similar to: [LLVMdev] adding a codegen pass into llvm

On 2/7/13 5:49 AM, marwayusuf at feng.bu.edu.eg wrote: > Dear All > in this - quite old - message > http://lists.cs.uiuc.edu/pipermail/llvmdev/2011-January/037496.html > I followed all the steps except the last one, as I don't think my pass depends on anything else, and I didn't fully understand it. > And I tried to rebuild the llvm to make it take my pass into account. I

Thanks a million for replying. Your answer helped me. I inspected the errors more closely - with the help of my supervisor - and managed to solve some problems. Finally the making was successful, by just following steps 1,2 and 3. I found that one of the problems was that I modify in a wrong Passes.h, as there are 2. My question now is: How I can apply this pass? (call, use ... etc) >From what

Dear All Is there is a way I can get the original hogh level code (e.g c++ code) of an IR function within MCJIT? Regards, Marwa Yusuf Teaching Assistant - Computer Engineering Department Faculty of Engineering - Benha University E-JUST PhD Student Computer Science & Engineering Dept. -------------- next part -------------- An HTML attachment was scrubbed... URL:

Dear All I need to create a pass that is triggered to run at a specific point during runtime. Also, this pass needs to have access to object file and to be able to modify the original IR. Any idea how to make such pass? Regards, Marwa Yusuf Teaching Assistant - Computer Engineering Department Faculty of Engineering - Benha University E-JUST PhD Student Computer Science & Engineering Dept.

Dear All Is there is a way to recompile a function at runtime? It seems that I need to -like - invalidate the recompiled function to force the MCJIT to recompile it, cause when I just call runFunction, it neglects any modifications in the function code. Thanks in advance. Regards, Marwa Yusuf Teaching Assistant - Computer Engineering Department Faculty of Engineering - Benha University E-JUST

Trying to run make to build llvm, I faced the following error: Linking CXX shared library ../../lib/libLTO.so collect2: error: ld returned 1 exit status make[2]: *** [lib/libLTO.so.3.8.0svn] Error 1 make[1]: *** [tools/lto/CMakeFiles/LTO.dir/all] Error 2 make: *** [all] Error 2 So, what's the problem here? Regards, Marwa Yusuf Teaching Assistant - Computer Engineering Department

Dear all I have a question and I really appreciate any help. I'm trying to extract variable values at interrupt of some program. And I want to do that from JIT. The situation is that the JIT will interrupt at certain point of executing some code and tries to get the variables' values. Any help about that? Regards, Marwa Yusuf Teaching Assistant - Computer Engineering Department Faculty

Hi Marwa, MCJIT doesn't provide any support for recompilation. The best you can do is use multiple MCJIT instances and destroy whichever one contains the function "foo" that you want to recompile. MCJIT does not do any dependence tracking either, so you'll also need to destroy and recompile any direct callers of "foo" (and their direct callers, and so on). If you know

Dear All I'm making an instrumentation pass. The pass is supposed to modify the given IR in a specefic way. One of the required modifications is to insert a call to a function at a specific location. This is the signature of the called function: void myclass::foo(Function *f, BasicBlock* b) This function's prototype is in an foofile.h file in include/llvm And the function

Dear All I'm making a transformation pass that inserts a new basic block at the start of a loop. However when I try to change predecessor/successor relations, it does not consider the new block in the loop at all. So I got that just inserting a loop in a function before another loop is not enough. So how exactly to do this job? Regards, Marwa Yusuf Teaching Assistant - Computer Engineering

Dear All I've made un update to the llvm to the latest. At configuration, I've met some new errors to me, like it needs lldb, lld, dragonegg, llgo and so. However, I've checked them and configuration went on. Now I'm stuck at build stage, specifically at dragonegg. First it could not find gmp.h, I've solved this by installing libgmp-dev. Now, it can't find auto-host.h.

This is effectively exactly what a debugger does when introspecting the state of a running program. Your JIT'ed code will need to use the program's debug information to get what it wants. -Jim On Apr 3, 2013, at 5:24 AM, marwayusuf at feng.bu.edu.eg wrote: > Dear all > I have a question and I really appreciate any help. > I'm trying to extract variable values at interrupt

Dear All My MSc was about implementing some kind of On stack replacement in LLVM to enhance security by obfuscating running code periodically. I've tried to implement this job in JIT, but I couldn't. So I implemented it in interpreter, as an initial step. However, now I'm trying to make it in JIT. My first step is to extract function state at an arbitrary point. So, I interrupt

Dear All I'm trying to create storeInst using StoreInst *SI = new StoreInst(val, AI, BB); AI is an alloca instruction previously created, BB is the basicBlock I want to put instructionsi in. Val is the value to store. I've created val from a genericValue like this: Value* val = ConstantInt::get(getGlobalContext(), Result.IntVal); Result is the genericValue However I keep getting

Dear All Here is another question. I'm trying now t install vmkit. According to this http://vmkit.llvm.org/index.html I have to install classpath. But in trying to configure it, I get the following error: configure: error: GTK+ peers requested but no X library available googling, I found that I can use --disable-gtk-peer --disable-gconf-peer to come over this error. Then, I got this error

Dear All I need to read the stackmap section at runtime of the inspected application. I've read the documentation of stackmap and found this paragraph: "A JIT compiler can easily access this section by providing its own memory manager via the LLVM C API LLVMCreateSimpleMCJITMemoryManager(). When creating the memory manager, the JIT provides a callback:

Dear All I have some questions about how to use patchpoint intrinsic: 1- The function that is provided as argument to the intrinsic call, where does it belong? I'm making a pass to change a given IR file and add call to patchpoint. Where should I put this function? In the pass or in the same IR file or in another IR and include it or what? 2- When actually this code is patched into the

Dear All Here is another question and any help is appreciated. Is there any way to get the state of the running program interrupt of JIT? I've managed to interrupt JIT while running some function and direct it to some function that plays as a handler. However, at this function I need to get the exact state of the running function. By state I mean which instruction, which basicblock, variables

Dear All, Did anyone read these 2 books? https://www.packtpub.com/application-development/llvm-cookbook LLVM Cookbook | PACKT Books Over 80 engaging recipes that will help you build a compiler frontend, optimizer, and code generator using LLVM Read more...<https://www.packtpub.com/application-development/llvm-cookbook>

Dear all, I'm trying to get openmp to work with llvm, either with static compiler or with jit, however with no success. I've followed this: http://blog.llvm.org/2015/05/openmp-support_22.html [http://llvm.org/img/DragonSmall.png]<http://blog.llvm.org/2015/05/openmp-support_22.html> LLVM Project Blog: OpenMP Support OpenMP support in Clang compiler is completed! Every pragma and