similar to: [LLVMdev] Boehm GC + static variables?

I've implemented this by adding calls to GC_add_roots(<first global in module>,<last global in module>+1) to the llvm.global_ctors before any other static initialization code for the module. This should be safe assuming that: - global variables are laid out in memory in the order they appear in their module (and ideally contiguously without being interleaved with any other values)

You should look at http://llvm.org/viewvc/llvm-project/llvm/trunk/include/llvm/ExecutionEngine/JITMemoryManager.h?view=markup and see if inheriting from that and overriding allocateGlobal() will do what you want. I'm a little surprised the boehm gc doesn't already see the globals, since there's a reference to their memory from the JMM, but maybe it doesn't scan mmap regions by

Hi, I want to use LLVM as replacement code generator for an existing self hosting compiler. I hope to replace the existing BURS code generator with LLVM in order to take advantage of LLVM's JIT, optimizations and wider range of targets. I'm planning on ditching my existing IR completely and using my language's native call mechanism to call the LLVM C bindings. I've got a couple

I'm having a curious issue with LLD/x86_64 linux/elf (Ubuntu 14.04); Where the l_addr of the link_map is invalid when linked with lld, but is fine with gnu ld. I'm using the libgc (boehm) code which when initializing reads the DYNAMIC/DEBUG link_map data, and crashes because the l_addr field has value out of readable memory. The strange this is that it happens only on some linux

I'm hoping to have a stab at getting a tiny language implementation up and running soon. I assume I can just pull in Boehm's GC to get a very rudimentary collector up and running. However, can this work using JIT compilation from OCaml or will the new GC try to traverse the OCaml program's own stack? -- Dr Jon D Harrop, Flying Frog Consultancy Ltd.

The virtual machine I'm implementing allocates objects on the heap using Boehm's garbage collector. I currently have an interpreter working for my language, and I'm working on a JIT that will compile some of the functions that would otherwise get interpreted. What I'm wondering is whether or not I need to perform special adjustments for the GC to work properly. My understanding is

As the Boehm GC is an available option, do I take that to mean it's beneficial for some users but not most? -- Daniel

Hello, I've been able to get J3 working with the single and multimap garbage collectors but have hit several snags with Boehm and MMTk. For MMTk I get here: Buildfile: /cs/student/kyleklein/vmkit/mmtk/java/build.xml main: [mkdir] Created dir: /cs/student/kyleklein/vmkit/mmtk/java/classes [javac] Compiling 373 source files to /cs/student/kyleklein/vmkit/mmtk/java/classes

For anyone who encounters this issue in the future, my issue was the configuration of llvm-gcc. Configure with --with-llvmgccdir=YOUR/PATH doesn't work, instead you are supposed to use --with-llvmgcc=PATH/TO/llvm-gcc and --with-llvmgxx=/PATH/TO/llvm-g++ That will at least get you past my error. Best, Kyle Quoting nicolas geoffray <nicolas.geoffray at gmail.com>: > I am

best described here: https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=64689 Confirmed also present with the rpm build at http://rsync.samba.org/ftp/rsync/binaries/redhat/rsync-2.4.6-1.i386.rpm Please cc: me on replies (I'm not on the list, yet - my procmailrc's in a major state of flux as I'm switching machines) and/or add comments to the above bugzilla entry James -- James

Hello, It would be nice if we could get a full list of packages which have a wrong sha1 checksum. I don't have the bandwidth for a full mirror of all of Dag's rpms and i also don't have shell access to such a mirror. So a small request for a mirror admin: The following checks the files repodata/*.xml.gz against the sha1 sums in repomd.xml and checks the RPMS/*.rpm files against

I saw some unexpected behavior in the interaction of hosts allow and hosts deny on Samba 3.0.28. I built Samba 3.2.4 just to be sure it wasn't something that had been fixed. I saw the same behavior. I'm not sure if it is a bug or a failure on my part to understand the documentation or misleading documentation. If I have a share defined as [export] comment = exported

Hi, I'd like to fix PR19590, which is about llvm.global_ctors containing functions that end up being empty after optimization (which causes the linker to add useless init_array entries to the output binary). globalopt removes empty functions from llvm.global_ctors, but by the time the function becomes empty globalopt has already run and it doesn't run again. I'm wondering what the

Can anyone explain how llc translates "appending linkage" global variables like llvm.global_ctors into assembly? In the case I am examining, the global_ctor variables are in multiple bitcode object modules produced by the llvm compiler as arrays of pointers. As documented, the arrays seem to be combined when the different bitcode object modules are linked into one, so now there is one

Hello, A strange problem appears when upgrading from release_34 to testing. Some transformations to bitcode files cause registered global_ctors to not be called. Here's an example (I've also attached the complete example and pasted it below): This works: clang -fsanitize=address -flto -c -o sum.o sum.c clang -fsanitize=address -o sum sum.o This doesn't work: clang

I'm a bit puzzled. I am able to map an account without any problem on Samba 2.2.8a (security=domain). However, access fails with Samba 3.0.14a when everything else is the same (same configuration files). Any advice as to the cause of the problems (and its solution) would be appreciated. >From 2.2.8a logs [2005/08/24 14:59:51, 3, pid=7767] smbd/reply.c:(880) Domain=[americase]

Hi all, I tried to modify Kaleidoscope Tutorial (toy.cpp from llvm/examples/Kaleidoscope/Chapter7, LLVM 3.3 release branch) in order to use MCJIT instead of JIT. I get segmentation fault when running toy.cpp with fibonacci example from the tutorial. My modified toy.cpp is in attachment and still works with JIT (when #define USE_MCJIT line is commented out). I read discussions regarding

I found the following problem when I try to debug "target does not support mc emission" in linux (the same code executes OK in windows): Below is a snippet extracted from this method, if (UseMCJIT && ExecutionEngine::MCJITCtor) { ExecutionEngine *EE = ExecutionEngine::MCJITCtor(M, ErrorStr, JMM, OptLevel,

On Mon, Sep 2, 2013 at 5:18 AM, Duncan Sands <baldrick at free.fr> wrote: > Hi Reid, > > On 26/08/13 23:43, Reid Kleckner wrote: > >> To implement http://llvm.org/PR16959 >> <http://llvm.org/bugs/show_**bug.cgi?id=16959<http://llvm.org/bugs/show_bug.cgi?id=16959>>, >> I need to add a new field to >> >> global_ctors. >> >>

Am 04.06.2013 16:05, schrieb David Tweed: > | I am curious about JMM->invalidInstructionCache(), which I found in > | lli.cpp implementation. lli.cpp contains also call finalizeObject(), I > | just overlooked it. lli.cpp calls finalizeObject(), which calls > | applyPermissions, which in turn calls invalidateInstructionCache. So why > | lli.cpp does call