--- Comment #16 from howarth at nitro dot med dot uc dot edu 2009-10-01
15:41 ---
It looks like this problem isn't strictly target specific (although we are
the only target currently impacted). If TARGET_ASM_EMIT_UNWIND_LABEL as defined
needs to support passing SECOND, both the
--- Comment #17 from howarth at nitro dot med dot uc dot edu 2009-10-01
15:55 ---
(In reply to comment #15)
Yeah, the patch in #11 is about the right for half the problem (darwin10), if
it weren't for ld's warning message. I don't know quite why it is doing that,
so guess we'll
--- Comment #18 from mrs at apple dot com 2009-10-01 19:14 ---
Yes. If someone wants to propose a better solution, they will...
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=41313
--- Comment #19 from howarth at nitro dot med dot uc dot edu 2009-10-01
23:54 ---
I'm not sure that hot_cold=0 would work as a grep of gcc trunk doesn't show any
thing matching hot_cold. I did find the original commit of the problematic
code...
--- Comment #20 from mrs at apple dot com 2009-10-02 02:03 ---
flag_reorder_blocks_and_partition was the complete name of the variable last I
looked. If flag_unwind_tables is true and this variable is true, that's the
bug.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=41313
--- Comment #21 from howarth at nitro dot med dot uc dot edu 2009-10-02
02:36 ---
We seemed to have changed the original code from gcc 4.2.1...
/* If user requested unwind info, then turn off the partitioning
optimization. */
if (flag_unwind_tables !
--- Comment #22 from howarth at nitro dot med dot uc dot edu 2009-10-02
02:44 ---
This change came from...
Author: jakub
Date: Fri Aug 7 06:23:42 2009
New Revision: 150553
URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=150553
Log:
* dwarf2out.c (output_fde): When doing
--- Comment #23 from howarth at nitro dot med dot uc dot edu 2009-10-02
05:00 ---
Oddly...
Index: config/darwin.c
===
--- config/darwin.c (revision 152389)
+++ config/darwin.c (working copy)
@@ -1697,6 +1697,9 @@
--- Comment #24 from howarth at nitro dot med dot uc dot edu 2009-10-02
05:11 ---
Oh, I didn't regress out all of the changes from...
http://gcc.gnu.org/viewcvs/trunk/gcc/opts.c?r1=150553r2=150552view=patchpathrev=150553
when I do that, the tree-prof tests pass fine...
Native
--- Comment #25 from howarth at nitro dot med dot uc dot edu 2009-10-02
05:34 ---
The following change achieves the same test results as in Comment 24...
--- /Users/howarth/gcc/gcc/opts.c 2009-09-28 19:33:18.0 -0400
+++ opts.c 2009-10-02 01:30:48.0 -0400
@@
--- Comment #26 from howarth at nitro dot med dot uc dot edu 2009-10-02
05:40 ---
I should note that if I don't wrapper the line ...
(USING_SJLJ_EXCEPTIONS
with the #if !defined(__MACH__), I get the following results...
=== g++ tests ===
Schedule of
--- Comment #27 from howarth at nitro dot med dot uc dot edu 2009-10-02
05:44 ---
I should note that with the proposed change in Comment 25, the 2 unsupported
tests properly show up in g++.log as...
cc1plus: note: -freorder-blocks-and-partition does not work with exceptions on
this
--- Comment #9 from howarth at nitro dot med dot uc dot edu 2009-09-30
20:26 ---
(In reply to comment #8)
darwin has -fpic by default.
Oddly, if I build gcc trunk on x86_64 Fedora 10 with CPPFLAGS set to -fPIC and
run the testsuite with -fPIC, I don't see the .eh symbols...
grep
--- Comment #10 from mrs at apple dot com 2009-09-30 20:54 ---
When targeting darwin10, the .eh symbols don't have to be generated anymore.
They are from darwin.c, darwin_emit_unwind_label.
if (darwin_macosx_version_min
strverscmp (darwin_macosx_version_min, 10.6) = 0)
can be
--- Comment #11 from howarth at nitro dot med dot uc dot edu 2009-09-30
22:36 ---
Mike,
I am a fuzzy on the best way to avoid darwin_emit_unwind_label since it is
assigned to a define in darwin.h. The only thing that makes sense to me is...
Index: gcc/config/darwin.c
--- Comment #12 from jakub at gcc dot gnu dot org 2009-09-30 22:46 ---
It is unclear what those labels are good for, but if darwin is to support
hot/cold partitioning and FDEs covering it, the emit unwind_label hook (which
is apparently darwin specific only) should be passed also the
--- Comment #13 from howarth at nitro dot med dot uc dot edu 2009-10-01
00:18 ---
Actually, the patch in Comment 11 leaves two of the original failures...
FAIL: g++.dg/tree-prof/partition1.C compilation, -g -fprofile-use
FAIL: g++.dg/tree-prof/partition1.C compilation, -O3 -g
--- Comment #14 from howarth at nitro dot med dot uc dot edu 2009-10-01
00:29 ---
Jakub,
If you could propose a patch to the darwin_emit_unwind_label subroutine in
gcc/config/darwin.c to implement the changes you propose in Comment 12, I would
be happy to test it on
--- Comment #15 from mrs at apple dot com 2009-10-01 00:54 ---
Yeah, the patch in #11 is about the right for half the problem (darwin10), if
it weren't for ld's warning message. I don't know quite why it is doing that,
so guess we'll just have to nix that idea. So, the next best thing
--- Comment #5 from howarth at nitro dot med dot uc dot edu 2009-09-26
18:14 ---
Jakub,
Can you check what --save-temps produces for g++.dg/tree-prof/partition1.C
in partition1.s on linux? Does it produce two sets of global statements as
well?
grep __Z3bari.eh partition1.S
.globl
--- Comment #6 from dominiq at lps dot ens dot fr 2009-09-26 19:10 ---
I also see similar failures on i686-apple-darwin9 in 32-bit mode (but not with
-m64) for
gcc.dg/tree-prof/bb-reorg.c and gcc.dg/tree-prof/pr34999.c:
output is:
/var/tmp//cc3DMvjA.s:230:FATAL:Symbol _foo.eh already
--- Comment #7 from hjl dot tools at gmail dot com 2009-09-26 20:20 ---
(In reply to comment #5)
Jakub,
Can you check what --save-temps produces for g++.dg/tree-prof/partition1.C
in partition1.s on linux? Does it produce two sets of global statements as
well?
grep __Z3bari.eh
--- Comment #8 from rguenth at gcc dot gnu dot org 2009-09-26 20:43 ---
darwin has -fpic by default.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=41313
--- Comment #1 from howarth at nitro dot med dot uc dot edu 2009-09-10
03:49 ---
Created an attachment (id=18558)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=18558action=view)
preprocessed source for g++.dg/tree-prof/partition1.C compilation -g
-fprofile-use
--
--- Comment #2 from howarth at nitro dot med dot uc dot edu 2009-09-10
03:50 ---
Created an attachment (id=18559)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=18559action=view)
assembly file for g++.dg/tree-prof/partition1.C compilation -g -fprofile-use
--
--- Comment #3 from howarth at nitro dot med dot uc dot edu 2009-09-10
03:51 ---
Created an attachment (id=18560)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=18560action=view)
gcda file for g++.dg/tree-prof/partition1.C compilation -g -fprofile-use
--
--- Comment #4 from howarth at nitro dot med dot uc dot edu 2009-09-10
03:52 ---
Attached files were generated under r151584 with the command...
/sw/src/fink.build/gcc45-4.4.999-20090909/darwin_objdir/gcc/testsuite/g++/../../g++
27 matches
Mail list logo
