[Bug java/37068] New: [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[Bug java/37068] New: [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[danglin at gcc dot gnu dot org]

Executing on host:
/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/libjava/testsuite/
../libtool --silent --tag=GCJ --mode=link /test/gnu/gcc/objdir/gcc/gcj
-B/test/g
nu/gcc/objdir/hppa2.0w-hp-hpux11.11/libjava/ -B/test/gnu/gcc/objdir/gcc/
--encod
ing=UTF-8 -B/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/libjava/testsuite/../
/te
st/gnu/gcc/gcc/libjava/testsuite/libjava.cni/PR9577.jar natPR9577.o  -w 
-no-ins
tall --main=PR9577 -g 
-L/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/./libjava/.l
ibs -lm   -o PR9577    (timeout = 300)
PASS: linking PR9577
set_ld_library_path_env_vars:
ld_library_path=.:/test/gnu/gcc/objdir/hppa2.0w-hp
-hpux11.11/./libjava/.libs
Setting LD_LIBRARY_PATH to
.:/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/./libjav
a/.libs:.:/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/./libjava/.libs
libgcj failure: gcj linkage error.
Incorrect library ABI version detected.  Aborting.

FAIL: PR9577 run

This error breaks most gcj tests.  See
<http://gcc.gnu.org/ml/gcc-testresults/2008-08/msg00893.html>.  See
<http://gcc.gnu.org/ml/gcc-testresults/2008-08/msg00807.html> for results
from the previous revision.

-bash-3.2$ ./xgcc -B. -v
Reading specs from ./specs
Target: hppa2.0w-hp-hpux11.11
Configured with: ../gcc/configure --with-gnu-as --with-as=/opt/gnu/bin/as
--enable-shared --disable-nls --with-local-prefix=/opt/gnu
--prefix=/opt/gnu/gcc/gcc-4.4.0 --enable-debug=no --enable-threads=posix
--with-mpfr=/opt/gnu/gcc/gcc-4.4.0 --with-gmp=/opt/gnu/gcc/gcc-4.3.0
--disable-libmudflap --enable-languages=c++,java
Thread model: posix
gcc version 4.4.0 20080715 (experimental) [trunk revision 137834] (GCC) 

The bug was introduced by the following change:

2008-07-15  Jan Hubicka  <jh@suse.cz>

        * lang.c (java_init_options): Enable unit-at-a-time by default.

So, presumably this change has broken reading the library ABI information.


-- 
           Summary: [4.4 Regression] libgcj linkage failure: Incorrect
                    library ABI version detected
           Product: gcc
           Version: 4.4.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: java
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: danglin at gcc dot gnu dot org
 GCC build triplet: hppa2.0w-hp-hpux11.11
  GCC host triplet: hppa2.0w-hp-hpux11.11
GCC target triplet: hppa2.0w-hp-hpux11.11


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[pinskia at gcc dot gnu dot org]

------- Comment #1 from pinskia at gcc dot gnu dot org  2008-08-10 19:55 -------
I want to say something is being miscompiled.


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |pinskia at gcc dot gnu dot
                   |                            |org
           Keywords|                            |wrong-code
   Target Milestone|---                         |4.4.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[rguenth at gcc dot gnu dot org]

-- 

rguenth at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |build
           Priority|P3                          |P2


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[hubicka at gcc dot gnu dot org]

------- Comment #2 from hubicka at gcc dot gnu dot org  2008-09-09 12:54 -------
Quite obviously :(
I will need some asistance here as I don't have HP-PA system and don't know
much about ISA either.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[dje at gcc dot gnu dot org]

------- Comment #3 from dje at gcc dot gnu dot org  2008-10-16 20:12 -------
I see the same failure on AIX that just restored Java support.

On AIX, _Jv_CheckABI is invoked with version argument of 0 instead of 404000
GCC version constant.


-- 

dje at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |dje at gcc dot gnu dot org
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|0                           |1
   Last reconfirmed|0000-00-00 00:00:00         |2008-10-16 20:12:35
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[dave at hiauly1 dot hia dot nrc dot ca]

------- Comment #4 from dave at hiauly1 dot hia dot nrc dot ca  2008-10-30 23:27 -------
Subject: Re:  [4.4 Regression] libgcj linkage failure:
        Incorrect library ABI version detected

> On AIX, _Jv_CheckABI is invoked with version argument of 0 instead of 404000
> GCC version constant.

Same on hppa2.0w-hp-hpux11.11.

(gdb) bt
#0  _Jv_CheckABIVersion (value=0)
    at ../../../gcc/libjava/java/lang/natClassLoader.cc:273
#1  0xc3f81ab4 in _Jv_RegisterClasses (classes=0x7eff2c78)
    at ../../../gcc/libjava/java/lang/natClassLoader.cc:295
#2  0xc3f81b48 in _Jv_RegisterClass (klass=0x0)
    at ../../../gcc/libjava/java/lang/natClassLoader.cc:441
#3  0xc4e37894 in global constructors keyed to
65535_0__var_tmp__ccxwp2Cmjx_4B3F0796_6CD682F9 ()
    at
../../../../../gcc/libjava/classpath/external/sax/org/xml/sax/SAXNotRecognizedException.java:49
#4  0xc3f2497c in _GLOBAL () at ../../../gcc/libjava/gcj/cni.h:61
#5  0xc003a560 in shl_invoke_initfini () from /usr/lib/dld.sl
#6  0xc0037640 in invoke_init_elab () from /usr/lib/dld.sl
#7  0xc0037610 in invoke_init_elab () from /usr/lib/dld.sl
#8  0xc0037610 in invoke_init_elab () from /usr/lib/dld.sl
#9  0xc0028d94 in finish_dld_main () from /usr/lib/dld.sl
#10 0xc002ba6c in _dld_main () from /usr/lib/dld.sl
#11 0x00002d2c in __map_dld ()
#12 0x000021dc in $START$ ()

Dave


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[dave at hiauly1 dot hia dot nrc dot ca]

------- Comment #5 from dave at hiauly1 dot hia dot nrc dot ca  2008-10-31 04:22 -------
Subject: Re:  [4.4 Regression] libgcj linkage failure: Incorrect library ABI
version detected

> Same on hppa2.0w-hp-hpux11.11.
> 
> (gdb) bt
> #0  _Jv_CheckABIVersion (value=0)
>     at ../../../gcc/libjava/java/lang/natClassLoader.cc:273
> #1  0xc3f81ab4 in _Jv_RegisterClasses (classes=0x7eff2c78)
>     at ../../../gcc/libjava/java/lang/natClassLoader.cc:295
> #2  0xc3f81b48 in _Jv_RegisterClass (klass=0x0)
>     at ../../../gcc/libjava/java/lang/natClassLoader.cc:441
> #3  0xc4e37894 in global constructors keyed to
> 65535_0__var_tmp__ccxwp2Cmjx_4B3F0796_6CD682F9 ()
>     at
> ../../../../../gcc/libjava/classpath/external/sax/org/xml/sax/SAXNotRecognizedException.java:49

The backtrace is somewhat inaccurate.  However, it seems that some klass
objects are not being correctly initialized:

...
klass=0x77afa164
klass->next_or_version=0xc0061a81
klass=0x77afa1fc
klass->next_or_version=0xc0061a81
klass=0x77af8b50
klass->next_or_version=0x0

(gdb) p ((jclass) 0x77af8b50)->next_or_version
$9 = (jclass) 0x0
(gdb) p *((jclass) 0x77af8b50)                
$10 = {<> = {<No data fields>}, static class$ = {<> = {<No data fields>}, 
    static class$ = <same as static member of an already seen type>, 
    next_or_version = 0x40062a20, name = 0x7714d2cc, accflags = 49, 

(gdb) x/4x 0x77af8b50
0x77af8b50:     0x7714d084      0x00000000      0x00000000      0x77afa530

(gdb) x/4x 0x7714d084
0x7714d084:     0x7714d190      0x00000002      0x76f421d6      0x400013aa

(gdb) x/4x 0x7714d190
0x7714d190:     0x7714d084      0x00000000      0x40062a20      0x7714d2cc

Dave


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[ro at gcc dot gnu dot org]

------- Comment #6 from ro at gcc dot gnu dot org  2008-10-31 17:18 -------
alpha-dec-osf5.1b (with the workarounds mentioned in PR bootstrap/37426) is
affected as well. _Jv_CheckABIVersion is invoked with 3221625473.


-- 

ro at gcc dot gnu dot org changed:

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[dave at hiauly1 dot hia dot nrc dot ca]

------- Comment #7 from dave at hiauly1 dot hia dot nrc dot ca  2008-10-31 19:50 -------
Subject: Re:  [4.4 Regression] libgcj linkage failure: Incorrect library ABI
version detected

> alpha-dec-osf5.1b (with the workarounds mentioned in PR bootstrap/37426) is
> affected as well. _Jv_CheckABIVersion is invoked with 3221625473.

It seems that when using collect2 the global constructor for libgcj
calls _Jv_RegisterClass more than once for the same class.  The default
class hook, _Jv_RegisterClassHookDefault, messes with klass->next_or_version.
The attempt at reregistration is not detected before _Jv_CheckABIVersion
is called causing the abort.

Dave


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[dave at hiauly1 dot hia dot nrc dot ca]

------- Comment #8 from dave at hiauly1 dot hia dot nrc dot ca  2008-10-31 23:31 -------
Subject: Re:  [4.4 Regression] libgcj linkage failure: Incorrect library ABI
version detected

> It seems that when using collect2 the global constructor for libgcj
> calls _Jv_RegisterClass more than once for the same class.

This seems to be because org-xml.o, for example,  contains two global
constructors:
_GLOBAL__I_65535_0__var_tmp__cc4hyJYpjx_4B3F0796_410DDDB4
_GLOBAL__I_65535_1__var_tmp__cc4hyJYpjx_4B3F0796_410DDDB4

Both register the same classes.

Dave


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[danglin at gcc dot gnu dot org]

------- Comment #9 from danglin at gcc dot gnu dot org  2008-11-01 16:18 -------
Breakpoint 1, cgraph_build_static_cdtor (which=-52 '?', body=0x400555d8, 
    priority=1074005676) at ../../gcc/gcc/cgraphunit.c:1353
1353      sprintf (which_buf, "%c_%.5d_%d", which, priority, counter++);
(gdb) bt
#0  cgraph_build_static_cdtor (which=-52 '?', body=0x400555d8, 
    priority=1074005676) at ../../gcc/gcc/cgraphunit.c:1353
#1  0x00077010 in java_parse_file (set_yydebug=73)
    at ../../gcc/gcc/java/jcf-parse.c:1702
#2  0x001ff4e8 in toplev_main (argc=1073746884, argv=0x400555d8)
    at ../../gcc/gcc/toplev.c:968
#3  0x00082a6c in main (argc=73, argv=0x7aff79e0) at ../../gcc/gcc/main.c:35
(gdb) c
Continuing.

Breakpoint 1, cgraph_build_static_cdtor (which=1 '\001', body=0x7eff0a08, 
    priority=1) at ../../gcc/gcc/cgraphunit.c:1353
1353      sprintf (which_buf, "%c_%.5d_%d", which, priority, counter++);
(gdb) bt
#0  cgraph_build_static_cdtor (which=1 '\001', body=0x7eff0a08, priority=1)
    at ../../gcc/gcc/cgraphunit.c:1353
#1  0x00289a20 in build_cdtor (ctor_p=1 '\001', cdtors=0x7afec8a8, len=1)
    at ../../gcc/gcc/cgraphunit.c:219
#2  0x0028bdc0 in cgraph_optimize () at ../../gcc/gcc/cgraphunit.c:289
#3  0x00077030 in java_parse_file (set_yydebug=73)
    at ../../gcc/gcc/java/jcf-parse.c:1961
#4  0x001ff4e8 in toplev_main (argc=1073746884, argv=0x400555d8)
    at ../../gcc/gcc/toplev.c:968
#5  0x00082a6c in main (argc=73, argv=0x7a7af860) at ../../gcc/gcc/main.c:35


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[dave at hiauly1 dot hia dot nrc dot ca]

------- Comment #10 from dave at hiauly1 dot hia dot nrc dot ca  2008-11-01 16:49 -------
Subject: Re:  [4.4 Regression] libgcj linkage failure:
        Incorrect library ABI version detected

On Sat, 01 Nov 2008, danglin at gcc dot gnu dot org wrote:

Possibly the attached change will fix the problem.  However, I'm
not sure if cgraphunit.c will detect all the java constructors that
need calling.  Someone who knows a little about java will need to
look at this.

Dave


------- Comment #11 from dave at hiauly1 dot hia dot nrc dot ca  2008-11-01 16:49 -------
Created an attachment (id=16608)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=16608&action=view)


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[dave at hiauly1 dot hia dot nrc dot ca]

------- Comment #12 from dave at hiauly1 dot hia dot nrc dot ca  2008-11-02 02:33 -------
Subject: Re:  [4.4 Regression] libgcj linkage failure: Incorrect library ABI
version detected

> On Sat, 01 Nov 2008, danglin at gcc dot gnu dot org wrote:
> 
> Possibly the attached change will fix the problem.  However, I'm
> not sure if cgraphunit.c will detect all the java constructors that
> need calling.  Someone who knows a little about java will need to
> look at this.

The patch doesn't work.  It appears some classes are not registered.

Starting program:
/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/libjava/.libs/gcj-dbtool -n
classmap.db
warning: Private mapping of shared library text was not specified
by the executable; setting a breakpoint in a shared library which
is not privately mapped will not work.  See the HP-UX 11i v3 chatr
manpage for methods to privately map shared library text.
[New process 9031, lwp 4498783]
[process 9031, lwp 4498783 exited]

Program received signal SIGSEGV, Segmentation fault.
0xc3eca160 in _Jv_Linker::verify_class (klass=0x7a5f7428)
    at ../../../gcc/libjava/link.cc:1904
1904      klass->engine->verify(klass);
(gdb) p klass->engine
$5 = (_Jv_ExecutionEngine *) 0x0

void
_Jv_RegisterClassHookDefault (jclass klass)
{
  // This is bogus, but there doesn't seem to be a better place to do
  // it.
  if (! klass->engine)
    klass->engine = &_Jv_soleCompiledEngine;

Dave


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[aph at gcc dot gnu dot org]

------- Comment #13 from aph at gcc dot gnu dot org  2008-11-03 10:18 -------
As a Java maintainer I'm happy to have a look at this, but I have no access to
HP/UX hardware.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[jakub at gcc dot gnu dot org]

------- Comment #14 from jakub at gcc dot gnu dot org  2008-11-03 11:54 -------
Does this cause a bootstrap failure anywhere, or just testsuite regressions and
(ignored) gcj-dbtool failure?  If the latter, this should be P4.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[dave at hiauly1 dot hia dot nrc dot ca]

------- Comment #15 from dave at hiauly1 dot hia dot nrc dot ca  2008-11-03 14:27 -------
Subject: Re:  [4.4 Regression] libgcj linkage failure: Incorrect library ABI
version detected

> Does this cause a bootstrap failure anywhere, or just testsuite regressions and
> (ignored) gcj-dbtool failure?  If the latter, this should be P4.

It just causes testsuite regressions.  As you note, the gcj-dbtool failure
is ignored.  Essentially, libgcj is broken and isn't usable.

Dave


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[aph at gcc dot gnu dot org]

------- Comment #16 from aph at gcc dot gnu dot org  2008-11-03 14:45 -------
Indeed.  If that doesn't work, libjava is hosed on that platform.
I'd like to get to the bottom of it.  Should the FE be calling
cgraph_build_static_cdtor() or not?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[jakub at gcc dot gnu dot org]

-- 

jakub at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P2                          |P4


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[dave at hiauly1 dot hia dot nrc dot ca]

------- Comment #17 from dave at hiauly1 dot hia dot nrc dot ca  2008-11-03 15:02 -------
Subject: Re:  [4.4 Regression] libgcj linkage failure: Incorrect library ABI
version detected

> ------- Comment #13 from aph at gcc dot gnu dot org  2008-11-03 10:18 -------
> As a Java maintainer I'm happy to have a look at this, but I have no access to
> HP/UX hardware.

I could provide access.  However, debugging this with gdb is tricky.
It can't set a breakpoint in a constructor in a shared library.  There's
some issue with load notifications.  It's also not possible to link
with -static.

I'm willing to look at anything you suggest.  There's a couple of
other PRs related to gcj-dbtool that probably relate to this problem.

The org-xml.list is one in which I see the same class registered twice
(e.g., _ZN3org3xml3sax3ext14LexicalHandler6class$E).

Dave


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[aph at redhat dot com]

------- Comment #18 from aph at redhat dot com  2008-11-03 15:12 -------
Subject: Re:  [4.4 Regression] libgcj linkage failure: Incorrect
 library ABI version detected

dave at hiauly1 dot hia dot nrc dot ca wrote:
> ------- Comment #17 from dave at hiauly1 dot hia dot nrc dot ca  2008-11-03 15:02 -------
> Subject: Re:  [4.4 Regression] libgcj linkage failure: Incorrect library ABI
> version detected
> 
>> ------- Comment #13 from aph at gcc dot gnu dot org  2008-11-03 10:18 -------
>> As a Java maintainer I'm happy to have a look at this, but I have no access to
>> HP/UX hardware.
> 
> I could provide access.  However, debugging this with gdb is tricky.
> It can't set a breakpoint in a constructor in a shared library.  There's
> some issue with load notifications.  It's also not possible to link
> with -static.

That doesn't matter, because it's not a runtime bug, it's a
compiler bug.  We have to debug the compiler.

> I'm willing to look at anything you suggest.  There's a couple of
> other PRs related to gcj-dbtool that probably relate to this problem.
> 
> The org-xml.list is one in which I see the same class registered twice
> (e.g., _ZN3org3xml3sax3ext14LexicalHandler6class$E).

Okay, the question is why is cgraph_build_static_cdtor() being called twice,
once from cgraph_optimize() and once from java_parse_file() ?

And, if the FE should not call cgraph_build_static_cdtor(), why does
the code generation fail if the call is removed ?

The entire solution to this problem lies in that answer.

Andrew.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[dje at gcc dot gnu dot org]

------- Comment #19 from dje at gcc dot gnu dot org  2008-11-03 17:35 -------
Characterizing this as "testsuite failures" trivializes the problem.  libgcj
builds but does not work properly on the platforms.  This PR should be
categorized as P2, not P4.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[mmitchel at gcc dot gnu dot org]

------- Comment #20 from mmitchel at gcc dot gnu dot org  2008-11-03 17:38 -------
Do we still think this is Java-specific, or do we have any evidence that this
can happen in C/C++ as well?


-- 

mmitchel at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P4                          |P3


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[jakub at gcc dot gnu dot org]

------- Comment #21 from jakub at gcc dot gnu dot org  2008-11-03 17:41 -------
Java is not a primary language, bugs that affect just Java and not C/C++
shouldn't be P1 or P2.  Fortran FE bugs are P4 as well.  That doesn't mean it
can't be fixed, just that the release will not be delayed because of it.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[aph at gcc dot gnu dot org]

------- Comment #22 from aph at gcc dot gnu dot org  2008-11-03 17:47 -------
I think think this is Java-specific.

It's quite likely that the Java FE should not be calling
cgraph_build_static_cdtor(), but when that call is removed some test
cases fail.  Rather than arguing what priority this should be, all we
need is someone who actually understands cgraph_build_static_cdtor(),
and can tell me when it should be called.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[mark at codesourcery dot com]

------- Comment #23 from mark at codesourcery dot com  2008-11-04 05:51 -------
Subject: Re:  [4.4 Regression] libgcj linkage failure: Incorrect
 library ABI version detected

aph at gcc dot gnu dot org wrote:

> It's quite likely that the Java FE should not be calling
> cgraph_build_static_cdtor(), but when that call is removed some test
> cases fail.  Rather than arguing what priority this should be, all we
> need is someone who actually understands cgraph_build_static_cdtor(),
> and can tell me when it should be called.

You shouldn't call that function.  Instead, you should set
DECL_STATIC_{CONSTRUCTOR,DESTRUCTOR}.  Then, cgraph will do the right
thing.  If necessary, you can also call decl_init_priority_insert.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[aph at gcc dot gnu dot org]

-- 

aph at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |aph at gcc dot gnu dot org
                   |dot org                     |
             Status|NEW                         |ASSIGNED
           Priority|P3                          |P2


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[aph at redhat dot com]

------- Comment #24 from aph at redhat dot com  2008-11-04 16:09 -------
Subject: Re:  [4.4 Regression] libgcj linkage failure: Incorrect
 library ABI version detected

mark at codesourcery dot com wrote:

> You shouldn't call that function.  Instead, you should set
> DECL_STATIC_{CONSTRUCTOR,DESTRUCTOR}.  Then, cgraph will do the right
> thing.  If necessary, you can also call decl_init_priority_insert.

This patch works on x86_64 GNU/Linux.

Please HP/UX, AIX, and OSF users make sure it works for them as well.

Thanks,
Andrew.



2008-11-04  Andrew Haley  <aph@redhat.com>

        * jcf-parse.c (java_emit_static_constructor): Don't call
        cgraph_build_static_cdtor.  Rewrite.

Index: jcf-parse.c
===================================================================
--- jcf-parse.c (revision 141575)
+++ jcf-parse.c (working copy)
@@ -1699,7 +1699,32 @@
   write_resource_constructor (&body);

   if (body)
-    cgraph_build_static_cdtor ('I', body, DEFAULT_INIT_PRIORITY);
+    {
+      tree name = get_identifier ("_Jv_global_static_constructor");
+
+      tree decl = build_decl (FUNCTION_DECL, name,
+                             build_function_type (void_type_node,
void_list_node));
+
+      tree resdecl = build_decl (RESULT_DECL, NULL_TREE, void_type_node);
+      DECL_ARTIFICIAL (resdecl) = 1;
+      DECL_RESULT (decl) = resdecl;
+      current_function_decl = decl;
+      allocate_struct_function (decl, false);
+
+      TREE_STATIC (decl) = 1;
+      TREE_USED (decl) = 1;
+      DECL_ARTIFICIAL (decl) = 1;
+      DECL_NO_INSTRUMENT_FUNCTION_ENTRY_EXIT (decl) = 1;
+      DECL_SAVED_TREE (decl) = body;
+      DECL_UNINLINABLE (decl) = 1;
+
+      DECL_INITIAL (decl) = make_node (BLOCK);
+      TREE_USED (DECL_INITIAL (decl)) = 1;
+
+      DECL_STATIC_CONSTRUCTOR (decl) = 1;
+      java_genericize (decl);
+      cgraph_finalize_function (decl, false);
+    }
 }


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[dje dot gcc at gmail dot com]

------- Comment #25 from dje dot gcc at gmail dot com  2008-11-04 18:11 -------
Subject: Re:  [4.4 Regression] libgcj linkage failure: Incorrect library ABI
version detected

> This patch works on x86_64 GNU/Linux.
>
> Please HP/UX, AIX, and OSF users make sure it works for them as well.

I recompiled jc1 and libjava with the patch.  I still encounter failures with
execution tests.  I will be interested to learn Dave's results on HP-UX.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[dje dot gcc at gmail dot com]

------- Comment #26 from dje dot gcc at gmail dot com  2008-11-04 18:15 -------
Subject: Re:  [4.4 Regression] libgcj linkage failure: Incorrect library ABI
version detected

The runtime failure has changed and no longer is complaining about ABI
versions on AIX.  As I mentioned, I will be interested to learn from Dave's
results this new runtime problem is AIX-specific or common to the other
platforms.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[aph at redhat dot com]

------- Comment #27 from aph at redhat dot com  2008-11-04 18:16 -------
Subject: Re:  [4.4 Regression] libgcj linkage failure: Incorrect
 library ABI version detected

dje dot gcc at gmail dot com wrote:
> ------- Comment #25 from dje dot gcc at gmail dot com  2008-11-04 18:11 -------
> Subject: Re:  [4.4 Regression] libgcj linkage failure: Incorrect library ABI
> version detected
> 
>> This patch works on x86_64 GNU/Linux.
>>
>> Please HP/UX, AIX, and OSF users make sure it works for them as well.
> 
> I recompiled jc1 and libjava with the patch.  I still encounter failures with
> execution tests.  I will be interested to learn Dave's results on HP-UX.

Oh dear: I followed mmitchel's instructions to the letter.
Sooner or later someone is going to have to be allowed access to
a suitable machine.

Andrew.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[aph at gcc dot gnu dot org]

------- Comment #28 from aph at gcc dot gnu dot org  2008-11-04 18:19 -------
So I take it that the libgcj linkage failure is fixed, but some other unrelated
bug now occurs?  Sounds like this one needs to be closed.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[dje at gcc dot gnu dot org]

------- Comment #29 from dje at gcc dot gnu dot org  2008-11-04 18:33 -------
> So I take it that the libgcj linkage failure is fixed, but some other unrelated
> bug now occurs?  Sounds like this one needs to be closed.

No.  "linkage failure" does not mean a link-edit failure.  It always has been a
runtime failure: "libgcj failure: gcj linkage error."

That error now is gone, but we may only have stepped to the next error on these
platforms.  We can decide if it should continue in this bug or a new bug should
be opened.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[dave at hiauly1 dot hia dot nrc dot ca]

------- Comment #30 from dave at hiauly1 dot hia dot nrc dot ca  2008-11-04 19:01 -------
Subject: Re:  [4.4 Regression] libgcj linkage failure: Incorrect library ABI
version detected

> That error now is gone, but we may only have stepped to the next error on these
> platforms.  We can decide if it should continue in this bug or a new bug should
> be opened.

If the error is releated to the running of constructors for class
registration or resource registration, then I think we should continue
with this bug.  Otherwise, it's a new bug.

Dave


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[aph at redhat dot com]

------- Comment #31 from aph at redhat dot com  2008-11-04 23:06 -------
Subject: Re:  [4.4 Regression] libgcj linkage failure: Incorrect
 library ABI version detected

dave at hiauly1 dot hia dot nrc dot ca wrote:
> ------- Comment #30 from dave at hiauly1 dot hia dot nrc dot ca  2008-11-04 19:01 -------
> Subject: Re:  [4.4 Regression] libgcj linkage failure: Incorrect library ABI
> version detected
> 
>> That error now is gone, but we may only have stepped to the next error on these
>> platforms.  We can decide if it should continue in this bug or a new bug should
>> be opened.
> 
> If the error is releated to the running of constructors for class
> registration or resource registration, then I think we should continue
> with this bug.  Otherwise, it's a new bug.

OK, but I need to know if my patch has been tested well enough for
me to check it in.  I'm fairly certain it doesn't break x86 GNU/Linux.

Andrew.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[dave at hiauly1 dot hia dot nrc dot ca]

------- Comment #32 from dave at hiauly1 dot hia dot nrc dot ca  2008-11-05 01:11 -------
Subject: Re:  [4.4 Regression] libgcj linkage failure: Incorrect library ABI
version detected

> OK, but I need to know if my patch has been tested well enough for
> me to check it in.  I'm fairly certain it doesn't break x86 GNU/Linux.

For me, the change fixes the PR.  There are still 10 fails in the
testsuite on hppa2.0w-hp-hpux11.11.  Four are expected because large
file support is not enabled.  I don't believe the other fails have
anything to do with this PR.

I checked that the duplicates are gone in org-xml.s.  gcj-dbtool
now runs successfully.  The testsuite on hppa-unknown-linux-gnu
is still clean.  There are no ABI version failures.

Thanks,
Dave


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[aph at gcc dot gnu dot org]

------- Comment #33 from aph at gcc dot gnu dot org  2008-11-05 10:06 -------
Subject: Bug 37068

Author: aph
Date: Wed Nov  5 10:05:03 2008
New Revision: 141604

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=141604
Log:
2008-11-04  Andrew Haley  <aph@redhat.com>

        PR java/37068
        * jcf-parse.c (java_emit_static_constructor): Don't call
        cgraph_build_static_cdtor.  Rewrite.

Modified:
    trunk/gcc/java/ChangeLog
    trunk/gcc/java/jcf-parse.c


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[jakub at gcc dot gnu dot org]

------- Comment #34 from jakub at gcc dot gnu dot org  2008-11-05 10:14 -------
This bug seems to be fixed then, if there are unrelated AIX or HPUX libjava
testsuite failures, please file separate bugs for them.


-- 

jakub at gcc dot gnu dot org changed:

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[ro at techfak dot uni-bielefeld dot de]

------- Comment #35 from ro at techfak dot uni-bielefeld dot de  2008-11-05 13:40 -------
Subject: Re:  [4.4 Regression] libgcj linkage failure: Incorrect library ABI
version detected

jakub at gcc dot gnu dot org writes:

> This bug seems to be fixed then, if there are unrelated AIX or HPUX libjava
> testsuite failures, please file separate bugs for them.

After rebuilding jc1 and rerunning the libjava testsuite on
alpha-dec-osf5.1b, I'm back at the same set of testsuite failures that I
had on 20080313.

        Rainer


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[aph at redhat dot com]

------- Comment #36 from aph at redhat dot com  2008-11-05 14:02 -------
Subject: Re:  [4.4 Regression] libgcj linkage failure: Incorrect
 library ABI version detected

ro at techfak dot uni-bielefeld dot de wrote:

> After rebuilding jc1 and rerunning the libjava testsuite on
> alpha-dec-osf5.1b, I'm back at the same set of testsuite failures that I
> had on 20080313.

Feel free to CC me with any of these.

Andrew.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[ro at techfak dot uni-bielefeld dot de]

------- Comment #37 from ro at techfak dot uni-bielefeld dot de  2008-11-05 15:07 -------
Subject: Re:  [4.4 Regression] libgcj linkage failure: Incorrect library ABI
version detected

aph at redhat dot com writes:

> Feel free to CC me with any of these.

Thanks.  A considerable number of them was already reported as PR
target/22224.

        Rainer


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068

[Bug java/37068] [4.4 Regression] libgcj linkage failure: Incorrect library ABI version detected

[dje at gcc dot gnu dot org]

------- Comment #38 from dje at gcc dot gnu dot org  2008-11-05 15:45 -------
I think the AIX testsuite problem is a problem with the testsuite harness.  If
I compile a trivial Java program with gcj, it works.  If I manually compile
programs from libjava.lang testsuite, they work.  If I manually run the
testsuite executables built by the testsuite, they work.  But the testsuite
reports that execution fails.  Where is libjava_load defined?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37068