search for: rootnodes

On Oct 30, 2010, at 4:38 AM, Duncan Sands wrote: >> Is there any facility in LLVM to identify recursion in a call graph? ... > use the facilities in SCCIterator.h, or declare your pass to be a > CallGraphSCCPass in which case it will work one strongly connected > component at a time. Converting my ModulePass to a CallGraphSCCPass doesn't seem feasible, so I'll use

Hi you basically need to find a cycles in the call graph. Do do this just search google for a graph algorithm, then make it for your problem. See http://en.wikipedia.org/wiki/Cycle_detection. Jeff Kunkel On Tue, Nov 2, 2010 at 4:27 AM, Duncan Sands <baldrick at free.fr> wrote: > Hi Trevor, > > > Converting my ModulePass to a CallGraphSCCPass doesn't seem feasible, so >

Hi Trevor, > Converting my ModulePass to a CallGraphSCCPass doesn't seem feasible, so I'll > use llvm::scc_iterator. Here's what I have so far: > > bool MyModulePass::isRecursive() { > CallGraphNode* rootNode = getAnalysis<CallGraph>().getRoot(); > for (scc_iterator<CallGraphNode*> SCCI = scc_begin(rootNode), E = > scc_end(rootNode); SCCI != E; ++SCCI)

Also, could you write this in a separate pass, and obtain the results from getAnalysis()? I think others would find it useful to discover if a Function may be called recursively. -Jeff Kunkel On Tue, Nov 2, 2010 at 2:38 PM, Jeff Kunkel <jdkunk3 at gmail.com> wrote: > Hi you basically need to find a cycles in the call graph. Do do this just > search google for a graph algorithm, then

Hi, Is there any facility in LLVM to identify recursion in a call graph? I realize this is undecidable in the general case due to function pointers, but at least the static cases could be identified. I don't even care about whole-program recursion, just looking at a single module would suffice. But I don't see anything like this already in LLVM, so do I simply write some code to

Hi Trevor, > Is there any facility in LLVM to identify recursion in a call graph? I > realize this is undecidable in the general case due to function > pointers, but at least the static cases could be identified. I don't > even care about whole-program recursion, just looking at a single > module would suffice. But I don't see anything like this already in > LLVM, so do

On Nov 2, 2010, at 12:53 PM, Jeff Kunkel wrote: > Also, could you write this in a separate pass, and obtain the > results from getAnalysis()? I think others would find it useful to > discover if a Function may be called recursively. I've modified the code so that it correctly identifies both direct and indirect recursion. I'm now trying to package it up as a patch for the

...an be used later on (Number is specified by numops), ty presumably specifies the type of node that this match can be attempted on. Is my understanding of this correct? The thing I'm still unsure about is roots, what exactly does this do? The comment above the definition specifies that 'RootNodes are the list of possible root nodes of the sub-dags to match' (RootsNodes is assigned to root so they're the same) but I can't make any sense of this. Cheers, Greg Chadwick

Trevor Harmon wrote: > On Nov 2, 2010, at 12:53 PM, Jeff Kunkel wrote: > >> Also, could you write this in a separate pass, and obtain the >> results from getAnalysis()? I think others would find it useful to >> discover if a Function may be called recursively. > > I've modified the code so that it correctly identifies both direct and > indirect recursion.

Hi all, I am trying to traverse a dominator tree and have encountered a weird runtime exception: Here's my simple code: virtual bool runOnFunction(Function &F) { DominatorTree& DT = getAnalysis<DominatorTree>(); * DomTreeNode* rootNode = DT.getRootNode();* return false. } Here's the documentation page:

Thanks for your reply, John. I am using opt to run my pass. Should opt by default link in the library? Would you please give me some idea on fixing it? Thanks! opt -load ../../../Debug+Asserts/lib/LLVMHello.dylib -hello < hello.bc > /dev/null My pass is just: #include "llvm/IR/Function.h" #include "llvm/Pass.h" #include "llvm/Support/raw_ostream.h" #include

On 4/11/13 10:30 PM, Bill He wrote: > Hi all, > > I am trying to traverse a dominator tree and have encountered a weird > runtime exception: > > Here's my simple code: > > virtual bool runOnFunction(Function &F) { > DominatorTree& DT = getAnalysis<DominatorTree>(); > > * DomTreeNode* rootNode = DT.getRootNode();* > return false.

On 4/12/13 9:51 AM, Bill He wrote: > Thanks for your reply, John. > > I am using opt to run my pass. Should opt by default link in the > library? Would you please give me some idea on fixing it? Thanks! Odd. I would think that opt would have that pass linked in statically. I'm not sure why it's not working. Does anyone else have an idea? -- John T. > > opt -load

Hi Weibo, I thought would be more efficient discussing here (I just saw that every info was replicated on both emails you were sending to me). I just run your example here on my working machine, and I didn't have any issues with the namespace. Just tell me one thing: which line (and flags) are you using when configuring and make your build llvm directory? Oh, and one more thing: what is on

Hi, Today I was trying to wath demo of LLVM on llvm.org page, but I have received an error "Internal Server Error". I was asked to notify server administrator which I am doing in this email. I am using Linux Debian Etch and Iceweasel in version 3.0.6 Best Regards, Lukasz

On Jan 17, 2010, at 2:56 AM, Mark Muir wrote: > I've developed a working back-end for a custom architecture, based on LLVM 2.6. I'm now trying to cover more of the unique features of this architecture. > > To make use of one such feature, I'm trying something cunning/crazy with the stack - implementing it in a type of memory that can only be addressed via immediates. >

...ly, this doesn't seem to be the problem. I have the following definition for the frameIndex: def frameIndex : Operand<i32>, ComplexPattern<i32 /*valueType*/, 1 /*numOperands*/, "SelectFrameIndex" /*selectFunction*/, [frameindex] /*rootNodes*/> { let PrintMethod = "printFrameIndexOperand"; let MIOperandInfo = (ops GPR); } And the following selection code: bool MyDAGToDAGISel:: SelectFrameIndex(SDValue Op, SDValue N, SDValue& Address) { if (FrameIndexSDNode* FIN = dyn_cast<FrameIndexSDNode>(N...

I've developed a working back-end for a custom architecture, based on LLVM 2.6. I'm now trying to cover more of the unique features of this architecture. To make use of one such feature, I'm trying something cunning/crazy with the stack - implementing it in a type of memory that can only be addressed via immediates. I've got this mostly working. However, I came across a problem

hello, I have a list of numbers that I want to convert into factors representing ranges of values. For example, if I have c(2.5, 1.6, 3.2) I might want a list of factors like c("<3", "<3", ">3") I think I can do this by writing a function and using apply, as.matrix, &c., but I'm looking for a nicer way. Any help is greatly appreciated. Thanks!

Hello, I have a quite common dovecot configuration with virtual users support: 1. User configuration is stored in MySQL database. 2. Mailboxes are owned by vmail:vmail with permission 0600 or drwx------. Is it possible to change mailbox mode to 0660? Dovecot version is 2.1.7 Part of SQL configuration: user_query = \ SELECT \ CONCAT_WS('/', home, u.uid, d.domain_name, u.user_name)