similar to: [LLVMdev] Assembler used in LLVM

Hi all, I have successfully built LLVM itself, as all the tools (llc et al) are working fine. I built the latest release version (2.7). I have been trying to build the llvm-gcc frontend on a sparc machine. I encountered the following issues : 1) During configuration, the default assembler (installed at /usr/ccs/bin) is chosen to be used, but I want to use the GNU assembler (installed at

On Sep 26, 2010, at 3:23 PMPDT, Raghu Prabhakar wrote: > 1) During configuration, the default assembler (installed at /usr/ccs/bin) is chosen to be used, but I want to use the GNU assembler (installed at /vm/GNU/bin in my machine). In spite of re-arranging paths on the PATH variable, and even after removing /usr/ccs/bin from PATH, configure still chooses the default assembler itself. Is there

To pass the MD5 checksum to the assembler, I added a new optional clause to the .file directive: md5 "checksum" where checksum is the 16-byte checksum in hex. It's quoted because the assembler doesn't have a way to parse a 16-byte integer. Also this is the same syntax Reid invented for the CodeView equivalent. To convey the root source filename, I allow the file number on the

Have you added the $LLVM_INSTALL_DIR/bin to your PATH environment variable? Regarding changing a particular backend, check the documentation in http://llvm.org/docs/WritingAnLLVMBackend.html http://llvm.org/docs/CodeGenerator.html A few LLVMdev meetings have good talks to give you a good enough idea. http://llvm.org/devmtg/ Cheers, Raghu. ----- Original Message ----- From: "Teklu

TL;DR: If I'm trying to define new assembler directive syntax to support DWARF v5, it seems like a good idea for all the various assemblers out there in the world to support the same syntax. How would I go about negotiating that syntax with other assembler providers? Is GNU as the only really relevant one? Long version: DWARF v5 introduces a couple of new features in the .debug_line section

Hi all, I have started using LLVM recently, and I have a basic question. There are two different ways in which an executable can be generated from a .c file - by just using llvm-gcc, and using a combination of llvm-gcc (to emit bitcode .bc), llc (to emit assembly code), as (to assemble the code) and collect2 (linker). For example, consider compiling test.c : 1) llvm-gcc -O3 test.c -o

Hi I am new to llvm and Linux (I am using ubuntu 10) My aim is to add backend to llvm. I install llvm using $ sudo apt-get install llvm and I am able to understand what I read about Writing an LLVM Compiler Backend. I have seen the assembly codes of the targets on llvm using llc. What I need now is to change some of the TableGen files of one of the target like SparcInstrInfo.td.  So I

Hello all, I am trying to modify the Sparc backend to do something for my team's project, to do the following. Whenever the backend encounters a call to one particular type of function names (functions like p0(), p1()..etc), I want the backend to generate a "sethi %g0, <number>", and NOT a "call p0, call p1..." instruction. However, the backend should work as usual

On 1/20/11 12:27 PM, Venkatraman Govindaraju wrote: Just out of curiosity, have either of you considered writing an LLVM transform that simply replaces these call instructions with inline assembly code that does what you want? If that works, it seems much simpler than modifying/enhancing the code generator. -- John T. > I have similar requirements for my project. This is what I do. >

We are already generating object directly. It is on my roadmap to implement the standalone assembler, but I was hoping for some suggestions ;-) My target is Mips and I have started looking at what ARM does. If there is a better model or how-to guide I'd like to know. Thanks, Jack ________________________________ From: Owen Anderson [resistor at mac.com] Sent: Wednesday, January 11, 2012

I have had references to llvm-mc in terms of a tester for MC level code changes and have just found and read the blog http://blog.llvm.org/2010/04/intro-to-llvm-mc-project.html. Am I correct that lib/Target/<target>/AsmParser/*.cpp needs to be completed to make a standalone assembler for a specific target and until that is done, llvm-mc can't be used to generate output object that is

On Jan 11, 2012, at 6:05 PM, Carter, Jack wrote: > I have had references to llvm-mc in terms of a tester for MC level code changes and have just found and read the blog http://blog.llvm.org/2010/04/intro-to-llvm-mc-project.html. > > Am I correct that lib/Target/<target>/AsmParser/*.cpp needs to be completed to make a standalone assembler for a specific target and until that is

ARM is the best model currently, though x86 does some interesting things as well and is worth looking at. If there's anything in particular that's causing head-scratching, just ask. There's not much available in terms of how-to documentation, unfortunately. That's definitely on my TODO list to write, but time constraints are rough right now. -Jim On Jan 11, 2012, at 8:09 PM,

Hi Raghu, > Thanks Duncan. Just to make sure I understand correctly, opt has more global information, due to which optimizations can have greater effect than in llvm-gcc, where optimizations are applied on a per-function basis. Is this correct ? if you look carefully at what "opt -O3" runs (using -debug-pass=Arguments) you will see that it first runs a small list of passes (the

I have similar requirements for my project. This is what I do. 1. Add a new intrinsic function to LLVM that corresponds to "functions with particular type of function names". The steps to create intrinsic function is documented in http://llvm.org/docs/ExtendingLLVM.html#intrinsic. 2. Create a lowering pass that lowers "functions with particular type" ( eg. p0, p1) to

Jim, What I am doing is modeling from ARM and X86 just enough to get "hello world" to produce correct ELF output. >From many years of experience I know this will leave me with little understanding of what is going on and I will need to have this understanding to really add long term value. Maybe my questions will help frame the how-to-and-why documentation. I think my first patch

>> To pass the MD5 checksum to the assembler, I added a new optional >> clause to the .file directive: >> md5 "checksum" >> where checksum is the 16-byte checksum in hex.  It's quoted because >> the assembler doesn't have a way to parse a 16-byte integer. > > I'd guess, long-term, that's probably not a great motivation for > choosing

For my case, I can't replace these call instructions with inline assembly code because I need to encode the registers into the "number". For instance, if the call instruction is %result = call i32 @foo(i32 %a) and the result is assigned to register %l0 and the variable "a" to register %l1, then I encode all foo, %l0 and %l1 and generate a sethi instruction. thanks,

Thanks Duncan. Just to make sure I understand correctly, opt has more global information, due to which optimizations can have greater effect than in llvm-gcc, where optimizations are applied on a per-function basis. Is this correct ? ----- Original Message ----- From: "Duncan Sands" <baldrick at free.fr> To: llvmdev at cs.uiuc.edu Sent: Tuesday, December 13, 2011 2:46:23 AM

Hi Tim, Thanks for the patch. I had a few questions though - just to make sure that I get the problem correctly. Currently, here is your flow, as I understand it : x86 llvm-gcc binary llc -march=sparc gcc cross/sparc source ---------------------> .bc ------------------> .s -----------------> sparc executable. Is this correct? Or did you manage to build the