# gcc GNU Compiler Collection is a suite of compilers and language front ends for multiple programming languages. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## How fast is C++23's std::flat\_map? DevFeed: [How fast is C++23's std::flat\_map?]() Original publisher: [Read original article]() Author: Daniel Lemire Published: 2026-09-16T20:26:36Z Content type: article Language: en Sources: [Daniel Lemire]() Topics: [Data structures](), [Library](), [gcc](), [LLVM](), [clang]() Tags: [array](), [arrays](), [clang](), [gcc](), [llvm](), [overhead](), [speed](), [standard-library]() ### AI overview This article benchmarks C++23's std::flat_map, a sorted container backed by parallel arrays of keys and values. It explains serialization and loading considerations, then compares insertion and lookup performance with std::map. Random-order insertion becomes quadratic as the container grows, while increasing-order or bulk insertion is much faster; random lookups can also be faster for large maps because std::flat_map uses less memory. ### Source excerpt C++23 added a new type to the standard library: std::flat_map. There is also a std::flat_set and other variants, but let me focus on std::flat_map. A flat map is a sorted vector of keys next to a vector of values. A query is a binary search over the sorted keys. You need a recent standard library: ... Continue reading How fast is C++23's std::flat_map? ## Announcing Fedora Linux 45 Beta DevFeed: [Announcing Fedora Linux 45 Beta]() Original publisher: [Read original article]() Author: Aoife Moloney Published: 2026-09-15T14:05:00Z Content type: release Language: en Sources: [Fedora Magazine]() Topics: [Fedora](), [Linux](), [dnf](), [podman](), [pandas](), [Python](), [gcc](), [Go Language](), [MariaDB](), [MySQL](), [Filesystems]() Tags: [2026](), [dnf](), [fedora-project-community](), [filesystems](), [gcc](), [go](), [linux](), [mysql](), [pandas](), [partitioning](), [podman](), [pre-release](), [python](), [release](), [toolchains]() ### AI overview Fedora Linux 45 Beta is available for download or upgrade on September 15, 2026. The release includes changes to the kernel console, package signature verification, desktop secret management, software versions, Atomic Desktop image builds, Fedora CoreOS swap support, Anaconda installation flows, and Stratis filesystem partitioning. ### Source excerpt Fedora Linux 45 Beta is here! On Tuesday, 15 September 2026 you can download or upgrade your systems from your usual spots to start enjoying what F45 has to offer early. How to get the beta release You can download F45 Beta, or our pre-release edition versions, from any of the following places: You can [...] ## GCC 13.5 Released With 265+ Bug Fixes DevFeed: [GCC 13.5 Released With 265+ Bug Fixes]() Original publisher: [Read original article]() Author: Michael Larabel Published: 2026-09-11T09:47:06Z Content type: news Language: en Sources: [Phoronix]() Topics: [gcc](), [Compiler](), [bug](), [cpu](), [intel]() Tags: [amd](), [bug-fixes](), [bugs](), [compiler](), [cpu](), [desktop-linux](), [gcc](), [intel](), [linux-benchmarking](), [linux-hardware-benchmarks](), [linux-hardware-reviews](), [linux-how-to](), [linux-performance](), [linux-server-benchmarks](), [open-source-graphics](), [phoronix](), [phoronix-test-suite](), [production](), [release](), [ubuntu-benchmarks](), [ubuntu-hardware]() ### AI overview GCC 13.5 has been released with more than 265 bug fixes. It is the final planned point release for the GCC 13 branch, which is now closed to further development. ### Source excerpt For those continuing to rely on the GCC 13 compiler in production that debuted more than three years ago, GCC 13.5 released today with more than 265 bug fixes. GCC 13.5 also caps off the GCC 13 series as the last planned point release... ## Parsing JSON at compile time with C++26 static reflection DevFeed: [Parsing JSON at compile time with C++26 static reflection]() Original publisher: [Read original article]() Author: Daniel Lemire Published: 2026-06-14T14:59:44Z Content type: tutorial Language: en Sources: [Daniel Lemire]() Topics: [C++](), [JSON](), [C](), [Compiler](), [configuration](), [gcc]() Tags: [c](), [c-plus-plus](), [compiler](), [configuration](), [gcc](), [json](), [library](), [parsing]() ### AI overview This tutorial demonstrates parsing a fixed JSON configuration file at compile time with C++26 static reflection and GCC 16. It explains embedding the file's bytes in the executable, using simdjson to synthesize a typed C++ object, and detecting malformed JSON during compilation instead of at startup. ### Source excerpt Suppose that you have a configuration file in JSON. Something like this: { "width": 1920, "height": 1080, "fullscreen": true, "title": "My Game", "volume": 0.8 } Normally you ship this file alongside your program, open it at startup, read it, and parse it. That is a lot of work for data that never changes. What if ... Continue reading Parsing JSON at compile time with C++26 static reflection ## wastrel milestone: full hoot support, with generational gc as a treat DevFeed: [wastrel milestone: full hoot support, with generational gc as a treat]() Original publisher: [Read original article]() Author: Andy Wingo Published: 2026-04-09T13:48:04Z Content type: article Language: en Sources: [wingolog]() Topics: [WebAssembly](), [scheme](), [toolchain](), [gcc]() Tags: [binaries](), [garbage-collection](), [gc](), [gcc](), [hoot](), [optimization](), [repl](), [scheme](), [spidermonkey](), [toolchain](), [treats](), [v8](), [wasm](), [wastrel](), [web](), [webassembly]() ### AI overview The article describes a Wastrel milestone: compiling WebAssembly files produced by the Hoot Scheme toolchain into native binaries, including a console-based read-eval-print loop. It covers the build process, runtime modules, binary size, dependencies, and initial performance observations. ### Source excerpt Hear ye, hear ye: Wastrel and Hoot means REPL! Which is to say, Wastrel can now make native binaries out of WebAssembly files as produced by the Hoot Scheme toolchain, up to and including a full read-eval-print loop. Like the REPL on the Hoot web page, but instead of requiring a browser, you can just run it on your console. Amazing stuff! try it at home First, we need the latest Hoot. Build it from source, then compile a simple REPL: echo '(import (hoot repl)) (spawn-repl)' > repl.scm ./pre-inst-env hoot compile -fruntime-modules -o repl.wasm repl.scm This takes about a minute. The resulting wasm file has a pretty full standard library including a full macro expander and evaluator. Normally Hoot would do some aggressive tree-shaking to discard any definitions not used by the program, but with a REPL we don't know what we might need. So, we pass -fruntime-modules to instruct Hoot to record all modules and their bindings in a central registry, so they can be looked up at run-time. This results in a 6.6 MB Wasm file; with tree-shaking we would have been at 1.2 MB. Next, build Wastrel from source, and compile our new repl.wasm: wastrel compile -o repl repl.wasm This takes about 5 minutes on my machine: about 3 minutes to generate all the C, about 6.6MLOC all in all, split into a couple hundred files of about 30KLOC each, and then 2 minutes to compile with GCC and link-time optimization (parallelised over 32 cores in my case). I have some ideas to golf the first part down a bit, but the the GCC side will resist improvements. Finally, the moment of truth: $ ./repl Hoot 0.8.0 Enter `,help' for help. (hoot user)> "hello, world!" => "hello, world!" (hoot user)> statics When I first got the REPL working last week, I gasped out loud: it's alive, it's alive!!! Now that some days have passed, I am finally able to look a bit more dispassionately at where we're at. Firstly, let's look at the compiled binary itself. By default, Wastrel passes the -g flag to GCC, which results in bi ## The 64-bit Hurd is Here! DevFeed: [The 64-bit Hurd is Here!]() Original publisher: [Read original article]() Author: Janneke Nieuwenhuizen, Yelninei Published: 2026-03-01T10:00:00Z Content type: article Language: en Sources: [GNU Guix -- Blog]() Topics: [x86](), [bug](), [patches](), [gcc]() Tags: [bug](), [gcc](), [gnu-hurd](), [patches](), [x86](), [x86-64]() ### AI overview Guix has added x86_64 support for the Hurd. The article reviews roughly 1.5 years of related work, including installer support, build and dependency fixes, package updates, and the merge of the 64-bit Hurd changes. ### Source excerpt Fifteen months have passed since our last Guix/Hurd on a Thinkpad X60 post and a lot has happened with respect to the Hurd . And most of you will have guessed, unless you skipped the title of this post, the rumored x86_64 support has landed in Guix! Here is a not-so-short overview of our Hurd work over the past 1.5 years: The build daemon fails when invoking guix authenticate on the Hurd bug was fixed. This was our most pressing problem as it meant that we could not keep... ## Bit-field layout DevFeed: [Bit-field layout]() Original publisher: [Read original article]() Published: 2026-02-22T08:00:00Z Content type: tutorial Language: en Sources: [MaskRay]() Topics: [implementation](), [C](), [C++](), [Compiler](), [clang](), [gcc](), [LLVM](), [MSVC]() Tags: [c](), [c-plus-plus](), [clang](), [codegen](), [compiler](), [compiler-optimization](), [gcc](), [implementation](), [llvm](), [msvc]() ### AI overview This article explains how C and C++ bit-field layout is implementation-defined and governed primarily by platform ABIs. It distinguishes ABI-defined storage layout from compiler code generation, focusing on the Itanium ABI and describing differences in the Microsoft ABI. ### Source excerpt The C and C++ standards leave nearly every detail to the implementation. C23 §6.7.3.2: An implementation may allocate any addressable storage unit large enough to hold a bit-field. If enough space remains, a bit-field that immediately follows another bit-field in a structure shall be packed into adjacent bits of the same unit. If insufficient space remains, whether a bit-field that does not fit is put into the next unit or overlaps adjacent units is implementation-defined. The order of allocation of bit-fields within a unit (high-order to low-order or low-order to high-order) is implementation-defined. The alignment of the addressable storage unit is unspecified C++ is also terse -- [class.bit]p1: Allocation of bit-fields within a class object is implementation-defined. Alignment of bit-fields is implementation-defined. Bit-fields are packed into some addressable allocation unit. ## \[RevEng\]\[x86\]\[x64\]\[C\]\[newbies\] Software reset in embedded devices DevFeed: [\[RevEng\]\[x86\]\[x64\]\[C\]\[newbies\] Software reset in embedded devices]() Original publisher: [Read original article]() Published: 2025-12-31T23:00:00Z Content type: tutorial Language: en Sources: [Dennis Yurichev]() Topics: [C](), [Embedded Software Dev](), [Code](), [x86](), [gcc]() Tags: [blog-post](), [c](), [code](), [compilation](), [embedded](), [embedded-devices](), [gcc](), [software](), [x86]() ### AI overview This tutorial explains a software-reset technique for embedded devices using pure C. It shows how calling a function pointer at address zero can jump to firmware on systems that start execution there after power-on or reset, and discusses the behavior of the stack and local variables. ### Source excerpt [RevEng][x86][x64][C][newbies] Software reset in embedded devices ## Progress update: fixing the ReactOS test suite DevFeed: [Progress update: fixing the ReactOS test suite]() Original publisher: [Read original article]() Published: 2025-11-04T00:00:00Z Content type: opinion Language: en Sources: [Front Page on ReactOS Website]() Topics: [Testing](), [WINE](), [Windows](), [gcc](), [MSVC](), [x86]() Tags: [changes](), [developer](), [free](), [gcc](), [msvc](), [open-source](), [os](), [react](), [reactos](), [tests](), [update](), [win32](), [winapi](), [windows](), [wine](), [x86]() ### AI overview A ReactOS project developer describes work to clean up the neglected test suite. The update reports fixes to 55 existing test modules, five new modules, broader testing across Windows versions and compiler toolchains, and changes intended to improve synchronization with Wine 10.0. The supplied text also states that the full test suite can now run on Windows Vista without a bugcheck. ### Source excerpt For many years, the ReactOS test suite was neglected. It was a random collection of our own tests and old Wine tests that were only checked against Windows Server 2003 and sometimes a random newer version of Windows up to the discretion of the contributor. The Wine tests we imported were heavily modified and the changes made weren't always well documented. I'm here to clean up this mess. I have been deeply involved with the ReactOS project since 2023 and in May of 2024 I became an official project developer. ## This Shit is Hard: Hardening glibc DevFeed: [This Shit is Hard: Hardening glibc]() Original publisher: [Read original article]() Published: 2025-08-20T00:00:00Z Content type: article Language: en Sources: [Chainguard: Unchained]() Topics: [Security](), [Compiler](), [gcc](), [toolchain](), [chainguard](), [chainguard packages](), [C](), [Linux]() Tags: [c](), [c-c-plus-plus](), [chainguard](), [chainguard-containers](), [chainguard-security](), [compiler](), [compiler-flags](), [gcc](), [glibc](), [hardening](), [openssf](), [security](), [toolchain]() ### AI overview Chainguard describes its effort to build glibc with hardened compiler flags. The work involved collaboration with the upstream GCC and glibc projects to resolve complex issues and ship a hardened version of the library. ### Source excerpt Chainguard uses compiler flags to be proactive in the security of our products. See how our compiler flag usage helped us catch a complex bug in glibc. ## Using \`make\` to compile C programs (for non-C-programmers) DevFeed: [Using \`make\` to compile C programs (for non-C-programmers)]() Original publisher: [Read original article]() Author: Julia Evans Published: 2025-06-10T00:00:00Z Content type: tutorial Language: en Sources: [Julia Evans]() Topics: [C](), [make](), [Compiler](), [Package manager](), [gcc](), [apt](), [Xcode](), [C++]() Tags: [apt](), [building](), [c](), [c-plus-plus](), [command-line](), [compiler](), [debian](), [dependencies](), [gcc](), [libraries](), [linux](), [mac](), [make](), [packages](), [xcode]() ### AI overview A beginner-oriented guide to compiling C programs from source, covering compiler installation, dependency discovery, and build tools on Ubuntu and macOS. ### Source excerpt I have never been a C programmer but every so often I need to compile a C/C++ program from source. This has been kind of a struggle for me: for a long time, my approach was basically "install the dependencies, run make, if it doesn't work, either try to find a binary someone has compiled or give up". "Hope someone else has compiled it" worked pretty well when I was running Linux but since I've been using a Mac for the last couple of years I've been running into more situations where I have to actually compile programs myself. So let's talk about what you might have to do to compile a C program! I'll use a couple of examples of specific C programs I've compiled and talk about a few things that can go wrong. Here are three programs we'll be talking about compiling: paperjam sqlite qf (a pager you can run to quickly open files from a search with rg -n THING | qf) step 1: install a C compiler This is pretty simple: on an Ubuntu system if I don't already have a C compiler I'll install one with: sudo apt-get install build-essential This installs gcc, g++, and make. The situation on a Mac is more confusing but it's something like "install xcode command line tools". step 2: install the program's dependencies Unlike some newer programming languages, C doesn't have a dependency manager. So if a program has any dependencies, you need to hunt them down yourself. Thankfully because of this, C programmers usually keep their dependencies very minimal and often the dependencies will be available in whatever package manager you're using. There's almost always a section explaining how to get the dependencies in the README, for example in paperjam's README, it says: To compile PaperJam, you need the headers for the libqpdf and libpaper libraries (usually available as libqpdf-dev and libpaper-dev packages). You may need a2x (found in AsciiDoc) for building manual pages. So on a Debian-based system you can install the dependencies like this. sudo apt install -y libqpdf-dev libpaper-dev ## NixOS 25.05 released DevFeed: [NixOS 25.05 released]() Original publisher: [Read original article]() Author: NixOS Published: 2025-05-23T12:00:00Z Content type: release Language: en Sources: [NixOS Announcements]() Topics: [Nix](), [Release notes](), [releases](), [Linux](), [configuration](), [modules](), [Linux Kernel](), [distributed-systems](), [gcc](), [LLVM](), [upgrade]() Tags: [bugfixes](), [compilers](), [deprecated](), [end-of-life](), [features](), [gcc](), [kernel](), [linux](), [linux-kernel](), [llvm](), [nix](), [nixos](), [release](), [release-notes](), [updated](), [updates](), [version]() ### AI overview NixOS 25.05, codenamed "Warbler," is publicly available. The release adds and updates packages, introduces configuration modules and options, updates GNOME and the default Linux kernel, and updates LLVM and GCC. NixOS 25.05 receives bugfixes and security updates through 2025-12-31, while NixOS 24.11 reaches end of life after 2025-06-30. ### Source excerpt Hey everyone, we are Leona Maroni and Tristan Ross, the release managers of the newest release of NixOS. We are very proud to announce the public availability of NixOS 25.05 "Warbler". NixOS is a Linux distribution. Its underlying package repository Nixpkgs can also be used on other Linux systems and macOS with the Nix package manager. This release will receive bugfixes and security updates for seven months (up until 2025-12-31). The old release 24.11 "Vicuña" is now officially deprecated and will reach its end-of-life and stop receiving security updates after 2025-06-30. NixOS Release Notes Highlights New Services Backward Incompatibilities Other Notable Changes Nixpkgs Release Notes Highlights Backward Incompatibilities Other Notable Changes Nixpkgs Library Upgrade Instructions ISO Image Downloads The 25.05 release was made possible due to the efforts of 2857 contributors, who authored 57054 commits since the previous release. Highlights NixOS is already known as the most up to date distribution while also being the distribution with the most packages. This release of Nixpkgs Added 7840 new packages Updated 28054 existing packages Removed 1694 outdated packages, in an effort to keep the package set maintainable and secure. In addition to packages from Nixpkgs, the NixOS Linux distribution also features composable configuration modules and integration tests for distributed systems. This release of NixOS Added 137 new modules and 1930 configuration options Removed 10 outdated modules and 215 configuration options. GNOME 48 GNOME has been updated to version 48 "Bengaluru", which introduces notification stacking, a new music player, HDR support, and more changes. Refer to the release notes for more details. Kernel The default Linux kernel has been updated from 6.6 to 6.12. All supported kernels remain available. C compilers LLVM has been updated to version 19. GCC has been updated to version 14. Special Thanks We want to personally thank Seth and raf for editorializin ## Performance of the Python 3.14 tail-call interpreter DevFeed: [Performance of the Python 3.14 tail-call interpreter]() Original publisher: [Read original article]() Author: Nelson Elhage Published: 2025-03-09T22:00:00Z Content type: article Language: en Sources: [Nelson Elhage]() Topics: [Python 3.14](), [benchmarking](), [gcc](), [LLVM](), [Software Engineering]() Tags: [benchmark](), [benchmarking](), [benchmarks](), [bytecode](), [clang](), [compilers](), [cpu](), [gcc](), [intel](), [llvm](), [performance](), [performance-engineering](), [python](), [python-3-14](), [software-engineering]() ### AI overview This article examines the performance gains attributed to CPython's Python 3.14 tail-call bytecode interpreter. It finds that the initial 10-15% improvement was primarily caused by inadvertently working around an LLVM 19 regression; with better baselines, the gain is closer to 1-5%, depending on the setup. ### Source excerpt About a month ago, the CPython project merged a new implementation strategy for their bytecode interpreter. The initial headline results were very impressive, showing a 10-15% performance improvement on average across a wide range of benchmarks across a variety of platforms. Unfortunately, as I will document in this post, these impressive performance gains turned out to be primarily due to inadvertently working around a regression in LLVM 19. When benchmarked against a better baseline (such GCC, clang-18, or LLVM 19 with certain tuning flags), the performance gain drops to 1-5% or so depending on the exact setup. ## Bazel caching and compressed debug info DevFeed: [Bazel caching and compressed debug info]() Original publisher: [Read original article]() Author: Keith Smiley Published: 2025-02-14T18:00:00Z Content type: tutorial Language: en Sources: [Keith Smiley]() Topics: [C++](), [Caching](), [LLVM](), [gcc](), [C]() Tags: [bazel](), [build](), [c-plus-plus](), [caching](), [clang](), [cmake](), [config](), [configuration]() ### AI overview This article explains how compressing debug information can reduce C and C++ binary sizes and improve Bazel remote-cache efficiency. Using llvm-objcopy, the example reduces a roughly 536 MB binary to about 290 MB, and the authors report nearly 60% fewer cache reads after deployment. ### Source excerpt One of bazel's most attractive features is the ability for it to remotely cache artifacts to reduce unnecessary work for large builds. Unfortunately users quickly discover this comes with non-trivial financial and bandwidth implications. There are many ways, of varying difficulty, to try and improve your cache usage. From breaking unnecessary dependencies, to adding larger local storage for CI workers, builds without the bytes, build avoidance, etc. For codebases with lots of C or C++ one of the potentially easiest wins is to enable compressed debug information1. Let's look at an example from our codebase. Looking at the size of a non-trivial C++ binary built with -g -O2 (similar to cmake's RelWithDebInfo configuration), or binary clocks in at ~530mbs: % du -sh bin 536M bin To get a sense of what percentage of this binary is debug info, we can use llvm-objcopy to strip the debug info entirely: % llvm-objcopy --strip-debug bin strippedbin % du -sh strippedbin 159M strippedbin This shows us that almost 70%(!!) of the binary size is taken up with debug info. In release configurations we can eliminate this entirely with bazel's --strip argument, but for developer builds, or other use cases where you need debug info, we can still improve this. If we use llvm-objcopy again, this time to compress the debug info, we can immediately see our potential gains: % llvm-objcopy --compress-debug-sections bin compressedbin % du -sh compressedbin 290M compressedbin This shows us we can get an almost 50%(!!) improvement in binary size in this example. To enable this in bazel, assuming you're using a relatively recent version of gcc or clang, you can add something like this to your .bazelrc2: build --enable_platform_specific_config build:linux --copt=-gz --host_copt=-gz build:linux --linkopt=-gz --host_linkopt=-gz In practice we saw cache reads drop by nearly 60% when we rolled out this change. Reducing binary size with this approach has a lot of benefits, but it's even more pronounced ## Using Compiler Flags to Secure Your Code DevFeed: [Using Compiler Flags to Secure Your Code]() Original publisher: [Read original article]() Published: 2025-01-22T00:00:00Z Content type: tutorial Language: en Sources: [Chainguard: Unchained]() Topics: [Compiler](), [Code](), [Memory Safety](), [Security](), [C](), [C++](), [gcc](), [Dockerfile]() Tags: [c](), [c-plus-plus](), [chainguard-images](), [code](), [compiler](), [compiler-flags](), [container](), [gcc](), [memory-safety](), [open-ssf](), [security]() ### AI overview This tutorial examines OpenSSF-recommended compiler flags for C and C++ code, focusing on how hardened standard-library calls and related protections affect memory safety and buffer-overflow behavior. It demonstrates the examples using a Docker-based development environment. ### Source excerpt Open SSF's recommended compiler flags are a great way to improve memory safety and security. Check out what effect using these flags can have. ## C++ exception performance three years later DevFeed: [C++ exception performance three years later]() Original publisher: [Read original article]() Author: Thomas Neumann (noreply@blogger.com) Published: 2024-12-10T14:44:00Z Content type: article Language: en Sources: [Database Architects]() Topics: [C++](), [Exception](), [Scalability](), [gcc](), [JIT](), [LLVM]() Tags: [benchmark](), [c-plus-plus](), [clang](), [exception](), [gcc](), [glibc](), [jit](), [llvm](), [performance](), [series]() ### AI overview The article reviews improvements to C++ exception unwinding performance. Lock-free lookup mechanisms in glibc and libgcc improve scalability for statically generated and JIT-generated code, although clang's implementation may still have scaling limitations. ### Source excerpt About three years ago we noticed serious performance problems in C++ exception unwinding. Due to contention on the unwinding path these became more and more severe the more cores a system had, and unwinding could slow down by orders of magnitude. Due to the constraints of backwards compatibility this contention was not easy to eliminate, and P2544 discussed ways to fix this problem via language changes in C++. But fortunately people found less invasive solutions. First, Florian Weimer changed the glibc to provide a lock-free mechanism to find the (static) unwind tables for a given shared object. Which eliminates the most serious contention for "simple" C++ programs. For example in a micro-benchmark that calls a function with some computations (100 calls to sqrt per function invocation), and which throws with a certain probability, we previously had very poor scalability with increasing core count. With his patch we now see with gcc 14.2 on a dual-socket EPYC 7713 the following performance development (runtime in ms): 1 2 4 8 16 32 64 128 threads 0% failure 29 29 29 29 29 29 29 42 0.1% failure 29 29 29 29 29 29 29 32 1% failure 29 30 30 30 30 30 32 34 10% failure 36 36 37 37 37 37 47 65 Which is more or less perfect. 128 threads are a bit slower, but that is to be expected as one EPYC only has 64 cores. With higher failure rates unwinding itself becomes slower but that is still acceptable here. Thus most C++ programs are just fine. For our use case that is not enough, though. We dynamically generate machine code at runtime, and we want to be able to pass exceptions through generated code. The _dl_find_object mechanism of glibc is not used for JITed code, instead libgcc maintains its own lookup structure. Historically this was a simple list with a global lock, which of course had terrible performance. But through a series of patches we managed to change libgcc into using a lock-free b-tree for maintaining the dynamic unwinding frames. Using a similar experiment to the ## GNU/Hurd development update for Q2 2024 DevFeed: [GNU/Hurd development update for Q2 2024]() Original publisher: [Read original article]() Published: 2024-06-27T06:07:00Z Content type: news Language: en Sources: [GNU Hurd]() Topics: [GNU/Hurd](), [Debian](), [gcc](), [Compiler](), [Unix](), [test](), [Hardware]() Tags: [compiler](), [debian](), [gcc](), [gnu-hurd](), [hardware](), [ssh](), [tests](), [unix]() ### AI overview A Q2 2024 update on GNU/Hurd development covers the AArch64 GNU Mach port, merged GCC support for compiling GNU/Hurd programs on AArch64, bug fixes and tests, experimental SMP support, Rust compiler availability, and Debian package compatibility. ### Source excerpt Hello! Welcome to a new qoth. This qoth covers new and interesting GNU/Hurd developments in Q2 of 2024! Details. Sergey Bugaev committed public headers for the GNU Mach AArch64 port. He writes that ...there is now a real port of GNU Mach to AArch64, using these headers as its actual API/ABI. We got the Mach port to run glibc, several Hurd servers, and simple Unix programs, including things like fork/exec and signal delivery & handling working, which exercises these architecture-specific definitions (thread state & exceptions). We have also managed to do some testing on real hardware; although not everything is working yet, we have seen thread state manipulation & Mach handling an unaligned SP fault work as expected. His email also mentions that the GCC patches that enable GCC to compile GNU/Hurd programs on AArch64 have been merged! This apparently will make it easier to merge his AArch64 specific glibc patches. He also added new tests to check that threads handle signals well, and he also fixed a use-after-free in vmmappageable_scan(). He also hosted a lengthy Hurd code jam (apologies for the poor audio quality). He also very notably added support to copy a send once right to Mach and MIG. Some time ago, Sergey also wrote the [terrible-mdns-responder][hurd/terrible-mdns-responder], and if you would like to be able to type in ssh HOSTNAME.local and connect to a locally running Hurd, then you may want to try it! Flávio Cruz fixed some issues with the Hurd compiling on GCC 14. Luca Dariz fixed message sizes, where the size was not set by userspace, and he added another test to check message sizes on various code paths. Debian GNU/Hurd now offers an experimental SMP GNU Mach kernel (32-bit only) and the official rustc compiler! Now that we have ported rustc to Debian GNU/Hurd, we can compile important packages like librsvg. Debian GNU/Hurd now can compile 71% of the packages from the Debian archive. Now for something trivial but fun! I updated the guide on the Hurd wik ## GNU/Hurd development updates from Q4 2023 DevFeed: [GNU/Hurd development updates from Q4 2023]() Original publisher: [Read original article]() Published: 2024-01-05T22:22:00Z Content type: news Language: en Sources: [GNU Hurd]() Topics: [gcc](), [Linux](), [CI/CD](), [Testing](), [qemu](), [x86](), [Kernel](), [make]() Tags: [continuous-integration](), [gcc](), [kernel](), [linux](), [make](), [qemu](), [testing](), [x86]() ### AI overview A quarterly update on GNU/Hurd development describes GCC and GNU Mach improvements, proposed continuous integration, expanded testing across environments, x86_64 build infrastructure, IPC changes, and efforts to improve browser support. ### Source excerpt Hello! Welcome to a new qoth. This qoth covers new and interesting GNU/Hurd developments in Q4 of 2023! Details. Samuel Thibault fixed gcc's Hurd's default pie and added static pie support. He also added a whatis command to GNU Mach's kernel debugger, which can determine what an address points to (a stack? a port? some kalloc? ...). He also added hurd-amd64 support to GCC. Samuel requested that the Hurd team set up a continuous integration, so that when developers make code changes, they can be certain that they did not break anything. It turns out that the Hurd supports several different environments: 32 bit, 64 bit, 32-on-64 bit, ACPI, non-ACPI, SMP, non-SMP, Xen, etc. Apparently Flavio has a personal CI, but it is set up in a Debian independent way. If you are interested in helping the Hurd project set up a CI, then please get in touch! Luca Dariz worked on adding some simple GNU Mach user-space tests . With a working MiG, a GNU/Linux machine can run make check in the GNU Mach source code, which will launch qemu to ensure that 32 bit (PAE and non PAE), 32 on 64 bit, and full 64 bit GNU Mach works. We currently do this testing on GNU/Linux, because qemu does not run on the Hurd. Many people worked on the Hurd's new x86_64 bit support. A 64-bit debian buildd is set up, and we can bootstrap a chroot! The hurd-amd64 wanna-build infrastructure is also set up. We are having issues reliably building packages on a 64-bit Hurd, which lead Samuel to uncover and fix a proc leak. Flavio Cruz improved GNU Mach's IPC by reordering mach_msg_type_t fields to byte align msgt_name and msgt_size. He also created a patch series to avoid message resizing for x86_64. He also removed untyped mach RPC code. GNU Mach uses typed IPC. The Hurd could support both typed and untyped, but it appears that the Hurd only uses typed RPC. So it seems best to remove any untyped RPC code. Sergey Bugaev added GNU Mach entry re-coalescing support. Essentially, Mach was not always able to merge two vm e ## Newsletter 103 - Late 2023 news DevFeed: [Newsletter 103 - Late 2023 news]() Original publisher: [Read original article]() Published: 2023-11-05T00:00:00Z Content type: news Language: en Sources: [Front Page on ReactOS Website]() Topics: [ReactOS](), [debug](), [Kernel](), [gcc](), [MSVC](), [Pull Request]() Tags: [bugfixes](), [bugs](), [debug](), [debugger](), [free](), [gcc](), [msvc](), [open-source](), [os](), [react](), [reactos](), [release](), [version](), [win32](), [winapi]() ### AI overview ReactOS Newsletter 103 reports that version 0.4.15 is still planned but delayed by unresolved regressions and bugs, including a large-file copying regression. It also discusses an overhaul of the ReactOS kernel debugging setup and the use of GCC, MSVC, and WinDbg. ### Source excerpt Salutations from the ReactOS project team! In previous posts, we talked about the ReactOS releasing process and the development status of the project, as well as the hiring of our long-term developer Hermès Bélusca-Maïto (HBelusca). We are making an effort to publish at least 3 newsletters per year, depending on how the development workflow goes. In this newsletter we will highlight some of the contributions made by project developers and contributors, as well as future plans and headlines. ## Fully bootstrapping Go from source in Wolfi DevFeed: [Fully bootstrapping Go from source in Wolfi]() Original publisher: [Read original article]() Published: 2023-08-11T00:00:00Z Content type: article Language: en Sources: [Chainguard: Unchained]() Topics: [Go](), [Go Language](), [Compiler](), [toolchain](), [gcc](), [C++](), [backdoor]() Tags: [backdoor](), [binaries](), [bootstrapping](), [c-plus-plus](), [gcc](), [go](), [provenance](), [source](), [supply-chain](), [toolchain](), [wolfi]() ### AI overview The article explains how to fully bootstrap Go from source in Wolfi to improve provenance and reduce trust risks associated with pre-built binaries. It describes using the C++-based gcc-go or gollvm implementations to build the official Go toolchain without requiring an existing Go compiler. ### Source excerpt Dive into fully bootstrapping Go from source in Wolfi, paving the way for secure, independent development. ## Fully bootstrapping Java from source in Wolfi DevFeed: [Fully bootstrapping Java from source in Wolfi]() Original publisher: [Read original article]() Published: 2023-06-02T00:00:00Z Content type: article Language: en Sources: [Chainguard: Unchained]() Topics: [Java](), [openjdk](), [supply-chain-security](), [chainguard](), [Compiler](), [Programming](), [gcc]() Tags: [bootstrapping](), [chainguard](), [chainguard-images](), [compiler](), [distribution](), [eclipse-compiler-for-java](), [gcc](), [gnu-linux](), [java](), [jdk](), [linux](), [openjdk](), [programming](), [programming-languages](), [provenance](), [security](), [software](), [software-supply-chain](), [software-supply-chain-security](), [supply-chain-security](), [wolfi]() ### AI overview Chainguard describes how it is fully bootstrapping Java from source in Wolfi. The process builds a chain of OpenJDK packages with provenance extending from source code to final binary packages, using GNU Classpath, GNU Compiler for Java, and the Eclipse Compiler for Java to overcome historical bootstrapping dependencies. ### Source excerpt Learn how Chainguard engineers created a chain of OpenJDK packages to provide full provenance from pure source code for the entire Java ecosystem in Wolfi. ## Making unwinding through JIT-ed code scalable - Replacing the gcc hooks DevFeed: [Making unwinding through JIT-ed code scalable - Replacing the gcc hooks]() Original publisher: [Read original article]() Author: Thomas Neumann (noreply@blogger.com) Published: 2022-06-26T08:48:00Z Content type: article Language: en Sources: [Database Architects]() Topics: [gcc](), [patches](), [JIT]() Tags: [atomic](), [gcc](), [jit](), [lock-free](), [mutex](), [patches]() ### AI overview This article explains patches to GCC that replace a globally locked list of unwinding frames with a read-optimized B-tree. The change enables concurrent registration, deregistration, and lock-free lookups, while keeping frames immutable during unwinding on platforms that support atomics. ### Source excerpt This article is part of the series about scalable unwinding that starts here. As discussed in the previous article, the gcc mechanism does not scale because it uses a global lock to protect its list of unwinding frames. To solve that problem, we replace that list with a read-optimized b-tree that allows for concurrent reads and writes. In this article we just discuss the patches to gcc necessary to enable that mechanism, the b-tree itself is discussed in subsequent articles. We start by replacing the old fast path mechanism with a b-tree root: index 8ee55be5675..d546b9e4c43 100644 --- a/libgcc/unwind-dw2-fde.c +++ b/libgcc/unwind-dw2-fde.c @@ -42,15 +42,34 @@ see the files COPYING3 and COPYING.RUNTIME respectively. If not, see #endif #endif +#ifdef ATOMIC_FDE_FAST_PATH +#include "unwind-dw2-btree.h" + +static struct btree registered_frames; + +static void +release_registered_frames (void) __attribute__ ((destructor (110))); +static void +release_registered_frames (void) +{ + /* Release the b-tree and all frames. Frame releases that happen later are + * silently ignored */ + btree_destroy (®istered_frames); +} + +static void +get_pc_range (const struct object *ob, uintptr_t *range); +static void +init_object (struct object *ob); + +#else + /* The unseen_objects list contains objects that have been registered but not yet categorized in any way. The seen_objects list has had its pc_begin and count fields initialized at minimum, and is sorted by decreasing value of pc_begin. */ static struct object *unseen_objects; static struct object *seen_objects; -#ifdef ATOMIC_FDE_FAST_PATH -static int any_objects_registered; -#endif #ifdef __GTHREAD_MUTEX_INIT static __gthread_mutex_t object_mutex = __GTHREAD_MUTEX_INIT; @@ -78,6 +97,7 @@ init_object_mutex_once (void) static __gthread_mutex_t object_mutex; #endif #endif +#endif When the platform supports atomics (ATOMIC_FDE_FAST_PATH), we replace the whole mechanism with one b-tree, whose root is registered_frames. Neither the ## Making unwinding through JIT-ed code scalable DevFeed: [Making unwinding through JIT-ed code scalable]() Original publisher: [Read original article]() Author: Thomas Neumann (noreply@blogger.com) Published: 2022-06-26T08:46:00Z Content type: tutorial Language: en Sources: [Database Architects]() Topics: [C++](), [Exception](), [gcc](), [Concurrency](), [Jule](), [systems]() Tags: [c-plus-plus](), [compiler](), [concurrent](), [exception](), [gcc](), [hooks](), [jit](), [lock-free](), [mutex](), [series]() ### AI overview This article explains why C++ exception unwinding remains effectively single-threaded when JIT-ed code is registered. It describes limitations in gcc and glibc mechanisms and introduces a gcc patch using a read-optimized lock-free b-tree to support parallel unwinding without atomic writes. ### Source excerpt Exceptions are a very handy mechanism to propagate errors in C++ programs, but unfortunately they do not scale very well. In all common C++ implementations the unwinding mechanism takes global lock during unwinding, which has disastrous consequences when the number of threads is high. On a machine with 256 hardware context we see worse-than-single-threaded behavior even for relatively modest failure rates. Fortunately the Florian Weimer fixed one contention point in gcc 12 on systems with glibc 2.35 or newer, which gives us scalable exceptions as long as no JIT-ed code has been registered. Unfortunately our system does register JIT-ed code... Which means exception unwinding in our code base is still single-threaded in practice. But we can fix that by teaching gcc to store the unwinding information in a read-optimized b-tree, which allows for fully parallel unwinding without any atomic writes. There is a gcc patch that does just that, but unfortunately it is quite involved and difficult to review. This article series thus explains all parts of the patch and shows how a read-optimized b-tree can be implemented lock-free. In order to keep the article length somewhat reasonable, the discusses is broken into parts: The problem (this article) Replacing the gcc hooks Optimistic Lock Coupling The b-tree b-tree operations When unwinding exceptions, the compiler has to find the corresponding unwinding information for every call frame on the stack between the throw and the catch. gcc uses two different mechanisms for that: For ahead-of-time compiled code it asks glibc to find the unwinding information using either dl_iterate_phdr (on older systems) or _dl_find_object (on systems with glibc 2.35 or newer). Note that this mapping is not static, as shared libraries could be added or removed at any time, potentially during a concurrent unwind. For that reason dl_iterate_phdr was protected by a global mutex, which clearly does not scale. _dl_find_object avoids that mutex by using a ## Newsletter 101 - July 2021 news DevFeed: [Newsletter 101 - July 2021 news]() Original publisher: [Read original article]() Published: 2021-08-01T00:00:00Z Content type: news Language: en Sources: [Front Page on ReactOS Website]() Topics: [ReactOS](), [Kernel](), [gcc](), [MSVC](), [Linux](), [macOS](), [Windows]() Tags: [free](), [gcc](), [kernel](), [linux](), [macos](), [msvc](), [news](), [open-source](), [os](), [react](), [reactos](), [release](), [win32](), [winapi]() ### AI overview ReactOS Newsletter 101 reports project changes from February through July 2021. It highlights improved stability for amd64 builds, completed Structured Exception Handling support, a fix enabling semi-stable virtual-machine booting, planned GCC support in RosBE 2.3, restored i686 APIC support, and ongoing SMP work for multicore systems. ### Source excerpt Hello ReactOS followers! This report covers changes in the project during February-July 2021. And we definitely have some things to highlight! amd64 build is getting more stable Timo Kreuzer (tkreuzer) worked hard on various parts of the kernel and HAL, fixing issues here and there. Structured Exception Handling (SEH) support for the amd64 architecture was finished, various bugs around the kernel are fixed. A major issue with interrupt handling in HAL was also fixed in May, which finally allowed a semi-stable boot in a virtual environment. [Next page]()