mcpField
K
kernel-memory-management

kernel-memory-management

github.com/mohitmishra786Development269308 installsUpdated 18w ago
Repository

About kernel-memory-management

A guide for agents choosing and debugging Linux kernel memory allocators, covering physical pages, SLUB, virtual mappings, DMA, and NUMA. It combines selection guidance with allocation patterns and diagnostic commands.

Key features

  1. 01Allocator selection for DMA and buffers
  2. 02kmalloc, SLUB, vmalloc, and page allocation
  3. 03GFP flags and execution contexts
  4. 04NUMA and memory-zone basics
  5. 05OOM and slab diagnostics

Use cases

  1. 01Choosing allocation for a driver buffer
  2. 02Investigating kernel OOM or slab corruption
  3. 03Checking DMA allocation requirements
  4. 04Inspecting memory and slab statistics

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/mohitmishra786/low-level-dev-skills --skill kernel-memory-management

After installing, type/kernel-memory-managementin Claude Code to confirm it loaded.

Details

From the project README

Kernel Memory Management

Purpose

Guide agents through Linux kernel memory allocation: physical page allocator (buddy), SLUB kmalloc caches, vmalloc virtual mappings, zones (DMA/Normal), and safe allocation patterns in drivers — extending skills/kernel/kernel-internals.

When to Use

  • Choosing kmalloc vs vmalloc vs get_free_pages
  • Debugging kernel OOM or slab corruption
  • DMA buffer allocation requirements
  • Understanding /proc/meminfo and slab stats

Workflow

1. Allocator decision tree
Need physically contiguous memory for DMA?
├── Yes → dma_alloc_coherent() or CMA pool
└── No
    ├── Size ≤ ~128K (arch-dependent) → kmalloc / kzalloc
    ├── Large or non-contiguous OK → vmalloc / vzalloc
    └── Whole pages → alloc_pages() / __get_free_pages()
2. kmalloc and SLUB

SLUB is the default general-purpose kmalloc backend (per kernel memory allocation guide); the legacy SLAB allocator was removed in Linux 6.12.

void *buf = kmalloc(size, GFP_KERNEL);
if (!buf)
    return -ENOMEM;
/* ... */
kfree(buf);
GFP flagContext
GFP_KERNELProcess context, may sleep
GFP_ATOMICIRQ / holding spinlock — smaller pools
GFP_DMADMA zone (legacy 32-bit devices)

Check /proc/slabinfo and cat /sys/kernel/slab/*/object_size.

3. vmalloc
void *v = vmalloc(size);  /* virtually contiguous, physically scattered */
vfree(v);

Use for large buffers where contiguous physical pages are unnecessary. Do not use for DMA without dma_map_*.

4. Buddy page allocator
struct page *page = alloc_pages(GFP_KERNEL, order);
void *addr = page_address(page);
__free_pages(page, order);

order n allocates 2^n pages. Zone selection: ZONE_DMA, ZONE_NORMAL, ZONE_MOVABLE (see include/linux/mmzone.h).

5. NUMA basics

On NUMA systems, kmalloc prefers local node; use alloc_pages_node() or mbind policies for HPC paths — see skills/allocators/numa-programming.

6. Debugging
cat /proc/meminfo
cat /proc/slabinfo
dmesg | grep -i oom

Kernel config: CONFIG_SLUB_DEBUG, CONFIG_KASAN for use-after-free.

7. Agent usage
/kernel-memory-management Choose allocator for 2 MiB driver ring buffer in process context

Common Problems

SymptomCauseFix
kmalloc fails largeExceeds kmalloc limitUse vmalloc or pages
DMA buffer wrongUsed vmalloc for HW DMAdma_alloc_coherent
OOM killerUnbounded cache growthLimit pools; use shrinker
Slab corruptionUse-after-freeKASAN; audit kfree pairing
GFP_KERNEL in IRQSleeping allocator in atomicGFP_ATOMIC

Related Skills

  • skills/kernel/kernel-internals — page cache, OOM killer
  • skills/kernel/device-drivers — dma_alloc_coherent
  • skills/allocators/custom-allocators — userspace allocator design
  • skills/allocators/numa-programming — NUMA-aware allocation
  • skills/runtimes/sanitizers — KASAN overlap

Comments

Markdown supported