similar to: Testing an LLVM pass

On Mon, Mar 02, 2015 at 12:12:34AM -0500, John Criswell wrote: > On 3/2/15 12:07 AM, Haopeng Liu wrote: > >Got it, thanks. But in my pass, I use function name to locate. Can I > >disable mangling in clang? > > No, but you can probably fine a library that can either mangle the original > name or demangle the name you're seeing in the LLVM bitcode. > > As an FYI,

> >> You can look at AddOptimizationPasses() in opt.cpp. > > > > As far as I understand, the two passmanager do not interleave their > > passes. It first runs all the function passes and below. Then all the > > module passes. So if you specify: > > > > opt -mymodulepass0 -myfunctionpass -mymodulepass1 > > > > What you actually get is:

Within the LLVM pass manager infrastructure, suppose we have two transformation passes, pass A makes some improvements, then pass B does likewise, but this creates opportunities for pass A to create further improvements (e.g. suppose B was function inlining) so it's desirable to run A again. How does the LLVM pass manager currently deal with this? -------------- next part -------------- An

Thanks, Vedant. Yes, we have done a lot of testing of Clang/LLVM (and GCC) in the past several years (more details at https://people.inf.ethz.ch/suz/emi/index.html): [GCC/LLVM bugs: *1,602* (total) / *1,007* (fixed)] [Reports: GCC (link1 <http://tinyurl.com/ojzdt78>, link2 <http://tinyurl.com/oxlkmjc>, link3 <http://tinyurl.com/jd3jnl3>, link4

You may also be interested in the following resources on compilers correctness (articles, software, and talks -- from the general topics to the ones specifically focused on testing, validation, and verification): https://github.com/MattPD/cpplinks/blob/master/compilers.correctness.md Best regards, Matt P. Dziubinski On 12/7/2018 20:19, Philip Reames via llvm-dev wrote: > There's also

On Fri, Nov 30, 2018 at 03:25:10PM +0100, Serge Guelton via llvm-dev wrote: > On Thu, Feb 01, 2018 at 04:47:03PM -0500, James Y Knight via llvm-dev wrote: > > That said, I do think it could make sense to prepare llvm for the world in > > which "python" is python3 on some systems. So, I'd propose the following: > > 1. Change all #! lines to say

Hi folks (CCing llvm-dev, but that's probably more of a cfe-dev topic), As a follow-up to that old thread about -D_FORTIFY_SOURCE=n http://lists.llvm.org/pipermail/cfe-dev/2015-November/045845.html And, more recently, to this fedora thread where clang/llvm -D_FORTIFY_SOURCE support is claimed to be only partial: https://pagure.io/fesco/issue/2020 I dig into the glibc headers in

Hi folks, some more information on this feature - as a reminder I started one month ago to work on an expensive check that would verify that pass return status is correctly reported by passes, i.e. no pass return « IR not modified » while actually modifying it. It took ~20 pass fixes to achieve that goal, as many passes were not respectful of that contract, but as of

Hi folks, today, I wanted to concatenate two .ll, and naively typed: $ llvm-cat -help This is the not-so-helpful output I had: OVERVIEW: Module concatenation USAGE: llvm-cat [options] <input files> OPTIONS: General options: -aarch64-neon-syntax - Choose style of NEON code to emit from AArch64 backend: =generic

> Out of curiosity, does change here include changes to names, and other semantically-irrelevant changes (e.g., changing the order of operands in a PHI)? The hashing function used to detect changes is currently very simple: it only accounts for instruction opcode and order. So some semantically-irrelevant changes are ignored (as well as some relevant changes), and some are not. Permuting two

Hi folks, Python2 has reached end of support[0], and many core Python packages are dropping pthon 2 support[1]. This is a subject that raises periodically on this mlist, with a rather strong no in 2018 [-1] and a slow move in 2019 [-2, -3]. Even if Python is not a core build requirement, it's used during some configurations steps (e.g. in the cmake export_executable_symbols function), for

Hi Folks, Although LLVM now relies on cmake for its build configuration, it still calls an external Python script, namely llvm-build, to manage component dependecies, activation or deactivation of target componenents and generating description for llvm-config. This system is documented in http://llvm.org/docs/CommandGuide/llvm-build.html and https://llvm.org/docs/LLVMBuild.html. Several

Hi folks, Per the documentation[0], whenever an LLVM pass doesn't modify the IR it's run on, it should return `false`--it's okay to return `true` if no change happen, just less optimal. In the New PM area, this is generally translated into a `PreservedAnalyses::all()`. https://reviews.llvm.org/D80916 provides an `EXPENSIVE_CHECK` that computes a hash of the IR before and after the

Hi llvm-dev, Every now and then, when building a direct ``CallInst`` either from IRBuilder or through ``Create*``, I have a latent bug because the calling convention (CC) expected by the function is not set on the call, and I need to manually call ``setCallingConv``. Now, from LangRef#calling-conventions, The calling convention of any pair of dynamic caller/callee must match, or the

The next LLVM social in Paris will happen on January 30th, 2018. Everyone interested in LLVM, Clang, lldb, Polly, lld, ... is invited to join. Event details, including registration (free but mandatory) at http://www.meetup.com/LLVM-Clang-social For this meetup, Adrien Guinet, Serge Guelton and Juan Manuel Martinez will talk about the "Challenges when building an LLVM bitcode

Hi Folks, Now that LLVM 11.0.0 has been released, it's time to prepare for the final step envisionned in the previous RFC named *[RFC] Python 2 / Python 3 status* [0], ie. requiring Python3.6 for LLVM 12.0.0, to be released in 2021. At least Fedora already only ships Python3 and we didn't have much bugs reported wrt. Python compatibility for the LLVM toolchain. Indeed, all Python scripts

Yeah I have already see pintool but I need this to work with ARM and Intel. If I remember correctly, pintool does not work on ARM (or quite bad). My goal with the instrumentation is, once I have the information (time + call), to choose (during a second compilation) if some of my passes are applicable or not. Greetings, Johan On 17 Jul 2015, at 16:09, Kenneth Adam Miller <kennethadammiller

The PGO was my first guess but I can get a lot of information. At first, I follow the explanation at http://clang.llvm.org/docs/UsersManual.html#profiling-with-instrumentation but instead of llvm-profdata merge, I used llvm-profdata show *.profraw. Sadly, the information I get is the total number of function, the maximum function count and the maximum internal block count. Do you know if you

Hello everyone! We are working on writing a paper about testing the reliability of C compilers by using Csmith (a random C99 program generator). A previous testing effort, using Csmith, found 202 LLVM bugs, which represented 2% of all reported bugs at that time (PDF: https://www.flux.utah.edu/download?uid=114 <https://www.flux.utah.edu/download?uid=114>): . However, after this paper was

I am pleased to announce that LLVM 10 is now available. Get it here: https://llvm.org/releases/download.html#10.0.0 This release is the result of the LLVM community's work over the past six months (up to to e26a78e7085 on master plus commits up to d32170dbd5b on the release/10.x branch). Some highlights include: - C++ Concepts support in Clang - Clang no longer runs in a separate process