Conversation

It's kind of interesting how many general Linux bugs and problems we're running into and fixing as Asahi. Nothing huge, but there's the whole >4K page support issue in random software (which we're pushing on purpose), BTI issues in mesa, lots of random kernel bugs...

768

Hector Martin

@marcan42

Aug 1

A lot of this is really just "real people are now using Linux on a real, modern ARM64 platform". Up until now there just hasn't been anything *modern* running real distros and a near-upstream kernel. Apple machines are ARMv8.5-A, everything else desktop is stuck on <=ARMv8.2-A.

282

Daniel Micay

@DanielMicay

Replying to

@marcan42

Non-Apple mobile devices were only using ARMv8.2 until this year. Snapdragon is on ARMv9 now but most vendors aren't interested in MTE, PAC, BTI, etc. AOSP supports them but they'll end up launching with Pixels and Pixels launch in October so they haven't been moved to ARMv9 yet.

2:38 AM · Aug 2, 2022Twitter Web App

Replying to

and

And, sadly, as an OS vendor with a minuscule budget, I can't buy anything (!Apple and > ARMv8.0) for sub-$5,000USD. And even now, Cavium ThunderX systems have completely disappeared on auction sites. Last one I purchased was $1500USD. But, years later, I'm still hoping for > 8.1

Replying to

and

In order to truly support arm64, I need my own fab. Situation sucks worse than my moldy shopvac.

Replying to

and

Android works with upstream kernels including mainline but it can't be shipped in production because it's missing CTS mandated security features such as type-based CFI on x86_64 (not upstream) / arm64 (incomplete upstream) and also some minor things like perf_event_paranoid=3.

Replying to

and

The non-hardening stuff was either landed upstream or replaced with BPF such as uid (per-profile-per-app) based network statistics, gid-based socket restrictions for the INTERNET permission (no longer gid-based though) and a bunch of other stuff including most use of netfilter.

Show replies