kmalloc.org/content/en/why-kmalloc.md
squid fe64a75c77 feat: add the Why kmalloc page, linked from the home page
Explains the name kmalloc() from the Linux kernel: no profiling, no
silent fallback, exactly what you asked for or a plain failure, plus
the honest nerd reason for picking it.
2026-11-15 10:55:18 +01:00

13 lines
1.3 KiB
Markdown

Title: kmalloc.org: Why kmalloc
ID: why-kmalloc
Description: Why this project is called kmalloc, after a function in the Linux kernel.
Layout: bare
-
# Why kmalloc
In the Linux kernel, `kmalloc()` is the function that hands a piece of memory to whatever asks for it: a driver, a filesystem, a network buffer. It does not profile the caller, does not resell the allocation to someone else, does not fail silently and hand you something smaller instead. It either gives you exactly what you asked for, or it tells you plainly that it could not.
That is the model. Infrastructure that works like a kernel primitive: give people what they ask for, honestly, with no conditions attached that were never part of the request. No engagement to optimize, no data to monetize behind your back, no "free" that is paid for with you.
We are also, simply, nerds. That is the other half of the reason for the name: not a word a branding agency would ever choose, but a function signature straight out of the kernel source, the kind of thing you only name something after if you have actually read it. The name is a kind of password. Anyone who has read kernel source recognizes it on sight; anyone who hasn't can look it up in five minutes. It does not promise to be mysterious. It promises to be exact.