Fuck, I’ve removed TPM_RC_SIZE at some point while drilling and hammering… I’ll add it back!

#linux #kernel #tpm #rust

"Who needs an App Store? You've got the Nokia PC Suite software on a CD-ROM! Just plug the proprietary cable into your Symbian S60 device, boot up your Windows PC, and away you go. THIS is the smartphone experience for people who are serious about mobile computing."

#Nokia #Symbian #smartphone #retrophone

The next thing with tpm2_protocol is obviously the figures!

I.e. creating a stress that measures the delta between compile-time estimated size and run-time realized size.

This spun up from my friend Philip Tricca asking about stack usage. I definitely want to know the truth and also catch regressions on this side that might be caused by new commits.

Obviously during construction time this has not been a priority but now it is time to level it up.

One of first commits to wiping away the excess fat is (already in 0.10.0):
https://git.kernel.org/pub/scm/linux/kernel/git/jarkko/tpm2-protocol.git/commit/?id=cd6641bf9e8c8fde8726bece9eb6cdc630d893c2

#linux #kernel #tpm #rust

hot:

https://shop.compiler-explorer.com/en-eur/collections/abi-mugs

As of today can we do this:

1. drivers/char/tpm (C code)
2. drivers/char/tpm/protocol (imported tpm2_protocol)

?

And then build FFI from C to Rust for building commands that we need today etc.

There's one particular challenge where this could help: early boot code for D-RTM (i.e., Trenchboot) as given my crate is just a thing in stack with no deps, it could be linked also to that payload.

#linux #kernel #rust #tpm

should usage normally cause exit code 1 or 0?

right now e.g.,

❯ sudo target/debug/tpm2sh unseal
tpm2sh-unseal
Unseals a keyedhash object

USAGE:
tpm2sh unseal [OPTIONS]
OPTIONS:
--password <PASSWORD> Authorization value
-h, --help Print help information


~/work/github.com/puavo-org/tpm2sh main
❯ echo $status
1

It checks if stdin is open for the sake of pipelien and if not it shows usage.

the very last spam alert:

https://lore.kernel.org/rust-for-linux/aKfaR-h6Itc38qfl@kernel.org/T/#u

moving to on hold as tpm2_protocol is/will be mailing list based project.

tpm2sh has a new github location: https://github.com/puavo-org/tpm2sh

#linux #kernel #tpm #rust

The first independent release of the protocol: https://crates.io/crates/tpm2-protocol

Git: https://git.kernel.org/pub/scm/linux/kernel/git/jarkko/tpm2-protocol.git

Release notes:

tag 0.10.0
Tagger: Jarkko Sakkinen <jarkko.sakkinen@iki.fi>
Date: Fri Aug 22 04:45:40 2025 +0300

Release 0.10.0

- chore: refactor into standalone crate
- refactor(tpm2_protocol): reduce defaults
- refactor(tpm2_protocol): merge TpmuSigScheme and TpmuAsymScheme
- refactor(tpm2_protocol): decouple command building
- fix(tpm2_protocol): trailing data
- refactor(tpm2_protocol): remove MAC definitions
- tests(tpm2_protocol)
- fix(tpm2-protocol): TpmuAttest error code
- tests: migrate dyn trait test to tpm2_protocol
- fix(tpm2_protocol): correct serialization logic
- refactor: InternalError -> Unreachable
- tests(tpm2_protocol): fix compilation errors
- fix(tpm2_protocol): StartAuthSession response
- refactor!(tpm2_protocol): drop tpm_response! and TpmParameters
- fix!(tpm2_protocol): DO NOT export submodules
- refactor(tpm2_protocol): adjust buffer debug output
- fix(tpm2_protocol): TpmRc::base() return code
-----BEGIN PGP SIGNATURE-----

iHUEABYIAB0WIQRE6pSOnaBC00OEHEIaerohdGur0gUCaKfLxAAKCRAaerohdGur
0nY/AP9/4HMLP+wY0h5tQSnbzmIajNzzBAoWIA7nA8dIkcQ8RQEAxhK/MrKKT7iQ
j2rEvvKdgWPdHtPhZWzUahuZnW6LIgA=
=RfaI
-----END PGP SIGNATURE-----

#linux #kernel #tpm #rust

@monsieuricon possible to get a ML?

https://git.kernel.org/pub/scm/linux/kernel/git/jarkko/tpm2-protocol.git

See https://github.com/tpm-rs/tpm-rs/issues/196#issuecomment-3207383832

I migrated dyn trait (or Box<dyn TpmObject>) as part of tpm2_protocol test suite just to demonstrate that on-wire TPM2 protocol can be dynamically detected without any spurious dependencies ;-)

Screencast demonstrates also the time that it takes to run the full kselftest compatible test suite.

#linux #kernel #tpm #rust

My standalone and kselftest exit code compatible test program might be archaic yes but it sure executes fast as hell compared to running cargo test :-)

On first run it also fully compiles the test program togehter with crate's source code linked into it.

#linux #kernel #tpm #rust

I've finally just submitted the RFC series for USB3 support on Apple Silicon machines. RFC because the entire Type-C machinery spans multiple subsystems and has quite a few quirks which makes the entire thing rather annoying to deal with and I'm not 100% convinced about the current approach.

This also includes some initial work for DisplayPort-altmode and USB4/Thunderbolt but both will require much more work after this is upstream, so don't get your hopes up just yet.

https://lore.kernel.org/asahi/20250821-atcphy-6-17-v1-0-172beda182b8@kernel.org/

#asahilinux #linux

I'm dying here. Just had a customer come in, asking where his drink was. We never got any mobile orders for him.

"I asked ChatGPT to order it for me and it said I could pick it up."

#smdh #FuckAI

I learned a new thing in macro_rules! i.e., pattern matching:

tpm_struct! {
    #[derive(Debug, Default, PartialEq, Eq, Clone)]
    kind: Command,
    name: TpmStartAuthSessionCommand,
    cc: TpmCc::StartAuthSession,
    no_sessions: true,
    with_sessions: true,
    handles: {
        pub tpm_key: crate::data::TpmiDhObject,
        pub bind: crate::data::TpmiDhObject,
    },
    parameters: {
        pub nonce_caller: Tpm2bNonce,
        pub encrypted_salt: Tpm2b,
        pub session_type: TpmSe,
        pub symmetric: TpmtSymDefObject,
        pub auth_hash: TpmAlgId,
    }
}

tpm_struct! {
    #[derive(Debug, Default, PartialEq, Eq, Clone)]
    kind: Response,
    name: TpmStartAuthSessionResponse,
    cc: TpmCc::StartAuthSession,
    no_sessions: true,
    with_sessions: true,
    handles: {
        pub session_handle: TpmSession,
    },
    parameters: {
        pub nonce_tpm: Tpm2bNonce,
    }
}

I can do this:

rustc --crate-type lib --crate-name tpm2_protocol tpm2_protocol/src/lib.rs --edition=2021 --emit=mir

Cool. Can I compile a single file from a crate project? This would be great when doing large refactor and not get your terminal DOS'd :-)

It's pain really. ATM, I'm refactoring out a macro called "tpm_response!" (it is replaced with pre-existing "tpm_struct!" that is extended) and I need to do similar changes to few dozen files.

I'd like to:

1. Do the edit to a single file.
2. Try if that file compiles.
3. Move on to next files.

Lack of knowledge with this makes refactors in Rust living hell TBH ...

Some constrains in my project that might help (possibly):

1. no_std
2. no deps
3. no alloc

Another way to express this: what is "gcc -c" of rustc?

#rust

Can you specify calling convention for assembly (with asm, global_asm etc.)? I mean things like stdcall and cdecl in C and GCC...

I could not find anything and it'd be a huge limitation...

#rust #assembly

[video] Blastromen (a.k.a. Sami Koskivaara & Mika Rosenberg) - Live @ Hellsinki Industrial Festival, Tiivistämö, Helsinki, Finland, 8 November 2024.

https://www.youtube.com/watch?v=g10owNrqthA

#Blastromen #electro #electronic #DominanceElectricity #HellsinkiIndustrialFestival #SamiKoskivaara #MikaRosenberg #Tiivistämö #Helsinki #Finland

Real or @NanoRaptor #datArena

Well this an IBM 5120 https://en.wikipedia.org/wiki/IBM_5120 a preliminary system to the IBM PC.

Very cool, I start to participate to tpm-rs meetings :-) My project is not a conflicting entity, it's just the protocol layer isolated and made for import (a design choice driven by kernel ofc).

I was just worried what are the thoughts when I come from the woods and say that here's 7000 lines of source code, please repeal and replace but initial feedback is actually quite positive.

For kernel (and tpm-rs) if/when kernel has TPM2 support on Rust side the model could be a bit similar as it is for ACPICA. You have a fast policy-free upstream that tracks the protocol spec as it develops and import cycle.

Would be great to see tpm-rs story to continue but obviously they make their choices.

IMHO, good design goals for any Rust crate would be that:

1. It should run on your toaster.
2. It should run on #Amiga.

Once I stopped using Cargo and 3rd party crates when making my own crates I've started to enjoy Rust for real :-) IMHO, programming should be fun, not efficient or "productive". If I have to start to be too efficient I'll immediately stop this career.

#rust