@bortzmeyer Excellent. J'ai expérimenté avec cette idée et je me demandais pourquoi elle est si peu utilisée. Je vais juste en conclure qu'elle est moins triviale à mettre en place qu'un cron :P

@bortzmeyer En parlant de déléguer l'obtention de certificats à un service tiers… As-tu un avis sur le fait de confier la terminaison TLS de tous les services du serveur à un démon ?

@lynne In the meantime, if the Phoronix article doesn't say it I will: thanks for all your work on the ffmpeg Vulkan code!

@lynne Nice! Sent an email anyways, hope he'll amend the article.

@lynne Or maybe I should get in touch, just so that he doesn't understand this as a random trying to get attention -_-

@lynne As always, phoronix pretty bad at getting things right... Maybe get in touch with Michael?

So, yeah, *&%*($ to silently ignoring errors, XCB.

Show thread

It turns out sending a closed FD via XCB doesn't fail. The other side (X.Org server) just receives a random FD number, which I think by chance referred to an unrelated FD which was opened.

But why was I sending a closed FD via XCB in the first place? It turns out sending a FD via XCB also closes it. So it worked fine the first time, and caused the bug the second time.

Show thread

This morning I spent a few hours tracking down a bug in wlroots' X11 backend. Importing a DMA-BUF via the X11 DRI3 extension failed, and I couldn't understand why.

Went from X.Org server core to Xwayland to GBM to Gallium to radeonsi to libdrm to the kernel. The FD seemed valid, since stat() didn't return an error, but drmPrimeFDToHandle failed.

Simon Ser boosted

A few ways to make money in FOSS



@wolf480pl Yeah, I assumed you'd use KMS to accelerate for these use-cases, not a separate API

@wolf480pl Yeah, that doesnt surprise me too much. When you have a 3D engine I don't know if there's any point in using a separate thing for 2D. Maybe some specific embedded use-cases could take advantage of it, dunno.

@wolf480pl The KMS API (with its writeback connectors) could be abused for that (but not drawing triangles and stuff). Probably a bad idea.

Last time I heard about 2D accel it was from Paul Kocialkowski. Maybe you can ask him more details, he's way more knowledgeable than me.

@wolf480pl I've heard some embedded platforms have types of 2D accel blocks not exposed via KMS. ATM these aren't exposed at all by the kernel, there's no uAPI.

I'm not sure what the 2D accel in XFree86 looked like. Would need to read the source code.

Related: blog.ffwll.ch/2018/08/no-2d-in

@wolf480pl In gaming scenarios, 2D accel still useful to reduce latency and leave the whole 3D pipeline for the game (see: gamescope). The other big goal is to reduce power consumption by being able to make the 3D engine sleep more often, but that doesn't apply here.

Show more

The social network of the future: No ads, no corporate surveillance, ethical design, and decentralization! Own your data with Mastodon!