From bugzilla-daemon at cs.uiuc.edu  Wed Oct  1 09:09:11 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 1 Oct 2008 09:09:11 -0500
Subject: [LLVMbugs] [Bug 2846] New: LLVM does not build (Linux AMD64 GCC
	3.4.6)
Message-ID: <bug-2846-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2846

           Summary: LLVM does not build (Linux AMD64 GCC 3.4.6)
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: blocker
          Priority: P2
         Component: Core LLVM classes
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: Axel.Naumann at cern.ch
                CC: llvmbugs at cs.uiuc.edu


I tried to build LLVM with GCC 3.4.6 on a Linux 64bit installation, today's
trunk. Relevant cmake flags of the out-of-source-build:
CMAKE_BUILD_TYPE: RelWithDebInfo
LLVM_TARGETS_TO_BUILD: CBackend;CppBackend;X86

Here are the issues I encountered:

/build/axel/build/llvm/src/llvm_svn/lib/VMCore/AsmWriter.cpp: In member
function `int <unnamed>::SlotTracker::getGlobalSlot(const llvm::GlobalValue*)':
/build/axel/build/llvm/src/llvm_svn/lib/VMCore/AsmWriter.cpp:388: warning:
converting of negative value `-0x00000000000000001' to `unsigned int'
/build/axel/build/llvm/src/llvm_svn/lib/VMCore/AsmWriter.cpp: In member
function `int <unnamed>::SlotTracker::getLocalSlot(const llvm::Value*)':
/build/axel/build/llvm/src/llvm_svn/lib/VMCore/AsmWriter.cpp:400: warning:
converting of negative value `-0x00000000000000001' to `unsigned int'

/build/axel/build/llvm/src/llvm_svn/lib/CodeGen/SelectionDAG/SelectionDAG.cpp:
In function `llvm::SDValue getMemcpyLoadsAndStores(llvm::SelectionDAG&,
llvm::SDValue, llvm::SDValue, llvm::SDValue, uint64_t, unsigned int, bool,
const llvm::Value*, uint64_t, const llvm::Value*, uint64_t)':
/build/axel/build/llvm/src/llvm_svn/lib/CodeGen/SelectionDAG/SelectionDAG.cpp:2929:
warning: converting of negative value `-0x00000000000000001' to `uint64_t'
/build/axel/build/llvm/src/llvm_svn/lib/CodeGen/SelectionDAG/SelectionDAG.cpp:
In function `llvm::SDValue getMemmoveLoadsAndStores(llvm::SelectionDAG&,
llvm::SDValue, llvm::SDValue, llvm::SDValue, uint64_t, unsigned int, bool,
const llvm::Value*, uint64_t, const llvm::Value*, uint64_t)':
/build/axel/build/llvm/src/llvm_svn/lib/CodeGen/SelectionDAG/SelectionDAG.cpp:2987:
warning: converting of negative value `-0x00000000000000001' to `uint64_t'

/build/axel/build/llvm/src/llvm_svn/lib/Target/X86/X86InstrInfo.cpp: In
constructor `llvm::X86InstrInfo::X86InstrInfo(llvm::X86TargetMachine&)':
/build/axel/build/llvm/src/llvm_svn/lib/Target/X86/X86InstrInfo.cpp:212:
warning: cast to pointer from integer of different size
/build/axel/build/llvm/src/llvm_svn/lib/Target/X86/X86InstrInfo.cpp:216:
warning: cast to pointer from integer of different size
/build/axel/build/llvm/src/llvm_svn/lib/Target/X86/X86InstrInfo.cpp:303:
warning: cast to pointer from integer of different size
/build/axel/build/llvm/src/llvm_svn/lib/Target/X86/X86InstrInfo.cpp:310:
warning: cast to pointer from integer of different size
/build/axel/build/llvm/src/llvm_svn/lib/Target/X86/X86InstrInfo.cpp:431:
warning: cast to pointer from integer of different size
/build/axel/build/llvm/src/llvm_svn/lib/Target/X86/X86InstrInfo.cpp:436:
warning: cast to pointer from integer of different size
/build/axel/build/llvm/src/llvm_svn/lib/Target/X86/X86InstrInfo.cpp:638:
warning: cast to pointer from integer of different size
/build/axel/build/llvm/src/llvm_svn/lib/Target/X86/X86InstrInfo.cpp:642:
warning: cast to pointer from integer of different size
/build/axel/build/llvm/src/llvm_svn/lib/Target/X86/X86InstrInfo.cpp: In member
function `llvm::MachineInstr*
llvm::X86InstrInfo::foldMemoryOperand(llvm::MachineFunction&,
llvm::MachineInstr*, unsigned int, llvm::SmallVector<llvm::MachineOperand,
4u>&) const':
/build/axel/build/llvm/src/llvm_svn/lib/Target/X86/X86InstrInfo.cpp:1989:
warning: cast to pointer from integer of different size
/build/axel/build/llvm/src/llvm_svn/lib/Target/X86/X86InstrInfo.cpp: In member
function `virtual bool
llvm::X86InstrInfo::canFoldMemoryOperand(llvm::MachineInstr*,
llvm::SmallVectorImpl<unsigned int>&) const':
/build/axel/build/llvm/src/llvm_svn/lib/Target/X86/X86InstrInfo.cpp:2156:
warning: cast to pointer from integer of different size
/build/axel/build/llvm/src/llvm_svn/lib/Target/X86/X86InstrInfo.cpp: In member
function `virtual bool
llvm::X86InstrInfo::unfoldMemoryOperand(llvm::MachineFunction&,
llvm::MachineInstr*, unsigned int, bool, bool,
llvm::SmallVectorImpl<llvm::MachineInstr*>&) const':
/build/axel/build/llvm/src/llvm_svn/lib/Target/X86/X86InstrInfo.cpp:2167:
warning: cast to pointer from integer of different size
/build/axel/build/llvm/src/llvm_svn/lib/Target/X86/X86InstrInfo.cpp: In member
function `virtual bool
llvm::X86InstrInfo::unfoldMemoryOperand(llvm::SelectionDAG&, llvm::SDNode*,
llvm::SmallVectorImpl<llvm::SDNode*>&) const':
/build/axel/build/llvm/src/llvm_svn/lib/Target/X86/X86InstrInfo.cpp:2273:
warning: cast to pointer from integer of different size
/build/axel/build/llvm/src/llvm_svn/lib/Target/X86/X86InstrInfo.cpp: In member
function `virtual unsigned int
llvm::X86InstrInfo::getOpcodeAfterMemoryUnfold(unsigned int, bool, bool)
const':
/build/axel/build/llvm/src/llvm_svn/lib/Target/X86/X86InstrInfo.cpp:2351:
warning: cast to pointer from integer of different size

/build/axel/build/llvm/src/llvm_svn/lib/Target/X86/X86TargetAsmInfo.cpp:41:
error: duplicate explicit instantiation of `bool
llvm::X86TargetAsmInfo<BaseTAI>::ExpandInlineAsm(llvm::CallInst*) const [with
BaseTAI = llvm::TargetAsmInfo]'
/build/axel/build/llvm/src/llvm_svn/lib/Target/X86/X86TargetAsmInfo.cpp:43:
error: duplicate explicit instantiation of `bool
llvm::X86TargetAsmInfo<BaseTAI>::LowerToBSwap(llvm::CallInst*) const [with
BaseTAI = llvm::TargetAsmInfo]'
gmake[2]: ***
[lib/Target/X86/CMakeFiles/LLVMX86CodeGen.dir/X86TargetAsmInfo.cpp.o] Error 1
gmake[2]: *** Waiting for unfinished jobs....

Could the compilation warnings and especially the error be fixed? Thanks!

Cheers, Axel.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct  1 13:32:19 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 1 Oct 2008 13:32:19 -0500
Subject: [LLVMbugs] [Bug 2847] New: make check failures in r56923
Message-ID: <bug-2847-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2847

           Summary: make check failures in r56923
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


With SVN r56923, on x86_64 Linux:

FAIL: /home/edwin/llvm-svn/llvm/test/Bindings/Ocaml/analysis.ml
Failed with unknown error (or has stderr output) at line 1
while running: /usr/bin/ocamlc.opt -cc g++ -I
/home/edwin/llvm-svn/llvm/Release/lib/ocaml -warn-error A llvm.cma
llvm_analysis.cma /home/edwin/llvm-svn/llvm/test/Bindings/Ocaml/analysis.ml -o
analysis.ml.tmp

FAIL: /home/edwin/llvm-svn/llvm/test/Bindings/Ocaml/bitreader.ml
Failed with unknown error (or has stderr output) at line 1
while running: /usr/bin/ocamlc.opt -cc g++ -I
/home/edwin/llvm-svn/llvm/Release/lib/ocaml -warn-error A llvm.cma
llvm_bitreader.cma llvm_bitwriter.cma
/home/edwin/llvm-svn/llvm/test/Bindings/Ocaml/bitreader.ml -o bitreader.ml.tmp

FAIL: /home/edwin/llvm-svn/llvm/test/Bindings/Ocaml/bitwriter.ml
Failed with unknown error (or has stderr output) at line 1
while running: /usr/bin/ocamlc.opt -cc g++ -I
/home/edwin/llvm-svn/llvm/Release/lib/ocaml -warn-error A llvm.cma
llvm_bitwriter.cma /home/edwin/llvm-svn/llvm/test/Bindings/Ocaml/bitwriter.ml
-o bitwriter.ml.tmp

FAIL: /home/edwin/llvm-svn/llvm/test/Bindings/Ocaml/executionengine.ml
Failed with unknown error (or has stderr output) at line 1
while running: /usr/bin/ocamlc.opt -cc g++ -I
/home/edwin/llvm-svn/llvm/Release/lib/ocaml -warn-error A llvm.cma
llvm_target.cma llvm_executionengine.cma
/home/edwin/llvm-svn/llvm/test/Bindings/Ocaml/executionengine.ml -o
executionengine.ml.tmp

FAIL: /home/edwin/llvm-svn/llvm/test/Bindings/Ocaml/scalar_opts.ml
Failed with unknown error (or has stderr output) at line 1
while running: /usr/bin/ocamlc.opt -cc g++ -I
/home/edwin/llvm-svn/llvm/Release/lib/ocaml -warn-error A llvm.cma
llvm_scalar_opts.cma llvm_target.cma
/home/edwin/llvm-svn/llvm/test/Bindings/Ocaml/scalar_opts.ml -o
scalar_opts.ml.tmp

FAIL: /home/edwin/llvm-svn/llvm/test/Bindings/Ocaml/target.ml
Failed with unknown error (or has stderr output) at line 1
while running: /usr/bin/ocamlc.opt -cc g++ -I
/home/edwin/llvm-svn/llvm/Release/lib/ocaml -warn-error A llvm.cma
llvm_target.cma /home/edwin/llvm-svn/llvm/test/Bindings/Ocaml/target.ml -o
target.ml.tmp

FAIL: /home/edwin/llvm-svn/llvm/test/Bindings/Ocaml/vmcore.ml
Failed with unknown error (or has stderr output) at line 1
while running: /usr/bin/ocamlc.opt -cc g++ -I
/home/edwin/llvm-svn/llvm/Release/lib/ocaml -warn-error A llvm.cma
llvm_analysis.cma llvm_bitwriter.cma /home/edwin
/llvm-svn/llvm/test/Bindings/Ocaml/vmcore.ml -o vmcore.ml.tmp


FAIL: /home/edwin/llvm-svn/llvm/test/CodeGen/Generic/i128-and-beyond.ll
Failed with exit(1) at line 2
while running: llvm-as <
/home/edwin/llvm-svn/llvm/test/CodeGen/Generic/i128-and-beyond.ll | llc
-march=ppc32 | grep 4294967295 | count 28
count: expected 28 lines and got 0.
llc: for the -march option: : Cannot find option named 'ppc32'!

FAIL: /home/edwin/llvm-svn/llvm/test/FrontendC/2008-01-25-ByValReadNone.c
Failed with exit(1) at line 1
while running: /usr/local/bin/llvm-gcc  -emit-llvm -w -O3 -S -o - -emit-llvm
/home/edwin/llvm-svn/llvm/test/FrontendC/2008-01-25-ByValReadNone.c | not grep
readonly
define i32 @h(%struct.S* byval %x) nounwind readonly {
        %0 = tail call i32 @g(%struct.S* byval %x) nounwind readonly           
; <i32> [#uses=1]
declare i32 @g(%struct.S* byval) readonly
child process exited abnormally

FAIL: /home/edwin/llvm-svn/llvm/test/FrontendC/Atomics-no64bit.c
Failed with exit(1) at line 1
while running: /usr/local/bin/llvm-gcc  -emit-llvm -w -S
/home/edwin/llvm-svn/llvm/test/FrontendC/Atomics-no64bit.c -o - | grep atomic |
count 192
count: expected 192 lines and got 200.
child process exited abnormally

FAIL: /home/edwin/llvm-svn/llvm/test/FrontendC/Atomics.c
Failed with exit(1) at line 4
while running: /usr/local/bin/llvm-gcc  -emit-llvm -w -S
/home/edwin/llvm-svn/llvm/test/FrontendC/Atomics.c -o - | grep p0i32 | count 92
count: expected 92 lines and got 50.
child process exited abnormally

===  Summary ===

# of expected passes            2676
# of unexpected failures        11
# of expected failures          9

I have:
$ g++ -v
Using built-in specs.
Target: x86_64-linux-gnu
Configured with: ../src/configure -v --with-pkgversion='Debian 4.3.2-1'
--with-bugurl=file:///usr/share/doc/gcc-4.3/README.Bugs
--enable-languages=c,c++,fortran,objc,obj-c++ --prefix=/usr --enable-shared
--with-system-zlib --libexecdir=/usr/lib --without-included-gettext
--enable-threads=posix --enable-nls --with-gxx-include-dir=/usr/include/c++/4.3
--program-suffix=-4.3 --enable-clocale=gnu --enable-libstdcxx-debug
--enable-objc-gc --enable-mpfr --enable-cld --enable-checking=release
--build=x86_64-linux-gnu --host=x86_64-linux-gnu --target=x86_64-linux-gnu
Thread model: posix
gcc version 4.3.2 (Debian 4.3.2-1)

$ llvm-gcc -v
Using built-in specs.
Target: x86_64-unknown-linux-gnu
Configured with: ../llvm-gcc4.2/configure
--prefix=/home/edwin/llvm-svn/obj42/../install --program-prefix=llvm-
--enable-llvm=/home/edwin/llvm-svn/llvm/ --disable-multilib
--enable-checking=release --enable-languages=c,c++ --disable-bootstrap
Thread model: posix
gcc version 4.2.1 (Based on Apple Inc. build 5623) (LLVM build)

$ ocamlopt -v
The Objective Caml native-code compiler, version 3.10.2
Standard library directory: /usr/lib/ocaml/3.10.2

I configured LLVM with:
./configure --enable-targets=x86,cbe,cpp --enable-optimized

Let me know if you need some further info.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct  1 21:42:37 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 1 Oct 2008 21:42:37 -0500
Subject: [LLVMbugs] [Bug 2848] New: new volatile regression
Message-ID: <bug-2848-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2848

           Summary: new volatile regression
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu


Seen on r56938 on Ubuntu Hardy on ia32.

This is a strange one: wrong code at -O0, correct at other optimization levels.

I don't think this problem was there 24 hours ago.

regehr at john-home:~/volatile/tmp50$ llvm-gcc -O0 small.c -S -o - | grep g_126
        .type   g_126, at object
        .section        .gnu.linkonce.b.g_126,"aw", at nobits
        .comm   g_126,4,4               # g_126

The x86 code contains no load from the volatile location!  The LLVM looks OK.



int mod_rhs (int x)
{
  return x;
}

volatile unsigned int g_126;

int func_93 (void);
int func_93 (void)
{
  return (g_126 > (mod_rhs (1)) < 0);
}

int main (void)
{
  func_93 ();
  return 0;
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct  1 23:16:32 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 1 Oct 2008 23:16:32 -0500
Subject: [LLVMbugs] [Bug 2849] New: llc fails an assertion in ScheduleDAG
Message-ID: <bug-2849-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2849

           Summary: llc fails an assertion in ScheduleDAG
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Keywords: ABI
          Severity: normal
          Priority: P2
         Component: llc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: santosh.nagarakatte at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2048)
 --> (http://llvm.org/bugs/attachment.cgi?id=2048)
bitcode file of the spec 2006 lbm benchmark with the custom transformation
causing llc to fail

llc fails an assertion in the ScheduleDAG at line number 885. The following
assert is not satisfied.
Assertion '0 && This target-independent node should have been selected'

I am using llvm-2.3 with svn revision number 53702
and llvm-gcc with svn revision number 53705.
Host machine is x86-64 base and run suse linux.
llvm is compiled using gcc-4.2.0

I have attached the reduced bitcode generated by bugpoint, for the spec 2006
benchmark 470.lbm with a custom transformation run through opt. The bitcode has
been verified after the custom transformation using -verify-each argument to
opt.


Santosh


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct  2 09:57:53 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 2 Oct 2008 09:57:53 -0500
Subject: [LLVMbugs] [Bug 2850] New: MOVDDUP instruction selection asserts
Message-ID: <bug-2850-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2850

           Summary: MOVDDUP instruction selection asserts
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicolas at capens.net
                CC: llvmbugs at cs.uiuc.edu


I'm hitting an assert "Cannot BIT_CONVERT between types of different sizes!"
with the following call stack:

SelectionDAG::getNode()  Line 2218
CanonicalizeMovddup()  Line 2994
X86TargetLowering::LowerVECTOR_SHUFFLE()  Line 3948
X86TargetLowering::LowerOperation()  Line 6039
...

This is with revision 56944 (HEAD at the moment of posting).

Removing the lines in X86ISelLowering.cpp that "Canonicalize movddup shuffles"
fixes the issue (but obviously prevents the selection of movddup when desirable
as well).


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct  2 10:08:12 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 2 Oct 2008 10:08:12 -0500
Subject: [LLVMbugs] [Bug 2848] new volatile regression
In-Reply-To: <bug-2848-206@http.llvm.org/bugs/>
Message-ID: <200810021508.m92F8CHx008323@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2848


Dan Gohman <gohman at apple.com> changed:

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




--- Comment #2 from Dan Gohman <gohman at apple.com>  2008-10-02 10:08:11 ---
Confirmed, it was fast-isel. Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080929/067919.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct  2 10:38:30 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 2 Oct 2008 10:38:30 -0500
Subject: [LLVMbugs] [Bug 2851] New: libiberty fails to build with GCC 4.3.1
Message-ID: <bug-2851-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2851

           Summary: libiberty fails to build with GCC 4.3.1
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: blocker
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: Axel.Naumann at cern.ch
                CC: llvmbugs at cs.uiuc.edu


I hope this is the right product - please move as needed.

I want to build llvm-gcc-4.2 with GCC 4.3.1 AMD64. I ran

CXX=`which g++` ../../src/llvm-gcc-4.2_svn/configure --prefix=/build/axel/local
--enable-checking --program-prefix=llvm-
--enable-llvm=/build/axel/build/llvm/obj/llvm-configure
--enable-languages=c,c++ --disable-bootstrap


Compilation with
  make LLVM_VERSION_INFO=trunk-20081001-axel
stops here:

/build/axel/build/llvm/obj/llvm-gcc-4.2/./prev-gcc/xgcc
-B/build/axel/build/llvm/obj/llvm-gcc-4.2/./prev-gcc/
-B/build/axel/local/x86_64-unknown-linux-gnu/bin/ -c -DHAVE_CONFIG_H -g -O2 -I.
-I../../../src/llvm-gcc-4.2_svn/libiberty/../include  -W -Wall -pedantic
-Wwrite-strings -Wstrict-prototypes -Wc++-compat
../../../src/llvm-gcc-4.2_svn/libiberty/cplus-dem.c -o cplus-dem.o
../../../src/llvm-gcc-4.2_svn/libiberty/cplus-dem.c:55: warning: function
declaration isn't a prototype
../../../src/llvm-gcc-4.2_svn/libiberty/cplus-dem.c:55: error: conflicting
types for 'malloc'
../../../src/llvm-gcc-4.2_svn/libiberty/cplus-dem.c:56: warning: function
declaration isn't a prototype
../../../src/llvm-gcc-4.2_svn/libiberty/cplus-dem.c: In function
'code_for_qualifier':
../../../src/llvm-gcc-4.2_svn/libiberty/cplus-dem.c:582: warning: implicit
declaration of function 'abort'
../../../src/llvm-gcc-4.2_svn/libiberty/cplus-dem.c:582: warning: incompatible
implicit declaration of built-in function 'abort'
../../../src/llvm-gcc-4.2_svn/libiberty/cplus-dem.c: In function
'qualifier_string':
../../../src/llvm-gcc-4.2_svn/libiberty/cplus-dem.c:622: warning: incompatible
implicit declaration of built-in function 'abort'
../../../src/llvm-gcc-4.2_svn/libiberty/cplus-dem.c: In function
'squangle_mop_up':
../../../src/llvm-gcc-4.2_svn/libiberty/cplus-dem.c:1082: warning: implicit
declaration of function 'free'
../../../src/llvm-gcc-4.2_svn/libiberty/cplus-dem.c: In function
'demangle_qualified':
../../../src/llvm-gcc-4.2_svn/libiberty/cplus-dem.c:3177: warning: implicit
declaration of function 'atoi'
make[3]: *** [cplus-dem.o] Error 1
make[3]: Leaving directory `/build/axel/build/llvm/obj/llvm-gcc-4.2/libiberty'
make[2]: *** [all-stage2-libiberty] Error 2
make[2]: Leaving directory `/build/axel/build/llvm/obj/llvm-gcc-4.2'
make[1]: *** [stage2-bubble] Error 2
make[1]: Leaving directory `/build/axel/build/llvm/obj/llvm-gcc-4.2'
make: *** [all] Error 2

What am I doing wrong? Or could these functions' declarations be added?


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct  2 11:54:28 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 2 Oct 2008 11:54:28 -0500
Subject: [LLVMbugs] [Bug 2851] libiberty fails to build with GCC 4.3.1
In-Reply-To: <bug-2851-206@http.llvm.org/bugs/>
Message-ID: <200810021654.m92GsShF012013@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2851


Anton Korobeynikov <asl at math.spbu.ru> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |asl at math.spbu.ru
             Status|NEW                         |RESOLVED
         Resolution|                            |WORKSFORME




--- Comment #1 from Anton Korobeynikov <asl at math.spbu.ru>  2008-10-02 11:54:27 ---
This was llvm-gcc bug, not your native gcc.
The fix was commited todays morning in r56946


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct  3 00:37:24 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 3 Oct 2008 00:37:24 -0500
Subject: [LLVMbugs] [Bug 2852] New: fold load (strchr/memchr ptr, c) --> c
Message-ID: <bug-2852-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2852

           Summary: fold load (strchr/memchr ptr, c) --> c
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Scalar Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicholas at mxc.ca
                CC: llvmbugs at cs.uiuc.edu


The pointer returned by the strchr/memchr family is usually tested for validity
(null-ness) and then dereferenced, possibly as part of a loop.

We know the value that the pointer points to; it's the one passed in to as the
second parameter to the function call.

There's no LLVM intrinsic for memchr though. Where should this sort of thing be
done? In simplify-libcalls?


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct  3 09:43:20 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 3 Oct 2008 09:43:20 -0500
Subject: [LLVMbugs] [Bug 2853] New: __gxx_personality_v0 causes overflow in
 R_X86_64_PC32 relocation
Message-ID: <bug-2853-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2853

           Summary: __gxx_personality_v0 causes overflow in R_X86_64_PC32
                    relocation
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: minor
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: Axel.Naumann at cern.ch
                CC: llvmbugs at cs.uiuc.edu


LLVM from the trunk from Oct 1, llvm-gcc-4.2's trunk from today, built with GCC
4.3.1 on an AMD64 bit linux. When building the source

int main(int, char*[])
{
}

with llvm-g++ test.cxx

I see 44 lines of output:
/build/axel/local/libexec/gcc/x86_64-unknown-linux-gnu/4.2.1/cc1plus: Symbol
`__gxx_personality_v0' causes overflow in R_X86_64_PC32 relocation

The binary still works despite the message. Some more info on the build:

llvm-g++ --verbose test.cxx
Using built-in specs.
Target: x86_64-unknown-linux-gnu
Configured with: ../../src/llvm-gcc-4.2_svn/configure
--prefix=/build/axel/local --enable-checking --program-prefix=llvm-
--enable-llvm=/build/axel/build/llvm/obj/llvm-configure --disable-bootstrap
--enable-languages=c,c++
Thread model: posix
gcc version 4.2.1 (Based on Apple Inc. build 5623) (LLVM build)
 /build/axel/local/libexec/gcc/x86_64-unknown-linux-gnu/4.2.1/cc1plus -quiet -v
-D_GNU_SOURCE test.cxx -quiet -dumpbase test.cxx -mtune=generic -auxbase test
-version -o /tmp/axel/ccn2jNGn.s
/build/axel/local/libexec/gcc/x86_64-unknown-linux-gnu/4.2.1/cc1plus: Symbol
`__gxx_personality_v0' causes overflow in R_X86_64_PC32 relocation
[last line repeated 43 times, i.e. it shows up 44 times in total]
ignoring nonexistent directory
"/build/axel/local/lib/gcc/x86_64-unknown-linux-gnu/4.2.1/../../../../x86_64-unknown-linux-gnu/include"
#include "..." search starts here:
#include <...> search starts here:

/build/axel/local/lib/gcc/x86_64-unknown-linux-gnu/4.2.1/../../../../include/c++/4.2.1

/build/axel/local/lib/gcc/x86_64-unknown-linux-gnu/4.2.1/../../../../include/c++/4.2.1/x86_64-unknown-linux-gnu

/build/axel/local/lib/gcc/x86_64-unknown-linux-gnu/4.2.1/../../../../include/c++/4.2.1/backward
 /usr/local/include
 /build/axel/local/include
 /build/axel/local/lib/gcc/x86_64-unknown-linux-gnu/4.2.1/include
 /usr/include
End of search list.
GNU C++ version 4.2.1 (Based on Apple Inc. build 5623) (LLVM build)
(x86_64-unknown-linux-gnu)
        compiled by GNU C version 4.3.1.
GGC heuristics: --param ggc-min-expand=100 --param ggc-min-heapsize=8192
Compiler executable checksum: 2c8907e8ae02f677319d96cae46b99d5
 as -V -Qy -o /tmp/axel/ccqvfExV.o /tmp/axel/ccn2jNGn.s
GNU assembler version 2.15.92.0.2 (x86_64-redhat-linux) using BFD version
2.15.92.0.2 20040927
 /build/axel/local/libexec/gcc/x86_64-unknown-linux-gnu/4.2.1/collect2
--eh-frame-hdr -m elf_x86_64 -dynamic-linker /lib64/ld-linux-x86-64.so.2
/usr/lib/../lib64/crt1.o /usr/lib/../lib64/crti.o
/build/axel/local/lib/gcc/x86_64-unknown-linux-gnu/4.2.1/crtbegin.o
-L/build/axel/local/lib/gcc/x86_64-unknown-linux-gnu/4.2.1
-L/build/axel/local/lib/gcc/x86_64-unknown-linux-gnu/4.2.1/../../../../lib64
-L/lib/../lib64 -L/usr/lib/../lib64
-L/build/axel/local/lib/gcc/x86_64-unknown-linux-gnu/4.2.1/../../..
/tmp/axel/ccqvfExV.o -lstdc++ -lm -lgcc_s -lgcc -lc -lgcc_s -lgcc
/build/axel/local/lib/gcc/x86_64-unknown-linux-gnu/4.2.1/crtend.o
/usr/lib/../lib64/crtn.o

What causes the message, is it a problem with the build of llvm or with a part
of llvm itself? Could the message be flagged as 'warning:' or 'info:'?


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct  3 10:47:14 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 3 Oct 2008 10:47:14 -0500
Subject: [LLVMbugs] [Bug 2854] New: SimplifyCFG outputs an incorrect IR
Message-ID: <bug-2854-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2854

           Summary: SimplifyCFG outputs an incorrect IR
           Product: libraries
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicolas.geoffray at lip6.fr
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2052)
 --> (http://llvm.org/bugs/attachment.cgi?id=2052)
Test-case for the bug

The attached test-case, after being opt'ed with -simplifycfg, makes llc
unhappy:

llc: SelectionDAGISel.cpp:331: virtual bool
llvm::SelectionDAGISel::runOnFunction(llvm::Function&): Assertion
`FuncInfo->CatchInfoFound.size() == FuncInfo->CatchInfoLost.size() && "Not all
catch info was assigned to a landing pad!"' failed.

If the test-case is not cfg simplified, llc is happy.

Steps to reproduce the bug:
llvm-as testcase.ll
opt -simplifycfg testcase.bc | llc

The testcase is an output from a Java function translated to LLVM IR with
vmkit. FYI, here's the (meaningless) Java function.

public static void main(String[] args) throws Exception {
  try {   
    synchronized(args) {
      while (true)
        abort();
    }
  }
}

Apparently, it's the smallest testcase i could get in Java, even so the
generated IR is quite complex :/


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct  3 11:09:37 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 3 Oct 2008 11:09:37 -0500
Subject: [LLVMbugs] [Bug 2855] New: simplifycfg aborts
Message-ID: <bug-2855-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2855

           Summary: simplifycfg aborts
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: opt
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: rafael.espindola at gmail.com
                CC: llvmbugs at cs.uiuc.edu, nlewycky at google.com


opt -simplifycfg bugpoint-passes.bc  -o opt.bc -f
------------------------------------------------------------
PHI nodes not grouped at top of basic block!
        %iftmp.1.0 = phi i1 [ true, %bb7 ], [ false, %bb3 ]             ; <i1>
[#uses=1]
label %bb7
Instruction does not dominate all uses!
        %iftmp.1.0 = phi i1 [ true, %bb7 ], [ false, %bb3 ]             ; <i1>
[#uses=1]
        %. = select i1 %iftmp.1.0, i32 %2, i32 %8               ; <i32>
[#uses=1]
Broken module found, compilation aborted!
------------------------------------------------------------

This is a reduction of a C++ test:
-----------------------------------------------------------
int f(int *b, int f) {
  int s = 0;

  int c;
  for (;
       ;
       ) {
    c=b[s++];
    if(c) {
      int __c2;
      if(s<f && (__c2=b[s])) {
        ++s;
        c=c+__c2;
      }
    }
  }
}
------------------------------------------------------


cc1plus -O2 testcase.ii -o test.s  
 int f(int*, int)
Analyzing compilation unitPerforming interprocedural optimizations
Assembling functions:
 int f(int*, int)cc1plus:
/usr/local/google/tmp/crosstool.t25222/gcc-llvm-glibc-2.3.6-grte/rpmbuild/BUILD/crosstoolv12-gcc-llvm-glibc-2.3.6-grte-0.1/llvm/lib/CodeGen/LiveVariables.cpp:129:
void llvm::LiveVariables::HandleVirtRegUse(unsigned int,
llvm::MachineBasicBlock*, llvm::MachineInstr*): Assertion `MRI->getVRegDef(reg)
&& "Register use before def!"' failed.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct  3 12:06:41 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 3 Oct 2008 12:06:41 -0500
Subject: [LLVMbugs] [Bug 2856] New: x86_fp80 NaN constants emitted
	inconsistently
Message-ID: <bug-2856-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2856

           Summary: x86_fp80 NaN constants emitted inconsistently
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Keywords: miscompilation
          Severity: normal
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: kamm-removethis at incasoftware.de
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2054)
 --> (http://llvm.org/bugs/attachment.cgi?id=2054)
test case

The compiling and executing the attached LLVM IR prints 

7FFF400000000
7FFFC00000000

even though the x86_fp80 constants used are identical
(0xK7FFF4000000000000000).

The relevant sections in the asm output are

the in-line constant (0x7FC00000):
.LCPI2_0:
        .long   2143289344      # float nan

the global (0x7FFF400000000000)
_D3nan4rvale:
        .size   _D3nan4rvale, 12
        .short  0       # long double least significant halfword of ~nan
        .short  0       # long double next halfword
        .short  0       # long double next halfword
        .short  16384   # long double next halfword
        .short  32767   # long double most significant halfword
        .zero   2

and they should describe the same value, but don't.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct  3 13:58:02 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 3 Oct 2008 13:58:02 -0500
Subject: [LLVMbugs] [Bug 2855] simplifycfg aborts
In-Reply-To: <bug-2855-206@http.llvm.org/bugs/>
Message-ID: <200810031858.m93Iw28g009221@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2855


Devang Patel <dpatel at apple.com> changed:

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




--- Comment #3 from Devang Patel <dpatel at apple.com>  2008-10-03 13:58:01 ---
I tried to simplify this earlier. It is a micro-optimizatin that makes
fast-isel and other passes' life easier.
Your patch is good. Applied.
Rev 57025.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct  3 14:59:23 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 3 Oct 2008 14:59:23 -0500
Subject: [LLVMbugs] [Bug 2846] LLVM does not build (Linux AMD64 GCC 3.4.6)
In-Reply-To: <bug-2846-206@http.llvm.org/bugs/>
Message-ID: <200810031959.m93JxNOK011844@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2846


??scar Fuentes <ofv at wanadoo.es> changed:

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




--- Comment #12 from ??scar Fuentes <ofv at wanadoo.es>  2008-10-03 14:59:22 ---
(In reply to comment #10)
> Concerning the error: I let cmake regenerate verbose Makefiles. This showed me
> that it invokes /usr/bin/c++ (GCC 3.4.6 in my case), not what is set as CXX
> when running cmake, nor what's in the $PATH (both was pointing to GCC 4.3.1):

You mean that there is a `c++' executable on your PATH that comes before than
the one on /usr/bin, and that it was ignored? If the answer is yes, it is a bug
on CMake. Please report it on http://www.cmake.org

[snip]
> So that explains, and you could rightfully call it my fault. I switched the
> cmake variables CMAKE_CXX_COMPILER, CMAKE_C_COMPILER to GCC 4.3.1 and the
> compilation succeeded.
> 

(In reply to comment #11)
> Ok, thanks for clarification. I'd consider this behaviour really error-prone,
> cmake should honour user-supplied standard CC, CXX, LD, etc variables

CMake uses the CXX environment variable. But if you want to specify a compiler
on the CMake command line invocation, you need to use CMAKE_CXX_COMPILER. This
makes some sense, because `make -DCXX=foo` sets the CXX *variable*, but there
is no CXX variable within CMake, it has CMAKE_CXX_COMPILER, which default value
is taken from the CXX environment variable, if defined. You can argue that this
is a bit verbose, but it is consistent with CMake's naming conventions.

One more thing to keep in mind is that CMake != make. If there is a confusing
thing here is CMake's name itself :-).


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct  3 16:41:45 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 3 Oct 2008 16:41:45 -0500
Subject: [LLVMbugs] [Bug 2857] New: Attempt to use a SCEVCouldNotCompute
	object!
Message-ID: <bug-2857-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2857

           Summary: Attempt to use a SCEVCouldNotCompute object!
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Loop Optimizer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nlewycky at google.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2056)
 --> (http://llvm.org/bugs/attachment.cgi?id=2056)
testcase

This testcase crashes:

$ llvm-as < b2.ll | opt -indvars
WARNING: You're attempting to print out a bitcode file.
This is inadvisable as it may cause display problems. If
you REALLY want to taste LLVM bitcode first-hand, you
can force output with the `-f' option.

opt: ScalarEvolution.cpp:136: virtual bool
llvm::SCEVCouldNotCompute::isLoopInvariant(const llvm::Loop*) const: Assertion
`0 && "Attempt to use a SCEVCouldNotCompute object!"' failed.
opt[0x86b068e]
opt[0x86b08f0]
[0xffffe500]
[0x6]
/lib/tls/i686/cmov/libc.so.6(abort+0xe9)[0x49739619]
/lib/tls/i686/cmov/libc.so.6(__assert_fail+0x101)[0x49731031]
opt(_ZNK4llvm19SCEVCouldNotCompute7getTypeEv+0x0)[0x8576c34]
opt(_ZNK4llvm19SCEVCommutativeExpr15isLoopInvariantEPKNS_8LoopBaseINS_10BasicBlockEEE+0x4c)[0x85875fa]
opt(_ZNK4llvm19SCEVCommutativeExpr15isLoopInvariantEPKNS_8LoopBaseINS_10BasicBlockEEE+0x4c)[0x85875fa]
opt[0x8410d0a]
opt[0x841181c]
opt(_ZN4llvm13LPPassManager13runOnFunctionERNS_8FunctionE+0x392)[0x8563cec]
opt(_ZN4llvm13FPPassManager13runOnFunctionERNS_8FunctionE+0x102)[0x863c5bc]
opt(_ZN4llvm13FPPassManager11runOnModuleERNS_6ModuleE+0x6e)[0x863c766]
opt(_ZN4llvm13MPPassManager11runOnModuleERNS_6ModuleE+0xcd)[0x863c289]
opt(_ZN4llvm15PassManagerImpl3runERNS_6ModuleE+0x74)[0x863c426]
opt(_ZN4llvm11PassManager3runERNS_6ModuleE+0x1b)[0x863c481]
opt(main+0x105d)[0x83712cd]
/lib/tls/i686/cmov/libc.so.6(__libc_start_main+0xd2)[0x49723ec2]
opt(__gxx_personality_v0+0x341)[0x8362911]
Aborted

The stacktrace is:

#0  0xffffe410 in __kernel_vsyscall ()
#1  0x49737c51 in raise () from /lib/tls/i686/cmov/libc.so.6
#2  0x49739619 in abort () from /lib/tls/i686/cmov/libc.so.6
#3  0x49731031 in __assert_fail () from /lib/tls/i686/cmov/libc.so.6
#4  0x08576c34 in llvm::SCEVCouldNotCompute::isLoopInvariant (this=0x87d0088, 
    L=0x87cfc80) at ScalarEvolution.cpp:136
#5  0x085875fa in llvm::SCEVCommutativeExpr::isLoopInvariant (this=0x87d0a30, 
    L=0x87cfc80)
    at
/usr/local/src/llvm/include/llvm/Analysis/ScalarEvolutionExpressions.h:230
#6  0x085875fa in llvm::SCEVCommutativeExpr::isLoopInvariant (this=0x87d0a98, 
    L=0x87cfc80)
    at
/usr/local/src/llvm/include/llvm/Analysis/ScalarEvolutionExpressions.h:230
#7  0x08410d0a in (anonymous namespace)::IndVarSimplify::RewriteLoopExitValues
    (this=0x87cb408, L=0x87cfc80, IterationCount=0x87d0430)
    at IndVarSimplify.cpp:379
#8  0x0841181c in (anonymous namespace)::IndVarSimplify::runOnLoop (
    this=0x87cb408, L=0x87cfc80, LPM=@0x87c7a98) at IndVarSimplify.cpp:463
#9  0x08563cec in llvm::LPPassManager::runOnFunction (this=0x87c7a98, 
    F=@0x87c9810) at LoopPass.cpp:226
#10 0x0863c5bc in llvm::FPPassManager::runOnFunction (this=0x87c7778, 
    F=@0x87c9810) at PassManager.cpp:1310
#11 0x0863c766 in llvm::FPPassManager::runOnModule (this=0x87c7778, 
    M=@0x87c6010) at PassManager.cpp:1333
#12 0x0863c289 in llvm::MPPassManager::runOnModule (this=0x87c6100, 
    M=@0x87c6010) at PassManager.cpp:1380
#13 0x0863c426 in llvm::PassManagerImpl::run (this=0x87c78b0, M=@0x87c6010)
    at PassManager.cpp:1454
#14 0x0863c481 in llvm::PassManager::run (this=0xffffd5ec, M=@0x87c6010)
    at PassManager.cpp:1486
#15 0x083712cd in main (argc=3, argv=0xffffd7e4) at opt.cpp:582


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct  3 19:36:28 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 3 Oct 2008 19:36:28 -0500
Subject: [LLVMbugs] [Bug 2716] integer comparison bug
In-Reply-To: <bug-2716-206@http.llvm.org/bugs/>
Message-ID: <200810040036.m940aSCb021766@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2716


Dan Gohman <gohman at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |gohman at apple.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Dan Gohman <gohman at apple.com>  2008-10-03 19:36:27 ---
It was a bug in the "local" register allocator, which is enabled in
llvm-gcc at -O0.  Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080929/068014.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct  3 20:53:18 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 3 Oct 2008 20:53:18 -0500
Subject: [LLVMbugs] [Bug 2833] AtomicSwap64 not supported by x86 backend
In-Reply-To: <bug-2833-206@http.llvm.org/bugs/>
Message-ID: <200810040153.m941rIPa024342@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2833


Dale Johannesen <dalej at apple.com> changed:

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




--- Comment #4 from Dale Johannesen <dalej at apple.com>  2008-10-03 20:53:17 ---
This is fixed.  All the 64-bit atomics should work now in 32-bit mode.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Oct  4 05:03:52 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 4 Oct 2008 05:03:52 -0500
Subject: [LLVMbugs] =?utf-8?q?=5BBug_2858=5D_New=3A_CBE=3A_error=3A__confl?=
 =?utf-8?b?aWN0aW5nIHR5cGVzIGZvciDigJhfbG9uZ2ptcOKAmQ==?=
Message-ID: <bug-2858-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2858

           Summary: CBE: error: conflicting types for ???_longjmp???
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Try to compile the following with SVN r57045:

$ cat foo.c
#include <setjmp.h>
jmp_buf buf;
int foo()
{
        _longjmp(buf, 1);
        return 0;
}
$ llvm-gcc foo.c -c -o foo.bc -emit-llvm
$ llc -march=c foo.bc -o foo.cbe.c
$ gcc foo.cbe.c -c
foo.cbe.c:149: error: conflicting types for ???_longjmp???
/usr/include/setjmp.h:91: error: previous declaration of ???_longjmp??? was
here
foo.cbe.c:150: warning: conflicting types for built-in function ???malloc???

Bugpoint reduced bitcode:
; ModuleID = 'bugpoint-reduced-simplified.bc'
target datalayout =
"e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128"
target triple = "x86_64-unknown-linux-gnu"
        %struct.__jmp_buf_tag = type { [8 x i64], i32, %struct.__sigset_t }
        %struct.__sigset_t = type { [16 x i64] }
@buf = common global [1 x %struct.__jmp_buf_tag] zeroinitializer, align 32     
        ; <[1 x %struct.__jmp_buf_tag]*> [#uses=1]

define i32 @foo() nounwind {
entry:
        %retval = alloca i32            ; <i32*> [#uses=1]
        %0 = alloca i32         ; <i32*> [#uses=0]
        %"alloca point" = bitcast i32 0 to i32          ; <i32> [#uses=0]
        call void @_longjmp(%struct.__jmp_buf_tag* getelementptr ([1 x
%struct.__jmp_buf_tag]* @buf, i32 0, i32 0), i32 1) noreturn nounwind
        unreachable

return:         ; No predecessors!
        %retval1 = load i32* %retval            ; <i32> [#uses=1]
        ret i32 %retval1
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Oct  4 08:28:10 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 4 Oct 2008 08:28:10 -0500
Subject: [LLVMbugs] [Bug 2859] New: Clang revision 57049 won't build
Message-ID: <bug-2859-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2859

           Summary: Clang revision 57049 won't build
           Product: clang
           Version: unspecified
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: blocker
          Priority: P2
         Component: Basic
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: cfkaran2 at gmail.com
                CC: llvmbugs at cs.uiuc.edu


I built LLVM and Clang with the following on a macintosh.  I have built LLVM
and Clang before, but this is the first time I built it with the options that I
did (see below for the options I passed to configure).  Note that I did NOT do
'make clean' or anything else; I simply updated my working copy, and then did
the following commands.

LLVM:
URL: http://llvm.org/svn/llvm-project/llvm/trunk/tools
Repository Root: http://llvm.org/svn/llvm-project
Repository UUID: 91177308-0d34-0410-b5e6-96231b3b80d8
Revision: 57049
Node Kind: directory
Schedule: normal
Last Changed Author: foldr
Last Changed Rev: 57031
Last Changed Date: 2008-10-03 17:26:27 -0400 (Fri, 03 Oct 2008)

Built with the following options:

./configure --enable-optimized --enable-assertions --enable-expensive-checks
--enable-debug-runtime

CLANG:
Path: .
URL: http://llvm.org/svn/llvm-project/cfe/trunk
Repository Root: http://llvm.org/svn/llvm-project
Repository UUID: 91177308-0d34-0410-b5e6-96231b3b80d8
Revision: 57049
Node Kind: directory
Schedule: normal
Last Changed Author: kremenek
Last Changed Rev: 57045
Last Changed Date: 2008-10-04 02:53:56 -0400 (Sat, 04 Oct 2008)

After updating, ran the following:

s2:clang cfkaran2$ make
make[2]: Nothing to be done for `all'.
llvm[2]: Compiling Diagnostic.cpp for Release+Checks build 
/usr/include/c++/4.0.0/debug/formatter.h: In constructor
???__gnu_debug::_Error_formatter::_Parameter::_Parameter(const
__gnu_debug::_Safe_iterator<_Iterator, _Sequence>&, const char*,
__gnu_debug::_Error_formatter::_Is_iterator)???:
/usr/include/c++/4.0.0/debug/formatter.h:214: error: cannot use typeid with
-fno-rtti
/usr/include/c++/4.0.0/debug/formatter.h:220: error: cannot use typeid with
-fno-rtti
/usr/include/c++/4.0.0/debug/formatter.h: In constructor
???__gnu_debug::_Error_formatter::_Parameter::_Parameter(const _Type*&, const
char*, __gnu_debug::_Error_formatter::_Is_iterator)???:
/usr/include/c++/4.0.0/debug/formatter.h:243: error: cannot use typeid with
-fno-rtti
/usr/include/c++/4.0.0/debug/formatter.h: In constructor
???__gnu_debug::_Error_formatter::_Parameter::_Parameter(_Type*&, const char*,
__gnu_debug::_Error_formatter::_Is_iterator)???:
/usr/include/c++/4.0.0/debug/formatter.h:256: error: cannot use typeid with
-fno-rtti
/usr/include/c++/4.0.0/debug/formatter.h: In constructor
???__gnu_debug::_Error_formatter::_Parameter::_Parameter(const _Iterator&,
const char*, __gnu_debug::_Error_formatter::_Is_iterator)???:
/usr/include/c++/4.0.0/debug/formatter.h:269: error: cannot use typeid with
-fno-rtti
/usr/include/c++/4.0.0/debug/formatter.h: In constructor
???__gnu_debug::_Error_formatter::_Parameter::_Parameter(const
__gnu_debug::_Safe_sequence<_Sequence>&, const char*,
__gnu_debug::_Error_formatter::_Is_sequence)???:
/usr/include/c++/4.0.0/debug/formatter.h:285: error: cannot use typeid with
-fno-rtti
/usr/include/c++/4.0.0/debug/formatter.h: In constructor
???__gnu_debug::_Error_formatter::_Parameter::_Parameter(const _Sequence&,
const char*, __gnu_debug::_Error_formatter::_Is_sequence)???:
/usr/include/c++/4.0.0/debug/formatter.h:294: error: cannot use typeid with
-fno-rtti
make[2]: ***
[/Users/cfkaran2/Desktop/Programming/llvm/tools/clang/lib/Basic/Release+Checks/Diagnostic.o]
Error 1
make[1]: *** [Basic/.makeall] Error 2
make: *** [all] Error 1


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Oct  4 12:04:05 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 4 Oct 2008 12:04:05 -0500
Subject: [LLVMbugs] [Bug 2859] Clang revision 57049 won't build
In-Reply-To: <bug-2859-206@http.llvm.org/bugs/>
Message-ID: <200810041704.m94H45E6030111@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2859


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-10-04 12:04:05 ---
This is a problem with your version of the checking libstdc++.  It apparently
requires rtti information, which clang does not build with.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Oct  4 12:48:51 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 4 Oct 2008 12:48:51 -0500
Subject: [LLVMbugs] =?utf-8?q?=5BBug_2858=5D_CBE=3A_error=3A__conflicting_?=
 =?utf-8?b?dHlwZXMgZm9yIOKAmF9sb25nam1w4oCZ?=
In-Reply-To: <bug-2858-206@http.llvm.org/bugs/>
Message-ID: <200810041748.m94HmpR8031582@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2858


Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |daniel at zuster.org
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #1 from Daniel Dunbar <daniel at zuster.org>  2008-10-04 12:48:50 ---


*** This bug has been marked as a duplicate of bug 1124 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Oct  4 13:25:06 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 4 Oct 2008 13:25:06 -0500
Subject: [LLVMbugs] [Bug 2857] Attempt to use a SCEVCouldNotCompute object!
In-Reply-To: <bug-2857-206@http.llvm.org/bugs/>
Message-ID: <200810041825.m94IP6Lr000342@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2857


Nick Lewycky <nicholas at mxc.ca> changed:

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




--- Comment #5 from Nick Lewycky <nicholas at mxc.ca>  2008-10-04 13:25:06 ---
Yes, I just forgot to close it. Fixed in r57049.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Oct  4 17:33:41 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 4 Oct 2008 17:33:41 -0500
Subject: [LLVMbugs] [Bug 2860] New: llvm-gcc ICE when building libgcc
Message-ID: <bug-2860-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2860

           Summary: llvm-gcc ICE when building libgcc
           Product: tools
           Version: trunk
          Platform: Other
        OS/Version: NetBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: krister.walfridsson at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2058)
 --> (http://llvm.org/bugs/attachment.cgi?id=2058)
source code that makes the compiler ICE

llvm-gcc-4.2 reports an internal compiler error when compiling libgcc on
x86_64-unknown-netbsd:

/usr/local/tmp/llvm081004/gbuild/./gcc/xgcc
-B/usr/local/tmp/llvm081004/gbuild/./gcc/
-B/home/cato/llvmgcc/x86_64-unknown-netbsd4.99.72/bin/
-B/home/cato/llvmgcc/x86_64-unknown-netbsd4.99.72/lib/ -isystem
/home/cato/llvmgcc/x86_64-unknown-netbsd4.99.72/include -isystem
/home/cato/llvmgcc/x86_64-unknown-netbsd4.99.72/sys-include  -O2  -O2 -g -O2 
-DIN_GCC    -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes
-Wold-style-definition  -isystem ./include  -fPIC -g -DHAVE_GTHR_DEFAULT
-DIN_LIBGCC2 -D__GCC_FLOAT_NOT_NEEDED  -I. -I. -I../../llvm-gcc-4.2/gcc
-I../../llvm-gcc-4.2/gcc/. -I../../llvm-gcc-4.2/gcc/../include -I./../intl
-I../../llvm-gcc-4.2/gcc/../libcpp/include 
-I../../llvm-gcc-4.2/gcc/../libdecnumber -I../libdecnumber
-I/usr/local/tmp/llvm081004/build/include
-I/usr/local/tmp/llvm081004/llvm/include -DL_popcountsi2 -fvisibility=hidden
-DHIDE_EXPORTS -c ../../llvm-gcc-4.2/gcc/libgcc2.c -o libgcc/./_popcountsi2.o
assertion "V == V2 && "Didn't find key?"" failed: file
"/usr/local/tmp/llvm081004/llvm/lib/Support/StringMap.cpp", line 177, function
"void llvm::StringMapImpl::RemoveKey(llvm::StringMapEntryBase*)"
../../llvm-gcc-4.2/gcc/libgcc2.c:809: internal compiler error: Abort trap
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://developer.apple.com/bugreporter> for instructions.

(btw. shouldn't it point me to http://llvm.org/bugs/ instead?)

The ICE can be reproduced by the slightly simplified attached file.

I configured/built llvm as:
  ../llvm/configure --prefix=/home/kristerw/llvm
  gmake ENABLE_OPTIMIZED=1
and llvm-gcc as:
  ../llvm-gcc-4.2/configure --program-prefix=llvm-
--prefix=/home/kristerw/llvmgcc --enable-languages="c,c++" --enable-checking
--enable-llvm=/usr/local/tmp/llvm081004/build --disable-bootstrap
--disable-multilib
  gmake


The failure was introduced by revision 55063
(http://llvm.org/viewvc/llvm-project?view=rev&revision=55063)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Oct  4 20:40:11 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 4 Oct 2008 20:40:11 -0500
Subject: [LLVMbugs] [Bug 2861] New: generic thunk code fails
Message-ID: <bug-2861-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2861

           Summary: generic thunk code fails
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-g++
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicholas at mxc.ca
                CC: llvmbugs at cs.uiuc.edu, nlewycky at google.com


Created an attachment (id=2059)
 --> (http://llvm.org/bugs/attachment.cgi?id=2059)
testcase

This delta-reduced testcase triggers a problem with llvm-g++ that doesn't
happen with FSF g++ 4.2:

$ g++ thunk_error.cpp -c
thunk_error.cpp:21: error: generic thunk code fails for method ???virtual void
RPCCore::Ev(const char*, int, int, const char*, ...) const??? which uses
???...???
$ g++-4.2 thunk_error.cpp -c
$ g++-4.2 -v
[...]
gcc version 4.2.4 (Debian 4.2.4-3)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Oct  4 21:14:23 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 4 Oct 2008 21:14:23 -0500
Subject: [LLVMbugs] [Bug 2842] instcombine should generate bswap
In-Reply-To: <bug-2842-206@http.llvm.org/bugs/>
Message-ID: <200810050214.m952ENpg017399@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2842


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #5 from Chris Lattner <clattner at apple.com>  2008-10-04 21:14:23 ---
Implemented here, thanks!
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080929/068071.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Oct  4 23:12:06 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 4 Oct 2008 23:12:06 -0500
Subject: [LLVMbugs] [Bug 2682] llvm-gcc fails to find that an argument to
 __builtin_prefetch is a constant
In-Reply-To: <bug-2682-206@http.llvm.org/bugs/>
Message-ID: <200810050412.m954C68O020598@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2682


Nick Lewycky <nicholas at mxc.ca> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |nicholas at mxc.ca
             Status|RESOLVED                    |REOPENED
         Resolution|FIXED                       |




--- Comment #3 from Nick Lewycky <nicholas at mxc.ca>  2008-10-04 23:12:06 ---
This is not fixed:

$ llvm-g++ 004.c -c
$ llvm-g++ 004.c -c -O2
004.c: In function ???void prefetch(const char*, int)???:
004.c:1: error: third argument to ???__builtin_prefetch??? must be a constant
$ 


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Oct  5 00:54:11 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 5 Oct 2008 00:54:11 -0500
Subject: [LLVMbugs] [Bug 2862] New: [sema] vector types differing only in
 signedness should be compatible
Message-ID: <bug-2862-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2862

           Summary: [sema] vector types differing only in signedness should
                    be compatible
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: daniel at zuster.org
                CC: llvmbugs at cs.uiuc.edu


clang rejects the following code which gcc accepts
--
extern void f1(char a0 __attribute__((vector_size(8))));

void f0(signed char a0 __attribute__((vector_size(8)))) {
  f1(a0);
}
--
ddunbar at ddunbar2:clang$ clang test/CodeGen/f.c
test/CodeGen/f.c:6:6: error: incompatible type passing 'signed char 
__attribute__((__vector_size__(32)))', expected 'char 
__attribute__((__vector_size__(32)))'
  f1(a0);
     ^~
1 diagnostic generated.
--


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Oct  5 02:19:16 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 5 Oct 2008 02:19:16 -0500
Subject: [LLVMbugs] [Bug 2863] New: llvm-gcc ICEs on __builtin_ia32_punpckldq
Message-ID: <bug-2863-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2863

           Summary: llvm-gcc ICEs on __builtin_ia32_punpckldq
           Product: libraries
           Version: trunk
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: daniel at zuster.org
                CC: llvmbugs at cs.uiuc.edu


llvm-gcc ICEs on
--
typedef signed int V2i __attribute__((vector_size(8)));
void f0() {
  V2i    tmp_V2i;
  tmp_V2i = __builtin_ia32_punpckldq(tmp_V2i, tmp_V2i);
}
--
ddunbar at ddunbar2:CodeGen$ gcc -v && gcc -c -mmmx m.c
Using built-in specs.
Target: i686-apple-darwin10
Configured with: /var/tmp/gcc/gcc-5616.3~3/src/configure --disable-checking
--enable-werror --prefix=/usr --mandir=/usr/share/man
--enable-languages=c,objc,c++,obj-c++
--program-transform-name=/^[cg][^.-]*$/s/$/\
-4.2/ --with-slibdir=/usr/lib --build=i686-apple-darwin10
--with-gxx-include-dir=/usr/include/c++/4.2.1 --host=i686-apple-darwin10
--target=i686-apple-darwin10
Thread model: posix
gcc version 4.2.1 (Apple Inc. build 5616) (dot 3)


ddunbar at ddunbar2:CodeGen$ llvm-gcc -v && llvm-gcc -c -mmmx m.c
Using built-in specs.
Target: i686-apple-darwin10
Configured with: ../llvm-gcc-4.2/configure
--prefix=/Users/ddunbar/llvm-gcc/obj/../install --program-prefix=llvm-
--enable-llvm=/Users/ddunbar/llvm/ --enable-languages=c,c++,objc,obj-c++
--with-arch=nocona --wit\
h-tune=generic --with-gxx-include-dir=/usr/include/c++/4.2.1
--build=i686-apple-darwin10 --host=i686-apple-darwin10
--target=i686-apple-darwin10
Thread model: posix
gcc version 4.2.1 (Based on Apple Inc. build 5623) (LLVM build)
Assertion failed: (VT.getSizeInBits() == Operand.getValueType().getSizeInBits()
&& "Cannot BIT_CONVERT between types of different sizes!"), function getNode,
file SelectionDAG.cpp, line 2218.
m.c:5: internal compiler error: Abort trap
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://developer.apple.com/bugreporter> for instructions.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Oct  5 10:08:15 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 5 Oct 2008 10:08:15 -0500
Subject: [LLVMbugs] [Bug 2830] Cpp Backend doesn' t generate correct code
 for Switch instructions
In-Reply-To: <bug-2830-206@http.llvm.org/bugs/>
Message-ID: <200810051508.m95F8FOH008903@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2830


Anton Korobeynikov <asl at math.spbu.ru> changed:

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




--- Comment #4 from Anton Korobeynikov <asl at math.spbu.ru>  2008-10-05 10:08:14 ---
(In reply to comment #1)
> Created an attachment (id=2040)
 --> (http://llvm.org/bugs/attachment.cgi?id=2040) [details]
> Patch to fix bug.
> 
> One way to fix this bug is to remove that code that generates calls to
> 'Constant::getNullValue', this will allow the correct constant type to be
> generated.
I agree, that having sane constant type is a right way to fix, otherwise we
will need to end with cast<ConstantInt> in bunch of places.

Patch applied as r57110. Thanks!


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Oct  5 13:22:52 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 5 Oct 2008 13:22:52 -0500
Subject: [LLVMbugs] [Bug 2849] llc fails an assertion in ScheduleDAG
In-Reply-To: <bug-2849-206@http.llvm.org/bugs/>
Message-ID: <200810051822.m95IMqpx015158@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2849


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #2 from Chris Lattner <clattner at apple.com>  2008-10-05 13:22:51 ---
This now works for me:

$ llc bugpoint-reduced-simplified.bc

I think that this was fixed on mainline recently.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Oct  5 13:32:22 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 5 Oct 2008 13:32:22 -0500
Subject: [LLVMbugs] [Bug 2827] CodeEmitterGen produce wrong values in
 generated function getBinaryCodeForInstr (...)
In-Reply-To: <bug-2827-206@http.llvm.org/bugs/>
Message-ID: <200810051832.m95IWMNI015507@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2827


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-10-05 13:32:21 ---
Fixed with a slightly different approach, please verify that this works for
you:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080929/068087.html

Thanks!


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Oct  5 13:48:36 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 5 Oct 2008 13:48:36 -0500
Subject: [LLVMbugs] [Bug 2776] possible volatile bug -- too many loads
In-Reply-To: <bug-2776-206@http.llvm.org/bugs/>
Message-ID: <200810051848.m95Ima0s016099@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2776


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #8 from Chris Lattner <clattner at apple.com>  2008-10-05 13:48:36 ---
Yep, I'm pretty sure this is fixed.  John, please reopen if it is still causing
problems, thanks! 


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Oct  5 14:23:27 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 5 Oct 2008 14:23:27 -0500
Subject: [LLVMbugs] [Bug 2773] Define __USER_LABEL_PREFIX__
In-Reply-To: <bug-2773-206@http.llvm.org/bugs/>
Message-ID: <200810051923.m95JNRLc017311@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2773


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-10-05 14:23:26 ---
Implemented here:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20080929/007973.html

I have no intention of implementing support for  -fno-leading-underscore.  If
someone needs it, they can make sure to fix this macro as well.  I don't think
the llvm backend supprots it at all for example.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Oct  5 16:56:08 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 5 Oct 2008 16:56:08 -0500
Subject: [LLVMbugs] [Bug 2761] Analyser: command line option for checking GC
	code
In-Reply-To: <bug-2761-206@http.llvm.org/bugs/>
Message-ID: <200810052156.m95Lu8kp022044@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2761


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #2 from Chris Lattner <clattner at apple.com>  2008-10-05 16:56:08 ---
Make sure to build your project with -fobjc-gc-only.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Oct  5 17:01:18 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 5 Oct 2008 17:01:18 -0500
Subject: [LLVMbugs] [Bug 2236] Implement defined but not used diagnostic for
	statics
In-Reply-To: <bug-2236-206@http.llvm.org/bugs/>
Message-ID: <200810052201.m95M1IeM022300@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2236


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #12 from Chris Lattner <clattner at apple.com>  2008-10-05 17:01:15 ---
The original issue is fixed


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Oct  5 17:04:11 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 5 Oct 2008 17:04:11 -0500
Subject: [LLVMbugs] [Bug 2301] Internal compiler error when doing inline
 optimization in llvm-gcc4.2-2.2
In-Reply-To: <bug-2301-206@http.llvm.org/bugs/>
Message-ID: <200810052204.m95M4BUj022427@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2301


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #6 from Chris Lattner <clattner at apple.com>  2008-10-05 17:04:10 ---
This works now, I don't know if it is because it is fixed or just hidden now
though.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Oct  5 17:53:55 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 5 Oct 2008 17:53:55 -0500
Subject: [LLVMbugs] [Bug 2682] llvm-gcc fails to find that an argument to
 __builtin_prefetch is a constant
In-Reply-To: <bug-2682-206@http.llvm.org/bugs/>
Message-ID: <200810052253.m95MrtAK023898@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2682


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|                            |WONTFIX




--- Comment #5 from Chris Lattner <clattner at apple.com>  2008-10-05 17:53:54 ---
This is one of those cases where GCC changes what code it accepts based on how
its optimizer evolves.  It will basically accept any builtin_prefetch as long
as the GIMPLE level optimizers are able to eliminate it, and does the checking
after the full optimizer pass runs.  A better way to do this is with macros and
with __builtin_choose_expr. __builtin_choose_expr( __builtin_constant_p, ...)
is guaranteed to not evaluate the subexpression that is not chosen.

The user should change their code, sorry.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Oct  6 12:38:09 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 6 Oct 2008 12:38:09 -0500
Subject: [LLVMbugs] [Bug 2704] Constructor-style initialization of bool fails
In-Reply-To: <bug-2704-206@http.llvm.org/bugs/>
Message-ID: <200810061738.m96Hc9lA005774@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2704


Argiris Kirtzidis <akyrtzi at gmail.com> changed:

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




--- Comment #2 from Argiris Kirtzidis <akyrtzi at gmail.com>  2008-10-06 12:38:08 ---
This has now been fixed (rev. 57178).


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Oct  6 13:24:52 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 6 Oct 2008 13:24:52 -0500
Subject: [LLVMbugs] [Bug 2856] x86_fp80 NaN constants emitted inconsistently
In-Reply-To: <bug-2856-206@http.llvm.org/bugs/>
Message-ID: <200810061824.m96IOqSb007338@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2856


Dale Johannesen <dalej at apple.com> changed:

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




--- Comment #3 from Dale Johannesen <dalej at apple.com>  2008-10-06 13:24:51 ---
I'm not really sure how these values ought to behave, but I've marked
conversions that shorten a NaN as Inexact; that fixes this problem(?), and
can't be too far off.
http://llvm.org/viewvc/llvm-project?view=rev&revision=57185


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Oct  6 15:47:38 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 6 Oct 2008 15:47:38 -0500
Subject: [LLVMbugs] [Bug 2864] New: instcombine removes live stores and calls
Message-ID: <bug-2864-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2864

           Summary: instcombine removes live stores and calls
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Keywords: miscompilation
          Severity: major
          Priority: P2
         Component: Scalar Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: alenhar2 at uiuc.edu
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2061)
 --> (http://llvm.org/bugs/attachment.cgi?id=2061)
butpoint-tooptimize.bc

Instcombine removes stores and recursive calls from this function.

opt bugpoint-tooptimize.bc -instcombine -o foo.bc

--- bugpoint-tooptimize.ll      2008-10-06 15:41:47.000000000 -0500
+++ foo.ll      2008-10-06 15:41:51.000000000 -0500
@@ -1,4 +1,4 @@
-; ModuleID = 'bugpoint-tooptimize.bc'
+; ModuleID = 'foo.bc'
 target datalayout =
"e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128"
 target triple = "x86_64-unknown-linux-gnu"
        %PoolDescriptor = type [16 x i8*]
@@ -50,23 +50,14 @@
        br label %bb1

 bb1.ret.exitStub:              ; preds = %bb1
-       store %struct.tree_t* %1, %struct.tree_t** %.out
+       %0 = bitcast i8* %1 to %struct.tree_t*          ; <%struct.tree_t*>
[#uses=1]
+       store %struct.tree_t* %0, %struct.tree_t** %.out
        ret void

 bb1:           ; preds = %newFuncRoot
-       %0 = call i8* @poolalloc(%PoolDescriptor* %PDa, i32 24)         ; <i8*>
[#uses=1]
-       %1 = bitcast i8* %0 to %struct.tree_t*          ; <%struct.tree_t*>
[#uses=4]
-       %2 = sdiv i32 %proc, 2          ; <i32> [#uses=3]
-       %3 = add i32 %2, %lo            ; <i32> [#uses=1]
-       %4 = add i32 %level, -1         ; <i32> [#uses=2]
-       %5 = call %struct.tree_t* @TreeAlloc2(%PoolDescriptor* %PDa, i32 %4,
i32 %3, i32 %2)            ; <%struct.tree_t*> [#uses=1]
-       %6 = call %struct.tree_t* @TreeAlloc2(%PoolDescriptor* %PDa, i32 %4,
i32 %lo, i32 %2)           ; <%struct.tree_t*> [#uses=1]
-       %7 = getelementptr %struct.tree_t* %1, i64 0, i32 0             ;
<i32*> [#uses=1]
-       store i32 1, i32* %7, align 8
-       %8 = getelementptr %struct.tree_t* %1, i64 0, i32 1             ;
<%struct.tree_t**> [#uses=1]
-       store %struct.tree_t* %5, %struct.tree_t** %8, align 8
-       %9 = getelementptr %struct.tree_t* %1, i64 0, i32 2             ;
<%struct.tree_t**> [#uses=1]
-       store %struct.tree_t* %6, %struct.tree_t** %9, align 8
+       %1 = call i8* @poolalloc(%PoolDescriptor* %PDa, i32 24)         ; <i8*>
[#uses=2]
+       %2 = bitcast i8* %1 to i32*             ; <i32*> [#uses=1]
+       store i32 1, i32* %2, align 8
        br label %bb1.ret.exitStub
 }


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Oct  6 16:13:55 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 6 Oct 2008 16:13:55 -0500
Subject: [LLVMbugs] [Bug 2850] MOVDDUP instruction selection asserts
In-Reply-To: <bug-2850-206@http.llvm.org/bugs/>
Message-ID: <200810062113.m96LDtLm014347@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2850


Evan Cheng <evan.cheng at apple.com> changed:

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




--- Comment #4 from Evan Cheng <evan.cheng at apple.com>  2008-10-06 16:13:54 ---
Fixed.
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081006/068139.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Oct  6 16:14:11 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 6 Oct 2008 16:14:11 -0500
Subject: [LLVMbugs] [Bug 2863] llvm-gcc ICEs on __builtin_ia32_punpckldq
In-Reply-To: <bug-2863-206@http.llvm.org/bugs/>
Message-ID: <200810062114.m96LEBZU014372@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2863


Evan Cheng <evan.cheng at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #3 from Evan Cheng <evan.cheng at apple.com>  2008-10-06 16:14:11 ---


*** This bug has been marked as a duplicate of bug 2850 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Oct  6 16:34:25 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 6 Oct 2008 16:34:25 -0500
Subject: [LLVMbugs] [Bug 2864] instcombine removes live stores and calls
In-Reply-To: <bug-2864-206@http.llvm.org/bugs/>
Message-ID: <200810062134.m96LYPAx015231@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2864


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-10-06 16:34:24 ---
This is because the callee is marked fastcc and the call site is not.  This
code has undefined behavior.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Oct  6 17:53:30 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 6 Oct 2008 17:53:30 -0500
Subject: [LLVMbugs] [Bug 2865] New: LoopIndexSplit assertion failure
Message-ID: <bug-2865-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2865

           Summary: LoopIndexSplit assertion failure
           Product: libraries
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Keywords: compile-fail
          Severity: normal
          Priority: P2
         Component: Loop Optimizer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: gohman at apple.com
                CC: dpatel at apple.com, llvmbugs at cs.uiuc.edu


Created an attachment (id=2062)
 --> (http://llvm.org/bugs/attachment.cgi?id=2062)
bugpoint-reduced testcase

opt -loop-index-split aborts on the attached testcase with
this message:

Assertion failed: (0 && "Unexpected split condition predicate"), function
updateLoopBounds, file lib/Transforms/Scalar/LoopIndexSplit.cpp, line 917.

This is is a testcase reduced from a failure observed in 400.perlbench.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Oct  6 18:12:20 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 6 Oct 2008 18:12:20 -0500
Subject: [LLVMbugs] [Bug 2866] New: fast-isel triggers infinite loop in
	linearscan
Message-ID: <bug-2866-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2866

           Summary: fast-isel triggers infinite loop in linearscan
           Product: libraries
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: gohman at apple.com
                CC: llvmbugs at cs.uiuc.edu


The attached testcase causes llc -fast-isel to hang in the
linearscan phase.  It was reduced from 403.gcc compiled
with regular optimizations.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Oct  6 18:23:08 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 6 Oct 2008 18:23:08 -0500
Subject: [LLVMbugs] [Bug 2865] LoopIndexSplit assertion failure
In-Reply-To: <bug-2865-206@http.llvm.org/bugs/>
Message-ID: <200810062323.m96NN8H1019658@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2865


Devang Patel <dpatel at apple.com> changed:

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




--- Comment #1 from Devang Patel <dpatel at apple.com>  2008-10-06 18:23:08 ---
Fixed. rev 57221.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Oct  6 18:30:17 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 6 Oct 2008 18:30:17 -0500
Subject: [LLVMbugs] [Bug 2867] New: ICE on invalid;
 class definiton inside method definition
Message-ID: <bug-2867-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2867

           Summary: ICE on invalid; class definiton inside method definition
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-g++
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nlewycky at google.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2064)
 --> (http://llvm.org/bugs/attachment.cgi?id=2064)
testcase

This testcase is delta-reduced from valid code. I'm not sure how to build it up
into a valid example.

Anyways, here's the goods:

  $ llvm-gcc ice.cc
  ice.cc: In member function ???void LogFileObject::SetBasename(const
char*)???:
  ice.cc:8: internal compiler error: in lookup_name_real, at
cp/name-lookup.c:4043
  Please submit a full bug report,
  with preprocessed source if appropriate.
  See <URL:http://developer.apple.com/bugreporter> for instructions.

The backtrace is useless:

  Breakpoint 2, 0x083771f0 in internal_error ()
  (gdb) bt
  #0  0x083771f0 in internal_error ()
  #1  0x083777c8 in fancy_abort ()
  #2  0xf7f72f00 in ?? ()
  #3  0xf7f6b780 in ?? ()
  #4  0xf7ecef00 in ?? ()
  #5  0xf7f7763c in ?? ()
  #6  0x00000000 in ?? ()
  (gdb)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Oct  6 21:32:40 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 6 Oct 2008 21:32:40 -0500
Subject: [LLVMbugs] [Bug 2868] New: ADDC ADDE not expanded
Message-ID: <bug-2868-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2868

           Summary: ADDC ADDE not expanded
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: alenhar2 at uiuc.edu
                CC: llvmbugs at cs.uiuc.edu


ADDC, ADDE, SUBC, and SUBE are not expanded in legalize.
testcase:
test/CodeGen/Alpha/add128.ll

Alpha specifies all 4 operations as Expand, but there is no code in legalize to
do so.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct  7 09:18:49 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 7 Oct 2008 09:18:49 -0500
Subject: [LLVMbugs] [Bug 2868] ADDC ADDE not expanded
In-Reply-To: <bug-2868-206@http.llvm.org/bugs/>
Message-ID: <200810071418.m97EInKk028766@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2868


Andrew Lenharth <alenhar2 at uiuc.edu> changed:

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




--- Comment #6 from Andrew Lenharth <alenhar2 at uiuc.edu>  2008-10-07 09:18:48 ---
fixed:
http://llvm.org/viewvc/llvm-project?view=rev&revision=57243


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct  7 13:10:45 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 7 Oct 2008 13:10:45 -0500
Subject: [LLVMbugs] [Bug 2765] CloneFunction Fails to correctly set
	Attributes
In-Reply-To: <bug-2765-206@http.llvm.org/bugs/>
Message-ID: <200810071810.m97IAjvX004920@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2765


Andrew Lenharth <alenhar2 at uiuc.edu> changed:

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




--- Comment #3 from Andrew Lenharth <alenhar2 at uiuc.edu>  2008-10-07 13:10:44 ---
fix in:
http://llvm.org/viewvc/llvm-project?view=rev&revision=57254


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct  7 15:34:05 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 7 Oct 2008 15:34:05 -0500
Subject: [LLVMbugs] [Bug 2869] New: Miscompilation with -loop-index-split
Message-ID: <bug-2869-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2869

           Summary: Miscompilation with -loop-index-split
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: dalej at apple.com
                CC: llvmbugs at cs.uiuc.edu


Attachment is bugpoint reduced from gcc.c-torture/execute/930603-2.c in gcc
testsuite.
(To execute, you will need a definition of 'w' which is int[2][2] .)
Top loops should store 1 into elements [0][0] and [1][1].  After running
through -loop-index-split, it does only the first store.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct  7 18:07:11 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 7 Oct 2008 18:07:11 -0500
Subject: [LLVMbugs] [Bug 2870] New: a modified pointer should be considered
 as being read in the future
Message-ID: <bug-2870-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2870

           Summary: a modified pointer should be considered as being read in
                    the future
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: ricardo_barbano at hotmail.com
                CC: llvmbugs at cs.uiuc.edu


if you have:

void foo(char *buffer){

buffer += 20;
}

llvm says: "Value stored to 'buffer' is never read"

well, you can't assume that because the modified value can be read from the
calling function.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct  7 18:27:29 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 7 Oct 2008 18:27:29 -0500
Subject: [LLVMbugs] [Bug 2870] a modified pointer should be considered as
 being read in the future
In-Reply-To: <bug-2870-206@http.llvm.org/bugs/>
Message-ID: <200810072327.m97NRTfQ030035@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2870


Anton Korobeynikov <asl at math.spbu.ru> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |asl at math.spbu.ru
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct  9 01:55:54 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 9 Oct 2008 01:55:54 -0500
Subject: [LLVMbugs] [Bug 2871] New: bugpoint erroneously assumes exit code 1
 is bad by default
Message-ID: <bug-2871-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2871

           Summary: bugpoint erroneously assumes exit code 1 is bad by
                    default
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: bugpoint
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicholas at mxc.ca
                CC: resistor at mac.com, llvmbugs at cs.uiuc.edu


Try bugpoint'ing spiff some time. It happens to trigger this code in
ExecutionDriver.cpp:

  // If we're checking the program exit code, assume anything nonzero is bad.
  if (CheckProgramExitCode && ProgramExitedNonzero) {
    Output.eraseFromDisk();
    if (RemoveBitcode)
      sys::Path(BitcodeFile).eraseFromDisk();
    return true;
  }

which is really horrible because ProgramExitedNonzero is true even when running
the native version, but bugpoint doesn't check for that. CheckProgramExitCode
is on by default, but you can pass in -check-exit-code=false which makes it
work.

I think with the advent of -append-exit-code, this check can just be removed,
or at least changed somehow. I'm not really sure up front what to do about it.
Thoughts?


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct  9 02:49:59 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 9 Oct 2008 02:49:59 -0500
Subject: [LLVMbugs] [Bug 2872] New: bugpoint needs to try removing fastcc
	when using CBE
Message-ID: <bug-2872-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2872

           Summary: bugpoint needs to try removing fastcc when using CBE
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: bugpoint
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicholas at mxc.ca
                CC: llvmbugs at cs.uiuc.edu


The CBE doesn't (can't, never will) support the fastcc calling convention.
Currently, when it splits the module it will emit the LLVM 'to-optimize' side
with fastcc intact, but the CBE-side will silently drop fastcc. This causes an
obvious ABI mismatch and a not very useful reduction.

Before bugpoint starts splitting across the CBE, it should try removing all
fastcc's. It also needs to make sure that step doesn't, on its own, 'fix' the
problem.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct  9 06:13:03 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 9 Oct 2008 06:13:03 -0500
Subject: [LLVMbugs] [Bug 2702] Assertion failed: (InReg && "Value not in
 map!"), function getValue, file SelectionDAGISel.cpp, line 1261.
In-Reply-To: <bug-2702-206@http.llvm.org/bugs/>
Message-ID: <200810091113.m99BD3CM023966@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2702


Nicolas Capens <nicolas at capens.net> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |nicolas at capens.net
             Status|RESOLVED                    |REOPENED
         Resolution|WORKSFORME                  |




--- Comment #3 from Nicolas Capens <nicolas at capens.net>  2008-10-09 06:13:02 ---
I'm hitting this assert as well. I'll provide an LLVM IR testcase.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct  9 10:24:16 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 9 Oct 2008 10:24:16 -0500
Subject: [LLVMbugs] [Bug 2702] Assertion failed: (InReg && "Value not in
 map!"), function getValue, file SelectionDAGISel.cpp, line 1261.
In-Reply-To: <bug-2702-206@http.llvm.org/bugs/>
Message-ID: <200810091524.m99FOF86002773@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2702


Nicolas Capens <nicolas at capens.net> changed:

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




--- Comment #5 from Nicolas Capens <nicolas at capens.net>  2008-10-09 10:24:07 ---
Never mind, I'm hitting this assert due to a wrongly customized optimization
pass. I should have done some simple testing before reopening this. Sorry for
the trouble.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct  9 18:17:26 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 9 Oct 2008 18:17:26 -0500
Subject: [LLVMbugs] [Bug 2874] New: missed optimization opportunity: imull
	-> shl+add
Message-ID: <bug-2874-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2874

           Summary: missed optimization opportunity: imull -> shl+add
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: enhancement
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nlewycky at google.com
                CC: llvmbugs at cs.uiuc.edu


LLVM will optimize (x << 8) + x into x*257, but on some CPUs the imull is
actually slower than the shift and add. The backend should take a moment to
unfurl those multiplies when possible.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct  9 18:54:01 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 9 Oct 2008 18:54:01 -0500
Subject: [LLVMbugs] [Bug 2875] New: Compiling LLVM with SPARC target
 (sparc-elf) fails on crtstuff.c
Message-ID: <bug-2875-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2875

           Summary: Compiling LLVM with SPARC target (sparc-elf)  fails on
                    crtstuff.c
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: hirenvt at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2068)
 --> (http://llvm.org/bugs/attachment.cgi?id=2068)
Preprocessed file using -save-temps

Configured with:
../configure --target=sparc-elf --enable-languages=c
--prefix=$HOME/sw/llvm/sparc-llvm/ --enable-llvm=$HOME/sw/llvm/llvm-2.3

Make gives the following error:
 ./xgcc -B./ -B/home/hiren/sw/llvm/sparc-llvm//sparc-elf/bin/ -isystem
/home/hiren/sw/llvm/sparc-llvm//sparc-elf/include -isystem
/home/hiren/sw/llvm/sparc-llvm//sparc-elf/sys-include
-L/home/hiren/sw/llvm/llvm-gcc4.2-2.3.source/obj/gcc/../ld -O2  -DIN_GCC
-DCROSS_DIRECTORY_STRUCTURE   -W -Wall -Wwrite-strings -Wstrict-prototypes
-Wmissing-prototypes -Wold-style-definition  -isystem ./include  -I. -I.
-I../../gcc -I../../gcc/. -I../../gcc/../include -I../../gcc/../libcpp/include 
-I../../gcc/../libdecnumber -I../libdecnumber
-I/home/hiren/sw/llvm/llvm-2.3/include -I/home/hiren/sw/llvm/llvm-2.3/include 
-g0 -finhibit-size-directive -fno-inline-functions -fno-exceptions
-fno-zero-initialized-in-bss -fno-toplevel-reorder -Dinhibit_libc -save-temps  
-c ../../gcc/crtstuff.c -DCRT_BEGIN   -o crtbegin.o 
../../gcc/crtstuff.c:378: internal compiler error: Segmentation fault
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://developer.apple.com/bugreporter> for instructions.

Using -save-temps, the crtstuff.i is attached.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct  9 22:44:42 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 9 Oct 2008 22:44:42 -0500
Subject: [LLVMbugs] [Bug 2849] llc fails an assertion in ScheduleDAG
In-Reply-To: <bug-2849-206@http.llvm.org/bugs/>
Message-ID: <200810100344.m9A3ig40024428@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2849


Santosh <santosh.nagarakatte at gmail.com> changed:

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




-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct  9 22:59:55 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 9 Oct 2008 22:59:55 -0500
Subject: [LLVMbugs] [Bug 2874] missed optimization opportunity: imull ->
	shl+add
In-Reply-To: <bug-2874-206@http.llvm.org/bugs/>
Message-ID: <200810100359.m9A3xtkD024972@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2874


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |LATER




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-10-09 22:59:54 ---
Please move this to a README.txt file.  I think discussion of this is already
in one.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 10 00:36:28 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 10 Oct 2008 00:36:28 -0500
Subject: [LLVMbugs] [Bug 2876] New: simplifycfg won't simplify switch to
 identical blocks
Message-ID: <bug-2876-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2876

           Summary: simplifycfg won't simplify switch to identical blocks
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicholas at mxc.ca
                CC: llvmbugs at cs.uiuc.edu


Given a branch where the two target blocks are identical ("ret i32 %b" in
both), simplifycfg will simplify them away. But not so for a switch statement:

define i32 @f(i32 %a, i32 %b) nounwind readnone {
entry:
        switch i32 %a, label %bb3 [
                i32 4, label %bb
                i32 6, label %bb
        ]

bb:             ; preds = %entry, %entry
        ret i32 %b

bb3:            ; preds = %entry
        ret i32 %b
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 10 04:37:07 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 10 Oct 2008 04:37:07 -0500
Subject: [LLVMbugs] [Bug 2877] New: Clang parser ignores protocol
 information of referenced properties
Message-ID: <bug-2877-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2877

           Summary: Clang parser ignores protocol information of referenced
                    properties
           Product: clang
           Version: unspecified
          Platform: Macintosh
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: parser
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nikita at zhuk.fi
                CC: llvmbugs at cs.uiuc.edu


If a protocol declares a property and a class declares property which
implementes that protocol, then referencing the property which is declared in
the protocol through property declared in the class causes parser to reject the
source code.

A test case which succeeds with GCC but fails with clang (revision 57238):


#import <Foundation/Foundation.h>

@protocol MyProtocol <NSObject>
@property(readonly) NSString *stringValue;
@end

@interface MyClass : NSObject {
  id<MyProtocol> _myIvar;
}
@property(readonly) id<MyProtocol> ivar;
@end

@implementation MyClass
@synthesize ivar = _myIvar;
@end

int main (int argc, const char * argv[]) {
  NSAutoreleasePool * pool = [[NSAutoreleasePool alloc] init];

  MyClass *obj = [[[MyClass alloc] init] autorelease];
  NSLog(@"String via property = '%@'", obj.ivar.stringValue);

  [pool drain];
  return 0;
}


Running the above code with clang -fsyntax-only produces the following error:

ProtocolPropertyTest.m:21:51: error: member reference base type
('id<MyProtocol>') is not a structure or union
    NSLog(@"String via property = '%@'", obj.ivar.stringValue);
                                         ~~~~~~~~ ^
1 diagnostic generated.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 10 06:19:39 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 10 Oct 2008 06:19:39 -0500
Subject: [LLVMbugs] [Bug 2878] New: Calling makeLLVMModuleContents twice
 gives memory errors
Message-ID: <bug-2878-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2878

           Summary: Calling makeLLVMModuleContents twice gives memory errors
           Product: tools
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicolas.geoffray at lip6.fr
                CC: llvmbugs at cs.uiuc.edu


When calling two times the makeLLVMModuleContents function generated by the cpp
backend with -cppgen=contents, I get a memory corruption.

You can find the .ll file to compile using:
llvm-as -f essai.ll -o - | llc -march=cpp -cppgen=contents -f -o generated.cpp

And then compile essai.cpp with the appropriate llvm libraries.

The error is probably due to the use of opaque types in the
makeLLVMModuleContents function.

The error message is:
*** glibc detected *** ./a.out: malloc(): memory corruption (fast): 0x087dabe8
***
======= Backtrace: =========
/lib/tls/i686/cmov/libc.so.6[0xb7d54962]
/lib/tls/i686/cmov/libc.so.6(__libc_malloc+0x8d)[0xb7d55cad]
/usr/lib/libstdc++.so.6(_Znwj+0x27)[0xb7f20447]
./a.out(_ZNSt6vectorIPKN4llvm4TypeESaIS3_EEC1ERKS5_+0x4e)[0x85ff81e]

The trace with the core dump.

#0  0xb7f8b410 in __kernel_vsyscall ()
#1  0xb7d13085 in raise () from /lib/tls/i686/cmov/libc.so.6
#2  0xb7d14a01 in abort () from /lib/tls/i686/cmov/libc.so.6
#3  0xb7d4bb7c in ?? () from /lib/tls/i686/cmov/libc.so.6
#4  0xb7d54962 in ?? () from /lib/tls/i686/cmov/libc.so.6
#5  0xb7d55cad in malloc () from /lib/tls/i686/cmov/libc.so.6
#6  0xb7f20447 in operator new () from /usr/lib/libstdc++.so.6
#7  0x085ff81e in std::vector<llvm::Type const*, std::allocator<llvm::Type
const*> >::vector ()
#8  0x085fc3fc in llvm::FunctionType::get ()
#9  0x0865b455 in makeLLVMModuleContents (mod=0x87da320) at generated.cpp:26
#10 0x0865b737 in main () at essai.cpp:19


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 10 15:31:54 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 10 Oct 2008 15:31:54 -0500
Subject: [LLVMbugs] [Bug 1557] Inline assembly problems for generic tests on
	SPARC
In-Reply-To: <bug-1557-206@http.llvm.org/bugs/>
Message-ID: <200810102031.m9AKVsBM005947@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1557


Anton Korobeynikov <asl at math.spbu.ru> changed:

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




--- Comment #6 from Anton Korobeynikov <asl at math.spbu.ru>  2008-10-10 15:31:52 ---
This was implemented in series of patches from r57358 to r57364, please reopen
if stuff still fails for you.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 10 15:35:01 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 10 Oct 2008 15:35:01 -0500
Subject: [LLVMbugs] [Bug 2875] [meta] Unbreak SPARC target for llvm-gcc
In-Reply-To: <bug-2875-206@http.llvm.org/bugs/>
Message-ID: <200810102035.m9AKZ1b2006081@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2875


Anton Korobeynikov <asl at math.spbu.ru> changed:

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




--- Comment #7 from Anton Korobeynikov <asl at math.spbu.ru>  2008-10-10 15:35:00 ---
As of r57364 llvm-gcc builds fine for me with newlib for sparc-elf. I have no
idea,  whether it generates sane output, and leave bug hunting for someone
else.

Please, however, note, that we are not ABI-compliant with the stuff compiled by
native gcc in bunch of cases (mostly - passing / returning structs, etc), but
nowadays llvm has support of aggregates as first class things and hopefully all
these problems can be fixed by interested party.

Der Mohr hat seine Arbeit getan, der Mohr kann gehen! :)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 10 17:03:10 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 10 Oct 2008 17:03:10 -0500
Subject: [LLVMbugs] [Bug 2869] Miscompilation with -loop-index-split
In-Reply-To: <bug-2869-206@http.llvm.org/bugs/>
Message-ID: <200810102203.m9AM3ADL009735@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2869


Devang Patel <dpatel at apple.com> changed:

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




--- Comment #4 from Devang Patel <dpatel at apple.com>  2008-10-10 17:03:09 ---
Fixed.
Committed revision 57369.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 10 17:53:36 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 10 Oct 2008 17:53:36 -0500
Subject: [LLVMbugs] [Bug 2881] New: Small leak in PluginLoader.cpp.
Message-ID: <bug-2881-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2881

           Summary: Small leak in PluginLoader.cpp.
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: minor
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: jlerouge at apple.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2077)
 --> (http://llvm.org/bugs/attachment.cgi?id=2077)
Fix a leak in PluginLoader.cpp

Valgrind is showing a small leak in the PluginLoader: a static pointer is used
to store the allocated list of plugins, but is never freed. The patch below
fixes that leak by using ManagedStatic.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 10 20:51:11 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 10 Oct 2008 20:51:11 -0500
Subject: [LLVMbugs] [Bug 2882] New: scan-build doesn't handle ~/
Message-ID: <bug-2882-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2882

           Summary: scan-build doesn't handle ~/
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: ricardo_barbano at hotmail.com
                CC: llvmbugs at cs.uiuc.edu


hi,

i have the path to clang and scan-build starting with a ~/ on my profile file.
when i do scan-build it says: "can't exec 'clang': file not found in scan-build
line 95"
if i change my path to /home/user/ it works ok.
the shell doesn't expand ~/ since it comes from a file.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 10 21:40:52 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 10 Oct 2008 21:40:52 -0500
Subject: [LLVMbugs] [Bug 2882] scan-build doesn't handle ~/
In-Reply-To: <bug-2882-206@http.llvm.org/bugs/>
Message-ID: <200810110240.m9B2eqpv021120@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2882


Gordon Henriksen <gordonhenriksen at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |gordonhenriksen at mac.com
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #1 from Gordon Henriksen <gordonhenriksen at mac.com>  2008-10-10 21:40:51 ---
~ expansion is the exclusive purview of of the shell. You need to fix your
profile to set the expanded path in the environment variable. Make sure the ~
is not enclosed in quotes.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Oct 11 00:27:51 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 11 Oct 2008 00:27:51 -0500
Subject: [LLVMbugs] [Bug 2883] New: linking of alias symbol with weak symbol
	fails
Message-ID: <bug-2883-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2883

           Summary: linking of alias symbol with weak symbol fails
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Linker
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: daniel at zuster.org
                CC: llvmbugs at cs.uiuc.edu


Given a.ll
--
define weak void @foo() {
ret void
}
--
and b.ll
--
@foo = alias void ()* @fooreal
define void @fooreal() {
}
--
then linking a.ll with b.ll succeeds, but linking b.ll with
a.ll fails:
--
ddunbar at bohr:~/foo$ llvm-as -f a.ll && llvm-as -f b.ll && llvm-link -f -o c.bc
a.bc b.bc
ddunbar at bohr:~/foo$ llvm-as -f a.ll && llvm-as -f b.ll && llvm-link -f -o c.bc
b.bc a.bc
llvm-link: link error in 'a.bc': Function-Alias Collision on 'foo': symbol
multiple defined
ddunbar at bohr:~/foo$ 
--


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Oct 11 17:09:22 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 11 Oct 2008 17:09:22 -0500
Subject: [LLVMbugs] [Bug 2735] ice: Cannot yet select
In-Reply-To: <bug-2735-206@http.llvm.org/bugs/>
Message-ID: <200810112209.m9BM9MUB008557@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2735


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED
            Summary|ISD::CALL clean up needed - |ice: Cannot yet select
                   |ice: Cannot yet select      |




--- Comment #3 from Chris Lattner <clattner at apple.com>  2008-10-11 17:09:21 ---
No major change is needed, just use TargetConstant instead of Constant.  Fixed
here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081006/068353.html

Thanks John!


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Oct 11 17:55:39 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 11 Oct 2008 17:55:39 -0500
Subject: [LLVMbugs] [Bug 2697] apparent int64 bug
In-Reply-To: <bug-2697-206@http.llvm.org/bugs/>
Message-ID: <200810112255.m9BMtdUS010433@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2697


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #10 from Chris Lattner <clattner at apple.com>  2008-10-11 17:55:38 ---
Eli, you patch look great.  I applied it here after giggling it around to apply
to mainline:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081006/068355.html


John, one very interesting experiment would be to exhaustively check whether we
do the right thing for code like this:

define i1 @x(i32 %x) nounwind {
        %div = sdiv i32 %x, 65536               ; <i32> [#uses=1]
        %cmp = icmp slt i32 %div, -65536
        ret i1 %cmp
}

where you could use i16 (instead of i32 to speed things up) and check for all
constants, all compares, and both sdiv and udiv.

-Chris


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Oct 11 22:12:48 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 11 Oct 2008 22:12:48 -0500
Subject: [LLVMbugs] [Bug 2882] scan-build doesn't handle ~/
In-Reply-To: <bug-2882-206@http.llvm.org/bugs/>
Message-ID: <200810120312.m9C3CmSi020350@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2882


Ted Kremenek <kremenek at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassignedbugs at nondot.org   |kremenek at apple.com
             Status|RESOLVED                    |ASSIGNED
         Resolution|INVALID                     |




--- Comment #3 from Ted Kremenek <kremenek at apple.com>  2008-10-11 22:12:47 ---
Created an attachment (id=2080)
 --> (http://llvm.org/bugs/attachment.cgi?id=2080)
patch to get the "real path" of Clang

Ricardo: Please try this patch on scan-build to see if it correctly handles the
case where you specify the path to clang using '~'.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Oct 11 22:12:48 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 11 Oct 2008 22:12:48 -0500
Subject: [LLVMbugs] [Bug 2882] scan-build doesn't handle ~/
In-Reply-To: <bug-2882-206@http.llvm.org/bugs/>
Message-ID: <200810120312.m9C3CmxG020351@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2882


Ted Kremenek <kremenek at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassignedbugs at nondot.org   |kremenek at apple.com
             Status|RESOLVED                    |ASSIGNED
         Resolution|INVALID                     |




--- Comment #3 from Ted Kremenek <kremenek at apple.com>  2008-10-11 22:12:47 ---
Created an attachment (id=2080)
 --> (http://llvm.org/bugs/attachment.cgi?id=2080)
patch to get the "real path" of Clang

Ricardo: Please try this patch on scan-build to see if it correctly handles the
case where you specify the path to clang using '~'.


--- Comment #4 from Ted Kremenek <kremenek at apple.com>  2008-10-11 22:12:47 ---
Created an attachment (id=2081)
 --> (http://llvm.org/bugs/attachment.cgi?id=2081)
patch to get the "real path" of Clang

Ricardo: Please try this patch on scan-build to see if it correctly handles the
case where you specify the path to clang using '~'.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Oct 12 03:19:28 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 12 Oct 2008 03:19:28 -0500
Subject: [LLVMbugs] [Bug 2885] New: assertion triggered in DwarfWriter
Message-ID: <bug-2885-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2885

           Summary: assertion triggered in DwarfWriter
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nlewycky at google.com
                CC: rafael.espindola at gmail.com, llvmbugs at cs.uiuc.edu


Created an attachment (id=2082)
 --> (http://llvm.org/bugs/attachment.cgi?id=2082)
c++ testcase

This attachment compiles fine unless you specify -g on the command line:

  $ llvm-g++ -g die.cc
  cc1plus: DwarfWriter.cpp:2064: void
llvm::DwarfDebug::ConstructDefaultScope(llvm::MachineFunction*): Assertion `0
&& "Couldn't find DIE for machine function!"' failed.
  die.cc:32: internal compiler error: Aborted
  Please submit a full bug report,
  with preprocessed source if appropriate.
  See <URL:http://developer.apple.com/bugreporter> for instructions.

This happens on 32 and 64-bit Linux.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Oct 12 03:41:01 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 12 Oct 2008 03:41:01 -0500
Subject: [LLVMbugs] [Bug 2886] New: 2.4 pre-release test failures: ocaml
	bindings
Message-ID: <bug-2886-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2886

           Summary: 2.4 pre-release test failures: ocaml bindings
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: gohman at apple.com, gordonhenriksen at mac.com,
                    llvmbugs at cs.uiuc.edu
        Depends on: 2847


+++ This bug was initially created as a clone of Bug #2847 +++

Some of these affect the 2.4 pre-release.

This bug is for the .ml test failures (all of them fail)

FAIL: /home/edwin/llvm2.4/llvm-2.4/test/Bindings/Ocaml/analysis.ml
Failed with unknown error (or has stderr output) at line 1
while running: /usr/bin/ocamlc.opt -cc g++ -I
/home/edwin/llvm2.4/llvm-2.4/Release//lib/ocaml -warn-error A llvm.cma
llvm_analysis.cma /home/edwin/llvm2.4/llvm-2.4/test/Bindings/Ocaml/analysis.ml
-o analysis.ml.tmp
/tmp/camlprim60e068.c:914: warning: deprecated conversion from string constant
to ???char*???
/tmp/camlprim60e068.c:914: warning: deprecated conversion from string constant
to ???char*???
/tmp/camlprim60e068.c:914: warning: deprecated conversion from string constant
to ???char*???
/tmp/camlprim60e068.c:914: warning: deprecated conversion from string constant
to ???char*???
/tmp/camlprim60e068.c:914: warning: deprecated conversion from string constant
to ???char*???
/tmp/camlprim60e068.c:914: warning: deprecated conversion from string constant
to ???char*
...

$ ocamlc.opt -v
The Objective Caml compiler, version 3.10.2
Standard library directory: /usr/lib/ocaml/3.10.2

The test fails due to those warnings being output, otherwise the exitcode of
that command is 0.
Those warnings are an ocaml issue, but the test shouldn't fail due to them.
Thoughts?


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Oct 12 03:47:15 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 12 Oct 2008 03:47:15 -0500
Subject: [LLVMbugs] [Bug 2887] New: 2.4 pre-release test failures: Atomics
Message-ID: <bug-2887-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2887

           Summary: 2.4 pre-release test failures: Atomics
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: gohman at apple.com, llvmbugs at cs.uiuc.edu
        Depends on: 2847


+++ This bug was initially created as a clone of Bug #2847 +++

Using 2.4 prerelease:

FAIL: /home/edwin/llvm2.4/llvm-2.4/test/FrontendC/Atomics-no64bit.c
Failed with exit(1) at line 1
while running: /usr/local/bin/llvm-gcc  -emit-llvm -w -S
/home/edwin/llvm2.4/llvm-2.4/test/FrontendC/Atomics-no64bit.c -o - | grep
atomic | count 192
count: expected 192 lines and got 200.
child process exited abnormally

There are both declarations and use of llvm.atomic.* in the output.
Perhaps new atomic declarations got added and this test was not updated?

FAIL: /home/edwin/llvm2.4/llvm-2.4/test/FrontendC/Atomics.c
Failed with exit(1) at line 4
while running: /usr/local/bin/llvm-gcc  -emit-llvm -w -S
/home/edwin/llvm2.4/llvm-2.4/test/FrontendC/Atomics.c -o - | grep p0i32 | count
92
count: expected 92 lines and got 50.
child process exited abnormally


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Oct 12 13:41:46 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 12 Oct 2008 13:41:46 -0500
Subject: [LLVMbugs] [Bug 2887] 2.4 pre-release test failures: Atomics
In-Reply-To: <bug-2887-206@http.llvm.org/bugs/>
Message-ID: <200810121841.m9CIfk9v019840@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2887


Dale Johannesen <dalej at apple.com> changed:

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




--- Comment #5 from Dale Johannesen <dalej at apple.com>  2008-10-12 13:41:45 ---
OK, these tests are assuming a 32-bit host.  I've removed the "long" cases
which should make them work for you.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Oct 13 13:34:50 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 13 Oct 2008 13:34:50 -0500
Subject: [LLVMbugs] [Bug 2888] New: linux regressions with 2.4
Message-ID: <bug-2888-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2888

           Summary: linux regressions with 2.4
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: blocker
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: tonic at nondot.org
                CC: llvmbugs at cs.uiuc.edu


We have the following regressions on Linux with 2.4:
External/SPEC/CINT2000/176.gcc/176.gcc (LLC, JIT)
External/SPEC/CINT95/126.gcc/126.gcc (LLC, JIT)
External/SPEC/CINT95/132.ijpeg/132.ijpeg  (JIT)
MultiSource/Benchmarks/Trimaran/netbench-crc/netbench-crc (JIT)

I'm currently bugpointing them and will attach test cases when they are done.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Oct 13 13:36:15 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 13 Oct 2008 13:36:15 -0500
Subject: [LLVMbugs] [Bug 2775] ice: invalid register-register move
	instruction
In-Reply-To: <bug-2775-206@http.llvm.org/bugs/>
Message-ID: <200810131836.m9DIaF9S021724@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2775


Evan Cheng <evan.cheng at apple.com> changed:

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




--- Comment #12 from Evan Cheng <evan.cheng at apple.com>  2008-10-13 13:36:14 ---
Fixed.
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081013/068404.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Oct 13 13:46:43 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 13 Oct 2008 13:46:43 -0500
Subject: [LLVMbugs] [Bug 2794] ISD:: SIGN_EXTEND_INREG added during
 Legalization and not Legalized
In-Reply-To: <bug-2794-206@http.llvm.org/bugs/>
Message-ID: <200810131846.m9DIkhoi022193@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2794


Evan Cheng <evan.cheng at apple.com> changed:

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




--- Comment #6 from Evan Cheng <evan.cheng at apple.com>  2008-10-13 13:46:43 ---
Thanks. Patch and test case are committed.
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081013/068406.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Oct 13 14:50:09 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 13 Oct 2008 14:50:09 -0500
Subject: [LLVMbugs] [Bug 2890] New:
 FrontendObjC++/2007-10-03-MetadataPointers. mm assert failure
Message-ID: <bug-2890-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2890

           Summary: FrontendObjC++/2007-10-03-MetadataPointers.mm assert
                    failure
           Product: new-bugs
           Version: unspecified
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: baldrick at free.fr
                CC: llvmbugs at cs.uiuc.edu


This test fails on x86-32-linux with:

while running: /usr/local/bin/llvm-gcc  -emit-llvm -w -w -x objective-c++ -c
llvm/test/FrontendObjC++/2007-10-03-MetadataPointers.mm -o /dev/null            
cc1objplus: llvm/include/llvm/ADT/ilist.h:141:
llvm::ilist_iterator<NodeTy>::operator typename bidirectional_iterator<NodeTy,
int>::pointer() const [with NodeTy = llvm::Argument]: Assertion
`Traits::getNext(NodePtr) != 0 && "Dereferencing end()!"' failed.
llvm/test/FrontendObjC++/2007-10-03-MetadataPointers.mm: In function 'void
__objc_gnu_init()':                    
llvm/test/FrontendObjC++/2007-10-03-MetadataPointers.mm:7: internal compiler
error: Aborted


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Oct 13 15:33:30 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 13 Oct 2008 15:33:30 -0500
Subject: [LLVMbugs] [Bug 2891] New: darwin-x86 regressions with 2.4
Message-ID: <bug-2891-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2891

           Summary: darwin-x86 regressions with 2.4
           Product: new-bugs
           Version: unspecified
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: blocker
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: tonic at nondot.org
                CC: llvmbugs at cs.uiuc.edu


There are currently two regressions on darwin, x86:
External/SPEC/CFP2000/188.ammp/188.ammp (CBE)
MultiSource/Benchmarks/NPB-serial/is/is (JIT)

I'm not sure if we care about the CBE failure, but the JIT one is more
interesting. Can anyone look at these?


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Oct 13 18:39:06 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 13 Oct 2008 18:39:06 -0500
Subject: [LLVMbugs] [Bug 2849] llc fails an assertion in ScheduleDAG
In-Reply-To: <bug-2849-206@http.llvm.org/bugs/>
Message-ID: <200810132339.m9DNd6qC002966@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2849


Evan Cheng <evan.cheng at apple.com> changed:

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




--- Comment #5 from Evan Cheng <evan.cheng at apple.com>  2008-10-13 18:39:05 ---
Dan fixed it:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081013/068414.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 00:20:44 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 00:20:44 -0500
Subject: [LLVMbugs] [Bug 2895] New: Include paths seem to be wrong in some
 files in lib/target/ x86/ASMPrinter
Message-ID: <bug-2895-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2895

           Summary: Include paths seem to be wrong in some files in
                    lib/target/x86/ASMPrinter
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: other
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nick at indigorenderer.com
                CC: llvmbugs at cs.uiuc.edu


Include paths seem to be wrong in some files in lib/target/x86/ASMPrinter
Built with visual studio 2008 on Vista 64 bit.
Should be fixed by this patch:


Index: Target/X86/AsmPrinter/X86ATTAsmPrinter.cpp
===================================================================
--- Target/X86/AsmPrinter/X86ATTAsmPrinter.cpp  (revision 57552)
+++ Target/X86/AsmPrinter/X86ATTAsmPrinter.cpp  (working copy)
@@ -15,11 +15,11 @@

 #define DEBUG_TYPE "asm-printer"
 #include "X86ATTAsmPrinter.h"
-#include "X86.h"
-#include "X86COFF.h"
-#include "X86MachineFunctionInfo.h"
-#include "X86TargetMachine.h"
-#include "X86TargetAsmInfo.h"
+#include "../X86.h"
+#include "../X86COFF.h"
+#include "../X86MachineFunctionInfo.h"
+#include "../X86TargetMachine.h"
+#include "../X86TargetAsmInfo.h"
 #include "llvm/CallingConv.h"
 #include "llvm/DerivedTypes.h"
 #include "llvm/Module.h"
Index: Target/X86/AsmPrinter/X86IntelAsmPrinter.cpp
===================================================================
--- Target/X86/AsmPrinter/X86IntelAsmPrinter.cpp        (revision 57552)
+++ Target/X86/AsmPrinter/X86IntelAsmPrinter.cpp        (working copy)
@@ -15,9 +15,9 @@

 #define DEBUG_TYPE "asm-printer"
 #include "X86IntelAsmPrinter.h"
-#include "X86InstrInfo.h"
-#include "X86TargetAsmInfo.h"
-#include "X86.h"
+#include "../X86InstrInfo.h"
+#include "../X86TargetAsmInfo.h"
+#include "../X86.h"
 #include "llvm/CallingConv.h"
 #include "llvm/Constants.h"
 #include "llvm/DerivedTypes.h"
Index: Target/X86/AsmPrinter/X86AsmPrinter.cpp
===================================================================
--- Target/X86/AsmPrinter/X86AsmPrinter.cpp     (revision 57552)
+++ Target/X86/AsmPrinter/X86AsmPrinter.cpp     (working copy)
@@ -16,7 +16,7 @@

 #include "X86ATTAsmPrinter.h"
 #include "X86IntelAsmPrinter.h"
-#include "X86Subtarget.h"
+#include "../X86Subtarget.h"
 using namespace llvm;

 /// createX86CodePrinterPass - Returns a pass that prints the X86 assembly
code


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 01:17:20 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 01:17:20 -0500
Subject: [LLVMbugs] [Bug 2733] Can't use named primitive type in structure
	definition
In-Reply-To: <bug-2733-206@http.llvm.org/bugs/>
Message-ID: <200810150617.m9F6HKoW012353@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2733


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-10-15 01:17:20 ---
Fixed, patch here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081013/068491.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 02:12:00 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 02:12:00 -0500
Subject: [LLVMbugs] [Bug 2895] Include paths seem to be wrong in some files
 in lib/target/x86/ ASMPrinter
In-Reply-To: <bug-2895-206@http.llvm.org/bugs/>
Message-ID: <200810150712.m9F7C0fa014401@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2895


Anton Korobeynikov <asl at math.spbu.ru> changed:

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




--- Comment #2 from Anton Korobeynikov <asl at math.spbu.ru>  2008-10-15 02:11:59 ---
Paths are ok. There was some discussion in llvm-commits mailing list about this
this summer. You should tune your project file for right include directories.
AFAIR, current project files do the right thing


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct 14 01:11:01 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 14 Oct 2008 01:11:01 -0500
Subject: [LLVMbugs] [Bug 2892] New: crash on redefinition of builtin
Message-ID: <bug-2892-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2892

           Summary: crash on redefinition of builtin
           Product: clang
           Version: unspecified
          Platform: All
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: daniel at zuster.org
                CC: llvmbugs at cs.uiuc.edu


clang asserts on the following input:
--
void __builtin_object_size();

int a[10];

int f0() {
  return __builtin_object_size(&a); // expected-error {{too few arguments to
function}}
}
--
this is because although we emit an error on the redefinition of the builtin,
we continue parsing. However, the builtin code which hand checks builtin object
size expects the number of arguments to have already been checked.

Its not clear what the right fix is. Its easy to fix in the builtin code,
however, it is also nice to rely on builtins always obeying certain invariants.
Perhaps we should disallow redefinition of builtins?


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct 14 01:43:19 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 14 Oct 2008 01:43:19 -0500
Subject: [LLVMbugs] [Bug 2893] New: Parsing error (type of property does not
 match type of ivar ) if property conforms to protocol and ivar doesn't
Message-ID: <bug-2893-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2893

           Summary: Parsing error (type of property does not match type of
                    ivar) if property conforms to protocol and ivar doesn't
           Product: clang
           Version: unspecified
          Platform: Macintosh
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: parser
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nikita at zhuk.fi
                CC: llvmbugs at cs.uiuc.edu


The following test case produces clang parser error, but compiles without
warnings with GCC 4.0:

#import <Foundation/Foundation.h>

@protocol MyProtocol
- (NSString *)stringValue;
@end

@interface MyClass : NSObject {
  id  _myIvar;
}
@property (readwrite, retain) id<MyProtocol> myIvar;
@end

@implementation MyClass
@synthesize myIvar = _myIvar;
@end

/tmp/iVarProtTest/MyClass.m:22:13: error: type of property 'myIvar'  does not
match type of ivar '_myIvar'
@synthesize myIvar = _myIvar;
            ^
1 diagnostic generated.


However, when the property is changed from:
@property (readwrite, retain)
to:
@property (readonly)

then this error goes away.

I'm not completely sure if this is intended behavior or not. However, since GCC
doesn't give any warnings about this, I assume that this could be something you
might be interested in.

Clang revision: 57479


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct 14 01:50:16 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 14 Oct 2008 01:50:16 -0500
Subject: [LLVMbugs] [Bug 2894] New: space missing in .ll for invoke
Message-ID: <bug-2894-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2894

           Summary: space missing in .ll for invoke
           Product: new-bugs
           Version: unspecified
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: baldrick at free.fr
                CC: llvmbugs at cs.uiuc.edu


Disassemble bitcode containing invoke and observe the missing spaces:
        invoke void at g()

Here's a testcase.  Run llvm-as testcase.ll -o - | llvm-dis -o -
to watch the space disappear!

declare void @g()
define void @f() {
        invoke void @g()
                        to label %c unwind label %c
c:
        ret void
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct 14 02:00:19 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 14 Oct 2008 02:00:19 -0500
Subject: [LLVMbugs] [Bug 2797] infinite recursion in fold
In-Reply-To: <bug-2797-206@http.llvm.org/bugs/>
Message-ID: <200810140700.m9E70JPm017918@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2797


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED
            Summary|Internal error: Segmentation|infinite recursion in fold
                   |fault (program cc1)         |




--- Comment #15 from Chris Lattner <clattner at apple.com>  2008-10-14 02:00:18 ---
Fixed in r57482.
Testcase here: test/FrontendC/2008-10-13-FrontendCrash.c

For the record, fold (and its poorly documented invariants) is a horrible
horrible thing to have to wrap your brain around.

-Chris


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 10:40:26 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 10:40:26 -0500
Subject: [LLVMbugs] [Bug 732] ABI mismatch passing/ receiving non-native GCC
 generic vectors to functions
In-Reply-To: <bug-732-206@http.llvm.org/bugs/>
Message-ID: <200810151540.m9FFeQsk009040@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=732


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #2 from Chris Lattner <clattner at apple.com>  2008-10-15 10:40:25 ---
This was fixed a long time ago


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 10:41:32 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 10:41:32 -0500
Subject: [LLVMbugs] [Bug 781] llvm-gcc4: bool on darwin/ppc turns into uint
 for function args/ results
In-Reply-To: <bug-781-206@http.llvm.org/bugs/>
Message-ID: <200810151541.m9FFfWpj009079@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=781


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #5 from Chris Lattner <clattner at apple.com>  2008-10-15 10:41:32 ---
This is incredibly minor and fixed in clang, closing.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 10:46:04 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 10:46:04 -0500
Subject: [LLVMbugs] [Bug 1112] llvm-objdump needed?
In-Reply-To: <bug-1112-206@http.llvm.org/bugs/>
Message-ID: <200810151546.m9FFk4Xo009267@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1112


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #3 from Chris Lattner <clattner at apple.com>  2008-10-15 10:46:03 ---
Configuring llvm-gcc works fine for me, I haven't heard reports of this since
early 2007.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 10:47:47 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 10:47:47 -0500
Subject: [LLVMbugs] [Bug 1249] Miscompilation passing struct by value to
 varargs function
In-Reply-To: <bug-1249-206@http.llvm.org/bugs/>
Message-ID: <200810151547.m9FFllxA009338@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1249


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #18 from Chris Lattner <clattner at apple.com>  2008-10-15 10:47:46 ---
This works for me now.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 10:49:19 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 10:49:19 -0500
Subject: [LLVMbugs] [Bug 1168] LLVM doesn't support "complex" type in "asm"
	statements
In-Reply-To: <bug-1168-206@http.llvm.org/bugs/>
Message-ID: <200810151549.m9FFnJGl009421@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1168


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #4 from Chris Lattner <clattner at apple.com>  2008-10-15 10:49:19 ---
GCC also rejects this:

$ gcc-4.0 t.c -S -o -
t.c: In function ???f???:
t.c:4: error: can't find a register in class ???GENERAL_REGS??? while reloading
???asm???
$ gcc-4.2 t.c -S -o -
t.c: In function ???f???:
t.c:4: error: can't find a register in class ???GENERAL_REGS??? while reloading
???asm???
t.c:4: error: ???asm??? operand has impossible constraints


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 10:53:46 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 10:53:46 -0500
Subject: [LLVMbugs] [Bug 1369] [libsystem] Implement better support for new
 MemoryBuffer class
In-Reply-To: <bug-1369-206@http.llvm.org/bugs/>
Message-ID: <200810151553.m9FFrkYQ009621@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1369


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #6 from Chris Lattner <clattner at apple.com>  2008-10-15 10:53:46 ---
I fixed this a long time ago when I ripped out the mmapfile abstraction.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 10:55:01 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 10:55:01 -0500
Subject: [LLVMbugs] [Bug 1373] Some method to represent alias information in
 LLVM would be useful
In-Reply-To: <bug-1373-206@http.llvm.org/bugs/>
Message-ID: <200810151555.m9FFt1vl009675@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1373


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #11 from Chris Lattner <clattner at apple.com>  2008-10-15 10:55:01 ---
We now support the noalias attribute, and have another PR to track the work for
TBAA, closing as fixed.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 10:58:10 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 10:58:10 -0500
Subject: [LLVMbugs] [Bug 1407] BitCodeFormat documentation missing/incomplete
In-Reply-To: <bug-1407-206@http.llvm.org/bugs/>
Message-ID: <200810151558.m9FFwAc2009861@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1407


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #2 from Chris Lattner <clattner at apple.com>  2008-10-15 10:58:10 ---
The doc has a lot of details about the bitcode format now.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 11:02:29 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 11:02:29 -0500
Subject: [LLVMbugs] [Bug 1488] bithacks: optimize code that counts bits
In-Reply-To: <bug-1488-206@http.llvm.org/bugs/>
Message-ID: <200810151602.m9FG2TKL010059@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1488


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |LATER




--- Comment #10 from Chris Lattner <clattner at apple.com>  2008-10-15 11:02:28 ---
I moved these ideas into lib/Target/README.txt


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 11:03:46 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 11:03:46 -0500
Subject: [LLVMbugs] [Bug 1560] --version should include svn revision number
In-Reply-To: <bug-1560-206@http.llvm.org/bugs/>
Message-ID: <200810151603.m9FG3koe010130@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1560


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #4 from Chris Lattner <clattner at apple.com>  2008-10-15 11:03:46 ---
See above,


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 11:04:36 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 11:04:36 -0500
Subject: [LLVMbugs] [Bug 1578] JIT doesn't support exception handling
In-Reply-To: <bug-1578-206@http.llvm.org/bugs/>
Message-ID: <200810151604.m9FG4a47010191@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1578


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #4 from Chris Lattner <clattner at apple.com>  2008-10-15 11:04:36 ---
Nicolas G implemented JIT EH support, it apparently works on linux and darwin. 
He's using it for vmkit.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 11:06:25 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 11:06:25 -0500
Subject: [LLVMbugs] [Bug 1168] LLVM doesn't support "complex" type in "asm"
	statements
In-Reply-To: <bug-1168-206@http.llvm.org/bugs/>
Message-ID: <200810151606.m9FG6PK0010366@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1168


Anton Korobeynikov <asl at math.spbu.ru> changed:

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




--- Comment #6 from Anton Korobeynikov <asl at math.spbu.ru>  2008-10-15 11:06:25 ---
(In reply to comment #4)
> GCC also rejects this:
> 
> $ gcc-4.0 t.c -S -o -
> t.c: In function ???f???:
> t.c:4: error: can't find a register in class ???GENERAL_REGS??? while reloading
> ???asm???
> $ gcc-4.2 t.c -S -o -
> t.c: In function ???f???:
> t.c:4: error: can't find a register in class ???GENERAL_REGS??? while reloading
> ???asm???
> t.c:4: error: ???asm??? operand has impossible constraints
This is bug in apple gcc then.

I have:
1. fsf gcc 4.2.4 on x86-64/linux:
<compiles fine, output attached>
2. llvm-gcc 4.2
Couldn't allocate output reg for constraint 'r'!
3. Both gcc 4.0 and 4.2 on darwin
t.c:4: error: can't find a register in class ???GENERAL_REGS??? while reloading


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 11:06:44 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 11:06:44 -0500
Subject: [LLVMbugs] [Bug 1604] need better mod/ref info for scanf
In-Reply-To: <bug-1604-206@http.llvm.org/bugs/>
Message-ID: <200810151606.m9FG6ikC010427@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1604


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |LATER




--- Comment #3 from Chris Lattner <clattner at apple.com>  2008-10-15 11:06:43 ---
No, LibCallAliasAnalysis is not an AA impl.  This bug is a very minor thing,
and requires knowing that "%d" stores to 4 bytes etc.  I moved this to
lib/Target/README.txt, which is where we keep assorted minor hacks.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 11:07:51 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 11:07:51 -0500
Subject: [LLVMbugs] [Bug 1624] OPT_BETA
In-Reply-To: <bug-1624-206@http.llvm.org/bugs/>
Message-ID: <200810151607.m9FG7pHu010505@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1624


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #3 from Chris Lattner <clattner at apple.com>  2008-10-15 11:07:51 ---
OPT_BETA in the makefiles is all we "need".  We don't need the nightly tester
to track OPT_BETA (or LLC_BETA for that matter) over time.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 11:09:42 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 11:09:42 -0500
Subject: [LLVMbugs] [Bug 1703] Teach the coalescing how to remove subreg
	instructions
In-Reply-To: <bug-1703-206@http.llvm.org/bugs/>
Message-ID: <200810151609.m9FG9gYX010685@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1703


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-10-15 11:09:42 ---
This bug has no contents, and the coalescer handles at least some cases now.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 11:10:12 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 11:10:12 -0500
Subject: [LLVMbugs] [Bug 1704] Add tablegen syntax support for selecting to
 subreg nodes in . td files
In-Reply-To: <bug-1704-206@http.llvm.org/bugs/>
Message-ID: <200810151610.m9FGACmS010752@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1704


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-10-15 11:10:12 ---
We have extract_subreg/insert_subreg now.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 11:10:56 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 11:10:56 -0500
Subject: [LLVMbugs] [Bug 1711] linux/x84-64 codegen problem (relocation
	R_X86_64_PC32)
In-Reply-To: <bug-1711-206@http.llvm.org/bugs/>
Message-ID: <200810151610.m9FGAuTP010819@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1711


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #31 from Chris Lattner <clattner at apple.com>  2008-10-15 11:10:56 ---
PIC reportedly works on linux/x86-64 now.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 11:14:29 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 11:14:29 -0500
Subject: [LLVMbugs] [Bug 1736] Frontend incompatibility precluding use of
 llvm-gcc as a drop-in replacement of gcc
In-Reply-To: <bug-1736-206@http.llvm.org/bugs/>
Message-ID: <200810151614.m9FGETvi011233@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1736


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|                            |WONTFIX




--- Comment #9 from Chris Lattner <clattner at apple.com>  2008-10-15 11:14:28 ---
We don't have plans to add the old broken semantics yet, I suggest upgrading to
a newer and less buggy glibc if this is important to you.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 11:17:04 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 11:17:04 -0500
Subject: [LLVMbugs] [Bug 1753] instcombine not transforming inttoptr/
 ptrtoint to getelementptr for global variable
In-Reply-To: <bug-1753-206@http.llvm.org/bugs/>
Message-ID: <200810151617.m9FGH4tK011523@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1753


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #7 from Chris Lattner <clattner at apple.com>  2008-10-15 11:17:04 ---
instcombine now compiles these to the same code.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 11:19:04 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 11:19:04 -0500
Subject: [LLVMbugs] [Bug 1798] wrong integer overflow calculation on signed/
 unsigned int when using optimizer
In-Reply-To: <bug-1798-206@http.llvm.org/bugs/>
Message-ID: <200810151619.m9FGJ4QM011765@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1798


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #25 from Chris Lattner <clattner at apple.com>  2008-10-15 11:19:04 ---
The miscompilation in this bug is fixed.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 11:28:26 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 11:28:26 -0500
Subject: [LLVMbugs] [Bug 1856] atomic builtins
In-Reply-To: <bug-1856-206@http.llvm.org/bugs/>
Message-ID: <200810151628.m9FGSQSW012225@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1856


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #11 from Chris Lattner <clattner at apple.com>  2008-10-15 11:28:26 ---
Yep, this should be closed now.  Thanks Dale!


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 11:30:52 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 11:30:52 -0500
Subject: [LLVMbugs] [Bug 1877] add needs to be commuted to eliminate copy in
 critical loops
In-Reply-To: <bug-1877-206@http.llvm.org/bugs/>
Message-ID: <200810151630.m9FGUqL7012325@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1877


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #9 from Chris Lattner <clattner at apple.com>  2008-10-15 11:30:52 ---
Evan fixed this a long time ago, I now get this inner loop:

        .align  4,0x90
LBB1_1: ## bb
        cvtsi2ss        (%esi,%edx,4), %xmm1
        incl    %edx
        cmpl    %ecx, %edx
        addss   %xmm1, %xmm0
        jne     LBB1_1  ## bb


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 11:31:46 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 11:31:46 -0500
Subject: [LLVMbugs] [Bug 1910] llvm-ld does not support linking archives
 with mixed .bc and . o objects
In-Reply-To: <bug-1910-206@http.llvm.org/bugs/>
Message-ID: <200810151631.m9FGVkSm012371@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1910


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WONTFIX




--- Comment #5 from Chris Lattner <clattner at apple.com>  2008-10-15 11:31:46 ---
We don't plan to extend llvm-ld to support this.  Doing so would require it to
understand tons of details of native .o formats.  It is better to extend your
native linker to use liblto.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 11:34:00 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 11:34:00 -0500
Subject: [LLVMbugs] [Bug 1941] instcombine doesn't merge mixed sign-unsign
	comparisons
In-Reply-To: <bug-1941-206@http.llvm.org/bugs/>
Message-ID: <200810151634.m9FGY0Tu012464@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1941


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |LATER




--- Comment #2 from Chris Lattner <clattner at apple.com>  2008-10-15 11:33:59 ---
Moved this to target/readme.txt


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 11:38:56 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 11:38:56 -0500
Subject: [LLVMbugs] [Bug 2031] dominator info correctly is not updated
	correctly
In-Reply-To: <bug-2031-206@http.llvm.org/bugs/>
Message-ID: <200810151638.m9FGcu7J012692@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2031


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #4 from Chris Lattner <clattner at apple.com>  2008-10-15 11:38:56 ---
This works now.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 11:39:59 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 11:39:59 -0500
Subject: [LLVMbugs] [Bug 2037] LLVM-GCC4. 2 instructions forget to mention
 --program-prefix option.
In-Reply-To: <bug-2037-206@http.llvm.org/bugs/>
Message-ID: <200810151639.m9FGdxEV012762@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2037


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Depends on|1436                        |
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Chris Lattner <clattner at apple.com>  2008-10-15 11:39:58 ---
This is fixed, and doesn't depend on 1436.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 11:40:43 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 11:40:43 -0500
Subject: [LLVMbugs] [Bug 2038] Ruby-1. 9 builds and passes light checks but
 explodes in the regression suite.
In-Reply-To: <bug-2038-206@http.llvm.org/bugs/>
Message-ID: <200810151640.m9FGeh2M012837@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2038


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #2 from Chris Lattner <clattner at apple.com>  2008-10-15 11:40:43 ---
Is this still broken?  My understanding was that this was a longjmp problem
that was fixed.  Please reverify and reopen if still broken.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 11:42:37 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 11:42:37 -0500
Subject: [LLVMbugs] [Bug 2064] redefinition of typedef'ed types
In-Reply-To: <bug-2064-206@http.llvm.org/bugs/>
Message-ID: <200810151642.m9FGgbUk012940@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2064


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED
            Summary|redifinition of typedef'ed  |redefinition of typedef'ed
                   |types                       |types




--- Comment #3 from Chris Lattner <clattner at apple.com>  2008-10-15 11:42:37 ---
I believe that this was a problem tracking whether we were parsing a "system
header" or not.  This still won't work with old .i files, but should work with
the original sources now.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 11:56:35 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 11:56:35 -0500
Subject: [LLVMbugs] [Bug 2115] please document llvm::Use
In-Reply-To: <bug-2115-206@http.llvm.org/bugs/>
Message-ID: <200810151656.m9FGuZJn013519@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2115


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #2 from Chris Lattner <clattner at apple.com>  2008-10-15 11:56:35 ---
Fixed.
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081013/068526.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 11:59:18 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 11:59:18 -0500
Subject: [LLVMbugs] [Bug 2117] Codegen constraint improvements
In-Reply-To: <bug-2117-206@http.llvm.org/bugs/>
Message-ID: <200810151659.m9FGxI4r013664@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2117


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #9 from Chris Lattner <clattner at apple.com>  2008-10-15 11:59:17 ---
I assume this works, we support atomic ops now, and use physreg constraints all
over the place.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 12:09:36 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 12:09:36 -0500
Subject: [LLVMbugs] [Bug 2211] Regenerating LibDeps.txt.tmp fails on AIX 5.3
In-Reply-To: <bug-2211-206@http.llvm.org/bugs/>
Message-ID: <200810151709.m9FH9aBQ014188@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2211


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |LATER




--- Comment #4 from Chris Lattner <clattner at apple.com>  2008-10-15 12:09:35 ---
Noone has the ability to fix this, if you have a patch, please let us know.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 12:11:29 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 12:11:29 -0500
Subject: [LLVMbugs] [Bug 2184] internal compiler error in
	lhd_set_decl_assembler_name
In-Reply-To: <bug-2184-206@http.llvm.org/bugs/>
Message-ID: <200810151711.m9FHBTu1014310@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2184


Devang Patel <dpatel at apple.com> changed:

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




--- Comment #4 from Devang Patel <dpatel at apple.com>  2008-10-15 12:11:29 ---
Now, I can not reproduce this.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 12:26:07 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 12:26:07 -0500
Subject: [LLVMbugs] [Bug 1803] Debug info doesn't work on x86-64
In-Reply-To: <bug-1803-206@http.llvm.org/bugs/>
Message-ID: <200810151726.m9FHQ7CZ014767@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1803


T??r??k Edwin <edwintorok at gmail.com> changed:

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




--- Comment #27 from T??r??k Edwin <edwintorok at gmail.com>  2008-10-15 12:26:06 ---
(In reply to comment #26)
> What is missing here?  Debug info at -O1+ is covered by another bug.  EH issues
> are covered by other bugs.  Does debug info at -O0 work now?
> 

Thanks for the ping ;)

Debug info looks reasonable now.
I loaded some apps in gdb with llvm-gcc -g, and the backtrace looks ok.

If I find any more bugs about debug info, I'll open another bug.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 13:16:12 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 13:16:12 -0500
Subject: [LLVMbugs] [Bug 1835] Register pressure reduction list schedulers
 are extremely slow on very long basic blocks
In-Reply-To: <bug-1835-206@http.llvm.org/bugs/>
Message-ID: <200810151816.m9FIGCOa016493@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1835


Dan Gohman <gohman at apple.com> changed:

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




--- Comment #7 from Dan Gohman <gohman at apple.com>  2008-10-15 13:16:12 ---
IsReachable was rewritten a while ago to be non-recursive, to
use a topological ordering to eliminate many unreachable cases
quickly, and to not use a SmallPtrSet. Compile time on the given
testcase is reasonable, with scheduling taking a reasonable
portion of the overall time.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 13:27:41 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 13:27:41 -0500
Subject: [LLVMbugs] [Bug 1638] PPC condition register (CR) spilling bug
In-Reply-To: <bug-1638-206@http.llvm.org/bugs/>
Message-ID: <200810151827.m9FIRflF016900@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1638


Bill Wendling <wendling at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |wendling at apple.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #11 from Bill Wendling <wendling at apple.com>  2008-10-15 13:27:40 ---
This appears to be fixed now. I added testcase CodeGen/PowerPC/cr_spilling.ll


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 15:23:43 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 15:23:43 -0500
Subject: [LLVMbugs] [Bug 2883] linking of alias symbol with weak symbol fails
In-Reply-To: <bug-2883-206@http.llvm.org/bugs/>
Message-ID: <200810152023.m9FKNhVe021276@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2883


Anton Korobeynikov <asl at math.spbu.ru> changed:

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




--- Comment #2 from Anton Korobeynikov <asl at math.spbu.ru>  2008-10-15 15:23:38 ---
(In reply to comment #1)
> Possibly just a matter of changing LinkModules:988 from
Yes, you're right. Somehow comment was correct, but not code :)

Fixed in:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081013/068529.html

also, same problem was fixed for global variables


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 15:27:38 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 15:27:38 -0500
Subject: [LLVMbugs] [Bug 2463] linking weak and alias fails
In-Reply-To: <bug-2463-206@http.llvm.org/bugs/>
Message-ID: <200810152027.m9FKRcOV021451@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2463


Anton Korobeynikov <asl at math.spbu.ru> changed:

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




--- Comment #5 from Anton Korobeynikov <asl at math.spbu.ru>  2008-10-15 15:27:38 ---
This was fixed in

http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081013/068529.html
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081013/068530.html

Please reopen, if it still fails for you


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 15 23:51:22 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Oct 2008 23:51:22 -0500
Subject: [LLVMbugs] [Bug 2897] New: Miscompilation when SSE2 is disabled
Message-ID: <bug-2897-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2897

           Summary: Miscompilation when SSE2 is disabled
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: andersca at mac.com
                CC: llvmbugs at cs.uiuc.edu


The following reduction:

define <4 x float> @_mm_loadh_pi(i64* %p) nounwind readonly {
entry:
        %conv2 = load i64* %p                ; <i64> [#uses=1]
        %insert = insertelement <2 x i64> undef, i64 %conv2, i32 0             
  ; <<2 x i64>> [#uses=1]
        %cast3 = bitcast <2 x i64> %insert to <4 x float>                ; <<4
x float>> [#uses=1]
        ret <4 x float> %cast3
}

is (with sse2 disabled) compiled as 

MacBook-Pro:clang andersca$ llc -mattr=-sse2 test.bc -o -


        .text
        .align        4,0x90
        .globl        __mm_loadh_pi
__mm_loadh_pi:
        subl        $36, %esp
        ## implicit-def: mm0
        movq        %mm0, 16(%esp)
        movl        40(%esp), %eax
        movl        4(%eax), %ecx
        movl        %ecx, 4(%esp)
        movl        (%eax), %eax
        movl        %eax, (%esp)
        movq        (%esp), %mm0
        movq        %mm0, 8(%esp)
        movaps        8(%esp), %xmm0
        addl        $36, %esp
        ret


and when SSE2 is enabled, compiled as

MacBook-Pro:clang andersca$ llc test.bc -o -


        .text
        .align        4,0x90
        .globl        __mm_loadh_pi
__mm_loadh_pi:
        subl        $28, %esp
        movl        32(%esp), %eax
        movl        4(%eax), %ecx
        movl        %ecx, 4(%esp)
        movl        (%eax), %eax
        movl        %eax, (%esp)
        movaps        (%esp), %xmm0
        addl        $28, %esp
        ret

        .subsections_via_symbols

Note that the second result does have no SSE2 instructions, and the first
result is also incorrect.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 00:27:08 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 00:27:08 -0500
Subject: [LLVMbugs] [Bug 2165] Constant folding should fold "<1 x i64>
 bitcast (i64 63 to < 1 x i64>)"
In-Reply-To: <bug-2165-206@http.llvm.org/bugs/>
Message-ID: <200810160527.m9G5R8Yg008199@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2165


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #4 from Chris Lattner <clattner at apple.com>  2008-10-16 00:27:08 ---
Fixed:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081013/068559.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 00:35:11 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 00:35:11 -0500
Subject: [LLVMbugs] [Bug 2366] clamav miscompiles on linux/x86
In-Reply-To: <bug-2366-206@http.llvm.org/bugs/>
Message-ID: <200810160535.m9G5ZB3q008626@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2366


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #4 from Chris Lattner <clattner at apple.com>  2008-10-16 00:35:11 ---
Works for me


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 00:36:49 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 00:36:49 -0500
Subject: [LLVMbugs] [Bug 2409] Fail to compile on SunOs5.8 with gcc-4.1.2
In-Reply-To: <bug-2409-206@http.llvm.org/bugs/>
Message-ID: <200810160536.m9G5anMG008704@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2409


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |LATER




--- Comment #3 from Chris Lattner <clattner at apple.com>  2008-10-16 00:36:48 ---
This is easy to fix, but noone is apparently interested in making the obvious
patch, and I can't test it.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 00:38:05 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 00:38:05 -0500
Subject: [LLVMbugs] [Bug 2557] Anders pass asserts on
	extractvalue/insertvalue
In-Reply-To: <bug-2557-206@http.llvm.org/bugs/>
Message-ID: <200810160538.m9G5c5Mt008772@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2557


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-10-16 00:38:05 ---


*** This bug has been marked as a duplicate of bug 2451 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 00:40:18 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 00:40:18 -0500
Subject: [LLVMbugs] [Bug 2467] Invalid instcombine of inttoptr -> gep
In-Reply-To: <bug-2467-206@http.llvm.org/bugs/>
Message-ID: <200810160540.m9G5eIfm008913@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2467


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-10-16 00:40:18 ---


*** This bug has been marked as a duplicate of bug 2831 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 00:44:48 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 00:44:48 -0500
Subject: [LLVMbugs] [Bug 2511] Assertion `!OpInfo.AssignedRegs.Regs.empty()
 && "Couldn' t allocate input reg!"' failed.
In-Reply-To: <bug-2511-206@http.llvm.org/bugs/>
Message-ID: <200810160544.m9G5im8x009157@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2511


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #3 from Chris Lattner <clattner at apple.com>  2008-10-16 00:44:48 ---


*** This bug has been marked as a duplicate of bug 1779 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 00:53:50 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 00:53:50 -0500
Subject: [LLVMbugs] [Bug 2559] Finish first-class aggregates support
In-Reply-To: <bug-2559-206@http.llvm.org/bugs/>
Message-ID: <200810160553.m9G5roes009601@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2559


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Depends on|2558                        |
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #7 from Chris Lattner <clattner at apple.com>  2008-10-16 00:53:48 ---
I don't consider Bug 2558 as a blocker for this bug.  It is a feature
enhancement request that stands on its own, the feature is done.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 00:55:19 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 00:55:19 -0500
Subject: [LLVMbugs] [Bug 2565] assertion "Src->getType()->isFloatingPoint()
 && " Unknown real conversion"" failed
In-Reply-To: <bug-2565-206@http.llvm.org/bugs/>
Message-ID: <200810160555.m9G5tJPh009690@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2565


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #2 from Chris Lattner <clattner at apple.com>  2008-10-16 00:55:18 ---
This got fixed somewhere along the way.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 00:57:29 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 00:57:29 -0500
Subject: [LLVMbugs] [Bug 2566] Combine of bitcast and shuffle assumes vector
 to vector cast
In-Reply-To: <bug-2566-206@http.llvm.org/bugs/>
Message-ID: <200810160557.m9G5vTi8009800@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2566


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|                            |LATER




--- Comment #11 from Chris Lattner <clattner at apple.com>  2008-10-16 00:57:28 ---
Please provide a testcase.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 01:01:04 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 01:01:04 -0500
Subject: [LLVMbugs] [Bug 2576] Type check llvm.prefetch intrinsic
In-Reply-To: <bug-2576-206@http.llvm.org/bugs/>
Message-ID: <200810160601.m9G614hP009941@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2576


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-10-16 01:01:04 ---
Alright, here ya go:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081013/068562.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 01:09:53 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 01:09:53 -0500
Subject: [LLVMbugs] [Bug 2604] mfocrf definition in PPCInstrInfo.td
 contradicts PowerPC spec
In-Reply-To: <bug-2604-206@http.llvm.org/bugs/>
Message-ID: <200810160609.m9G69rcL010325@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2604


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #2 from Chris Lattner <clattner at apple.com>  2008-10-16 01:09:53 ---
This is correct on darwin, and we don't support AIX.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 01:17:29 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 01:17:29 -0500
Subject: [LLVMbugs] [Bug 2617] [ia64] code generation error
In-Reply-To: <bug-2617-206@http.llvm.org/bugs/>
Message-ID: <200810160617.m9G6HTHe010715@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2617


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |LATER




--- Comment #2 from Chris Lattner <clattner at apple.com>  2008-10-16 01:17:29 ---
I don't think this bug is useful.  The IA64 target has serious problems and
this is just the tip of the iceburg.  I changed the llc --help output to list
it as experimental.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 01:18:03 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 01:18:03 -0500
Subject: [LLVMbugs] [Bug 2616] [CellSPU] Code generation failure
In-Reply-To: <bug-2616-206@http.llvm.org/bugs/>
Message-ID: <200810160618.m9G6I3IY010757@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2616


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #2 from Chris Lattner <clattner at apple.com>  2008-10-16 01:18:03 ---
I don't think this bug is useful.  The Cell target has serious missing features
and this is just the tip of the iceburg.  I changed the llc --help output to
list it as experimental.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 01:21:10 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 01:21:10 -0500
Subject: [LLVMbugs] [Bug 2636] Assertion failed: (false && " Serializization
 for type not supported."), function EmitImpl, file TypeSerialization.cpp,
 line 56.
In-Reply-To: <bug-2636-206@http.llvm.org/bugs/>
Message-ID: <200810160621.m9G6LAaS010965@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2636


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #2 from Chris Lattner <clattner at apple.com>  2008-10-16 01:21:10 ---
Serialization isn't done yet, we'll work on it in the future with or without
this bug.  Thanks for filing it though!


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 01:22:29 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 01:22:29 -0500
Subject: [LLVMbugs] [Bug 2642] Constant floating-point vector multiplies
	don't fold
In-Reply-To: <bug-2642-206@http.llvm.org/bugs/>
Message-ID: <200810160622.m9G6MTrN011042@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2642


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #12 from Chris Lattner <clattner at apple.com>  2008-10-16 01:22:28 ---
The optimizer pieces of this are committed, please file another bug for
remaining issues that need to be done in the codegen.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 01:35:16 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 01:35:16 -0500
Subject: [LLVMbugs] [Bug 2668] Assertion failed: (isa<FunctionDecl>(d) && "
 DeclRefExpr can only own FunctionDecls for implicitly def. funcs."),
 function EmitImpl, file StmtSerialization.cpp, line 504.
In-Reply-To: <bug-2668-206@http.llvm.org/bugs/>
Message-ID: <200810160635.m9G6ZG8Y011673@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2668


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |LATER




--- Comment #3 from Chris Lattner <clattner at apple.com>  2008-10-16 01:35:15 ---
No bugs on serialization until it works for simple testcases :)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 01:36:07 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 01:36:07 -0500
Subject: [LLVMbugs] [Bug 2669] Branch Folding pass leaves double
 unconditional branches to a BB .
In-Reply-To: <bug-2669-206@http.llvm.org/bugs/>
Message-ID: <200810160636.m9G6a76K011723@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2669


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |LATER




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-10-16 01:36:06 ---
Heikki, please provide a testcase and reopen.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 01:37:09 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 01:37:09 -0500
Subject: [LLVMbugs] [Bug 2673] ExpandLibCall function does not try to
 resolve libcalls to currently lowered module .
In-Reply-To: <bug-2673-206@http.llvm.org/bugs/>
Message-ID: <200810160637.m9G6b957011788@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2673


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WONTFIX




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-10-16 01:37:08 ---
We don't want the llvm backend to implicitly serve as a linker, sorry. 


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 01:41:46 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 01:41:46 -0500
Subject: [LLVMbugs] [Bug 2683] x86 Cannot yet select v4i16 = fp_to_sint
In-Reply-To: <bug-2683-206@http.llvm.org/bugs/>
Message-ID: <200810160641.m9G6fkm7011977@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2683


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |LATER




--- Comment #2 from Chris Lattner <clattner at apple.com>  2008-10-16 01:41:45 ---
Moving to lib/Target/X86/README-UNIMPLEMENTED.txt


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 01:42:38 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 01:42:38 -0500
Subject: [LLVMbugs] [Bug 2684] x86 Cannot yet select v2i32 = fp_to_sint
In-Reply-To: <bug-2684-206@http.llvm.org/bugs/>
Message-ID: <200810160642.m9G6gcRa012036@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2684


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |LATER




--- Comment #2 from Chris Lattner <clattner at apple.com>  2008-10-16 01:42:38 ---
Moving to lib/Target/X86/README-UNIMPLEMENTED.txt


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 01:43:11 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 01:43:11 -0500
Subject: [LLVMbugs] [Bug 2685] x86 Cannot yet select v4f32 = sint_to_fp
	(from v4i16)
In-Reply-To: <bug-2685-206@http.llvm.org/bugs/>
Message-ID: <200810160643.m9G6hB3Q012078@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2685


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |LATER




--- Comment #3 from Chris Lattner <clattner at apple.com>  2008-10-16 01:43:10 ---
Moving to lib/Target/X86/README-UNIMPLEMENTED.txt


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 01:44:48 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 01:44:48 -0500
Subject: [LLVMbugs] [Bug 2686] x86 Cannot yet select v2f32 = sint_to_fp
In-Reply-To: <bug-2686-206@http.llvm.org/bugs/>
Message-ID: <200810160644.m9G6imMZ012172@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2686


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |LATER




--- Comment #2 from Chris Lattner <clattner at apple.com>  2008-10-16 01:44:47 ---
Moving to lib/Target/X86/README-UNIMPLEMENTED.txt


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 01:45:32 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 01:45:32 -0500
Subject: [LLVMbugs] [Bug 2688] x86 legalize error on v4f32 = concat_vectors
In-Reply-To: <bug-2688-206@http.llvm.org/bugs/>
Message-ID: <200810160645.m9G6jWVG012254@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2688


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |LATER




--- Comment #3 from Chris Lattner <clattner at apple.com>  2008-10-16 01:45:31 ---
Moving to lib/Target/X86/README-UNIMPLEMENTED.txt


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 01:47:28 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 01:47:28 -0500
Subject: [LLVMbugs] [Bug 2693] ice on valid; Segmentation fault
In-Reply-To: <bug-2693-206@http.llvm.org/bugs/>
Message-ID: <200810160647.m9G6lSqA012388@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2693


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #2 from Chris Lattner <clattner at apple.com>  2008-10-16 01:47:27 ---
works for me


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 01:50:29 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 01:50:29 -0500
Subject: [LLVMbugs] [Bug 2712] Hello,
	World in `Writing an LLVM Pass' does not work
In-Reply-To: <bug-2712-206@http.llvm.org/bugs/>
Message-ID: <200810160650.m9G6oTZG012640@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2712


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-10-16 01:50:29 ---
This works for me:
opt --load ~/llvm/Debug/lib/LLVMHello.dylib --help


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 01:51:07 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 01:51:07 -0500
Subject: [LLVMbugs] [Bug 2898] New: Invalid removal of "dead" stores.
Message-ID: <bug-2898-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2898

           Summary: Invalid removal of "dead" stores.
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: lhames at gmail.com
                CC: llvmbugs at cs.uiuc.edu


The ldecod benchmark (MultiSource/Applications/JM/ldecod) is miscompiled for
AMD64. I believe this is due to a bug in LocalSpiller where certain stores are
assumed to be "dead", and thus are removed, when they are in fact still
necessary.

The problem is exposed when the PBQP allocator is used on AMD64. It can be
demonstrated by adding -regalloc=pbqp to LLCBETA_OPTIONS, and running, from
projects/test-suite/MultiSource/Applications/JM/ldecod/Output,

make ENABLE_LLC=1 ENABLE_LLCBETA=1

The resulting program segfaults when run on the test input:

bash-3.2$ ./Output/ldecod.llc-beta -i data/test.264 -o Output/test_dec.yuv -r
data/test_rec.yuv 

----------------------------- JM 12.1 (FRExt) -----------------------------
 Decoder config file                    : (null) 
--------------------------------------------------------------------------
 Input H.264 bitstream                  : data/test.264 
 Output decoded YUV                     : Output/test_dec.yuv 
 Output status file                     : log.dec 
 Input reference file                   : data/test_rec.yuv 
--------------------------------------------------------------------------
POC must = frame# or field# for SNRs to be correct
--------------------------------------------------------------------------
  Frame       POC   Pic#   QP   SnrY    SnrU    SnrV   Y:U:V  Time(ms)
--------------------------------------------------------------------------
Segmentation fault


The test case can be reduced with the following:

bugpoint -run-llc ldecod.llvm.bc --args -i ../data/test.264 -r
../data/test_rec.yuv --tool-args -regalloc=pbqp

Faulting basic block is itrans8x8_bb15_bb15_2E_ce.

After register allocation, before spilling bb15.ce contains the following two
instructions (among many others):

        %reg1170<def> = LEA64r %reg1024, 1, %reg1103<kill>, 1384
        MOV32mr %reg1171<kill>, 4, %reg1089, 0, %reg1167<kill>, Mem:ST(4,4)
[0x15c0628 + 0]

reg167, reg1170 and reg1171 are spill intervals. Pertinent allocations are:
reg1170 = R8, reg1171 = R9, reg1089 = R10, reg 1167 = R8D

During spilling by LocalSpiller the following problem occurs:

        %reg1170<def> = LEA64r %R12, 1, %R8<kill>, 1384
Store:  MOV64mr <fi#10>, 1, %reg0, 0, %R8<kill>
Remembering SS#10 in physreg R8
        MOV64mr <fi#10>, 1, %reg0, 0, %R8<kill>
Reusing SS#10 from physreg R8 for vreg1171 instead of reloading into physreg R9
Removed dead store:     MOV64mr <fi#10>, 1, %reg0, 0, %R8<kill>
PhysReg R9 clobbered, invalidating SS#9
PhysReg R8 clobbered, invalidating SS#10
Remembering SS#10 in physreg R9 
        %R9<def> = MOV64rm <fi#10>, 1, %reg0, 0
Reuse undone!
Remembering SS#8 in physreg R8D
        %R8D<def> = MOV32rm <fi#8>, 1, %reg0, 0
        MOV32mr %R9<kill>, 4, %R10, 0, %R8D<kill>, Mem:ST(4,4) [0x15c0628 + 0]

In attempting to Reuse SS#10 in physreg R8 for the second instruction the store
of reg1170/R8 is treated as dead by the block starting it VirtRegMap.cpp:1441.
When it is discovered that R8 is clobbered by a later parameter the load into
R9 is re-instated, but the store is not, leading to junk in R9. In general I do
not think dead stores cannot be eliminated until the validity of the re-uses
are confirmed.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 01:51:36 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 01:51:36 -0500
Subject: [LLVMbugs] [Bug 2714] [sema] obj-c properties are not searched for
 messages to id
In-Reply-To: <bug-2714-206@http.llvm.org/bugs/>
Message-ID: <200810160651.m9G6pa6Q012755@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2714


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-10-16 01:51:36 ---
Works for me, I think steve fixed this.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 01:56:40 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 01:56:40 -0500
Subject: [LLVMbugs] [Bug 2767] -C output inconsistent with -E for trigraphs
In-Reply-To: <bug-2767-206@http.llvm.org/bugs/>
Message-ID: <200810160656.m9G6ueCD013142@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2767


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WONTFIX




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-10-16 01:56:40 ---
This is because clang doesn't track whether comments need cleaning.  I don't
think that this is worth the perf cost to track.  If you know of a real world
problem this causes, let me know and I will reconsider.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 01:59:57 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 01:59:57 -0500
Subject: [LLVMbugs] [Bug 2789] module does not pass 'verifierPass' after
	'loop unroll'
In-Reply-To: <bug-2789-206@http.llvm.org/bugs/>
Message-ID: <200810160659.m9G6xvgq013287@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2789


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #7 from Chris Lattner <clattner at apple.com>  2008-10-16 01:59:56 ---


*** This bug has been marked as a duplicate of bug 2253 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 02:03:56 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 02:03:56 -0500
Subject: [LLVMbugs] [Bug 2804] Cannot select bitcast of <2 x i32> to double
	on x86
In-Reply-To: <bug-2804-206@http.llvm.org/bugs/>
Message-ID: <200810160703.m9G73ukN013561@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2804


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |LATER




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-10-16 02:03:56 ---
Moving to X86/README-UNIMPLEMENTED.txt


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 02:06:00 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 02:06:00 -0500
Subject: [LLVMbugs] [Bug 2807] CellSPU cannot select fdiv
In-Reply-To: <bug-2807-206@http.llvm.org/bugs/>
Message-ID: <200810160706.m9G760CL013693@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2807


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |LATER




--- Comment #4 from Chris Lattner <clattner at apple.com>  2008-10-16 02:05:59 ---
Closing this as LATER for now.  The Cell backend has some serious deficiencies
that prevent it from working on real world code.  Until it works on more simple
cases, I don't think it makes sense to track the individual issues.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 02:06:50 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 02:06:50 -0500
Subject: [LLVMbugs] [Bug 2809] Avoid execution domain bypass
In-Reply-To: <bug-2809-206@http.llvm.org/bugs/>
Message-ID: <200810160706.m9G76omS013745@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2809


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #6 from Chris Lattner <clattner at apple.com>  2008-10-16 02:06:50 ---
If we ever see a reason to do this, we can clearly consider it :)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 02:11:26 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 02:11:26 -0500
Subject: [LLVMbugs] [Bug 2825] cannot codegen __builtin_va_arg
In-Reply-To: <bug-2825-206@http.llvm.org/bugs/>
Message-ID: <200810160711.m9G7BQ6i013958@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2825


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #4 from Chris Lattner <clattner at apple.com>  2008-10-16 02:11:26 ---


*** This bug has been marked as a duplicate of bug 1740 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 02:14:54 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 02:14:54 -0500
Subject: [LLVMbugs] [Bug 2860] llvm-gcc ICE when building libgcc
In-Reply-To: <bug-2860-206@http.llvm.org/bugs/>
Message-ID: <200810160714.m9G7EslR014137@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2860


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #8 from Chris Lattner <clattner at apple.com>  2008-10-16 02:14:54 ---
Ok, plz reopen if you find that this doesn't fix it.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 02:17:01 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 02:17:01 -0500
Subject: [LLVMbugs] [Bug 2894] space missing in .ll for invoke
In-Reply-To: <bug-2894-206@http.llvm.org/bugs/>
Message-ID: <200810160717.m9G7H1tu014250@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2894


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #2 from Chris Lattner <clattner at apple.com>  2008-10-16 02:17:01 ---
No really, you fixed it :)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 02:56:15 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 02:56:15 -0500
Subject: [LLVMbugs] [Bug 2557] Anders pass asserts on
	extractvalue/insertvalue
In-Reply-To: <bug-2557-206@http.llvm.org/bugs/>
Message-ID: <200810160756.m9G7uF6h021851@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2557


Matthijs Kooijman <matthijs at stdin.nl> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|DUPLICATE                   |




-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 03:54:26 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 03:54:26 -0500
Subject: [LLVMbugs] [Bug 2762] LegalizeTypes abort on vector cast
In-Reply-To: <bug-2762-206@http.llvm.org/bugs/>
Message-ID: <200810160854.m9G8sQTB026905@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2762


Duncan Sands <baldrick at free.fr> changed:

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




--- Comment #5 from Duncan Sands <baldrick at free.fr>  2008-10-16 03:54:25 ---
I fixed the LegalizeTypes part some time ago, and it looks like some
kind soul added the missing patterns.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 04:46:38 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 04:46:38 -0500
Subject: [LLVMbugs] [Bug 2825] cannot codegen __builtin_va_arg
In-Reply-To: <bug-2825-206@http.llvm.org/bugs/>
Message-ID: <200810160946.m9G9kcH2029395@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2825


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Depends on|                            |1740
             Status|RESOLVED                    |REOPENED
         Resolution|DUPLICATE                   |




--- Comment #5 from Eli Friedman <sharparrow1 at yahoo.com>  2008-10-16 04:46:37 ---
This is a clang bug; this can be fixed without any modifications to LLVM by
expanding it out like llvm-gcc does.  Of course, it would be nice if the va_arg
LLVM instruction worked, but I'm not expecting any movement there anytime soon.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 10:01:50 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 10:01:50 -0500
Subject: [LLVMbugs] [Bug 2893] Parsing error (type of property does not
 match type of ivar) if property conforms to protocol and ivar doesn't
In-Reply-To: <bug-2893-206@http.llvm.org/bugs/>
Message-ID: <200810161501.m9GF1oll008070@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2893


snaroff at apple.com <snaroff at apple.com> changed:

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




--- Comment #2 from snaroff at apple.com  2008-10-16 10:01:48 ---
Fixed in http://llvm.org/viewvc/llvm-project?view=rev&revision=57640.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 11:24:47 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 11:24:47 -0500
Subject: [LLVMbugs] [Bug 2557] Anders pass asserts on
	extractvalue/insertvalue
In-Reply-To: <bug-2557-206@http.llvm.org/bugs/>
Message-ID: <200810161624.m9GGOl4I011052@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2557


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #3 from Chris Lattner <clattner at apple.com>  2008-10-16 11:24:46 ---
Updated 2451 to include extractvalue, so this really is a dupe of it.  Thanks
for pointing out my mistake Matthijs!

*** This bug has been marked as a duplicate of bug 2451 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 14:46:09 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 14:46:09 -0500
Subject: [LLVMbugs] [Bug 2901] New: C-Backend crashes on compilng with O, O2,
 O3: "  Expected FP register!"
Message-ID: <bug-2901-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2901

           Summary: C-Backend crashes on compilng with O, O2, O3: "Expected
                    FP register!"
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Windows NT
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: khabkr at yahoo.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2092)
 --> (http://llvm.org/bugs/attachment.cgi?id=2092)
Preprocessed C source code archive

SUMMARY
llvm-gcc chokes on six module files in Python's 2.6 source tree, namely:

1. ../Modules/cmathmodule.c 
2. ../Modules/mathmodule.c
3. ../Objects/complexobject.c 
4  ../Objects/floatobject.c
5. ../Objects/longobject.c
6. ../Python/pymath.c


(and do note the math connection), as follows:

Assertion failed: Reg >= X86::FP0 && Reg <= X86::FP6 && "Expected FP
register!", file c:/proj/llvm/src/lib/Target/X86/X86FloatingPoint.cpp, line 174

This application has requested the Runtime to terminate it in an unusual way.
Please contact the application's support team for more information.
../Objects/floatobject.c: At top level:
../Objects/floatobject.c:2521: internal compiler error: Aborted
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://developer.apple.com/bugreporter> for instructions.

The same error is repeated for all six files. Details follow together with an
attached zip file containing the preprocessed C source code.



PLATFORM
OS: Windows Vista 32 SP1 on Intel (Centrino) Core 2 Due


With the following command issued in MSYS:
alias gcc=/c/dev/llvm-gcc4.2-2.2/bin/llvm-gcc

we get:

gcc -v
Using built-in specs.
Target: i386-mingw32
Configured with: ../src/configure --build=i386-mingw32 --host=i386-mingw32
--target=i386-mingw32 --disable-shared --disable-bootstrap --disable-nls
--with-arch=i486 --with-cpu=generic --prefix=c:/proj/llvm/llvm-gcc-4.2/install
--enable-llvm=c:/proj/llvm/build --disable-win32-registry
--enable-languages=c,c++ --program-prefix=llvm- --with-sjlj-exceptions=no
Thread model: win32
gcc version 4.2.1 (Based on Apple Inc. build 5546) (LLVM build)


LLVM FILES
The llvm files (LLVM-GCC 4.2 Front End Binaries for Mingw32/x86 (22M)) were
downloaded from here:
http://llvm.org/releases/2.2/llvm-gcc4.2-2.2-x86-mingw32.tar.bz2



COMMENTS
All six files compile fine if the optimization options (-O, -O2, -O3) are
removed. 



PROBLEM FILES LIST AND COMPILER OUTPUT
[1] --------------------------------------
$ gcc -I"../Include" -I"../Pc" -I"../Python" -I"../Modules/zlib" -Wall -DNDEBUG
-DPy_BUILD_CORE -DPy_ENABLE_SHARED -O3 -fomit-fram
e-pointer -fwrapv -c -o../Modules/cmathmodule.o ../Modules/cmathmodule.c 
-save-temps
Assertion failed: Reg >= X86::FP0 && Reg <= X86::FP6 && "Expected FP
register!", file c:/proj/llvm/src/lib/Target/X86/X86FloatingPoint.cpp, line 174

This application has requested the Runtime to terminate it in an unusual way.
Please contact the application's support team for more information.
../Modules/cmathmodule.c:1209: internal compiler error: Aborted
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://developer.apple.com/bugreporter> for instructions.


[2] --------------------------------------
$ gcc -I"../Include" -I"../Pc" -I"../Python" -I"../Modules/zlib" -Wall -DNDEBUG
-DPy_BUILD_CORE -DPy_ENABLE_SHARED -O3 -fomit-fram
e-pointer -fwrapv -c -o ../Modules/mathmodule.o ../Modules/mathmodule.c
-save-temps
Assertion failed: Reg >= X86::FP0 && Reg <= X86::FP6 && "Expected FP
register!", file c:/proj/llvm/src/lib/Target/X86/X86FloatingPoint.cpp, line 174

This application has requested the Runtime to terminate it in an unusual way.
Please contact the application's support team for more information.
../Modules/mathmodule.c:1063: internal compiler error: Aborted
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://developer.apple.com/bugreporter> for instructions.


[3] --------------------------------------
$ gcc -I"../Include" -I"../Pc" -I"../Python" -I"../Modules/zlib" -Wall -DNDEBUG
-DPy_BUILD_CORE -DPy_ENABLE_SHARED -O3 -fomit-fram
e-pointer -fwrapv -c -o ../Objects/complexobject.o ../Objects/complexobject.c
-save-temps
../Objects/complexobject.c: In function 'complex_richcompare':
../Objects/complexobject.c:776: warning: dereferencing type-punned pointer will
break strict-aliasing rules
../Objects/complexobject.c:778: warning: dereferencing type-punned pointer will
break strict-aliasing rules
../Objects/complexobject.c: In function 'complex_new':
../Objects/complexobject.c:1051: warning: dereferencing type-punned pointer
will break strict-aliasing rules
Assertion failed: Reg >= X86::FP0 && Reg <= X86::FP6 && "Expected FP
register!", file c:/proj/llvm/src/lib/Target/X86/X86FloatingPoint.cpp, line 174

This application has requested the Runtime to terminate it in an unusual way.
Please contact the application's support team for more information.
../Objects/complexobject.c: At top level:
../Objects/complexobject.c:1271: internal compiler error: Aborted
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://developer.apple.com/bugreporter> for instructions.


[4] --------------------------------------
$ gcc -I"../Include" -I"../Pc" -I"../Python" -I"../Modules/zlib" -Wall -DNDEBUG
-DPy_BUILD_CORE -DPy_ENABLE_SHARED -O3 -fomit-fram
e-pointer -fwrapv -c -o ../Objects/floatobject.o ../Objects/floatobject.c
-save-temps
../Objects/floatobject.c: In function 'float_is_integer':
../Objects/floatobject.c:1038: warning: dereferencing type-punned pointer will
break strict-aliasing rules
../Objects/floatobject.c:1038: warning: dereferencing type-punned pointer will
break strict-aliasing rules
../Objects/floatobject.c:1041: warning: dereferencing type-punned pointer will
break strict-aliasing rules
../Objects/floatobject.c:1041: warning: dereferencing type-punned pointer will
break strict-aliasing rules
../Objects/floatobject.c: In function 'float_new':
../Objects/floatobject.c:1626: warning: dereferencing type-punned pointer will
break strict-aliasing rules
Assertion failed: Reg >= X86::FP0 && Reg <= X86::FP6 && "Expected FP
register!", file c:/proj/llvm/src/lib/Target/X86/X86FloatingPoint.cpp, line 174

This application has requested the Runtime to terminate it in an unusual way.
Please contact the application's support team for more information.
../Objects/floatobject.c: At top level:
../Objects/floatobject.c:2521: internal compiler error: Aborted
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://developer.apple.com/bugreporter> for instructions.


[5] --------------------------------------
$ gcc -I"../Include" -I"../Pc" -I"../Python" -I"../Modules/zlib" -Wall -DNDEBUG
-DPy_BUILD_CORE -DPy_ENABLE_SHARED -O3 -fomit-fram
e-pointer -fwrapv -c -o ../Objects/longobject.o ../Objects/longobject.c
-save-temps
../Objects/longobject.c: In function '_PyLong_Format':
../Objects/longobject.c:1360: warning: dereferencing type-punned pointer will
break strict-aliasing rules
Assertion failed: Reg >= X86::FP0 && Reg <= X86::FP6 && "Expected FP
register!", file c:/proj/llvm/src/lib/Target/X86/X86FloatingPoint.cpp, line 174

This application has requested the Runtime to terminate it in an unusual way.
Please contact the application's support team for more information.
../Objects/longobject.c: At top level:
../Objects/longobject.c:3591: internal compiler error: Aborted
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://developer.apple.com/bugreporter> for instructions.


[6] --------------------------------------
$ gcc -I"../Include" -I"../Pc" -I"../Python" -I"../Modules/zlib" -Wall -DNDEBUG
-DPy_BUILD_CORE -DPy_ENABLE_SHARED -O3 -fomit-fram
e-pointer -fwrapv -c -o ../Python/pymath.o ../Python/pymath.c -save-temps
Assertion failed: Reg >= X86::FP0 && Reg <= X86::FP6 && "Expected FP
register!", file c:/proj/llvm/src/lib/Target/X86/X86FloatingPoint.cpp, line 174

This application has requested the Runtime to terminate it in an unusual way.
Please contact the application's support team for more information.
../Python/pymath.c:237: internal compiler error: Aborted
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://developer.apple.com/bugreporter> for instructions.


Regards
Khalid


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 15:12:50 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 15:12:50 -0500
Subject: [LLVMbugs] [Bug 2902] New: -std-compile-opt produces a CFG w/
	infinite loop
Message-ID: <bug-2902-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2902

           Summary: -std-compile-opt produces a CFG w/ infinite loop
           Product: tools
           Version: 2.3
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P2
         Component: opt
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: micah.villmow at amd.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2093)
 --> (http://llvm.org/bugs/attachment.cgi?id=2093)
Test case for above bug

The attached test case produces an infinite loop in the CFG when run w/
std-compile-opt

The method that this is reproducible is as follows

clang -emit-llvm test_llvm_bug_report.c
llvm-as < test_llvm_bug_report.ll | opt -std-compile-opts -print-cfg-only >
test.bc

To figure out that it was this specific test I ran through the expanded list of
options with the following set not producing the infinite loop:
-preverify  -domtree  -verify  -lowersetjmp  -raiseallocs  -simplifycfg 
-domtree  -domfrontier  -mem2reg  -globalopt  -globaldce  -ipconstprop 
-deadargelim  -instcombine  -simplifycfg  -basiccg  -prune-eh  -inline 
-argpromotion  -simplify-libcalls  -instcombine  -jump-threading  -simplifycfg 
-domtree  -domfrontier  -scalarrepl  -instcombine  -break-crit-edges  -condprop
 -tailcallelim  -simplifycfg  -reassociate  -domtree  -loops  -loopsimplify 
-domfrontier  -scalar-evolution  -lcssa  -loop-rotate  -licm  -lcssa
-simplifycfg

but adding -loopunswitch between the final -lcssa and -simplifycfg causes the
problem.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 16:28:56 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 16:28:56 -0500
Subject: [LLVMbugs] [Bug 2901] C-Backend crashes on compilng with O, O2,
 O3: "  Expected FP register!"
In-Reply-To: <bug-2901-206@http.llvm.org/bugs/>
Message-ID: <200810162128.m9GLSu4h023767@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2901


Anton Korobeynikov <asl at math.spbu.ru> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |asl at math.spbu.ru
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #1 from Anton Korobeynikov <asl at math.spbu.ru>  2008-10-16 16:28:55 ---
This is known problem. Mingw includes contains inline FP math and do not
provide and alternative w/o inline assembler. For example, I had to patch
mingw's complex.h in order to build libstdc++.

*** This bug has been marked as a duplicate of bug 879 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 17:37:33 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 17:37:33 -0500
Subject: [LLVMbugs] [Bug 2903] New: Optimization opportunity - group
 similarly initialized stack slots and bulk initialize
Message-ID: <bug-2903-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2903

           Summary: Optimization opportunity - group similarly initialized
                    stack slots and bulk initialize
           Product: libraries
           Version: 2.3
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: maxburke at gmail.com
                CC: llvmbugs at cs.uiuc.edu


I've run into a case (x86) where I have a bunch of zero-initialized stack slots
(rooted GC pointers) and the initialization code for them, of the form "mov
mem, 0" is 47kb. 

Moving the immediate value to a register and using this register as the source
would save ~20% (8bytes/opcode instead of 10), ie:

xor eax, eax
mov mem, eax

But I was wondering if it'd be possible to group all similarly initialized
stack slots together so that "rep stos" could be used to initialize them all in
one go.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 18:21:11 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 18:21:11 -0500
Subject: [LLVMbugs] [Bug 2904] New: Error initializing va_list with another
	variable.
Message-ID: <bug-2904-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2904

           Summary: Error initializing va_list with another variable.
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: parser
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: alexei.svitkine at gmail.com
                CC: llvmbugs at cs.uiuc.edu


In a function:

void do_stuff(va_list arglist)
{
        va_list temp_args= arglist;
        ....
}

clang gives an error:

error: initialization with "{...}" expected for array
        va_list temp_args= arglist;

GCC and VC7 compile the code fine.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 19:36:27 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 19:36:27 -0500
Subject: [LLVMbugs] [Bug 2905] New: Implement a pretty printer for LLVM
	assembly.
Message-ID: <bug-2905-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2905

           Summary: Implement a pretty printer for LLVM assembly.
           Product: libraries
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: enhancement
          Priority: P2
         Component: LLVM assembly language parser
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: daniel at zuster.org
                CC: llvmbugs at cs.uiuc.edu


LLVM IR files can be pretty unreadable when they have large constant data
structures. It would be really great if these would get pretty printed in some
simple but reasonable fashion.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 20:10:28 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 20:10:28 -0500
Subject: [LLVMbugs] [Bug 2784] Assert violation when generating debug info
 for pointer to opaque type
In-Reply-To: <bug-2784-206@http.llvm.org/bugs/>
Message-ID: <200810170110.m9H1ASMQ031720@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2784


Daniel Dunbar <daniel at zuster.org> changed:

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




--- Comment #1 from Daniel Dunbar <daniel at zuster.org>  2008-10-16 20:10:27 ---
Fixed here: http://llvm.org/viewvc/llvm-project?view=rev&revision=57661

The emitted debug information for this case needs review but I will do that as
part of a forthcoming review of the debug info generation.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 16 21:41:03 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Oct 2008 21:41:03 -0500
Subject: [LLVMbugs] [Bug 2611] smax scev is redundant with umax
In-Reply-To: <bug-2611-206@http.llvm.org/bugs/>
Message-ID: <200810170241.m9H2f3iX002650@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2611


Nick Lewycky <nicholas at mxc.ca> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WONTFIX




--- Comment #5 from Nick Lewycky <nicholas at mxc.ca>  2008-10-16 21:41:02 ---
I don't think so. Instead, we should peep-hole optimize combinations of smax
and umax at the time they're constructed.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 17 01:12:31 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 17 Oct 2008 01:12:31 -0500
Subject: [LLVMbugs] [Bug 2907] New: CBE can't handle ppc long double
	constants
Message-ID: <bug-2907-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2907

           Summary: CBE can't handle ppc long double constants
           Product: libraries
           Version: 1.0
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: C
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: llvmbugs at cs.uiuc.edu


This is a 2.4 regression, llc -march=c dies on the attached testcase.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 17 07:43:48 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 17 Oct 2008 07:43:48 -0500
Subject: [LLVMbugs] [Bug 2908] New: C binding doesn't bind VICmp, VFCmp,
 ExtractValue, InsertValue
Message-ID: <bug-2908-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2908

           Summary: C binding doesn't bind VICmp, VFCmp, ExtractValue,
                    InsertValue
           Product: new-bugs
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: fvbommel at wxs.nl
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2101)
 --> (http://llvm.org/bugs/attachment.cgi?id=2101)
Patch to add said instructions to the binding.

The C binding doesn't contain bindings for the VICmp, VFCmp, ExtractValue and
InsertValue instructions. Patch attached.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 17 12:36:01 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 17 Oct 2008 12:36:01 -0500
Subject: [LLVMbugs] [Bug 2909] New: testsuite should compile Fortran using
 llvm-gfortran if available
Message-ID: <bug-2909-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2909

           Summary: testsuite should compile Fortran using llvm-gfortran if
                    available
           Product: new-bugs
           Version: unspecified
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: enhancement
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: baldrick at free.fr
                CC: tonic at nondot.org, llvmbugs at cs.uiuc.edu


In order to give the llvm-gcc Fortran front-end more testing,
how about having the testsuite use llvm-gfortran, if it's
available, when compiling Fortran (rather than using f2c +
llvm-gcc C frontend as I think is done now).


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 17 13:03:47 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 17 Oct 2008 13:03:47 -0500
Subject: [LLVMbugs] [Bug 2356] Crash in llc with asm (regression)
In-Reply-To: <bug-2356-206@http.llvm.org/bugs/>
Message-ID: <200810171803.m9HI3lcK015527@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2356


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #3 from Chris Lattner <clattner at apple.com>  2008-10-17 13:03:46 ---
Fixed.  This was much more of a pain than expected, and involved many patches,
however, here's the most relevant:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081013/068616.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 17 13:20:52 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 17 Oct 2008 13:20:52 -0500
Subject: [LLVMbugs] [Bug 2661] ICE in cp/decl.c
In-Reply-To: <bug-2661-206@http.llvm.org/bugs/>
Message-ID: <200810171820.m9HIKqfR016436@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2661


Dale Johannesen <dalej at apple.com> changed:

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




--- Comment #8 from Dale Johannesen <dalej at apple.com>  2008-10-17 13:20:51 ---
That worked better.  Fixed by Mike Stump.  Fix applied here:
http://llvm.org/viewvc/llvm-project?view=rev&revision=57711


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 17 13:49:51 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 17 Oct 2008 13:49:51 -0500
Subject: [LLVMbugs] [Bug 2885] assertion triggered in DwarfWriter
In-Reply-To: <bug-2885-206@http.llvm.org/bugs/>
Message-ID: <200810171849.m9HInpEI017867@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2885


Bill Wendling <wendling at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |wendling at apple.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #5 from Bill Wendling <wendling at apple.com>  2008-10-17 13:49:50 ---
(In reply to comment #3)
> Bill, reverting r56628 fixes the assertion, but I get Dwarf Errors out of GDB.
> 
> I understand that patch was probably making a bad situation problem better, but
> is there anything you can do about this assertion? I'd be happy to trade the
> compile failure for bad debug info.
> 
I looked at it and it seems as if there's a problem with comparing mangled
names with their unmangled counter parts. I've been mucking through the debug
code and it appears that linking the subprogram description with the
appropriate machine function isn't feasible. At least not with the current
debugging infrastructure we have.

I can disable the assertion, which will get you past this point.

"Fixed"
http://llvm.org/viewvc/llvm-project?rev=57714&view=rev


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 17 15:53:29 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 17 Oct 2008 15:53:29 -0500
Subject: [LLVMbugs] [Bug 2910] New: clang codegen doesn't handle
	non-constant offsetof
Message-ID: <bug-2910-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2910

           Summary: clang codegen doesn't handle non-constant offsetof
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: llvmbugs at cs.uiuc.edu, ddunbar at stanford.edu


This crashes -emit-llvm:

struct sockaddr_un {
 unsigned char sun_len;
 char sun_path[104];
};

int test(int len) {
  return __builtin_offsetof(struct sockaddr_un, sun_path[len+1]);
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 17 15:57:07 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 17 Oct 2008 15:57:07 -0500
Subject: [LLVMbugs] [Bug 2898] Invalid removal of "dead" stores.
In-Reply-To: <bug-2898-206@http.llvm.org/bugs/>
Message-ID: <200810172057.m9HKv7oR023023@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2898


Evan Cheng <evan.cheng at apple.com> changed:

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




--- Comment #5 from Evan Cheng <evan.cheng at apple.com>  2008-10-17 15:57:06 ---
Thanks! I've applied your patch.
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081013/068635.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 17 17:01:59 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 17 Oct 2008 17:01:59 -0500
Subject: [LLVMbugs] [Bug 2911] New: Assert triggered when complex GC roots
	initialized
Message-ID: <bug-2911-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2911

           Summary: Assert triggered when complex GC roots initialized
           Product: libraries
           Version: 2.3
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: maxburke at gmail.com
                CC: llvmbugs at cs.uiuc.edu


In some GC systems where fat pointers are used to keep track of references
within objects, instances of fat pointers may be rooted on the stack to make
them visible to the garbage collector. When the root initializer code in the
collector base class hits these it will assert because it is not a pointer and
a cast<> operation cannot be performed safely. This happens for example with
this code:

%FatPointerType = type { i8*, i32 }
declare void @llvm.gcroot(i8**, i8*) nounwind 

define i32 @TestFunc() gc "somegc"
{
    %local_fatpointer = alloca %FatPointerType
    %local_fatpointer_ptr = bitcast %FatPointerType* %local_fatpointer to i8**
    call void @llvm.gcroot(i8** %local_fatpointer_ptr, i8* null)

    ret i32 0
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 17 17:26:35 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 17 Oct 2008 17:26:35 -0500
Subject: [LLVMbugs] [Bug 2912] New: MultiSource/Applications/lua cannot be
 run if srcdir != objdir
Message-ID: <bug-2912-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2912

           Summary: MultiSource/Applications/lua cannot be run if srcdir !=
                    objdir
           Product: Test Suite
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Nightly Tester
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: jlerouge at apple.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2104)
 --> (http://llvm.org/bugs/attachment.cgi?id=2104)
Run lua test with srcdir != objdir

I'd like to be able to run this test without touching my source tree. I believe
the attached patch allows this, it force copying the test data and script to
generate additional data into the current objdir, before running the tests.
Source tree is left untouched.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Oct 18 02:06:17 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 18 Oct 2008 02:06:17 -0500
Subject: [LLVMbugs] [Bug 2818] Segmentation fault on PPC64 generated code
In-Reply-To: <bug-2818-206@http.llvm.org/bugs/>
Message-ID: <200810180706.m9I76H9j011200@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2818


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #11 from Chris Lattner <clattner at apple.com>  2008-10-18 02:06:16 ---
ok thanks


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Oct 18 08:51:32 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 18 Oct 2008 08:51:32 -0500
Subject: [LLVMbugs] [Bug 2913] New: [llvm2.4-prerelease]
 MultiSource/Benchmarks/MallocBench/gs/ gs JIT regression
Message-ID: <bug-2913-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2913

           Summary: [llvm2.4-prerelease]
                    MultiSource/Benchmarks/MallocBench/gs/gs JIT regression
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2105)
 --> (http://llvm.org/bugs/attachment.cgi?id=2105)
bugpoint.safe.bc

$ cat Output/gs.out-jit
Reading ghost.ps...
/home/edwin/llvm2.4/release/llvm-2.4/Release/bin/lli[0x95602f]
/lib/libc.so.6[0x7f27703bdf60]
[0x7f276f41c2d8]
exit 139

Same happens if I run without RunSafely.sh (i.e. without ulimit)

*** The following function is being miscompiled:  zvmstatus_bb1
<cbe><gcc>You can reproduce the problem with the command line:
  lli -load ./bugpoint.safe.bc.cbe.c.so bugpoint.test.bc -DNODISPLAY
/home/edwin/llvm2.4/release/llvm-2.4/projects/test-suite/MultiSource/Benchmarks/MallocBench/gs/INPUT/large.ps
The shared object was created with:
  llc -march=c bugpoint.safe.bc -o temporary.c
  gcc -xc temporary.c -O2 -o ./bugpoint.safe.bc.cbe.c.so -fPIC -shared
-fno-strict-aliasing


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Oct 18 08:54:01 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 18 Oct 2008 08:54:01 -0500
Subject: [LLVMbugs] [Bug 2914] New: [llvm2.4-prerelease]
 MultiSource/Benchmarks/MallocBench/ cfrac/cfrac JIT regression
Message-ID: <bug-2914-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2914

           Summary: [llvm2.4-prerelease]
                    MultiSource/Benchmarks/MallocBench/cfrac/cfrac JIT
                    regression
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2107)
 --> (http://llvm.org/bugs/attachment.cgi?id=2107)
bugpoint.test.bc

cat Output/cfrac.out-jit
/home/edwin/llvm2.4/release/llvm-2.4/Release/bin/lli[0x95602f]
/lib/libc.so.6[0x7fa405ee9f60]
[0x7fa404f3f3e8]
exit 139

*** The following function is being miscompiled:  ptoa_bb45
<cbe><gcc>You can reproduce the problem with the command line:
  lli -load ./bugpoint.safe.bc-x5kgHK.cbe.c.so bugpoint.test.bc-Hcmrp6
41757646344123832613190542166099121
The shared object was created with:
  llc -march=c bugpoint.safe.bc-x5kgHK -o temporary.c
  gcc -xc temporary.c -O2 -o ./bugpoint.safe.bc-x5kgHK.cbe.c.so -fPIC -shared
-fno-strict-aliasing


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Oct 18 08:57:16 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 18 Oct 2008 08:57:16 -0500
Subject: [LLVMbugs] [Bug 2915] New: [llvm2.4-prerelease]
 MultiSource/Benchmarks/MallocBench/gs/ gs JIT regression
Message-ID: <bug-2915-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2915

           Summary: [llvm2.4-prerelease]
                    MultiSource/Benchmarks/MallocBench/gs/gs JIT regression
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


There is an ulimit of 250M, but the JITed netbench-crc doesn't fit there.
I 

$ cat Output/netbench-crc.out-jit
Not enough space to create packet data
exit 1

$ cat Output/netbench-crc.out-llc
CRC completed for 12000 packets
exit 154

This patch fixes it for me (x86-64 Linux):
Index: RunSafely.sh
===================================================================
--- RunSafely.sh        (revision 57766)
+++ RunSafely.sh        (working copy)
@@ -94,7 +94,7 @@
     ULIMITCMD="$ULIMITCMD ulimit -f 10485760;"

     # virtual memory: 250 MB should be enough for anybody. ;)
-    ULIMITCMD="$ULIMITCMD ulimit -v 250000;"
+    ULIMITCMD="$ULIMITCMD ulimit -v 350000;"
 esac
 rm -f core core.*


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Oct 18 08:58:43 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 18 Oct 2008 08:58:43 -0500
Subject: [LLVMbugs] [Bug 2916] New: SubReg implementation can not be used to
 insert the hi sub-register.
Message-ID: <bug-2916-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2916

           Summary: SubReg implementation can not be used to insert the hi
                    sub-register.
           Product: tools
           Version: trunk
          Platform: PC
               URL: http://lists.cs.uiuc.edu/pipermail/llvmdev/2008-
                    October/017786.html
        OS/Version: Linux
            Status: NEW
          Keywords: portability
          Severity: normal
          Priority: P2
         Component: llc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: sanjiv.gupta at microchip.com
                CC: llvmbugs at cs.uiuc.edu


We have 16-bit register class and want to set both the lo and high parts using
INSERT_SUBREG. The hi part can not be inserted using SubIdx 2 unless you
declare the sub register class twice in the SubRegClassList.

def FSR16: RegisterClass <"PIC16", [i16], 8, [FSR0, FSR1]> {
  let SubRegClassList = [FSR8, FSR8];   // HERE.
}

For more details refer to the email thread.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Oct 18 13:21:52 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 18 Oct 2008 13:21:52 -0500
Subject: [LLVMbugs] [Bug 2912] MultiSource/Applications/lua cannot be run if
 srcdir != objdir
In-Reply-To: <bug-2912-206@http.llvm.org/bugs/>
Message-ID: <200810181821.m9IILqZw011931@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2912


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-10-18 13:21:51 ---
Awesome, thanks Julien!
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081013/068674.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Oct 18 13:30:44 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 18 Oct 2008 13:30:44 -0500
Subject: [LLVMbugs] [Bug 2917] New: [llvm2.4-prerelease]
 MultiSource/Benchmarks/tramp3d-v4/ LLC , CBE, JIT fails
Message-ID: <bug-2917-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2917

           Summary: [llvm2.4-prerelease] MultiSource/Benchmarks/tramp3d-v4/
                    LLC,CBE,JIT fails
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2109)
 --> (http://llvm.org/bugs/attachment.cgi?id=2109)
testcase-min.i

On x86_64/Linux all of LLC, CBE, JIT fail. 
Even with llvm-gcc -O0 I get a crash.

exitcode is 139, it crashes.

$ llvm-g++ -O0 -g tramp.i -o tramp

(gdb) r --cartvis 1.0 0.0 --rhomin 1e-8
Starting program:
/home/edwin/llvm2.4/release/llvm-2.4/projects/test-suite/MultiSource/Benchmarks/tramp3d-v4/tramp
--cartvis 1.0 0.0 --rhomin 1e-8

Backtrace:
Program received signal SIGSEGV, Segmentation fault.
0x00007feb4e9bfa05 in free () from /lib/libc.so.6
(gdb) bt
#0  0x00007feb4e9bfa05 in free () from /lib/libc.so.6
#1  0x000000000047f089 in __gnu_cxx::new_allocator<Observer<GridLayoutData<3>
>*>::deallocate (this=0x1d4a0b8, __p=0x1d470c000000000,
    unnamed_arg=18430262291329974272) at tramp.i:6236
#2  0x000000000047f0d7 in std::_Vector_base<Observer<GridLayoutData<3> >*,
std::allocator<Observer<GridLayoutData<3> >*> >::_M_deallocate (
    this=0x1d4a0b8, __p=0x1d470c000000000, __n=18430262291329974272) at
tramp.i:28970
#3  0x000000000047f11e in std::_Vector_base<Observer<GridLayoutData<3> >*,
std::allocator<Observer<GridLayoutData<3> >*> >::~_Vector_base (
    this=0x1d4a0b8) at tramp.i:28959
#4  0x00000000004835ea in std::vector<Observer<GridLayoutData<3> >*,
std::allocator<Observer<GridLayoutData<3> >*> >::~vector (this=0x1d4a0b8)
    at tramp.i:29031
#5  0x00000000004836d2 in Observable<GridLayoutData<3> >::~Observable
(this=0x1d4a0b0) at tramp.i:31727
#6  0x0000000000558b6b in GridLayoutData<3>::~GridLayoutData (this=0x1d49fb0)
at tramp.i:57533
#7  0x00000000005662f4 in RefCountedPtr<GridLayoutData<3> >::operator=
(this=0x7fff5765d5e0, rhs=@0x7fff5765bbd0) at tramp.i:32111
#8  0x00000000005663f1 in GridLayout<3>::operator= (this=0x7fff5765d5d8,
model=@0x7fff5765bbc8) at tramp.i:58369
#9  0x000000000057d707 in ParallelTraits<3, UniformRectilinearTag,
CartesianTag>::createLayout (l=@0x7fff5765d5d8, grid=@0x7fff5765d590,
    unnamed_arg=@0x89a720, igc=@0x7fff5765c000, egc=@0x7fff5765bfe8) at
tramp.i:78445
#10 0x000000000040bee4 in main (argc=6, argv=0x7fff5765e9e8) at tramp.i:83417


Since I cannot use bugpoint here, I used delta.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Oct 18 21:18:54 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 18 Oct 2008 21:18:54 -0500
Subject: [LLVMbugs] [Bug 2919] New: gcc regards struct foo and const struct
 foo as __builtin_types_compatible_p
Message-ID: <bug-2919-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2919

           Summary: gcc regards struct foo and const struct foo as
                    __builtin_types_compatible_p
           Product: clang
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: daniel at zuster.org
                CC: llvmbugs at cs.uiuc.edu


gcc accepts the following code which clang rejects:
--
int a0[__builtin_types_compatible_p(struct foo, 
                                    const struct foo) ? 1 : -1];
--

Reported as problem compiling Dovecot.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Oct 18 23:28:50 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 18 Oct 2008 23:28:50 -0500
Subject: [LLVMbugs] [Bug 2920] New: JIT uses rip-relative addressing for
 absolute addresses.
Message-ID: <bug-2920-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2920

           Summary: JIT uses rip-relative addressing for absolute addresses.
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: critical
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: ofv at wanadoo.es
                CC: llvmbugs at cs.uiuc.edu, ofv at wanadoo.es


This is for x86-64. The problem is not present on x86.

This llvm code uses intttoptr for reading and writing some memory addresses:

define i32 @main() {
        %1 = alloca i1          ; <i1*> [#uses=2]
        %2 = alloca i32         ; <i32*> [#uses=2]
        %3 = inttoptr i64 17349872 to i1*               ; <i1*> [#uses=1]
        %4 = load i1* %3                ; <i1> [#uses=1]
        br i1 %4, label %9, label %5

; <label>:5             ; preds = %0
        %6 = inttoptr i64 17349872 to i1*               ; <i1*> [#uses=1]
        store i1 true, i1* %1
        %7 = load i1* %1                ; <i1> [#uses=1]
        store i1 %7, i1* %6
        %8 = inttoptr i64 17349876 to i32*              ; <i32*> [#uses=1]
        store i32 10, i32* %8
        br label %9

; <label>:9             ; preds = %5, %0
        %10 = inttoptr i64 17349876 to i32*             ; <i32*> [#uses=1]
        %11 = load i32* %10             ; <i32> [#uses=1]
        store i32 %11, i32* %2
        %12 = load i32* %2              ; <i32> [#uses=1]
        ret i32 %12
}

the JIT generates this machine code (taken from a gdb session):

0x00007f15bb92e010:     sub    $0x8,%rsp                     
0x00007f15bb92e014:     mov    0x108bcf0(%rip),%al        # 0x7f15bc9b9d0a
0x00007f15bb92e01a:     test   %al,%al
0x00007f15bb92e01c:     jne    0x7f15bb92e038
0x00007f15bb92e022:     movb   $0x1,0x7(%rsp)
0x00007f15bb92e027:     movb   $0x1,0x108bcf0(%rip)        # 0x7f15bc9b9d1e
0x00007f15bb92e02e:     movl   $0xa,0x108bcf4(%rip)        # 0x7f15bc9b9d2c
<getutline_r+12>
0x00007f15bb92e038:     mov    0x108bcf4(%rip),%eax        # 0x7f15bc9b9d32
<getutline_r+18>
0x00007f15bb92e03e:     mov    %eax,(%rsp)
0x00007f15bb92e041:     add    $0x8,%rsp
0x00007f15bb92e045:     retq

As you can see, rip-relative addressing is used combined with the absolute
addresses.

No optimizations applied. I can't reproduce the problem with llc. The code it
generates uses absolute addresses.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Oct 19 12:34:27 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 19 Oct 2008 12:34:27 -0500
Subject: [LLVMbugs] [Bug 2921] New: alloca >4k stack probes on
	windows/visual studio
Message-ID: <bug-2921-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2921

           Summary: alloca >4k stack probes on windows/visual studio
           Product: libraries
           Version: 2.3
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: ubub at gmx.net
                CC: llvmbugs at cs.uiuc.edu


llvm already contains code to insert the required >4k stack probes aka
alloca_probe/chkstk on cygwin/mingw builds.
but if llvm is build with visual studio the insertion is omitted.

I've changed the test in lib/Target/X86/X86RegisterInfo.cpp:

if (NumBytes >= 4096 && Subtarget->isTargetCygMing()) {

to:

if (NumBytes >= 4096 && (Subtarget->isTargetCygMing() ||
Subtarget->isTargetWindows())) {

to avoid crashes for stack allocation bigger 4k.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Oct 20 13:15:58 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 20 Oct 2008 13:15:58 -0500
Subject: [LLVMbugs] [Bug 2922] New: Bad codegen in LLVM JIT on ppc.
Message-ID: <bug-2922-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2922

           Summary: Bad codegen in LLVM JIT on ppc.
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: jlerouge at apple.com
                CC: llvmbugs at cs.uiuc.edu


The attached IR generates bad output when going through lli. llc output is
fine.

$ lli ppc-globals.bc
Result: 43

$ llc -march=ppc32 -f -o ppc-globals.S ppc-globals.bc 
$ gcc -o ppc-globals ppc-globals.S 
$ ./ppc-globals 
Result: 50

Issue seems to have appeared between rev 57669 and 57773, possibly related to
57748 (fold constant offsets into GlobalAddress nodes) ?

Thanks,
Julien


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct 21 09:23:40 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 21 Oct 2008 09:23:40 -0500
Subject: [LLVMbugs] [Bug 2924] New: select instruction with struct result
 type asserts in SDNode::getValueType
Message-ID: <bug-2924-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2924

           Summary: select instruction with struct result type asserts in
                    SDNode::getValueType
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: tomas.l.olsen at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2114)
 --> (http://llvm.org/bugs/attachment.cgi?id=2114)
select on struct failure

I've attached a bugpoint reduced test case that trigger an assertion failure on
a select instruction with { i32, i8* } type.

Pasting .ll here as well for convenience:

; ModuleID = 'bugpoint-reduced-simplified.bc'
target datalayout =
"e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:32:32"
target triple = "i686-pc-linux-gnu"

define x86_stdcallcc { i32, i8* } @_D3std6string7toupperFAaZAa({ i32, i8* } %s)
{
entry_std.string.toupper:
        %tmp58 = load i32* null         ; <i32> [#uses=1]
        %tmp59 = icmp eq i32 %tmp58, 0          ; <i1> [#uses=1]
        %r.val = load { i32, i8* }* null, align 8               ; <{ i32, i8*
}> [#uses=1]
        %condtmp.0 = select i1 %tmp59, { i32, i8* } undef, { i32, i8* } %r.val 
        ; <{ i32, i8* }> [#uses=1]
        ret { i32, i8* } %condtmp.0
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct 21 11:02:22 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 21 Oct 2008 11:02:22 -0500
Subject: [LLVMbugs] [Bug 2922] Bad codegen in LLVM JIT on ppc.
In-Reply-To: <bug-2922-206@http.llvm.org/bugs/>
Message-ID: <200810211602.m9LG2MJU018639@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2922


Julien Lerouge <jlerouge at apple.com> changed:

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




--- Comment #2 from Julien Lerouge <jlerouge at apple.com>  2008-10-21 11:02:21 ---
Seems to have been fixed in rev 57886, cannot reproduce anymore.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct 21 12:30:23 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 21 Oct 2008 12:30:23 -0500
Subject: [LLVMbugs] [Bug 2925] New: mem2reg Painfully Slow on Large
	BasicBlocks
Message-ID: <bug-2925-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2925

           Summary: mem2reg Painfully Slow on Large BasicBlocks
           Product: libraries
           Version: trunk
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Scalar Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: greened at obbligato.org
                CC: llvmbugs at cs.uiuc.edu


mem2reg has a serious slowdown in RewriteSingleStoreAlloca when working on
large blocks.  The search loops to find matching load and store instructions
dominate compile time.

Flat profile:

  %   cumulative   self              self     total           
 time   seconds   seconds    calls   s/call   s/call  name    
 93.23      1.24     1.24        1     1.24     1.25  (anonymous
namespace)::PromoteMem2Reg::run()
  0.75      1.25     0.01  9425833     0.00     0.00 
llvm::cast_retty<llvm::LoadInst, llvm::ilist_iterator<llvm::Instruction>
>::ret_type llvm::cast<llvm::LoadInst, llvm::ilist_iterator<llvm::Instruction>
>(llvm::ilist_iterator<llvm::Instruction> const&)

The call graph profile does not dive beneath the run()  method, presmably
because calls have been inlined.  Profiling of third-party compilers indictes
that RewriteSingleStoreAlloca is the culprit.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct 21 15:01:49 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 21 Oct 2008 15:01:49 -0500
Subject: [LLVMbugs] [Bug 2924] select instruction with struct result type
 asserts in SDNode:: getValueType
In-Reply-To: <bug-2924-206@http.llvm.org/bugs/>
Message-ID: <200810212001.m9LK1nae028225@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2924


Dan Gohman <gohman at apple.com> changed:

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




--- Comment #3 from Dan Gohman <gohman at apple.com>  2008-10-21 15:01:48 ---
This is fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081020/068782.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct 21 15:11:54 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 21 Oct 2008 15:11:54 -0500
Subject: [LLVMbugs] [Bug 2926] New: Assertions fail when linking symbol
	aliases
Message-ID: <bug-2926-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2926

           Summary: Assertions fail when linking symbol aliases
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: torvald at se.inf.tu-dresden.de
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2115)
 --> (http://llvm.org/bugs/attachment.cgi?id=2115)
Source for one.bc

I saw two cases of assertion failures when linking modules with slightly
different symbol aliases (same name, but different aliasee types). As several
other cases are reported as errors in the code, I think these two cases should
also be reported as errors.

The system is an AMD64 Debian lenny, LLVM is built from the 2.4 release branch,
llvm-gcc is built from the 2.3 sources (I wasn't able to build it from 2.4).

Here is what fails (one.ll, two.ll, three.ll are attached).

llvm-ld -link-as-library -internalize -o foo.bc one.bc three.bc

llvm-ld:
/home/staff/torvald/STM/stm/code/cpp/llvm-2.4/include/llvm/Support/Casting.h:199:
typename llvm::cast_retty<To, From>::ret_type llvm::cast(const Y&) [with X =
llvm::GlobalValue, Y = llvm::Value*]: Assertion `isa<X>(Val) && "cast<Ty>()
argument of incompatible type!"' failed.
llvm-ld[0x8ca3d7]
llvm-ld[0x8ca4f8]
/lib/libc.so.6[0x7f2cbbdf8f60]
/lib/libc.so.6(gsignal+0x35)[0x7f2cbbdf8ed5]
/lib/libc.so.6(abort+0x183)[0x7f2cbbdfa3f3]
/lib/libc.so.6(__assert_fail+0xe9)[0x7f2cbbdf1dc9]
llvm-ld(_ZN4llvm12cast_or_nullINS_11GlobalValueENS_5ValueEEENS_10cast_rettyIT_PT0_E8ret_typeES6_+0x0)[0x68e11b]
llvm-ld[0x688c90]
llvm-ld(_ZN4llvm6Linker11LinkModulesEPNS_6ModuleES2_PSs+0x639)[0x68abdf]
llvm-ld(_ZN4llvm6Linker12LinkInModuleEPNS_6ModuleEPSs+0x28)[0x68593a]
llvm-ld(_ZN4llvm6Linker10LinkInFileERKNS_3sys4PathERb+0x4d4)[0x684fe4]
llvm-ld(_ZN4llvm6Linker11LinkInFilesERKSt6vectorINS_3sys4PathESaIS3_EE+0x36)[0x68514a]
llvm-ld(main+0x359)[0x6628c9]
/lib/libc.so.6(__libc_start_main+0xe6)[0x7f2cbbde51a6]
llvm-ld(_ZNSt8ios_base4InitD1Ev+0x39)[0x65a0b9]
Aborted


llvm-ld -link-as-library -internalize -o foo.bc two.bc three.bc

llvm-ld:
/home/staff/torvald/STM/stm/code/cpp/llvm-2.4/lib/VMCore/Globals.cpp:193:
llvm::GlobalAlias::GlobalAlias(const llvm::Type*,
llvm::GlobalValue::LinkageTypes, const std::string&, llvm::Constant*,
llvm::Module*): Assertion `aliasee->getType() == Ty && "Alias and aliasee types
should match!"' failed.
llvm-ld[0x8ca3d7]
llvm-ld[0x8ca4f8]
/lib/libc.so.6[0x7f537d344f60]
/lib/libc.so.6(gsignal+0x35)[0x7f537d344ed5]
/lib/libc.so.6(abort+0x183)[0x7f537d3463f3]
/lib/libc.so.6(__assert_fail+0xe9)[0x7f537d33ddc9]
llvm-ld(_ZN4llvm11GlobalAliasC1EPKNS_4TypeENS_11GlobalValue12LinkageTypesERKSsPNS_8ConstantEPNS_6ModuleE+0xa2)[0x839fd2]
llvm-ld[0x6895a1]
llvm-ld(_ZN4llvm6Linker11LinkModulesEPNS_6ModuleES2_PSs+0x639)[0x68abdf]
llvm-ld(_ZN4llvm6Linker12LinkInModuleEPNS_6ModuleEPSs+0x28)[0x68593a]
llvm-ld(_ZN4llvm6Linker10LinkInFileERKNS_3sys4PathERb+0x4d4)[0x684fe4]
llvm-ld(_ZN4llvm6Linker11LinkInFilesERKSt6vectorINS_3sys4PathESaIS3_EE+0x36)[0x68514a]
llvm-ld(main+0x359)[0x6628c9]
/lib/libc.so.6(__libc_start_main+0xe6)[0x7f537d3311a6]
llvm-ld(_ZNSt8ios_base4InitD1Ev+0x39)[0x65a0b9]
Aborted


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct 21 17:28:25 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 21 Oct 2008 17:28:25 -0500
Subject: [LLVMbugs] [Bug 2927] New: Calling a function from the JIT
	segfaults.
Message-ID: <bug-2927-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2927

           Summary: Calling a function from the JIT segfaults.
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Windows 2000
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: ofv at wanadoo.es
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2119)
 --> (http://llvm.org/bugs/attachment.cgi?id=2119)
C++ tests case that crashes when calling C function.

On the x86-64 JIT, calling a C function registered via
ExecutionEngine::addGlobalMapping, produces a crash.

Inpecting the JIT'ed generated machine code shows that the call does not refer
the address of the C function. It is like the global mapping for the function
didn't exist.

Attached is a C++ test case that crashes on x86-64 and succeeds on x86.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct 21 21:18:26 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 21 Oct 2008 21:18:26 -0500
Subject: [LLVMbugs] [Bug 2928] New: make install shall ignore *.cmake files.
Message-ID: <bug-2928-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2928

           Summary: make install shall ignore *.cmake files.
           Product: Build scripts
           Version: 2.4
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Makefiles
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: ofv at wanadoo.es
                CC: llvmbugs at cs.uiuc.edu


Those files are headers used by the CMake build. Current "configure && make
install" treats them as regular files and are installed.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct 21 23:14:43 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 21 Oct 2008 23:14:43 -0500
Subject: [LLVMbugs] [Bug 2914] [llvm2.4-prerelease]
 MultiSource/Benchmarks/MallocBench/cfrac/ cfrac JIT regression
In-Reply-To: <bug-2914-206@http.llvm.org/bugs/>
Message-ID: <200810220414.m9M4Eh0d010446@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2914


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #2 from Chris Lattner <clattner at apple.com>  2008-10-21 23:14:43 ---
This is a misreduction (bugpoint.test.bc is 'empty'), we can't debug this.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct 21 23:53:47 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 21 Oct 2008 23:53:47 -0500
Subject: [LLVMbugs] [Bug 2907] CBE can't handle ppc long double constants
In-Reply-To: <bug-2907-206@http.llvm.org/bugs/>
Message-ID: <200810220453.m9M4rlmI011768@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2907


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #2 from Chris Lattner <clattner at apple.com>  2008-10-21 23:53:47 ---
Fixed, patch here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081020/068806.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 22 01:16:18 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Oct 2008 01:16:18 -0500
Subject: [LLVMbugs] [Bug 2929] New: Bugpoint should preserve the nounwind
 attribute when code extracting
Message-ID: <bug-2929-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2929

           Summary: Bugpoint should preserve the nounwind attribute when
                    code extracting
           Product: tools
           Version: 2.2
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: bugpoint
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: llvmbugs at cs.uiuc.edu


When extracting functions/loops/blocks etc, bugpoint does not put the nounwind
attr on the generated functions.  This causes the compiler to slow down and can
hide bugs.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 22 05:30:18 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Oct 2008 05:30:18 -0500
Subject: [LLVMbugs] [Bug 2930] New: clang does not allow VLA in
 unions/structs as gcc does ( violates C99)
Message-ID: <bug-2930-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2930

           Summary: clang does not allow VLA in unions/structs as gcc does
                    (violates C99)
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Keywords: missing-feature
          Severity: normal
          Priority: P2
         Component: parser
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: rdivacky at freebsd.org
                CC: llvmbugs at cs.uiuc.edu


gcc accepts this code:

int
__xuname(int namesize, void *namebuf)
{
        struct xutsname {
                char    sysname[namesize];      /* Name of this OS. */
        } *name;
}

which clang refuses with:

const.c:5:25: error: arrays with static storage duration must have constant
integer length
                char    sysname[namesize];      /* Name of this OS. */
                                        ^~~~~~~

this violates C99 but it is needed to get full gcc compatibility


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 22 09:55:44 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Oct 2008 09:55:44 -0500
Subject: [LLVMbugs] [Bug 2931] New: [llvm2.4-prerelease] r57979 failed to
	build
Message-ID: <bug-2931-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2931

           Summary: [llvm2.4-prerelease] r57979 failed to build
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


CBackend.cpp: In member function
???void<unnamed>::CWriter::printConstant(llvm::Constant*, bool)???:
CBackend.cpp:1146: error: no matching function for call to
???llvm::APFloat::convert(const llvm::fltSemantics&,
llvm::APFloat::roundingMode, bool*)???
/home/edwin/llvm2.4/release/llvm-2.4/include/llvm/ADT/APFloat.h:219: note:
candidates are: llvm::APFloat::opStatus llvm::APFloat::convert(const
llvm::fltSemantics&, llvm::APFloat::roundingMode)
CBackend.cpp: In member function
???void<unnamed>::CWriter::printFloatingPointConstants(const
llvm::Constant*)???:
CBackend.cpp:2063: error: ???const class llvm::APFloat??? has no member named
???bitcastToAPInt???
CBackend.cpp:2069: error: ???const class llvm::APFloat??? has no member named
???bitcastToAPInt???
CBackend.cpp:2076: error: ???const class llvm::APFloat??? has no member named
???bitcastToAPInt???
CBackend.cpp:2084: error: ???const class llvm::APFloat??? has no member named
???bitcastToAPInt???
make[3]: ***
[/home/edwin/llvm2.4/release/llvm-2.4/lib/Target/CBackend/Release/CBackend.o]
Error 1
make[3]: Leaving directory
`/home/edwin/llvm2.4/release/llvm-2.4/lib/Target/CBackend'


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 22 12:23:55 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Oct 2008 12:23:55 -0500
Subject: [LLVMbugs] [Bug 2881] Small leak in PluginLoader.cpp.
In-Reply-To: <bug-2881-206@http.llvm.org/bugs/>
Message-ID: <200810221723.m9MHNto7017165@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2881


Julien Lerouge <jlerouge at apple.com> changed:

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




-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 22 17:14:51 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Oct 2008 17:14:51 -0500
Subject: [LLVMbugs] [Bug 2933] New: Inline attribute on function gives the
 wrong linkage attribute in LLVM IR
Message-ID: <bug-2933-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2933

           Summary: Inline attribute on function gives the wrong linkage
                    attribute in LLVM IR
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Basic
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nadav256 at gmail.com
                CC: llvmbugs at cs.uiuc.edu


When I define an inlined C function like:  "inline int f(...) {...}"   clang
compiles it to the following LLVM IR:

define weak i32 @f(i32 %input) nounwind {  ...

This actually prevents f from being inlined. 

llvm-gcc will create the correct result:

define linkonce i32 @f(i32 %input) { ....

which will be inlined correctly.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 23 00:54:31 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 23 Oct 2008 00:54:31 -0500
Subject: [LLVMbugs] [Bug 2886] 2.4 pre-release test failures: ocaml bindings
In-Reply-To: <bug-2886-206@http.llvm.org/bugs/>
Message-ID: <200810230554.m9N5sVVv008019@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2886


Tanya Lattner <tonic at nondot.org> changed:

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




--- Comment #12 from Tanya Lattner <tonic at nondot.org>  2008-10-23 00:54:30 ---
Merged into 2.4.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 23 00:55:12 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 23 Oct 2008 00:55:12 -0500
Subject: [LLVMbugs] [Bug 2931] [llvm2.4-prerelease] r57979 failed to build
In-Reply-To: <bug-2931-206@http.llvm.org/bugs/>
Message-ID: <200810230555.m9N5tCPc008066@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2931


Tanya Lattner <tonic at nondot.org> changed:

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




--- Comment #5 from Tanya Lattner <tonic at nondot.org>  2008-10-23 00:55:12 ---
Applied to 2.4


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 23 07:51:52 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 23 Oct 2008 07:51:52 -0500
Subject: [LLVMbugs] [Bug 2934] New: llvm2.4-prerelease: 1 unexpected success
Message-ID: <bug-2934-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2934

           Summary: llvm2.4-prerelease: 1 unexpected success
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


XPASS:
/home/edwin/llvm2.4/release/llvm-2.4/test/Linker/2008-07-06-AliasWeakDest.ll
for PR2463

===  Summary ===

# of expected passes            3155
# of unexpected successes       1
# of expected failures          9

Well, its not a bug, but if that test works, shouldn't the XPASS be removed?

This is on x86_64 Linux.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 23 07:53:08 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 23 Oct 2008 07:53:08 -0500
Subject: [LLVMbugs] [Bug 2847] make check failures in r56923
In-Reply-To: <bug-2847-206@http.llvm.org/bugs/>
Message-ID: <200810231253.m9NCr8Tr029967@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2847


T??r??k Edwin <edwintorok at gmail.com> changed:

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




--- Comment #3 from T??r??k Edwin <edwintorok at gmail.com>  2008-10-23 07:53:08 ---
These failures have been fixed.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 23 08:16:59 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 23 Oct 2008 08:16:59 -0500
Subject: [LLVMbugs] [Bug 2935] New: mention that GetResultInst has been
 removed in ReleaseNotes
Message-ID: <bug-2935-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2935

           Summary: mention that GetResultInst has been removed in
                    ReleaseNotes
           Product: Documentation
           Version: 2.4
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: General docs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


I've just come accross this while upgrading some code to LLVM 2.4, I think
there should be some mention of it in the release notes.


------------------------------------------------------------------------
r53941 | djg | 2008-07-23 03:34:11 +0300 (Wed, 23 Jul 2008) | 10 lines

Enable first-class aggregates support.

Remove the GetResultInst instruction. It is still accepted in LLVM assembly
and bitcode, where it is now auto-upgraded to ExtractValueInst. Also, remove
support for return instructions with multiple values. These are auto-upgraded
to use InsertValueInst instructions.

The IRBuilder still accepts multiple-value returns, and auto-upgrades them
to InsertValueInst instructions.

------------------------------------------------------------------------


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 23 08:23:00 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 23 Oct 2008 08:23:00 -0500
Subject: [LLVMbugs] [Bug 2936] New: wrong x86 asm symbol name decoration
Message-ID: <bug-2936-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2936

           Summary: wrong x86 asm symbol name decoration
           Product: libraries
           Version: 2.4
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: ubub at gmx.net
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2124)
 --> (http://llvm.org/bugs/attachment.cgi?id=2124)
testcase

when dllexport'ing functions with llvm on the x86 cygwin platform two bugs
arose:

1. the X86ATTAsmPrinter uses the properly decorated function names in the
function definition, but exports the undecorated names.

2. both, the X86ATTAsmPrinter and X86IntelAsmPrinter use an incorrectly
decorated name when calling a fastcall/stdcall function. this only happens if
the function is called before it is defined.

I'm attaching a single testcase to reproduce both problems.
It is only reproduceable with the 2.4 branch on cygwin or mingw.

Additionally I'm providing a workaround (not a fully tested patch).


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 23 08:32:49 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 23 Oct 2008 08:32:49 -0500
Subject: [LLVMbugs] [Bug 2937] New: more API changes in 2.4 not listed in
	ReleaseNotes
Message-ID: <bug-2937-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2937

           Summary: more API changes in 2.4 not listed in ReleaseNotes
           Product: Documentation
           Version: 2.4
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: General docs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


I'll add them here as soon as I find them.

------------------------------------------------------------------------
r53269 | nicholas | 2008-07-09 06:21:51 +0300 (Wed, 09 Jul 2008) | 3 lines

Remove getValueRange from SCEV. It wasn't doing anything there anyways, and a
more complete version is now available from the LoopVR pass.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 23 08:43:01 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 23 Oct 2008 08:43:01 -0500
Subject: [LLVMbugs] [Bug 2938] New: [llvm2.4-prerelease] cannot print
 ConstantRange to DOUT
Message-ID: <bug-2938-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2938

           Summary: [llvm2.4-prerelease] cannot print ConstantRange to DOUT
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


This code doesn't compile anymore:

ConstantRange Bounds = VR->get(V);
DOUT << "Bounds: " << Bounds;

I think the problem is that DOUT is not a raw_ostream.
I've commented out that code for now.

/home/edwin/llvm-svn/llvm/include/llvm/Support/Streams.h: In member function
???llvm::BaseStream<StreamTy>& llvm::BaseStream<StreamTy>::operator<<(const
Ty&) [with Ty = llvm::ConstantRange, StreamTy = std::basic_ostream<char,
std::char_traits<char> >]???:
Solver.cpp:440:   instantiated from here
/home/edwin/llvm-svn/llvm/include/llvm/Support/Streams.h:57: error: no match
for ???operator<<??? in ???*((llvm::BaseStream<std::basic_ostream<char,
std::char_traits<char> > >*)this)->llvm::BaseStream<std::basic_ostream<char,
std::char_traits<char> > >::Stream << Thing???
/usr/include/c++/4.3/ostream:112: note: candidates are:
std::basic_ostream<_CharT, _Traits>& std::basic_ostream<_CharT,
_Traits>::operator<<(std::basic_ostream<_CharT, _Traits>&
(*)(std::basic_ostream<_CharT, _Traits>&)) [with _CharT = char, _Traits =
std::char_traits<char>]
/usr/include/c++/4.3/ostream:121: note:                
std::basic_ostream<_CharT, _Traits>& std::basic_ostream<_CharT,
_Traits>::operator<<(std::basic_ios<_CharT, _Traits>&
(*)(std::basic_ios<_CharT, _Traits>&)) [with _CharT = char, _Traits =
std::char_traits<char>]
/usr/include/c++/4.3/ostream:131: note:                
std::basic_ostream<_CharT, _Traits>& std::basic_ostream<_CharT,
_Traits>::operator<<(std::ios_base& (*)(std::ios_base&)) [with _CharT = char,
_Traits = std::char_traits<char>]
/usr/include/c++/4.3/ostream:169: note:                
std::basic_ostream<_CharT, _Traits>& std::basic_ostream<_CharT,
_Traits>::operator<<(long int) [with _CharT = char, _Traits =
std::char_traits<char>]
/usr/include/c++/4.3/ostream:173: note:                
std::basic_ostream<_CharT, _Traits>& std::basic_ostream<_CharT,
_Traits>::operator<<(long unsigned int) [with _CharT = char, _Traits =
std::char_traits<char>]
/usr/include/c++/4.3/ostream:177: note:                
std::basic_ostream<_CharT, _Traits>& std::basic_ostream<_CharT,
_Traits>::operator<<(bool) [with _CharT = char, _Traits =
std::char_traits<char>]
/usr/include/c++/4.3/bits/ostream.tcc:97: note:                
std::basic_ostream<_CharT, _Traits>& std::basic_ostream<_CharT,
_Traits>::operator<<(short int) [with _CharT = char, _Traits =
std::char_traits<char>]
/usr/include/c++/4.3/ostream:184: note:                
std::basic_ostream<_CharT, _Traits>& std::basic_ostream<_CharT,
_Traits>::operator<<(short unsigned int) [with _CharT = char, _Traits =
std::char_traits<char>]
/usr/include/c++/4.3/bits/ostream.tcc:111: note:                
std::basic_ostream<_CharT, _Traits>& std::basic_ostream<_CharT,
_Traits>::operator<<(int) [with _CharT = char, _Traits =
std::char_traits<char>]
/usr/include/c++/4.3/ostream:195: note:                
std::basic_ostream<_CharT, _Traits>& std::basic_ostream<_CharT,
_Traits>::operator<<(unsigned int) [with _CharT = char, _Traits =
std::char_traits<char>]
/usr/include/c++/4.3/ostream:204: note:                
std::basic_ostream<_CharT, _Traits>& std::basic_ostream<_CharT,
_Traits>::operator<<(long long int) [with _CharT = char, _Traits =
std::char_traits<char>]
/usr/include/c++/4.3/ostream:208: note:                
std::basic_ostream<_CharT, _Traits>& std::basic_ostream<_CharT,
_Traits>::operator<<(long long unsigned int) [with _CharT = char, _Traits =
std::char_traits<char>]
/usr/include/c++/4.3/ostream:213: note:                
std::basic_ostream<_CharT, _Traits>& std::basic_ostream<_CharT,
_Traits>::operator<<(double) [with _CharT = char, _Traits =
std::char_traits<char>]
/usr/include/c++/4.3/ostream:217: note:                
std::basic_ostream<_CharT, _Traits>& std::basic_ostream<_CharT,
_Traits>::operator<<(float) [with _CharT = char, _Traits =
std::char_traits<char>]
/usr/include/c++/4.3/ostream:225: note:                
std::basic_ostream<_CharT, _Traits>& std::basic_ostream<_CharT,
_Traits>::operator<<(long double) [with _CharT = char, _Traits =
std::char_traits<char>]
/usr/include/c++/4.3/ostream:229: note:                
std::basic_ostream<_CharT, _Traits>& std::basic_ostream<_CharT,
_Traits>::operator<<(const void*) [with _CharT = char, _Traits =
std::char_traits<char>]
/usr/include/c++/4.3/bits/ostream.tcc:125: note:                
std::basic_ostream<_CharT, _Traits>& std::basic_ostream<_CharT,
_Traits>::operator<<(std::basic_streambuf<_CharT, _Traits>*) [with _CharT =
char, _Traits = std::char_traits<char>]
/home/edwin/llvm-svn/llvm/include/llvm/Value.h:247: note:                
std::ostream& llvm::operator<<(std::ostream&, const llvm::Value&)
/home/edwin/llvm-svn/llvm/include/llvm/Value.h:251: note:                
llvm::raw_ostream& llvm::operator<<(llvm::raw_ostream&, const llvm::Value&)
/home/edwin/llvm-svn/llvm/include/llvm/System/Path.h:708: note:                
std::ostream& llvm::operator<<(std::ostream&, const llvm::sys::Path&)
/home/edwin/llvm-svn/llvm/include/llvm/System/Path.h:710: note:                
std::ostream& llvm::operator<<(std::ostream&, const llvm::sys::PathWithStatus&)
/home/edwin/llvm-svn/llvm/include/llvm/Pass.h:205: note:                
std::ostream& llvm::operator<<(std::ostream&, const llvm::Pass&)
/home/edwin/llvm-svn/llvm/include/llvm/Type.h:457: note:                
std::ostream& llvm::operator<<(std::ostream&, const llvm::Type&)
/home/edwin/llvm-svn/llvm/include/llvm/Type.h:458: note:                
llvm::raw_ostream& llvm::operator<<(llvm::raw_ostream&, const llvm::Type&)
/home/edwin/llvm-svn/llvm/include/llvm/ADT/APInt.h:1386: note:                
llvm::raw_ostream& llvm::operator<<(llvm::raw_ostream&, const llvm::APInt&)
/home/edwin/llvm-svn/llvm/include/llvm/Analysis/ScalarEvolution.h:105: note:   
             std::ostream& llvm::operator<<(std::ostream&, const llvm::SCEV&)
/home/edwin/llvm-svn/llvm/include/llvm/Support/ConstantRange.h:188: note:      
          llvm::raw_ostream& llvm::operator<<(llvm::raw_ostream&, const
llvm::ConstantRange&)
Constraint.h:58: note:                 std::ostream&
llvm::operator<<(std::ostream&, const llvm::Relation&)
Constraint.h:134: note:                 std::ostream&
llvm::operator<<(std::ostream&, const llvm::Constraint&)
/home/edwin/llvm-svn/llvm/include/llvm/Analysis/AliasSetTracker.h:245: note:   
             std::ostream& llvm::operator<<(std::ostream&, const
llvm::AliasSet&)
/home/edwin/llvm-svn/llvm/include/llvm/Analysis/AliasSetTracker.h:384: note:   
             std::ostream& llvm::operator<<(std::ostream&, const
llvm::AliasSetTracker&)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 23 11:29:22 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 23 Oct 2008 11:29:22 -0500
Subject: [LLVMbugs] [Bug 2934] llvm2.4-prerelease: 1 unexpected success
In-Reply-To: <bug-2934-206@http.llvm.org/bugs/>
Message-ID: <200810231629.m9NGTMMq006055@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2934


Tanya Lattner <tonic at nondot.org> changed:

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




--- Comment #2 from Tanya Lattner <tonic at nondot.org>  2008-10-23 11:29:21 ---
Yes, he did not ask me to merge in that change. I've merged in 57596.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 23 12:33:02 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 23 Oct 2008 12:33:02 -0500
Subject: [LLVMbugs] [Bug 2939] New: Assertion failed: (0 && "Unknown type!"),
 function getMVT, file ValueTypes.cpp, line 109.
Message-ID: <bug-2939-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2939

           Summary: Assertion failed: (0 && "Unknown type!"), function
                    getMVT, file ValueTypes.cpp, line 109.
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Core LLVM classes
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: rdivacky at freebsd.org
                CC: llvmbugs at cs.uiuc.edu


this simple program crashes clang, but clang -emit-llvm works so I believe the
problem is else where:

witten /tmp# cat semctl.c
#include <stdarg.h>

union semun {
        int             val;            /* value for SETVAL */
        unsigned short  *array;         /* array for GETALL & SETALL */
};


int semctl(int semid, int semnum, int cmd, ...)
{
        va_list ap;
        union semun semun;

        va_start(ap, cmd);
                semun = va_arg(ap, union semun);
        va_end(ap);
}
witten /tmp# ccc semctl.c
Assertion failed: (0 && "Unknown type!"), function getMVT, file ValueTypes.cpp,
line 109.

.ll file attached


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 23 13:26:05 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 23 Oct 2008 13:26:05 -0500
Subject: [LLVMbugs] [Bug 2939] Assertion failed: (0 && "Unknown type!"),
 function getMVT, file ValueTypes.cpp, line 109.
In-Reply-To: <bug-2939-206@http.llvm.org/bugs/>
Message-ID: <200810231826.m9NIQ5Fa010481@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2939


Dan Gohman <gohman at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #3 from Dan Gohman <gohman at apple.com>  2008-10-23 13:26:04 ---
Thanks. This is PR2594.

*** This bug has been marked as a duplicate of bug 2594 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 23 13:31:07 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 23 Oct 2008 13:31:07 -0500
Subject: [LLVMbugs] [Bug 2938] [llvm2.4-prerelease] cannot print
	ConstantRange to DOUT
In-Reply-To: <bug-2938-206@http.llvm.org/bugs/>
Message-ID: <200810231831.m9NIV7iF010764@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2938


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WONTFIX




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-10-23 13:31:07 ---
The plan is to convert everything over to raw_ostream.  Sorry, we won't be
adding new operator<<'s for std::ostream unless needed in mainline.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 23 15:32:42 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 23 Oct 2008 15:32:42 -0500
Subject: [LLVMbugs] [Bug 2940] New: Assertion fails in getLowBitsSet when
	optimizing
Message-ID: <bug-2940-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2940

           Summary: Assertion fails in getLowBitsSet when optimizing
           Product: libraries
           Version: trunk
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Transformation Utilities
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: wesleypeck at gmail.com
                CC: llvmbugs at cs.uiuc.edu


The LLVM bitcode optimizer fails with an assertion error inside of
getLowBitsSet. The problems seems to arise from the "ptrtoint" line and the ret
that follows. The program compiles and executes without problem when
optimization is not performed.

The original .ll file is as follows (this is the exact file contents):
define i8* @id(i8* %x) gc "ocaml"
{
entry:
        ret i8* %x
}

define i32 @tstid() gc "ocaml"
{
entry:
        %var0 = inttoptr i32 1 to i8*
        %var1 = tail call i8* @id(i8* %var0)
        %var2 = ptrtoint i8* %var1 to i32
        ret i32 %var2
}

define i32 @main() gc "ocaml"
{
entry:
        %var0 = tail call i32 @tstid()
        ret i32 %var0
}

The commands used to compile and optimize the file were:
llvm-as tyvars.ll
opt -f -o=tyvars.bc.opt.bc -std-compile-opts tyvars.bc

The command "opt --version" returns:
Low Level Virtual Machine (http://llvm.org/):
  llvm version 2.4svn
  Optimized build with assertions.

The error message printed out follows:
===============================================================================
Assertion failed: (loBitsSet <= numBits && "Too many bits to set!"), function
getLowBitsSet, file
/opt/local/var/macports/build/_opt_local_var_macports_sources_rsync.macports.org_release_ports_lang_llvm-devel/work/llvm-53722/include/llvm/ADT/APInt.h,
line 416.
0   opt                                 0x003ae37c
_ZNSt8_Rb_treeIN4llvm3sys4PathES2_St9_IdentityIS2_ESt4lessIS2_ESaIS2_EE13insert_uniqueERKS2_
+ 7948
1   libSystem.B.dylib                   0x92fc709b _sigtramp + 43
2   ???                                 0xffffffff 0x0 + 4294967295
3   libSystem.B.dylib                   0x9303fec2 raise + 26
4   libSystem.B.dylib                   0x9304f47f abort + 73
5   libSystem.B.dylib                   0x93041063 __assert_rtn + 101
6   opt                                 0x0022d3b4
_ZN4llvm15callDefaultCtorIN183_GLOBAL__N__opt_local_var_macports_build__opt_local_var_macports_sources_rsync.macports.org_release_ports_lang_llvm_devel_work_llvm_53722_lib_Analysis_CFGPrinter.cpp_00000000_FBA5535913CFGOnlyViewerEEEPNS_4PassEv
+ 10948
7   opt                                 0x0022dba5
_ZN4llvm15callDefaultCtorIN183_GLOBAL__N__opt_local_var_macports_build__opt_local_var_macports_sources_rsync.macports.org_release_ports_lang_llvm_devel_work_llvm_53722_lib_Analysis_CFGPrinter.cpp_00000000_FBA5535913CFGOnlyViewerEEEPNS_4PassEv
+ 12981
8   opt                                 0x000e71e2
_ZNSt8_Rb_treeIjSt4pairIKjPN4llvm5ValueEESt10_Select1stIS5_ESt4lessIjESaIS5_EE8_M_eraseEPSt13_Rb_tree_nodeIS5_E
+ 336226
9   opt                                 0x000f33ca
_ZNSt8_Rb_treeIjSt4pairIKjPN4llvm5ValueEESt10_Select1stIS5_ESt4lessIjESaIS5_EE8_M_eraseEPSt13_Rb_tree_nodeIS5_E
+ 385866
10  opt                                 0x000f4d5c
_ZNSt8_Rb_treeIjSt4pairIKjPN4llvm5ValueEESt10_Select1stIS5_ESt4lessIjESaIS5_EE8_M_eraseEPSt13_Rb_tree_nodeIS5_E
+ 392412
11  opt                                 0x0033bce6 _ZN4llvm12FunctionPassD1Ev +
25590
12  opt                                 0x0033c15a _ZN4llvm12FunctionPassD1Ev +
26730
13  opt                                 0x0033c781 _ZN4llvm12FunctionPassD1Ev +
28305
14  opt                                 0x0033cb94 _ZN4llvm12FunctionPassD1Ev +
29348
15  opt                                 0x0033cc0b _ZN4llvm12FunctionPassD1Ev +
29467
16  opt                                 0x00008b12
_ZN4llvm12scc_iteratorIPNS_8FunctionENS_11GraphTraitsIS2_EEED1Ev + 4514
17  opt                                 0x00001cd6 _mh_execute_header + 3286
/bin/sh: line 1:  6817 Abort trap              opt -f -o=tyvars.bc.opt.bc
-std-compile-opts tyvars.bc
===============================================================================


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 24 01:17:14 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 24 Oct 2008 01:17:14 -0500
Subject: [LLVMbugs] [Bug 2940] Assertion fails in getLowBitsSet when
	optimizing
In-Reply-To: <bug-2940-206@http.llvm.org/bugs/>
Message-ID: <200810240617.m9O6HEBR001572@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2940


Nick Lewycky <nicholas at mxc.ca> changed:

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




--- Comment #2 from Nick Lewycky <nicholas at mxc.ca>  2008-10-24 01:17:14 ---
Fixed in r58075.

(This fix won't be in LLVM 2.4 by the way. We're too late in the release
process for that.)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 24 03:08:55 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 24 Oct 2008 03:08:55 -0500
Subject: [LLVMbugs] [Bug 2919] gcc regards struct foo and const struct foo
 as __builtin_types_compatible_p
In-Reply-To: <bug-2919-206@http.llvm.org/bugs/>
Message-ID: <200810240808.m9O88tqu014214@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2919


Daniel Dunbar <daniel at zuster.org> changed:

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




-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 24 12:05:03 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 24 Oct 2008 12:05:03 -0500
Subject: [LLVMbugs] [Bug 2942] New: FunctionDecls by typedef crash the C++
	front-end
Message-ID: <bug-2942-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2942

           Summary: FunctionDecls by typedef crash the C++ front-end
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: akyrtzi at gmail.com
                CC: llvmbugs at cs.uiuc.edu


e.g:

typedef void fn(int);
fn f;

There are no ParmVarDecls created for the 'f' FunctionDecl, and it crashes the
C++ front-end when trying to use the ParmVarDecls (mostly for handling default
args).

In C there is no problem since ParmVarDecls are used only for definitions,
otherwise FunctionTypeProto is used.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 24 13:29:31 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 24 Oct 2008 13:29:31 -0500
Subject: [LLVMbugs] [Bug 2942] FunctionDecls by typedef crash the C++
	front-end
In-Reply-To: <bug-2942-206@http.llvm.org/bugs/>
Message-ID: <200810241829.m9OITVAg004383@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2942


Douglas Gregor <dgregor at apple.com> changed:

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




--- Comment #2 from Douglas Gregor <dgregor at apple.com>  2008-10-24 13:29:31 ---
Fixed on the trunk.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 24 15:28:06 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 24 Oct 2008 15:28:06 -0500
Subject: [LLVMbugs] [Bug 2945] New: always-inline removes internal functions
 w/ o always inline attribute
Message-ID: <bug-2945-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2945

           Summary: always-inline removes internal functions w/o always
                    inline attribute
           Product: libraries
           Version: trunk
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: minor
          Priority: P2
         Component: Interprocedural Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: daniel at zuster.org
                CC: llvmbugs at cs.uiuc.edu


Test case (attached):
--
; RUN: llvm-as < %s | opt -inline | llvm-dis | grep {@foo}

define internal i32 @foo() nounwind {
  ret i32 0
}
--

This is blocking always running the always inline pass (or we should remove the
test for this from the GCC test suite). clang, for example, never ever emits
(trivially) unused static definitions.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 24 17:05:47 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 24 Oct 2008 17:05:47 -0500
Subject: [LLVMbugs] [Bug 2946] New: CellSPU Error: unknown pseudo-op:
	`.cstring'
Message-ID: <bug-2946-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2946

           Summary: CellSPU Error: unknown pseudo-op: `.cstring'
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: bwatt at us.ibm.com
                CC: llvmbugs at cs.uiuc.edu


When using the CellSPU backend the following program generates a "Error:
unknown pseudo-op: `.cstring'" when compiling with GNU Cell SDK assembler,
spu-as. It appears the assembler directive, .cstring, is not supported. One
possible solution may be to generate ".section .rodata" instead.

Input, main.c:

char *s = "test";
int main(unsigned long long id) {
  return 0;
}

Script:

clang -emit-llvm main.c
llvm-as main.ll -f 
opt -std-compile-opts main.bc -o main.opt.bc -f
llc -march=cellspu main.opt.bc -f -o main.s
spu-as -g -o main.o main.s

Assembler file, main.s

        .text


        .align  3
        .global main
        .type   main, @function
main:
        il      $3, 0
        bi      $lr
        .size   main,.-main
        .cstring
        .align  4
.str:                           # '.str'
        .asciz  "test"

        .global s
        .data
        .align  3
s:                              # 's'
        .long   .str


        .ident  "(llvm 2.2+) STI CBEA Cell SPU backend"


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 24 17:19:46 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 24 Oct 2008 17:19:46 -0500
Subject: [LLVMbugs] [Bug 2947] New: CellSPU assert "Register use before def!"
Message-ID: <bug-2947-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2947

           Summary: CellSPU assert "Register use before def!"
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: bwatt at us.ibm.com
                CC: llvmbugs at cs.uiuc.edu


When using the CellSPU backend the following program gets an assert failure:
`MRI->getVRegDef(reg) && "Register use before def!"'.

Input file, main.c:

void func2();
void func() {
  func2();
}

Script:

clang -emit-llvm main.c
llvm-as main.ll -f
opt -std-compile-opts main.bc -o main.opt.bc -f
llc -march=cellspu main.opt.bc -f -o main.s

Output:

llc: LiveVariables.cpp:129: void llvm::LiveVariables::HandleVirtRegUse(unsigned
int, llvm::MachineBasicBlock*, llvm::MachineInstr*): Assertion
`MRI->getVRegDef(reg) && "Register use before def!"' failed.
llc[0xfa9d3b]
llc[0xfa9e5c]
/lib64/libc.so.6[0x358c8322a0]
/lib64/libc.so.6(gsignal+0x35)[0x358c832215]
/lib64/libc.so.6(abort+0x183)[0x358c833d83]
/lib64/libc.so.6(__assert_fail+0xe9)[0x358c82b039]
llc(_ZN4llvm13LiveVariables16HandleVirtRegUseEjPNS_17MachineBasicBlockEPNS_12MachineInstrE+0xd5)[0xd6b14d]
llc(_ZN4llvm13LiveVariables20runOnMachineFunctionERNS_15MachineFunctionE+0x5ce)[0xd6c628]
llc(_ZN4llvm19MachineFunctionPass13runOnFunctionERNS_8FunctionE+0x31)[0x820171]
llc(_ZN4llvm13FPPassManager13runOnFunctionERNS_8FunctionE+0xef)[0xf39395]
llc(_ZN4llvm23FunctionPassManagerImpl3runERNS_8FunctionE+0x74)[0xf39ed8]
llc(_ZN4llvm19FunctionPassManager3runERNS_8FunctionE+0x80)[0xf3a036]
llc(main+0x8ba)[0x7bdbc8]
/lib64/libc.so.6(__libc_start_main+0xfa)[0x358c81e32a]
llc(atan+0xf1)[0x7bc5d9]
./x: line 4: 17751 Aborted                 llc -march=cellspu main.opt.bc -f -o
main.s


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 24 21:35:05 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 24 Oct 2008 21:35:05 -0500
Subject: [LLVMbugs] [Bug 2618] Dependency files not supported
In-Reply-To: <bug-2618-206@http.llvm.org/bugs/>
Message-ID: <200810250235.m9P2Z5a5021853@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2618


Daniel Dunbar <daniel at zuster.org> changed:

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




--- Comment #4 from Daniel Dunbar <daniel at zuster.org>  2008-10-24 21:35:04 ---
Patch applied here: http://llvm.org/viewvc/llvm-project?view=rev&revision=58122


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Oct 25 10:08:01 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 25 Oct 2008 10:08:01 -0500
Subject: [LLVMbugs] [Bug 2911] Assert triggered when complex GC roots
	initialized
In-Reply-To: <bug-2911-206@http.llvm.org/bugs/>
Message-ID: <200810251508.m9PF80Wo022870@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2911


Gordon Henriksen <gordonhenriksen at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |gordonhenriksen at mac.com
             Status|NEW                         |RESOLVED
         Resolution|                            |LATER




--- Comment #1 from Gordon Henriksen <gordonhenriksen at mac.com>  2008-10-25 10:07:48 ---
This is by design. That is, the system only supports using pointers as gcroots.
I don't see this changing before a major redesign of the GC system (but patches
are welcome!).

A workaround would be to create (locally constant) pointers to your structs and
pass those to llvm.gcroot:

    %fp = alloca %FatPointerType
    store %FatPointerType { i8* null, i32 0 }, %FatPointerType* %fp
    %fp_root = alloca %FatPointerType*
    %fp_root2 = bitcast %FatPointerType** %fp_root to i8**
    call void @llvm.gcroot(i8** %fp_root2, i8* null)
    store %FatPointerType* %fp, %FatPointerType** %fp

This will cost an extra indirection in the collector and increase function call
overhead.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Oct 25 10:09:04 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 25 Oct 2008 10:09:04 -0500
Subject: [LLVMbugs] [Bug 2950] New: clang -g crashing
Message-ID: <bug-2950-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2950

           Summary: clang -g crashing
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: rdivacky at freebsd.org
                CC: llvmbugs at cs.uiuc.edu


clang -g crashes for me on the attached input with the attached backtrace


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Oct 25 15:28:02 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 25 Oct 2008 15:28:02 -0500
Subject: [LLVMbugs] [Bug 2951] New: llvm-gcc error on static initializer
Message-ID: <bug-2951-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2951

           Summary: llvm-gcc error on static initializer
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Keywords: compile-fail
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: gohman at apple.com
                CC: llvmbugs at cs.uiuc.edu


GCC 4.2 (and other versions) accept this C testcase without error:


void *p = ((void*)&((char*)0)[0]);


llvm-gcc-4.2 rejects it with this message:

t.c:1: error: initializer element is not constant

This is reproducible on both trunk and the 2.4 prerelease.

This prevents llvm-gcc from compiling SQLite version 3.6.4,
which is the current version at this time.  The specific
error message is:

sqlite3.c:12108: error: initializer element is not constant


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Oct 25 18:27:06 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 25 Oct 2008 18:27:06 -0500
Subject: [LLVMbugs] [Bug 2952] New: typo in ExprMapKeyType::operator ==()
Message-ID: <bug-2952-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2952

           Summary: typo in ExprMapKeyType::operator ==()
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: fvbommel at wxs.nl
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2138)
 --> (http://llvm.org/bugs/attachment.cgi?id=2138)
Patch to fix the typo.

There's a typo in ExprMapKeyType::operator ==(). This is the code:
-----
  bool operator==(const ExprMapKeyType& that) const {
    return this->opcode == that.opcode &&
           this->predicate == that.predicate &&
           this->operands == that.operands;
           this->indices == that.indices;
  }
-----
Note the ';' instead of '&&' on the 'operands' line.

Trivial patch attached.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Oct 25 19:20:24 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 25 Oct 2008 19:20:24 -0500
Subject: [LLVMbugs] [Bug 2952] typo in ExprMapKeyType::operator ==()
In-Reply-To: <bug-2952-206@http.llvm.org/bugs/>
Message-ID: <200810260020.m9Q0KODn008252@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2952


Bill Wendling <isanbard at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |isanbard at gmail.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Bill Wendling <isanbard at gmail.com>  2008-10-25 19:20:24 ---
Fixed. Thanks!

http://llvm.org/viewvc/llvm-project?rev=58175&view=rev


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Oct 26 04:35:58 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 26 Oct 2008 04:35:58 -0500
Subject: [LLVMbugs] [Bug 2953] New: raw_ostream operator<< resize logic buggy
Message-ID: <bug-2953-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2953

           Summary: raw_ostream operator<< resize logic buggy
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Using SVN r58183 I got this error:

a.out: raw_ostream.cpp:182: llvm::raw_ostream&
llvm::raw_ostream::operator<<(const llvm::format_object_base&): Assertion
`BytesUsed > NextBufferSize && "Didn't grow buffer!?"' failed.
Aborted

Looking at the code that assertion will always trigger if that execution path
is reached.

at line raw_ostream.cpp:163
NextBufferSize = BytesUsed;

at line raw_ostream.cpp:178
   if (BytesUsed < NextBufferSize)
      return write(&V[0], BytesUsed);

   // Otherwise, try again with a new size.
   assert(BytesUsed > NextBufferSize && "Didn't grow buffer!?");
   NextBufferSize = BytesUsed;


In this case BytesUsed == NextBufferSize, write is not called, and the
assertion fails.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Oct 26 08:58:54 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 26 Oct 2008 08:58:54 -0500
Subject: [LLVMbugs] [Bug 2954] New: Type of enumerators is wrong in C++
Message-ID: <bug-2954-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2954

           Summary: Type of enumerators is wrong in C++
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: dgregor at apple.com
                CC: llvmbugs at cs.uiuc.edu, sebastian.redl at getdesigned.at


The type of an enum constant expression is currently the underlying type
of the enum (usually int). While this is correct in C, it's incorrect in
C++ and breaks overload resolution. The type of the enumerators should be the
type of the enumeration once the enumeration has been defined.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Oct 26 14:21:17 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 26 Oct 2008 14:21:17 -0500
Subject: [LLVMbugs] [Bug 2953] raw_ostream operator<< resize logic buggy
In-Reply-To: <bug-2953-206@http.llvm.org/bugs/>
Message-ID: <200810261921.m9QJLHpD016277@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2953


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #2 from Chris Lattner <clattner at apple.com>  2008-10-26 14:21:16 ---
Fixed:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081020/068953.html

Thanks for the great testcase!


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Oct 26 18:59:43 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 26 Oct 2008 18:59:43 -0500
Subject: [LLVMbugs] [Bug 2871] bugpoint erroneously assumes exit code 1 is
 bad by default
In-Reply-To: <bug-2871-206@http.llvm.org/bugs/>
Message-ID: <200810262359.m9QNxhuB025683@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2871


Nick Lewycky <nicholas at mxc.ca> changed:

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




--- Comment #3 from Nick Lewycky <nicholas at mxc.ca>  2008-10-26 18:59:43 ---
Fixed in r58221.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Oct 26 19:17:59 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 26 Oct 2008 19:17:59 -0500
Subject: [LLVMbugs] [Bug 2644] miscompile with -loop-rotate -indvars
In-Reply-To: <bug-2644-206@http.llvm.org/bugs/>
Message-ID: <200810270017.m9R0HxmP026313@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2644


Nick Lewycky <nicholas at mxc.ca> changed:

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




--- Comment #5 from Nick Lewycky <nicholas at mxc.ca>  2008-10-26 19:17:59 ---
Nope.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Oct 26 19:54:47 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 26 Oct 2008 19:54:47 -0500
Subject: [LLVMbugs] [Bug 2468] clang doesn' t use the right calling
 convention for complex numbers
In-Reply-To: <bug-2468-206@http.llvm.org/bugs/>
Message-ID: <200810270054.m9R0sloE027659@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2468


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #2 from Chris Lattner <clattner at apple.com>  2008-10-26 19:54:47 ---
Daniel fixed this.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Oct 26 20:12:00 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 26 Oct 2008 20:12:00 -0500
Subject: [LLVMbugs] [Bug 2904] Error initializing va_list with another
	variable.
In-Reply-To: <bug-2904-206@http.llvm.org/bugs/>
Message-ID: <200810270112.m9R1C0RV028287@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2904


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #4 from Chris Lattner <clattner at apple.com>  2008-10-26 20:12:00 ---
Fixed:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20081020/008584.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Oct 26 20:34:01 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 26 Oct 2008 20:34:01 -0500
Subject: [LLVMbugs] [Bug 2955] New: can't write llvm assembler with values
 containing quotes
Message-ID: <bug-2955-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2955

           Summary: can't write llvm assembler with values containing quotes
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: tomas.l.olsen at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2139)
 --> (http://llvm.org/bugs/attachment.cgi?id=2139)
bitcode that doesn't disassemble

I'm seeing a problem with the AsmWriter, the attached module shows it.
>From my compiler, I'm able to generate native assembler and the bitcode
attached, but trying to output LLVM assembly, or disassemble the module makes
AsmWriter barf with:

AsmWriter.cpp:161: void PrintLLVMName(llvm::raw_ostream&, const char*, unsigned
int, PrefixType): Assertion `C != '"' && "Illegal character in LLVM value
name!"' failed.

Using the release_24 branch.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Oct 26 23:34:02 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 26 Oct 2008 23:34:02 -0500
Subject: [LLVMbugs] [Bug 2955] can't write llvm assembler with values
	containing quotes
In-Reply-To: <bug-2955-206@http.llvm.org/bugs/>
Message-ID: <200810270434.m9R4Y2sc018726@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2955


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-10-26 23:34:01 ---
This is fixed on mainline, producing (e.g.):

"entry_foo.test!(A!(\22foobar\22)).test":
        ret void


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Oct 26 23:35:26 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 26 Oct 2008 23:35:26 -0500
Subject: [LLVMbugs] [Bug 2937] more API changes in 2.4 not listed in
	ReleaseNotes
In-Reply-To: <bug-2937-206@http.llvm.org/bugs/>
Message-ID: <200810270435.m9R4ZQWn018785@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2937


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WONTFIX




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-10-26 23:35:25 ---
The goal is to list only the major changes that bite people in practice.  We
don't want to list every tiny little change.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Oct 26 23:39:27 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 26 Oct 2008 23:39:27 -0500
Subject: [LLVMbugs] [Bug 2935] mention that GetResultInst has been removed
 in ReleaseNotes
In-Reply-To: <bug-2935-206@http.llvm.org/bugs/>
Message-ID: <200810270439.m9R4dRVt019004@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2935


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-10-26 23:39:27 ---
done, thanks


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Oct 27 01:06:17 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 27 Oct 2008 01:06:17 -0500
Subject: [LLVMbugs] [Bug 2925] mem2reg Painfully Slow on Large BasicBlocks
In-Reply-To: <bug-2925-206@http.llvm.org/bugs/>
Message-ID: <200810270606.m9R66HaV022256@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2925


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #2 from Chris Lattner <clattner at apple.com>  2008-10-27 01:06:16 ---
Fixed, patch here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081027/068967.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Oct 27 10:06:39 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 27 Oct 2008 10:06:39 -0500
Subject: [LLVMbugs] [Bug 2957] New: Change vector_shuffle into a variadic
	SDNode
Message-ID: <bug-2957-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2957

           Summary: Change vector_shuffle into a variadic SDNode
           Product: new-bugs
           Version: unspecified
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: enhancement
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: baldrick at free.fr
                CC: llvmbugs at cs.uiuc.edu


Currently codegen represents a vector_shuffle
as a node with three operands.  The third operand
is a mask, which *must* be a BUILD_VECTOR node of
constant integers (a different kind of node that
produces a result of the right vector type isn't
good enough).  This causes problems during type
legalization because it is not always possible to
create a legal BUILD_VECTOR usable as the vector
shuffle mask.  (LegalizeDAG ignores the legality
or not of the mask argument, but LegalizeTypes
insists on legalizing the types of all nodes,
which I think is the right thing to do).

I think the best plan is to incorporate the mask
directly into the vector shuffle node, by turning
into a node with 2+N operands: the first two the
same as now, followed by the N mask elements.

This seems fairly straightforward to do, but involves
a lot of tedious work.  I don't have time for it right
now so I've opened this PR in the hope that someone
else will take it on :)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Oct 27 11:04:17 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 27 Oct 2008 11:04:17 -0500
Subject: [LLVMbugs] [Bug 2958] New: ICE with global initialised to address
 of incomplete static
Message-ID: <bug-2958-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2958

           Summary: ICE with global initialised to address of incomplete
                    static
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: richard at xmos.com
                CC: llvmbugs at cs.uiuc.edu


llvm-gcc dies trying to compile the following (invalid) program:

static struct foo s;
struct foo *p = &s;

cc1: /homelocal/richard/llvm-upstream/llvm/lib/VMCore/Constants.cpp:128: static
llvm::Constant* llvm::Constant::getNullValue(const llvm::Type*): Assertion
`!"Cannot create a null constant of that type!"' failed.
test.c:2: internal compiler error: Aborted
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://developer.apple.com/bugreporter> for instructions.

Instead llvm-gcc should give the error "storage size of 's' isn't known" and
halt compilation, as gcc does.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Oct 27 11:58:05 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 27 Oct 2008 11:58:05 -0500
Subject: [LLVMbugs] [Bug 2959] New: Miscompilation of expression with
 multiple side-effects
Message-ID: <bug-2959-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2959

           Summary: Miscompilation of expression with multiple side-effects
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: richard at xmos.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2141)
 --> (http://llvm.org/bugs/attachment.cgi?id=2141)
Resulting bitcode

When the following code is compiled with llvm-gcc and run

int *p;

int f() {
        (*p)++;
        return 1;
}

int main() {
        int b;
        p = &b;
        b++ + f();
        return (b == 2) ? 0 : 1;
}

it returns 1. Even though the order of evaluation of the operands of add is
undefined, I believe the result is well defined since the call to f is a
sequence point. Either b++ is executed first in which case the increment of b
must complete before the call to f() or f() is executed first in which case the
increment of (*p) must complete by the end of the 1st statement in f(). Either
way b should end up with the value 2, giving an expected return value of 0.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Oct 27 12:21:25 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 27 Oct 2008 12:21:25 -0500
Subject: [LLVMbugs] [Bug 2959] Miscompilation of expression with multiple
	side-effects
In-Reply-To: <bug-2959-206@http.llvm.org/bugs/>
Message-ID: <200810271721.m9RHLPmo023616@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2959


Dale Johannesen <dalej at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |dalej at apple.com
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #2 from Dale Johannesen <dalej at apple.com>  2008-10-27 12:21:23 ---
First, b in uninitialized, so the program has undefined behavior.  However,
initializing b to 0 does result in the behavior you describe (as it does with
gcc).  Second, I think this behavior is valid; in C99, see 6.5p2 and p3, and
example 7 (p15) in 5.1.2.3.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Oct 27 12:30:36 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 27 Oct 2008 12:30:36 -0500
Subject: [LLVMbugs] [Bug 2960] New: static llvm::MVT llvm::MVT::getMVT(const
 llvm::Type*, bool)   - "Unknown type!"
Message-ID: <bug-2960-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2960

           Summary: static llvm::MVT llvm::MVT::getMVT(const llvm::Type*,
                    bool) - "Unknown type!"
           Product: new-bugs
           Version: unspecified
          Platform: Sun
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: wilsonkk at shaw.ca
                CC: llvmbugs at cs.uiuc.edu


llc crash. This bug occurs on Sparc-unknown-linux-gnu for first-class
aggrgates. Not sure if this is a duplicate of bug 2594.

See attachment for reduced test case.

Thanks,
K.Wilson


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Oct 27 12:39:29 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 27 Oct 2008 12:39:29 -0500
Subject: [LLVMbugs] [Bug 2634] Remat applied to updated value
In-Reply-To: <bug-2634-206@http.llvm.org/bugs/>
Message-ID: <200810271739.m9RHdTIG024316@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2634


David Greene <greened at obbligato.org> changed:

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




--- Comment #5 from David Greene <greened at obbligato.org>  2008-10-27 12:39:28 ---
Fixed with rev. 58255.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Oct 27 17:25:12 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 27 Oct 2008 17:25:12 -0500
Subject: [LLVMbugs] [Bug 2959] Miscompilation of expression with multiple
	side-effects
In-Reply-To: <bug-2959-206@http.llvm.org/bugs/>
Message-ID: <200810272225.m9RMPCiD002636@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2959


Neil Booth <neil at daikokuya.co.uk> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |neil at daikokuya.co.uk
             Status|RESOLVED                    |REOPENED
         Resolution|INVALID                     |




--- Comment #4 from Neil Booth <neil at daikokuya.co.uk>  2008-10-27 17:25:10 ---
Dale - you quote several sections of the standard without explaining how they
relate to this issue.

My reading of the standard (assuming b is initialized to 0) sections you quote
-- I only find 6.5p2 relevant -- tells me that this is indeed a bug and b
should be 2.  You seem to think that b is stored twice between sequence points;
could you point out how / where?


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Oct 27 18:58:22 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 27 Oct 2008 18:58:22 -0500
Subject: [LLVMbugs] [Bug 2961] New: llvm-ld doesn't accept -march option
Message-ID: <bug-2961-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2961

           Summary: llvm-ld doesn't accept -march option
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: minor
          Priority: P2
         Component: llvm-ld
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: yliang6 at illinois.edu
                CC: llvmbugs at cs.uiuc.edu


llvm-ld does not accept -march option which the llvm-ld man page says it
should...

Either such support should be added or the document should be revised.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct 28 00:37:09 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 28 Oct 2008 00:37:09 -0500
Subject: [LLVMbugs] [Bug 2962] New: stack realignment code does not work
 with alloca in function
Message-ID: <bug-2962-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2962

           Summary: stack realignment code does not work with alloca in
                    function
           Product: libraries
           Version: 1.0
          Platform: PC
        OS/Version: All
            Status: NEW
          Keywords: miscompilation
          Severity: normal
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: llvmbugs at cs.uiuc.edu


Stack realignment does not work properly for this case:

target datalayout =
"e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:32:32"
target triple = "i386-pc-linux-gnu"

define void @foo(i32 %t) nounwind {
  %tmp1210 = alloca i8, i32 40, align 16
  store i8 0, i8* %tmp1210

  %x = alloca i8, i32 %t
  ret void
}


If you comment out "x", the stack realignment does work.  This causes
miscompilations when the code generator infers memsets using xmm regs e.g. Bug
2888


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct 28 00:47:17 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 28 Oct 2008 00:47:17 -0500
Subject: [LLVMbugs] [Bug 2963] New: "enumerator value is not an integer
 constant" error when using macros
Message-ID: <bug-2963-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2963

           Summary: "enumerator value is not an integer constant" error when
                    using macros
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: parser
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: alexei.svitkine at gmail.com
                CC: llvmbugs at cs.uiuc.edu


I have some code:

typedef short short_fixed;

enum
{
        // 8.8 short_fixed
        SHORT_FIXED_FRACTIONAL_BITS= 8,
        SHORT_FIXED_ONE= 1<<SHORT_FIXED_FRACTIONAL_BITS
};

#define FLOAT_TO_SHORT_FIXED(f) ((short_fixed)((f)*SHORT_FIXED_ONE))

enum
{
        SOME_VALUE= FLOAT_TO_SHORT_FIXED(0.1)
};

This code causes clang to output the following error:

error: enumerator value for 'SOME_VALUE' is not an integer constant
        SOME_VALUE = FLOAT_TO_SHORT_FIXED(0.1)

However, the same code is accepted fine by both GCC and VC++ 7.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct 28 01:08:00 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 28 Oct 2008 01:08:00 -0500
Subject: [LLVMbugs] [Bug 2888] linux regressions with 2.4
In-Reply-To: <bug-2888-206@http.llvm.org/bugs/>
Message-ID: <200810280608.m9S680Zr029154@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2888


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #23 from Chris Lattner <clattner at apple.com>  2008-10-28 01:07:57 ---
That fixed 176.gcc.  I'm assuming the same problem was behind 126.gcc.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct 28 01:22:47 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 28 Oct 2008 01:22:47 -0500
Subject: [LLVMbugs] [Bug 2963] "enumerator value is not an integer constant"
 error when using macros
In-Reply-To: <bug-2963-206@http.llvm.org/bugs/>
Message-ID: <200810280622.m9S6MlUv029745@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2963


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #2 from Chris Lattner <clattner at apple.com>  2008-10-28 01:22:47 ---
Your code is valid.  A floating point multiply is not a valid
integer-constant-expression.  The only floating point expression that is, is a
floating point constant that is immediately cast to an integer.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct 28 01:53:39 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 28 Oct 2008 01:53:39 -0500
Subject: [LLVMbugs] [Bug 2964] New: Prolog/Epilog Insertion going crazy on
	PPC
Message-ID: <bug-2964-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2964

           Summary: Prolog/Epilog Insertion going crazy on PPC
           Product: libraries
           Version: 1.0
          Platform: PC
        OS/Version: All
            Status: NEW
          Keywords: code-quality
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: llvmbugs at cs.uiuc.edu


go into test/CodeGen/PowerPC and do this:

$ llvm-as < compare-fcmp-ord.ll | llvm-extract -func test | llc -march=ppc32
-print-machineinstrs -debug-pass=Executions | & less

Before PEI, the code looks like this:

entry: 0x1812ab4, LLVM BB @0x14021a0, ID#0:
Live Ins: %F1 %F2
        %CR7<def> = FCMPUD %F1<kill>, %F2<kill>
        %R2<def> = MFCR %CR0<imp-use,kill>, %CR1<imp-use,kill>,
%CR2<imp-use,kill>, %CR3<imp-use,kill>, 
%CR4<imp-use,kill>, %CR5<imp-use,kill>, %CR6<imp-use,kill>, %CR7<imp-use,kill>
        %R3<def> = RLWINM %R2, 29, 31, 31
        %R2<def> = RLWINM %R2<kill>, 31, 31, 31
        %R3<def> = OR %R2<kill>, %R3<kill>
        BLR 20, %reg0, %R3<imp-use,kill>, %LR<imp-use>

Afterwards, it looks like this:

entry: 0x1812ab4, LLVM BB @0x14021a0, ID#0:
Live Ins: %F1 %F2 %CR2 %CR3 %CR4
        %R0<def> = MFCR %CR0<imp-use>, %CR1<imp-use>, %CR2<imp-use>,
%CR3<imp-use>, %CR4<imp-use>, %CR5<
imp-use>, %CR6<imp-use>, %CR7<imp-use>
        %R0<def> = RLWINM %R0, 8, 0, 31
        STW %R0<kill>, -4, %R1
        %R0<def> = MFCR %CR0<imp-use>, %CR1<imp-use>, %CR2<imp-use>,
%CR3<imp-use>, %CR4<imp-use>, %CR5<
imp-use>, %CR6<imp-use>, %CR7<imp-use>
        %R0<def> = RLWINM %R0, 12, 0, 31
        STW %R0<kill>, -8, %R1
        %R0<def> = MFCR %CR0<imp-use>, %CR1<imp-use>, %CR2<imp-use>,
%CR3<imp-use>, %CR4<imp-use>, %CR5<
imp-use>, %CR6<imp-use>, %CR7<imp-use>
        %R0<def> = RLWINM %R0, 16, 0, 31
        STW %R0<kill>, -12, %R1
        %CR7<def> = FCMPUD %F1<kill>, %F2<kill>
        %R2<def> = MFCR %CR0<imp-use,kill>, %CR1<imp-use,kill>,
%CR2<imp-use,kill>, %CR3<imp-use,kill>, 
%CR4<imp-use,kill>, %CR5<imp-use,kill>, %CR6<imp-use,kill>, %CR7<imp-use,kill>
        %R3<def> = RLWINM %R2, 29, 31, 31
        %R2<def> = RLWINM %R2<kill>, 31, 31, 31
        %R3<def> = OR %R2<kill>, %R3<kill>
        %R0<def> = LWZ -12, %R1
        %R0<def> = RLWINM %R0, 16, 0, 31
        %CR4<def> = MTCRF %R0
        %R0<def> = LWZ -8, %R1
        %R0<def> = RLWINM %R0, 20, 0, 31
        %CR3<def> = MTCRF %R0
        %R0<def> = LWZ -4, %R1
        %R0<def> = RLWINM %R0, 24, 0, 31
        %CR2<def> = MTCRF %R0
        BLR 20, %reg0, %R3<imp-use,kill>, %LR<imp-use>


This is a huge code quality regression from the past.  It looks like
individuals CR is being saved and restored.  CR7 (which is def'd by this code)
is a call clobber register according to this:

PPCInstrInfo.td:422.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct 28 07:05:24 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 28 Oct 2008 07:05:24 -0500
Subject: [LLVMbugs] [Bug 2966] New: double va_end crashes clang
Message-ID: <bug-2966-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2966

           Summary: double va_end crashes clang
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: rdivacky at freebsd.org
                CC: llvmbugs at cs.uiuc.edu


when you use va_end twice as in:

#include <stdarg.h>

void
run_err(const char *fmt, ...)
{
        va_list ap;

        va_start(ap, fmt);
        va_end(ap);

        va_end(ap);
}

it crashes clang with:

Assertion failed: (E && !hasAggregateLLVMType(E->getType()) && "Invalid scalar
expression to emit"), function EmitScalarExpr, file CGExprScalar.cpp, line
1179.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct 28 08:41:30 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 28 Oct 2008 08:41:30 -0500
Subject: [LLVMbugs] [Bug 2963] "enumerator value is not an integer constant"
 error when using macros
In-Reply-To: <bug-2963-206@http.llvm.org/bugs/>
Message-ID: <200810281341.m9SDfUT1021701@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2963


Neil Booth <neil at daikokuya.co.uk> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |neil at daikokuya.co.uk
             Status|RESOLVED                    |REOPENED
         Resolution|INVALID                     |




--- Comment #4 from Neil Booth <neil at daikokuya.co.uk>  2008-10-28 08:41:29 ---
My front end gives

"/tmp/bug.c", line 14: error: a floating constant in an integer constant
        expression must be cast to integer type
        SOME_VALUE= FLOAT_TO_SHORT_FIXED(0.1)
                    ^

1 error found compiling "/tmp/bug.c".

which is perhaps slightly more helpful.  There isn't a good way around this -
the standard is reluctant to impose any hard requirements on compile-time
calculation of target floating point arithmetic, viewing it as too much of a
burden in general on implementations, although for IEEE targets it's not too
hard to do perfectly with about 1000 lines of code (and clang has this code). 
But that is the reason for this constraint :(

Perhaps clang should accept with a warning, unless pedantic?

Reopening as it seems others agree.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct 28 08:48:48 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 28 Oct 2008 08:48:48 -0500
Subject: [LLVMbugs] [Bug 2967] New: instcombine is removing a traping
	instruction
Message-ID: <bug-2967-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2967

           Summary: instcombine is removing a traping instruction
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Scalar Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: alenhar2 at uiuc.edu
                CC: llvmbugs at cs.uiuc.edu


In this code, instcombine reduces it to ret void.  The trapping load is
removed.  Essentially I would contend the load should be treated in the same
way a noreturn function before an unreachable is treated.


target datalayout =
"e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-$
target triple = "i386-pc-linux-gnu"

define void @foo(i32 %x) nounwind {
entry:
        %0 = icmp eq i32 %x, 0          ; <i1> [#uses=1]
        br i1 %0, label %bb, label %return

bb:             ; preds = %entry
        %1 = volatile load i32* null            ; <i32> [#uses=0]
        unreachable
        br label %return
return:         ; preds = %entry
        ret void
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct 28 08:56:40 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 28 Oct 2008 08:56:40 -0500
Subject: [LLVMbugs] [Bug 2959] Miscompilation of expression with multiple
	side-effects
In-Reply-To: <bug-2959-206@http.llvm.org/bugs/>
Message-ID: <200810281356.m9SDueID022181@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2959


Neil Booth <neil at daikokuya.co.uk> changed:

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




--- Comment #7 from Neil Booth <neil at daikokuya.co.uk>  2008-10-28 08:56:39 ---
I apologise, I misread the code entirely.  It is indeed unspecified whether the
result is 1 or 2: it depends which operand is evaluated first.  I'd initially
misread it and thought it didn't matter.  In particular

  b++ + f()

is 0 + 1 if the "b++" is calculated before the f() call; it is  1 + 1 if the
"b++" is calculated after the f() call.

Sorry for the noise.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct 28 15:05:51 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 28 Oct 2008 15:05:51 -0500
Subject: [LLVMbugs] [Bug 2961] llvm-ld doesn't accept -march option
In-Reply-To: <bug-2961-206@http.llvm.org/bugs/>
Message-ID: <200810282005.m9SK5p8s004317@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2961


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-10-28 15:05:51 ---
Fixed the documentation:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081027/069064.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct 28 15:24:36 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 28 Oct 2008 15:24:36 -0500
Subject: [LLVMbugs] [Bug 2867] ICE on invalid;
	class definiton inside method definition
In-Reply-To: <bug-2867-206@http.llvm.org/bugs/>
Message-ID: <200810282024.m9SKOaFV005006@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2867


Dale Johannesen <dalej at apple.com> changed:

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




--- Comment #6 from Dale Johannesen <dalej at apple.com>  2008-10-28 15:24:33 ---
Fix applied here (courtesy of Mike Stump).
http://llvm.org/viewvc/llvm-project?view=rev&revision=58347
Nick, please see if your original code works.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct 28 16:17:12 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 28 Oct 2008 16:17:12 -0500
Subject: [LLVMbugs] [Bug 2968] New: ppcf128->i32 conversion could go in PPC
 specific code
Message-ID: <bug-2968-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2968

           Summary: ppcf128->i32 conversion could go in PPC specific code
           Product: libraries
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: PowerPC
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: dalej at apple.com
                CC: llvmbugs at cs.uiuc.edu


Currently this conversion is handled in LegalizeDAG and is now duplicated in
LegalizeTypes.  It would be more aesthetically pleasing to put in the PPC
specific area, by custom lowering somehow.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct 28 16:33:42 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 28 Oct 2008 16:33:42 -0500
Subject: [LLVMbugs] [Bug 2969] New: clang turns const arrays inside
 functions into allocas/ stores instead of "internal constant" arrays
Message-ID: <bug-2969-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2969

           Summary: clang turns const arrays inside functions into
                    allocas/stores instead of "internal constant" arrays
           Product: clang
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: richard at godbee.net
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2144)
 --> (http://llvm.org/bugs/attachment.cgi?id=2144)
C source code + complete LLVM assembly output from llvm-gcc and clang

int f1(int n) {
        const int array_one[] = { 1, 2, 3 };
        return array_one[n];
}

int f2(int n) {
        static const int array_two[] = { 7, 8, 9 };
        return array_two[n];
}

/* Note that f1() and f2() are identical, except f1()'s int[] is "const" and
f2()'s int[] is "static const". */

**********

llvm-gcc (as shipped with Xcode 3.1.1) will emit "internal constant" arrays for
both array_one and array_two...

@array_one.1538 = internal constant [3 x i32] [ i32 1, i32 2, i32 3 ]          
; <[3 x i32]*> [#uses=1]
@array_two.1545 = internal constant [3 x i32] [ i32 7, i32 8, i32 9 ]          
; <[3 x i32]*> [#uses=1]

Like llvm-gcc, clang (trunk r58037) will emit an "internal constant" array for
array_two...

@f2.array_two = internal global [3 x i32] [ i32 7, i32 8, i32 9 ]              
; <[3 x i32]*> [#uses=1]

However, for array_one, clang will alloca space for the array and store each
element one-by-one inside of f1()...

define i32 @f1(i32 %n) nounwind {
entry:
        %array_one = alloca [3 x i32], align 4          ; <[3 x i32]*>
[#uses=4]
        %.array = getelementptr [3 x i32]* %array_one, i32 0, i32 0            
; <i32*> [#uses=1]
        store i32 1, i32* %.array
        %.array1 = getelementptr [3 x i32]* %array_one, i32 0, i32 1           
; <i32*> [#uses=1]
        store i32 2, i32* %.array1
        %.array2 = getelementptr [3 x i32]* %array_one, i32 0, i32 2           
; <i32*> [#uses=1]
        store i32 3, i32* %.array2
        %arrayidx = getelementptr [3 x i32]* %array_one, i32 0, i32 %n         
; <i32*> [#uses=1]
        %tmp3 = load i32* %arrayidx             ; <i32> [#uses=1]
        ret i32 %tmp3
}

opt -std-compile-args doesn't transform the alloca + stores into an "internal
constant" array, either.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct 28 17:07:21 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 28 Oct 2008 17:07:21 -0500
Subject: [LLVMbugs] [Bug 2970] New: -march flag doesn't work.
Message-ID: <bug-2970-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2970

           Summary: -march flag doesn't work.
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Basic
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: scooter.phd at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Built clang per directions in README, placed in the usual tools/ top-level
directory.

$ Debug/bin/clang -arch=ppc32 -emit-llvm /tmp/zzz.c
Sorry, I don't know what target this is: ppc32-unknown-linux-gnu
Please use -triple or -arch.
$

If this doesn't work for ppc32, it'll make it just that much more interesting
for debugging CellSPU. :-)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct 28 19:05:25 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 28 Oct 2008 19:05:25 -0500
Subject: [LLVMbugs] [Bug 2971] New: &((char*)0)[0] not regarded as static
	initializer
Message-ID: <bug-2971-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2971

           Summary: &((char*)0)[0] not regarded as static initializer
           Product: clang
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nunoplopes at sapo.pt
                CC: llvmbugs at cs.uiuc.edu


The following code crashes clang:

typedef struct FuncDef FuncDef;
struct FuncDef {};

void sqlite3RegisterDateTimeFunctions(void){
  static FuncDef aDateTimeFuncs = {((void*)&((char*)0)[0])};
}


With the error message:
clang: CGDecl.cpp:86: llvm::GlobalValue*
clang::CodeGen::CodeGenFunction::GenerateStaticBlockVarDecl(const
clang::VarDecl&, bool, const char*): Assertion
`getContext().getLangOptions().CPlusPlus && "only C++ supports non-constant
static initializers!"' failed.

(I'm not sure if this a duplicate or not. I have the strange feeling it's not
the first time I'm seeing this bug..)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct 28 19:12:59 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 28 Oct 2008 19:12:59 -0500
Subject: [LLVMbugs] [Bug 2970] -march flag doesn't work.
In-Reply-To: <bug-2970-206@http.llvm.org/bugs/>
Message-ID: <200810290012.m9T0Cx9x013356@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2970


Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |daniel at zuster.org
             Status|NEW                         |RESOLVED
         Resolution|                            |WORKSFORME




--- Comment #1 from Daniel Dunbar <daniel at zuster.org>  2008-10-28 19:12:59 ---
clang accepts -arch=ppc as 32-bit ppc (ppc64 also is recognized).
--
ddunbar at ddunbar2:tmp$ echo "int x;" | clang -arch=ppc -emit-llvm
; ModuleID = '-'
target datalayout =
"E-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v128:128:128"
target triple = "ppc-apple-darwin10.0.0d2"
@x = common global i32 0, align 4               ; <i32*> [#uses=0]
--

Is there a reason you expect the arch to be ppc32?


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct 28 19:13:20 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 28 Oct 2008 19:13:20 -0500
Subject: [LLVMbugs] [Bug 2970] -march flag doesn't work.
In-Reply-To: <bug-2970-206@http.llvm.org/bugs/>
Message-ID: <200810290013.m9T0DK1f013386@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2970


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|WORKSFORME                  |INVALID




--- Comment #2 from Chris Lattner <clattner at apple.com>  2008-10-28 19:13:19 ---
This is because clang has no target support for that triple, as it just told
you.  You have to explicitly teach clang about targets.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Oct 28 23:48:41 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 28 Oct 2008 23:48:41 -0500
Subject: [LLVMbugs] [Bug 2966] double va_end crashes clang
In-Reply-To: <bug-2966-206@http.llvm.org/bugs/>
Message-ID: <200810290448.m9T4mf5c023051@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2966


Eli Friedman <sharparrow1 at yahoo.com> changed:

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




--- Comment #5 from Eli Friedman <sharparrow1 at yahoo.com>  2008-10-28 23:48:41 ---
Fixed in r58366, although note that clang/LLVM can't do codegen for va_arg yet.
 See bug 2825 and bug 1740.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 29 01:24:53 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 29 Oct 2008 01:24:53 -0500
Subject: [LLVMbugs] [Bug 2973] New: very poor choice by the inliner (code
	size + perf)
Message-ID: <bug-2973-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2973

           Summary: very poor choice by the inliner (code size + perf)
           Product: libraries
           Version: 1.0
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Interprocedural Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: llvmbugs at cs.uiuc.edu


SPASS has code that looks like this:

LIST clause_NumberSort(LIST List) {
  return list_Sort(List, (BOOL (*) (POINTER, POINTER)) clause_NumberLower);
}

LIST pcheck_ClauseNumberMergeSort(LIST L) {
  return clause_NumberSort(L);
}

Because the inliner is running bottom up, it inlines clause_NumberSort into
pcheck_ClauseNumberMergeSort, duplicating it!  It would be much better to not
inline clause_NumberSort in this case, and just inline
pcheck_ClauseNumberMergeSort into its callers instead.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 29 05:01:40 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 29 Oct 2008 05:01:40 -0500
Subject: [LLVMbugs] [Bug 2974] New: Invalid debug information for array types
Message-ID: <bug-2974-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2974

           Summary: Invalid debug information for array types
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: richard at xmos.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2146)
 --> (http://llvm.org/bugs/attachment.cgi?id=2146)
source file

If a program with a variable with an array type is compiled with -g in llvm-gcc
then the type of the variable is incorrectly given as it's element type.

For example the attached program declares x as follows:

int x[10];

but the dwarf information shows x is of type int:

gdb a.out
GNU gdb Red Hat Linux (6.3.0.0-1.153.el4rh)
Copyright 2004 Free Software Foundation, Inc.
GDB is free software, covered by the GNU General Public License, and you are
welcome to change it and/or distribute copies of it under certain conditions.
Type "show copying" to see the conditions.
There is absolutely no warranty for GDB.  Type "show warranty" for details.
This GDB was configured as "x86_64-redhat-linux-gnu"...Using host libthread_db
library "/lib64/tls/libthread_db.so.1".

(gdb) whatis x
type = int
(gdb)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 29 05:17:16 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 29 Oct 2008 05:17:16 -0500
Subject: [LLVMbugs] [Bug 2975] New: memory leak in Type.cpp
Message-ID: <bug-2975-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2975

           Summary: memory leak in Type.cpp
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: minor
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


SVN r58373.

==6329== 928 (440 direct, 488 indirect) bytes in 5 blocks are definitely lost
in loss record 20 of 23
==6329==    at 0x4C21849: operator new(unsigned long) (vg_replace_malloc.c:230)
==6329==    by 0xA43776: llvm::PointerType::get(llvm::Type const*, unsigned)
(Type.cpp:1333)
==6329==    by 0x9C75DF: llvm::BitcodeReader::ParseTypeTable()
(BitcodeReader.cpp:490)
==6329==    by 0x9CE030: llvm::BitcodeReader::ParseModule(std::string const&)
(BitcodeReader.cpp:1070)
==6329==    by 0x9CF294: llvm::BitcodeReader::ParseBitcode()
(BitcodeReader.cpp:1359)
==6329==    by 0x9CF334: llvm::getBitcodeModuleProvider(llvm::MemoryBuffer*,
std::string*) (BitcodeReader.cpp:2070)
==6329==    by 0x9CF3B2: llvm::ParseBitcodeFile(llvm::MemoryBuffer*,
std::string*) (BitcodeReader.cpp:2086)
==6329==    by 0x7399FE: main (opt.cpp:435)
==6329==
==6329==
==6329== 1,872 (1,280 direct, 592 indirect) bytes in 18 blocks are definitely
lost in loss record 22 of 23
==6329==    at 0x4C21849: operator new(unsigned long) (vg_replace_malloc.c:230)
==6329==    by 0xA57051: llvm::User::operator new(unsigned long, unsigned)
(Value.cpp:384)
==6329==    by 0xA007C8: llvm::ConstantInt::operator new(unsigned long)
(Constants.h:55)
==6329==    by 0x9F2D7E: llvm::ConstantInt::get(llvm::APInt const&)
(Constants.cpp:272)
==6329==    by 0x9F2DFF: llvm::ConstantInt::get(llvm::Type const*, unsigned
long, bool) (Constants.cpp:254)
==6329==    by 0x9C91DB: llvm::BitcodeReader::ParseConstants()
(BitcodeReader.cpp:849)
==6329==    by 0x9CE090: llvm::BitcodeReader::ParseModule(std::string const&)
(BitcodeReader.cpp:1082)
==6329==    by 0x9CF294: llvm::BitcodeReader::ParseBitcode()
(BitcodeReader.cpp:1359)
==6329==    by 0x9CF334: llvm::getBitcodeModuleProvider(llvm::MemoryBuffer*,
std::string*) (BitcodeReader.cpp:2070)
==6329==    by 0x9CF3B2: llvm::ParseBitcodeFile(llvm::MemoryBuffer*,
std::string*) (BitcodeReader.cpp:2086)
==6329==    by 0x7399FE: main (opt.cpp:435)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 29 06:51:27 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 29 Oct 2008 06:51:27 -0500
Subject: [LLVMbugs] [Bug 2976] New: Mips backend problem
Message-ID: <bug-2976-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2976

           Summary: Mips backend problem
           Product: new-bugs
           Version: unspecified
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: ben.swire at tandberg.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2152)
 --> (http://llvm.org/bugs/attachment.cgi?id=2152)
disassembled bugpoint output

Using llvm 2.3, and llvm-gcc 4.2-2.3-x86-linux-RHEL4, I get a problem with the
Mips backend. I am basing my own backend on this one, so tried a test with it
first. A simple program outputs "Hello world", creates a variable and counts to
5. 

Running through instructions in example at the bottom of
http://www.llvm.org/docs/GettingStarted.html I run into trouble at step 6,
running llc with the argument -march=mips.

I have run bugpoint to reduce and llvm-dis on the output, and the problem seems
to be in returning an i32. (assuming I understand the system correctly!)

When running llc normally, I get the following error:

[bens at BenSCentOS llvm]$ llc hello.bc -f -o hello.s -march=mips
terminate called after throwing an instance of 'std::bad_alloc'
  what():  St9bad_alloc
llc[0x88409bf]
/lib/libc.so.6(abort+0x101)[0xc7a631]
/usr/lib/libstdc++.so.6(_ZN9__gnu_cxx27__verbose_terminate_handlerEv+0x150)[0x7cbec60]
/usr/lib/libstdc++.so.6[0x7cbc6c5]
/usr/lib/libstdc++.so.6[0x7cbc702]
/usr/lib/libstdc++.so.6[0x7cbc83a]
/usr/lib/libstdc++.so.6(_Znwj+0x7e)[0x7cbccbe]
llc(_ZNSt6vectorISt4pairIPKN4llvm5SUnitEjESaIS5_EE13_M_insert_auxEN9__gnu_cxx17__normal_iteratorIPS5_S7_EERKS5_+0xb8)[0x8503708]
[0x77f56008]
Aborted

Any ideas?


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 29 07:18:13 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 29 Oct 2008 07:18:13 -0500
Subject: [LLVMbugs] [Bug 2977] New: new type legalizer fails to codegen some
	va_args
Message-ID: <bug-2977-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2977

           Summary: new type legalizer fails to codegen some va_args
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Windows NT
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nunoplopes at sapo.pt
                CC: llvmbugs at cs.uiuc.edu


The following code crashes llc:

define i8* @ap_php_conv_p2(){
entry:
        %ap.addr = alloca i8*           ; <i8**> [#uses=36]
        br label %sw.bb301
sw.bb301:
        %0 = va_arg i8** %ap.addr, i64          ; <i64> [#uses=1]
        br label %sw.bb301
}


with the error:

ExpandIntegerResult #0: 0x906fc34: i64,ch = vaarg 0x906fa08, 0x906fbac,
0x906fb24
llc: LegalizeIntegerTypes.cpp:951: void
llvm::DAGTypeLegalizer::ExpandIntegerResult(llvm::SDNode*, unsigned int):
Assertion `0 && "Do not know how to expand the result of this operator!"'
failed.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From neil at daikokuya.co.uk  Wed Oct 29 07:51:49 2008
From: neil at daikokuya.co.uk (Neil Booth)
Date: Wed, 29 Oct 2008 21:51:49 +0900
Subject: [LLVMbugs] [Bug 2971] New: &((char*)0)[0] not regarded
	as	static initializer
In-Reply-To: <bug-2971-206@http.llvm.org/bugs/>
References: <bug-2971-206@http.llvm.org/bugs/>
Message-ID: <20081029125149.GA5199@daikokuya.co.uk>

bugzilla-daemon at cs.uiuc.edu wrote:-

> http://llvm.org/bugs/show_bug.cgi?id=2971
> 
>            Summary: &((char*)0)[0] not regarded as static initializer
>            Product: clang
>            Version: unspecified
>           Platform: All
>         OS/Version: All
>             Status: NEW
>           Severity: normal
>           Priority: P2
>          Component: LLVM Codegen
>         AssignedTo: unassignedbugs at nondot.org
>         ReportedBy: nunoplopes at sapo.pt
>                 CC: llvmbugs at cs.uiuc.edu
> 
> 
> The following code crashes clang:
> 
> typedef struct FuncDef FuncDef;
> struct FuncDef {};
> 
> void sqlite3RegisterDateTimeFunctions(void){
>   static FuncDef aDateTimeFuncs = {((void*)&((char*)0)[0])};
> }
> 
> 
> With the error message:
> clang: CGDecl.cpp:86: llvm::GlobalValue*
> clang::CodeGen::CodeGenFunction::GenerateStaticBlockVarDecl(const
> clang::VarDecl&, bool, const char*): Assertion
> `getContext().getLangOptions().CPlusPlus && "only C++ supports non-constant
> static initializers!"' failed.
> 
> (I'm not sure if this a duplicate or not. I have the strange feeling it's not
> the first time I'm seeing this bug..)

It's not a bug; casts must be to arithmetic type (it would be nice if
Clang were more explicit about the real problem, rather than just saying
it's no good...).

Neil.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 29 09:28:30 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 29 Oct 2008 09:28:30 -0500
Subject: [LLVMbugs] [Bug 2977] new type legalizer fails to codegen some
	va_args
In-Reply-To: <bug-2977-206@http.llvm.org/bugs/>
Message-ID: <200810291428.m9TESU3x019335@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2977


Duncan Sands <baldrick at free.fr> changed:

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




--- Comment #1 from Duncan Sands <baldrick at free.fr>  2008-10-29 09:28:29 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081027/069093.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 29 10:13:37 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 29 Oct 2008 10:13:37 -0500
Subject: [LLVMbugs] [Bug 2979] New: lua assumes bash shell
Message-ID: <bug-2979-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2979

           Summary: lua assumes bash shell
           Product: Test Suite
           Version: trunk
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Programs Tests
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: baldrick at free.fr
                CC: llvmbugs at cs.uiuc.edu


The lua test's generate_inputs.sh contains some bashisms
(in spite of starting #!/bin/sh) which cause it to fail
here:

        declare -i count=$2

^ declare is a bashism (sh error: "declare: not found")

        output=${1/.txt/${3}.txt}
^ sh error: "bad substitution"

                let count=$((count - 50))
^ I don't think sh allows this construction.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 29 11:08:52 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 29 Oct 2008 11:08:52 -0500
Subject: [LLVMbugs] [Bug 2980] New: llc crashes with Cannot yet select:
 0xa60d794: ch = declare 0xa60d568, 0xa60d70c, 0xa60d8a4
Message-ID: <bug-2980-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2980

           Summary: llc crashes with Cannot yet select: 0xa60d794: ch =
                    declare 0xa60d568, 0xa60d70c, 0xa60d8a4
           Product: tools
           Version: 2.4
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: martin.nowack at inf.tu-dresden.de
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2156)
 --> (http://llvm.org/bugs/attachment.cgi?id=2156)
Bugpoint example

Tried to generate assembler code with a few linked bc files.

llc filename.bc -f

Crash in:
        SelectionDAGISel::CodeGenAndEmitDAG()

Thanks for help.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 29 12:47:15 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 29 Oct 2008 12:47:15 -0500
Subject: [LLVMbugs] [Bug 2967] instcombine is removing a traping instruction
In-Reply-To: <bug-2967-206@http.llvm.org/bugs/>
Message-ID: <200810291747.m9THlFjp026710@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2967


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #4 from Chris Lattner <clattner at apple.com>  2008-10-29 12:47:15 ---
I agree that this is marginal behavior.  Considering how poorly defined
'volatile' is though, it seems reasonable to err on the side of conservatism. 
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081027/069096.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 29 13:08:41 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 29 Oct 2008 13:08:41 -0500
Subject: [LLVMbugs] [Bug 2917] [llvm2.4-prerelease]
 MultiSource/Benchmarks/tramp3d-v4/ LLC, CBE, JIT fails
In-Reply-To: <bug-2917-206@http.llvm.org/bugs/>
Message-ID: <200810291808.m9TI8f1J027654@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2917


Duncan Sands <baldrick at free.fr> changed:

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




--- Comment #20 from Duncan Sands <baldrick at free.fr>  2008-10-29 13:08:40 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081027/069098.html

If this survives a round in the nightly testers, hopefully it can be pulled
into the release.  The danger is that it might break something since it is
touching a fragile part of llvm-gcc.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 29 13:42:41 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 29 Oct 2008 13:42:41 -0500
Subject: [LLVMbugs] [Bug 2964] Prolog/Epilog Insertion going crazy on PPC
In-Reply-To: <bug-2964-206@http.llvm.org/bugs/>
Message-ID: <200810291842.m9TIgfK7028829@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2964


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #4 from Chris Lattner <clattner at apple.com>  2008-10-29 13:42:41 ---
Awesome, thanks Dale.  I added a testcase here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081027/069103.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 29 20:22:19 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 29 Oct 2008 20:22:19 -0500
Subject: [LLVMbugs] [Bug 2981] New: Vector logic does not get optimized
Message-ID: <bug-2981-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2981

           Summary: Vector logic does not get optimized
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Windows NT
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicolas at capens.net
                CC: llvmbugs at cs.uiuc.edu


The attached testcase contains a logical and operation that does not get
optimized away. Here's the code in 'pseudo vector C':

void foo(int4 *in, int count, int4 *out)
{
    int4 x = *in;
    int4 a[1];
    a[0] = {0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF};

    int t = 0;

    for(int i = 0; i < count; i++)
    {
        x = x & a[t];   // Optimize me

        *out = x;
    }
}

The combination of the array and the loop somehow stop LLVM from eliminating
the and operation. Replacing the vectors by scalars does optimize it away
though. So possibly some code isn't treating scalars and vectors equally for
this type of optimization. I suspect that several other vector operations don't
get fully optimized either.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 29 20:53:49 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 29 Oct 2008 20:53:49 -0500
Subject: [LLVMbugs] [Bug 2947] CellSPU assert "Register use before def!"
In-Reply-To: <bug-2947-206@http.llvm.org/bugs/>
Message-ID: <200810300153.m9U1rnYr011493@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2947


Scott Michel <scooter.phd at gmail.com> changed:

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




--- Comment #6 from Scott Michel <scooter.phd at gmail.com>  2008-10-29 20:53:48 ---
R3-R79 are now properly spilled to stack for vararg-marked functions; see svn
revision 58415.

Note, however, that full testing of this code cannot proceed much farther:
(a) clang apparently cannot generate code for "va_arg(argptr, type)" (as of
this afternoon)
(b) llvm-gcc-4.2 doesn't yet fully build for CellSPU.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Oct 29 21:19:21 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 29 Oct 2008 21:19:21 -0500
Subject: [LLVMbugs] [Bug 2982] New: ice: This target-independent node should
 have been selected!
Message-ID: <bug-2982-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2982

           Summary: ice: This target-independent node should have been
                    selected!
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu


Seen on Ubuntu Hardy on x86 using LLVM r58399.

regehr at john-home:~/volatile/tmp54$ llvm-gcc -O3 small.c
0x8e392f4: i32 = and 0x8e391e4, 0x8e397bccc1: ScheduleDAGEmit.cpp:545: void
llvm::ScheduleDAG::EmitNode(llvm::SDNode*, bool, llvm::DenseMap<llvm::SDValue,
unsigned int, llvm::DenseMapInfo<llvm::SDValue>, llvm::DenseMapInfo<unsigned
int> >&): Assertion `0 && "This target-independent node should have been
selected!"' failed.
small.c:13: internal compiler error: Aborted
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://developer.apple.com/bugreporter> for instructions.

regehr at john-home:~/volatile/tmp54$ cat small.c

char g_3;
int g_265;
int g_279;

void foo (char p_56)
{
  rshift_u_u (p_56 * (g_279 & 1) + g_265 + g_3, 0);
}

void bar (void)
{
  foo (2);
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 30 04:08:00 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 30 Oct 2008 04:08:00 -0500
Subject: [LLVMbugs] [Bug 2983] New: Missing memory leak report in -init
 methods that do not return self.
Message-ID: <bug-2983-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2983

           Summary: Missing memory leak report in -init methods that do not
                    return self.
           Product: clang
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: devlists at shadowlab.org
                CC: llvmbugs at cs.uiuc.edu


Generaly, an initializer returns self, but in some case, it may decide to
return something else (it's a common practice to return nil when something goes
wrong)

In these cases, 'self' must be release before the method returns else
it will leaks.
(http://www.cocoabuilder.com/archive/message/cocoa/2008/2/11/198549
)

for example:

- (id)initWithString:(NSString *)str {
        if (!str) {
                [self release];
                return nil; // analyzer does not report a leak, but it should
        }
        if (self = [super init]) {
                // do something with str
        }
        return self;
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 30 11:21:32 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 30 Oct 2008 11:21:32 -0500
Subject: [LLVMbugs] [Bug 2984] New: Thumb target should use register
	scavenger
Message-ID: <bug-2984-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2984

           Summary: Thumb target should use register scavenger
           Product: libraries
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: major
          Priority: P2
         Component: Backend: ARM
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: evan.cheng at apple.com
                CC: llvmbugs at cs.uiuc.edu


Currently the Thumb target reserves r3. It is not available for general
register allocation but is only used to materialize stack offsets. See
ARMRegisterInfo.cpp.

We should teach it to use the register scavenger to get a register to
materialize stack offset instead (same as ARM target). This is a critical step
towards making it a first class target.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 30 11:28:26 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 30 Oct 2008 11:28:26 -0500
Subject: [LLVMbugs] [Bug 2985] New: Invalid code generated for cpu2000/gcc
	with -mthumb
Message-ID: <bug-2985-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2985

           Summary: Invalid code generated for cpu2000/gcc with -mthumb
           Product: libraries
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: major
          Priority: P2
         Component: Backend: ARM
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: evan.cheng at apple.com
                CC: llvmbugs at cs.uiuc.edu


llvm generates illegal assembly in Thumb:
sub r5,sp,r2

The issue is the Thumb backend is treating all integer registers in the same
register class. The register coalescer doesn't know it cannot coalesce a copy
from sp into a use.

I have attached a reduced test case. Just run llc on it to see the bad
assembly.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 30 12:55:32 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 30 Oct 2008 12:55:32 -0500
Subject: [LLVMbugs] [Bug 2986] New: Crash in llc with asm (regression)
Message-ID: <bug-2986-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2986

           Summary: Crash in llc with asm (regression)
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: dalej at apple.com
                CC: llvmbugs at cs.uiuc.edu


llc < bugpoint-reduced-simplified.bc
PromoteIntegerOperand Op #1: 0x18084d4: i32 = shl 0x180822c, 0x180844c
Assertion failed: (0 && "Do not know how to promote this operator's operand!"),
function PromoteIntegerOperand, file LegalizeIntegerTypes.cpp, line 584.

This is reduced from a case in the gcc testsuite (ppc32 Darwin).  Apparently
came in with LegalizeTypes.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 30 13:23:59 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 30 Oct 2008 13:23:59 -0500
Subject: [LLVMbugs] [Bug 2987] New: Crash from llvm-types
Message-ID: <bug-2987-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2987

           Summary: Crash from llvm-types
           Product: libraries
           Version: trunk
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Support Libraries
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: dalej at apple.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2165)
 --> (http://llvm.org/bugs/attachment.cgi?id=2165)
.i file

This happens on llvm-gcc on x86 darwin (or ppc32).  Was introduced by 58387.

now what? $ld/gcc/llvm-gcc-4.2.SL.obj/gcc/xgcc -B
$ld/gcc/llvm-gcc-4.2.SL.obj/gcc/ t024_x.i -w
Assertion failed: ((isBitfield(Field) || FieldNo == ~0U || FieldOffsetInBits ==
8*Info->ElementOffsetInBytes[FieldNo]) && "Wrong LLVM field offset!"), function
ConvertRECORD, file ../../llvm-gcc-4.2/gcc/llvm-types.cpp, line 2150.
In file included from
/Volumes/MacOS9/gcc/llvm-gcc-4.2.SL.obj/gcc/testsuite/gcc/gcc.dg-struct-layout-1/t024_x.c:6:
/Volumes/MacOS9/gcc/llvm-gcc-4.2.SL.obj/gcc/testsuite/gcc/gcc.dg-struct-layout-1/t024_test.h:
In function ???test2045???:
/Volumes/MacOS9/gcc/llvm-gcc-4.2.SL.obj/gcc/testsuite/gcc/gcc.dg-struct-layout-1/t024_test.h:46:
internal compiler error: Abort trap

Test from gcc testsuite.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 30 14:52:18 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 30 Oct 2008 14:52:18 -0500
Subject: [LLVMbugs] [Bug 2976] Mips backend problem
In-Reply-To: <bug-2976-206@http.llvm.org/bugs/>
Message-ID: <200810301952.m9UJqI5x028490@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2976


Bruno Cardoso Lopes <bruno.cardoso at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #1 from Bruno Cardoso Lopes <bruno.cardoso at gmail.com>  2008-10-30 14:52:17 ---
Hi, this bug is fixed in trunk!

Thanks,

*** This bug has been marked as a duplicate of bug 2543 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 30 15:36:28 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 30 Oct 2008 15:36:28 -0500
Subject: [LLVMbugs] [Bug 2986] Crash in llc with asm (regression)
In-Reply-To: <bug-2986-206@http.llvm.org/bugs/>
Message-ID: <200810302036.m9UKaS1b030093@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2986


Duncan Sands <baldrick at free.fr> changed:

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




--- Comment #2 from Duncan Sands <baldrick at free.fr>  2008-10-30 15:36:22 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081027/069146.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 30 16:14:20 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 30 Oct 2008 16:14:20 -0500
Subject: [LLVMbugs] [Bug 2987] Crash from llvm-types
In-Reply-To: <bug-2987-206@http.llvm.org/bugs/>
Message-ID: <200810302114.m9ULEKrm021246@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2987


Duncan Sands <baldrick at free.fr> changed:

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




--- Comment #4 from Duncan Sands <baldrick at free.fr>  2008-10-30 16:14:14 ---
This was an old bug caught by recently added sanity checking.
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081027/069152.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 30 18:03:25 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 30 Oct 2008 18:03:25 -0500
Subject: [LLVMbugs] [Bug 2988] New: ppcf128 intrinsics broken by
	LegalizeTypes
Message-ID: <bug-2988-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2988

           Summary: ppcf128 intrinsics broken by LegalizeTypes
           Product: new-bugs
           Version: unspecified
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: dalej at apple.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2166)
 --> (http://llvm.org/bugs/attachment.cgi?id=2166)
reduction from gcc testsuite

now what? llc < yy.bc
ExpandFloatResult #0: 0x180842c: ppcf128 = fsqrt 0x18083a4
Assertion failed: (0 && "Do not know how to expand the result of this
operator!"), function ExpandFloatResult, file LegalizeFloatTypes.cpp, line 566.

There are other ppcf128 problems as well.  Do you have a ppc machine?  I'll
isolate them if not.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 30 19:08:26 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 30 Oct 2008 19:08:26 -0500
Subject: [LLVMbugs] [Bug 2888] linux regressions with 2.4
In-Reply-To: <bug-2888-206@http.llvm.org/bugs/>
Message-ID: <200810310008.m9V08QTA019431@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2888


Tanya Lattner <tonic at nondot.org> changed:

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




--- Comment #24 from Tanya Lattner <tonic at nondot.org>  2008-10-30 19:08:25 ---
External/SPEC/CINT95/132.ijpeg/132.ijpeg  (JIT) is still failing.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 30 20:06:24 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 30 Oct 2008 20:06:24 -0500
Subject: [LLVMbugs] [Bug 2888] linux regressions with 2.4
In-Reply-To: <bug-2888-206@http.llvm.org/bugs/>
Message-ID: <200810310106.m9V16O2a030800@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2888


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #25 from Chris Lattner <clattner at apple.com>  2008-10-30 20:06:23 ---
The 132.ijpeg JIT failure is caused by it running out of memory.  This is a
known issue that we're both planning on fixing, and shouldn't block 2.4. 
Thanks!


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 30 20:25:11 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 30 Oct 2008 20:25:11 -0500
Subject: [LLVMbugs] [Bug 2647] extractps selected too eagerly
In-Reply-To: <bug-2647-206@http.llvm.org/bugs/>
Message-ID: <200810310125.m9V1PBYL004394@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2647


Dan Gohman <gohman at apple.com> changed:

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




--- Comment #9 from Dan Gohman <gohman at apple.com>  2008-10-30 20:25:10 ---
With this commit:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081027/069165.html
codegen avoids selecting extractps for a store of element 0, and uses movss
instead.
The provided testcase now compiles to three instructions even with
-mcpu=penryn.

The extra setup is a known issue.  The code that decides if stack-pointer
alignment
is necessary is over-conservative in some cases. See
lib/Target/X86/README-SSE.txt,
the entry starting with "LLVM currently generates stack realignment code".

Thanks!


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 30 22:32:04 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 30 Oct 2008 22:32:04 -0500
Subject: [LLVMbugs] [Bug 2989] New: weak attribute improper handled when
 return type is pointer type
Message-ID: <bug-2989-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2989

           Summary: weak attribute improper handled when return type is
                    pointer type
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: yliang6 at illinois.edu
                CC: llvmbugs at cs.uiuc.edu


int * __attribute__((weak)) foo(int * x) {
    return x;
}

llvm-gcc warns ``'weak' attribute ignored'', and llvm-nm shows "T foo".


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Oct 30 23:05:39 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 30 Oct 2008 23:05:39 -0500
Subject: [LLVMbugs] [Bug 2950] clang -g crashing
In-Reply-To: <bug-2950-206@http.llvm.org/bugs/>
Message-ID: <200810310405.m9V45d5h006443@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2950


Daniel Dunbar <daniel at zuster.org> changed:

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




--- Comment #4 from Daniel Dunbar <daniel at zuster.org>  2008-10-30 23:05:38 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20081027/008763.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 31 03:08:31 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 31 Oct 2008 03:08:31 -0500
Subject: [LLVMbugs] [Bug 2990] New: Requiring a ModulePass and
 InstructionNamerID causes PassManager to infloop
Message-ID: <bug-2990-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2990

           Summary: Requiring a ModulePass and InstructionNamerID causes
                    PassManager to infloop
           Product: new-bugs
           Version: unspecified
          Platform: PC
               URL: http://lists.cs.uiuc.edu/pipermail/llvmdev/2008-
                    October/018091.html
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


To reproduce:
* using LLVM TOT, unarchive the attached Hello.zip in lib/Transforms/Hello
* run make
* ../../../Release/bin/opt -load ../../../Release/lib/LLVMHello.so -tstfunc
foo.bc

foo.bc can be any bitcode

If I make the modulepass be a functionpass it no longer infloops, or if I
remove the InstructionNamerID from the requierements.

Stacktrace below:

Program received signal SIGINT, Interrupt.
[Switching to Thread 0x7f358a99e6f0 (LWP 19611)]
0x0000000000a3a113 in std::map<llvm::PassInfo const*, llvm::Pass*,
std::less<llvm::PassInfo const*>, std::al
locator<std::pair<llvm::PassInfo const* const, llvm::Pass*> > >::find
(this=0x1dfcdf0, __x=@0x7fff929c5760)
at /usr/include/c++/4.3/bits/stl_map.h:615
(gdb) bt
#0  0x0000000000a3a113 in std::map<llvm::PassInfo const*, llvm::Pass*,
std::less<llvm::PassInfo const*>, std
::allocator<std::pair<llvm::PassInfo const* const, llvm::Pass*> > >::find
(this=0x1dfcdf0, __x=@0x7fff929c57
60) at /usr/include/c++/4.3/bits/stl_map.h:615
#1  0x0000000000a34fa3 in llvm::PMDataManager::findAnalysisPass
(this=0x1dfcd10, AID=0xf2e220, SearchParent=
false) at PassManager.cpp:936
#2  0x0000000000a34e74 in llvm::PMTopLevelManager::findAnalysisPass
(this=0x19a4900, AID=0xf2e220) at PassMa
nager.cpp:517
#3  0x0000000000a353bd in llvm::PMTopLevelManager::schedulePass
(this=0x19a4900, P=0x19a70d0) at PassManager
.cpp:474
warning: (Internal error: pc 0xa4021d in read in psymtab, but not in symtab.)
warning: (Internal error: pc 0xa4021d in read in psymtab, but not in symtab.)

#4  0x0000000000a4021e in llvm::PassManagerImpl::add (this=warning: (Internal
error: pc 0xa4021d in read in
psymtab, but not in symtab.)

warning: (Internal error: pc 0xa401fa in read in psymtab, but not in symtab.)

warning: (Internal error: pc 0xa4021d in read in psymtab, but not in symtab.)

0x19a4750, P=warning: (Internal error: pc 0xa4021d in read in psymtab, but not
in symtab.)

0x19a70d0) at PassManager.cpp:280
#5  0x0000000000a354dd in llvm::PassManager::add (this=0x7fff929c5a40,
P=0x19a70d0) at PassManager.cpp:1490
#6  0x0000000000738c6a in addPass (PM=@0x7fff929c5a40, P=0x19a70d0) at
opt.cpp:258
#7  0x000000000073a61f in main (argc=5, argv=0x7fff929c5dd8) at opt.cpp:532


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 31 05:17:35 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 31 Oct 2008 05:17:35 -0500
Subject: [LLVMbugs] [Bug 2992] New: Assertion failed: (Ty && "GEP indices
 invalid!"), function getGetElementPtr, file Constants.cpp, line 2147.
Message-ID: <bug-2992-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2992

           Summary: Assertion failed: (Ty && "GEP indices invalid!"),
                    function getGetElementPtr, file Constants.cpp, line
                    2147.
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: rdivacky at freebsd.org
                CC: llvmbugs at cs.uiuc.edu


when compiling the attached with clang I am getting this:

Assertion failed: (Ty && "GEP indices invalid!"), function getGetElementPtr,
file Constants.cpp, line 2147.

strange thing is that 

clang exdump.i works, while ccc exdump.i crashes on this assert


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 31 06:13:34 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 31 Oct 2008 06:13:34 -0500
Subject: [LLVMbugs] [Bug 2993] New: instcombine causes miscompilation with
	rems -1
Message-ID: <bug-2993-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2993

           Summary: instcombine causes miscompilation with rems -1
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: richard at xmos.com
                CC: llvmbugs at cs.uiuc.edu


Running "opt -instcombine" on the following testcase:

define i1 @main_bb_bb_2E_ce(i32 %i.0.reg2mem.0) {
newFuncRoot:
        br label %bb.ce

bb2.exitStub:           ; preds = %bb.ce
        ret i1 true

codeRepl1.exitStub:             ; preds = %bb.ce
        ret i1 false

bb.ce:          ; preds = %newFuncRoot
        %0 = srem i32 %i.0.reg2mem.0, -1                ; <i32> [#uses=1]
        %1 = icmp eq i32 %0, 0          ; <i1> [#uses=1]
        br i1 %1, label %bb2.exitStub, label %codeRepl1.exitStub
}

turns it into:

define i1 @main_bb_bb_2E_ce(i32 %i.0.reg2mem.0) {
newFuncRoot:
        br label %bb.ce

bb2.exitStub:           ; preds = %bb.ce
        ret i1 true

codeRepl1.exitStub:             ; preds = %bb.ce
        ret i1 false

bb.ce:          ; preds = %newFuncRoot
        %0 = icmp eq i32 %i.0.reg2mem.0, 0              ; <i1> [#uses=1]
        br i1 %0, label %bb2.exitStub, label %codeRepl1.exitStub
}

This transformation is incorrect. Instruction Combine should not be turning

        %0 = srem i32 %i.0.reg2mem.0, -1                ; <i32> [#uses=1]
        %1 = icmp eq i32 %0, 0          ; <i1> [#uses=1]
into 
        %0 = icmp eq i32 %i.0.reg2mem.0, 0              ; <i1> [#uses=1]


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 31 08:03:37 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 31 Oct 2008 08:03:37 -0500
Subject: [LLVMbugs] [Bug 2994] New: assertion in indvars pass
Message-ID: <bug-2994-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2994

           Summary: assertion in indvars pass
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: richard at xmos.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2170)
 --> (http://llvm.org/bugs/attachment.cgi?id=2170)
testcase

Running opt bugpoint-reduced-simplified.bc -indvars with the attached .bc file
causes the following assertion:

opt: /homelocal/richard/8_11/tools_llvm/src/lib/Support/APInt.cpp:1718:
llvm::APInt llvm::APInt::udiv(const llvm::APInt&) const: Assertion `RHS.VAL !=
0 && "Divide by zero?"' failed.
opt[0xac07da]
opt[0xac08f1]
/lib64/tls/libc.so.6[0x356b22e2f0]
/lib64/tls/libc.so.6(gsignal+0x3d)[0x356b22e25d]
/lib64/tls/libc.so.6(abort+0xfe)[0x356b22fa5e]
/lib64/tls/libc.so.6(__assert_fail+0xf1)[0x356b227ae1]
opt(_ZNK4llvm5APInt4udivERKS0_+0x6c)[0xa96f80]
opt(_ZNK4llvm5APInt4sdivERKS0_+0x11a)[0x89c02c]
opt[0x9b1f98]
opt(_ZNK4llvm14SCEVAddRecExpr23getNumIterationsInRangeENS_13ConstantRangeERNS_15ScalarEvolutionE+0x636)[0x9b322c]
opt(_ZNK4llvm14SCEVAddRecExpr23getNumIterationsInRangeENS_13ConstantRangeERNS_15ScalarEvolutionE+0x1a5)[0x9b2d9b]
opt(_ZN40_GLOBAL__N__ZN4llvm15ScalarEvolution2IDE20ScalarEvolutionsImpl21ComputeIterationCountEPKN4llvm8LoopBaseINS1_10BasicBlockEEE+0x6c5)[0x9af7d1]
opt(_ZN40_GLOBAL__N__ZN4llvm15ScalarEvolution2IDE20ScalarEvolutionsImpl17getIterationCountEPKN4llvm8LoopBaseINS1_10BasicBlockEEE+0x67)[0x9aefcf]
opt(_ZNK4llvm15ScalarEvolution17getIterationCountEPKNS_8LoopBaseINS_10BasicBlockEEE+0x28)[0x9b3832]
opt(_ZN108_GLOBAL__N__homelocal_richard_8_11_tools_llvm_src_lib_Transforms_Scalar_IndVarSimplify.cpp_DD76033A_235439B414IndVarSimplify9runOnLoopEPN4llvm8LoopBaseINS1_10BasicBlockEEERNS1_13LPPassManagerE+0x9b)[0x85ebe7]
opt(_ZN4llvm13LPPassManager13runOnFunctionERNS_8FunctionE+0x38e)[0x9942c6]
opt(_ZN4llvm13FPPassManager13runOnFunctionERNS_8FunctionE+0x111)[0xa559a9]
opt(_ZN4llvm13FPPassManager11runOnModuleERNS_6ModuleE+0x61)[0xa55af5]
opt(_ZN4llvm13MPPassManager11runOnModuleERNS_6ModuleE+0xd8)[0xa55c00]
opt(_ZN4llvm15PassManagerImpl3runERNS_6ModuleE+0x94)[0xa55f98]
opt(_ZN4llvm11PassManager3runERNS_6ModuleE+0x21)[0xa56249]
opt(main+0xf29)[0x7c6591]
/lib64/tls/libc.so.6(__libc_start_main+0xdb)[0x356b21c3fb]
opt(fmod+0xb2)[0x7b7e3a]
Aborted


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 31 08:58:06 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 31 Oct 2008 08:58:06 -0500
Subject: [LLVMbugs] [Bug 2995] New: gcc/config/i386//darwin.h should not set
 -Wformat, -Wformat-security, -Wno-format-extra-args,
 -Wno-format-zero-length and -Wno-nonnull
Message-ID: <bug-2995-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2995

           Summary: gcc/config/i386//darwin.h should not set -Wformat, -
                    Wformat-security, -Wno-format-extra-args, -Wno-format-
                    zero-length and -Wno-nonnull
           Product: new-bugs
           Version: unspecified
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: howarth at bromo.msbb.uc.edu
                CC: llvmbugs at cs.uiuc.edu


Currently in llvm-gcc 4.2 from llvm 2.4 and llvm svn, the
gcc/config/i386/darwin.h header is inappropriately setting CC1_SPEC to always
pass...

-Wformat -Wformat-security -Wno-format-extra-args -Wno-format-zero-length
-Wno-nonnull 

This causes the llvm-gfortran compiler to spew warning messages of the form...

f951: warning: command line option "-Wformat" is valid for C/C++/ObjC/ObjC++
but not for Fortran
f951: warning: command line option "-Wformat-security" is valid for
C/C++/ObjC/ObjC++ but not for Fortran
f951: warning: command line option "-Wno-format-extra-args" is valid for
C/C++/ObjC/ObjC++ but not for Fortran
f951: warning: command line option "-Wno-format-zero-length" is valid for
C/C++/ObjC/ObjC++ but not for Fortran
f951: warning: command line option "-Wno-nonnull" is valid for
C/C++/ObjC/ObjC++ but not for Fortran

when any fortran source is compiled. These should either be removed from the
definition of CC1_SPECS or recrafted to not be used in the case of compilers
like gfortran that don't grok them.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 31 09:08:29 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 31 Oct 2008 09:08:29 -0500
Subject: [LLVMbugs] [Bug 2988] ppcf128 intrinsics broken by LegalizeTypes
In-Reply-To: <bug-2988-206@http.llvm.org/bugs/>
Message-ID: <200810311408.m9VE8Tem003845@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2988


Duncan Sands <baldrick at free.fr> changed:

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




--- Comment #2 from Duncan Sands <baldrick at free.fr>  2008-10-31 09:08:28 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081027/069182.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 31 11:58:38 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 31 Oct 2008 11:58:38 -0500
Subject: [LLVMbugs] [Bug 2996] New: Assertion `New->getType() == getType()
 && " replaceAllUses of value with new value of different type!"' failed.
Message-ID: <bug-2996-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2996

           Summary: Assertion `New->getType() == getType() &&
                    "replaceAllUses of value with new value of different
                    type!"' failed.
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu


Seen on Ubuntu Hardy on x86 using LLVM r58509.

[regehr at babel tmp17]$ llvm-gcc -O1 small.c
cc1: Value.cpp:319: void llvm::Value::replaceAllUsesWith(llvm::Value*):
Assertion `New->getType() == getType() && "replaceAllUses of value with new
value of different type!"' failed.
small.c:12: internal compiler error: Aborted
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://developer.apple.com/bugreporter> for instructions.

[regehr at babel tmp17]$ cat small.c

int func_77 (char);
int func_38 (int);

int func_24 (void)
{
  return -1;
}

void func_100 (void)
{
  func_77 ((func_38 (1) > func_24 ()) * -1);
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 31 14:25:11 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 31 Oct 2008 14:25:11 -0500
Subject: [LLVMbugs] [Bug 2616] [CellSPU] Code generation failure
In-Reply-To: <bug-2616-206@http.llvm.org/bugs/>
Message-ID: <200810311925.m9VJPBL4015688@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2616


Scott Michel <scooter.phd at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|LATER                       |




-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 31 15:14:03 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 31 Oct 2008 15:14:03 -0500
Subject: [LLVMbugs] [Bug 2613] Debugging labels from virtual destructors are
 not being invalidated even without optimization
In-Reply-To: <bug-2613-206@http.llvm.org/bugs/>
Message-ID: <200810312014.m9VKE3YG017443@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2613


Anton Korobeynikov <asl at math.spbu.ru> changed:

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




--- Comment #12 from Anton Korobeynikov <asl at math.spbu.ru>  2008-10-31 15:14:01 ---
(In reply to comment #10)
> The problem is now, that there is no housekeeping to invalidate also the debug
> informations in the MachineModuleInfo.

Patch applied at:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081027/069215.html

Testcase at:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081027/069216.html

Actually, we need to invalidate all labels (EH, GC), not only debug ones. Thus
I slightly modified your patch.

I consider debug information & optimization as separate issue and close this
PR. Feel free to file new PR for other problems.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 31 15:58:12 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 31 Oct 2008 15:58:12 -0500
Subject: [LLVMbugs] [Bug 2616] [CellSPU] Code generation failure
In-Reply-To: <bug-2616-206@http.llvm.org/bugs/>
Message-ID: <200810312058.m9VKwCbW019110@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2616


Scott Michel <scooter.phd at gmail.com> changed:

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




--- Comment #4 from Scott Michel <scooter.phd at gmail.com>  2008-10-31 15:58:10 ---
I just tested this with "llc -march=cellspu -o - zzz.bc" (where zzz.ll has the
example code in the initial bug report, which was then llvm-as-ed to zzz.bc).

Other than the ellipsis that indicates a vararg function, llc accepts it just
fine and generates code w/o an assert.

Heisenbug caused by something else other than the backend?

If you take out the ellipsis, i.e.:

; ModuleID = 't0038.ubc'

define i32 @main() {
entry:
        %retval = alloca i32            ; <i32*> [#uses=2]
        store i32 0, i32* %retval
        %i = alloca i32         ; <i32*> [#uses=0]
        br label %return

return:         ; preds = %entry
        load i32* %retval               ; <i32>:0 [#uses=1]
        ret i32 %0
}

llc produces the following (it isn't the greatest code produced, but it is
assert-free):

        .text


        .align  3
        .global main
        .type   main, @function
main:
        stqd    $lr, 16($sp)
        stqd    $sp, -64($sp)
        ai      $sp, $sp, -64
        lqd     $3, 0($sp)
        cwd     $4, 0($sp)
        il      $5, 0
        shufb   $3, $5, $3, $4
        stqd    $3, 0($sp)
.LBB1_1:        # return
        lqd     $3, 0($sp)
        rotqby  $3, $3, $sp
        lqd     $lr, 80($sp)
        ai      $sp, $sp, 64
        bi      $lr
        .size   main,.-main

        .ident  "(llvm 2.2+) STI CBEA Cell SPU backend"

Note: This is similar to what would be produced by llc with the ellipsis, but
without the spillage of 48 registers to the stack:

        .text


        .align  3
        .global main
        .type   main, @function
main:
        stqd    $lr, 16($sp)
        stqd    $sp, -64($sp)
        ai      $sp, $sp, -64
        stqd    $79, 1312($sp)
        stqd    $78, 1296($sp)
        stqd    $77, 1280($sp)
        stqd    $76, 1264($sp)
        stqd    $75, 1248($sp)
        stqd    $74, 1232($sp)
        stqd    $73, 1216($sp)
        stqd    $72, 1200($sp)
        stqd    $71, 1184($sp)
        stqd    $70, 1168($sp)
        stqd    $69, 1152($sp)
        stqd    $68, 1136($sp)
        stqd    $67, 1120($sp)
        stqd    $66, 1104($sp)
        stqd    $65, 1088($sp)
        stqd    $64, 1072($sp)
        stqd    $63, 1056($sp)
        stqd    $62, 1040($sp)
        stqd    $61, 1024($sp)
        stqd    $60, 1008($sp)
        stqd    $59, 992($sp)
        stqd    $58, 976($sp)
        stqd    $57, 960($sp)
        stqd    $56, 944($sp)
        stqd    $55, 928($sp)
        stqd    $54, 912($sp)
        stqd    $53, 896($sp)
        stqd    $52, 880($sp)
        stqd    $51, 864($sp)
        stqd    $50, 848($sp)
        stqd    $49, 832($sp)
        stqd    $48, 816($sp)
        stqd    $47, 800($sp)
        stqd    $46, 784($sp)
        stqd    $45, 768($sp)
        stqd    $44, 752($sp)
        stqd    $43, 736($sp)
        stqd    $42, 720($sp)
        stqd    $41, 704($sp)
        stqd    $40, 688($sp)
        stqd    $39, 672($sp)
        stqd    $38, 656($sp)
        stqd    $37, 640($sp)
        stqd    $36, 624($sp)
        stqd    $35, 608($sp)
        stqd    $34, 592($sp)
        stqd    $33, 576($sp)
        stqd    $32, 560($sp)
        stqd    $31, 544($sp)
        stqd    $30, 528($sp)
        stqd    $29, 512($sp)
        stqd    $28, 496($sp)
        stqd    $27, 480($sp)
        stqd    $26, 464($sp)
        stqd    $25, 448($sp)
        stqd    $24, 432($sp)
        stqd    $23, 416($sp)
        stqd    $22, 400($sp)
        stqd    $21, 384($sp)
        stqd    $20, 368($sp)
        stqd    $19, 352($sp)
        stqd    $18, 336($sp)
        stqd    $17, 320($sp)
        stqd    $16, 304($sp)
        stqd    $15, 288($sp)
        stqd    $14, 272($sp)
        stqd    $13, 256($sp)
        stqd    $12, 240($sp)
        stqd    $11, 224($sp)
        stqd    $10, 208($sp)
        stqd    $9, 192($sp)
        stqd    $8, 176($sp)
        stqd    $7, 160($sp)
        stqd    $6, 144($sp)
        stqd    $5, 128($sp)
        stqd    $4, 112($sp)
        stqd    $3, 96($sp)
        lqd     $3, 0($sp)
        cwd     $4, 0($sp)
        il      $5, 0
        shufb   $3, $5, $3, $4
        stqd    $3, 0($sp)
.LBB1_1:        # return
        lqd     $3, 0($sp)
        rotqby  $3, $3, $sp
        lqd     $lr, 80($sp)
        ai      $sp, $sp, 64
        bi      $lr
        .size   main,.-main

        .ident  "(llvm 2.2+) STI CBEA Cell SPU backend"


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 31 19:52:39 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 31 Oct 2008 19:52:39 -0500
Subject: [LLVMbugs] [Bug 2998] New: llvm-g++ EH code doesn't work with GCC
 4.3 libgcc_s.so.1
Message-ID: <bug-2998-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2998

           Summary: llvm-g++ EH code doesn't work with GCC 4.3 libgcc_s.so.1
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: gohman at apple.com
                CC: llvmbugs at cs.uiuc.edu


Compiling this test from the test-suite:
SingleSource/Regression/C++/EH/ctor_dtor_count.cpp

with llvm-g++ on an Ubuntu "Intrepid" box, where /lib/libgcc_s.so.1 is
built from gcc-4.3 sources, results in an executable that hangs when
executed.

I dug up a libgcc_s.so.1 from an older package compiled from gcc-4.2 sources,
and ran the same ctor_dtor_count executable without recompiling, forcing the
older library to be used using LD_PRELOAD, and it executed successfully.

Compiling ctor_dtor_count with regular g++-4.2 and running the executable with
the gcc-4.3 libgcc_s.so.1 does work.

I'm seeing this in both 32-bit and 64-bit executables.

Valgrind on the test executable found nothing interesting.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 31 20:35:36 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 31 Oct 2008 20:35:36 -0500
Subject: [LLVMbugs] [Bug 2999] New: happy halloween...
Message-ID: <bug-2999-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2999

           Summary: happy halloween...
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu


This testcase, unfortunately, is kind of ugly and scary.  It appears to be
miscompiled by LLVM r58541 on Ubuntu Hardy on x86.  This is a regression since
2.3.

regehr at john-home:~/volatile/tmp55$ llvm-gcc -O0 small.c -o small
regehr at john-home:~/volatile/tmp55$ ./small
0
regehr at john-home:~/volatile/tmp55$ llvm-gcc -O1 small.c -o small
regehr at john-home:~/volatile/tmp55$ ./small
1


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 31 21:09:21 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 31 Oct 2008 21:09:21 -0500
Subject: [LLVMbugs] [Bug 3000] New: clang: used type 'union u' where
 arithmetic or pointer type is required
Message-ID: <bug-3000-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3000

           Summary: clang: used type 'union u' where arithmetic or pointer
                    type is required
           Product: clang
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: richard at godbee.net
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 2778


union u { int a, b; };

void f2(union u);

void f1() {
    f2((union u) 0);
}


<stdin>:6:8: error: used type 'union u' where arithmetic or pointer type is
required
    f2((union u) 0);
       ^         ~
1 diagnostic generated.


llvm/clang r58543 (trunk)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Oct 31 21:28:32 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 31 Oct 2008 21:28:32 -0500
Subject: [LLVMbugs] [Bug 3001] New: clang: Assertion failed: ... "not sorted
 by InitIndex!", function CreateDesignation,
 include/clang/Parse/Designator.h:206
Message-ID: <bug-3001-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3001

           Summary: clang: Assertion failed: ... "not sorted by InitIndex!",
                    function CreateDesignation,
                    include/clang/Parse/Designator.h:206
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Keywords: crash-on-invalid
          Severity: normal
          Priority: P2
         Component: parser
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: richard at godbee.net
                CC: llvmbugs at cs.uiuc.edu


struct s1 s2 = {
    .a = sizeof(struct s3),
    .b = c
}


<stdin>:2:10: error: invalid application of 'sizeof' to an incomplete type
'struct s3'
    .a = sizeof(struct s3),
         ^     ~~~~~~~~~~~
Assertion failed: ((Designations.empty() || Designations.back().InitIndex <
Idx) && "not sorted by InitIndex!"), function CreateDesignation, file
/Users/rwg/llvm/svn-src/llvm-trunk/tools/clang/lib/Parse/../../include/clang/Parse/Designator.h,
line 206.


llvm/clang r58543 (trunk)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.