similar to: [LLVMdev] PassManager and dependencies

Chris Lattner wrote: >>I don't grok this error message. Of course, -opt-a and -opt-b both work >>fine in isolation. >> >> > >You're right, this error message is terrible. As it turns out, all of >your passes invalidate all of the other passes, so C doesn't get A (which >is invalidated by B). The problem turns out to be a really trivial bug:

On Mon, 9 Jan 2006, Saem Ghani wrote: > The patch below basically hammers out some ideas as to where I'd like > to take the passmanager in LLVM. I've tried thinking things through, > but I'm still a n00b, so some criticism would be more than welcome. =) > > Starting from line 191 down. If you're wondering why I created a > patch, well that's because I found

On Tue, 21 Mar 2006, Michael McCracken wrote: > opt: /home/mmccrack/lens/obj-llvm-darcslocal/../llvm-darcslocal/llvm/lib/VMCore/PassManagerT.h:426: > void llvm::PassManagerT<UnitType>::markPassUsed(const llvm::PassInfo*, > llvm::Pass*) [with UnitType = llvm::Module]: Assertion > `getAnalysisOrNullUp(P) && > dynamic_cast<ImmutablePass*>(getAnalysisOrNullUp(P))

A On 3/21/06, Michael McCracken <michael.mccracken at gmail.com> wrote: > On 3/21/06, Chris Lattner <sabre at nondot.org> wrote: > > On Mon, 20 Mar 2006, Michael McCracken wrote: > > > > > Hi, I'm trying to access an analysis pass from the Inliner pass, and > > > I'm having a lot of trouble getting that to work - I can verify that > >

Sorry in advance for taking so long to respond to this issue... > I can't seem to figure out how to tell the PassManager that one Pass > requires the results of two other Passes, in such a way that it will not > crash itself. Attached file is the simplest possible example of Passes > I don't grok this error message. Of course, -opt-a and -opt-b both work > fine in

On 1/10/06, Chris Lattner <sabre at nondot.org> wrote: > Interesting approach. :) Thanks. > Comments below, with ***'s before the notes: > +class LoopPass : public Pass {}; // Temporary. > > *** I wouldn't worry about loop passes yet. Sure. > +class PassUnit { > + Pass *pass; > + > + enum Traversal { > + LINEAR, // Standard top down

The patch below basically hammers out some ideas as to where I'd like to take the passmanager in LLVM. I've tried thinking things through, but I'm still a n00b, so some criticism would be more than welcome. =) Starting from line 191 down. If you're wondering why I created a patch, well that's because I found thinking in passmanagert.h the most productive. -- Regards.

I cannot figure out a particular PassManager error for what seem to be legal dependencies. Here is the situation. We have 5 passes, RegisterAllocator, FunctionLiveVarInfo, CoalesceCopies, DominanceForest, and UnionSSAVars, with dependencies as follows: class RegisterAllocator : public FunctionPass { . . . virtual void getAnalysisUsage(AnalysisUsage &AU) const {

Issue: GraphWriter includes <ostream>, which my gcc2 apparently thinks is <ostream.h>. Fix: Make a new <Support/ostream> that handles this discrepancy, ala <Support/hash_set>. -- Casey Carter Casey at Carter.net ccarter at uiuc.edu AIM: cartec69 -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: patch URL:

So, I had to make a few changes to the llvm sources to allow compilation on x86 redhat 7.3 (gcc-2.96, glibc 2.2.4). Is there any general interest in maintaining a port? I will happily submit patches. -- Casey Carter Casey at Carter.net ccarter at uiuc.edu AIM: cartec69

ISSUE: In Interpreter::getCurrentExecutablePath(), dladdr() is a Solarisism. Luckily, getCurrentExecutablePath isn't being currently used anywhere in lli. ACTION: Wrap the method contents with #ifdef __sun__ ... #else return ""; #endif. If this functionality is actually desired, it would be more portable to hack up main() to join getcwd() with basename(argv[0]) to find the

Chris Lattner wrote: >>>Interesting. INT64_MAX is supposed to be provided by >>>include/Support/DataTypes.h. Do you know of a reliable preprocessor >>>symbol that can be used to determine whether we're on a linux box, or >>> >>> > > > >>Well, there is always __linux__, but that doesn't necessarily imply that >>we

ISSUE: This code: %bob = type { int } int %alias() { %pbob1 = alloca %bob %pbob2 = getelementptr %bob* %pbob1 ;pbob2 aliases pbob1 %pbobel = getelementptr %bob* %pbob2, long 0, ubyte 0 %rval = load int* %pbobel ret int %rval } Crashes when run through opt -instcombine. InstCombiner visits instructions in reverse declaration order, but

ISSUE: INT64_MAX undefined in InstrSelectionSupport.cpp and InstructionCombining.cpp. I'm not completely sure where INT64_MAX comes from on Solaris, but C99 says that INT64_MAX is defined in stdint.h, but, for C++, only if __STDC_LIMIT_MACROS is #defined. Solaris (at least in CSIL) unfortunately does not have stdint.h, but it does have the old inttypes.h - and so does Linux. ACTION: In

Hi, Reid Spencer wrote: >> That's interesting! So, one only needs to add a 2-arg function called >> RunOptimizations to the module (can't check it right now)? >> > > That is correct. That function and only that function will be called. > What happens in that function is up to you :) > So, I tried this the last two days, but to no avail. I first

llvm-gcc is producing segmentation fault when running with -emit-llvm. Running cc1 in valgind produces the warning: ==12768== at 0x8D042C: llvm::llvm_ostream& llvm::llvm_ostream::operator<< <llvm::Module>(llvm::Module const&) (Streams.h:41) ==12768== by 0x8D0464: llvm::PrintModulePass::runOnModule(llvm::Module&) (PrintModulePass.h:41) ==12768== by 0xED6EB9:

Okay, I've been studying passmanagert and company for quite a while. Here are my conclusions. It's actually not nearly as bad as I thought. I'll outline the problem as I see it. We want to be able to manage the way passes traverse things, in particular a module's worth of functions. As was in your example, we want inliner and mem2reg calls interleaved, as we traverse functions

Chris Lattner wrote: >>ISSUE: INT64_MAX undefined in InstrSelectionSupport.cpp and >>InstructionCombining.cpp. I'm not completely sure where INT64_MAX comes >>from on Solaris, but C99 says that INT64_MAX is defined in stdint.h, >>but, for C++, only if __STDC_LIMIT_MACROS is #defined. Solaris (at >>least in CSIL) unfortunately does not have stdint.h, but it does

Bram, I looked over your patch and it looks good. I applied a patch based on yours. The llvm-ld tool now uses the PluginLoader just like the opt tool. It will also run some cleanup passes after the loaded plugins run to ensure cruft is removed. See this patch for details: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20060814/036882.html Thanks for the patch! And, yes, you are on

On Wed, 12 Apr 2006 22:48:16 -0500 Patrick Meredith <pmeredit at uiuc.edu> wrote: > > On Apr 12, 2006, at 10:23 PM, Simon Burton wrote: > > > > > Is it possible to get llvm to generate native machine code > > without using gcc and friends ? Do I use lli ? > > llc. llc --help lists all the options. it compiles llvm bytecode > files. It seems this