From bugzilla-daemon at llvm.org Sat Jan 1 05:00:37 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 1 Jan 2011 05:00:37 -0600 (CST)
Subject: [LLVMbugs] [Bug 8879] New: -Wtype-limits not implemented
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8879
Summary: -Wtype-limits not implemented
Product: clang
Version: trunk
Platform: PC
OS/Version: Linux
Status: NEW
Severity: enhancement
Priority: P
Component: Frontend
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: tss at iki.fi
CC: llvmbugs at cs.uiuc.edu
No warnings are given when comparing a type against impossible values.
Example:
int main(void) {
unsigned short a = 0;
if (a < 0) return 1;
if (a > 65535) return 1;
if (a == -1) return 1;
return 0;
}
gcc warns:
test2.c:3: warning: comparison is always false due to limited range of data
type
test2.c:4: warning: comparison is always false due to limited range of data
type
test2.c:5: warning: comparison is always false due to limited range of data
type
clang doesn't warn anything.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sat Jan 1 13:07:38 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 1 Jan 2011 13:07:38 -0600 (CST)
Subject: [LLVMbugs] [Bug 8878] [MC] regression: unknown fixup kind in ELF
In-Reply-To:
References:
Message-ID: <20110101190738.2F2582A6C12E@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8878
Rafael ?vila de Esp?ndola changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
--- Comment #2 from Rafael ?vila de Esp?ndola 2011-01-01 13:07:37 CST ---
Fixed in 122658.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sat Jan 1 13:55:42 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 1 Jan 2011 13:55:42 -0600 (CST)
Subject: [LLVMbugs] [Bug 8880] New: clang dies processing a for loop
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8880
Summary: clang dies processing a for loop
Product: clang
Version: trunk
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: Frontend
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: bagnara at cs.unipr.it
CC: llvmbugs at cs.uiuc.edu
$ cat f3.c
int f3() {
int first = 1;
for ( ; ({ if (first) { first = 0; continue; } 0; }); )
return 0;
return 1;
}
$ ~/llvm.debug/Debug+Asserts/bin/clang /tmp/f3.c
0 clang 0x00000000020d461a
1 clang 0x00000000020d440d
2 libpthread.so.0 0x0000003dcd80f3c0
3 libc.so.6 0x0000003dccd34245
4 clang 0x00000000016a5d53
clang::BumpVector::grow(clang::BumpVectorContext&, unsigned
long) + 165
5 clang 0x00000000016a419f
clang::BumpVector::push_back(clang::CFGBlock* const&,
clang::BumpVectorContext&) + 127
6 clang 0x00000000016a361e
clang::CFGBlock::addSuccessor(clang::CFGBlock*, clang::BumpVectorContext&) + 64
7 clang 0x000000000169517e
8 clang 0x000000000169a8db
9 clang 0x00000000016968c6
10 clang 0x0000000001694fe1
11 clang 0x0000000001697be1
12 clang 0x000000000169687d
13 clang 0x0000000001694fe1
14 clang 0x0000000001698725
15 clang 0x0000000001696ae5
16 clang 0x0000000001694fe1
17 clang 0x0000000001697be1
18 clang 0x000000000169aa6f
19 clang 0x0000000001696cbf
20 clang 0x0000000001694fe1
21 clang 0x0000000001698fe9
22 clang 0x0000000001696a9f
23 clang 0x0000000001694fe1
24 clang 0x0000000001697be1
25 clang 0x000000000169687d
26 clang 0x0000000001694fe1
27 clang 0x0000000001695568
28 clang 0x000000000169c4c7 clang::CFG::buildCFG(clang::Decl const*,
clang::Stmt*, clang::ASTContext*, clang::CFG::BuildOptions) + 103
29 clang 0x000000000168e6a1 clang::AnalysisContext::getCFG() + 233
30 clang 0x0000000001555a0c
31 clang 0x0000000001556535
32 clang 0x0000000001556a0a
clang::sema::AnalysisBasedWarnings::IssueWarnings(clang::sema::AnalysisBasedWarnings::Policy,
clang::Decl const*, clang::QualType) + 558
33 clang 0x0000000001556b21
clang::sema::AnalysisBasedWarnings::IssueWarnings(clang::sema::AnalysisBasedWarnings::Policy,
clang::FunctionDecl const*) + 57
34 clang 0x0000000001385bca
clang::Sema::ActOnFinishFunctionBody(clang::Decl*, clang::Stmt*, bool) + 2164
35 clang 0x000000000138534e
clang::Sema::ActOnFinishFunctionBody(clang::Decl*, clang::Stmt*) + 60
36 clang 0x00000000012da581
clang::Parser::ParseFunctionStatementBody(clang::Decl*) + 271
37 clang 0x00000000012e04d9
clang::Parser::ParseFunctionDefinition(clang::Parser::ParsingDeclarator&,
clang::Parser::ParsedTemplateInfo const&) + 1001
38 clang 0x00000000012e6458
clang::Parser::ParseDeclGroup(clang::Parser::ParsingDeclSpec&, unsigned int,
bool, clang::SourceLocation*) + 522
39 clang 0x00000000012e0065
clang::Parser::ParseDeclarationOrFunctionDefinition(clang::Parser::ParsingDeclSpec&,
clang::AttributeList*, clang::AccessSpecifier) + 1079
40 clang 0x00000000012e00d1
clang::Parser::ParseDeclarationOrFunctionDefinition(clang::AttributeList*,
clang::AccessSpecifier) + 95
41 clang 0x00000000012dfa2c
clang::Parser::ParseExternalDeclaration(clang::CXX0XAttributeList,
clang::Parser::ParsingDeclSpec*) + 2364
42 clang 0x00000000012df061
clang::Parser::ParseTopLevelDecl(clang::OpaquePtr&) + 343
43 clang 0x00000000012c9468 clang::ParseAST(clang::Sema&, bool) + 328
44 clang 0x000000000103d3ed clang::ASTFrontendAction::ExecuteAction()
+ 263
45 clang 0x00000000011675af clang::CodeGenAction::ExecuteAction() +
969
46 clang 0x000000000103d040 clang::FrontendAction::Execute() + 320
47 clang 0x0000000001025c11
clang::CompilerInstance::ExecuteAction(clang::FrontendAction&) + 745
48 clang 0x0000000000fd52f5
clang::ExecuteCompilerInvocation(clang::CompilerInstance*) + 878
49 clang 0x0000000000fc763f cc1_main(char const**, char const**, char
const*, void*) + 990
50 clang 0x0000000000fd0890 main + 484
51 libc.so.6 0x0000003dccc1ee7d __libc_start_main + 253
52 clang 0x0000000000fc6cd9
Stack dump:
0. Program arguments: /home/roberto/llvm.debug/Debug+Asserts/bin/clang -cc1
-triple x86_64-unknown-linux-gnu -emit-obj -mrelax-all -disable-free
-main-file-name f3.c -mrelocation-model static -mdisable-fp-elim -masm-verbose
-mconstructor-aliases -munwind-tables -target-cpu x86-64 -target-linker-version
2.20.51.0.7 -resource-dir
/home/roberto/llvm.debug/Debug+Asserts/bin/../lib/clang/2.9 -ferror-limit 19
-fmessage-length 80 -fgnu-runtime -fdiagnostics-show-option -fcolor-diagnostics
-o /tmp/cc-91LqrB.o -x c /tmp/f3.c
1. parser at end of file
2. /tmp/f3.c:1:10: parsing function body 'f3'
clang: error: unable to execute command: Segmentation fault
clang: error: clang frontend command failed due to signal 1 (use -v to see
invocation)
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sat Jan 1 14:38:58 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 1 Jan 2011 14:38:58 -0600 (CST)
Subject: [LLVMbugs] [Bug 2313] ARM: Rd and Rm registers should be different
in MUL on ARMv5
In-Reply-To:
References:
Message-ID: <20110101203858.44E192A6C130@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=2313
Anton Korobeynikov changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
--- Comment #12 from Anton Korobeynikov 2011-01-01 14:38:56 CST ---
Should be fixed in r122663
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sat Jan 1 14:39:07 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 1 Jan 2011 14:39:07 -0600 (CST)
Subject: [LLVMbugs] [Bug 8157] ARM: rdhi,
rdlo and rm registers should be different in SMULL on ARMv5
In-Reply-To:
References:
Message-ID: <20110101203907.F321E2A6C130@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8157
Anton Korobeynikov changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
--- Comment #4 from Anton Korobeynikov 2011-01-01 14:39:07 CST ---
Should be fixed in r122663
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sat Jan 1 15:56:09 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 1 Jan 2011 15:56:09 -0600 (CST)
Subject: [LLVMbugs] [Bug 8881] New: We need warnings or errors for
unsupported uses of 'asm'
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8881
Summary: We need warnings or errors for unsupported uses of
'asm'
Product: clang
Version: unspecified
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: Frontend
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: rafael.espindola at gmail.com
CC: llvmbugs at cs.uiuc.edu
Currently clang uses the asm in a local variable declaration only when that
variable is used in a asm statement. We should produce warnings when we can.
Some options:
1) Produce warnings when a variable is not use in an asm statement. This would
warn in
void foo(void) {
register unsigned long long bar asm("rax"); //Warning
}
2) Warn when a variable is used in a non asm statement. This would warn in
int foo(void) {
register unsigned long long bar asm("rax");
return bar + 1; // Warning
}
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sat Jan 1 16:40:10 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 1 Jan 2011 16:40:10 -0600 (CST)
Subject: [LLVMbugs] [Bug 8882] New: Poor scalar optimization loses trip
count info
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8882
Summary: Poor scalar optimization loses trip count info
Product: libraries
Version: 1.0
Platform: PC
OS/Version: All
Status: NEW
Severity: normal
Priority: P
Component: Scalar Optimizations
AssignedTo: unassignedbugs at nondot.org
ReportedBy: clattner at apple.com
CC: llvmbugs at cs.uiuc.edu
We compile this:
template
void fill(Iterator first, Iterator last, T value) {
while (first != last) *first++ = value;
}
int X[1000];
void foo() {
fill(X, X+1000, 0);
}
with: clang t.cpp -S -o - -emit-llvm -O3
into:
while.body.i: ; preds = %while.body.i,
%entry
%indvar.i = phi i64 [ %tmp.i, %while.body.i ], [ 0, %entry ]
%tmp.i = add i64 %indvar.i, 1
%first.addr.05.i = getelementptr [1000 x i32]* @X, i64 0, i64 %indvar.i
store i32 0, i32* %first.addr.05.i, align 4, !tbaa !0
%ptrincdec.i.idx.mask = and i64 %tmp.i, 4611686018427387903
%cmp.i = icmp eq i64 %ptrincdec.i.idx.mask, 1000
br i1 %cmp.i, label %_Z4fillIPiiEvT_S1_T0_.exit, label %while.body.i
This loop executes exactly 1000 times. This appears to be being pessimized by
the -indvars pass, as can be seen by:
clang t.cpp -S -o - -emit-llvm -O0 | opt -mem2reg -S -inline -loop-rotate
-instcombine -indvars
which introduces the 'and'. The and is not needed because the GEP can't wrap
around.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sat Jan 1 16:44:06 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 1 Jan 2011 16:44:06 -0600 (CST)
Subject: [LLVMbugs] [Bug 8883] New: SCEV doing a poor job simplifying
constant expressions
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8883
Summary: SCEV doing a poor job simplifying constant expressions
Product: libraries
Version: 1.0
Platform: PC
OS/Version: All
Status: NEW
Severity: normal
Priority: P
Component: Scalar Optimizations
AssignedTo: unassignedbugs at nondot.org
ReportedBy: clattner at apple.com
CC: llvmbugs at cs.uiuc.edu
This code:
void fill(Iterator first, Iterator last, T value) {
while (first != last) *first++ = value;
}
char X[1000];
void foo() {
fill(X, X+1000, 0);
}
Compiles with: clang t.cpp -S -o - -emit-llvm -O3
into:
...
while.body.i: ; preds = %while.body.i,
%entry
%indvar.i = phi i64 [ 0, %entry ], [ %indvar.next.i, %while.body.i ]
%first.addr.05.i = getelementptr [1000 x i8]* @X, i64 0, i64 %indvar.i
store i8 0, i8* %first.addr.05.i, align 1, !tbaa !0
%indvar.next.i = add i64 %indvar.i, 1
%exitcond = icmp eq i64 %indvar.next.i, ptrtoint (i8* getelementptr ([1000 x
i8]* @X, i64 1, i64 sub (i64 0, i64 ptrtoint ([1000 x i8]* @X to i64))) to i64)
br i1 %exitcond, label %_Z4fillIPciEvT_S1_T0_.exit, label %while.body.i
Notice the poor icmp, which should be comparing the indvar against 1000. SCEV
should simplify this away. This can also be seen in
simple_types_constant_folding.llvm.bc in
Nehalem:SingleSource/Benchmarks/Adobe-C++ where SCEV is expanding an expression
for the loop-idiom pass, resulting in this crazy memset:
call void @llvm.memset.p0i8.i64(i8* getelementptr inbounds ([8000 x i8]*
@data8, i64 0, i64 0), i8 %11, i64 ptrtoint (i8* getelementptr ([8000 x i8]*
@data8, i64 1, i64 sub (i64 0, i64 ptrtoint ([8000 x i8]* @data8 to i64))) to
i64), i32 16, i1 false)
Codegen is able to simplify the constant to "8000", but the expression
simplification logic in SCEV should do this. This may be just a matter of
using lib/Analysis/ConstantFolding APIs with target data in scev.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sat Jan 1 17:49:22 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 1 Jan 2011 17:49:22 -0600 (CST)
Subject: [LLVMbugs] [Bug 6394] Optimizer swaps bitcast and getelementptr in
an invalid way
In-Reply-To:
References:
Message-ID: <20110101234922.7B1282A6C12C@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=6394
Henning Thielemann changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|RESOLVED |REOPENED
Version|2.6 |2.8
Resolution|FIXED |
--- Comment #7 from Henning Thielemann 2011-01-01 17:49:21 CST ---
(In reply to comment #6)
> Please update to LLVM 2.8, or better yet, mainline. We do not put out bug
> fixes for previous releases.
I'm continuing to comment on this problem, because it still exists. The only
difference between LLVM-2.6 and LLVM-2.8 is, that the problem now only occurs
when calling the optimizer via LLVM's C interface, and I cannot reproduce it
any longer via 'opt' in the shell. The ticket was closed as FIXED without a
comment how it was fixed. The last comment before closing the ticket was, that
I shall set a target string in LL file and I answered, that I have initially no
LL file, because I'm using JIT (via the C wrapper to LLVM).
I think the following phenomenon is just another instance of the same problem.
I have generated code with the JIT, written to a Bitcode file and disassembled.
Then I have optimized it, trying to match -O1 as good as possible as the C
wrapper around LLVM allows. In the optimized module some getelementptrs to a
struct are replaced by accesses via constant offsets and pointer casting, where
the offsets seem to be made for a 64 bit machine (but I have a 32 bit machine).
optimized by opt (correct):
%9 = getelementptr %0* %2, i32 0, i32 0, i32 0
store i1 true, i1* %9
%10 = getelementptr %0* %2, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0
store i32 1, i32* %10
%11 = getelementptr %0* %2, i32 0, i32 0, i32 1, i32 0, i32 0, i32 1, i32 0,
i32 1
store float 0x3FD99999A0000000, float* %11
%12 = getelementptr %0* %2, i32 0, i32 0, i32 1, i32 0, i32 1, i32 0, i32 0,
i32 1
store i32 %4, i32* %12
%13 = getelementptr %0* %2, i32 0, i32 1, i32 0
store float 0.000000e+00, float* %13
%14 = getelementptr %0* %2, i32 0, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1,
i32 0, i32 0, i32 0
store i32 %6, i32* %14
%15 = getelementptr %0* %2, i32 0, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1,
i32 0, i32 0, i32 1
store float* %8, float** %15
%16 = getelementptr %0* %2, i32 0, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1,
i32 0, i32 1, i32 1
store float 0.000000e+00, float* %16
optimized by JIT (invalid):
%9 = bitcast i8* %1 to i1*
store i1 true, i1* %9, align 1
%10 = getelementptr i8* %1, i64 4
%11 = bitcast i8* %10 to i32*
store i32 1, i32* %11, align 4
%12 = getelementptr i8* %1, i64 8
%13 = bitcast i8* %12 to float*
store float 0x3FD99999A0000000, float* %13, align 4
%14 = getelementptr i8* %1, i64 12
%15 = bitcast i8* %14 to i32*
store i32 %4, i32* %15, align 4
%16 = getelementptr i8* %1, i64 16
%17 = bitcast i8* %16 to float*
store float 0.000000e+00, float* %17, align 4
%18 = getelementptr i8* %1, i64 24
%19 = bitcast i8* %18 to i32*
store i32 %6, i32* %19, align 4
%20 = getelementptr i8* %1, i64 32
%21 = bitcast i8* %20 to float**
store float* %8, float** %21, align 8
%22 = getelementptr i8* %1, i64 40
%23 = bitcast i8* %22 to float*
store float 0.000000e+00, float* %23, align 4
The offsets 16, 24, 32, 40 should have been 16, 20, 24, 28.
The Release Notes of 2.7 say that in LLVM-2.6 no target string meant 'SparcV9'.
Now I tried to reproduce the JIT behaviour via 'opt' by placing
target triple = "sparcv9"
on top of a disassembled bitcode file. However 'opt' still does not generate
those accesses via constant offsets. The LLVM tools also do not complain about
fictional target names, what made me nervous.
In the end I still do not know, how to set the target specification via C
wrapper around LLVM. I hoped that calling LLVMInitializeX86Target and
LLVMInitializeX86TargetInfo would be enough, since these are the appropriate
functions for my machine. If I have to set something other (LLVMSetDataLayout?)
then I wonder how to do that (how to find out the appropriate settings for
every target?) and why its default does not match my information given at
initialization time.
Since you said, that the problem is fixed, I assumed that my current problem
must have another reason. Thus I have spent more than 15 hours with trying to
simulate 'opt's behaviour via LLVM's C interface and since JIT-opt and
Shell-opt still differed, I tried to simulate the JIT optimizer behaviour by
'opt'. Without success. I flooded my code with debug output. I tried hard to
export my JIT constructed functions and typical parameter records to disk
together with a driving main program, that allows to narrow the bug using
bugpoint. No luck: JIT fails, LLVM shell tools work. I dissected my code with
GDB in order to see what actually goes wrong.
I am really frustrated now. I can well imagine that I use LLVM the wrong way.
But if it is that easy, then there must be something improved in the JIT or in
the C wrapper. At the very least the JIT optimizer behaviour must be better
documented.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sat Jan 1 19:17:40 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 1 Jan 2011 19:17:40 -0600 (CST)
Subject: [LLVMbugs] [Bug 8884] New: Surprising behavior with 'friend' and
'extern "C"'
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8884
Summary: Surprising behavior with 'friend' and 'extern "C"'
Product: clang
Version: unspecified
Platform: PC
OS/Version: All
Status: NEW
Severity: normal
Priority: P
Component: C++
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: rafael.espindola at gmail.com
CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com
Clang rejects
----------
extern "C" {
//struct foo;
struct bar {
friend struct foo;
private:
void zed () const;
const struct foo& baz () const;
};
struct foo {
bar x;
void zed () const;
};
void foo::zed () const {
x.zed ();
}
}
----------
but any of the following makes it accept it:
*) Uncommenting the forward declaration of foo.
*) Removing the extern "C".
*) Commenting out the declaration of baz.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sat Jan 1 21:41:20 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 1 Jan 2011 21:41:20 -0600 (CST)
Subject: [LLVMbugs] [Bug 6394] Optimizer swaps bitcast and getelementptr in
an invalid way
In-Reply-To:
References:
Message-ID: <20110102034120.55BB82A6C124@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=6394
Chris Lattner changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|REOPENED |RESOLVED
Resolution| |WORKSFORME
--- Comment #8 from Chris Lattner 2011-01-01 21:41:19 CST ---
Please provide a testcase that reproduces the problem. Without a testcase, we
cannot 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 llvm.org Sun Jan 2 06:32:49 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 2 Jan 2011 06:32:49 -0600 (CST)
Subject: [LLVMbugs] [Bug 8885] New: CMake Generator "Visual Studio 10 Win64"
fails
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8885
Summary: CMake Generator "Visual Studio 10 Win64" fails
Product: Build scripts
Version: trunk
Platform: PC
OS/Version: Windows NT
Status: NEW
Keywords: build-problem
Severity: normal
Priority: P
Component: cmake
AssignedTo: unassignedbugs at nondot.org
ReportedBy: vanboxem.ruben at gmail.com
CC: llvmbugs at cs.uiuc.edu
running the following commands:
cmake ../../Source/LLVM -G"Visual Studio 10 Win64"
causes following error:
>-- Targeting X86
>CMake Error at M:/Development/utils/cmake/share/cmake-2.8/Modules/CMakeDetermineASMCompiler.cmake:68 (CMAKE_DETERMINE_COMPILER_ID_VENDOR):
> Unknown CMake command "CMAKE_DETERMINE_COMPILER_ID_VENDOR".
>Call Stack (most recent call first):
> M:/Development/utils/cmake/share/cmake-2.8/Modules/CMakeDetermineASM_MASMCompiler.cmake:26 (INCLUDE)
> lib/Target/X86/CMakeLists.txt:44 (enable_language)
>
>
>CMake Error: Could not find cmake module file:M:/Development/x64-msvc/llvm/CMakeFiles/CMakeASM_MASMCompiler.cmake
>-- Configuring incomplete, errors occurred!
The generator "Visual Studio 10" does not exhibit this issue.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sun Jan 2 07:29:26 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 2 Jan 2011 07:29:26 -0600 (CST)
Subject: [LLVMbugs] [Bug 8886] New: GCC testcases that clang fails to
devirtualize
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8886
Summary: GCC testcases that clang fails to devirtualize
Product: clang
Version: trunk
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: C++
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: baldrick at free.fr
CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com
The attached testcases were recently posted to the gcc mailing list. GCC from
top-of-tree apparently manages to completely devirtualize the code in "main"
for
each of them, but clang does not at -O4.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sun Jan 2 07:39:31 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 2 Jan 2011 07:39:31 -0600 (CST)
Subject: [LLVMbugs] [Bug 8702] opt -loop-rotate never completes
In-Reply-To:
References:
Message-ID: <20110102133931.232DC2A6C12D@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8702
Duncan Sands changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
--- Comment #13 from Duncan Sands 2011-01-02 07:39:30 CST ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20101227/114494.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 llvm.org Sun Jan 2 12:55:59 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 2 Jan 2011 12:55:59 -0600 (CST)
Subject: [LLVMbugs] [Bug 8885] CMake Generator "Visual Studio 10 Win64" fails
In-Reply-To:
References:
Message-ID: <20110102185559.EB1CD2A6C12C@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8885
?scar Fuentes changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
--- Comment #4 from ?scar Fuentes 2011-01-02 12:55:59 CST ---
This is a cmake 2.8.3 bug.
If you put
INCLUDE(CMakeDetermineCompilerId)
at the beginning of
M:/Development/utils/cmake/share/cmake-2.8/Modules/CMakeDetermineASMCompiler.cmake,
it works fine.
It is already fixed on cmake's git repo.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sun Jan 2 15:27:59 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 2 Jan 2011 15:27:59 -0600 (CST)
Subject: [LLVMbugs] [Bug 8887] New: [deadargelim] should be reworked as a
CGSCC pass
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8887
Summary: [deadargelim] should be reworked as a CGSCC pass
Product: libraries
Version: 1.0
Platform: PC
OS/Version: All
Status: NEW
Severity: normal
Priority: P
Component: Interprocedural Optimizations
AssignedTo: unassignedbugs at nondot.org
ReportedBy: clattner at apple.com
CC: llvmbugs at cs.uiuc.edu
The DeadArgElim is an old-school Module pass. It should be reworked to be a
CGSCC pass, which would allow it to benefit from things like "iterate after
devirtualization" as well as making it more powerful since it will benefit from
things done by the scalar optimizer.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sun Jan 2 17:40:38 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 2 Jan 2011 17:40:38 -0600 (CST)
Subject: [LLVMbugs] [Bug 8888] New: [MS Bugs] char/signed char decl
compatibility.
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8888
Summary: [MS Bugs] char/signed char decl compatibility.
Product: clang
Version: trunk
Platform: All
OS/Version: All
Status: NEW
Severity: normal
Priority: P
Component: -New Bugs
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: bigcheesegs at gmail.com
CC: llvmbugs at cs.uiuc.edu, pichet2000 at gmail.com
MSVC allows the following in C mode (in C++ it's always legal because of
overloading) while clang (correctly) does not. Clang should support this in
ms-extensions mode, although I do not yet know the exact semantics (does it use
signed or unsigned?).
int __cdecl _access( const char * _Filename, int _AccessMode);
int __cdecl _access( const signed char *path,int amode)
{
return 0;
}
C:\Users\Michael\Projects\msvcrt\research>clang -fsyntax-only double-decl.c
double-decl.c(2) : error: conflicting types for '_access'
int __cdecl _access( const signed char *path,int amode)
^
double-decl.c(1) : note: previous declaration is here
int __cdecl _access( const char * _Filename, int _AccessMode);
^
1 error generated.
C:\Users\Michael\Projects\msvcrt\research>clang++ -fsyntax-only -x c++
double-decl.c
C:\Users\Michael\Projects\msvcrt\research>cl -Zs double-decl.c -nologo
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sun Jan 2 21:26:39 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 2 Jan 2011 21:26:39 -0600 (CST)
Subject: [LLVMbugs] [Bug 8889] New: codegenprepare takes about 10% of llc
time on 403.gcc
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8889
Summary: codegenprepare takes about 10% of llc time on 403.gcc
Product: new-bugs
Version: unspecified
Platform: PC
OS/Version: All
Status: NEW
Severity: normal
Priority: P
Component: new bugs
AssignedTo: unassignedbugs at nondot.org
ReportedBy: zwarich at apple.com
CC: llvmbugs at cs.uiuc.edu
codegenprepare takes about 10% of llc time on 403.gcc. Even when I remove the
fixed point iteration, it still uses a bit over half of that. Something is
wrong here, because there is no way codegenprepare should take that long.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sun Jan 2 23:10:18 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 2 Jan 2011 23:10:18 -0600 (CST)
Subject: [LLVMbugs] [Bug 8890] New: build.llvm.org should redirect to
http://google1.osuosl.org:8011/
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8890
Summary: build.llvm.org should redirect to
http://google1.osuosl.org:8011/
Product: Website
Version: unspecified
Platform: PC
OS/Version: All
Status: NEW
Severity: normal
Priority: P
Component: General Website
AssignedTo: unassignedbugs at nondot.org
ReportedBy: zwarich at apple.com
CC: llvmbugs at cs.uiuc.edu
I am so used to typing build.webkit.org. It would be nice to have something as
easy to remember for LLVM.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sun Jan 2 23:33:05 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 2 Jan 2011 23:33:05 -0600 (CST)
Subject: [LLVMbugs] [Bug 8891] New: Early phases of backend recompute
dominators too often
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8891
Summary: Early phases of backend recompute dominators too often
Product: new-bugs
Version: unspecified
Platform: PC
OS/Version: All
Status: NEW
Severity: normal
Priority: P
Component: new bugs
AssignedTo: unassignedbugs at nondot.org
ReportedBy: zwarich at apple.com
CC: llvmbugs at cs.uiuc.edu
Machine Module Information
ModulePass Manager
FunctionPass Manager
Preliminary module verification
Dominator Tree Construction
Module Verifier
Natural Loop Information
Loop Pass Manager
Canonicalize natural loops
Scalar Evolution Analysis
Loop Pass Manager
Induction Variable Users
Canonicalize natural loops
Induction Variable Users
Loop Strength Reduction
Lower Garbage Collection Instructions
Remove unreachable blocks from the CFG
Dominator Tree Construction <--- Here
Exception handling preparation
Optimize for code generation
Insert stack protectors
Preliminary module verification
Dominator Tree Construction <--- Here
Module Verifier
The last dominator tree computation should be easy to get rid of, since
DwarfEHPrepare already preserves dominators, CodeGenPrepare only messes with
blocks that have a single successor, and StackProtector only needs to be
concerned with the two blocks that it adds (which both have the same idom).
The one before it might not be so easy, but "UnreachableBlockElim" is entirely
dominator based, so it won't be 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 llvm.org Mon Jan 3 00:12:47 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 3 Jan 2011 00:12:47 -0600 (CST)
Subject: [LLVMbugs] [Bug 8892] New: CGSCCPassManager does not visit all
functions.
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8892
Summary: CGSCCPassManager does not visit all functions.
Product: libraries
Version: trunk
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: Interprocedural Optimizations
AssignedTo: unassignedbugs at nondot.org
ReportedBy: nlewycky at google.com
CC: llvmbugs at cs.uiuc.edu
Created an attachment (id=5945)
--> (http://llvm.org/bugs/attachment.cgi?id=5945)
testcase
The call graph SCC pass manager does not visit any of the functions in this
example .ll file.
For fun, try running:
opt -instcombine -debug x.ll
to see that it visits every function, then
opt -functionattrs -instcombine -debug x.ll
to see that it visits none.
The pass' runOnSCC is being called only once with a single CallGraphNode, being
the external one (CGN->getFunction() == NULL).
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Mon Jan 3 00:29:34 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 3 Jan 2011 00:29:34 -0600 (CST)
Subject: [LLVMbugs] [Bug 8892] CGSCCPassManager does not visit all functions.
In-Reply-To:
References:
Message-ID: <20110103062934.C28252A6C124@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8892
Chris Lattner changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |INVALID
--- Comment #1 from Chris Lattner 2011-01-03 00:29:34 CST ---
All your functions are invalid, therefore they are not reachable from the
callgraph. We're optimizing compile 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 llvm.org Mon Jan 3 02:28:10 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 3 Jan 2011 02:28:10 -0600 (CST)
Subject: [LLVMbugs] [Bug 8893] New: Build warning in APInt.cpp while
building trunk on DragonflyBSD current.
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8893
Summary: Build warning in APInt.cpp while building trunk on
DragonflyBSD current.
Product: new-bugs
Version: trunk
Platform: PC
OS/Version: DragonFly BSD
Status: NEW
Severity: normal
Priority: P
Component: new bugs
AssignedTo: unassignedbugs at nondot.org
ReportedBy: eocallaghan at auroraux.org
CC: llvmbugs at cs.uiuc.edu
APInt.cpp: In member function 'llvm:APInt llvm:APInt:sqrt() const':
APInt.cpp:1378: warning: passing double for argument 2 to
'llvm:APInt:APInt(unsigned int, uint64_t, bool)'
Building with GCC 4.1.2 and llvm/clang svn rev 122743.
Regards,
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Mon Jan 3 05:11:27 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 3 Jan 2011 05:11:27 -0600 (CST)
Subject: [LLVMbugs] [Bug 8894] New: Clang++ misses an invalid construct in
Sema and assert in codegen.
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8894
Summary: Clang++ misses an invalid construct in Sema and assert
in codegen.
Product: clang
Version: trunk
Platform: All
OS/Version: All
Status: NEW
Severity: normal
Priority: P
Component: C++
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: devlists at shadowlab.org
CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com
Created an attachment (id=5946)
--> (http://llvm.org/bugs/attachment.cgi?id=5946)
Make clang assert
Attached is a sample where clang++ does not emit any diagnostic (where it
should) and so try to codegen invalid code (and assert).
> clang++ --version
clang version 2.9 (trunk 122743)
Target: x86_64-apple-darwin10
Thread model: posix
> clang++ Metadata.cpp
Assertion failed: (E && !hasAggregateLLVMType(E->getType()) && "Invalid scalar
expression to emit"), function EmitScalarExpr, file
/Volumes/MacPro/Projects/OpenSource/llvm/tools/clang/lib/CodeGen/CGExprScalar.cpp,
line 2515.
0 clang 0x0000000100eb6aa2 PrintStackTrace(void*) + 34
1 clang 0x0000000100eb7049 SignalHandler(int) + 857
2 libSystem.B.dylib 0x00007fff8326667a _sigtramp + 26
3 libSystem.B.dylib 0x010000010204ba01 _sigtramp + 2128499617
4 clang 0x0000000100011b06 abort + 22
5 clang 0x0000000100011ac8 __assert_rtn + 56
6 clang 0x0000000100186f5e
clang::CodeGen::CodeGenFunction::EmitScalarExpr(clang::Expr const*, bool) + 222
7 clang 0x000000010019139a clang::StmtVisitor<(anonymous
namespace)::ScalarExprEmitter, llvm::Value*>::Visit(clang::Stmt*) + 21242
8 clang 0x0000000100186f3d
clang::CodeGen::CodeGenFunction::EmitScalarExpr(clang::Expr const*, bool) + 189
[?]
1. parser at end of file
2. Metadata.cpp:15:13: LLVM IR generation of declaration 'test'
3. Metadata.cpp:15:13: Generating code for declaration 'test'
4. Metadata.cpp:15:20: LLVM IR generation of compound statement ('{}')
clang: error: unable to execute command: Illegal instruction
clang: error: clang frontend command failed due to signal 1 (use -v to see
invocation)
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Mon Jan 3 06:30:44 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 3 Jan 2011 06:30:44 -0600 (CST)
Subject: [LLVMbugs] [Bug 8895] New: [REGRESSION] Assertion failed:
(HT.TopLevelMap[ThisEntry->getKey()] == ThisEntry && "Scope imbalance!")
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8895
Summary: [REGRESSION] Assertion failed:
(HT.TopLevelMap[ThisEntry->getKey()] == ThisEntry &&
"Scope imbalance!")
Product: clang
Version: trunk
Platform: Macintosh
OS/Version: MacOS X
Status: NEW
Severity: normal
Priority: P
Component: -New Bugs
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: ismail at namtrac.org
CC: llvmbugs at cs.uiuc.edu
Trying to compile snow.c with clang trunk;
clang -I. -I"/Users/cartman/Sources/ffmpeg" -D_ISOC99_SOURCE
-D_POSIX_C_SOURCE=200112 -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE -DPIC
-DHAVE_AV_CONFIG_H -std=c99 -mdynamic-no-pic -fomit-frame-pointer -fPIC -g
-Wdeclaration-after-statement -Wall -Wno-parentheses -Wno-switch
-Wdisabled-optimization -Wpointer-arith -Wredundant-decls -Wno-pointer-sign
-Wcast-qual -Wwrite-strings -Wtype-limits -Wundef -Wmissing-prototypes -O3
-fno-math-errno -fno-signed-zeros -mllvm -stack-alignment=16 -Qunused-arguments
-MMD -c -o libavcodec/snow.o libavcodec/snow.c
Assertion failed: (HT.TopLevelMap[ThisEntry->getKey()] == ThisEntry && "Scope
imbalance!"), function ~ScopedHashTableScope, file
/Users/cartman/Sources/llvm/include/llvm/ADT/ScopedHashTable.h, line 224.
0 clang 0x00000001014b4df2
llvm::SmallVectorTemplateBase::grow(unsigned long) +
754
1 clang 0x00000001014b5c43
llvm::SmallVectorTemplateBase::grow(unsigned long) +
4419
2 libSystem.B.dylib 0x00007fff81fe867a _sigtramp + 26
3 libSystem.B.dylib 0x0000000102afc080 _sigtramp + 2159098400
4 clang 0x000000010001a472
std::vector
>::_M_insert_aux(__gnu_cxx::__normal_iterator > >,
llvm::PassRegistrationListener* const&) + 3474
5 clang 0x00000001010fc920 llvm::cast_retty >::ret_type
llvm::cast
>(llvm::ilist_iterator const&) + 5984
6 clang 0x00000001010fcf12 llvm::cast_retty >::ret_type
llvm::cast
>(llvm::ilist_iterator const&) + 7506
7 clang 0x00000001010fcedf llvm::cast_retty >::ret_type
llvm::cast
>(llvm::ilist_iterator const&) + 7455
8 clang 0x00000001010fcedf llvm::cast_retty >::ret_type
llvm::cast
>(llvm::ilist_iterator const&) + 7455
9 clang 0x00000001010fcedf llvm::cast_retty >::ret_type
llvm::cast
>(llvm::ilist_iterator const&) + 7455
10 clang 0x00000001010fcedf llvm::cast_retty >::ret_type
llvm::cast
>(llvm::ilist_iterator const&) + 7455
11 clang 0x00000001010fe19c llvm::cast_retty >::ret_type
llvm::cast
>(llvm::ilist_iterator const&) + 12252
12 clang 0x00000001013e8af0 llvm::BasicBlockPass::~BasicBlockPass()
+ 31584
13 clang 0x0000000101256eb1 std::_Rb_tree,
std::_Select1st >,
std::less, std::allocator >
>::_M_insert_unique(std::_Rb_tree_iterator >, std::pair const&) + 7473
14 clang 0x0000000101257c81 std::_Rb_tree,
std::_Select1st >,
std::less, std::allocator >
>::_M_insert_unique(std::_Rb_tree_iterator >, std::pair const&) + 11009
15 clang 0x00000001013ea7ff llvm::BasicBlockPass::~BasicBlockPass()
+ 39023
16 clang 0x00000001013eaaf3 llvm::BasicBlockPass::~BasicBlockPass()
+ 39779
17 clang 0x00000001013eabcd llvm::BasicBlockPass::~BasicBlockPass()
+ 39997
18 clang 0x0000000100177d6f clang::PCHGenerator::~PCHGenerator() +
5823
19 clang 0x000000010029f2fe llvm::IRBuilder
>::CreateIsNull(llvm::Value*, llvm::Twine const&) + 1822
20 clang 0x00000001002e4dc9 llvm::IRBuilder
>::CreateGEP(llvm::Value*, llvm::Value*, llvm::Twine const&) + 1065
21 clang 0x000000010029fefc llvm::IRBuilder
>::CreateIsNull(llvm::Value*, llvm::Twine const&) + 4892
22 clang 0x0000000100052829 llvm::DenseMap,
llvm::DenseMapInfo >::grow(unsigned int) + 7561
23 clang 0x0000000100024612 std::_Rb_tree, std::less,
std::allocator >::_M_insert_unique(std::string const&) + 2002
24 clang 0x000000010001c29a
std::vector
>::_M_insert_aux(__gnu_cxx::__normal_iterator > >,
llvm::PassRegistrationListener* const&) + 11194
25 clang 0x0000000100023644 std::vector >::operator=(std::vector > const&) + 12196
26 clang 0x000000010001abf4
std::vector
>::_M_insert_aux(__gnu_cxx::__normal_iterator > >,
llvm::PassRegistrationListener* const&) + 5396
27 clang 0x0000000000000046
std::vector
>::_M_insert_aux(__gnu_cxx::__normal_iterator > >,
llvm::PassRegistrationListener* const&) + 4294863206
Stack dump:
0. Program arguments: /usr/local/bin/clang -cc1 -triple
x86_64-apple-darwin10.0.0 -emit-obj -disable-free -main-file-name snow.c
-pic-level 2 -masm-verbose -munwind-tables -target-cpu core2
-target-linker-version 97.14 -g -resource-dir /usr/local/bin/../lib/clang/2.9
-dependency-file libavcodec/snow.d -MT libavcodec/snow.o -D _ISOC99_SOURCE -D
_POSIX_C_SOURCE=200112 -D _FILE_OFFSET_BITS=64 -D _LARGEFILE_SOURCE -D PIC -D
HAVE_AV_CONFIG_H -I . -I /Users/cartman/Sources/ffmpeg -O3
-Wdeclaration-after-statement -Wall -Wno-parentheses -Wno-switch
-Wdisabled-optimization -Wpointer-arith -Wredundant-decls -Wno-pointer-sign
-Wcast-qual -Wwrite-strings -Wtype-limits -Wundef -Wmissing-prototypes -std=c99
-ferror-limit 19 -fmessage-length 120 -stack-protector 1 -fblocks
-fdiagnostics-show-option -fcolor-diagnostics -mllvm -stack-alignment=16 -o
libavcodec/snow.o -x c libavcodec/snow.c
1. parser at end of file
2. Per-module optimization passes
3. Running pass 'CallGraph Pass Manager' on module 'libavcodec/snow.c'.
4. Running pass 'Early CSE' on function '@mc_block'
clang: error: unable to execute command: Illegal instruction
clang: error: clang frontend command failed due to signal 1 (use -v to see
invocation)
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Mon Jan 3 11:41:15 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 3 Jan 2011 11:41:15 -0600 (CST)
Subject: [LLVMbugs] [Bug 8654] clang-as-a-library's preprocessor doesn't
trigger Elif, Else and Endif callbacks at the right times
In-Reply-To:
References:
Message-ID: <20110103174115.668CF2A6C12D@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8654
Chandler Carruth changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
CC| |chandlerc at gmail.com
Resolution| |FIXED
--- Comment #4 from Chandler Carruth 2011-01-03 11:41:14 CST ---
Fixed in r122755. Thanks for the patch, and sorry for delays committing 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 llvm.org Mon Jan 3 12:28:30 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 3 Jan 2011 12:28:30 -0600 (CST)
Subject: [LLVMbugs] [Bug 8895] [REGRESSION] Assertion failed:
(HT.TopLevelMap[ThisEntry->getKey()] == ThisEntry && "Scope imbalance!")
In-Reply-To:
References:
Message-ID: <20110103182830.49B5E2A6C12E@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8895
Chris Lattner changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
--- Comment #4 from Chris Lattner 2011-01-03 12:28:29 CST ---
Fixed in r122758, thanks for the 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 llvm.org Mon Jan 3 13:28:04 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 3 Jan 2011 13:28:04 -0600 (CST)
Subject: [LLVMbugs] [Bug 8841] Spurious "unused function" diagnostic with
inline friend definition
In-Reply-To:
References:
Message-ID: <20110103192804.264002A6C124@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8841
Chandler Carruth changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
--- Comment #3 from Chandler Carruth 2011-01-03 13:28:03 CST ---
Submitted this patch in r122763 w/ dgregor's sign off.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Mon Jan 3 15:32:58 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 3 Jan 2011 15:32:58 -0600 (CST)
Subject: [LLVMbugs] [Bug 8896] New: dragonegg: Assertion `bitPosition <
getBitWidth() && "Bit position out of bounds!"' failed.
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8896
Summary: dragonegg: Assertion `bitPosition < getBitWidth() &&
"Bit position out of bounds!"' failed.
Product: new-bugs
Version: trunk
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: new bugs
AssignedTo: unassignedbugs at nondot.org
ReportedBy: regehr at cs.utah.edu
CC: llvmbugs at cs.uiuc.edu, chenyang at cs.utah.edu
We've not done any dragonegg testing before today. If it's not ready for a bit
of stress testing just say.
The -v output doesn't say the dragonegg/LLVM versions; they are 122764 and
122768.
[regehr at gamow tmp439]$ gcc-4.5
-fplugin=/home/regehr/z/compiler-source/dragonegg/dragonegg.so -v
Using built-in specs.
COLLECT_GCC=gcc-4.5
COLLECT_LTO_WRAPPER=/uusoc/exports/scratch/regehr/z/bin/../libexec/gcc/x86_64-unknown-linux-gnu/4.5.3/lto-wrapper
Target: x86_64-unknown-linux-gnu
Configured with: ../configure --prefix=/home/regehr/z --enable-languages=c,c++
--program-suffix=-4.5 --enable-plugin --enable-lto
Thread model: posix
gcc version 4.5.3 20110103 (prerelease) (GCC)
[regehr at gamow tmp439]$ gcc-4.5
-fplugin=/home/regehr/z/compiler-source/dragonegg/dragonegg.so -O2 small.c
small.c:17:1: warning: overflow in implicit constant conversion
small.c: In function 'func_91':
small.c:31:6: warning: comparison between pointer and integer
cc1: APInt.cpp:486: bool llvm::APInt::operator[](unsigned int) const: Assertion
`bitPosition < getBitWidth() && "Bit position out of bounds!"' failed.
Stack dump:
0. Running pass 'Combine redundant instructions' on function '@func_91'
gcc-4.5: Internal error: Aborted (program cc1)
Please submit a full bug report.
See for instructions.
[regehr at gamow tmp439]$ cat small.c
static short foo (short left, int right)
{
return left || right || right >= 0 ? left : left >> right;
}
static unsigned char bar (unsigned long ui1, unsigned long ui2)
{
ui2 ? : (ui1 % ui2);
return 0;
}
short *g_22;
short g_38;
short g_54[7] = {
0x1F32AFCBL
};
short **volatile g = &g_22;
short g_265[4] = {
0x4F2AL
};
int func_14 (short*, short*, short*, int);
void func_91 (int * const p_92)
{
short l_125;
short *l_264 = &l_125;
*g >= bar (g_38,
(1
&& *p_92) ^
foo (func_14 (l_264, &g_54[0], &g_54[6], 0), 0x9DADCB93L)),
g_265[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 llvm.org Mon Jan 3 19:01:59 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 3 Jan 2011 19:01:59 -0600 (CST)
Subject: [LLVMbugs] [Bug 8897] New: Clang fails to find crtbegin.o
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8897
Summary: Clang fails to find crtbegin.o
Product: clang
Version: trunk
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: Driver
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: greened at obbligato.org
CC: llvmbugs at cs.uiuc.edu
I get the following for the Driver/hello.c test on SuSE Linux 10.1:
"/ptmp/dag/build.llvm.trunk.official.opt/x86_64-unknown-linux-gnu/Release+Asserts/bin/clang"
-cc1 -triple x86_64-unknown-linux-gnu -emit-obj -mrelax-all -disable-free
-main-file-name hello.c -mrelocation-model static -mdisable-fp-elim
-masm-verbose -mconstructor-aliases -munwind-tables -target-cpu x86-64
-target-linker-version 2.17 -resource-dir
/ptmp/dag/build.llvm.trunk.official.opt/x86_64-unknown-linux-gnu/Release+Asserts/bin/../lib/clang/2.9
-ferror-limit 19 -fmessage-length 0 -fgnu-runtime -fdiagnostics-show-option -o
/tmp/cc-cppG60.o -x c
/ptmp/dag/llvm-project.official/llvm/trunk/tools/clang/test/Driver/hello.c
"/opt/cpkg/v6/binutils/2.17/bin/ld" --eh-frame-hdr -m elf_x86_64
-dynamic-linker /lib64/ld-linux-x86-64.so.2 -o
/ptmp/dag/build.llvm.trunk.official.opt/x86_64-unknown-linux-gnu/tools/clang/test/Driver/Output/hello.c.tmp
/usr/lib/../lib64/crt1.o /usr/lib/../lib64/crti.o crtbegin.o -L
-L/../../../../lib64 -L/lib/../lib64 -L/usr/lib/../lib64 -L/../../..
/tmp/cc-cppG60.o -lgcc --as-needed -lgcc_s --no-as-needed -lc -lgcc --as-needed
-lgcc_s --no-as-needed crtend.o /usr/lib/../lib64/crtn.o
/opt/cpkg/v6/binutils/2.17/bin/ld: crtbegin.o: No such file: No such file or
directory
clang: error: linker command failed with exit code 1 (use -v to see invocation)
I believe this is happening because the gcc used to build clang is not in a
standard location:
/opt/gcc/4.5.1/bin/g++
Or alternatively, the system gcc is not listed in the ToolChains.cpp
GccVersions list:
const char* GccVersions[] = {"4.5.1", "4.5", "4.4.5", "4.4.4", "4.4.3",
"4.4",
"4.3.4", "4.3.3", "4.3.2"};
dag at royale:/ptmp/dag/compiler_ref$ which gcc
/usr/bin/gcc
dag at royale:/ptmp/dag/compiler_ref$ gcc --version
gcc (GCC) 4.1.2 20070115 (prerelease) (SUSE Linux)
Copyright (C) 2006 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
In any case, the driver should not rely on a small set of hardcoded guesses as
to where crtbegin.o is. A simple solution would be to use the build gcc's
-print-search-dirs to get a list of candidate directories and fall back on the
hardcoded ones as a last resort. Obviously different handling would be
required if gcc were not the build compiler but for Linux, most of the time it
will be.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Mon Jan 3 21:39:17 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 3 Jan 2011 21:39:17 -0600 (CST)
Subject: [LLVMbugs] [Bug 8898] New: dragonegg bug: nondeterministic wrong
code
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8898
Summary: dragonegg bug: nondeterministic wrong code
Product: new-bugs
Version: trunk
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: new bugs
AssignedTo: unassignedbugs at nondot.org
ReportedBy: regehr at cs.utah.edu
CC: llvmbugs at cs.uiuc.edu, chenyang at cs.utah.edu
The correct answer is 1. No idea what's going on, perhaps memory unsafety +
ASLR. GCC-4.5, dragonegg, and LLVM are all from today.
[regehr at gamow tmp440]$ gcc-4.5
-fplugin=/home/regehr/z/compiler-source/dragonegg/dragonegg.so -O0 small.c -o
small
[regehr at gamow tmp440]$ ./small
g_321[1].f0 = 252
[regehr at gamow tmp440]$ gcc-4.5
-fplugin=/home/regehr/z/compiler-source/dragonegg/dragonegg.so -O0 small.c -o
small
[regehr at gamow tmp440]$ ./small
g_321[1].f0 = 235
[regehr at gamow tmp440]$ gcc-4.5
-fplugin=/home/regehr/z/compiler-source/dragonegg/dragonegg.so -O0 small.c -o
small
[regehr at gamow tmp440]$ ./small
g_321[1].f0 = 156
[regehr at gamow tmp440]$ clang -O0 small.c -o small[regehr at gamow tmp440]$ ./small
g_321[1].f0 = 1
[regehr at gamow tmp440]$ gcc-4.5
-fplugin=/home/regehr/z/compiler-source/dragonegg/dragonegg.so -v
Using built-in specs.
COLLECT_GCC=gcc-4.5
COLLECT_LTO_WRAPPER=/uusoc/exports/scratch/regehr/z/bin/../libexec/gcc/x86_64-unknown-linux-gnu/4.5.3/lto-wrapper
Target: x86_64-unknown-linux-gnu
Configured with: ../configure --prefix=/home/regehr/z --enable-languages=c,c++
--program-suffix=-4.5 --enable-plugin --enable-lto
Thread model: posix
gcc version 4.5.3 20110103 (prerelease) (GCC)
[regehr at gamow tmp440]$ cat small.c
#pragma pack(push)
#pragma pack(1)
struct S1 {
int f0 : 23;
};
#pragma pack(pop)
static struct S1 g_36 = {0};
static struct S1 g_321[2] = {{1}, {1}};
static struct S1 func_31(void)
{
return g_36;
}
int printf(const char *format, ...);
int main (void)
{
g_321[0] = func_31();
printf("g_321[1].f0 = %d\n", g_321[1].f0);
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 llvm.org Tue Jan 4 03:23:24 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 4 Jan 2011 03:23:24 -0600 (CST)
Subject: [LLVMbugs] [Bug 8899] New: including fenv.h fails on linux
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8899
Summary: including fenv.h fails on linux
Product: clang
Version: 2.8
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: Frontend
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: quintino at vki.ac.be
CC: llvmbugs at cs.uiuc.edu
Created an attachment (id=5953)
--> (http://llvm.org/bugs/attachment.cgi?id=5953)
raising fe exceptions
When including the header:
#include
in Linux, the compiler loses track of the headers being included inside this
header...
---- /usr/include/c++/4.4/fenv.h has this snippet:
#include
#if _GLIBCXX_HAVE_FENV_H
# include_next
#endif
and I think that the #include_next is treated incorrectly, because it should
then parse the /usr/include/fenv.h and apparently it does not.
For the attached example, the errors are:
main.cxx:6:17: error: use of undeclared identifier 'FE_DIVBYZERO'
if ( flags & FE_DIVBYZERO ) printf(" division by zero exception\n");
^
main.cxx:7:17: error: use of undeclared identifier 'FE_ALL_EXCEPT'
if (!(flags & FE_ALL_EXCEPT)) printf(" no exceptions\n");
^
main.cxx:12:3: error: use of undeclared identifier 'fexcept_t'
fexcept_t flags;
^
main.cxx:14:20: error: use of undeclared identifier 'flags'
fegetexceptflag(&flags, FE_ALL_EXCEPT);
^
main.cxx:16:15: error: use of undeclared identifier 'flags'
print_flags(flags);
^
main.cxx:18:17: error: use of undeclared identifier 'FE_DIVBYZERO'
feraiseexcept(FE_DIVBYZERO); /* raise divide-by-zero exception */
^
main.cxx:19:30: error: use of undeclared identifier 'FE_DIVBYZERO'
int excepts = fetestexcept(FE_DIVBYZERO | FE_INEXACT);
I can attest that it works on:
- Mac OS X (Snow Leopard) + clang++ 2.8 (but of course the fenv.h headers are
different)
- Linux Ubuntu 10.10 + gcc 4.4
It fails so far in:
- Linux Ubuntu 10.10 + clang++ 2.8 (but of course the fenv.h headers are
different)
I also tried the svn trunk around one month ago and the problem was the same.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Tue Jan 4 08:11:44 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 4 Jan 2011 08:11:44 -0600 (CST)
Subject: [LLVMbugs] [Bug 8900] New: Shufflevector of doubles generates
incorrect code (using float shuff)
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8900
Summary: Shufflevector of doubles generates incorrect code
(using float shuff)
Product: new-bugs
Version: 2.8
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: new bugs
AssignedTo: unassignedbugs at nondot.org
ReportedBy: nadav256 at gmail.com
CC: llvmbugs at cs.uiuc.edu
In the code below, the shuffle instruction is first selected to be unpckld,
which is correct. However, during the ISel pattern matching phase, it is
replaced with the unpackls instruction (which is incorrect).
define <2 x double> @test_even_double4(<4 x double>* %srcA, <2 x double>*
%dst){
%i5 = getelementptr inbounds <4 x double> * %srcA, i32 3
%i6 = load <4 x double>* %i5, align 32
%m387 = shufflevector <4 x double> %i6, <4 x double> undef, <2 x i32>
ret <2 x double> %m387
}
movaps XMM0, XMMWORD PTR [RDI + 96]
unpcklps XMM0, XMMWORD PTR [RDI + 112] # xmm0 = xmm0[0],mem[0],xmm0[1],mem[1]
ret
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Tue Jan 4 08:29:19 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 4 Jan 2011 08:29:19 -0600 (CST)
Subject: [LLVMbugs] [Bug 8901] New: Attribute mode rejected for enum types.
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8901
Summary: Attribute mode rejected for enum types.
Product: clang
Version: unspecified
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: -New Bugs
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: zaffanella at cs.unipr.it
CC: llvmbugs at cs.uiuc.edu
gcc allows mode attribute to be applied to (a declaration of) an enumerated
type:
typedef enum { a, b } E __attribute__((__mode__(__QI__)));
clang rejects the code above with the following message:
error: mode attribute only supported for integer and floating-point types
typedef enum { a, b } E __attribute__((__mode__(__QI__)));
In C99 (6.2.5 p17) an enumerated type is an integer type.
Also note that gcc allows for the mode attribute to be directly specified on
the enumeration type declaration (i.e., not necessarily through a typedef
declaration). Namely, the following variants are accepted too:
enum __attribute__((__mode__(__QI__))) E { a, b };
typedef enum __attribute__((__mode__(__QI__))) { a, b } E;
clang instead complains as follows:
error: 'mode' attribute invalid on this declaration, requires typedef or value
typedef enum __attribute__((__mode__(__QI__))) { a, b } E;
^ ~~~~~~~~
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Tue Jan 4 08:43:28 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 4 Jan 2011 08:43:28 -0600 (CST)
Subject: [LLVMbugs] [Bug 8902] New: Wrong type source location for
CXXMethodDecl
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8902
Summary: Wrong type source location for CXXMethodDecl
Product: clang
Version: unspecified
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: C++
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: alr at google.com
CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com
CXXMethodDecl *decl;
When I call
decl->getResultType().getAsString()
I get:
const string &
However:
decl->getTypeSourceInfo()->getTypeLoc().getBeginLoc().dump()
gives me location of string not const
No matter what type location methods I try to use, I can't seem to get to
correct source pos (the one for const).
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Tue Jan 4 08:48:14 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 4 Jan 2011 08:48:14 -0600 (CST)
Subject: [LLVMbugs] [Bug 8903] New: Wrong source location returned by
FieldDecl::getLocEnd
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8903
Summary: Wrong source location returned by FieldDecl::getLocEnd
Product: clang
Version: unspecified
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: C++
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: alr at google.com
CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com
FieldDecl::getLocEnd() returns position after type decl, not semicolon
class A {
bool field1;
// ^the value returned by getLocEnd()
bool field2;
// ^ expected
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Tue Jan 4 08:53:17 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 4 Jan 2011 08:53:17 -0600 (CST)
Subject: [LLVMbugs] [Bug 8903] Wrong source location returned by
FieldDecl::getLocEnd
In-Reply-To:
References:
Message-ID: <20110104145317.4A7BE2A6C124@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8903
Douglas Gregor changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |INVALID
--- Comment #1 from Douglas Gregor 2011-01-04 08:53:16 CST ---
Clang doesn't keep track of semicolons. This was essentially a design decision,
so it doesn't make sense for us to hold a bug open for any specific case of
missing-semicolon information (they're all missing). If the lack of semicolon
location information is a problem, please open a discussion on the mailing
list.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Tue Jan 4 09:06:30 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 4 Jan 2011 09:06:30 -0600 (CST)
Subject: [LLVMbugs] [Bug 8903] Wrong source location returned by
FieldDecl::getLocEnd
In-Reply-To:
References:
Message-ID: <20110104150630.835E32A6C124@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8903
Alex changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|RESOLVED |REOPENED
Resolution|INVALID |
--- Comment #2 from Alex 2011-01-04 09:06:30 CST ---
(In reply to comment #1)
> Clang doesn't keep track of semicolons. This was essentially a design decision,
> so it doesn't make sense for us to hold a bug open for any specific case of
> missing-semicolon information (they're all missing). If the lack of semicolon
> location information is a problem, please open a discussion on the mailing
> list.
The problem is not about semicolons. getLocEnd() position returns position
after type declaration, not after name (the end of token before semicolon).
Please take another look at the description.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Tue Jan 4 09:08:19 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 4 Jan 2011 09:08:19 -0600 (CST)
Subject: [LLVMbugs] [Bug 8903] Wrong source location returned by
FieldDecl::getLocEnd
In-Reply-To:
References:
Message-ID: <20110104150819.BA9F62A6C12E@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8903
Douglas Gregor changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|REOPENED |RESOLVED
Resolution| |INVALID
--- Comment #3 from Douglas Gregor 2011-01-04 09:08:19 CST ---
Clang is providing the location at the beginning of the field name, which is
the last token in the declaration (except for the semicolon, which is not in
the design). Please see
http://clang.llvm.org/docs/InternalsManual.html#SourceRange for more
information on this design.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Tue Jan 4 10:59:12 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 4 Jan 2011 10:59:12 -0600 (CST)
Subject: [LLVMbugs] [Bug 7696] SelectionDAGBuilder doing bad things on
certain architectures
In-Reply-To:
References:
Message-ID: <20110104165912.7034D2A6C12C@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=7696
Micah Villmow changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
--- Comment #2 from Micah Villmow 2011-01-04 10:59:11 CST ---
This was checked in by chris lattner.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Tue Jan 4 12:19:28 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 4 Jan 2011 12:19:28 -0600 (CST)
Subject: [LLVMbugs] [Bug 8896] instcombine crash on overlarge ashr
In-Reply-To:
References:
Message-ID: <20110104181928.A5C162A6C124@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8896
Chris Lattner changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
--- Comment #4 from Chris Lattner 2011-01-04 12:19:28 CST ---
Fixed in r122814, 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 llvm.org Tue Jan 4 12:59:01 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 4 Jan 2011 12:59:01 -0600 (CST)
Subject: [LLVMbugs] [Bug 8902] Clang does not track the source locations of
cv-qualifiers
In-Reply-To:
References:
Message-ID: <20110104185901.A4CD52A6C12F@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8902
Douglas Gregor changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |WONTFIX
--- Comment #5 from Douglas Gregor 2011-01-04 12:59:01 CST ---
We decided not to include source-location information for cv-qualifiers because
we're worried that it will greatly increase Clang's memory footprint. Zhanyong
brings up a good point, though: I'm closing this bug as WONTFIX. If someone
believes that this is the wrong choose, please discuss it on the mailing list
rather than in a 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 llvm.org Tue Jan 4 13:03:37 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 4 Jan 2011 13:03:37 -0600 (CST)
Subject: [LLVMbugs] [Bug 8853] Clang goes into an endless loop while trying
to compile ScummVM
In-Reply-To:
References:
Message-ID: <20110104190337.BD3162A6C12F@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8853
Dale Johannesen changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
--- Comment #11 from Dale Johannesen 2011-01-04 13:03:36 CST ---
This case is fixed by 122821.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Tue Jan 4 15:40:50 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 4 Jan 2011 15:40:50 -0600 (CST)
Subject: [LLVMbugs] [Bug 8904] New: gnu screen with/without InstCombine
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8904
Summary: gnu screen with/without InstCombine
Product: libraries
Version: trunk
Platform: PC
OS/Version: FreeBSD
Status: NEW
Severity: normal
Priority: P
Component: Scalar Optimizations
AssignedTo: unassignedbugs at nondot.org
ReportedBy: rdivacky at freebsd.org
CC: llvmbugs at cs.uiuc.edu
these three files show difference:
(first one is with InstCombine, second one is without)
-rw------- 1 root wheel 64904 Jan 4 22:37 work/screen-4.0.3/ansi.o
-rw------- 1 root wheel 64920 Jan 4 22:40 work/screen-4.0.3/ansi.o
-rw------- 1 root wheel 9464 Jan 4 22:37 work/screen-4.0.3/input.o
-rw------- 1 root wheel 10880 Jan 4 22:40 work/screen-4.0.3/input.o
-rw------- 1 root wheel 68744 Jan 4 22:37 work/screen-4.0.3/screen.o
-rw------- 1 root wheel 68792 Jan 4 22:40 work/screen-4.0.3/screen.o
all the files are compiled with simple:
clang -c -I. -I. -O2 -pipe -fno-strict-aliasing foo.c
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Tue Jan 4 16:31:28 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 4 Jan 2011 16:31:28 -0600 (CST)
Subject: [LLVMbugs] [Bug 8905] New: Clang rejects partial specialization
with non-type template argument of different type
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8905
Summary: Clang rejects partial specialization with non-type
template argument of different type
Product: clang
Version: trunk
Platform: PC
OS/Version: All
Status: NEW
Severity: normal
Priority: P
Component: C++
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: dgregor at apple.com
CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com
terfin:clang dgregor$ cat t.cpp
template
struct X;
template
struct X;
terfin:clang dgregor$ clang++ t.cpp
t.cpp:5:10: error: non-type template argument depends on a template parameter
of
the partial specialization
struct X;
^~
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Tue Jan 4 17:37:14 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 4 Jan 2011 17:37:14 -0600 (CST)
Subject: [LLVMbugs] [Bug 8905] Clang rejects partial specialization with
non-type template argument of different type
In-Reply-To:
References:
Message-ID: <20110104233714.402072A6C12D@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8905
Douglas Gregor changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
--- Comment #1 from Douglas Gregor 2011-01-04 17:37:14 CST ---
FUN! Fixed in Clang r122851.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Tue Jan 4 17:51:43 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 4 Jan 2011 17:51:43 -0600 (CST)
Subject: [LLVMbugs] [Bug 8906] New: Overly aggressive memset optimization.
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8906
Summary: Overly aggressive memset optimization.
Product: new-bugs
Version: trunk
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: new bugs
AssignedTo: unassignedbugs at nondot.org
ReportedBy: rich at pennware.com
CC: llvmbugs at cs.uiuc.edu
I updated to 122785 and compiled the my modified NetBSD standard library.
Unfortunately the memset() function blew up the stack. Its body was replaced by
a call to memset(). Ouch.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Tue Jan 4 18:13:56 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 4 Jan 2011 18:13:56 -0600 (CST)
Subject: [LLVMbugs] [Bug 7851] clang++ can't compile CryptoPP 5.6.0
In-Reply-To:
References:
Message-ID: <20110105001356.2ABEE2A6C12F@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=7851
Douglas Gregor changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|REOPENED |RESOLVED
Resolution| |FIXED
--- Comment #10 from Douglas Gregor 2011-01-04 18:13:55 CST ---
Fixed in Clang r122853.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Tue Jan 4 19:03:54 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 4 Jan 2011 19:03:54 -0600 (CST)
Subject: [LLVMbugs] [Bug 8906] Overly aggressive memset optimization.
In-Reply-To:
References:
Message-ID: <20110105010354.B56C72A6C12C@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8906
Chris Lattner changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
--- Comment #2 from Chris Lattner 2011-01-04 19:03:54 CST ---
Fixed in r122854. If you don't want the compiler to form a memset, please
build with -fno-builtin.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Tue Jan 4 22:43:36 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 4 Jan 2011 22:43:36 -0600 (CST)
Subject: [LLVMbugs] [Bug 8907] New: [meta] build firefox 4
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8907
Summary: [meta] build firefox 4
Product: clang
Version: trunk
Platform: PC
OS/Version: All
Status: NEW
Severity: normal
Priority: P
Component: C++
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: rafael.espindola at gmail.com
CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com
Currently clang cannot build firefox 4 because of issues in Firefox's code that
are tracked in https://bugzilla.mozilla.org/show_bug.cgi?id=574346 and because
of clang issues that are tracked by 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 llvm.org Wed Jan 5 00:39:38 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 5 Jan 2011 00:39:38 -0600 (CST)
Subject: [LLVMbugs] [Bug 8908] New: call ModRefInfo not sensitive to ordering
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8908
Summary: call ModRefInfo not sensitive to ordering
Product: libraries
Version: trunk
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: Scalar Optimizations
AssignedTo: unassignedbugs at nondot.org
ReportedBy: nlewycky at google.com
CC: llvmbugs at cs.uiuc.edu
Created an attachment (id=5960)
--> (http://llvm.org/bugs/attachment.cgi?id=5960)
.ll testcase (run opt -basicaa -gvn)
This is a breakout bug from bug 8886 comment 2. The problem is that we have in
pseudo:
%x = alloca
store @CONSTANT, %x
call @external1()
%DEAD = load %x
call @external2(%x)
and the load isn't being eliminated by GVN. MemDep does its backwards scan but
when it asks BasicAA about the call to external1, BasicAA looks at the alloca's
address leaking into external2 and bails returning ModRef.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Wed Jan 5 08:00:28 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 5 Jan 2011 08:00:28 -0600 (CST)
Subject: [LLVMbugs] [Bug 8911] New: VMLA/VMLS codegen error
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8911
Summary: VMLA/VMLS codegen error
Product: libraries
Version: trunk
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: Backend: ARM
AssignedTo: unassignedbugs at nondot.org
ReportedBy: rengolin at gmail.com
CC: llvmbugs at cs.uiuc.edu
When faced with the following code:
#define __ARM_NEON__
#include
float32x2_t f32d;
void vmul() {
f32d = vmla_f32(f32d, f32d, f32d);
}
Clang produces the IR sequence:
%tmp.i = load <2 x float>* %__a.addr.i, align 8
%tmp1.i = load <2 x float>* %__b.addr.i, align 8
%tmp2.i = load <2 x float>* %__c.addr.i, align 8
%mul.i = fmul <2 x float> %tmp1.i, %tmp2.i
%add.i = fadd <2 x float> %tmp.i, %mul.i
Which used to be translated into:
vmla.f32 d0, d1, d2
A while ago, this was expanded wrongly to a pair of vmul+vadd but in the latest
trunk I'm getting the assert:
clang: /work/llvm/src/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp:810:
llvm::SDValue::SelectionDAGLegalize::LegalizeOp(llvm::SDValue):
Assertion `(isTypeLegal(Node->getOperand(i).getValueType()) ||
Node->getOperand(i).getOpcode() == ISD::TargetConstant) && "Unexpected illegal
type!"' failed.
With Last error in the stack:
4. Running pass 'ARM Instruction Selection' on function '@vmul'
I believe there are two bugs here, fixing the codegen assert will clear to
expose the real error, not matching mul+add to vmla.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Wed Jan 5 09:14:39 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 5 Jan 2011 09:14:39 -0600 (CST)
Subject: [LLVMbugs] [Bug 8912] New: Non-type template argument parse error
(with a functional cast expression).
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8912
Summary: Non-type template argument parse error (with a
functional cast expression).
Product: clang
Version: trunk
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: C++
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: zaffanella at cs.unipr.it
CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com
Parsing the following program
template struct S {};
S 1)> s;
produces the following error:
bug.cc:2:10: error: expected ')'
S 1)> s;
^
bug.cc:2:7: note: to match this '('
S 1)> s;
^
NOTE: no error is obtained if the functional cast is changed into a C-style
cast.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Wed Jan 5 09:45:30 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 5 Jan 2011 09:45:30 -0600 (CST)
Subject: [LLVMbugs] [Bug 8913] New: clang++ asserts when compiling vmac.cpp
from CryptoPP 5.6.1
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8913
Summary: clang++ asserts when compiling vmac.cpp from CryptoPP
5.6.1
Product: clang
Version: trunk
Platform: Macintosh
OS/Version: MacOS X
Status: NEW
Severity: normal
Priority: P
Component: C++
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: mclow at qualcomm.com
CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com
Created an attachment (id=5961)
--> (http://llvm.org/bugs/attachment.cgi?id=5961)
Preprocessed source code
After Doug fixed bug #7851, I decided to make sure that CryptoPP would, in
fact, build.
Clang threw several errors about "dependent base classes", so I fixed those in
the CryptoPP sources.
However, there was one file that it didn't handle.
This was done with clang/llvm built from revision 122872
attached file was created via:
clang++ -DNDEBUG -g -O2 -DCRYPTOPP_DISABLE_ASM -pipe -E vmac.cpp >
vmac.i
-----
clang++ -DNDEBUG -g -O2 -arch x86_64 -arch i386 -DCRYPTOPP_DISABLE_ASM -pipe -c
vmac.cpp
In file included from vmac.cpp:5:
In file included from ./vmac.h:4:
In file included from ./iterhash.h:5:
In file included from ./secblock.h:7:
./misc.h:414:8: warning: comparison of unsigned expression < 0 is always false
[-Wtautological-compare]
if (a < 0)
~ ^ ~
In file included from vmac.cpp:5:
In file included from ./vmac.h:4:
In file included from ./iterhash.h:7:
./simple.h:35:110: note: in instantiation of function template specialization
'CryptoPP::IntToString' requested here
...length) : InvalidArgument(algorithm + ": " + IntToString(length) + " is
not a ...
^
In file included from vmac.cpp:5:
In file included from ./vmac.h:4:
In file included from ./iterhash.h:5:
In file included from ./secblock.h:7:
./misc.h:414:8: warning: comparison of unsigned expression < 0 is always false
[-Wtautological-compare]
if (a < 0)
~ ^ ~
In file included from vmac.cpp:5:
In file included from ./vmac.h:4:
In file included from ./iterhash.h:7:
./simple.h:42:113: note: in instantiation of function template specialization
'CryptoPP::IntToString' requested here
...rounds) : InvalidArgument(algorithm + ": " + IntToString(rounds) + " is
not a ...
^
Assertion failed: (getMinSignedBits() <= 64 && "Too many bits for int64_t"),
function getSExtValue, file /Volumes/Bay2/LLVM/llvm/include/llvm/ADT/APInt.h,
line 1158.
0 clang 0x00000001012bd3ae PrintStackTrace(void*) + 38
1 clang 0x00000001012bd969 SignalHandler(int) + 254
2 libSystem.B.dylib 0x00007fff8630467a _sigtramp + 26
3 clang 0x000000010107cd29 llvm::isa_impl_wrap::doit(llvm::Value const&) + 21
4 clang 0x000000010001fc09 raise + 27
5 clang 0x000000010001fc19 abort + 14
6 clang 0x000000010001fca6 __gnu_cxx::new_allocator >::new_allocator() + 0
7 clang 0x00000001009a0094 llvm::APInt::getSExtValue() const + 148
8 clang 0x0000000100a4695d llvm::ConstantInt::getSExtValue() const
+ 25
9 clang 0x0000000100d9688e
llvm::DwarfDebug::constructGlobalVariableDIE(llvm::MDNode const*) + 1706
10 clang 0x0000000100d97447
llvm::DwarfDebug::beginModule(llvm::Module*) + 423
11 clang 0x0000000100d9bb3a
llvm::DwarfDebug::DwarfDebug(llvm::AsmPrinter*, llvm::Module*) + 1048
12 clang 0x0000000100d8277b
llvm::AsmPrinter::doInitialization(llvm::Module&) + 853
13 clang 0x00000001011ebeab
llvm::FPPassManager::doInitialization(llvm::Module&) + 65
14 clang 0x00000001011efb5b
llvm::FPPassManager::runOnModule(llvm::Module&) + 29
15 clang 0x00000001011ef68c
llvm::MPPassManager::runOnModule(llvm::Module&) + 384
16 clang 0x00000001011f0e3d
llvm::PassManagerImpl::run(llvm::Module&) + 111
17 clang 0x00000001011f0e9f llvm::PassManager::run(llvm::Module&) +
33
18 clang 0x000000010017deea (anonymous
namespace)::EmitAssemblyHelper::EmitAssembly(clang::BackendAction,
llvm::raw_ostream*) + 748
19 clang 0x000000010017df9e
clang::EmitBackendOutput(clang::Diagnostic&, clang::CodeGenOptions const&,
clang::TargetOptions const&, llvm::Module*, clang::BackendAction,
llvm::raw_ostream*) + 115
20 clang 0x0000000100261526 (anonymous
namespace)::BackendConsumer::HandleTranslationUnit(clang::ASTContext&) + 436
21 clang 0x000000010029f969 clang::ParseAST(clang::Sema&, bool) +
534
22 clang 0x000000010008a1c7
clang::ASTFrontendAction::ExecuteAction() + 233
23 clang 0x0000000100261b98 clang::CodeGenAction::ExecuteAction() +
794
24 clang 0x000000010008a2ce clang::FrontendAction::Execute() + 256
25 clang 0x000000010006b4f8
clang::CompilerInstance::ExecuteAction(clang::FrontendAction&) + 710
26 clang 0x000000010002d2ad
clang::ExecuteCompilerInvocation(clang::CompilerInstance*) + 725
27 clang 0x0000000100021e98 cc1_main(char const**, char const**,
char const*, void*) + 941
28 clang 0x0000000100029c3b main + 450
29 clang 0x00000001000215d8 start + 52
30 clang 0x0000000000000028 start + 4294830724
Stack dump:
0. Program arguments: /Volumes/Bay2/LLVM/llvm-build/Debug+Asserts/bin/clang
-cc1 -triple x86_64-apple-darwin10.0.0 -emit-obj -disable-free -main-file-name
vmac.cpp -pic-level 1 -mdisable-fp-elim -masm-verbose -munwind-tables
-target-cpu core2 -target-linker-version 97.17 -g -resource-dir
/Volumes/Bay2/LLVM/llvm-build/Debug+Asserts/bin/../lib/clang/2.9 -D NDEBUG -D
CRYPTOPP_DISABLE_ASM -O2 -ferror-limit 19 -fmessage-length 87 -stack-protector
1 -fblocks -fexceptions -fdiagnostics-show-option -fcolor-diagnostics -o
/var/folders/E-/E-Uv6JW8Hni-PjXnXWJGbU+++TI/-Tmp-/cc-TfbSZn.o -x c++ vmac.cpp
1. parser at end of file
2. Code generation
3. Running pass 'Function Pass Manager' on module 'vmac.cpp'.
clang: error: unable to execute command: Illegal instruction
clang: error: clang frontend command failed due to signal 1 (use -v to see
invocation)
make: *** [vmac.o] Error 255
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Wed Jan 5 09:57:44 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 5 Jan 2011 09:57:44 -0600 (CST)
Subject: [LLVMbugs] [Bug 8914] New: __dead2 not respected
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8914
Summary: __dead2 not respected
Product: clang
Version: trunk
Platform: All
OS/Version: All
Status: NEW
Severity: normal
Priority: P
Component: Static Analyzer
AssignedTo: kremenek at apple.com
ReportedBy: erik at cederstrand.dk
CC: llvmbugs at cs.uiuc.edu
FreeBSD run the analyzer regularly on our source code. The scan reports contain
many false positives caused by the analyzer not knowing that a call to e.g.
errx() or exit() doesn't return. Instead it proceeds to find e.g. a bogus null
dereference. Examples:
http://scan.freebsd.your.org/freebsd-head/usr.sbin.jail/2011-01-05-amd64/report-W2zjAu.html#EndPath:
Step 12 (line 192) calls usage() (line 552) which in turn calls exit()
http://scan.freebsd.your.org/freebsd-head/usr.sbin.pw/2011-01-05-amd64/report-KkilQ3.html#EndPath:
Step 6 (line 167) calls cmdhelp() (line 285) which in turn calls exit()
Many such functions in FreeBSD are marked with "__dead2" to indicate that the
function never returns (e.g. exit() in
http://svn.freebsd.org/viewvc/base/head/include/stdlib.h?revision=206997&view=markup
and errx() in
http://svn.freebsd.org/viewvc/base/head/include/err.h?revision=203964&view=markup).
If the analyzer was able to find and honour this flag we would be able to
ignore a huge number of false positive reports, increasing the value of the
reports.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Wed Jan 5 10:41:29 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 5 Jan 2011 10:41:29 -0600 (CST)
Subject: [LLVMbugs] [Bug 5314] [Linux Kernel] Inline asm asmprinter crash
with P modifier
In-Reply-To:
References:
Message-ID: <20110105164129.A53EB2A6C12F@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=5314
Rob changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|RESOLVED |REOPENED
CC| |rob1weld at aol.com
Resolution|FIXED |
--- Comment #5 from Rob 2011-01-05 10:41:28 CST ---
(In reply to comment #4)
> Crash fixed in 108545. Generated code made correct in 108548 (I think).
I downloaded the Attachment and I can reproduce this BR, I 'REOPENED' Tiago's
BR to avoid a Dupe:
# uname -a
Linux debian 2.6.33.7 #1 SMP PREEMPT Tue Jan 4 19:10:07 PST 2011 i686 GNU/Linux
# cd /usr/local/linux-2.6.33.7
# make-kpkg clean
# MAKEFLAGS="CC=llvm-gcc-4.2" make-kpkg --initrd kernel_image
...
HOSTCC scripts/kallsyms
HOSTCC scripts/pnmtologo
HOSTCC scripts/conmakehash
HOSTCC scripts/bin2c
CC init/main.o
UNREACHABLE executed!
init/main.c:897: internal compiler error: Aborted
Please submit a full bug report,
with preprocessed source if appropriate.
See for instructions.
make[2]: *** [init/main.o] Error 1
make[1]: *** [init] Error 2
make[1]: Leaving directory `/usr/local/linux-2.6.33.7'
make: *** [debian/stamp/build/kernel] Error 2
root at debian:/usr/local/linux-2.6.33.7#
(Note: Line init/main.c:897 is EOF - nothing to see but "}")
# llc-2.7 -version
Low Level Virtual Machine (http://llvm.org/):
llvm version 2.7 (Debian 2.7-6)
Optimized build.
Built Sep 23 2010 (21:14:53).
Host: i386-pc-linux-gnu
Host CPU: athlon64
...
# llc-2.7 t.bc -o -
.file "t.bc"
.text
.globl init_post
.type init_post, at function
init_post: # @init_post
# BB#0: # %entry
subl $4, %esp
xorb %al, %al
testb %al, %al
movl $per_cpu__current_task, %eax
#APP
movl %fs:UNREACHABLE executed!
0 libLLVM-2.7.so.1 0xb7554c08
Stack dump:
0. Program arguments: llc-2.7 t.bc -o -
1. Running pass 'X86 AT&T-Style Assembly Printer' on function '@init_post'
Aborted
I will try and grab LLVM's Trunk and see if _another_ BR (and associated fix)
has resolved this.
Thanks,
Rob
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Wed Jan 5 10:46:05 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 5 Jan 2011 10:46:05 -0600 (CST)
Subject: [LLVMbugs] [Bug 5314] [Linux Kernel] Inline asm asmprinter crash
with P modifier
In-Reply-To:
References:
Message-ID: <20110105164605.BFB522A6C12E@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=5314
Anton Korobeynikov changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|REOPENED |RESOLVED
Resolution| |FIXED
--- Comment #6 from Anton Korobeynikov 2011-01-05 10:46:04 CST ---
(In reply to comment #5)
> I will try and grab LLVM's Trunk and see if _another_ BR (and associated fix)
> has resolved this.
Please reopen only in case if the problem can be reproduced with LLVM top of
the tree.
LLVM 2.7 is pretty ancient right 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 llvm.org Wed Jan 5 11:26:45 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 5 Jan 2011 11:26:45 -0600 (CST)
Subject: [LLVMbugs] [Bug 8915] New: Assertion failure when instantiating
member class of class template.
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8915
Summary: Assertion failure when instantiating member class of
class template.
Product: clang
Version: unspecified
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: C++
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: zaffanella at cs.unipr.it
CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com
The following test (from the g++ test suite) causes a clang assertion to
trigger.
template struct A {
struct B {
struct C {};
};
};
template void foo() {
class A::B::C X;
}
void bar() {
foo<0>();
}
clang: include/clang/AST/DeclTemplate.h:1283: void
clang::ClassTemplateSpecializationDecl::setPointOfInstantiation(clang::SourceLocation):
Assertion `Loc.isValid() && "point of instantiation must be valid!"' 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 llvm.org Wed Jan 5 11:38:45 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 5 Jan 2011 11:38:45 -0600 (CST)
Subject: [LLVMbugs] [Bug 8911] VMLA/VMLS codegen error
In-Reply-To:
References:
Message-ID: <20110105173845.AA9D02A6C124@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8911
Bob Wilson changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |WORKSFORME
--- Comment #5 from Bob Wilson 2011-01-05 11:38:45 CST ---
This works fine for me -- no assertion. We're not producing a VMLA
instruction, but as Anton already pointed out, that is intentional.
Renato, please reopen if you have a reproducible testcase. The IR excerpt you
copied into the PR is not usable and does not match the C source you provided.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Wed Jan 5 11:59:30 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 5 Jan 2011 11:59:30 -0600 (CST)
Subject: [LLVMbugs] [Bug 8916] New: clang tot takes a long,
long time (maybe forever) to compile Cryptopp's sha.cpp
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8916
Summary: clang tot takes a long, long time (maybe forever) to
compile Cryptopp's sha.cpp
Product: clang
Version: trunk
Platform: Macintosh
OS/Version: MacOS X
Status: NEW
Severity: normal
Priority: P
Component: LLVM Codegen
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: mclow at qualcomm.com
CC: llvmbugs at cs.uiuc.edu
Created an attachment (id=5962)
--> (http://llvm.org/bugs/attachment.cgi?id=5962)
Preprocessed source code for sha.cpp
When trying to compile CryptoPP , there was a file
that appeared to hang the compiler. I don't know for sure, but I got impatient
and killed it after 30 minutes.
This was done with clang/llvm built from revision 122872
attached file was created via:
clang++ -Wno-tautological-compare -DNDEBUG -g -DCRYPTOPP_DISABLE_ASM -E
sha.cpp > sha.i
Here are some timings:
$ time clang++ -Wno-tautological-compare -DNDEBUG -g -arch x86_64 -arch
i386 -DCRYPTOPP_DISABLE_ASM -pipe -fsyntax-only sha.cpp
real 6.234s
$ time clang++ -Wno-tautological-compare -DNDEBUG -g -arch x86_64 -arch
i386 -DCRYPTOPP_DISABLE_ASM -pipe -c sha.cpp
real 7.863s
$ time clang++ -Wno-tautological-compare -DNDEBUG -g -O1 -arch x86_64 -arch
i386 -DCRYPTOPP_DISABLE_ASM -pipe -c sha.cpp
real 17.882s
$ time clang++ -Wno-tautological-compare -DNDEBUG -g -O2 -arch x86_64 -arch
i386 -DCRYPTOPP_DISABLE_ASM -pipe -c sha.cpp
.... long, long time; I killed it after 30 minutes.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Wed Jan 5 12:03:23 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 5 Jan 2011 12:03:23 -0600 (CST)
Subject: [LLVMbugs] [Bug 8917] New: clang much slower than gcc compiling
CryptoPP sources
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8917
Summary: clang much slower than gcc compiling CryptoPP sources
Product: clang
Version: trunk
Platform: Macintosh
OS/Version: MacOS X
Status: NEW
Severity: normal
Priority: P
Component: C++
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: mclow at qualcomm.com
CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com
Created an attachment (id=5965)
--> (http://llvm.org/bugs/attachment.cgi?id=5965)
preprocessed source code for sha.cpp
When trying to compile CryptoPP , I noticed that the
compile times were much, much slower than gcc. Some of this is the optimizer,
but there's something going on in the front end, too.
This was done with clang/llvm built from revision 122872
attached file was created via:
clang++ -Wno-tautological-compare -DNDEBUG -g -DCRYPTOPP_DISABLE_ASM -E
sha.cpp > sha.i
Compare:
$ time clang++ -Wno-tautological-compare -DNDEBUG -g -arch x86_64 -arch
i386 -DCRYPTOPP_DISABLE_ASM -pipe -fsyntax-only sha.cpp
real 6.234s
$ time clang++ -Wno-tautological-compare -DNDEBUG -g -arch x86_64 -arch
i386 -DCRYPTOPP_DISABLE_ASM -pipe -c sha.cpp
real 7.863s
with:
$ time g++ -DNDEBUG -g -O2 -arch x86_64 -arch i386 -DCRYPTOPP_DISABLE_ASM -pipe
-c sha.cpp
real 1.662s
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Wed Jan 5 12:20:36 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 5 Jan 2011 12:20:36 -0600 (CST)
Subject: [LLVMbugs] [Bug 8911] VMLA/VMLS codegen error
In-Reply-To:
References:
Message-ID: <20110105182036.2CF7D2A6C12E@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8911
Bob Wilson changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|RESOLVED |REOPENED
Resolution|WORKSFORME |
--- Comment #9 from Bob Wilson 2011-01-05 12:20:35 CST ---
OK, I can reproduce it now. It's quite a bit different with EABI.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Wed Jan 5 12:36:28 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 5 Jan 2011 12:36:28 -0600 (CST)
Subject: [LLVMbugs] [Bug 8917] clang much slower than gcc compiling CryptoPP
sources
In-Reply-To:
References:
Message-ID: <20110105183628.F21CE2A6C12E@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8917
Marshall Clow (work) changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |INVALID
--- Comment #1 from Marshall Clow (work) 2011-01-05 12:36:28 CST ---
More timings; this one with a non-debug build.
Looks like pilot error to me.
$ time clang++ -Wno-tautological-compare -DNDEBUG -g -arch x86_64 -arch i386
-DCRYPTOPP_DISABLE_ASM -pipe -c sha.cpp
real 0m0.764s
$ time clang++ -Wno-tautological-compare -DNDEBUG -g -arch x86_64 -arch i386
-DCRYPTOPP_DISABLE_ASM -pipe -fsyntax-only sha.cpp
real 0m0.581s
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Wed Jan 5 14:59:52 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 5 Jan 2011 14:59:52 -0600 (CST)
Subject: [LLVMbugs] [Bug 8911] VMLA/VMLS codegen error
In-Reply-To:
References:
Message-ID: <20110105205952.1D2862A6C124@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8911
Bob Wilson changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|REOPENED |RESOLVED
Resolution| |WORKSFORME
--- Comment #10 from Bob Wilson 2011-01-05 14:59:51 CST ---
Clang doesn't know what to do with the "armv7-none-eabi" triple that you have
specified. It's warning about that:
clang: warning: unknown platform, assuming -mfloat-abi=soft
The "soft" ABI means that hardware floating-point instructions cannot be used.
Apparently no one has implemented that for vector operations. If you specify
-mfloat-abi=softfp, you'll get what you want.
If you want that triple to be recognized, you'll need to update the
Clang::AddARMTargetArgs() function in lib/Driver/Tools.cpp. I'm going to close
this again, since I don't think we need to keep a bug report open to track
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 llvm.org Wed Jan 5 15:00:11 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 5 Jan 2011 15:00:11 -0600 (CST)
Subject: [LLVMbugs] [Bug 8909] explicit move in move_iterator::operator*
In-Reply-To:
References:
Message-ID: <20110105210011.5D50C2A6C124@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8909
Nick Lewycky changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |llvmbugs at cs.uiuc.edu
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Wed Jan 5 15:00:45 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 5 Jan 2011 15:00:45 -0600 (CST)
Subject: [LLVMbugs] [Bug 8910] _LIBCPP_HAS_NO_ALWAYS_INLINE_VARIADICS and
g++-4.6
In-Reply-To:
References:
Message-ID: <20110105210045.5ADD32A6C12D@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8910
Nick Lewycky changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |llvmbugs at cs.uiuc.edu
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Wed Jan 5 16:29:06 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 5 Jan 2011 16:29:06 -0600 (CST)
Subject: [LLVMbugs] [Bug 8900] Shufflevector of doubles generates incorrect
code (using float shuff)
In-Reply-To:
References:
Message-ID: <20110105222906.A58232A6C12C@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8900
Chris Lattner changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|ASSIGNED |RESOLVED
Resolution| |FIXED
--- Comment #3 from Chris Lattner 2011-01-05 16:29:06 CST ---
Wow, that is really scary. Your patch looks great, applied in r122921. Thanks
a lot!
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Wed Jan 5 18:00:52 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 5 Jan 2011 18:00:52 -0600 (CST)
Subject: [LLVMbugs] [Bug 8918] New: LLVM generates incorrect __main call
with MinGW64
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8918
Summary: LLVM generates incorrect __main call with MinGW64
Product: libraries
Version: trunk
Platform: PC
OS/Version: other
Status: NEW
Severity: normal
Priority: P
Component: Backend: X86
AssignedTo: unassignedbugs at nondot.org
ReportedBy: sreeram at tachyontech.net
CC: llvmbugs at cs.uiuc.edu
Created an attachment (id=5966)
--> (http://llvm.org/bugs/attachment.cgi?id=5966)
Patch for fixing the above issue.
When used with MinGW64, LLVM generates a "calll __main" at the beginning of the
"main" function. The assembler complains about the invalid suffix for the
'call' instruction. The right instruction is "callq __main".
I've attached a patch that fixes 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 llvm.org Wed Jan 5 18:05:48 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 5 Jan 2011 18:05:48 -0600 (CST)
Subject: [LLVMbugs] [Bug 8919] New: LLVM generated code crashes if stack
frame > 4k on MinGW64
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8919
Summary: LLVM generated code crashes if stack frame > 4k on
MinGW64
Product: libraries
Version: trunk
Platform: PC
OS/Version: other
Status: NEW
Severity: normal
Priority: P
Component: Backend: X86
AssignedTo: unassignedbugs at nondot.org
ReportedBy: sreeram at tachyontech.net
CC: llvmbugs at cs.uiuc.edu
Created an attachment (id=5967)
--> (http://llvm.org/bugs/attachment.cgi?id=5967)
Patch for fixing the above issue.
LLVM incorrectly generates "_alloca" as the stack probing call. That works only
on MinGW32. On 64-bit, the function to call is "__chkstk". I've attached a
patch to 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 llvm.org Wed Jan 5 18:47:28 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 5 Jan 2011 18:47:28 -0600 (CST)
Subject: [LLVMbugs] [Bug 8918] LLVM generates incorrect __main call with
MinGW64
In-Reply-To:
References:
Message-ID: <20110106004728.AE5F22A6C12C@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8918
Bill Wendling changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
CC| |wendling at apple.com
Resolution| |FIXED
--- Comment #1 from Bill Wendling 2011-01-05 18:47:28 CST ---
Fixed r122933.
Thanks for the 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 llvm.org Wed Jan 5 18:50:50 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 5 Jan 2011 18:50:50 -0600 (CST)
Subject: [LLVMbugs] [Bug 8919] LLVM generated code crashes if stack frame >
4k on MinGW64
In-Reply-To:
References:
Message-ID: <20110106005050.C751F2A6C12C@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8919
Bill Wendling changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
CC| |wendling at apple.com
Resolution| |FIXED
--- Comment #1 from Bill Wendling 2011-01-05 18:50:50 CST ---
Applied r122934.
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 llvm.org Thu Jan 6 00:20:39 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 6 Jan 2011 00:20:39 -0600 (CST)
Subject: [LLVMbugs] [Bug 8883] constant folding doesn't simplify
constantexpr gep to constantint
In-Reply-To:
References:
Message-ID: <20110106062039.F2BFF2A6C12C@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8883
Chris Lattner changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
Summary|instcombine doesn't |constant folding doesn't
|simplify constantexpr |simplify constantexpr gep
|operands |to constantint
--- Comment #5 from Chris Lattner 2011-01-06 00:20:39 CST ---
Inliner is doing the right thing, constantfolding.cpp just didn't know how to
handle this case. Simple enough, implemented in r122950. Thanks for tracking
down what the root issue is Dan!
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Thu Jan 6 04:26:28 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 6 Jan 2011 04:26:28 -0600 (CST)
Subject: [LLVMbugs] [Bug 8920] New: clang-analyzer can't deal with FD_SET
and FD_ZERO
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8920
Summary: clang-analyzer can't deal with FD_SET and FD_ZERO
Product: clang
Version: 2.8
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: Static Analyzer
AssignedTo: kremenek at apple.com
ReportedBy: asn at cryptomilk.org
CC: llvmbugs at cs.uiuc.edu
clang-analyzer can't really deal with FD_SET and FD_ZERO. I think the problem
is cause these macros are assembler code.
Code:
rc = select(maxfd, &localset, NULL, NULL, timeout);
...
FD_ZERO(&localset2);
for (f = 0; f < maxfd; f++) {
if (FD_ISSET(f, readfds) && FD_ISSET(f, &localset)) {
FD_SET(f, &localset2);
}
}
http://test.libssh.org/clang-analyzer/
Result:
Within the expansion of the macro 'FD_SET':
The left expression of the compound assignment is an uninitialized value. The
computed value will also be garbage.
Test code:
If you're looking for some test code see the EXAMPLE section in the select(2)
manpage.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Thu Jan 6 04:39:39 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 6 Jan 2011 04:39:39 -0600 (CST)
Subject: [LLVMbugs] [Bug 8919] LLVM generated code crashes if stack frame >
4k on MinGW64
In-Reply-To:
References:
Message-ID: <20110106103939.49F492A6C12C@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8919
NAKAMURA Takumi changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|RESOLVED |REOPENED
CC| |geek4civic at gmail.com
Resolution|FIXED |
OS/Version|other |Windows XP
--- Comment #3 from NAKAMURA Takumi 2011-01-06 04:39:38 CST ---
I have to reopen this. Sreeram's patch would be effective only on TDM.
I think it would not be needed to revert r122934. Prior implementation (_alloca
by %eax) was bogus, too.
in http://tdm-gcc.tdragon.net/development
> TDM64 edition only: Includes a patch to remove a leading underscore from the 64-bit cygwin.asm symbols "__chkstk" and "_alloca", allowing them to be exported with the same rule as the 32-bit versions after underscoring rules are applied.
The symbol "__chkstk" is unique to tdm. gcc's cygwin.asm and mingw-w64-gcc (and
I) assume "___chkstk".
ps. see also bug 8777 ;)
Anton,
> Also, I believe mingw-w64 provided *both* __alloca and
> __chkstk.
We may prefer ___chkstk on win64 rather than __alloca.
___chkstk does not modify %rcx %rdx %r8 and %r9. (__alloca does not)
Or shall we implement our chkstk? :D
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Thu Jan 6 09:22:09 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 6 Jan 2011 09:22:09 -0600 (CST)
Subject: [LLVMbugs] [Bug 8921] New: ARM/Thumb interoperability broken in v4T
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8921
Summary: ARM/Thumb interoperability broken in v4T
Product: libraries
Version: trunk
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: Backend: ARM
AssignedTo: unassignedbugs at nondot.org
ReportedBy: rengolin at gmail.com
CC: llvmbugs at cs.uiuc.edu
The following code generates wrong assembly when compiled to ARM:
extern int c();
int b()
{
return c()+1;
}
int a()
{
return b();
}
EABI requires ARM/Thumb interoperability by default (unless you know *exactly*
what you're doing, which is not the case). LLVM back-end (either through clang
or clang+llc) is generating b() as:
b:
push {r11, lr}
mov r11, sp
bl c
add r0, r0, #1
ldmia sp!, {r11, pc}
which doesn't take into account the instruction set change. The correct way to
return is "bx lr" as a default case. We're having problems when linking against
7TDMI libraries, since it has both ARM and Thumb functions.
The command lines:
$ clang -ccc-host-triple armv4t-none-eabi -mfloat-abi=softfp -S test.c
or
$ clang -ccc-host-triple armv4t-none-eabi -mfloat-abi=softfp -emit-llvm -S
test.c -o test.ll
$ llc -march=arm test.ll -o -
Attaching source, IR and assembly generated with trunk clang+llc.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Thu Jan 6 09:43:45 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 6 Jan 2011 09:43:45 -0600 (CST)
Subject: [LLVMbugs] [Bug 8922] New: Win64 Does not codegen correctly alloca
align 64
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8922
Summary: Win64 Does not codegen correctly alloca align 64
Product: libraries
Version: trunk
Platform: PC
OS/Version: Windows XP
Status: NEW
Severity: normal
Priority: P
Component: Backend: X86
AssignedTo: unassignedbugs at nondot.org
ReportedBy: zvi.rackover at intel.com
CC: llvmbugs at cs.uiuc.edu
Created an attachment (id=5971)
--> (http://llvm.org/bugs/attachment.cgi?id=5971)
Reproducer test
The following test case produces the assertion below when running on the Win64
subtarget:
define noalias i32* @factorial() nounwind readnone {
entry:
%a = alloca i32, align 64
ret i32* %a
}
Assertion failed: (-(Offset + StackSize)) % Align == 0, file
..\..\..\..\lib\Target\X86\X86FrameInfo.cpp, line 867
Explanation:
Win64 CodeGen adds an extra 32 bytes to frame size, to be used as a Red Zone
for potential calls. This addition ruins the stack alignment in case the
required alignment is greater than 32 bytes.
Proposed fix is to add max(32, MaxAlignment) bytes instead.
Attached are the testcase and 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 llvm.org Thu Jan 6 10:57:46 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 6 Jan 2011 10:57:46 -0600 (CST)
Subject: [LLVMbugs] [Bug 8485] error: unknown target CPU 'cortex-m3'
In-Reply-To:
References:
Message-ID: <20110106165746.828E52A6C12C@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8485
Bob Wilson changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
--- Comment #6 from Bob Wilson 2011-01-06 10:57:46 CST ---
Thanks for the reminder. I applied the patch as svn r122965.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Thu Jan 6 13:26:54 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 6 Jan 2011 13:26:54 -0600 (CST)
Subject: [LLVMbugs] [Bug 8921] ARM/Thumb interoperability broken in v4T
In-Reply-To:
References:
Message-ID: <20110106192655.00E552A6C12D@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8921
Bob Wilson changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
--- Comment #5 from Bob Wilson 2011-01-06 13:26:54 CST ---
Fixed in svn 122970
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Thu Jan 6 13:49:22 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 6 Jan 2011 13:49:22 -0600 (CST)
Subject: [LLVMbugs] [Bug 8582] [legalize crash] Assertion
`N.getNode()->getNodeId() != NewNode && "Mapped to new node!"' failed.
In-Reply-To:
References:
Message-ID: <20110106194922.5D1F52A6C12E@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8582
John Regehr changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
--- Comment #8 from John Regehr 2011-01-06 13:49:21 CST ---
A couple hundred 1000 testcases failed to turn up anything triggering this on
x86 or x64, so let's call it good for 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 llvm.org Thu Jan 6 15:40:45 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 6 Jan 2011 15:40:45 -0600 (CST)
Subject: [LLVMbugs] [Bug 8913] clang++ asserts when compiling vmac.cpp from
CryptoPP 5.6.1
In-Reply-To:
References:
Message-ID: <20110106214045.744FE2A6C12D@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8913
Devang Patel changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
--- Comment #2 from Devang Patel 2011-01-06 15:40:44 CST ---
Fixed in r122971 and r122972.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Thu Jan 6 17:45:29 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 6 Jan 2011 17:45:29 -0600 (CST)
Subject: [LLVMbugs] [Bug 8629] clang++ fails an assertion
In-Reply-To:
References:
Message-ID: <20110106234529.192C42A6C12E@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8629
Elias Pipping changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|RESOLVED |REOPENED
Resolution|FIXED |
--- Comment #5 from Elias Pipping 2011-01-06 17:45:27 CST ---
This still fails for me.
bogus% clang++ --version
clang version 2.9 (trunk 122981)
Target: x86_64-pc-linux-gnu
Thread model: posix
bogus%
bogus% clang++ -c dgfparserblocks.ii
dgfparserblocks.ii:25:14: warning: 'template' keyword outside of a template
[-Wc++0x-extensions]
simplex_.template initialize<0>();
^~~~~~~~
dgfparserblocks.ii:9:24: warning: template argument uses local type
'GenericReferenceElement::type'
[-Wlocal-type-template-args]
typedef Initialize Init;
^~~~~~~~
dgfparserblocks.ii:15:5: note: in instantiation of function template
specialization
'GenericReferenceElement::initializeTopology::type>'
requested here
initializeTopology();
^
dgfparserblocks.ii:25:5: note: in instantiation of function template
specialization
'GenericReferenceElement::initialize<0>' requested here
simplex_.template initialize<0>();
^
clang:
/var/tmp/paludis/build/dev-lang-llvm-scm/work/llvm-scm/tools/clang/lib/Sema/../../include/clang/AST/DeclTemplate.h:1287:
void
clang::ClassTemplateSpecializationDecl::setPointOfInstantiation(clang::SourceLocation):
Assertion `Loc.isValid() && "point of instantiation must be valid!"' failed.
0 libLLVM-2.9svn.so 0x00007f44f88d2dcf
1 libLLVM-2.9svn.so 0x00007f44f88d3b61
2 libpthread.so.0 0x00007f44f7e19fe0
3 libc.so.6 0x00007f44f71376a5 gsignal + 53
4 libc.so.6 0x00007f44f71389a5 abort + 389
5 libc.so.6 0x00007f44f7130235 __assert_fail + 245
6 clang 0x0000000000969a52
clang::Sema::InstantiateClass(clang::SourceLocation, clang::CXXRecordDecl*,
clang::CXXRecordDecl*, clang::MultiLevelTemplateArgumentList const&,
clang::TemplateSpecializationKind, bool) + 1666
7 clang 0x000000000096a4cb
clang::Sema::InstantiateClassTemplateSpecialization(clang::SourceLocation,
clang::ClassTemplateSpecializationDecl*, clang::TemplateSpecializationKind,
bool) + 1147
8 clang 0x000000000099a4a9
clang::Sema::RequireCompleteType(clang::SourceLocation, clang::QualType,
clang::PartialDiagnostic const&, std::pair) + 1081
9 clang 0x000000000099ae94
clang::Sema::RequireCompleteType(clang::SourceLocation, clang::QualType,
clang::PartialDiagnostic const&) + 84
10 clang 0x00000000007bd46a
clang::Sema::RequireCompleteDeclContext(clang::CXXScopeSpec&,
clang::DeclContext*) + 378
11 clang 0x000000000092c080
clang::Sema::LookupTemplateName(clang::LookupResult&, clang::Scope*,
clang::CXXScopeSpec&, clang::QualType, bool, bool&) + 2032
12 clang 0x000000000092cace
clang::Sema::isTemplateName(clang::Scope*, clang::CXXScopeSpec&, bool,
clang::UnqualifiedId&, clang::OpaquePtr, bool,
clang::OpaquePtr&, bool&) + 366
13 clang 0x000000000092ce2e
clang::Sema::ActOnDependentTemplateName(clang::Scope*, clang::SourceLocation,
clang::CXXScopeSpec&, clang::UnqualifiedId&, clang::OpaquePtr,
bool, clang::OpaquePtr&) + 206
14 clang 0x000000000095a7e6
15 clang 0x000000000096d54b
16 clang 0x0000000000961e2b
17 clang 0x0000000000964bd0
18 clang 0x000000000096c770
19 clang 0x000000000096ca7e
20 clang 0x0000000000974d78
21 clang 0x00000000009602e5
22 clang 0x00000000009766b0
23 clang 0x0000000000976863
24 clang 0x000000000095fa85
25 clang 0x0000000000971f45
26 clang 0x0000000000972ca4
27 clang 0x000000000097209d
28 clang 0x000000000097354d clang::Sema::SubstStmt(clang::Stmt*,
clang::MultiLevelTemplateArgumentList const&) + 45
29 clang 0x00000000009866f6
clang::Sema::InstantiateFunctionDefinition(clang::SourceLocation,
clang::FunctionDecl*, bool, bool) + 790
30 clang 0x0000000000986c4d
clang::Sema::PerformPendingInstantiations(bool) + 365
31 clang 0x000000000098699f
clang::Sema::InstantiateFunctionDefinition(clang::SourceLocation,
clang::FunctionDecl*, bool, bool) + 1471
32 clang 0x0000000000986c4d
clang::Sema::PerformPendingInstantiations(bool) + 365
33 clang 0x00000000007aa8f2
clang::Sema::ActOnEndOfTranslationUnit() + 242
34 clang 0x0000000000771409
clang::Parser::ParseTopLevelDecl(clang::OpaquePtr&) + 201
35 clang 0x0000000000758a23 clang::ParseAST(clang::Sema&, bool) +
147
36 clang 0x00000000006305b4 clang::CodeGenAction::ExecuteAction() +
68
37 clang 0x000000000052e863
clang::CompilerInstance::ExecuteAction(clang::FrontendAction&) + 355
38 clang 0x000000000050c9ac
clang::ExecuteCompilerInvocation(clang::CompilerInstance*) + 1276
39 clang 0x0000000000504f40 cc1_main(char const**, char const**,
char const*, void*) + 672
40 clang 0x000000000050b9aa main + 4458
41 libc.so.6 0x00007f44f7123c7d __libc_start_main + 253
42 clang 0x0000000000503699
Stack dump:
0. Program arguments: /usr/bin/clang -cc1 -triple x86_64-pc-linux-gnu
-emit-obj -mrelax-all -disable-free -main-file-name dgfparserblocks.ii
-mrelocation-model static -mdisable-fp-elim -masm-verbose -mconstructor-aliases
-munwind-tables -target-cpu x86-64 -target-linker-version 2.21.51.0.2.20101206
-resource-dir /usr/bin/../lib/clang/2.9 -ferror-limit 19 -fmessage-length 118
-fexceptions -fgnu-runtime -fdiagnostics-show-option -fcolor-diagnostics -o
dgfparserblocks.o -x c++-cpp-output dgfparserblocks.ii
1. parser at end of file
2. dgfparserblocks.ii:12:24: instantiating function definition
'initialize'
3. dgfparserblocks.ii:7:33: instantiating function definition
'initializeTopology'
clang: error: unable to execute command: Aborted
clang: error: clang frontend command failed due to signal 1 (use -v to see
invocation)
bogus%
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Thu Jan 6 17:54:57 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 6 Jan 2011 17:54:57 -0600 (CST)
Subject: [LLVMbugs] [Bug 8923] New:
DAGTypeLegalizer::PerformExpensiveChecks(): Assertion
`NewVal.getNode()->getNodeId() != NewNode && "ReplacedValues maps to a new
node!"' failed.
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8923
Summary: DAGTypeLegalizer::PerformExpensiveChecks(): Assertion
`NewVal.getNode()->getNodeId() != NewNode &&
"ReplacedValues maps to a new node!"' failed.
Product: libraries
Version: 2.8
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: Common Code Generator Code
AssignedTo: unassignedbugs at nondot.org
ReportedBy: nadav.rotem at intel.com
CC: llvmbugs at cs.uiuc.edu
llc: lib/CodeGen/SelectionDAG/LegalizeTypes.cpp:106: void
llvm::DAGTypeLegalizer::PerformExpensiveChecks(): Assertion
`NewVal.getNode()->getNodeId() != NewNode && "ReplacedValues maps to a new
node!"' failed.
We hit an assert for the code below (both Linux and Windows):
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-f80:128:128-v64:64:64-v128:128:128-a0:0:64-f80:32:32
n8:16:32"
target triple = "i686-pc-win32"
define void @test() nounwind {
%i17 = icmp eq <4 x i8> undef, zeroinitializer
%cond = extractelement <4 x i1> %i17, i32 0
%_comp = select i1 %cond, i8 0, i8 undef
%merge = insertelement <4 x i8> undef, i8 %_comp, i32 0
%cond3 = extractelement <4 x i1> %i17, i32 1
%_comp4 = select i1 %cond3, i8 0, i8 undef
%merge5 = insertelement <4 x i8> %merge, i8 %_comp4, i32 1
%cond8 = extractelement <4 x i1> %i17, i32 2
%_comp9 = select i1 %cond8, i8 0, i8 undef
%m387 = insertelement <4 x i8> %merge5, i8 %_comp9, i32 2
store <4 x i8> %m387, <4 x i8>* undef
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 llvm.org Thu Jan 6 17:59:49 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 6 Jan 2011 17:59:49 -0600 (CST)
Subject: [LLVMbugs] [Bug 8924] New:
/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp:784:
llvm::SDValue::SelectionDAGLegalize::LegalizeOp(llvm::SDValue):
Assertion `getTypeAction(Node->getValueType(i)) == Legal && "Unexpected
illegal type!"' failed.
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8924
Summary: /lib/CodeGen/SelectionDAG/LegalizeDAG.cpp:784:
llvm::SDValue::SelectionDAGLegalize::Legalize
Op(llvm::SDValue): Assertion
`getTypeAction(Node->getValueType(i)) == Legal &&
"Unexpected illegal type!"' failed.
Product: libraries
Version: 2.8
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: Common Code Generator Code
AssignedTo: unassignedbugs at nondot.org
ReportedBy: nadav.rotem at intel.com
CC: llvmbugs at cs.uiuc.edu
We hit the assertion above on the code below. Tested on LLVM 2.8 on Windows and
Linux.
define void @m387() nounwind {
%1 = load <4 x i64> addrspace(1)* undef, align 32
%temp.vect37 = shufflevector <4 x i64> undef, <4 x i64> %1, <4 x i32>
%2 = mul <4 x i64> zeroinitializer, %temp.vect37
%3 = shufflevector <4 x i64> undef, <4 x i64> %2, <4 x i32>
%4 = shufflevector <4 x i64> %3, <4 x i64> undef, <4 x i32>
store <4 x i64> %4, <4 x i64> addrspace(1)* undef, align 32
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 llvm.org Thu Jan 6 18:09:53 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 6 Jan 2011 18:09:53 -0600 (CST)
Subject: [LLVMbugs] [Bug 8923] DAGTypeLegalizer::PerformExpensiveChecks():
Assertion `NewVal.getNode()->getNodeId() != NewNode && "ReplacedValues maps
to a new node!"' failed.
In-Reply-To:
References:
Message-ID: <20110107000953.EB9132A6C12C@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8923
Nadav Rotem changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
--- Comment #2 from Nadav Rotem 2011-01-06 18:09:53 CST ---
This is a duplicate of 8582. Duncan Sands reported that he commited the fix for
this issue.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Thu Jan 6 18:10:59 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 6 Jan 2011 18:10:59 -0600 (CST)
Subject: [LLVMbugs] [Bug 8923] DAGTypeLegalizer::PerformExpensiveChecks():
Assertion `NewVal.getNode()->getNodeId() != NewNode && "ReplacedValues maps
to a new node!"' failed.
In-Reply-To:
References:
Message-ID: <20110107001100.0D5EC2A6C12C@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8923
Nadav Rotem changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|RESOLVED |REOPENED
Resolution|FIXED |
--- Comment #3 from Nadav Rotem 2011-01-06 18:10:59 CST ---
My mistake. Duncan Sands fixed another issue. This one is still open.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Thu Jan 6 18:16:06 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 6 Jan 2011 18:16:06 -0600 (CST)
Subject: [LLVMbugs] [Bug 8923] DAGTypeLegalizer::PerformExpensiveChecks():
Assertion `NewVal.getNode()->getNodeId() != NewNode && "ReplacedValues maps
to a new node!"' failed.
In-Reply-To:
References:
Message-ID: <20110107001606.B67622A6C12C@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8923
Nadav Rotem changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|REOPENED |RESOLVED
Resolution| |FIXED
--- Comment #4 from Nadav Rotem 2011-01-06 18:16:06 CST ---
Fixed in
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20110103/114745.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 llvm.org Thu Jan 6 20:03:50 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 6 Jan 2011 20:03:50 -0600 (CST)
Subject: [LLVMbugs] [Bug 8925] New: Unnecessary munging of stack pointer
with guarantee tail calls
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8925
Summary: Unnecessary munging of stack pointer with guarantee
tail calls
Product: libraries
Version: trunk
Platform: PC
OS/Version: All
Status: NEW
Severity: normal
Priority: P
Component: Backend: X86
AssignedTo: unassignedbugs at nondot.org
ReportedBy: arplynn at gmail.com
CC: llvmbugs at cs.uiuc.edu
In the following test case:
define fastcc i32* @force.ri32(i32* %x) nounwind {
entry:
%nz = icmp eq i32* %x, null
br i1 %nz, label %Limpl1.pre, label %Limpl0.pre
Limpl0.pre: ; preds = %entry
ret i32* %x
Limpl1.pre: ; preds = %entry
tail call fastcc void @pr.error() noreturn nounwind
unreachable
}
declare fastcc void @pr.error() noreturn nounwind
LLVM currently generates:
_force.ri32: ## @force.ri32
## BB#0: ## %entry
subq $8, %rsp
testq %rdi, %rdi
je LBB0_2
## BB#1: ## %Limpl0.pre
movq %rdi, %rax
addq $8, %rsp
ret $8
LBB0_2: ## %Limpl1.pre
addq $8, %rsp
jmp _pr.error ## TAILCALL
The instructions that change %rsp are unnecessary.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Thu Jan 6 23:09:57 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 6 Jan 2011 23:09:57 -0600 (CST)
Subject: [LLVMbugs] [Bug 8411] ARM back end is not very clever with
shufflevector
In-Reply-To:
References:
Message-ID: <20110107050957.4CB1D2A6C12D@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8411
Bob Wilson changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
--- Comment #2 from Bob Wilson 2011-01-06 23:09:56 CST ---
I committed the Tim Northover's extract_subvector patch which fixes this. svn
122995. I'm also going to change clang's vget_low and vget_high intrinsics to
use __builtin_shufflevector.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Fri Jan 7 04:21:20 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 7 Jan 2011 04:21:20 -0600 (CST)
Subject: [LLVMbugs] [Bug 8924] /lib/CodeGen/SelectionDAG/LegalizeDAG.cpp:784:
llvm::SDValue::SelectionDAGLegalize::LegalizeOp(llvm::SDValue):
Assertion `getTypeAction(Node->getValueType(i)) == Legal && "Unexpected
illegal type!"' failed.
In-Reply-To:
References:
Message-ID: <20110107102120.84B182A6C12C@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8924
Nadav Rotem changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
--- Comment #3 from Nadav Rotem 2011-01-07 04:21:19 CST ---
I can only reproduce this on LLVM2.8. I guess that it is fixed in 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 llvm.org Fri Jan 7 06:34:48 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 7 Jan 2011 06:34:48 -0600 (CST)
Subject: [LLVMbugs] [Bug 8926] New: clang++ assert while compiling c++
source file
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8926
Summary: clang++ assert while compiling c++ source file
Product: clang
Version: trunk
Platform: Macintosh
OS/Version: MacOS X
Status: NEW
Severity: normal
Priority: P
Component: -New Bugs
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: felix.ritter at mevis.fraunhofer.de
CC: llvmbugs at cs.uiuc.edu
clang version 2.9 (trunk 122999)
Target: x86_64-apple-darwin10
Thread model: posix
I've attached the preprocessed source file, that crashes when compiled with the
trunk clang++:
clang++ -c PythonQtSlot.ii
Assertion failed: (!HasCachedLinkage || Linkage(CachedLinkage) ==
LI.linkage()), function getLinkageAndVisibility, file Decl.cpp, line 587.
0 clang 0x00000001014cb4c2 PrintStackTrace(void*) + 34
1 clang 0x00000001014cc313 SignalHandler(int) + 707
2 libSystem.B.dylib 0x00007fff8596a67a _sigtramp + 26
3 libSystem.B.dylib 0x00007fff5fbf9cf8 _sigtramp + 3660117656
4 clang 0x000000010001a2b2 __assert_rtn + 66
5 clang 0x00000001007d3695
clang::NamedDecl::getLinkageAndVisibility() const + 149
6 clang 0x00000001002adbd7
clang::CodeGen::CodeGenModule::SetFunctionAttributes(clang::CodeGen::GlobalDecl,
llvm::Function*, bool) + 87
7 clang 0x00000001002b1272
clang::CodeGen::CodeGenModule::GetOrCreateLLVMFunction(llvm::StringRef,
llvm::Type const*, clang::CodeGen::GlobalDecl) + 914
8 clang 0x00000001001f508e
clang::CodeGen::CodeGenFunction::EmitCXXMemberCallExpr(clang::CXXMemberCallExpr
const*, clang::CodeGen::ReturnValueSlot) + 1630
9 clang 0x00000001001dc7b8
clang::CodeGen::CodeGenFunction::EmitCallExpr(clang::CallExpr const*,
clang::CodeGen::ReturnValueSlot) + 104
10 clang 0x00000001002171ce clang::StmtVisitor<(anonymous
namespace)::ScalarExprEmitter, llvm::Value*>::Visit(clang::Stmt*) + 3374
11 clang 0x00000001002181b0 (anonymous
namespace)::ScalarExprEmitter::Visit(clang::Expr*) + 272
12 clang 0x0000000100218239
clang::CodeGen::CodeGenFunction::EmitScalarExpr(clang::Expr const*, bool) + 121
13 clang 0x00000001001dc4e6
clang::CodeGen::CodeGenFunction::EmitAnyExpr(clang::Expr const*,
clang::CodeGen::AggValueSlot, bool) + 86
14 clang 0x00000001001dc64c
clang::CodeGen::CodeGenFunction::EmitAnyExprToTemp(clang::Expr const*) + 76
15 clang 0x000000010019ed5e
clang::CodeGen::CodeGenFunction::EmitCallArg(clang::Expr const*,
clang::QualType) + 62
16 clang 0x00000001001ea1f8 void
clang::CodeGen::CodeGenFunction::EmitCallArgs(llvm::SmallVector, 16u>&, clang::FunctionProtoType const*,
clang::ConstExprIterator, clang::ConstExprIterator) + 440
17 clang 0x00000001001dace0
clang::CodeGen::CodeGenFunction::EmitCall(clang::QualType, llvm::Value*,
clang::CodeGen::ReturnValueSlot, clang::ConstExprIterator,
clang::ConstExprIterator, clang::Decl const*) + 352
18 clang 0x00000001001dc98a
clang::CodeGen::CodeGenFunction::EmitCallExpr(clang::CallExpr const*,
clang::CodeGen::ReturnValueSlot) + 570
19 clang 0x000000010021715f clang::StmtVisitor<(anonymous
namespace)::ScalarExprEmitter, llvm::Value*>::Visit(clang::Stmt*) + 3263
20 clang 0x00000001002181b0 (anonymous
namespace)::ScalarExprEmitter::Visit(clang::Expr*) + 272
21 clang 0x0000000100218239
clang::CodeGen::CodeGenFunction::EmitScalarExpr(clang::Expr const*, bool) + 121
22 clang 0x000000010027d08d
clang::CodeGen::CodeGenFunction::EmitReturnStmt(clang::ReturnStmt const&) + 173
23 clang 0x000000010027eb94
clang::CodeGen::CodeGenFunction::EmitStmt(clang::Stmt const*) + 964
24 clang 0x00000001002816ab
clang::CodeGen::CodeGenFunction::EmitCompoundStmt(clang::CompoundStmt const&,
bool, clang::CodeGen::AggValueSlot) + 283
25 clang 0x0000000100281a42
clang::CodeGen::CodeGenFunction::EmitSimpleStmt(clang::Stmt const*) + 498
26 clang 0x000000010027e800
clang::CodeGen::CodeGenFunction::EmitStmt(clang::Stmt const*) + 48
27 clang 0x00000001002813eb
clang::CodeGen::CodeGenFunction::EmitIfStmt(clang::IfStmt const&) + 459
28 clang 0x000000010027ec2f
clang::CodeGen::CodeGenFunction::EmitStmt(clang::Stmt const*) + 1119
29 clang 0x00000001002816ab
clang::CodeGen::CodeGenFunction::EmitCompoundStmt(clang::CompoundStmt const&,
bool, clang::CodeGen::AggValueSlot) + 283
30 clang 0x0000000100281a42
clang::CodeGen::CodeGenFunction::EmitSimpleStmt(clang::Stmt const*) + 498
31 clang 0x000000010027e800
clang::CodeGen::CodeGenFunction::EmitStmt(clang::Stmt const*) + 48
32 clang 0x00000001002813eb
clang::CodeGen::CodeGenFunction::EmitIfStmt(clang::IfStmt const&) + 459
33 clang 0x000000010027ec2f
clang::CodeGen::CodeGenFunction::EmitStmt(clang::Stmt const*) + 1119
34 clang 0x0000000100281457
clang::CodeGen::CodeGenFunction::EmitIfStmt(clang::IfStmt const&) + 567
35 clang 0x000000010027ec2f
clang::CodeGen::CodeGenFunction::EmitStmt(clang::Stmt const*) + 1119
36 clang 0x00000001002816ab
clang::CodeGen::CodeGenFunction::EmitCompoundStmt(clang::CompoundStmt const&,
bool, clang::CodeGen::AggValueSlot) + 283
37 clang 0x0000000100281a42
clang::CodeGen::CodeGenFunction::EmitSimpleStmt(clang::Stmt const*) + 498
38 clang 0x000000010027e800
clang::CodeGen::CodeGenFunction::EmitStmt(clang::Stmt const*) + 48
39 clang 0x00000001002a6d2f
clang::CodeGen::CodeGenFunction::GenerateCode(clang::CodeGen::GlobalDecl,
llvm::Function*) + 1071
40 clang 0x00000001002b3291
clang::CodeGen::CodeGenModule::EmitGlobalFunctionDefinition(clang::CodeGen::GlobalDecl)
+ 753
41 clang 0x00000001002b62eb
clang::CodeGen::CodeGenModule::EmitGlobalDefinition(clang::CodeGen::GlobalDecl)
+ 395
42 clang 0x00000001002b6466
clang::CodeGen::CodeGenModule::EmitGlobal(clang::CodeGen::GlobalDecl) + 150
43 clang 0x00000001002b70a0
clang::CodeGen::CodeGenModule::EmitTopLevelDecl(clang::Decl*) + 912
44 clang 0x00000001002da20c (anonymous
namespace)::CodeGeneratorImpl::HandleTopLevelDecl(clang::DeclGroupRef) + 60
45 clang 0x00000001002a0a2b (anonymous
namespace)::BackendConsumer::HandleTopLevelDecl(clang::DeclGroupRef) + 155
46 clang 0x00000001002e6206 clang::ParseAST(clang::Sema&, bool) +
182
47 clang 0x00000001002a130c clang::CodeGenAction::ExecuteAction() +
60
48 clang 0x0000000100052c89
clang::CompilerInstance::ExecuteAction(clang::FrontendAction&) + 393
49 clang 0x0000000100024392
clang::ExecuteCompilerInvocation(clang::CompilerInstance*) + 1602
50 clang 0x000000010001c01a cc1_main(char const**, char const**,
char const*, void*) + 586
51 clang 0x00000001000233c4 main + 5524
52 clang 0x000000010001a994 start + 52
Stack dump:
0. Program arguments: /Users/ritter/Temp/llvm/Release+Asserts/bin/clang -cc1
-triple x86_64-apple-darwin10.0.0 -emit-obj -mrelax-all -disable-free
-main-file-name PythonQtSlot.ii -pic-level 1 -mdisable-fp-elim -masm-verbose
-munwind-tables -target-cpu core2 -target-linker-version 97.17 -resource-dir
/Users/ritter/Temp/llvm/Release+Asserts/bin/../lib/clang/2.9 -ferror-limit 19
-fmessage-length 80 -stack-protector 1 -fblocks -fexceptions
-fdiagnostics-show-option -fcolor-diagnostics -o PythonQtSlot.o -x
c++-cpp-output PythonQtSlot.ii
1. src/PythonQtSlot.cpp:249:1: current parser token 'PyObject'
2. src/PythonQtSlot.cpp:197:11: LLVM IR generation of declaration
'PythonQtSlotFunction_Call'
3. src/PythonQtSlot.cpp:197:11: Generating code for declaration
'PythonQtSlotFunction_Call'
4. src/PythonQtSlot.cpp:198:1: LLVM IR generation of compound statement
('{}')
5. src/PythonQtSlot.cpp:210:64: LLVM IR generation of compound statement
('{}')
6. src/PythonQtSlot.cpp:212:35: LLVM IR generation of compound statement
('{}')
clang: error: unable to execute command: Illegal instruction
clang: error: clang frontend command failed due to signal 1 (use -v to see
invocation)
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Fri Jan 7 08:25:20 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 7 Jan 2011 08:25:20 -0600 (CST)
Subject: [LLVMbugs] [Bug 8927] New: Invalid constant merging
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8927
Summary: Invalid constant merging
Product: libraries
Version: trunk
Platform: PC
OS/Version: All
Status: NEW
Severity: normal
Priority: P
Component: Interprocedural Optimizations
AssignedTo: unassignedbugs at nondot.org
ReportedBy: rafael.espindola at gmail.com
CC: llvmbugs at cs.uiuc.edu
The C standard says that
Two pointers compare equal if and only if both are null pointers, both are
pointers to the same object (including a pointer to an object and a subobject
at its beginning) or function, both are pointers to one past the last element
of the same array object, or one is a pointer to one past the end of one array
object and the other is a pointer to the start of a different array object that
happens to immediately follow the first array object in the address space.
With -O0 and some cleanup passes clang compiles
struct foobar {
int x;
};
static const struct foobar* foo() {
static const struct foobar d = { 0 };
return &d;
}
static const struct foobar* bar() {
static const struct foobar d = { 0 };
return &d;
}
int zed(const struct foobar *a, const struct foobar *b);
int main() {
zed(foo(), bar());
}
to
%struct.foobar = type { i32 }
@bar.d = internal constant %struct.foobar zeroinitializer, align 4
@foo.d = internal constant %struct.foobar zeroinitializer, align 4
define i32 @main() nounwind ssp {
entry:
%call2 = tail call i32 @zed(%struct.foobar* @foo.d, %struct.foobar* @bar.d)
nounwind
ret i32 0
}
declare i32 @zed(%struct.foobar*, %struct.foobar*)
but constmerge (which is included in -O2) transforms it to
%struct.foobar = type { i32 }
@bar.d = internal constant %struct.foobar zeroinitializer, align 4
define i32 @main() nounwind ssp {
entry:
%call2 = tail call i32 @zed(%struct.foobar* @bar.d, %struct.foobar* @bar.d)
nounwind
ret i32 0
}
declare i32 @zed(%struct.foobar*, %struct.foobar*)
Since the function zed in unknown, it might compare the pointers and now it
will get a different result.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Fri Jan 7 11:45:56 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 7 Jan 2011 11:45:56 -0600 (CST)
Subject: [LLVMbugs] [Bug 8928] New: Assertion failed: (T && "getTerminator
returned null!"), function SuccIterator,
file /data/home/rdivacky/llvm/include/llvm/Support/CFG.h, line 105.
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8928
Summary: Assertion failed: (T && "getTerminator returned
null!"), function SuccIterator, file
/data/home/rdivacky/llvm/include/llvm/Support/CFG.h,
line 105.
Product: libraries
Version: trunk
Platform: PC
OS/Version: FreeBSD
Status: NEW
Severity: normal
Priority: P
Component: Global Analyses
AssignedTo: unassignedbugs at nondot.org
ReportedBy: rdivacky at freebsd.org
CC: llvmbugs at cs.uiuc.edu
Assertion failed: (T && "getTerminator returned null!"), function SuccIterator,
file /data/home/rdivacky/llvm/include/llvm/Support/CFG.h, line 105.
Stack dump:
0. Program arguments: opt boot2.bc -adce -basicaa -codegenprepare
-constprop -die -domtree -early-cse -globaldce -inline
-insert-optimal-edge-profiling -internalize -ipconstprop -jump-threading -lcssa
-lda -libcall-aa -licm -live-values -loop-deletion -loop-extract-single
-loop-idiom -loop-rotate -loop-unroll -loop-unswitch -loops -loweratomic
-mem2reg -mergefunc -no-aa -postdomfrontier -postdomtree -profile-estimator
-simplify-libcalls -simplifycfg -strip -strip-debug-declare -strip-nondebug
-tailcallelim -tbaa
1. Running pass 'Function Pass Manager' on module 'boot2.bc'.
2. Running pass 'Profiling information estimator' on function '@lookup'
and
Assertion failed: (T && "getTerminator returned null!"), function SuccIterator,
file /data/home/rdivacky/llvm/include/llvm/Support/CFG.h, line 105.
Stack dump:
0. Program arguments: opt boot2.bc -adce -always-inline -constprop
-correlated-propagation -dce -domfrontier -dse -early-cse -extract-blocks
-functionattrs -globalsmodref-aa -gvn -indvars -inline -instnamer -intervals
-ipsccp -iv-users -jump-threading -lcssa -live-values -loop-deletion
-loop-extract-single -loop-rotate -loop-unroll -lowersetjmp -lowerswitch
-mem2reg -mergereturn -module-debuginfo -partial-inliner -partialspecialization
-profile-estimator -scalar-evolution -sccp -scev-aa -simplify-libcalls
-simplify-libcalls-halfpowr -simplifycfg -sink -strip-dead-debug-info
-strip-dead-prototypes -strip-nondebug -tailduplicate -tbaa
the boot2.bc 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 llvm.org Fri Jan 7 11:49:08 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 7 Jan 2011 11:49:08 -0600 (CST)
Subject: [LLVMbugs] [Bug 8929] New: Assertion failed: (CallSites.empty() &&
"Dangling pointers found in call sites map"), function RefreshCallGraph,
file CallGraphSCCPass.cpp, line 333.
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8929
Summary: Assertion failed: (CallSites.empty() && "Dangling
pointers found in call sites map"), function
RefreshCallGraph, file CallGraphSCCPass.cpp, line 333.
Product: libraries
Version: trunk
Platform: PC
OS/Version: FreeBSD
Status: NEW
Severity: normal
Priority: P
Component: Interprocedural Analyses
AssignedTo: unassignedbugs at nondot.org
ReportedBy: rdivacky at freebsd.org
CC: llvmbugs at cs.uiuc.edu
Assertion failed: (CallSites.empty() && "Dangling pointers found in call sites
map"), function RefreshCallGraph, file CallGraphSCCPass.cpp, line 333.
Stack dump:
0. Program arguments: opt boot2.bc -always-inline -basicaa -basiccg
-break-crit-edges -codegenprepare -deadtypeelim -die -domtree -early-cse
-functionattrs -globaldce -globalopt -globalsmodref-aa -indvars -inline
-instcombine -instnamer -instsimplify -intervals -iv-users -lazy-value-info
-lda -libcall-aa -licm -lint -live-values -loop-deletion -loop-extract
-loop-instsimplify -loop-reduce -loop-rotate -loop-unswitch -loops -mem2reg
-memcpyopt -mergefunc -mergereturn -no-profile -postdomfrontier -postdomtree
-profile-estimator -reg2mem -scalar-evolution -sccp -scev-aa -simplify-libcalls
-simplify-libcalls-halfpowr -sink -split-geps -strip-dead-debug-info
-strip-debug-declare -tailduplicate -tbaa
1. Running pass 'CallGraph Pass Manager' on module 'boot2.bc'.
and
Assertion failed: (CallSites.empty() && "Dangling pointers found in call sites
map"), function RefreshCallGraph, file CallGraphSCCPass.cpp, line 333.
Stack dump:
0. Program arguments: opt boot2.bc -adce -block-placement
-break-crit-edges -constprop -die -dse -globalopt -indvars -inline -instcombine
-instsimplify -iv-users -jump-threading -lazy-value-info -lda -licm
-live-values -loop-deletion -loop-extract -loop-idiom -loop-unroll -loops
-loweratomic -lowerinvoke -lowersetjmp -mem2reg -memcpyopt -mergefunc
-no-profile -postdomtree -prune-eh -reassociate -reg2mem -scalar-evolution
-scalarrepl -sccp -scev-aa -simplify-libcalls -simplify-libcalls-halfpowr
-simplifycfg -sink -sretpromotion -strip-dead-debug-info -strip-debug-declare
-strip-nondebug -tbaa
1. Running pass 'CallGraph Pass Manager' on module 'boot2.bc'.
the test file is the same as in bug 8928
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Fri Jan 7 12:51:26 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 7 Jan 2011 12:51:26 -0600 (CST)
Subject: [LLVMbugs] [Bug 3099] multiple errors in gcc/config/rs6000/rs6000.c
In-Reply-To:
References:
Message-ID: <20110107185126.78BF62A6C12C@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=3099
Chris Lattner changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
CC| |baldrick at free.fr
Resolution| |WONTFIX
--- Comment #38 from Chris Lattner 2011-01-07 12:51:23 CST ---
There is no reason to fix this anymore. If you're interested in this, please
try out dragonegg.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Fri Jan 7 15:55:04 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 7 Jan 2011 15:55:04 -0600 (CST)
Subject: [LLVMbugs] [Bug 8866] instcombine does not get max if hidden by sext
In-Reply-To:
References:
Message-ID: <20110107215504.3BD752A6C12C@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8866
Tobias Grosser changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
--- Comment #4 from Tobias Grosser 2011-01-07 15:55:03 CST ---
(In reply to comment #3)
> Patch posted:
> http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20110103/114606.html
>
> Thanks Dan for the hint
An improved version of this patch committed:
Revision: 123034
URL: http://llvm.org/viewvc/llvm-project?rev=123034&view=rev
Log: InstCombine: Match min/max hidden by sext/zext
X = sext x; x >s c ? X : C+1 --> X = sext x; X X = sext x; X >s C-1 ? C-1 : X
X = zext x; x >u c ? X : C+1 --> X = zext x; X X = zext x; X >u C-1 ? C-1 : X
X = sext x; x >u c ? X : C+1 --> X = sext x; X X = sext x; X >u C-1 ? C-1 : X
Instead of calculating this with mixed types promote all to the
larger type. This enables scalar evolution to analyze this
expression. PR8866
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Fri Jan 7 21:30:01 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 7 Jan 2011 21:30:01 -0600 (CST)
Subject: [LLVMbugs] [Bug 8930] New: std::ios_base is not forward declared by
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8930
Summary: std::ios_base is not forward declared by
Product: libc++
Version: unspecified
Platform: PC
OS/Version: Windows NT
Status: NEW
Severity: normal
Priority: P
Component: All Bugs
AssignedTo: hhinnant at apple.com
ReportedBy: eric at boostpro.com
CC: llvmbugs at cs.uiuc.edu
title says it all
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Fri Jan 7 22:24:07 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 7 Jan 2011 22:24:07 -0600 (CST)
Subject: [LLVMbugs] [Bug 8931] New: Missing FPU type directives in generated
gas assembly for ARM ELF target
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8931
Summary: Missing FPU type directives in generated gas assembly
for ARM ELF target
Product: new-bugs
Version: trunk
Platform: Other
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: new bugs
AssignedTo: unassignedbugs at nondot.org
ReportedBy: siarhei.siamashka at gmail.com
CC: llvmbugs at cs.uiuc.edu
Created an attachment (id=5975)
--> (http://llvm.org/bugs/attachment.cgi?id=5975)
arm-fpu-type-fix.patch
This problem in LLVM causes failure to compile anything that uses floating
point by clang.
The issue be fixed by applying the attached patch to LLVM.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sat Jan 8 15:10:23 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 8 Jan 2011 15:10:23 -0600 (CST)
Subject: [LLVMbugs] [Bug 8932] New: "opt -O3" goes into an infinite loop
with the attached IR
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8932
Summary: "opt -O3" goes into an infinite loop with the attached
IR
Product: libraries
Version: trunk
Platform: Macintosh
OS/Version: MacOS X
Status: NEW
Severity: normal
Priority: P
Component: Scalar Optimizations
AssignedTo: unassignedbugs at nondot.org
ReportedBy: sreeram at tachyontech.net
CC: llvmbugs at cs.uiuc.edu
I'm using LLVM r123062 on Mac OS X 64-bit.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sat Jan 8 16:31:21 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 8 Jan 2011 16:31:21 -0600 (CST)
Subject: [LLVMbugs] [Bug 8933] New: multilib support not detected on Debian
squeeze
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8933
Summary: multilib support not detected on Debian squeeze
Product: clang
Version: trunk
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: Driver
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: michael.kostylev at gmail.com
CC: llvmbugs at cs.uiuc.edu
Since base-files 6.0 the /etc/debian_version file contains "6.0" instead of
"squeeze/sid" that leads to the misdetection of the multilib support.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sat Jan 8 17:01:02 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 8 Jan 2011 17:01:02 -0600 (CST)
Subject: [LLVMbugs] [Bug 8934] New: GCC ignores conversion function template
specializatons if a derived class' conversion function converts to the same
type
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8934
Summary: GCC ignores conversion function template
specializatons if a derived class' conversion function
converts to the same type
Product: clang
Version: trunk
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: C++
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: schaub.johannes at googlemail.com
CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com
Unless I'm missing something, the following is ambiguous:
template
struct A { };
struct base {
template
operator A();
};
struct derived : base {
template
operator A();
};
int main() {
derived().operator A();
}
However, clang always wants to take the derived class version.
The reason for ambiguity is:
A is another type than A (prior
to deduction. When collecting candidates, 14.5.2/6 says:
"A specialization of a conversion function template is not found by name
lookup. Instead, any conversion function templates visible in the context of
the use are considered. For each such operator, if argument deduction
succeeds
(14.8.2.3), the resulting specialization is used as if found by name lookup."
In the context of the use, both functions templates are visible, so both are
sent to argument deduction. Deduction will succeed, so *the specialization is
used as if found by name lookup*. Name lookup cannot find declarations that are
hidden, so this quite clearly requires the above to be ambiguous, and doesn't
allow the base-class version to be "filtered out" later on. Hiding only takes
place prior to deduction:
template
struct A { };
struct base {
template
operator A();
};
struct derived : base {
template
operator A();
};
int main() {
derived().operator A();
}
Now that is not ambiguous anymore, and will call the derived class version,
because in the context of the use, the base-class version is hidden (12.3/5 and
3/8).
Note that by constructing cases making the base-class version more specialized
than the derived class version, we could actually make this a rejects-valid
bug:
template
struct A { };
struct base {
template
operator A();
};
struct derived : base {
private:
template
operator A();
};
int main() {
derived().operator A();
}
Comeau accepts this one just fine, while clang/GCC rejects it, thinking the
base-version is hidden.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sat Jan 8 17:29:07 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 8 Jan 2011 17:29:07 -0600 (CST)
Subject: [LLVMbugs] [Bug 8935] New: [MC x86] does not know xgetbv instruction
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8935
Summary: [MC x86] does not know xgetbv instruction
Product: new-bugs
Version: trunk
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: new bugs
AssignedTo: unassignedbugs at nondot.org
ReportedBy: kaffeemonster at googlemail.com
CC: llvmbugs at cs.uiuc.edu
TOT at 123092
int test_cpu_feature_avx_callback(void)
{
unsigned low, high;
asm volatile(
"xgetbv\n"
: /* %0 */ "=a" (low),
/* %1 */ "=d" (high)
: /* %2 */ "c" (0)
);
if((low & 0x06) != 0x06)
return 0;
return 1;
};
$ clang -c xgetbv.c
xgetbv.c:5:3: error: invalid instruction mnemonic 'xgetbv'
"xgetbv\n"
^
:1:2: note: instantiated into assembly here
xgetbv
^
1 error generated.
Greetings
Jan
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sat Jan 8 17:36:34 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 8 Jan 2011 17:36:34 -0600 (CST)
Subject: [LLVMbugs] [Bug 8936] New: [MC x86] does not know xcryptecb
instruction
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8936
Summary: [MC x86] does not know xcryptecb instruction
Product: new-bugs
Version: trunk
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: new bugs
AssignedTo: unassignedbugs at nondot.org
ReportedBy: kaffeemonster at googlemail.com
CC: llvmbugs at cs.uiuc.edu
TOT at 123092
struct aes_encrypt_ctx
{
unsigned k[(11*16)/4];
} __attribute__((aligned(16)));
void aes_ecb_encrypt_padlock(const struct aes_encrypt_ctx *ctx, void *out,
const void *in)
{
static const unsigned control_word[4] __attribute__((aligned(16))) =
{138, 0, 0, 0};
unsigned c, S, D;
asm(
"rep xcryptecb\n\t"
: /* %0 */ "=S" (S),
/* %1 */ "=D" (D),
/* %2 */ "=c" (c),
/* %4 */ "=m" (*(char *)out)
: /* %4 */ "d" (control_word),
/* %5 */ "b" (&ctx->k),
/* %6 */ "2" (1),
/* %7 */ "0" (in),
/* %8 */ "1" (out),
/* %9 */ "m" (*(const char *)in)
);
}
$ clang -c xcryptecb.c
xcryptecb.c:13:4: error: invalid instruction mnemonic 'xcryptecb'
"rep xcryptecb\n\t"
^
:1:6: note: instantiated into assembly here
rep xcryptecb
^
1 error generated.
Greetings
Jan
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sat Jan 8 18:03:36 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 8 Jan 2011 18:03:36 -0600 (CST)
Subject: [LLVMbugs] [Bug 8937] New: [MC x86] gas knows a default size for
stack access?
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8937
Summary: [MC x86] gas knows a default size for stack access?
Product: new-bugs
Version: trunk
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: new bugs
AssignedTo: unassignedbugs at nondot.org
ReportedBy: kaffeemonster at googlemail.com
CC: llvmbugs at cs.uiuc.edu
TOT at 123092
int foo(int c)
{
asm (
"push %0\n\t"
"and $3, (%%esp)\n\t"
"pop %0\n\t"
: "=r" (c)
: "0" (c)
);
return c;
}
$ gcc -c clang_and.c
gcc or more correct gas eats it
$ as --version
GNU assembler (GNU Binutils) 2.20.1.20100303
$ clang -c clang_and.c
clang_and.c:4:14: error: ambiguous instructions require an explicit suffix
(could be 'andb', 'andw', 'andl', or 'andq')
"push %0\n\t"
^
:2:2: note: instantiated into assembly here
and $3, (%esp)
^
1 error generated.
clang/llvm does not
Greetings
Jan
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sat Jan 8 18:54:22 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 8 Jan 2011 18:54:22 -0600 (CST)
Subject: [LLVMbugs] [Bug 8938] New: wint_t definition clashes with GCC
compiled c++ standard library
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8938
Summary: wint_t definition clashes with GCC compiled c++
standard library
Product: clang
Version: unspecified
Platform: Macintosh
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: Headers
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: admin at thefireflyproject.us
CC: llvmbugs at cs.uiuc.edu
The definition of wint_t
(http://llvm.org/svn/llvm-project/cfe/trunk/lib/Basic/TargetInfo.cpp line 45,
`WIntType = SignedInt;') conflicts with the definition of wint_t provided by
GCC (GCC defines wint_t as an unsigned int).
This (subtly) breaks clang on Linux, unless one uses a non-GCC compiled GNU
stdlib (I can't remember when I last tried compiling GNU stdlib with clang, but
I didn't have success. libc++ uses a Mac OSX centric build set up and I've only
had moderate success getting it to function on Linux). Debian and Ubuntu's
clang packages depend on GCC compiled GNU stdlib packages; there is not a
clang-compiled GNU stdlib package that I know of. I believe other distros
(RPM-based, etc) also distribute clang with GCC-built GNU stdlib.
Subsequently, this leads clang to fail to pass Boost C++ regressions for a
number of Boost libraries: string_algo, asio, graph, iostreams, math, regex,
serialization, spirit, units and uuid. The point of failure is std::wstreambuf;
the pbackfail and overflow methods take a parameter of int_type, which is
wint_t in the case of the std::wstreambuf typedef. GCC compiles the shared
library with wint_t typedef'd to an unsigned int, thus leading to a mismatch in
the function signature of pbackfail and overflows declaration (in the headers,
compiled by clang) and the definition (in the shared library, compiled by GCC).
This definition was introduced on Oct 21st, 2009 - git-svn-id:
http://llvm.org/svn/llvm-project/cfe/trunk at 84740
91177308-0d34-0410-b5e6-96231b3b80d8.
This problem doesn't occur on darwin (our clang-linux tester has been down for
awhile, so I thought this was just a problem for me for a period of time). I'm
assuming that's because either 0.) Apple ships a modified version of gcc with
Xcode that defines wint_t as an int, and Apple's shipped GNU stdlib package is
compiled with said modified version of gcc 1.) Apple has been building GNU
stdlib with clang since this change was introduced 2.) Apple has been using
libc++ since this change was introduced.
This is not a problem that I have been able to solve on the Boost end of things
with a workaround (I've fixed other stream-related stdlib issues with clang +
linux + gnu stdlib + boost by hacking together an implementation of the broken
component and having Boost use it if in a clang linux environment, but I'd
pretty much have to reimplement the entire stream library for this one). I
suspect changing the definition of wint_t in clang for all platforms will break
some existing code/interoperability with existing binaries. OTOH, this seems to
be a pretty fundamental inconsistency between clang and GCC-4.2 throught
GCC-4.4 (at least on Debian and from GNU's VCS trunk, GCC-4.5/libstdc++-4.5
replaces previously working C++03 and C++98 parts of the GNU standard library
with C++0x code that only GCC can compile, so I didn't bother checking if
GCC-4.5 defines wint_t differently).
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sat Jan 8 19:18:55 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 8 Jan 2011 19:18:55 -0600 (CST)
Subject: [LLVMbugs] [Bug 8937] [MC x86] gas knows a default size for stack
access?
In-Reply-To:
References:
Message-ID: <20110109011855.C78012A6C12C@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8937
Chris Lattner changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |INVALID
--- Comment #1 from Chris Lattner 2011-01-08 19:18:55 CST ---
The MC assembler is behaving correctly. There is nothing that specifies the
size for the operation, and you'll see that it picks a "default size" even if
not referencing ESP. Please add an explicit suffix.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sat Jan 8 22:41:08 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 8 Jan 2011 22:41:08 -0600 (CST)
Subject: [LLVMbugs] [Bug 8939] New: AST/CMakeLists.txt update
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8939
Summary: AST/CMakeLists.txt update
Product: clang
Version: trunk
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: -New Bugs
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: oroppas at gmail.com
CC: llvmbugs at cs.uiuc.edu
I've got the following message:
At revision 123105.
Export complete.
-- Target triple: i686-pc-linux-gnu
-- Native target architecture is X86
-- Threads enabled.
-- Building with -fPIC
-- Targeting X86
-- Targeting PTX
-- Clang version: 2.9
CMake Error at cmake/modules/LLVMProcessSources.cmake:79 (message):
Found unknown source file
/home/ryuta/devel/llvm/src/llvm/tools/clang/lib/AST/FullExpr.cpp
Please update
/home/ryuta/devel/llvm/src/llvm/tools/clang/lib/AST/CMakeLists.txt
Call Stack (most recent call first):
cmake/modules/LLVMProcessSources.cmake:35 (llvm_check_source_file_list)
tools/clang/CMakeLists.txt:58 (llvm_process_sources)
tools/clang/lib/AST/CMakeLists.txt:5 (add_clang_library)
-- Configuring incomplete, errors occurred!
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sun Jan 9 02:49:15 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 9 Jan 2011 02:49:15 -0600 (CST)
Subject: [LLVMbugs] [Bug 8940] New: libclang API to access
CXCodeCompleteResults and CXCompletionResult structures
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8940
Summary: libclang API to access CXCodeCompleteResults and
CXCompletionResult structures
Product: clang
Version: trunk
Platform: All
OS/Version: All
Status: NEW
Severity: enhancement
Priority: P
Component: -New Bugs
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: lightside at safebox.ru
CC: llvmbugs at cs.uiuc.edu
There are no API to access CXCodeCompleteResults structure to get
CXCompletionResult and unsigned NumResults.
The same situation with CXCompletionResult structure (e.g. CXCursorKind
CursorKind, CXCompletionString CompletionString). They used inside
clang_sortCodeCompletionResults, clang_getCompletionChunkKind,
clang_getCompletionChunkText, clang_getCompletionChunkCompletionString,
clang_getNumCompletionChunks, etc. functions.
Currently, it possible to get them directly only, e.g.:
//...
CXIndex idx = clang_createIndex(NULL, NULL);
CXTranslationUnit tu = clang_createTranslationUnit(idx, ast_filename);
CXCodeCompleteResults *rs = clang_codeCompleteAt(tu, complete_filename,
complete_line, complete_column, unsaved_files, num_unsaved_files, options);
clang_sortCodeCompletionResults(rs->Results, rs->NumResults);
for (unsigned i = 0, n = rs->NumResults; i < n; ++i) {
CXCompletionResult r = rs->Results[i];
CXCursorKind kind = r.CursorKind;
CXCompletionString cs = r.CompletionString;
for (unsigned j = 0, m = clang_getNumCompletionChunks(cs); j < m; ++j) {
CXString text = clang_getCompletionChunkText(cs, j);
printf("%d: %s\n", i, text);
clang_disposeString(text);
}
}
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sun Jan 9 07:15:28 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 9 Jan 2011 07:15:28 -0600 (CST)
Subject: [LLVMbugs] [Bug 8939] AST/CMakeLists.txt update
In-Reply-To:
References:
Message-ID: <20110109131528.9DBC22A6C12E@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8939
?scar Fuentes changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
CC| |ofv at wanadoo.es
Resolution| |INVALID
--- Comment #1 from ?scar Fuentes 2011-01-09 07:15:28 CST ---
That source file does not exists on the clang sources. I have no idea why you
have it on your disk (does `snv status' report it as unknown? Maybe your
working copy is on a inconsistent state.) Just delete 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 llvm.org Sun Jan 9 17:02:53 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 9 Jan 2011 17:02:53 -0600 (CST)
Subject: [LLVMbugs] [Bug 8882] Need an "isexact" bit for shift left
In-Reply-To:
References:
Message-ID: <20110109230253.080212A6C12C@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8882
Chris Lattner changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |DUPLICATE
--- Comment #4 from Chris Lattner 2011-01-09 17:02:52 CST ---
Just duping this to PR8862, we'll track the issue there.
*** This bug has been marked as a duplicate of bug 8862 ***
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sun Jan 9 17:38:45 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 9 Jan 2011 17:38:45 -0600 (CST)
Subject: [LLVMbugs] [Bug 8942] New: Not aggressively optimizing std::fill
loop
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8942
Summary: Not aggressively optimizing std::fill loop
Product: libraries
Version: 1.0
Platform: PC
OS/Version: All
Status: NEW
Severity: normal
Priority: P
Component: Scalar Optimizations
AssignedTo: unassignedbugs at nondot.org
ReportedBy: clattner at apple.com
CC: llvmbugs at cs.uiuc.edu
Chandler observed that we don't optimize this function into a memset:
void f1(int* begin, int* end) {
std::fill(begin, end, 0);
}
With recent changes, we now compute a backedge taken count for this loop, of
"((-4 + (-1 * %__first) + %__last) /u 4)". The problem is that indvars has
some (very reasonable in general) code that Dan added at the top of
IndVarSimplify::LinearFunctionTestReplace:
// Special case: If the backedge-taken count is a UDiv, it's very likely a
// UDiv that ScalarEvolution produced in order to compute a precise
// expression, rather than a UDiv from the user's code. If we can't find a
// UDiv in the code with some simple searching, assume the former and forego
// rewriting the loop.
if (isa(BackedgeTakenCount)) {
If I hack out that code, we form a memset and delete the loop. There are two
problems here. First, if I disable the code, I get some pretty gross IR:
define void @_Z1fPiS_(i32* %begin, i32* %end) nounwind {
entry:
%cmp7.i.i = icmp eq i32* %begin, %end
br i1 %cmp7.i.i, label %_ZSt4fillIPiiEvT_S1_RKT0_.exit, label
%for.body.lr.ph.i.i
for.body.lr.ph.i.i: ; preds = %entry
%begin2 = bitcast i32* %begin to i8*
%__first10.i.i = ptrtoint i32* %begin to i64
%scevgep.i.i = getelementptr i32* %end, i64 -1
%scevgep9.i.i = bitcast i32* %scevgep.i.i to i8*
%tmp.i.i = sub i64 0, %__first10.i.i
%uglygep.i.i = getelementptr i8* %scevgep9.i.i, i64 %tmp.i.i
%uglygep11.i.i = ptrtoint i8* %uglygep.i.i to i64
%tmp12.i.i4 = add i64 %uglygep11.i.i, 4
%tmp3 = and i64 %tmp12.i.i4, -4
call void @llvm.memset.p0i8.i64(i8* %begin2, i8 0, i64 %tmp3, i32 4, i1
false)
ret void
_ZSt4fillIPiiEvT_S1_RKT0_.exit: ; preds = %entry
ret void
}
1. The "gep -1", and "add -4" should be merged together. It is unclear if it
should be SCEV doing this or instcombine.
2. The "and x, -4" is because we don't have an "isexact" bit on the udiv
generated by the trip count. We have no way to represent this in SCEV or IR
(PR8862).
The second issue is that just hacking out the code isn't the right thing to do.
We should use the "isexact" bit on the udiv scev to decide if the end result
will be simple enough to make it profitable.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sun Jan 9 17:50:25 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 9 Jan 2011 17:50:25 -0600 (CST)
Subject: [LLVMbugs] [Bug 8943] New: pow is transformed to exp2 in -std=gnu89
mode
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8943
Summary: pow is transformed to exp2 in -std=gnu89 mode
Product: clang
Version: trunk
Platform: PC
OS/Version: Windows NT
Status: NEW
Severity: normal
Priority: P
Component: -New Bugs
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: amine.khaldi at reactos.org
CC: llvmbugs at cs.uiuc.edu
I got undefined reference to `exp2' from this line of code:
r->f = a->f * pow(2., n.f);
exp2 doesn't seem to be in the C89 standard, so this transformation is a bug
most likely.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sun Jan 9 18:56:31 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 9 Jan 2011 18:56:31 -0600 (CST)
Subject: [LLVMbugs] [Bug 8944] New: Integrated assembler doesn't resolve
local symbol differences
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8944
Summary: Integrated assembler doesn't resolve local symbol
differences
Product: libraries
Version: trunk
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: Target Description Classes
AssignedTo: unassignedbugs at nondot.org
ReportedBy: joerg at NetBSD.org
CC: llvmbugs at cs.uiuc.edu
OpenSSL's aes-x86_64.s does something like
movq T-S, %rax
.globl T
T:
.globl S
S:
This results in a R_X86_64_PC32 text relocation, which GNU ld doesn't like.
Runnning the same code through GNU as doesn't result in a relocation, e.g. the
difference is handled internally. Same for using . in place of 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 llvm.org Sun Jan 9 19:06:30 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 9 Jan 2011 19:06:30 -0600 (CST)
Subject: [LLVMbugs] [Bug 8945] New: Support for -cxx-isystem
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8945
Summary: Support for -cxx-isystem
Product: clang
Version: trunk
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: Frontend
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: joerg at NetBSD.org
CC: llvmbugs at cs.uiuc.edu
GCC has a special option to override the builtin c++ include path in
combination with -nostdinc++. The option -cxx-isystem is different from
-isystem in so far as that it only applies to C++ compilation.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Sun Jan 9 20:01:05 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 9 Jan 2011 20:01:05 -0600 (CST)
Subject: [LLVMbugs] [Bug 8946] New: SSE2 : valid reg-to-reg MOVDQU rejected
by the x86 backend
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8946
Summary: SSE2 : valid reg-to-reg MOVDQU rejected by the x86
backend
Product: libraries
Version: trunk
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: Backend: X86
AssignedTo: unassignedbugs at nondot.org
ReportedBy: jmenon86 at gmail.com
CC: llvmbugs at cs.uiuc.edu
Created an attachment (id=5977)
--> (http://llvm.org/bugs/attachment.cgi?id=5977)
test case
Attached code is valid and compiles fine with gcc but with clang, I get
[jai at outrax tmp]$ ~/llvm-build/Debug+Asserts/bin/clang -msse2 -c sse2_test.c
sse2_test.c:3:13: error: invalid operand for instruction
__asm__("movdqu %%xmm0, %%xmm1\n" : );
^
:1:2: note: instantiated into assembly here
movdqu %xmm0, %xmm1
^
1 error 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 llvm.org Sun Jan 9 23:14:53 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 9 Jan 2011 23:14:53 -0600 (CST)
Subject: [LLVMbugs] [Bug 8947] New: Miscompilation when optimizing OpenSSL
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8947
Summary: Miscompilation when optimizing OpenSSL
Product: new-bugs
Version: trunk
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: new bugs
AssignedTo: unassignedbugs at nondot.org
ReportedBy: jhendrix at galois.com
CC: llvmbugs at cs.uiuc.edu
Created an attachment (id=5979)
--> (http://llvm.org/bugs/attachment.cgi?id=5979)
Test case reduced by bugpoint
In trying to build OpenSSL with clang, I ran into a crypto algorithm where the
LLVM optimizer is generating bad code. This error was found on LLVM 2.8, but I
have reproduced it on 2.9-svn.
I was able to extract a simple test case and use bugpoint to narrow down the
bug into LLVM's optimizer, and have attached LLVM code that triggered the bug.
To reproduce the bug use the following script:
opt -o bad-reduced.bc good-reduced.ll -no-aa -tbaa -basicaa -globalopt -ipsccp
-deadargelim -instcombine -simplifycfg -basiccg -prune-eh -inline
-functionattrs -domtree -domfrontier -scalarrepl -early-cse -simplify-libcalls
-lazy-value-info -jump-threading -correlated-propagation -simplifycfg
-instcombine -tailcallelim -simplifycfg -reassociate -domtree -loops
-loopsimplify -lcssa -loop-rotate -licm -lcssa -loop-unswitch -instcombine
-scalar-evolution -loopsimplify -lcssa -iv-users -indvars -loop-idiom
-loop-deletion -loop-unroll -instcombine -memdep -gvn -memdep -memcpyopt -sccp
-instcombine -lazy-value-info -jump-threading -correlated-propagation -domtree
-memdep -dse -adce -simplifycfg -strip-dead-prototypes -print-used-types
-deadtypeelim -constmerge -preverify -domtree -verify
echo "Running unoptimized"
clang -o good-reduced good-reduced.ll && ./good-reduced
echo "Running optimized"
clang -o bad-reduced bad-reduced.bc && ./bad-reduced
On my machine, the following output shows the expected result:
Running unoptimized
ok 10ce4c18
broken 4c
ok 527039d8
broken 39
Running optimized
ok 10ce4c18
broken 4e
ok 320ab844
broken ba
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Mon Jan 10 00:06:40 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 10 Jan 2011 00:06:40 -0600 (CST)
Subject: [LLVMbugs] [Bug 8948] New: .section parsed incorrectly
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8948
Summary: .section parsed incorrectly
Product: libraries
Version: trunk
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: LLVM assembly language parser
AssignedTo: unassignedbugs at nondot.org
ReportedBy: joerg at NetBSD.org
CC: llvmbugs at cs.uiuc.edu
On a ELF system, compile
.section "foo"
and check the output of readelf -S. Expected output is a section with name foo,
but created is a section with name "fo.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Mon Jan 10 01:53:40 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 10 Jan 2011 01:53:40 -0600 (CST)
Subject: [LLVMbugs] [Bug 6371] Assertion failed: (TInfo && "couldn't build
declarator info for anonymous struct/union")
In-Reply-To:
References:
Message-ID: <20110110075340.6EADA2A6C12D@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=6371
Chris Lattner changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
--- Comment #1 from Chris Lattner 2011-01-10 01:53:39 CST ---
This looks like it got fixed 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 llvm.org Mon Jan 10 09:50:39 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 10 Jan 2011 09:50:39 -0600 (CST)
Subject: [LLVMbugs] [Bug 8949] New: Redefining a weak symbol crashes with
integrated as
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8949
Summary: Redefining a weak symbol crashes with integrated as
Product: libraries
Version: trunk
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: Target Description Classes
AssignedTo: unassignedbugs at nondot.org
ReportedBy: joerg at NetBSD.org
CC: llvmbugs at cs.uiuc.edu
Consider:
asm("bar:\n.weak foo\nfoo = bar");
int foo(void)
{
return 0;
}
Assembling first and pushing the result through the integrated assembler works
and correctly flags the redefinition. Compiling this directly with the
integrated assembler crashes with:
#2 0x00007ffff6cada71 in __assert_fail (
assertion=0x170c300 "!Symbol->isVariable() && \"Cannot emit a variable
symbol!\"", file=, line=100,
function=0x1714780 "virtual void
llvm::MCObjectStreamer::EmitLabel(llvm::MCSymbol*)") at assert.c:81
#3 0x00000000010b0d89 in llvm::MCObjectStreamer::EmitLabel (
this=, Symbol=0x1d629f0)
at
/home/joerg/work/infrastructure/llvm.org/llvm/lib/MC/MCObjectStreamer.cpp:100
#4 0x00000000010a0bcc in (anonymous namespace)::MCELFStreamer::EmitLabel (
this=0x1d5fd60, Symbol=0x3fc)
at
/home/joerg/work/infrastructure/llvm.org/llvm/lib/MC/MCELFStreamer.cpp:191
#5 0x0000000000da248e in llvm::AsmPrinter::EmitFunctionEntryLabel (
this=0x1d631f0)
at
/home/joerg/work/infrastructure/llvm.org/llvm/lib/CodeGen/AsmPrinter/AsmPrinter.cpp:468
#6 0x0000000000da5333 in llvm::AsmPrinter::EmitFunctionHeader (
this=0x1d631f0)
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Mon Jan 10 11:53:31 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 10 Jan 2011 11:53:31 -0600 (CST)
Subject: [LLVMbugs] [Bug 8950] New: Crypto++ validation test crashes in IDEA
code
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8950
Summary: Crypto++ validation test crashes in IDEA code
Product: clang
Version: trunk
Platform: Macintosh
OS/Version: MacOS X
Status: NEW
Severity: normal
Priority: P
Component: C++
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: mclow at qualcomm.com
CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com
Created an attachment (id=5981)
--> (http://llvm.org/bugs/attachment.cgi?id=5981)
Replacement makefile (uses clang, removes -O2)
This looks like a codegen error; but I'm not 100% sure.
To reproduce:
1) Get crypto++ from here: "svn co
https://cryptopp.svn.sourceforge.net/svnroot/cryptopp/trunk/c5 cryptopp"
2) Replace GNUMakefile with one from this bug report.
3) make test
--> segmentation fault
Running under gdb shows:
Program received signal EXC_BAD_ACCESS, Could not access memory.
Reason: 13 at address: 0x0000000000000000
0x000000010017d7a0 in CryptoPP::IDEA::Base::DeKey (this=0x100d04ba0) at
idea.cpp:141
141 tempkey[i*6+5] = m_key[(ROUNDS-1-i)*6+5];
and some poking around shows that i == 0, and that the sizes of the arrays are
ok.
More questions? Drop me a line.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Mon Jan 10 12:59:19 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 10 Jan 2011 12:59:19 -0600 (CST)
Subject: [LLVMbugs] [Bug 8951] New: Support for .equiv in integrated
assembler
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8951
Summary: Support for .equiv in integrated assembler
Product: libraries
Version: trunk
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: LLVM assembly language parser
AssignedTo: unassignedbugs at nondot.org
ReportedBy: joerg at NetBSD.org
CC: llvmbugs at cs.uiuc.edu
.equiv works like .set / .equ, but disallows redefinition. Attached patch adds
support for 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 llvm.org Mon Jan 10 15:54:09 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 10 Jan 2011 15:54:09 -0600 (CST)
Subject: [LLVMbugs] [Bug 8916] clang tot takes a long,
long time (maybe forever) to compile Cryptopp's sha.cpp
In-Reply-To:
References:
Message-ID: <20110110215409.DE8E22A6C12D@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8916
Dale Johannesen changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
--- Comment #11 from Dale Johannesen 2011-01-10 15:54:09 CST ---
I've checked in the 1liner described above in 123191, which fixes this
particular problem, anyway.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Mon Jan 10 16:30:49 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 10 Jan 2011 16:30:49 -0600 (CST)
Subject: [LLVMbugs] [Bug 8952] New: extern "C" function returning pointer to
class in namespace{} silently drops symbol
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8952
Summary: extern "C" function returning pointer to class in
namespace{} silently drops symbol
Product: clang
Version: trunk
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: -New Bugs
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: ps at kr.tuwien.ac.at
CC: llvmbugs at cs.uiuc.edu
I compiled the following minimal example with clang trunk using "clang++ -c"
===
namespace{
class Foo {};
Foo f;
}
extern "C" Foo* bar() { return &f; }
extern "C" void* baz() { return reinterpret_cast(&f); }
===
The resulting object file has no symbol "bar" but it has a symbol "baz".
There is no warning and no error.
g++ 4.4.5 creates an object file with both symbols (and they are usable).
I honestly do not know which compiler is right, but probably if the symbol is
not generated there should at least be a warning.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Mon Jan 10 18:33:29 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 10 Jan 2011 18:33:29 -0600 (CST)
Subject: [LLVMbugs] [Bug 8912] Non-type template argument parse error (with
a functional cast expression).
In-Reply-To:
References:
Message-ID: <20110111003329.C6E372A6C124@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8912
Douglas Gregor changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
--- Comment #1 from Douglas Gregor 2011-01-10 18:33:29 CST ---
Fixed in Clang r123201.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Mon Jan 10 20:51:49 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 10 Jan 2011 20:51:49 -0600 (CST)
Subject: [LLVMbugs] [Bug 8953] New: Edge splitting should update
LiveIntervals
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8953
Summary: Edge splitting should update LiveIntervals
Product: libraries
Version: trunk
Platform: PC
OS/Version: All
Status: NEW
Severity: normal
Priority: P
Component: Common Code Generator Code
AssignedTo: unassignedbugs at nondot.org
ReportedBy: zwarich at apple.com
CC: llvmbugs at cs.uiuc.edu
Edge splitting should update LiveIntervals so that we can split edges on demand
throughout the register allocation pipeline.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Mon Jan 10 21:18:05 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 10 Jan 2011 21:18:05 -0600 (CST)
Subject: [LLVMbugs] [Bug 8954] New: non deterministic "-basicaa -loop-unroll
-gvn"
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8954
Summary: non deterministic "-basicaa -loop-unroll -gvn"
Product: libraries
Version: trunk
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: Scalar Optimizations
AssignedTo: unassignedbugs at nondot.org
ReportedBy: rafael.espindola at gmail.com
CC: llvmbugs at cs.uiuc.edu
Created an attachment (id=5984)
--> (http://llvm.org/bugs/attachment.cgi?id=5984)
testcase
Running
opt -basicaa -loop-unroll -gvn t.ll -o sched-vis2.ll
on the attached file multiple times produces different results if address space
randomization is enabled.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Mon Jan 10 22:44:46 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 10 Jan 2011 22:44:46 -0600 (CST)
Subject: [LLVMbugs] [Bug 8955] New: Assertion `DidIt && "Block merge
failed??"' failed.
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8955
Summary: Assertion `DidIt && "Block merge failed??"' failed.
Product: new-bugs
Version: trunk
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: new bugs
AssignedTo: unassignedbugs at nondot.org
ReportedBy: regehr at cs.utah.edu
CC: llvmbugs at cs.uiuc.edu, chenyang at cs.utah.edu
regehr at home:~/volatile/bugs/tmp344$ clang -c -w small.c -O1
clang: LoopRotation.cpp:343:
bool::LoopRotate::rotateLoop(llvm::Loop*): Assertion `DidIt && "Block
merge failed??"' failed.
0 clang 0x09476448
Stack dump:
0. Program arguments:
/mnt/z/z/compiler-install/llvm-gcc-r123209-install/bin/clang -cc1 -triple
i386-pc-linux-gnu -emit-obj -disable-free -main-file-name small.c
-mrelocation-model static -mdisable-fp-elim -masm-verbose -mconstructor-aliases
-target-cpu pentium4 -target-linker-version 2.20.51 -resource-dir
/mnt/z/z/compiler-install/llvm-gcc-r123209-install/bin/../lib/clang/2.9 -O1 -w
-ferror-limit 19 -fmessage-length 80 -fgnu-runtime -fdiagnostics-show-option
-fcolor-diagnostics -o small.o -x c small.c
1. parser at end of file
2. Per-module optimization passes
3. Running pass 'CallGraph Pass Manager' on module 'small.c'.
4. Running pass 'Loop Pass Manager' on function '@func_46'
5. Running pass 'Rotate Loops' on basic block '%lbl_283'
clang: error: unable to execute command: Aborted
clang: error: clang frontend command failed due to signal 1 (use -v to see
invocation)
regehr at home:~/volatile/bugs/tmp344$ clang -v
clang version 2.9 (trunk 123209)
Target: i386-pc-linux-gnu
Thread model: posix
regehr at home:~/volatile/bugs/tmp344$ cat small.c
int g_2;
int *g_37;
int **g = &g_37;
int g_57;
int *g_141 = &g_2;
int *g_143 = &g_57;
int func_64 (int *);
void func_46 (int p_47)
{
int l_166;
int *l_261 = &l_166;
int *l_286;
l_286 = &l_166;
lbl_283:
*g_143 |= 0;
if (p_47)
{
int **l_262 = &l_261;
if (**l_262)
{
for (; g; )
{
}
lbl_281:
*g = 0;
}
if (g_57) goto lbl_281;
goto lbl_283;
}
else
{
int **l_295 = &l_286;
if (func_64 (*l_295));
}
}
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Tue Jan 11 01:48:16 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 11 Jan 2011 01:48:16 -0600 (CST)
Subject: [LLVMbugs] [Bug 8955] Assertion `DidIt && "Block merge failed??"'
failed.
In-Reply-To:
References:
Message-ID: <20110111074816.10B862A6C12D@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8955
Chris Lattner changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
--- Comment #1 from Chris Lattner 2011-01-11 01:48:15 CST ---
Aha, great catch, fixed in r123219. 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 llvm.org Tue Jan 11 09:26:33 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 11 Jan 2011 09:26:33 -0600 (CST)
Subject: [LLVMbugs] [Bug 8956] New: asm labels are ignored for static local
variables
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8956
Summary: asm labels are ignored for static local variables
Product: clang
Version: trunk
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: LLVM Codegen
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: abramobagnara at tin.it
CC: llvmbugs at cs.uiuc.edu
$ cat t.c
int f() {
static int x asm("qq");
return x;
}
$ clang -c t.c
$ nm t.o
0000000000000000 T f
0000000000000000 b f.x
$ gcc -c t.c
$ nm t.o
0000000000000000 T f
0000000000000000 b qq
No symbol called qq has been created by clang.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Tue Jan 11 10:58:25 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 11 Jan 2011 10:58:25 -0600 (CST)
Subject: [LLVMbugs] [Bug 8957] New: Recognize ARM EABI triple
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8957
Summary: Recognize ARM EABI triple
Product: clang
Version: unspecified
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: Driver
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: rengolin at gmail.com
CC: llvmbugs at cs.uiuc.edu
The EABI triples are not being recognized by Clang (but are by llc), so we need
to add it to lib/Driver/Tools.cpp:Clang::AddARMTargetArgs() and possibly update
llvm/ADT/Triple.h with an "EABI" Environment (or OS).
See bug 8911 for full discussion.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Tue Jan 11 11:05:19 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 11 Jan 2011 11:05:19 -0600 (CST)
Subject: [LLVMbugs] [Bug 8946] [MC] SSE2 : valid reg-to-reg MOVDQU rejected
by the x86 backend
In-Reply-To:
References:
Message-ID: <20110111170519.E1F1F2A6C124@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8946
Chris Lattner changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
Summary|SSE2 : valid reg-to-reg |[MC] SSE2 : valid
|MOVDQU rejected by the x86 |reg-to-reg MOVDQU rejected
|backend |by the x86 backend
--- Comment #2 from Chris Lattner 2011-01-11 11:05:19 CST ---
Nice catch, fixed in r123242, 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 llvm.org Tue Jan 11 11:28:21 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 11 Jan 2011 11:28:21 -0600 (CST)
Subject: [LLVMbugs] [Bug 8958] New: Hit assertion in SelectionDAG.cpp for
the attached code
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8958
Summary: Hit assertion in SelectionDAG.cpp for the attached
code
Product: libraries
Version: trunk
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: Backend: X86
AssignedTo: unassignedbugs at nondot.org
ReportedBy: nadav.rotem at intel.com
CC: llvmbugs at cs.uiuc.edu
Assertion failed: !VT.isVector() && "getZeroExtendInReg should use the vector
element type instead of " "the vector type!", file
..\..\..\..\..\llvm\lib\CodeGen\SelectionDAG\SelectionDAG.cpp, line 855
Checked on today's trunk.
; ModuleID = 'bugpoint-reduced-simplified.bc'
target datalayout =
"e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-f80:128:128-v64:64:64-v128:128:128-a0:0:64-f80:32:32-n8:16:32"
target triple = "i686-pc-win32"
define void @m_387() nounwind {
%1 = load <4 x i8> addrspace(1)* undef, align 1
%2 = zext <4 x i8> %1 to <4 x i32>
store <4 x i32> %2, <4 x i32> addrspace(1)* undef, align 4
ret void
}
Fix forthcoming.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Tue Jan 11 11:34:59 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 11 Jan 2011 11:34:59 -0600 (CST)
Subject: [LLVMbugs] [Bug 8318] ARM EABI: inefficient code saving/restoring
sp around function call
In-Reply-To:
References:
Message-ID: <20110111173459.2A6472A6C12C@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8318
Bob Wilson changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |INVALID
--- Comment #3 from Bob Wilson 2011-01-11 11:34:58 CST ---
This is expected behavior. You need to compile with -fomit-frame-pointer. I
think GCC automatically enables that option for some targets when compiling
with optimization, but clang does not do that, as far as I 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 llvm.org Tue Jan 11 13:52:42 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 11 Jan 2011 13:52:42 -0600 (CST)
Subject: [LLVMbugs] [Bug 8959] New: inline asm "0" constraint doesn't
implicitly cast
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8959
Summary: inline asm "0" constraint doesn't implicitly cast
Product: new-bugs
Version: trunk
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: new bugs
AssignedTo: unassignedbugs at nondot.org
ReportedBy: nlewycky at google.com
CC: llvmbugs at cs.uiuc.edu, spectral at google.com
This is a breakout from bug 3373 comment 11. New testcase:
int test_and_set(volatile int *addr)
{
unsigned char oldval;
/* Note: the "xchg" instruction does not need a "lock" prefix */
__asm__ __volatile__("xchgb %0, %1"
: "=Q"(oldval), "=m"(*addr)
: "0"(0xff), "m"(*addr) : "memory");
return (int)oldval;
}
which produces:
$ clang t.c -S -o -
t.c:7:25: error: unsupported inline asm: input with type 'int' matching
output
with type 'unsigned char'
: "0"(0xff), "m"(*addr) : "memory");
^~~~
1 error generated.
Ian Taylor explains: "Clang is incorrectly assuming that when given an operand
number as a constraint that the type of a constant must be identical to the
type of the other operand. When the operand is a constant, it should be
converted to the relevant type as though in an assignment. When the operand is
a variable of some sort, it must of course have exactly the correct type."
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Tue Jan 11 14:32:47 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 11 Jan 2011 14:32:47 -0600 (CST)
Subject: [LLVMbugs] [Bug 8960] New: Wrong end location for CXXNamedCastExpr
Message-ID:
http://llvm.org/bugs/show_bug.cgi?id=8960
Summary: Wrong end location for CXXNamedCastExpr
Product: clang
Version: unspecified
Platform: PC
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P
Component: Frontend
AssignedTo: unassignedclangbugs at nondot.org
ReportedBy: alr at google.com
CC: llvmbugs at cs.uiuc.edu
For the following expression:
x = static_cast(f(y));
CXXStaticCastExpr::GetLocEnd()
returns:
x = static_cast(f(y));
^
expected:
x = static_cast(f(y));
^
The last token in static_cast<> should be the closing parenthesis, not last
token of sub-expression.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Tue Jan 11 16:27:12 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 11 Jan 2011 16:27:12 -0600 (CST)
Subject: [LLVMbugs] [Bug 8658] clang doesn't ignore '/' comments in
assembler .s files
In-Reply-To:
References:
Message-ID: <20110111222712.CC0DB2A6C12D@llvm.org>
http://llvm.org/bugs/show_bug.cgi?id=8658
Nick Lewycky changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |WONTFIX
--- Comment #10 from Nick Lewycky 2011-01-11 16:27:12 CST ---
After a lengthy discussion on IRC, the concensus was reached that this is an
undesirable extension to binutils and since the binutils documentation doesn't
list it as a valid way to write comments, we shouldn't accept it either.
I'm going to file bugs on both binutils and mozilla to get their sides fixed
respectively and will update this bug with link to those bugs as they're filed.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
From bugzilla-daemon at llvm.org Tue Jan 11 17:25:11 2011
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 11 Jan 2011 17:25:11 -0600 (CST)
Subject: [LLVMbugs] [Bug 8956] asm labels are ignored for static local
variables
In-Reply-To: