The Linux kernel's CPU scheduler has been overhauled for the first time in two decades. Episode 135 explains the Earliest Eligible Virtual Deadline First (EEVDF) algorithm that's replacing the Completely Fair Scheduler (CFS). Lucas and Luna break down how EEVDF balances latency for interactive tasks with throughput for compute-heavy workloads, and what it means for desktop responsiveness, server performance, and gaming under Proton. They discuss real-world trade-offs, including the scheduler's complexity and adoption pace across major distributions like Fedora and Ubuntu. No prior scheduler knowledge required—just a curiosity about how Linux decides which process gets the CPU next.

#LinuxKernel #EEVDF #CFS #CPUScheduling #OpenSource #LinuxPerformance #Latency #Throughput #FedoraLinux #Ubuntu #Proton #GamingOnLinux #KernelDevelopment #Server #Desktop #Technology #FexingoBusiness #BusinessPodcast

Keep every episode free: buymeacoffee.com/fexingo

Episode 134 dives into Unified Kernel Images (UKI), a new boot method that combines the Linux kernel, initramfs, and boot command line into a single signed EFI executable. Hosts Lucas and Luna explore how UKI streamlines Secure Boot by requiring only one signature instead of multiple for kernel and bootloader. They discuss real-world adoption: Fedora 40 ships UKI by default, and Ubuntu is testing it. The episode explains the technical mechanism—how systemd-stub converts a UKI into an EFI application, enabling measured boot with TPM2 and reducing attack surface. Lucas contrasts UKI with traditional GRUB + signed kernel setups, noting that UKI eliminates the need for a separate bootloader in many cases. Luna raises concerns about backward compatibility and fringe hardware. The conversation concludes with the trend toward simpler, more secure boot chains, and whether UKI could replace GRUB entirely.

#Linux #UnifiedKernelImage #SecureBoot #UEFI #Systemd #Kernel #Bootloader #EFI #TPM2 #Fedora #Initramfs #MeasuredBoot #LinuxBoot #Technology #FexingoBusiness #BusinessPodcast #OpenSource #SystemdBoot

Keep every episode free: buymeacoffee.com/fexingo

Memory pressure can cripple a Linux system once swap hits disk. This episode dives into zram, the kernel module that compresses swap pages in RAM instead of writing to slow storage. We explain how zram works under the hood—using algorithms like lz4 and zstd to achieve 2:1 compression ratios—and why it's a game-changer for devices with limited memory, from Raspberry Pis to cloud VMs and Android phones. We also compare zram with zswap, another in-kernel compression cache, and discuss trade-offs like CPU overhead and memory latency. If you manage Linux servers, containers, or edge devices, understanding zram can save you from swapping nightmares. Plus, we touch on community support that keeps shows like this ad-free. Tune in to learn how to enable zram and when to use it for peak performance.

#Linux #zram #Swap #MemoryManagement #Kernel #OpenSource #SystemAdministration #DevOps #Performance #Containers #IoT #RaspberryPi #CloudComputing #Technology #FexingoBusiness #BusinessPodcast #TechPodcast #TheLinuxPodcast

Keep every episode free: buymeacoffee.com/fexingo

When a critical kernel vulnerability hits, rebooting hundreds of servers causes massive downtime. This episode explores how Linux live patching—via kpatch, livepatch, and the mainline infrastructure—lets administrators apply security fixes in seconds without a reboot. Lucas breaks down how ftrace and stop_machine work under the hood, why some patches still require a reboot, and the trade-off between stability and rapid response. Luna asks about real-world adoption: which distros support it out of the box and what happens when a patch fails mid-flight. The conversation also touches on the role of live patching in zero-downtime environments like financial trading and cloud infrastructure. Plus, a listener-supported moment that keeps technical talk ad-free.

#Linux #Kernel #LivePatching #kpatch #livepatch #ftrace #stopMachine #Security #Downtime #RedHat #SUSE #Ubuntu #Canonical #Technology #Sysadmin #FexingoBusiness #BusinessPodcast #TechPodcast

Keep every episode free: buymeacoffee.com/fexingo

Lucas and Luna explore the ongoing integration of the Rust programming language into the Linux kernel. They discuss the motivations behind this shift — primarily memory safety — and highlight concrete milestones like over 30 Rust-based kernel drivers for GPU, networking, and file systems as of mid-2026. They examine the challenges of co-existing with C, the role of the Rust for Linux project led by Miguel Ojeda, and the cautious roadmap toward broader adoption. A candid look at an evolution that could make the world's most critical operating system kernel more secure without sacrificing performance.

#Rust #LinuxKernel #MemorySafety #OpenSource #KernelDevelopment #DriverDevelopment #MiguelOjeda #RustForLinux #CProgramming #SystemProgramming #Technology #FexingoBusiness #BusinessPodcast #LinuxPodcast #KernelTLS #AndroidKernel #Google #Microsoft

Keep every episode free: buymeacoffee.com/fexingo

Episode 130 of The Linux Podcast dives into the eXpress Data Path (XDP), a technology that accelerates network packet processing by attaching eBPF programs directly to network drivers. Lucas explains how XDP bypasses the kernel's networking stack to achieve millions of packets per second with minimal latency, making it essential for DDoS mitigation, edge computing, and high-frequency trading. Luna asks about real-world use cases and the challenges of writing safe, efficient XDP programs. The episode also touches on why container orchestrators and cloud providers are adopting XDP for load balancing and security filtering. A must-listen for any Linux professional working with high-throughput networking or performance tuning.

#XDP #eBPF #LinuxNetworking #PacketProcessing #KernelBypass #HighPerformance #NetworkAcceleration #DDoSMitigation #EdgeComputing #CloudNative #LinuxKernel #BPF #Technology #LinuxPodcast #FexingoBusiness #BusinessPodcast #OpenSource #SystemsProgramming

Keep every episode free: buymeacoffee.com/fexingo

Episode 129 of The Linux Podcast drills into Linux's CPU frequency scaling infrastructure — specifically the intel_pstate and amd_pstate drivers. Lucas and Luna walk through how the kernel decides when to boost a core and when to throttle it, using real-world examples like a web server under sudden load versus a database doing steady batch processing. They unpack the difference between passive, active, and guided autonomous modes, and why choosing the right governor — performance, powersave, or schedutil — can save a data center thousands in electricity without cratering latency. The episode closes with a concrete tip: checking your current scaling driver and governor with a simple cat command, and adjusting it via sysfs or tuned-adm for workload-specific tuning. A focused 10-minute look at the mechanics behind modern Linux power management, from the old acpi-cpufreq to today's hardware-driven scaling.

#Linux #CPUScaling #Pstate #intel_pstate #amd_pstate #Governor #Schedutil #PowerManagement #PerformanceTuning #Sysfs #Tuned #ServerOptimization #DataCenter #EnergyEfficiency #OpenSource #Kernel #Technology #FexingoBusiness

Keep every episode free: buymeacoffee.com/fexingo

Lucas and Luna dive into the Linux cpuset cgroup controller, a mechanism that allows system administrators to pin specific processes to specific CPU cores and memory nodes. They explain how this differs from general CPU affinity settings, walk through the sysfs interface and hierarchy rules, and discuss real-world use cases such as isolating real-time workloads on a telecom edge server and optimizing NUMA-aware database performance. The episode explains how cpuset interacts with cgroups v2, the implications for hotplug CPUs, and the trade-offs between strict partitioning and utilization. By the end, listeners understand when and why to pin workloads to cores, and how to configure cpuset on a modern Linux system.

#Linux #Cgroups #Cpuset #CPUAffinity #NUMA #RealTimeLinux #EdgeComputing #ServerOptimization #Kernel #Sysfs #PerformanceTuning #WorkloadIsolation #CgroupsV2 #LinuxAdmin #DataCenter #Telecom #Database #FexingoBusiness

Keep every episode free: buymeacoffee.com/fexingo

In episode 127 of The Linux Podcast, Lucas and Luna explore BPF iterators — a relatively recent addition to the eBPF ecosystem that lets you traverse kernel data structures efficiently without writing manual loops in C. They walk through a concrete example: using a BPF iterator to list all open TCP sockets with minimal overhead, comparing it to older methods like /proc and netlink. The hosts discuss how iterators reduce boilerplate, improve safety, and enable new observability use cases. They also cover the runtime cost difference and why projects like Cilium and Falco are adopting iterators. A focused, code-level conversation for anyone building or debugging Linux systems.

#Linux #eBPF #BPFIterators #Kernel #Observability #SystemsProgramming #Networking #TCP #Cilium #Falco #OpenSource #Technology #Performance #KernelDevelopment #LinuxKernel #FexingoBusiness #BusinessPodcast #Podcast

Keep every episode free: buymeacoffee.com/fexingo

In this episode, Lucas and Luna explore lockless data structures in the Linux kernel — specifically read-copy-update (RCU) and lock-free queues — and how they eliminate contention in multi-core systems. They walk through a concrete example: how a major CDN provider replaced a spinlock-protected hash table with an RCU-based design, cutting p99 latency from 12 milliseconds to under 200 microseconds. The episode also covers when NOT to use lockless techniques, the memory-ordering pitfalls that trip up developers, and why kernel maintainers are increasingly adopting these patterns for io_uring and networking stacks. A practical deep dive for systems programmers and performance engineers.

#Linux #Kernel #LocklessDataStructures #RCU #LockFreeQueues #Performance #MultiCore #Concurrency #CDN #MemoryOrdering #io_uring #NetworkingStack #SystemsProgramming #OpenSource #Technology #FexingoBusiness #BusinessPodcast #LinuxPodcast

Keep every episode free: buymeacoffee.com/fexingo