Deep DivesDeep Dives
From start_kernel() to Userspace: A Line-by-Line Walkthrough
Walk the path from start_kernel to userspace on arm64: rest_init, kernel_init, the initcall levels, and the fixed order the kernel searches for…
RB Oct 3 · 20 min
Linux kernel & Device drivers starts on 24th Oct 2026 enrollingCorporate on-site training - Submit proposal Pick your modulesSharpen your kernel skills: deep dives, drivers, Yocto, CVEs, careers — updated daily. Read the blog →Embedded Linux fast track starts 30th sept 2026 enrollingEmbedded Linux Mastery track starts 30th sept 2026 enrollingLinux systems engineering starts 30th sept 2026 enrolling
~/ blog
Deep Dives
Articles in this topic.
Deep Dives
Deep DivesDeep Dives
Linux nvmem Cells: How Per-Unit Data Reaches Your Driver
How Linux nvmem cells carry a MAC address or calibration data out of an EEPROM or fuse array and into a driver…
RB Sep 27 · 19 min
Deep DivesDeep Dives
What PREEMPT_RT Changes: Sleeping Spinlocks (Part 1)
Under PREEMPT_RT spinlock_t becomes an rtmutex. Sleeping spinlocks explain why _irq does nothing and why unlock order matters.
RB Sep 23 · 13 min
Deep DivesDeep Dives
Kernel Memory Attribution Is a Product Decision, Not a Debugging Trick
Memory allocation profiling attributes kernel RAM to a source line. What it costs to ship, how to leave it dormant, and why…
RB Sep 15 · 14 min
Deep DivesDeep Dives
Embedded Linux Observability: What to Instrument on a Device You Cannot Reach
eBPF observability for embedded Linux in three layers: in-kernel instrumentation, structured events, and the power regressions logs never show.
RB Sep 8 · 13 min
Deep DivesDeep Dives
UBI on Raw NAND: The Layer That Hides Bad Blocks and Wear
How UBI works on raw NAND: LEB to PEB mapping, the 20 per 1024 bad block reserve, wear levelling, attach time and…
RB Aug 30 · 14 min
Deep DivesDeep Dives
Linux genpd Power Domains: Why a Rail Won’t Power Off?
How genpd power domains decide to cut power in the Linux kernel: the checks in genpd_power_off(), the stay_on rule, and reading pm_genpd_summary.
RB Aug 25 · 18 min
Deep DivesDeep Dives
The Linux DMA API: Coherent and Streaming Mappings Explained
A working guide to the Linux DMA API: coherent vs streaming mappings, cache sync rules, the dma-coherent property, and debugging with CONFIG_DMA_API_DEBUG.
RB Aug 21 · 15 min
Deep DivesDeep Dives
Writing a Minimal sched_ext Scheduler
How to write a minimal sched_ext scheduler: the four callbacks, the dispatch queue model, and how to build, load and verify it…
RB Aug 17 · 11 min
Deep DivesDeep Dives
Embedded Linux Boot Time — Part 2: Measuring the Kernel and User Space
A long initcall is often a waiting initcall, and the log gives you no way to tell. How to read initcall_debug without…
RB Aug 13 · 20 min
Deep DivesDeep Dives
Embedded Linux Boot Time — Part 1: The Phase You Are Not Measuring
Part 1 of two. Boot time spans four phases with four independent clocks, and dmesg only shows the third. How to put…
RB Aug 7 · 19 min
Deep DivesDeep Dives
dm-verity: How a Read-Only Rootfs Stays Verified
dm-verity gives an embedded device a tamper-evident read-only rootfs. Learn how per-block hash-tree checking works and why the root hash must be…
RB Aug 3 · 11 min
Deep DivesDeep Dives
Devres: How Managed Device Resources Work in the Linux Kernel
How managed device resources (devres) work in the Linux kernel: the devres list, devm_* APIs, release order on unbind, pitfalls, and CONFIG_DEBUG_DEVRES.
RB Aug 1 · 10 min
Deep DivesDeep Dives
System Device Trees and Yocto for Heterogeneous SoCs
A system device tree plus Yocto multiconfig builds Linux, bare-metal, and firmware for heterogeneous SoCs like ZynqMP from one hardware description.
RB Jul 27 · 9 min
Deep DivesDeep Dives
The Linux Generic IRQ Subsystem: From Device Tree to Handler
A walkthrough of the Linux generic IRQ subsystem: why it exists, irq_domain hwirq-to-virq mapping, flow handlers, irq_chip, and Device Tree to handler.
RB Jul 26 · 18 min
The fortnightly digest
One email every two weeks — the best of our kernel & embedded writing, plus a hand-picked link or two. No spam.
// 1,200+ embedded engineers already read it
We email occasionally and never share your address.