similar to: [LLVMdev] StringMap question

In SmallVector.h: class SmallVectorBase { *protected:* void *BeginX, *EndX, *CapacityX; *protected:* SmallVectorBase(void *FirstEl, size_t Size) : BeginX(FirstEl), EndX(FirstEl), CapacityX((char*)FirstEl+Size) {} In StringMap.h: class StringMapImpl { *protected:* // Array of NumBuckets pointers to entries, null pointers are holes. // TheTable[NumBuckets] contains a sentinel value

This series of patches address several issues causing memory usage to grow indefinetely on a long lived process. These are not convenional leaks -- memory would have been freed when the LLVM context or/and JIT engine is destroyed -- but for as long as they aren't the memory is usage effectively ubounded. The issues were found using valgrind with '--show-reachable=yes' option: 1.

Thanks, I tried calling ResetAllOptionOccurrences after the run like this… // Compile the module TimeCompilations times to give better compile time // metrics. for (unsigned I = TimeCompilations; I; --I) if (int RetVal = compileModule(argv, Context)) return RetVal; if (YamlFile) YamlFile->keep(); cl::ResetAllOptionOccurrences(); return 0; } Unfortunately

From: José Fonseca <jfonseca at vmware.com> When the hash function uses object pointers all free entries eventually become tombstones as they are used at least once, regardless of the size. DenseMap cannot function with zero empty keys, so it double size to get get ridof the tombstones. However DenseMap never shrinks automatically unless it is cleared, so the net result is that certain

On OS X, this test: ------------------------------------------------------ #include <llvm/ADT/DenseSet.h> #include <new> #include <stdio.h> int main(int argc, char** argv) { llvm::DenseSet<unsigned> set(2); set.insert(0); set.insert(1); if (set.count(2)) printf("error\n"); return 0; } ------------------------------------------------------ saved

I'm trying to set up some automated testing, and I'd like to have multiple instances of ExecutionEngines, so that the state from the first test doesn't alter the second state. Right now I'm doing something along the lines of: Module *emptyModule = new Module("emptyModule"); ExecutionEngine executionEngine = ExecutionEngine::create(emptyModule);

Hi: I'm building a small tool on top of LLVM Core Library. ``` LLVMContext context; SMDiagnostic diag; Module *M = parseIRFile(InputIR, diag, context).get(); if (M == nullptr) { diag.print("LLVM", errs()); exit(-1); } assert(M->isMaterialized() && "Module not materialized!"); PointerType *ArrayPtrTy =

Hello, Lately, I have been working on moving Debian & Ubuntu packages to a stage2 build. This means that, instead of shipping llvm-toolchain packages built with gcc, we are rebuilding everything a second time using the newly built clang. Now, when pushed to Debian, it caused some unexpected issues in particular with rust reported here:

I get the error below (and have for a couple weeks now) when trying to build llvm-gcc on Ubuntu Feisty. In the meantime, on Ubuntu Gutsy, everything has been building fine. Both are release builds for x86. John cc1: StringMap.cpp:177: void llvm::StringMapImpl::RemoveKey(llvm::StringMapEntryBase*): Assertion `V == V2 && "Didn't find key?"' failed.

Hello everyone! I've found weird lines of code in tools/libclang/CIndex.cpp file in clang sources. Function clang_tokenize holds the following lines: *Tokens = (CXToken *)malloc(sizeof(CXToken) * CXTokens.size()); memmove(*Tokens, CXTokens.data(), sizeof(CXToken) * CXTokens.size()); Since malloc function can return null pointer and there are no checks for this case, memmove invocation

I'm trying to link a program with LLVM on Linux. I've managed to take out the test files from 'llvm-config --libs' but now it's getting errors with missing symbols. The same program successfully links with the same version of LLVM (3.8) on Windows, so it's not a problem with the code per se, it's some kind of configuration issue. Any ideas? g++ -std=c++11

Hi Russel, Can you try compiling aklo.o using -fno-rtti prior to linking? On Thu, Mar 24, 2016 at 11:44 PM, Russell Wallace via llvm-dev <llvm-dev at lists.llvm.org> wrote: > I'm trying to link a program with LLVM on Linux. I've managed to take out > the test files from 'llvm-config --libs' but now it's getting errors with > missing symbols. The same program

Tried that just now, same result. On Fri, Mar 25, 2016 at 6:58 AM, Snehasish Kumar <kumar.snehasish at gmail.com> wrote: > Hi Russel, > > Can you try compiling aklo.o using -fno-rtti prior to linking? > > On Thu, Mar 24, 2016 at 11:44 PM, Russell Wallace via llvm-dev > <llvm-dev at lists.llvm.org> wrote: > > I'm trying to link a program with LLVM on Linux.

Since the log tells you that llvm::outs() is not found, and since it is supposed to be found in libLLVMSupport.a, you can try: $ nm /home/a/build//lib/libLLVMSupport.a | grep ZN4llvm4outsEv ... 0000000000002070 T __ZN4llvm4outsEv ... The output should have the symbol defined. -- Mehdi > On Mar 25, 2016, at 12:18 AM, Russell Wallace via llvm-dev <llvm-dev at lists.llvm.org> wrote:

I'm getting the following when building llvm-gcc with an optimized set of LLVM libraries: /ptmp/dag/build.llvm-gcc-4.2.trunk.official.opt/x86_64-unknown-linux-gnu/./gcc/xgcc -B/ptmp/dag/build.llvm-gcc-4.2.trunk.official.opt/x86_64-unknown-linux-gnu/./gcc/ -B/cray/iss/compiler/cost/tools/llvm-tools/llvm/install.trunk.official.opt/x86_64-unknown-linux-gnu/x86_64-unknown-linux-gnu/bin/

Yeah, seems to. a at a-VirtualBox:~$ nm /home/a/build//lib/libLLVMSupport.a | grep ZN4llvm4outsEv U _ZN4llvm4outsEv U _ZN4llvm4outsEv 0000000000000000 b _ZGVZN4llvm4outsEvE1S 0000000000000000 T _ZN4llvm4outsEv 0000000000000000 r _ZZN4llvm4outsEvE19__PRETTY_FUNCTION__ 0000000000000000 b _ZZN4llvm4outsEvE1S U _ZN4llvm4outsEv I can try building

Looking again at your link line, I think static library has to appear after their uses. So try to put the .o before the list of library. -- Mehdi > On Mar 25, 2016, at 12:41 AM, Russell Wallace <russell.wallace at gmail.com> wrote: > > Yeah, seems to. > > a at a-VirtualBox:~$ nm /home/a/build//lib/libLLVMSupport.a | grep ZN4llvm4outsEv > U

Hi, I'm trying to autotune a good sequence of IR optimization passes and I seem to run into segfaults in opt (in LLVM5) with certain pass orderings. Is this expected behavior? If so, what would be the recommended way of determining pass dependencies so that I can encode them into the tuner? The test program can be found here: https://gist.github.com/kavon/92d153cdd54ce9b77162af3af47d4c95

While building a package, I see the following: * checking R code for possible problems ... NOTE cheat.fit: no visible binding for global variable 'Zobs' plot.jml: no visible binding for global variable 'Var1' I see the issue has come up before, but I'm having a hard time discerning how solutions applied elsewhere would apply here. The entire code for both functions is below,

Hi llvm-dev, Our code base has a ancient copy of llvm (ver 3.5.1), and it uses the LLVM code gen for some domain-specific language. The previous dev left a global lock around the usage of LLVM code gen stating that because LLVM code gen can only be accessed single-threaded it needs to be protected with this global lock. But now this lock has caused some perf issues as we pretty much lose