Re: libgcj/6092: sparc-sun-solaris2.7 gcc-3.1 has hundreds of libjava failures with -m64

Re: libgcj/6092: sparc-sun-solaris2.7 gcc-3.1 has hundreds of libjava failures with -m64

[Tom Tromey]

The following reply was made to PR libgcj/6092; it has been noted by GNATS.

From: Tom Tromey <tromey@redhat.com>
To: "Kaveh R. Ghazi" <ghazi@caip.rutgers.edu>
Cc: ghazi@gcc.gnu.org, davem@gcc.gnu.org, gcc-bugs@gcc.gnu.org,
        gcc-gnats@gcc.gnu.org, gcc-prs@gcc.gnu.org, java-prs@gcc.gnu.org
Subject: Re: libgcj/6092: sparc-sun-solaris2.7 gcc-3.1 has hundreds of libjava failures with -m64
Date: 24 Apr 2002 17:05:04 -0600

 >>>>> "Kaveh" == Kaveh R Ghazi <ghazi@caip.rutgers.edu> writes:
 
 Kaveh> I'm not sure what problem you are seeing or why.  Can you
 Kaveh> provide more detail?  Perhaps it makes sense to open a separate
 Kaveh> PR.
 
 I am seeing the same behavior I saw before the patch went in.
 The GC crashes in the same way.
 At this point I'm assuming there is just something wrong with my build.
 I'm not very concerned about it, and anyway I don't have the time to
 look into it.
 
 Tom

Re: libgcj/6092: sparc-sun-solaris2.7 gcc-3.1 has hundreds of libjava failures with -m64

[Kaveh R. Ghazi]

The following reply was made to PR libgcj/6092; it has been noted by GNATS.

From: "Kaveh R. Ghazi" <ghazi@caip.rutgers.edu>
To: tromey@redhat.com
Cc: davem@gcc.gnu.org, gcc-bugs@gcc.gnu.org, gcc-gnats@gcc.gnu.org,
        gcc-prs@gcc.gnu.org, ghazi@gcc.gnu.org, java-prs@gcc.gnu.org
Subject: Re: libgcj/6092: sparc-sun-solaris2.7 gcc-3.1 has hundreds of libjava failures with -m64
Date: Wed, 24 Apr 2002 19:12:57 -0400 (EDT)

  > From: Tom Tromey <tromey@redhat.com>
  > 
  > >>>>> "Kaveh" == Kaveh R Ghazi <ghazi@caip.rutgers.edu> writes:
  > 
  > Kaveh> I'm not sure what problem you are seeing or why.  Can you
  > Kaveh> provide more detail?  Perhaps it makes sense to open a separate
  > Kaveh> PR.
  > 
  > I am seeing the same behavior I saw before the patch went in.
  > The GC crashes in the same way.
  > At this point I'm assuming there is just something wrong with my build.
  > I'm not very concerned about it, and anyway I don't have the time to
  > look into it.
  > Tom
 
 Tom - Here's an independent example that leads me to believe its truly
 fixed.  A 4/18 build on a sparcv9 target had 257 failures, whereas a
 4/22 build had only 4 failures.  See:
 
 http://gcc.gnu.org/ml/gcc-testresults/2002-04/msg00685.html
 http://gcc.gnu.org/ml/gcc-testresults/2002-04/msg00820.html
 
 So I also suspect there's something wrong in your personal build.
 But it could also be something subtly different about your box
 like the kernel version, etc.  I can't say for sure.
 
 		--Kaveh
 --
 Kaveh R. Ghazi			Director of Systems Architecture
 ghazi@caip.rutgers.edu		Qwest Global Services

Re: libgcj/6092: sparc-sun-solaris2.7 gcc-3.1 has hundreds of libjava failures with -m64

[Kaveh R. Ghazi]

The following reply was made to PR libgcj/6092; it has been noted by GNATS.

From: "Kaveh R. Ghazi" <ghazi@caip.rutgers.edu>
To: ghazi@gcc.gnu.org, tromey@redhat.com
Cc: davem@gcc.gnu.org, gcc-bugs@gcc.gnu.org, gcc-gnats@gcc.gnu.org,
        gcc-prs@gcc.gnu.org, java-prs@gcc.gnu.org
Subject: Re: libgcj/6092: sparc-sun-solaris2.7 gcc-3.1 has hundreds of libjava failures with -m64
Date: Wed, 24 Apr 2002 16:05:47 -0400 (EDT)

  > From: Tom Tromey <tromey@redhat.com>
  > 
  > Kaveh> See -m64 java results here:
  > Kaveh> http://gcc.gnu.org/ml/gcc-testresults/2002-04/msg00875.html
  >     
  > I can't argue with those results.
  > I do wonder why my Solaris 2.7 build still fails.
  > Do the GC tests work for you?
  > (`make check' in the boehm-gc target build directory)
  > Tom
 
 
 The "make check" works for me in both the regular (m32) and sparcv9
 (m64) directories (output listed below.)  I'm not sure what problem
 you are seeing or why.  Can you provide more detail?  Perhaps it makes
 sense to open a separate PR.
 
 in boehm-gc:
  > Completed 3 tests
  > Allocated 5252866 collectable objects
  > Allocated 306 uncollectable objects
  > Allocated 3476354 atomic objects
  > Allocated 32854 stubborn objects
  > Finalized 6622/6622 objects - finalization is probably ok
  > Total number of bytes allocated is 185271172
  > Final heap size is 13213696 bytes
  > Collector appears to work
  > Completed 87 collections
  > PASS: gctest
  > ==================
  > All 1 tests passed
  > ==================
  > make[2]: Leaving directory
  > `/teal/caip5/ghazi/gcc-testing/branch/build/sparc-sun-solaris2.7/boehm-gc'
 
 in sparcv9/boehm-gc:
  > Completed 3 tests
  > Allocated 5124806 collectable objects
  > Allocated 306 uncollectable objects
  > Allocated 3094311 atomic objects
  > Allocated 32895 stubborn objects
  > Finalized 6675/6675 objects - finalization is probably ok
  > Total number of bytes allocated is 230972168
  > Final heap size is 23150592 bytes
  > Collector appears to work
  > Completed 63 collections
  > PASS: gctest
  > ==================
  > All 1 tests passed
  > ==================
  > make[2]: Leaving directory
  > `/teal/caip5/ghazi/gcc-testing/branch/build/sparc-sun-solaris2.7/sparcv9/boehm-gc'
 
 --
 Kaveh R. Ghazi			Director of Systems Architecture
 ghazi@caip.rutgers.edu		Qwest Global Services

Re: libgcj/6092: sparc-sun-solaris2.7 gcc-3.1 has hundreds of libjava failures with -m64

[David S. Miller]

The following reply was made to PR libgcj/6092; it has been noted by GNATS.

From: "David S. Miller" <davem@redhat.com>
To: tromey@redhat.com
Cc: ghazi@gcc.gnu.org, davem@gcc.gnu.org, gcc-bugs@gcc.gnu.org,
        gcc-prs@gcc.gnu.org, ghazi@caip.rutgers.edu, java-prs@gcc.gnu.org,
        gcc-gnats@gcc.gnu.org
Subject: Re: libgcj/6092: sparc-sun-solaris2.7 gcc-3.1 has hundreds of
 libjava failures with -m64
Date: Wed, 24 Apr 2002 12:49:13 -0700 (PDT)

    From: Tom Tromey <tromey@redhat.com>
    Date: 24 Apr 2002 13:58:23 -0600
 
    Kaveh> See -m64 java results here:
    Kaveh> http://gcc.gnu.org/ml/gcc-testresults/2002-04/msg00875.html
        
    I can't argue with those results.
    I do wonder why my Solaris 2.7 build still fails.
    Do the GC tests work for you?
    (`make check' in the boehm-gc target build directory)
 
 With that small number of failures, boehm-gc is working.
 If it isn't working you get at least 140 failures.

Re: libgcj/6092: sparc-sun-solaris2.7 gcc-3.1 has hundreds of libjava failures with -m64

[Tom Tromey]

The following reply was made to PR libgcj/6092; it has been noted by GNATS.

From: Tom Tromey <tromey@redhat.com>
To: ghazi@gcc.gnu.org
Cc: davem@gcc.gnu.org, gcc-bugs@gcc.gnu.org, gcc-prs@gcc.gnu.org,
        ghazi@caip.rutgers.edu, java-prs@gcc.gnu.org, gcc-gnats@gcc.gnu.org
Subject: Re: libgcj/6092: sparc-sun-solaris2.7 gcc-3.1 has hundreds of libjava failures with -m64
Date: 24 Apr 2002 13:58:23 -0600

 Kaveh> See -m64 java results here:
 Kaveh> http://gcc.gnu.org/ml/gcc-testresults/2002-04/msg00875.html
     
 I can't argue with those results.
 I do wonder why my Solaris 2.7 build still fails.
 Do the GC tests work for you?
 (`make check' in the boehm-gc target build directory)
 
 Tom

Re: libgcj/6092: sparc-sun-solaris2.7 gcc-3.1 has hundreds of libjava failures with -m64

[ghazi]

Synopsis: sparc-sun-solaris2.7 gcc-3.1 has hundreds of libjava failures with -m64

State-Changed-From-To: analyzed->closed
State-Changed-By: ghazi
State-Changed-When: Wed Apr 24 12:41:06 2002
State-Changed-Why:
    Fixed by:
    http://gcc.gnu.org/ml/gcc-patches/2002-04/msg01245.html
    
    See -m64 java results here:
    http://gcc.gnu.org/ml/gcc-testresults/2002-04/msg00875.html
    

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=6092

Re: libgcj/6092: sparc-sun-solaris2.7 gcc-3.1 has hundreds of libjava failures with -m64

[Tom Tromey]

The following reply was made to PR libgcj/6092; it has been noted by GNATS.

From: Tom Tromey <tromey@redhat.com>
To: "David S. Miller" <davem@redhat.com>
Cc: davem@gcc.gnu.org, gcc-gnats@gcc.gnu.org
Subject: Re: libgcj/6092: sparc-sun-solaris2.7 gcc-3.1 has hundreds of  libjava failures with -m64
Date: 19 Apr 2002 19:38:58 -0600

 >>>>> "David" == David S Miller <davem@redhat.com> writes:
 
 David> If it is a GC problem, is it Solaris specific.
 
 That's right.
 
 David> I bet it is a non-working definition of STACKBOTTOM or similar.
 David> The CPP_WORDSZ protection of DYNAMIC_LOADING for Solaris in
 David> gcconfig.h seems suspicious too.
 
 Thanks.  I've played with it a little but haven't seriously debugged
 it yet.  Hans has given a lot of good advice though.
 
 David> As Richard has told me numerous times, we need some description
 David> language for this stuff so we only need to do it in one file.
 
 I completely agree.  Right now libffi has two APIs.  Sometimes people
 write one but not the other.  For instance there is no Sparc port of
 the raw API; this means the libgcj bytecode interpreter doesn't work
 there.  I'd love it if doing a base gcc port meant that all of libffi
 (and thus the hardest-to-port parts of libgcj) were done
 automatically.
 
 Tom

Re: libgcj/6092: sparc-sun-solaris2.7 gcc-3.1 has hundreds of libjava failures with -m64

[Tom Tromey]

The following reply was made to PR libgcj/6092; it has been noted by GNATS.

From: Tom Tromey <tromey@redhat.com>
To: rth@gcc.gnu.org
Cc: davem@gcc.gnu.org, gcc-bugs@gcc.gnu.org, gcc-prs@gcc.gnu.org,
        ghazi@caip.rutgers.edu, java-prs@gcc.gnu.org, nobody@gcc.gnu.org,
        gcc-gnats@gcc.gnu.org
Subject: Re: libgcj/6092: sparc-sun-solaris2.7 gcc-3.1 has hundreds of libjava failures with -m64
Date: 19 Apr 2002 16:59:21 -0600

 >>>>> "rth" == rth  <rth@gcc.gnu.org> writes:
 
 rth>     Dave sez libffi for sparc64 is borked.
     
 This is a GC problem, not a libffi problem.  The boehm-gc gctest
 program crashes in the same way as every java program; also a libffi
 problem wouldn't show itself as a systematic failure of all the
 libjava execute tests (most programs never use libffi).
 
 rth>     I disbelieve this is a regression, so I've downgraded
 rth>     priority to medium.
 
 I don't think it is a regression either.  However, that leaves us in
 the odd situation where we build a compiler where the -m32 libraries
 work and the -m64 ones do not.  This is why I had left it as high
 priority.  That said, it isn't clear I'll actually be able to finish
 debugging this before the deadline, either.
 
 Tom

Re: libgcj/6092: sparc-sun-solaris2.7 gcc-3.1 has hundreds of libjava failures with -m64

[rth]

Synopsis: sparc-sun-solaris2.7 gcc-3.1 has hundreds of libjava failures with -m64

Responsible-Changed-From-To: unassigned->davem
Responsible-Changed-By: rth
Responsible-Changed-When: Fri Apr 19 11:54:29 2002
Responsible-Changed-Why:
    Sucker.  ;-)
State-Changed-From-To: open->analyzed
State-Changed-By: rth
State-Changed-When: Fri Apr 19 11:54:29 2002
State-Changed-Why:
    Dave sez libffi for sparc64 is borked.
    
    I disbelieve this is a regression, so I've downgraded
    priority to medium.

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=6092