search for: jitdebugregisterer

In several places, for example in JITDebugRegisterer::MakeELF, stream is opened and the error is ignored. When the error is actually returned by open There are two solutions (assuming there are no exceptions). 1. check error string after every raw_fd_stream::raw_fd_stream, fix all the places where the check is missing 2. make raw_fd_stream::raw_fd...

So far, for X86 identical ELF relocation flags show up in include/llvm/Support/ELF.h and lib/Target/X86/X86ELFWriterInfo.h Not to mention, there are two different files named ELF.h ./lib/CodeGen/ELF.h ./include/llvm/Support/ELF.h I'd think the latter should be renamed MCELFSupport.h Thanks! -jason

...to mention, there are two different files named ELF.h I would say that ELF.h should be the canonical one. > ./lib/CodeGen/ELF.h > ./include/llvm/Support/ELF.h > > I'd think the latter should be renamed MCELFSupport.h The first one should probably be removed. It is included from JITDebugRegisterer.cpp ELFCodeEmitter.cpp ELFWriter.cpp We might be able to remove the last two files today. Any missing information in the first can probably be inlined while the MCJIT is implemented. I will take a look at what is using this. > Thanks! > -jason Cheers, Rafael

the reason I built llvm with vc is that I am familiar with vc. At first i also thing that build llvm with MingW will solve this problem. but after I read the source code void JITDebugRegisterer::RegisterFunction(const Function *F, DebugInfo &I) { // TODO: Support non-ELF platforms. if (!TM.getELFWriterInfo()) return; ...... } I thing MinGW is helpless with debug jit on win. getELFWriterInfo returns true only on linux ,and the RegisterFunction gen symbol infomation only on linu...

Since FreeBSD is an ELF target, this should work fine. I've also tested that this works in 32-bit by building gdb and llvm in 32-bit mode and testing this stuff while running on a 64-bit OS. I would try setting a breakpoint in gdb on 'llvm::JITDebugRegisterer::RegisterFunction' to see that it is being called, and that getELFWriterInfo returns something. If that function actually installs the code entry, then it's a gdb problem. Reid On Mon, Jun 21, 2010 at 12:49 PM, Yuri <yuri at rawbw.com> wrote: > On 06/21/2010 11:14, Reid Kleckner...

On 06/21/2010 11:14, Reid Kleckner wrote: > Yes, I have some version of 7.1 installed on my workstation and it works for me. I repeted what you did step by step and it still doesn't work for me, see log below. Two differences: I am on FreeBSD and it's 32 bit. It's svn revision 105825. Plus I have few unrelated patched applied to it which shouldn't matter. Yuri $ lli

so as you said , there is no hop for debug jit on windows? 2011/6/7 ��f�� <chenwj at iis.sinica.edu.tw>: > Hi, Tang > >> but in the source code i find this comment. >> // ELF is a reasonably sane default and the only other X86 targets we >> // support are Darwin and Windows. Just use "not those". > > I guess you found above comment in

> so as you said , there is no hop for debug jit on windows? I have no idea. Maybe someone familiar with LLVM on Windows can answer your question. But I have a question for you, you said you built LLVM with Visual Studio and debugged your program with GDB. I doubt the Visual Studio produces the debugging information recognized by GDB. Why not build LLVM with MinGW or someting like that?

...eid Kleckner wrote: > Since FreeBSD is an ELF target, this should work fine. I've also > tested that this works in 32-bit by building gdb and llvm in 32-bit > mode and testing this stuff while running on a 64-bit OS. > > I would try setting a breakpoint in gdb on > 'llvm::JITDebugRegisterer::RegisterFunction' to see that it is being > called, and that getELFWriterInfo returns something. If that function > actually installs the code entry, then it's a gdb problem. > I am on FreeBSD and gdb-7.1 seems to be broken. It fails to set breakpoints in shared libs. using...