similar to: [LLVMdev] llvm pass INITIALIZE

Hi, I'm attempting to do some alias analysis & other memory inspection. I've written a pointless AliasAnalysis pass (that says everything must alias) to attempt to verify that my pass is getting picked up & run by opt. I run opt with: opt -load ~/Applications/llvm/lib/MustAA.so -must-aa -aa-eval -debug < trace0.ll I see my pass being initialized, but never being called (I see

I used LLVM already existent passes with no problems till I tried to use MemDepPrinter.cpp : http://llvm.org/doxygen/MemDepPrinter_8cpp_source.html. I got the following segfault: llvm[0]: Compiling MyMemDepPrinter.cpp for Release+Asserts build (PIC) llvm[0]: Linking Release+Asserts Loadable Module MyMemDepPrinter.so WARNING: You're attempting to print out a bitcode file. This

Hi Matthew, Did you add your alias analysis pass initializeEverythingMustAliasPass() into llvm::initializeAnalysis(PassRegistry &Registry) {} ? This will initialize it linked into the Analysis library. thanks, chen On Apr 28, 2014, at 8:43 PM, Jingyue Wu <jingyue at google.com> wrote: > > > > On Thu, Apr 24, 2014 at 4:38 PM, Matthew O'Connor <thegreendragon at

Hello everyone, I have written several complex passes till now, but I cannot write a MachineFunctionPass pass. It just gives me segfault. Hence I reduced the pass to the following form : using namespace llvm; namespace { class CFGexplorator : public MachineFunctionPass { public: static char ID; // Pass identification, replacement for typeid CFGexplorator() : MachineFunctionPass(ID)

I found that : "Code generator passes are registered and initialized specially by TargetMachine::addPassesToEmitFile and similar routines, so they cannot generally be run from the *opt* or *bugpoint* commands."...So how I can run a MachineFunctionPass? In the end, I just want to apply a DFS on a CFG of a function. And I need oriented edges and wanted to use

Hello LLVMdev, I’m using LLVM to do static analysis exclusively (without any code generation). To implement this analysis, I’m using multiple address spaces to disambiguate the purpose of the pointed memory. Since address spaces never alias in my model, I set on to implement an alias analysis pass that would exactly provide this information, as I’m seeing a couple of otherwise dead store that

Hi everyone, I am wondering what is the right way to register an alias analysis pass. My alias analysis pass is implemented in a dynamic load module. As I can see, a general way to register a pass looks like: llvm::RegisterPass<Analyzer> X("rci", "RCI Pass"); However, registering a type-based analysis uses: INITIALIZE_AG_PASS(...) I think, registering an

Hello all, I am working with someone else's LLVM code, which is about 8 months old. Part of this pass involves AliasAnalysis, and I'm getting the above assertion when the pass completes. The dump is as follows: -------------------------------- Program received signal SIGABRT, Aborted. 0xf7fdf430 in __kernel_vsyscall () (gdb) bt #0 0xf7fdf430 in __kernel_vsyscall () #1 0xf602e921 in

Yes, it is a pass. Here is a very general overview of the file structure as far as the AA is concerned. LLVM is not my strong-suit, I do hardware simulators, not compilers. using namespace llvm; char RelRecovery::ID = 0; static RegisterPass<RelRecovery> X("relRecovery", "Reliability transformation for lightweight recovery"); void

Hello all, I really, really, really need your help. This is my third email now, please don't ignore me! I understand this must be a trivial thing, but I've ground to a halt, and REALLY need some guidance. Please see below for the context of my problem. I'm not trying to be a waste of time, but I'm desperate here. I have a getAnalysisUsage method which does the following

Griffin Wright wrote: > > Hello all, > > I really, really, really need your help. This is my third email now, > please don't ignore me! I understand this must be a trivial thing, but > I've ground to a halt, and REALLY need some guidance. Please see below for > the context of my problem. I'm not trying to be a waste of time, but I'm > desperate here.

Hello ! I have a toolchain of two passes. First one, is a transformation pass that should add metadata to some structures (instructions/variables) and the second pass is an analyzing pass which needs to access the added metadata. The problem is with my adding metadata transformation pass. There might be two problems(or both): First, maybe I don't add correctly metadata.

Hello everyone, I am trying to write a trivial adding metadata transformation pass. I have Test.bc the target on which I want to perform transformations. First, I compile with clang++ my target program and I have my resulting Test.bc. I save it under Test1.bc. As a particular example, I have " %myDimension = alloca %class.ThreeDimension, align 4 ". Second, I apply my pass :

Dear All, What is the appropriate way to create analysis groups when using mainline LLVM? Following the directions in the "Writing an LLVM Pass" doesn't appear to work, and the INITIALIZE_ANALYSIS_GROUP() and INITIALIZE_AG_PASS() macros seem to each call a function defined by the other. -- John T.

Hi everyone, I am wondering what is the right way to register an alias analysis pass. My alias analysis pass is implemented in a dynamic load module. As I can see, a general way to register a pass looks like: llvm::RegisterPass<Analyzer> X("rci", "RCI Pass"); However, registering a type-based analysis uses: INITIALIZE_AG_PASS(...) I think, registering an

Hello fellow LLVM-ers, As those of you who follow llvm-commits have probably noticed, I've been hard at work reworking our pass infrastructure to perform pass initialization/registration without static constructors. This is a boon for clients of the LLVM libraries, as they can now control when and if initialization occurs, as opposed to being forced to incur the initialization cost at

Forgot to reply to the list. -- John T. -------- Original Message -------- Subject: Re: [LLVMdev] MAJOR API CHANGE: Pass initialization without static constructors Date: Fri, 8 Oct 2010 15:05:21 -0500 From: John Criswell <criswell at illinois.edu> To: Owen Anderson <resistor at mac.com> On 10/8/10 1:29 PM, Owen Anderson wrote: > Hello fellow LLVM-ers, > > As

Duncan, thanks for your answer! > In order to perform this transform the optimizers would have to work out > that sample does not modify any global state.  This cannot be done without > knowing the definition of sample, but you only provide a declaration. Which is why I am trying to supply this additional information in a custom alias analysis pass, but it doesn't seem to work. (The

Hi Stephan, > intrinsic float4 sample(int tex, float2 tc); > > float4 main(int tex, float2 tc) > { > float4 x = sample(tex, tc); > return 0.0; > } without additional information it would be wrong to remove the call to sample because it might write to a global variable. > As you can see, the call to the sample function is still present, > although the actual value

On Thu, Jan 15, 2015 at 1:26 PM, Nick Lewycky <nlewycky at google.com> wrote: > On 15 January 2015 at 13:10, Daniel Berlin <dberlin at dberlin.org> wrote: > >> Yes. >> I've attached an updated patch that does the following: >> >> 1. Fixes the partialalias of globals/arguments >> 2. Enables partialalias for cases where nothing has been unified to