Tag: #linux
Writing on GPUs, LLMs, MLOps, Kubernetes — and mindset · 70 posts
The eBPF Verifier Only Tells You Where It Stopped — Measuring the Diagnostic Gap Across 235 Reproduced Rejections
Anyone who has used eBPF has lived this: the verifier rejects your program, and the error message points at a line that looks perfectly innocent. A paper published in July 2026 is the first to put a number on that frustr
2026-07-16 · 18 min read #ebpf#linux#kernel#debugging#developer-experiencesched_ext Sub-Schedulers — the Per-cgroup Scheduler That Arrived Half-Done in Kernel 7.1
Linux 7.1 (released June 14, 2026) landed cgroup sub-scheduler support for schedext. The goal is clear — attach a BPF scheduler anywhere in the cgroup tree, let parent schedulers dynamically allocate CPU to their childre
2026-07-16 · 13 min read #linux#kernel#scheduler#cgroups#bpfbcachefs Drops the Experimental Tag from Erasure Coding — and What It Actually Means
On September 29, 2025, Linus Torvalds erased 117,483 lines of bcachefs code from the kernel in a single commit, and bcachefs became a filesystem shipped as a DKMS module, like ZFS. That was the end of the drama — the rea
2026-07-16 · 17 min read #linux#kernel#filesystem#bcachefs#storageEROFS Native Layers in containerd 2.3 — What You Get in Exchange for Removing Unpack
containerd 2.3, released on April 30, 2026, opened a path to pull an EROFS filesystem image straight from the registry and mount it as-is. Since there's no tar to unpack, the entire unpack step disappears, leaving behind
2026-07-16 · 18 min read #container#containerd#oci#storage#linuxFrom Docker to Podman — the Complete Guide to Switching to a Daemonless Container Engine
Podman differs from Docker at the level of architectural philosophy: daemonless (fork-exec) operation and rootless by default. Internals (conmon, crun), config file locations and meanings, CDI setup for GPUs, the two way
2026-07-08 · 8 min read #docker#podman#containers#devops#linuxReading htop and top — How to Decode Every Number
Open htop or top on Linux and a flood of numbers pours out. The difference between VIRT, RES, and SHR; the process state letters; why load average is not CPU usage; what the meter colors mean; and zombie processes — this
2026-07-05 · 7 min read #linux#performance#htop#sysadminKernel Tuning for Low-Latency Trading Systems — The War Against Microseconds
How to tune the Linux kernel in domains like HFT and market making where microseconds decide profit and loss. Covers CPU isolation recipes, NUMA alignment, the kernel-bypass spectrum, C-state pinning, PTP time synchroniz
2026-06-13 · 17 min read #linux#kernel#low-latency#trading#performanceContainers Are a Lie — The Kernel Truth Behind cgroups and Namespaces
There is no container object in the kernel. A tour of the seven namespaces, hands-on cgroup v2 file manipulation, building a mini container with unshare and pivotroot, overlayfs, capabilities, and seccomp — we dissect ea
2026-06-13 · 17 min read #linux#kernel#container#cgroups#namespaceseBPF Fundamentals — Programs, Maps, and the World of the Verifier
A ground-up tour of how eBPF changed Linux kernel programming. We cover program types and maps, how the verifier works, and build a first eBPF program with libbpf and CO-RE from scratch.
2026-06-13 · 19 min read #ebpf#linux#kernel#libbpf#observabilityThe Evolution of the Linux CPU Scheduler — From CFS to EEVDF
We follow the evolution of the Linux scheduler, from CFS implementing fairness with vruntime and a red-black tree, to EEVDF which became the default in kernel 6.6 to address latency-sensitive workloads. Covers cgroup CPU
2026-06-13 · 17 min read #linux#kernel#scheduler#cfs#eevdfeBPF Observability in Practice — Opening the Black Box with bpftrace
Ten battle-tested bpftrace one-liners and a tour of the BCC toolbox for inspecting live systems without touching application code. Covers latency histogram interpretation, flame graphs, container environments, and three
2026-06-13 · 15 min read #ebpf#bpftrace#observability#linux#profilingA Packet Journey — Tracing the Linux Kernel Networking Stack
We dissect the full path a packet takes from the NIC through DMA, ring buffers, NAPI, and the IP/TCP layers to the application socket. Covers RSS/RPS/RFS multi-core scaling, GRO/GSO/TSO offloads, XDP, socket buffer tunin
2026-06-13 · 19 min read #linux#kernel#networking#tcp#xdpHacking the Linux Memory Hierarchy — swap, zram, and the Wild Idea of Swapping to VRAM
The nbd-vram project, which exposes GPU VRAM as an NBD block device to use as swap, made waves on Hacker News. Starting from there, this post covers Linux memory management fundamentals, swappiness and zram/zswap tuning,
2026-06-12 · 16 min read #linux#memory#swap#zram#kernelMastering systemd Timers — Time to Graduate from cron
Sparked by the Hacker News hit You do not love systemd timers enough, this post walks through the structural limits of cron and everything systemd timers offer. OnCalendar syntax, Persistent catch-up, failure notificatio
2026-06-12 · 15 min read #linux#systemd#cron#scheduling#devopseBPF Is Eating Observability
The idea of safely running sandboxed programs inside the kernel is upending observability. Hook anywhere with kprobes, uprobes, tracepoints, and XDP; the verifier guarantees safety; trace without touching a line of appli
2026-06-11 · 13 min read #ebpf#observability#linux#kernelLinux Distros 2026 Complete Guide — Ubuntu 26.04 / Fedora 43 / NixOS 25.05 / Debian 13 / Arch / Asahi / Bazzite / Bluefin Deep Dive
Ubuntu 26.04 LTS, Fedora 43, NixOS 25.05, Debian 13 Trixie, the Arch / Manjaro / CachyOS / Garuda / EndeavourOS rolling-release camp, Asahi Linux on Apple Silicon, the explosive growth of immutable / atomic distros like
2026-05-16 · 26 min read #english#linux#ubuntu#fedora#nixosThe Linux Desktop in 2026 — COSMIC, GNOME 47, KDE Plasma 6.3, Wayland, Asahi, NixOS, and Immutable Distros, Deep Dive
As of May 2026, the Linux desktop is not a dead market — it is the most actively redesigned operating system in computing. System76 shipped COSMIC 1.0 alpha after two years of Rust-plus-iced rebuilding. GNOME 47 finally
2026-05-15 · 26 min read #linux#desktop#cosmic#gnome#kdeA Modern Understanding of the Operating System — io_uring, cgroups/namespaces, eBPF, NUMA, GPU UVM, EEVDF, Zero-Copy Complete Guide (2025)
iouring as the successor to epoll, the cgroups + namespaces that made Docker, the eBPF that injects code into the kernel safely, the hidden cost NUMA imposes, GPU drivers and UVM, the EEVDF scheduler that landed in Linux
2026-04-15 · 13 min read #operating-systems#linux#io-uring#cgroups#namespacesLinux I/O Evolution Deep Dive — blocking, select, poll, epoll, io_uring (2025)
How do you handle not 10,000 but 1,000,000 concurrent connections on a single server? 30 years of Linux I/O API evolution — from the 1970s blocking read, through the limits of select/poll, the epoll revolution, to iourin
2026-04-15 · 11 min read #linux#io#epoll#io-uring#asynceBPF Deep Dive — The VM, Verifier, XDP, and CO-RE That Made the Linux Kernel Programmable (2025)
Injecting safe code into the kernel — that is the eBPF revolution. This post dissects eBPF from internals to practice: the VM and bytecode, the Verifier that enforces safety, maps (Hash/Array/Ring Buffer/Perf), program t
2026-04-15 · 15 min read #ebpf#linux#kernel#xdp#observability