similar to: 'invalid subroutine type ref' when linking custom metadata

> On 2016-Jul-26, at 02:22, Mehdi Amini <mehdi.amini at apple.com> wrote: > > +CC Duncan``` > >> On Jul 25, 2016, at 8:53 PM, Robin Sommer via llvm-dev <llvm-dev at lists.llvm.org> wrote: >> >> >> With 3.9, llvm-link tells me 'invalid subroutine type ref' when >> linking the two code pieces below, and I don't quite understand

The language reference says: "identified [structure] types are never uniqued". But llvm-link seems to do just that at link-time: > cat x.ll %A = type {} %B = type {} %C = type { %A*, %B* } define void @foo(%C* %c) { ret void } > llvm-link x.ll | llvm-dis ; ModuleID = '<stdin>' source_filename = "llvm-link"

On Wed, Sep 25, 2013 at 17:14 +0000, you wrote: > What API did you use in 3.2? llvm::Linker used to have a method LinkInLibrary(). My understanding is that there's a larger reorganisation of the linking process going on but I haven't found a straight-forward replacement for that method yet. I might be missing something though. Or perhaps I could even just reuse some code from that old

Hi all, is there a way to link a native static library (.a) into code that runs through MCJIT? I was able to do it with LLVM 3.2, but it seems the API has changed and maybe was removed? Thanks, Robin -- Robin Sommer * Phone +1 (510) 722-6541 * robin at icir.org ICSI/LBNL * Fax +1 (510) 666-2956 * www.icir.org/robin

I need to do some analysis, in which I have to selectively convert IFs. I am trying to find a way to call this optimization pass from the JIT code, thus I can decide which branches to convert and which to keep. Apparently I cannot do this using a pass manager. -- View this message in context: http://llvm.1065342.n5.nabble.com/Machine-level-IfConversion-for-ARM-tp69513p69622.html Sent from the

On Wed, Sep 25, 2013 at 19:23 +0000, you wrote: > OK, so in the old implementation you linked a module against the > archive before passing it to MCJIT? Yes, I used a Linker instance to link together a number of my own modules and a static library, and then get a single new LLVM module out of that to hand over to MCJIT. > I can think of a way to solve this inside MCJIT, but it is

OK, so in the old implementation you linked a module against the archive before passing it to MCJIT? I'm not familiar with the llvm::Linker interface, so I hope someone else will jump in an address that. I can think of a way to solve this inside MCJIT, but it is almost certainly not what you'd want. -Andy -----Original Message----- From: Robin Sommer [mailto:robin at icir.org] Sent:

Hi all, the LLVM documentation lists the 'appending' linkage for having the linker combine two arrays. I'm wondering though if it's then possible to get the size of the combined array, e.g., for iterating over its elements? Or how would I otherwise use the combined array? Thanks, Robin -- Robin Sommer * Phone +1 (510) 666-2886 * robin at icir.org ICSI/LBNL * Fax +1

What API did you use in 3.2? I've never tried this, but I think it's probably possible. -Andy -----Original Message----- From: llvmdev-bounces at cs.uiuc.edu [mailto:llvmdev-bounces at cs.uiuc.edu] On Behalf Of Robin Sommer Sent: Tuesday, September 24, 2013 9:37 PM To: llvmdev at cs.uiuc.edu Subject: [LLVMdev] Linking static archives with MCJIT? Hi all, is there a way to link a native

I don't believe the changes have made it in yet. In addition to the previous patch set, we've got an independent implementation we've been using for a while: https://github.com/AzulSystems/llvm-late-safepoint-placement We really should get one or the other merged in tree. Could you point me to the previous patch set so I can ping it and review? Philip -------- Original Message

Hello, While trying ThinLTO, I ran into an assertion failure in IRMover: https://llvm.org/bugs/show_bug.cgi?id=28180. I found that the assertion failure is happening because IRMover tries to map the metadata that already mapped in the destination module, and it seems that this happens because two different IRMovers are used for the same destination (or composite) module. During LTO, an IRMover

Yes, I have the repro, though I can’t publish it externally. It would be great if you can upstream the patch so I can try it. Thank you for your explanation as well! -- Taewook From: <mehdi.amini at apple.com> on behalf of Mehdi Amini <mehdi.amini at apple.com> Date: Friday, July 22, 2016 at 2:16 PM To: Taewook Oh <twoh at fb.com> Cc: via llvm-dev <llvm-dev at

Any suggestions? It's not a problem for me, I'm mostly curious as it makes it harder to read the IR. Feel free to point me to other documentation if I'm missing something. Thanks, Robin On Wed, Jul 27, 2016 at 08:10 -0700, I wrote: > The language reference says: "identified [structure] types are never > uniqued". But llvm-link seems to do just that at link-time: >

Sorry for the spam, but we're looking for a postdoc with compiler expertise, and in particular LLVM, for ICSI's networking group. I thought this list might have somebody interested in such a position: http://www.icsi.berkeley.edu/about/netjob.html Feel free to forward further. Thanks, Robin -- Robin Sommer * Phone +1 (510) 722-6541 * robin at icsi.berkeley.edu ICSI/LBNL *

Hello, I am editing the LowerTypeTests pass in LLVM, and part of my additions include the following 3 lines of code: // newTypeName is a std::string MDString* newMD = MDString::get(M.getContext(), newTypeName); ArrayRef<Metadata*> mdArray {ConstantInt::get(Int64Ty, 0), newMD}; auto* node = MDTuple::get(M.getContext(), mdArray); Thus far, I have been developing on a version of Clang with

Hello all, I've run into a problem with returning a struct on Linux x86. Consider the following two pieces of code: # cat a.ll declare { i32, i32 } @foo(i32, i32, i32) define void @do_foo() { call { i32, i32 } @foo(i32 1, i32 2, i32 3) ret void } # cat b.c #include <stdio.h> #include <stdint.h> extern void

On Sat, Dec 05, 2009 at 22:56 -0800, I wrote: > the LLVM documentation lists the 'appending' linkage for having the > linker combine two arrays. I'm wondering though if it's then > possible to get the size of the combined array, e.g., for iterating > over its elements? Or how would I otherwise use the combined array? Any thoughts on this? Am I missing something?

It seems that llvm-ld is stripping debugging symbols even with optimizations disabled: > cat main.c int main() { int a = 42; return a; } > clang -c -g -emit-llvm -o main.bc main.c > llvm-ld -disable-opt -b=a.out.bc main.bc > clang -g a.out.bc -o a.out >gdb a.out [...] (gdb) break main Breakpoint 1 at 0x40045b:

I agree that the ArrayRef is likely the issue. I've debugged a crash caused by a temporary ArrayRef like that a couple times. Either do what David suggested or use a normal array: Metadata *mdArray[] = {ConstantInt::get(Int64Ty, 0), newMD}; ~Craig On Thu, Apr 30, 2020 at 9:56 AM David Blaikie via llvm-dev < llvm-dev at lists.llvm.org> wrote: > > > On Thu, Apr 30, 2020 at

I seem to remember seeing somewhere (on this list/on the web -- don't remember) that there was some ``Snort-like'' software that was available under the BSD license. Unfortunately, I'm unable to find any information about such software. Was I dreaming, or can anybody else jog my memory? :) Kind regards, Devon H. O'Dell