similar to: [LLVMdev] Crash with llc and vector code

Hi, Compiling the simple following function with llv (LLVM 2.8) : define <4 x i32> @foo(<4 x i32> %a, <4 x i32> %b) { %res = select <4 x i1> <i1 true, i1 false, i1 true, i1 true>, <4 x i32> %a, <4 x i32> %b ; ret <4 x i32> %res } gives : UNREACHABLE executed! 0 llc 0x0000000100927422 std::_Rb_tree<llvm::sys::Path,

I've hit this weird compiler error when building llvm/clang $ clang --version clang version 2.9 (trunk 125254) Target: x86_64-apple-darwin10 Thread model: posix Source rev is 125326 $ make llvm[1]: Compiling APFloat.cpp for Release build llvm[1]: Compiling APInt.cpp for Release build llvm[1]: Compiling APSInt.cpp for Release build llvm[1]: Compiling Allocator.cpp for Release build llvm[1]:

[snip] > The ABI boundary lowering requires types to be broken down further into > 'legal types' that can be mapped to registers. The secondary breakdown is > currently handled by TargetLowering::LowerCallTo() calling > getRegisterType() and getNumRegisters(). Most ABIs are defined in terms > of C types, not LLVM IR types, so there is a close connection between the C >

Hi > One would expect this, its a facility of the C++ language. The anonymous > namespace is, essentially, the same as declaring everything in it > static. That is, the symbols are not exported and not available for > linking. Yes, it was pretty clear after finding out that this isn't a linking error which i suspected... > > So for all those trying to add an analysis path:

Hey all, I need to use C bindings in a pass that I am authoring (specifically, I want to use the OCaml bindings that are tied to the C bindings to author the pass). However, this requires that I link in LLVMCore to provide said bindings to my library. From what I've read, there is a known issue with linking LLVMCore into a module that is fed to opt. Namely, you get the following assertion:

We have discovered the issue with our code. It seems that when the first call to doInitilization utilized one of the specified analyses, that analysis had never been properly added as required. Using gdb, we can see that addRequired is getting called (seen below) but when doInitalization is finally called, it still responds as if the analysis was not required. Is this a bug or is

Hey all, We have been working on a pass that uses another pass to count loads and stores prior to performing its own instrumentation. The second pass adds the first as required via the usual getAnalysisUsage function. On one machine, it has been tested and proven to function correctly. On another machine, whenever the second pass is run, it consistently fails the assertion: opt:

Hi Am Dienstag, 30. Mai 2006 19:21 schrieb Chris Lattner: > On Tue, 30 May 2006, Silken Tiger wrote: > > Everthing now compiles fine, but when running llc with invoking my own > > backend derived from the cbackend i get the following error: > > namespace llvm { > > class MParSchedule : public BasicBlockPass { > > public: > > > >

Hi > Right it does. However, does something *else* require MParSchedule? If > so, what? Ok, i am writing on a different backend based on the cbackend. The test usage of this pass looks like this: void getAnalysisUsage(AnalysisUsage &AU) const { AU.addRequired<LoopInfo>(); AU.addRequired<MParSchedule>(); AU.setPreservesAll(); } and then in the

Hello, I am trying to modify my LLVM-based compiler to perform an initial, no-optimization compilation synchronously on startup and then perform an asynchronous, optimized recompilation in the background, and I am getting in one of the optimization passes. - I am using the official release of LLVM 2.8 - I have compiled LLVM with threading enabled; I am running llvm::llvm_start_multithreaded() on

Hi everyone, I downloaded the latest version of LLVM PTX backend from http://www.prog.uni-saarland.de/projects/anysl and made the required changes to all the files mentioned in the README. But I get the following error when I compile it. llvm[3]: Compiling PTXBackend.cpp for Release build In file included from PTXBackend.h:70:0, from PTXBackend.cpp:36: PTXPasses.h: In constructor

On Wed, 31 May 2006, Silken Tiger wrote: >> that requires a BasicBlockPass, it will fail the same was as when a >> ModulePass requires a FunctionPass. > void MParSchedule::getAnalysisUsage(AnalysisUsage &AU) const { > AU.setPreservesAll(); > } > > MParSchedule requires nothing and changes nothing. So hopefully the above code > represents this fact? Right

Hi, Probably some of the leaks Valgrind reports are spurious, but the numbers seem to be significant enough to demand some attention: ==10132== LEAK SUMMARY: ==10132== definitely lost: 15,624 bytes in 558 blocks. ==10132== indirectly lost: 44,548 bytes in 1,591 blocks. ==10132== possibly lost: 37,576 bytes in 98 blocks. ==10132== still reachable: 1,336,876 bytes in 1,364 blocks.

Ether, Sounds great! Please keep in mind that, eventually, we'll also want to configure those options from TLI (or something similar). The patch looks good to me. -Hal On Wed, 4 Apr 2012 23:54:18 +0800 Hongbin Zheng <etherzhhb at gmail.com> wrote: > Hi Hal, > > I add a function named "vectorizeBasicBlock" which allow users to > perform basic block

Hi Hal, I add a function named "vectorizeBasicBlock" which allow users to perform basic block vectoirzation inside their pass. But i am not sure whether i missed something as no one use the function right now (But it will be used by Polly sometimes later[1]). In addition, we (tobi and me) also want to make the vectorizer being configured command line flags. To achieve this, we are

ether zhhb wrote: > hi John, > > sorry for reply so late. > > On Tue, Apr 13, 2010 at 10:38 PM, John Criswell <criswell at uiuc.edu > <mailto:criswell at uiuc.edu>> wrote: > > Devang Patel wrote: > > On Mon, Apr 12, 2010 at 6:41 PM, ether zhhb > <etherzhhb at gmail.com <mailto:etherzhhb at gmail.com>> wrote: > >

I'm not sure about the old pass manager, but I think the new pass manager solves this issue. See llvm::updateCGAndAnalysisManagerForFunctionPass where it updates the call graph to be in sync with edges deleted by function passes. So I suspect the right fix is to use the new pass manager. -- Sanjoy On Mon, May 7, 2018 at 7:32 AM, Björn Pettersson A via llvm-dev <llvm-dev at

Hi Björn, 1) The pass pipeline syntax is documented here: https://github.com/llvm-project/llvm/blob/master/include/llvm/Passes/PassBuilder.h#L378 -die is not implemented, since the new pass manager does not support BasicBlock passes. But you can use dce instead: "-passes=dce" 2) I don't have a qualified answer here, but if I recall correctly, the trouble to correctly update the

hi John, On Mon, Apr 19, 2010 at 10:27 PM, John Criswell <criswell at uiuc.edu> wrote: > ether zhhb wrote: > >> hi John, >> >> sorry for reply so late. >> >> On Tue, Apr 13, 2010 at 10:38 PM, John Criswell <criswell at uiuc.edu<mailto: >> criswell at uiuc.edu>> wrote: >> >> Devang Patel wrote: >> >> On

If I run: opt -globals-aa -die -inline -debug-pass=Details foo.ll -S then I will get this pass structure: Target Library Information Target Transform Information Target Pass Configuration Assumption Cache Tracker Profile summary info ModulePass Manager CallGraph Construction Globals Alias Analysis FunctionPass Manager BasicBlockPass Manager Dead Instruction