similar to: [LLVMdev] Creating and implementing an analysis group out of tree

Hi Eli, Thanks for the link! I'm able to compile a standalone pass outside of the source tree, and I'm obviously ably to compile analysis groups inside the source tree. However, the problem comes when I try to do what the tutorial suggests to create an analysis group outside of the source tree. My understanding is that building out of tree requires different methods to register the

The attached patch didn't fix it, but when I also changed another == to = a few lines further down it seemed to work. Jeremy On Tue, Jun 24, 2014 at 10:07 PM, Justin Bogner <mail at justinbogner.com> wrote: > Jeremy Salwen <jeremysalwen at gmail.com> writes: > > I followed the LNT quickstart guide > > (http://llvm.org/docs/lnt/quickstart.html) on my debian

Hello All, I have some data tied to the basic blocks in a function, and after inlining that function, I'd like to recover that data in the inlined version. Is there some way to match up the inlined version of the function with the original basic blocks? Thanks, Jeremy -------------- next part -------------- An HTML attachment was scrubbed... URL:

Hello all, I'm wondering as to the status of control flow profiling in llvm. From what I can gather there was an old system (using opt -insert-edge-profiling and the like) which was removed in this commit llvm.org/viewvc/llvm-project?view=revision&revision=191835 . The commit message mentions "modern PGO efforts", but I can't find anything in the source tree or

Hi Hal, Thank you for your email, that makes a lot of sense to me. I am working on some tools to use memory profiling to speculatively replace memory loads and stores with value forwarding in hardware implementations. I'd like to compare the profiled data to static alias analysis, so it would be super useful if there was a way to answer the questions about aliasing across backedges that

Thanks Daniel! I think you've cleared up some of my misconceptions, but I still am a bit confused about some of the corner cases. Suppose we had something like this std::vector<int> A(100); > int* x,y; > x=&A[0]; > for(int i=0; i<100; i++) { > y=&A[i]; > *y=*x; > x=&A[i+1]; > } > Would the load and store instructions be MustAlias? I

I followed the LNT quickstart guide ( http://llvm.org/docs/lnt/quickstart.html) on my debian machine, and it seems to have a problem running. When I consult test.log, every single test has a section like this # Pretend that the reference output was produced instantaneously. > cp /home/jeremy/hardware/llvm/test-suite/instant.txt > Output/sse.expandfft.out-nat.time >

Hi Daniel, Sorry for taking so long to respond. I spoke with a colleague more familiar with llvm who thought he could clear up my confusion, but we both came out of the conversation confused. I will try my best to explain the ambiguity. In an DAG, alias queries would be completely unambiguous. Every instruction would only be executed once, and every SSA value really would have a single static

Hey Daniel, Thanks again for the help. I'm still a bit confused about the interface to the alias analysis. It seems like we are talking about different interfaces. Has it changed from what the documentation says? As far as I can tell, the documentation takes a specific Value*, and no information about which dynamic execution it is talking about. When you say "Right. It's a

Hello all, I've read the documentation on alias analysis, and I think I understand it literally, but I just want to be sure, because it seems a bit strange. As it says on this web page <http://llvm.org/docs/AliasAnalysis.html>, The MayAlias response is used whenever the two pointers might refer to the > same object. > > The PartialAlias response is used when the two memory

On Thu, Aug 14, 2014 at 6:37 AM, Daniel Berlin <dberlin at dberlin.org> wrote: > On Wed, Aug 13, 2014 at 8:35 PM, Jeremy Salwen <jeremysalwen at gmail.com> wrote: >> Hey Daniel, >> >> Thanks again for the help. I'm still a bit confused about the interface to >> the alias analysis. It seems like we are talking about different >> interfaces. >

I have seen INITIALIZE_PASS(LiveVariables, "livevars", "Live Variable Analysis", false, false); in the llvm/lib/codegen/LiveVariables.cpp, where LiveVariables is a subclass of MachineFunctionPass, and #define INITIALIZE_PASS(passName, arg, name, cfg, analysis) \ static RegisterPass<passName> passName ## _info(arg, name, cfg, analysis) in PassSupport.h Is this code used

Hi list, I have several questions about LLVM pass. 1) Is building a custom LLVM pass out-of-source not recommended? The official document only contains instructions about in-source build (http://llvm.org/docs/WritingAnLLVMPass.html <http://llvm.org/docs/WritingAnLLVMPass.html>). 2) opt (ver >= 4) with custom pass libraries does not work as before. When I have a simple custom LLVM pass

Hi, I want to create a copy of DependenceAnalysis downstream. At first I tried the HelloWorld tutorial. But then I figured, maybe I should move on to create my pass by mimicking some other pass in LLVM. So, I tried copying DependenceAnalysis itself [1] Although after a lot of time of trying it compiled, I'm pretty sure I have done it completely in the wrong way. What is more, I can't

Hi everyone, I have a question that seems simple, but has been confounding me for several hours. I'd like to debug a binary produced with LLVM. For the life of me, I can't get any symbols into gdb and llvm-db won't even start the program nor load any useful information about it. Here's my current strategy (which isn't working): llvm-gcc -g -O0 -c -emit-llvm

On Fri, Sep 6, 2013 at 9:35 PM, Pete Cooper <peter_cooper at apple.com> wrote: > Hi Eli > > I don’t know a workaround which will keep support for coloured terminal > output, but if you can live without it then passing these to configure > should work. > > --enable-curses=no > --enable-terminfo=no > > Re-doing the configure & build in a new directory solved

Hello, I'm trying to build some code that relies on LLVM as a library (out-of-source). It uses llvm-config for the compilation & linkage flags. Some recent change breaks it, and with the newest trunk LLVM I get: /home/eliben/llvm/build_llvm_svn_rw/Debug+Asserts/lib/libLLVMSupport.a(Process.o): In function `terminalHasColors':

Hello, Clang is used before LLVM to generate IR, and is it possible to use it in a reverse way, i.e., after running LLVM passes, we can run clang again to use LLVM pass results? If so, should I write clang functions in LLVM passes, or write LLVM pass requirements in clang? Thank you in advance! Padefic -------------- next part -------------- An HTML attachment was scrubbed... URL:

I think you probably need to pass -O0 to llvm-ld. The link-time optimizations are probably killing your debug info. --Owen On May 5, 2008, at 8:21 PM, Mark Oskin wrote: > Hi everyone again, > > I did discover the following works (see below). However, does anyone > know of the "proper" way with LLVM? > > llvm-gcc -g -c -emit-llvm helloworld.c > opt

Hi Rui, I still fail to enable the lld in my Uefi firmware build to replace ld, and I found it is related to the wrong symbol values in the .data section, which are pointed by R_X86_64_64 relocation entries. I need your advices. My firmware uses a linker script https://github.com/tianocore/edk2/blob/master/BaseTools/Scripts/GccBase.lds to do the linking. We use position independent code with