similar to: [LLVMdev] Invalid code generated for on-stack class compares

Hi All, Recently I am using clang++ (3.8.0) to compile some c++ programs and find that some type names will be changed after linking different modules. Here is a simple example: I have two files: main.cpp and b.cpp main.cpp #include <iostream> using namespace std; class A { public: virtual void f() { cout << "A:\tf\n"; } }; int main () { A *p = new A;

I created a plugin to add simple Annotations to VarDecls and FieldDecls, and write this modified AST out to a file. I notice that when I use clang to compile this file I get different code then when I use the source directly. In both cases I'm compiling with -O4. Can anyone explain this? Thanks define i32 @somefunc(i32 (i32)* %ptr) nounwind uwtable { entry: %ptr.addr = alloca i32 (i32)*,

> I created a plugin to add simple Annotations to VarDecls and > FieldDecls, and write this modified AST out to a file. I notice that > when I use clang to compile this file I get different code then when I > use the source directly. In both cases I'm compiling with -O4. Can > anyone explain this? This is how clang implements annotations on local variables, that's all.

On Wed, Dec 12, 2012 at 01:59:55PM -0800, Dan Gohman wrote: > On Wed, Dec 12, 2012 at 1:26 PM, Joerg Sonnenberger > > The original issue is that clang maps restrict on function arguments to > > NoAlias and that makes compares against the address of global variables > > false. Minimal test case: > > > > @y = external global i32 > > > > define zeroext i1

Hi Bjorn, > currently the following two lines of code with integer variables get compiled completely differently: > > if (d0<tmp) d0=tmp; there must be something else going on because this void bar(int); void foo(int d0, int tmp) { if (d0 < tmp) d0=tmp; bar(d0); } when compiled like this clang -S -O4 -o - sel.c gives this define void @foo(i32 %d0, i32 %tmp) nounwind

On Wed, Dec 12, 2012 at 2:45 PM, Joerg Sonnenberger <joerg at britannica.bec.de> wrote: > On Wed, Dec 12, 2012 at 01:59:55PM -0800, Dan Gohman wrote: >> On Wed, Dec 12, 2012 at 1:26 PM, Joerg Sonnenberger >> > The original issue is that clang maps restrict on function arguments to >> > NoAlias and that makes compares against the address of global variables >>

> On 2015-May-15, at 13:15, Teresa Johnson <tejohnson at google.com> wrote: > > What isn't clear to me is what all uses the available > externally linkage type currently - do you happen to know? It's used for the `inline` keyword in the C language. If you do a `git grep available_externally -- test/` inside a clang checkout you might find some other uses. $ cat

On Thu, Dec 8, 2011 at 4:26 PM, Eli Friedman <eli.friedman at gmail.com> wrote: > On Thu, Dec 8, 2011 at 2:11 PM, Vitor Luis Menezes <vitor at utexas.edu> wrote: >> We've got the following test case: >> >> >> class A { >> public: >>   int x; >>   A(int x) : x(x) {} >>   int hoo() {return 4;} >>   virtual int foo() {return x;}

Hi, I’m looking at what the LLVM compilation chain does to optimize the following functions using logical operators. The one using integer type (t1 and t3) are optimized, but not the ones using float type (t2 and t4), even when using -O3 or -Ofast. Why is that ? Thera are those kind of optimizations done in the LLVM source code? Thanks. bool t1(int v1) { return (2 + v1) == 7; } bool

I did an update about 1/2 hr ago, and instead of failing to link with -O4, I'm getting an error elsewhere in the compile. I think I need to stop chasing the bleeding edge. pandora 142% svn update At revision 147455. pandora 143% uname -a Darwin pandora.psychiatry.uiowa.edu 10.8.0 Darwin Kernel Version 10.8.0: Tue Jun 7 16:33:36 PDT 2011; root:xnu-1504.15.3~1/RELEASE_I386 i386 pandora 144%

I decided to pump some IR through the optimizers and was surprised to find that we don't choose to canonicalize any of the following functions to the other even though they are all roughly the same. define i32 @f1(i32 %x, i32 %m) #0 { entry: %and = and i32 %m, 32896 %tobool = icmp ne i32 %and, 0 %cond = select i1 %tobool, i32 8, i32 0 %shr = lshr i32 %x, %cond ret i32 %shr } ;

Hi, All. Does anybody know about ConstantExpr in llvm? What's it? Since it always appears after llvm optimization such as -O2 level, what is it supposed to be to codegen? I am wondering it represents constant value which can be determined or computed at compile-time(actually is link-time) to improve performance. Although we do not know the actual constant value util the object file is linked.

Hi, I found a mysterious behavior of LLVM optimizer. I compiled the following code by clang -emit-llvm -S: #include <stddef.h> static const struct t {char t[4]; char s;} p = {{}, 'a'}; char f() { return ((char*)&p)[offsetof(struct t, s)]; } then I obtained the following LLVM IR: %struct.t = type { [4 x i8], i8 } @p = constant %struct.t { [4 x i8] zeroinitializer, i8 97

Hi, I've been following the instructions on how to use the LLVM Gold plugin at http://llvm.org/docs/GoldPlugin.html while building an multiple versions of WebKit. The documentation hasn't been updated since 2010 and hasn't really matched my experiences, so I'd like to ask if I'm doing these steps incorrectly. What I'm trying to do is force all compilation steps to

> First, I create libjscore.a with ar, where I've changed the command > line to load the plugin: > ar cqs --plugin > /home/tmjackso/multicompiler/llvm-3.0/release/lib/LLVMgold.so > libjscore.a <list of object files> Note that you shouldn't need to pass --plugin to ar, it searches bfd-plugins. > Then I link jsc against libjscore.a: >

Hi- Here is the simple C++ function: ----------- void foo() { } ----------- Let's say, above function is compiled to generate LLVM IR with -g flag using the command line `clang++ -g -O0 -S -emit-llvm foo.cpp`, we get below IR ----------- ; ModuleID = 'foo.cpp' source_filename = "foo.cpp" target datalayout =

I am trying to generate PTX code for 'nbody' sample program's kernel (nbody_kernel.cu) using clang/LLVM version 3.2. The nbody CUDA program is available in Nvidia's SDK. I am referring to https://github.com/jholewinski/llvm-ptx-samples project. Following are my commands, clang++ -O4 -S -I/usr/local/cuda/include -emit-llvm -target nvptx64 nbody_kernel.cu -o nbody_kernel.ll

Hello since my broad RFC request didn't catch any responses, let me get a bit more into the nitty-gritty: I tried to get llvm (3.7) to optimize superflous allocas away, but so far I haven't figured out how. Is there no optimizer for this ? Should I adorn something with some attributes ? As far as I can tell no, but I am no llvm expert... For what I want to do, i will probably have a

I installed LLVM gold plugin and can ran the example in http://llvm.org/docs/GoldPlugin.html#example1. Then I tried to use it to build a big project. I set the environment variables like: export CC="/usr/local/bin/clang -use-gold-plugin -Wl, -plugin-opt=also-emit-llvm" export CXX="/usr/local/bin/clang++ -use-gold-plugin -Wl, -plugin-opt=also-emit-llvm" export

Confirmed, this is a bug. This define i32* @get_pntr(i32* %p) nounwind uwtable { entry: ret i32* %p } define void @store(i32* %p) noinline nounwind uwtable { entry: %call = call i32* @get_pntr(i32* %p) store i32 10, i32* %call, align 4 ret void } run through opt -functionattrs gets a 'readnone' on @store's %p. That's wrong, it clearly stores to it. The bug is due to