similar to: [LLVMdev] What's the difference between using llvm-link and using llvm gold plugin?

On 26 June 2012 16:46, Tianyin Xu <tixu at cs.ucsd.edu> wrote: > Hi, > > I'm sorry if the question doesn't fit this mailing list. It seems that LLVM > does not have a user support mailing list so I post it here. > > We're using the LLVM tool chain as our static analysis infrastructure. So, > basically we use LLVM to generate an aggregated .bc file and analyze

Hi, Rafael, Thanks a lot for the explanation! 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 list of symbols passed to llvm's internalize pass" do? I checked a little bit the corresponding .ll file generated according to the

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

> 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 list of symbols passed to llvm's > internalize pass" do? > > I checked a little bit the corresponding .ll file generated according to the > .bc file. It

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

The module is shipped with the source code, so it's from the developers (not me). The path in the source tree is: llvm-3.4.2.src/lib/Transforms/Hello After compiled LLVM *#./configure --prefix=XXX --with-binutils-include=/path-to-binutils-2.24/include --enable-debug-symbols --enable-debug-runtime --enable-assertions --disable-optimized; make; make install* LLVMHello.so is generated at

I just follow the tutorial: http://llvm.org/docs/WritingAnLLVMPass.html Actually it's not only my module, I used the LLVMHello.so (a sample module in the source tree) and get the crash (I didn't do anything :-P). ~t On Tue, Aug 12, 2014 at 12:57 AM, Eric Christopher <echristo at gmail.com> wrote: > Weird, it definitely shouldn't be crashing. How did you create your

no... I guess verifier runs by default and it points to the doFinailization() function :-( ~t On Tue, Aug 12, 2014 at 12:53 AM, Eric Christopher <echristo at gmail.com> wrote: > No idea, I was suggesting to just run the verifier to see if it would > pinpoint what's broken about the module. > > -eric > > On Tue, Aug 12, 2014 at 12:52 AM, Tianyin Xu <tixu at

Oh, you mean opt -disable-verify? Yes, it does eliminate the crashing... Do you know what's the problem, Eric? Thanks a lot! ~t On Tue, Aug 12, 2014 at 12:42 AM, Eric Christopher <echristo at gmail.com> wrote: > Tried opt -verify on your module? > > -eric > > On Tue, Aug 12, 2014 at 12:28 AM, Tianyin Xu <tixu at cs.ucsd.edu> wrote: > > Hi all, > >

Hi all, I find all my passes are all broken with LLVM 3.4. Then I tried out the LLVMHello.so specified in the LLVM doc, http://llvm.org/docs/WritingAnLLVMPass.html and it also crashes. It seems all the functions in the pass do work, but LLVM crashes in the doFinalization() step. Does anyone know this problem? Thanks! Tianyin

Hi, all, I'm using LLVM-3.0. I want to use gold plugin, but I got the following error message: /home/tianyin/binutils-2.22/bin/ld: error: /home/tianyin/llvm-3.0/lib/LLVMgold.so: could not load plugin library: /home/tianyin/llvm-3.0/lib/LLVMgold.so: undefined symbol: lto_codegen_set_cpu Could anyone tell me what's the problem? Thanks a lot! Tianyin

Hi, In my work applying -flto to Chrome, I need to set some names to be skipped by the InternalizePass; otherwise, the linking stage fails (i.e., when building the chrome binary). In the past, I had a tiny patch that I hadn't submitted to LLVM: it was something like: Index: Internalize.cpp =================================================================== --- Internalize.cpp (revision

To compile an auto-tooled project to bitcode, I used to use plugin-opt=save-temps switch as explained in: http://gbalats.github.io/2015/12/10/compiling-autotooled-projects-to-LLVM-bitcode.html. This switch works with LLVM 3.8.0. However, it seems to have no effect with LLVM 5.0.0. I can see intermediate bitcode files on disk, but not the final one. Is the option removed altogether? Has the switch

On Wed, Dec 12, 2012 at 06:25:53PM -0800, Tianyin Xu wrote: > Hi, all, > > I'm using LLVM-3.0. I want to use gold plugin, but I got the following > error message: Why not use TOT, or latest release? :) Regards, chenwj -- Wei-Ren Chen (陳韋任) Computer Systems Lab, Institute of Information Science, Academia Sinica, Taiwan (R.O.C.) Tel:886-2-2788-3799 #1667 Homepage:

Ah! Okay. A quick test of llvm-lto gets an error: llvm-lto: error adding file 'main.bc': Linking COMDATs named '??_7_Iostream_error_category at std@@6B@': invalid selection kinds! Would it be the case that llvm-lto is also deprecated in favor of the gold plugin? On Tue, Feb 10, 2015 at 4:47 PM, Teresa Johnson <tejohnson at google.com> wrote: > LTOCodeGenerator is used

Following on from my last question, based on such information as I've been able to find on Google, I get the impression llvm-link is effectively deprecated / not actively maintained, because the focus of whole-program optimization is now on the gold plugin. Is that the case, or am I missing something? Running searches over the source tree, it seems the Linker class that actually implements

Hello, Is it still possible to produce the LLVM bitcode after linking using the LLVM gold plugin? I get the warning "ld: warning: Ignoring flag also-emit-llvm" and the LLVM bitcode is not dumped. I installed the LLVM gold plugin using the latest binutils from CVS and LLVM 2.6, since I read on the discussion list that this was possible with LLVM 2.6. This is the command I use:

Hi, all. I want to get the single bitcode file after link-time optimizations. After doing some searching, I get commands as follows: *llvm-gcc -O4 -c xxx.c -o xxx.o llvm-gcc xxx.o -use-gold-plugin -Wl,-debug -Wl,-plugin-opt=also-emit-llvm -lm -o xxx* But these command don't give me the wanted bitcode file. The used version of llvm is 2.7. Do you have any idea of this problem or any other

On 13 June 2011 15:37, Cristian Zamfir <cristian.zamfir at epfl.ch> wrote: > Hello, > > Is it still possible to produce the LLVM bitcode after linking using the > LLVM gold plugin? > I get the warning "ld: warning: Ignoring flag also-emit-llvm" and the LLVM > bitcode is not dumped. > > I installed the LLVM gold plugin using the latest binutils from CVS and

Hi,all I have installed the gold-plugin and it can run correctly,but I still have another two questions: 1.Can gold-plugin generate a bitcode file: >From the document and also my experience on using it,"ld -plugin...." will only generate an executable by link several LLVM bitcode file together,but how to make it generate a whole-program bitcode file?(even archieve it manually is