similar to: [LLVMdev] External names for LTO in gold plugin

Hi Hal, Thanks for your hint! $ /opt/llvm-svn/bin/opt -S -internalize -internalize-public-api-list=main -globaldce hello3.ll -o hello3.dce.ll    it works :) But I argue that `main` Function should be inserted into ExternalNames by default: Index: lib/Transforms/IPO/Internalize.cpp =================================================================== --- lib/Transforms/IPO/Internalize.cpp 

Hi LLVM developers, $ cat hello.c #include <stdio.h> void foo() { } int main(int argc, char *argv[]) {   for (int i = 0; i < 10; i++) {     printf("%d\n", i);   }   return 0; } $ /opt/llvm-svn/bin/clang --version Fedora clang version 6.0.0 (trunk 316308) (based on LLVM 6.0.0svn) Target: x86_64-redhat-linux Thread model: posix InstalledDir: /opt/llvm-svn/bin $

Hello, I''m using kernel 3.1.0 and I have both / and /home as btrfs. I used suspend to ram quite often and never had a problem, but yesterday I''ve suspended to get into a plane and when I resumed my /home was all about input/output errors. Reboot did not help either. My root (/) did not suffer any problems. Today I''ve upgraded btrfs-progs to latest GIT and tried scrub

I'm working on a whole program optimizer that uses LLVM as a library, and one of the things I want to do is eliminate dead global functions and variables even when they are not local to a module. (This understandably doesn't happen by default because the optimizer has to assume it could be compiling a library rather than a program.) I've actually written a function to do this, but

Dear Ben, Just a few comments on DSA: 1) I'll try out your example C++ code below and see if I can get the same results that you do. However, I'm at a conference right now (Usenix Security), so I don't know exactly when I'll get to it. 2) DSA can get pessimistic results when dealing with external code (as Andrew described). It is designed for whole program analysis, meaning

In an earlier message (http://lists.cs.uiuc.edu/pipermail/llvmdev/2011-August/042298.html), Andrew Lenharth suggested that EQTDDataStructures (from the poolalloc project) may only try to resolve indirect function calls. However, I am now finding that the generated DSCallGraph over-approximates the callees in a very simple indirect call. Some over-approximation is unavoidable, but this case

Hi, I was interested in telling the compiler not to inline a given function, and discovered that the __attribute__((noinline)) implements this using a global variable @llvm.noinline. It did not work for me initially. I noticed that @ llvm.noinline was being internalized by -internalize, dead-code-eliminated by -constmerge, causing this information not to reach the -inline pass. I am not sure if

Hi Xiantao, I general I think the code in this patch is fine. I have a couple of nit-picking comments: > + if (target_mask&0x1) { The formatting here isn't quite what most of the kernel does. It would be better if you added spaces so it's a little easier to read, ie: if (target_mask & 0x1) { > + p = &__per_cpu_idtrs[cpu][0][0]; > + for (i = IA64_TR_ALLOC_BASE;

Hi Xiantao, I general I think the code in this patch is fine. I have a couple of nit-picking comments: > + if (target_mask&0x1) { The formatting here isn't quite what most of the kernel does. It would be better if you added spaces so it's a little easier to read, ie: if (target_mask & 0x1) { > + p = &__per_cpu_idtrs[cpu][0][0]; > + for (i = IA64_TR_ALLOC_BASE;

I got your point! Thanks a lot, Rafael! Best, Tianyin On Tue, Jun 26, 2012 at 4:12 PM, Rafael Espíndola < rafael.espindola at gmail.com> wrote: > > Could you explain a bit more (sorry I'm not from a compiler background)? > So, > > what's the difference in the analysis perspective? What does "reading ELF > > files and archives" and "creating a

Hello everyone, 9.0.0-rc2 was tagged yesterday from the release_90 branch at r368683. In the Git monorepo it's available as the llvmorg-9.0.0-rc2 tag. Source code and docs are available at https://prereleases.llvm.org/9.0.0/#rc2 Binaries will be added as they become available. The tag went in roughly one week behind schedule (see "Upcoming Releases" at https://llvm.org), but

Hi everyone, 9.0.0-rc1 was just tagged from the release_90 branch at r367217 (tagged as llvmorg-9.0.0-rc1 in the Git monorepo). Source code and docs are available at https://prereleases.llvm.org/9.0.0/#rc1 Binaries will be added as they become available. Please file bug reports for any issues you find as blockers of https://llvm.org/PR42474 Release testers: please start your engines, run the

Hello everyone, 9.0.0-rc5 was just tagged from the release_90 branch at r371837. In the Git monorepo, it's tagged as llvmorg-9.0.0-rc5. Source code and docs are available at https://prereleases.llvm.org/9.0.0/#rc5 Binaries will be added as they become available. There is only a single change from rc4 to rc5. Once more, the hope is that this will be the last release candidate and that we can

Hi, I've just tagged LLVM 9.0.1-rc3. Testers can begin testing and uploading binaries. This will be the last release candidate unless there is a major problem. I'm planning to tag the final release on Dec 19. -Tom

Hello again, 9.0.0-rc4 was just tagged from the release_90 branch at r371490. In the Git monorepo, it's tagged as llvmorg-9.0.0-rc4. Source code and docs are available at https://prereleases.llvm.org/9.0.0/#rc4 Binaries will be added as they become available. There are not a lot of changes from rc3 to rc4, and there are again no open release blockers, so I'm hoping this will be the last

Hi, I have a vector like: r <- runif(100) Now I would like to split r into 10 pieces (each with 10 elements) ? but the 'pieces' should be roughly similar with regard to mean and sd. what is an efficient way to do this in R? thanks!

I like to propose a new loop multi versioning optimization for LICM. For now I kept this for LICM only, but it can be used in multiple places. The main motivation is to allow optimizations stuck because of memory alias dependencies. Most of the time when alias analysis is unsure about memory access and it says may-alias. This un surety from alias analysis restrict some of the memory based

I read the tutorial doc and some info of SSA, finally understand it. Thanks for your help. And the segmentation error of loading arguementation occurs, (gdb error info) Program received signal SIGSEGV, Segmentation fault. 0x0855bb68 in llvm::LoadInst::LoadInst(llvm::Value*, char const*, bool, llvm::Instruction*) () code is like follows: //#include necessary header files int main(){

Hi, I've tagged the LLVM 9.0.1-rc1 release. Testers can begin testing and upload binaries. I've also updated the test-release.sh script to pull from GitHub instead of SVN, if you run into any issues with the new script, let me know. -Tom

Hi, > I got your point! Thanks a lot, Rafael! one big difference is that llvm-link doesn't do any optimization, while the gold linker does. You can get something similar to gold by using llvm-link, and then running "opt -std-link-opts" on the linked bitcode file. However all symbols will be internalized by default (you can adjust this on the command line) while the gold linker