similar to: [LLVMdev] whether these transformations are doable and how?

Thanks for all the replies which are really helpful. one more question regarding transformation: 4. Given an instruction %x1 = I1, can I replace the rhs "I1" with "I2" and get a new instruction %x1 = I2? Alternatively, I can add a new instruction %x2 = I2, and replace all uses of %x1 with %x2, and then delete %x1 = I1, but it seems the former is simpler or faster if

> 3.  can I modify a function to take extra formal parameters? can I update > all calls of the original function to take extra actual paramters?  The > function might be called across multiple modules.   It seems this has to be > done at both ModulePass and FunctionPass levels. > Check out DeadArgumentElimination pass. It does opposite of what you want, but you'll get an idea.

Ning Wang wrote: > Hi folk, > > I'm a newbie to llvm, please first forgive my naive questions. I want > to instrument llvm code to do some run-time monitoring work. After > reading some of the llvm documentation, it begins clear to me that I > can do the instrumentation in a transformation pass. There are > several things I want to do in the transformation pass, but

On Wed, Apr 14, 2010 at 5:45 PM, Neal N. Wang <neal.wang at gmail.com> wrote: > Thanks for all the replies which are really helpful. > > one more question regarding transformation: > > 4.  Given an instruction  %x1 = I1,  can I replace the rhs "I1"  with  "I2" > and get a new instruction %x1 = I2?   Alternatively, I can add a new > instruction %x2 =

On 2011-08-09 16:48, John Criswell wrote: > On 8/9/11 6:49 AM, Bjorn Reese wrote: >> I have an optimization pass (FunctionPass) where I need to add global >> constructors. >> >> For cleaness sake I decided to add these in my own module. My module >> is created in my FunctionPass constructor: > > This is not how I would do it. A FunctionPass has

On 8/9/11 9:15 AM, Bjorn Reese wrote: > On 2011-08-09 16:48, John Criswell wrote: >> On 8/9/11 6:49 AM, Bjorn Reese wrote: >>> I have an optimization pass (FunctionPass) where I need to add global >>> constructors. >>> >>> For cleaness sake I decided to add these in my own module. My module >>> is created in my FunctionPass constructor:

Hello, I've recently changed AddressSanitizer (asan) compiler pass from ModulePass to FunctionPass and it could a bit of mayhem. The problem is that asan FunctionPass instruments a function foo, then foo gets inlined into bar, then bar gets instrumented and thus the code of foo gets instrumented twice (which causes run-time crash). This happens only at -O0; at -O1 we get the correct order of

Hi folk, In the Source Level Debugging document, llvm.dbg.declare takes two paramters. "The first argument is the alloca for the variable, cast to a {}*. The second argument is the llvm.dbg.variable containing the description of the variable." The second parameter corresponds to DIVariable, what is the corresponding LLVM class of the first parameter? I want to find which

Dear All, I'm having some problems using a function pass from a Module pass. My code is as follows: DominatorTree & domTree; ... Function &F = *I; // I is an interator from using Module::begin() ... domTree = getAnalysis<DominatorTree>(F); When I compile this code, I get the following error: /home/vadve/criswell/src/llvm22/include/llvm/Pass.h: In member function

Here's a basic patch which would solve it in sort of the same way as the other optimizations I was fixing (just special case the debug info & fix it up). I can work up a test case for this as well, or you can, if you like/this seems reasonable. On Tue, Oct 7, 2014 at 2:30 PM, Peter Collingbourne <peter at pcc.me.uk> wrote: > On Tue, Oct 07, 2014 at 12:20:55PM -0700, David

Hello all; So my goal is to insert some (self-written) function calls in the LLVM IR. I can achieve it with a ModulePass and using the getOrInsertFunction() call. For a ModulePass I follow the steps- 1. Compile the code (in which call instructions are to be inserted) to LLVM IR 2. Compile the file (which contains the *called* external function ) to LLVM IR 3. Link them together and run the

On Tue, Oct 7, 2014 at 2:51 PM, Peter Collingbourne <peter at pcc.me.uk> wrote: > Looks good, thanks! > > Can you write the test case, please? You probably have more experience > writing debug info tests than I do. > Sure - though how would I get the pre-dfsan .ll file to produce this behavior? I've tried compiling to a .ll file without dfsan, then feeling that .ll

On Oct 17, 2012, at 7:34 AM, Kostya Serebryany <kcc at google.com> wrote: > Hello, > > I've recently changed AddressSanitizer (asan) compiler pass from ModulePass to FunctionPass and it could a bit of mayhem. > > The problem is that asan FunctionPass instruments a function foo, then foo gets inlined into bar, then bar gets instrumented > and thus the code of foo

Hi everybody, I am writing an LLVM pass and I want to iterate over the whole module (including global variables), that's why I use ModulePass instead of FunctionPass. But I don't know how to do it. Using Module::iterator seams to iterate only over functions. But I need to iterate over all the Instructions in the module. How should I do such an iteration? Also, I would like to find all the

Hello, I am trying to add some thing into LLVM, while I encountered some problems. So my situation is that I have some information W, some transform passes may change it, after these passes, I need the new W. What I did is to create an analysis pass similar to scalar-evolution or loopinfo, I can get the information by using getAnalysis<W>(); and preserve this information W by using

On Nov 29, 2010, at 1:17 AM, Duncan Sands wrote: > Hi David, > >> Interesting -- LLVM does perform on the fly cleanups during inlining >> transformation -- this will make summary update precise. One thing I notice from >> the debug pass dump is that the 'deduce function attribute' pass happens before >> the clean up -- Is it intended? > > you are

On Sat, Jul 17, 2010 at 12:00 AM, Neal N. Wang <neal.wang at gmail.com> wrote: > Hi folk, > > In the Source Level Debugging document,  llvm.dbg.declare takes two > paramters.  "The first argument is the alloca for the variable, cast to a > {}*. The second argument is the llvm.dbg.variable containing the description > of the variable."  The second parameter

On Tue, Oct 7, 2014 at 12:18 PM, David Blaikie <dblaikie at gmail.com> wrote: > > > On Tue, Oct 7, 2014 at 12:10 PM, David Blaikie <dblaikie at gmail.com> wrote: > >> >> >> On Tue, Oct 7, 2014 at 11:48 AM, Peter Collingbourne <peter at pcc.me.uk> >> wrote: >> >>> On Tue, Oct 07, 2014 at 10:04:30AM -0700, David Blaikie wrote:

Hi, all: I wanted to use a FunctionPass (e.g. *MemoryDependenceAnalysis*) in a ModulePass, and then I used the method "getAnalysis<* MemoryDependenceAnalysis*>(llvm::Function *)" described at http://llvm.org/docs/WritingAnLLVMPass.html#ModulePass to get the FunctionPass. But , it still crashed when I invoked this pass in tool 'opt'. However, if I change my pass to

I am trying to write a ModulePass which requires PostDominator sets for every function in the module. Now finding post dominators is a function pass. The link on the llvm.org website says that : "Currently it is illegal for a ModulePass<http://llvm.org/docs/WritingAnLLVMPass.html#ModulePass>to require a FunctionPass <http://llvm.org/docs/WritingAnLLVMPass.html#FunctionPass>.