Posts
4417
Following
315
Followers
471
Linux kernel hacker and maintainer etc.

OpenPGP: 3AB05486C7752FE1

Jarkko Sakkinen

Edited 1 year ago
OK, now I have together ASN.1 parsing and new key subtype "tpm2_key_subtype", which is foundations for TPM2 asymmetric keys.

Only compile-tested ;-) This is now bare minimum to trial in QEMU. Lot's of dance relocating necessary bits of trusted keys code to the TPM driver (only ASN.1 parser at this point).
1
0
0
weekend as a kind of freelancer (at least for some time) is the best time for own patches. on plus, you don't have to care anything but the solving problem itself... paid or unpaid i enjoy this ;-)
1
0
1

SDL3 Adds PipeWire Camera Support

Adding to the growing list of features coming with the SDL3 release for this hardware/software abstraction layer commonly used by cross-platform games and other software is PipeWire camera capturing support...
https://www.phoronix.com/news/SDL3-PipeWire-Camera-Capture

0
3
1

Jarkko Sakkinen

#Amaranth sounds like a name of a black/death metal band from Scandinavia but is actually pretty neat hardware (#FPGA) synthesis framework:

https://amaranth-lang.org/docs/amaranth/latest/intro.html
0
0
0

Jarkko Sakkinen

working on RFC patch for TPM2 asymmetric keys (will use null seed encrypted session) for supporting x509: https://git.kernel.org/pub/scm/linux/kernel/git/jarkko/linux-tpmdd.git/commit/?h=tpm2_key

https://datatracker.ietf.org/doc/draft-woodhouse-cert-best-practice/
1
0
0
@jmorris FPGA is great optimization for latency/frequency (e.g. audio, router and such) rather than throughput, which can be often achieved just by upgrading the SoC (which is usually cheaper).
0
0
1
And it sort of makes sense too. Communicating complex technical topics through Internet is always a non-trivial task, no matter how it is orchestrated.
0
0
0
What is the most difficult task in kernel development? Emails ;-)

That said,it is still the best available in my opinion :-) Email archives are much nicer to backtrack even multiple years behind than these crazy Git issue trackers. I.e. there is a clear timeline in communications.
1
0
2

Jarkko Sakkinen

Edited 1 year ago
@jmorris off-topic but i have this sound card called "RME BabyFace Pro FS". it has pretty cool idea for FPGA's: the mixer is an FPGA. With this topology it can reach 1 ms latency for audio. And i.e. a firmware update contains also potentially update for the FPGA. Pretty clever hybrid approach IMHO.

Before I got this sound card I was not really sure where application wise it would make sense to have e.g. a SoC combined with FPGA but this product made me understand it a bit better.

I also own one small lattice FPGA. As soon as I get an idea how to use it e.g. with RPi for similar hybrid thing I'll definitely will PoC something. Lattice's have the benefit of having open source FPGA stack and since I also have a SoC FPGA does not have to have too many logic ports.
2
1
1

Jarkko Sakkinen

Edited 1 year ago
@jmorris off-topic but i have this sound card called "RME BabyFace Pro FS". it has pretty cool idea for FPGA's: the mixer is an FPGA. With this topology it can reach 1 ms latency for audio. And i.e. a firmware update contains also potentially update for the FPGA. Pretty clever hybrid approach IMHO.

Before I got this sound card I was not really sure where application wise it would make sense to have e.g. a SoC combined with FPGA but this product made me understand it a bit better.

I also own one small lattice FPGA. As soon as I get an idea how to use it e.g. with RPi for similar hybrid thing I'll definitely will PoC something. Lattice's have the benefit of having open source FPGA stack and since I also have a SoC FPGA does not have to have too many logic ports.
2
1
1

Jarkko Sakkinen

Edited 1 year ago
probably will get some feedback from linus but i'm in middle of changing my process (by request), so i'll just then apply learnings to 6.11 🤷
1
0
1

Jarkko Sakkinen

Splitting my pull request. For v6.10 not yet fully finished/polished but the idea is to have simple script to create the PR from signed git tag and then have a few #aerc templates for each subsystem:

https://lore.kernel.org/linux-integrity/D15DSV117DQZ.3GJOTXCTGZHE9@kernel.org/T/#u

Two first I did for keyring and trusted key did not yet use this scheme but learned along the way that I need to scale a bit. Slowly figuring out how to do this properly.

https://man.archlinux.org/man/aerc-templates.7.en
1
0
2
@countcol went too young only 52 years...
0
0
1
@countcol she died already back in 2004
1
0
1

Jarkko Sakkinen

this looks cool: https://aya-rs.dev/
0
0
0
@jstultz I'll check all tracing tools and debuggers as they come by :-) QA is fun!
0
0
0

(speaking for myself here) One of the cooler things about Google is getting just the slightest glimpse of the power of some of the tooling your wizard coworkers use.

https://perfetto.dev/ is one of those very cool tools. It's like kernel shark, but has really powerful SQL capabilities. It's well configured for use with Android and CrOS, but with classic Linux environments it can be a little daunting. So here are my notes on using it for upstream kernel development: https://gist.github.com/johnstultz-work/0ec4974e0929c4707bfd89c876ae4735

2
10
2
Show older