN4java4util7logging6Logger7getNameEv + 20
00000000000005db  0000004100000027 R_IA64_DIR64LSB        
0000000000000000 .gnu.linkonce.t._ZN4java4util7logging6Logger7getNameEv + 0
00000000000000e0  0000019500000081 R_IA64_IPLTLSB         0000000000000000 
_ZN4java4util7logging6Logger7getNameEv + 0
0000000000000640  0000011c00000047 R_IA64_FPTR64LSB       
0000000000000000 .L_ZN4java4util7logging6Logger7getNameEv13 + 0
000000000000033a  0000004100000027 R_IA64_DIR64LSB        
0000000000000000 .gnu.linkonce.t._ZN4java4util7logging6Logger7getNameEv + 0
0000000000000342  0000004100000027 R_IA64_DIR64LSB        
0000000000000000 .gnu.linkonce.t._ZN4java4util7logging6Logger7getNameEv + 2
00000000000000a0  0000004100000027 R_IA64_DIR64LSB        
0000000000000000 .gnu.linkonce.t._ZN4java4util7logging6Logger7getNameEv + 0
   284: 0000000000000000    32 FUNC    LOCAL  DEFAULT  
134 .L_ZN4java4util7logging6Logger7getNameEv13
   405: 0000000000000000    32 FUNC    WEAK   DEFAULT  134 
_ZN4java4util7logging6Logger7getNameEv

The ones in Logger.o will be discarded. The problem may 
be .L_ZN4java4util7logging6Logger7getNameEv13. Discarding definitions in 
Logger.o seems to cause problems on ia64.

-- 
           Summary: [4.1 Regression]: java compiler generates wrong code on
                    ia64
           Product: gcc
           Version: 4.1.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: java
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: hjl at lucon dot org
                CC: gcc-bugs at gcc dot gnu dot org,java-prs at gcc dot gnu
                    dot org
 GCC build triplet: ia64-unknown-linux-gnu
  GCC host triplet: ia64-unknown-linux-gnu
GCC target triplet: ia64-unknown-linux-gnu


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21070

[Bug middle-end/21070] [4.1 Regression]: java compiler generates wrong code on ia64

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-04-17 16:34 -------
Are you sure that this is not a linker problem.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|java                        |middle-end


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21070

[Bug middle-end/21070] [4.1 Regression]: java compiler generates wrong code on ia64

[hjl at lucon dot org]

------- Additional Comments From hjl at lucon dot org  2005-04-17 16:39 -------
I am investigating the ia64 linker. I think the ia64 linker should issue
an error if it can't get it to work at the run time.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21070

[Bug middle-end/21070] [4.1 Regression]: java compiler generates wrong code on ia64

[hjl at lucon dot org]

------- Additional Comments From hjl at lucon dot org  2005-04-17 18:01 -------
.L_ZN4java4util7logging6Logger7getNameEv13 is a local alias of
_ZN4java4util7logging6Logger7getNameEv. When
_ZN4java4util7logging6Logger7getNameEv is discarded, I am trying
to figure out what happened to .L_ZN4java4util7logging6Logger7getNameEv13.
My simple testcase indicates linker will issue an error. But gcc
shouldn't use .L_ZN4java4util7logging6Logger7getNameEv13 when
_ZN4java4util7logging6Logger7getNameEv is a link once symbol. It
should use _ZN4java4util7logging6Logger7getNameEv directly.

I think this bug affects all link once targets. It is just that it
shows up on ia64 immediately.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21070

[Bug middle-end/21070] [4.1 Regression]: java compiler generates wrong code on ia64

[hjl at lucon dot org]

------- Additional Comments From hjl at lucon dot org  2005-04-17 18:06 -------
Ok, linkder did complain:

/usr/local/bin/ld: `.L_ZN4java4util7logging6Logger7getNameEv13' referenced in 
section `.data.rel' of ./.libs/libgcj0_convenience.a(Logger.o): defined in 
discarded section `.gnu.linkonce.t._ZN4java4util7logging6Logger7getNameEv
[java::util::logging::Logger::getName()]' of ./.libs/libgcj0_convenience.a
(Logger.o)
/usr/local/bin/ld: BFD 2.16.90.0.2 20050414 assertion fail elf64-ia64.c:3371

It is a gcc bug.



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21070

[Bug java/21070] [4.1 Regression]: java compiler generates wrong code on ia64

[pinskia at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|middle-end                  |java


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21070

[Bug java/21070] [4.1 Regression]: java compiler generates wrong code on ia64

[hjl at lucon dot org]

------- Additional Comments From hjl at lucon dot org  2005-04-17 19:12 -------
A patch is posted at

http://gcc.gnu.org/ml/gcc-patches/2005-04/msg01915.html

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21070

[Bug java/21070] [4.1 Regression]: java compiler generates wrong code on ia64

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-04-19 20:57 -------
*** Bug 21109 has been marked as a duplicate of this bug. ***

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |schwab at suse dot de


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21070

[Bug java/21070] [4.1 Regression]: java compiler generates wrong code on ia64

[schwab at suse dot de]

------- Additional Comments From schwab at suse dot de  2005-05-05 14:14 -------
I'm seeing the same error in some C++-only packages.  The smallest example I 
could find is jikes 
<http://prdownloads.sourceforge.net/jikes/jikes-1.22.tar.bz2>: 
 
g++  -O2 -fmessage-length=0 -Wall   -o jikes  ast.o body.o bytecode.o case.o 
class.o code.o control.o decl.o definite.o depend.o diagnose.o double.o dump.o 
error.o expr.o incrmnt.o init.o javaact.o jikes.o jikesapi.o long.o lookup.o 
lpginput.o modifier.o op.o option.o parser.o platform.o scanner.o segment.o 
set.o stream.o symbol.o system.o tab.o unparse.o unzip.o zip.o 
/usr/lib/gcc/ia64-suse-linux/4.0.1/../../../../ia64-suse-linux/bin/ld: 
`.gnu.linkonce.t._ZN12ConstantPool9CPInvalidD1Ev' referenced in section 
`.gnu.linkonce.ia64unw._ZN12ConstantPool9CPInvalidD1Ev' of class.o: defined in 
discarded section 
`.gnu.linkonce.t._ZN12ConstantPool9CPInvalidD1Ev[ConstantPool::CPInvalid::~CPInvalid()]' 
of class.o 
/usr/lib/gcc/ia64-suse-linux/4.0.1/../../../../ia64-suse-linux/bin/ld: final 
link failed: Bad value 
 

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21070

[Bug java/21070] [4.1 Regression]: java compiler generates wrong code on ia64

[hjl at lucon dot org]

------- Additional Comments From hjl at lucon dot org  2005-05-06 02:48 -------
Which binutils are you using? It looks like this problem

http://sourceware.org/ml/binutils/2005-04/msg00011.html

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21070

[Bug java/21070] [4.1 Regression]: java compiler generates wrong code on ia64

[schwab at suse dot de]

------- Additional Comments From schwab at suse dot de  2005-05-06 09:33 -------
I'm using binutils 2.16.90.0.2. 

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21070

[Bug java/21070] [4.1 Regression]: java compiler generates wrong code on ia64

[hjl at lucon dot org]

------- Additional Comments From hjl at lucon dot org  2005-05-07 20:24 -------
That is a binutils bug:

http://sources.redhat.com/bugzilla/show_bug.cgi?id=940

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21070

[Bug java/21070] [4.1 Regression]: java compiler generates wrong code on ia64

[hjl at lucon dot org]

------- Comment #12 from hjl at lucon dot org  2005-10-26 04:42 -------
I didn't see those failures today.


-- 

hjl at lucon dot org changed:

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21070

[Bug java/21070] [4.1 Regression]: java compiler generates wrong code on ia64

[pinskia at gcc dot gnu dot org]

------- Comment #11 from pinskia at gcc dot gnu dot org  2005-10-25 19:39 -------
Is this now fixed?


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |WAITING


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21070