[llvm-bugs] [Bug 31161] New: Crash when using vector long long or vector double operator manipulation with altivec
https://llvm.org/bugs/show_bug.cgi?id=31161
Bug ID: 31161
Summary: Crash when using vector long long or vector double
operator manipulation with altivec
Product: clang
Version: trunk
Hardware: Other
OS: Linux
Status: NEW
Severity: normal
Priority: P
Component: -New Bugs
Assignee: unassignedclangb...@nondot.org
Reporter: westion...@gmail.com
CC: llvm-bugs@lists.llvm.org
Classification: Unclassified
int main(){
vector long long a = {1,2};
vector long long b = {1,2};
vector double c = {1.1,2.1};
vector double d = {1.1,2.1};
if(ab);
if(a==b);
if(cd);
if(c==d);
}
it crashes.
As vector int and vector float work well. It should call related llvm builtins.
example.C
int main(){
vector int a = {1,2};
vector int b = {1,2};
vector float c = {1.1,2.1};
vector float d = {1.1,2.1};
if(ab);
if(a==b);
if(cd);
if(c==d);
}
cat example.ll | grep llvm.
%8 = call i32 @llvm.ppc.altivec.vcmpgtsw.p(i32 2, <4 x i32> %7, <4 x i32> %6)
%13 = call i32 @llvm.ppc.altivec.vcmpgtsw.p(i32 2, <4 x i32> %11, <4 x i32>
%12)
%18 = call i32 @llvm.ppc.altivec.vcmpequw.p(i32 2, <4 x i32> %16, <4 x i32>
%17)
%23 = call i32 @llvm.ppc.altivec.vcmpgtfp.p(i32 2, <4 x float> %22, <4 x
float> %21)
%28 = call i32 @llvm.ppc.altivec.vcmpgtfp.p(i32 2, <4 x float> %26, <4 x
float> %27)
%33 = call i32 @llvm.ppc.altivec.vcmpeqfp.p(i32 2, <4 x float> %31, <4 x
float> %32)
declare i32 @llvm.ppc.altivec.vcmpgtsw.p(i32, <4 x i32>, <4 x i32>) #1
declare i32 @llvm.ppc.altivec.vcmpequw.p(i32, <4 x i32>, <4 x i32>) #1
declare i32 @llvm.ppc.altivec.vcmpgtfp.p(i32, <4 x float>, <4 x float>) #1
declare i32 @llvm.ppc.altivec.vcmpeqfp.p(i32, <4 x float>, <4 x float>) #1
--
You are receiving this mail because:
You are on the CC list for the bug.
___
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 26322] Core load hangs
https://llvm.org/bugs/show_bug.cgi?id=26322 lab...@google.com changed: What|Removed |Added Status|ASSIGNED|RESOLVED CC||lab...@google.com Resolution|--- |FIXED --- Comment #3 from lab...@google.com --- Marking the bug as closed. -- You are receiving this mail because: You are on the CC list for the bug. ___ llvm-bugs mailing list llvm-bugs@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 31162] New: LLVM (Clang) failed to optimized little-endian integer composition
https://llvm.org/bugs/show_bug.cgi?id=31162
Bug ID: 31162
Summary: LLVM (Clang) failed to optimized little-endian integer
composition
Product: new-bugs
Version: trunk
Hardware: PC
OS: Linux
Status: NEW
Severity: enhancement
Priority: P
Component: new bugs
Assignee: unassignedb...@nondot.org
Reporter: peetluge...@yandex.com
CC: llvm-bugs@lists.llvm.org
Classification: Unclassified
Clang/LLVM is unable to optimize this:
uint32_t read32le(const uint8_t* ptr) {
return (static_cast(ptr[0])& 0x00FFu) |
((static_cast(ptr[1]) << 8) & 0xFF00u) |
((static_cast(ptr[2]) << 16) & 0x00FFu) |
((static_cast(ptr[3]) << 24) & 0xFF00u);
}
LLVM code:
read32le(unsigned char const*):
movzx eax, byte ptr [rdi]
movzx ecx, byte ptr [rdi + 1]
shl ecx, 8
or ecx, eax
movzx edx, byte ptr [rdi + 2]
shl edx, 16
or edx, ecx
movzx eax, byte ptr [rdi + 3]
shl eax, 24
or eax, edx
ret
GCC (7) code:
read32le(unsigned char const*):
mov eax, DWORD PTR [rdi]
ret
--
You are receiving this mail because:
You are on the CC list for the bug.
___
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 31163] New: "using I = int" makes Clang crash if "I" is also a name of a constant in another scope.
https://llvm.org/bugs/show_bug.cgi?id=31163
Bug ID: 31163
Summary: "using I = int" makes Clang crash if "I" is also a
name of a constant in another scope.
Product: clang
Version: 3.9
Hardware: PC
OS: Linux
Status: NEW
Severity: normal
Priority: P
Component: -New Bugs
Assignee: unassignedclangb...@nondot.org
Reporter: officesamu...@gmail.com
CC: llvm-bugs@lists.llvm.org
Classification: Unclassified
The following code makes Clang crash if compiled with -std=c++14
#include
template
void foo()
{
constexpr int I = 0;
std::get(Tuple());
}
int main()
{
foo>();
}
// Remove this and the crash is gone
using I = int;
Output:
~/tmp$ clang++-3.9 ClangCrash.cpp -std=c++14 -stdlib=libc++
#0 0x0157a335 llvm::sys::PrintStackTrace(llvm::raw_ostream&)
(/home/brd/soft/clang+llvm-3.9.0-x86_64-linux-gnu-ubuntu-14.04/bin/clang-3.9+0x157a335)
#1 0x0157a996 SignalHandler(int)
(/home/brd/soft/clang+llvm-3.9.0-x86_64-linux-gnu-ubuntu-14.04/bin/clang-3.9+0x157a996)
#2 0x7fbecb4e1330 __restore_rt
(/lib/x86_64-linux-gnu/libpthread.so.0+0x10330)
#3 0x02923870
clang::NestedNameSpecifier::containsUnexpandedParameterPack() const
(/home/brd/soft/clang+llvm-3.9.0-x86_64-linux-gnu-ubuntu-14.04/bin/clang-3.9+0x2923870)
#4 0x027cc7f5
clang::ASTContext::getDependentNameType(clang::ElaboratedTypeKeyword,
clang::NestedNameSpecifier*, clang::IdentifierInfo const*, clang::QualType)
const
(/home/brd/soft/clang+llvm-3.9.0-x86_64-linux-gnu-ubuntu-14.04/bin/clang-3.9+0x27cc7f5)
#5 0x024cced7
clang::Sema::CheckTemplateTypeArgument(clang::TemplateTypeParmDecl*,
clang::TemplateArgumentLoc&, llvm::SmallVectorImpl&)
(/home/brd/soft/clang+llvm-3.9.0-x86_64-linux-gnu-ubuntu-14.04/bin/clang-3.9+0x24cced7)
#6 0x024cdefd clang::Sema::CheckTemplateArgument(clang::NamedDecl*,
clang::TemplateArgumentLoc&, clang::NamedDecl*, clang::SourceLocation,
clang::SourceLocation, unsigned int,
llvm::SmallVectorImpl&,
clang::Sema::CheckTemplateArgumentKind)
(/home/brd/soft/clang+llvm-3.9.0-x86_64-linux-gnu-ubuntu-14.04/bin/clang-3.9+0x24cdefd)
#7 0x024c61f1
clang::Sema::CheckTemplateArgumentList(clang::TemplateDecl*,
clang::SourceLocation, clang::TemplateArgumentListInfo&, bool,
llvm::SmallVectorImpl&)
(/home/brd/soft/clang+llvm-3.9.0-x86_64-linux-gnu-ubuntu-14.04/bin/clang-3.9+0x24c61f1)
#8 0x0250b5af
clang::Sema::SubstituteExplicitTemplateArguments(clang::FunctionTemplateDecl*,
clang::TemplateArgumentListInfo&,
llvm::SmallVectorImpl&,
llvm::SmallVectorImpl&, clang::QualType*,
clang::sema::TemplateDeductionInfo&)
(/home/brd/soft/clang+llvm-3.9.0-x86_64-linux-gnu-ubuntu-14.04/bin/clang-3.9+0x250b5af)
#9 0x0250dbd0
clang::Sema::DeduceTemplateArguments(clang::FunctionTemplateDecl*,
clang::TemplateArgumentListInfo*, llvm::ArrayRef,
clang::FunctionDecl*&, clang::sema::TemplateDeductionInfo&, bool)
(/home/brd/soft/clang+llvm-3.9.0-x86_64-linux-gnu-ubuntu-14.04/bin/clang-3.9+0x250dbd0)
#10 0x0245cccd
clang::Sema::AddTemplateOverloadCandidate(clang::FunctionTemplateDecl*,
clang::DeclAccessPair, clang::TemplateArgumentListInfo*,
llvm::ArrayRef, clang::OverloadCandidateSet&, bool, bool)
(/home/brd/soft/clang+llvm-3.9.0-x86_64-linux-gnu-ubuntu-14.04/bin/clang-3.9+0x245cccd)
#11 0x0246dfbf
clang::Sema::AddOverloadedCallCandidates(clang::UnresolvedLookupExpr*,
llvm::ArrayRef, clang::OverloadCandidateSet&, bool)
(/home/brd/soft/clang+llvm-3.9.0-x86_64-linux-gnu-ubuntu-14.04/bin/clang-3.9+0x246dfbf)
#12 0x0246e180 clang::Sema::buildOverloadedCallSet(clang::Scope*,
clang::Expr*, clang::UnresolvedLookupExpr*,
llvm::MutableArrayRef, clang::SourceLocation,
clang::OverloadCandidateSet*, clang::ActionResult*)
(/home/brd/soft/clang+llvm-3.9.0-x86_64-linux-gnu-ubuntu-14.04/bin/clang-3.9+0x246e180)
#13 0x0246e4aa clang::Sema::BuildOverloadedCallExpr(clang::Scope*,
clang::Expr*, clang::UnresolvedLookupExpr*, clang::SourceLocation,
llvm::MutableArrayRef, clang::SourceLocation, clang::Expr*, bool,
bool)
(/home/brd/soft/clang+llvm-3.9.0-x86_64-linux-gnu-ubuntu-14.04/bin/clang-3.9+0x246e4aa)
#14 0x022c9e42 clang::Sema::ActOnCallExpr(clang::Scope*, clang::Expr*,
clang::SourceLocation, llvm::MutableArrayRef,
clang::SourceLocation, clang::Expr*, bool)
(/home/brd/soft/clang+llvm-3.9.0-x86_64-linux-gnu-ubuntu-14.04/bin/clang-3.9+0x22c9e42)
#15 0x02547c79 clang::TreeTransform<(anonymous
namespace)::TemplateInstantiator>::TransformCallExpr(clang::CallExpr*)
(/home/brd/soft/clang+llvm-3.9.0-x86_64-linux-gnu-ubuntu-14.04/bin/clang-3.9+0x2547c79)
#16 0x02541668 clang::TreeTransform<(anonymous
namespace)::TemplateInstantiator>::TransformStmt(clang::Stmt*)
(/home/brd/soft/clang+llvm-3.9.0-x86_64-linux-gnu-ubuntu-14.04/bin/clang-3.9+0x2541668)
#17 0x02552271 clang::TreeTransform<(anonymous
[llvm-bugs] [Bug 31164] New: x64 backend creates unnecessary setcc instructions
https://llvm.org/bugs/show_bug.cgi?id=31164
Bug ID: 31164
Summary: x64 backend creates unnecessary setcc instructions
Product: tools
Version: trunk
Hardware: PC
OS: Linux
Status: NEW
Severity: normal
Priority: P
Component: llc
Assignee: unassignedb...@nondot.org
Reporter: e...@arbat.com
CC: llvm-bugs@lists.llvm.org
Classification: Unclassified
An i1 phi is cleverly placed in the processor status word, but then when we
want to branch on it, it is materialized in a setcc-testb-je sequence instead
of just a jne instruction.
Input .ll:
define i64 @foo(i64, i64, i64) {
%4 = icmp eq i64 %0, 0
br i1 %4, label %left, label %right
left:
%l = icmp eq i64 %1, 1
br label %join
right:
%r = icmp eq i64 %1, %2
br label %join
join:
%lr = phi i1 [%l, %left], [%r, %right]
%iszero = icmp eq i1 %lr, 0
br i1 %iszero, label %left2, label %right2
left2:
ret i64 %2
right2:
ret i64 %0
}
Output x64
foo:# @foo
.cfi_startproc
# BB#0:
testq %rdi, %rdi
je .LBB0_1
# BB#2: # %right
cmpq%rdx, %rsi
jmp .LBB0_3
.LBB0_1:# %left
cmpq$1, %rsi
.LBB0_3:# %join
sete%al
testb %al, %al
je .LBB0_4
# BB#5: # %right2
movq%rdi, %rax
retq
.LBB0_4:# %left2
movq%rdx, %rax
retq
See http://pastebin.com/UinMkFr1 and http://pastebin.com/Su1b1mxj for syntax
coloured versions.
The sete and testb instructions are superfluous.
--
You are receiving this mail because:
You are on the CC list for the bug.
___
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
