Just random note (and I think I stick to my mac for the time being e.g. because RME soundcards are proprietary as hell unfortunately).

One could make these quite easily a wine focused distribution for audio with wineasio for wiring into pipewire.

I.e. not Bitwig or other (available for) Linux DAW and yabridge but full-on everything inside wine. E.g. FL Studio performs better wine and is more stable than in Windows. For audio, wine is better Windows than current Windows. Even things like iLok (ugh) work these days without glitches (you need gnutls 32-bit for this, I found it by accident).

It has many other benefits:

1. One can snapshot the production environment because Windows is just a folder. This is useful for e.g. testing new plugins and making perfect clean ups after.
2. One can have multiple environments.

Making this an app ("manager' alike or something, this what Linux Steam is really) or a spin of distribution would be interesting idea. Pipewire was from get go a game changer in audio, it is actually just as a stack better than CoreAudio, but now that it has stabilized it is diamond.

Also, all RT just went to mainline, not sure if there is something that could improve from plain PREEMPT. And the biggest performance glitch was WaitForMultipleObjects (Win32 API), which was recently (either 6.10 or 6.11) recently fixed by having a driver and device just for this call, which bypasses nasty wineserver completely.

Personally I stick mac in audio (because I have RME BabyFace Pro FS) but Wine (with WineASIO) would actually be my option number 2, not because it is a "Linux way" but more like it has better user experience than Windows.

Just writing this down so that won't be forgotten :-)

#linux #audio #wine #note #musicproduction

Ooo, it's Mike Blumenkrantz's new post "My Wayland Your Wayland Our Wayland" https://www.supergoodcode.com/My-Wayland-Your-Wayland-Our-Wayland/ ✨

https://www.systeminit.com/

I want but does Telegram support SMS only binding? I can scrape Signal if it comes down to that ;-) Rad as hell.

https://www.hmd.com/en_int/hmd-barbie-phone?sku=1GF030APC1L01

@duanin2 I acked the patch, since it uses verify_pkcs7_signature() correctly :-)

@kernellogger Well it worked, thank you! Saved me few hours of time ;-)

@kernellogger So once I get that, I think the vanilla kernel copr will be very much useful.

@kernellogger Actually I did a typo, I was supposed to do system upgrade from 40 to 41 but typed by mistake 42 :-)

Oh well, I guess I will make fresh 41 beta install https://fedoraproject.org/workstation/download?beta

@kernellogger good to know thanks!

@duanin2 unsurprisingly yet another merkle tree ;-) read-only except system update so make sense in that context...

#GNOME 47 https://release.gnome.org/47/

I was asked to ack/nack a patch for it and I've just read LWN article about it in 2011. Netflix contributed it but AFAIK as of today Netflix uses FreeBSD not Linux.

Does anyone use dm-verity for anything? Looks almost like abaddonware.

#linux #kernel

This makes #Julia rad. You have like Jupyter notebook in a terminal! And with not much configuration (just add Plots and SixelTerm packages to the environment).

I don't care about learning a new language but it's all this other stuff in the environment that makes it so much more effective that it is worth it (and the language seems to be as trivial as Python anyway).

Game changer... ... ... in math ;-)

Deep Dive into RCU Race Condition: Analysis of TCP-AO UAF (CVE-2024–27394)

V4bel published an analysis of a race condition vulnerability in the TCP-AO subsystem caused by incorrect usage of the kernel RCU mechanism.

The researcher managed to trigger it reliably using the ExpRace technique.

Article: https://blog.theori.io/deep-dive-into-rcu-race-condition-analysis-of-tcp-ao-uaf-cve-2024-27394-f40508b84c42
ExpRace: https://www.usenix.org/conference/usenixsecurity21/presentation/lee-yoochan

Time for Paul's quiz session! Have been waiting for a new RCU article only because I love those quizzes even tho I'm not always that great solving them ;-)

Also the book "Is Parallel Programming Hard, And, If So, What Can You Do About It?" is worth reading.

https://lwn.net/Articles/988638/

Makes me wonder why don't netatalk just plain import those legacy ciphers? Their security has been compromised already by time and age, so it does not really matter. Nobody's after your Apple 2 LAN setup 🤶

In common case you never should reimplement ciphers given FIPS 140 crap etc.

https://lwn.net/Articles/989687/

#netatalk #wolfssl #openssl #fedora

When you think about it there are no hardware companies anymore in IT. There's just geeks doing weird stuff with their laptops and sending their designs to manufacturing plants :-)

A company like Nvidia for instance is a design company, IP holder and brand agency with a research department. [no means to slander Nvidia per se, sorry]

It's kind of interesting equation given that the state of the art manufacturing is located in places, which would not require a lot to turn into warzones.

I just find this fascinating...

@wamserma Still, if TSMC and Samsung facilities went down for some reason, it would be a game over for all consumer electronics. I would not actually mind computers getting slower and moving backwards to 16/22 nm if that ever happened :-)