Skip to content

Attempt to make kernel builds reproducible - #88

Draft
legoktm wants to merge 4 commits into
mainfrom
reproducibility
Draft

Attempt to make kernel builds reproducible#88
legoktm wants to merge 4 commits into
mainfrom
reproducibility

Conversation

@legoktm

@legoktm legoktm commented Aug 31, 2026

Copy link
Copy Markdown
Member

Refs #3.

This is useless because we use an ephemeral key that we immediately
throw away and don't distribute. On the negative side it makes the build
unreproducible.
Extra hardening that the workstation has that Claude noticed the server
was missing.
Because our kernels are public, using a random seed has minimal value
for a dedicated attacker because they can figure out e.g. randomized
struct layout by using the published headers or decompiling vmlinux. It
could have some value in forcing attackers to make their exploits work
with our layouts.

But since it interferes with reproducibility, use a fixed seed so we
still retain the advantage of a different layout than stock kernels but
builds will not actually be random.
To verify our code that injects a deterministic seed works as expected
and that it doesn't break reproducibility.
@legoktm legoktm moved this to Next sprint candidates in SecureDrop Aug 31, 2026
@nathandyer nathandyer moved this from Next sprint candidates to Backlog in SecureDrop Sep 3, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

Status: Backlog

Development

Successfully merging this pull request may close these issues.

2 participants