Login
You're viewing the functional.cafe public feed.

Replies

  • Nov 14, 2026, 11:11 PM

    @TTimo congrats on catching up with flatpak from 10 years ago by copying half of it. There could have been an easier way...

    💬 0🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • Nov 15, 2026, 1:14 AM

    @TTimo would have taken a day to prototype the libcapsule approach in flatpak (hey look, I even wrote about it blog.sebastianwick.net/posts/f). Instead we now have another copy of a bunch of flatpak internals. I know that because I work with Simon, and I wrote a lot of code that is or will be in pressure-vessel. It's just sad to see valve making bad decisions and not talking with the people who actually work in that space...

    💬 0🔄 0⭐ 0
  • Nov 15, 2026, 5:18 AM

    @swick @TTimo I'm not involved in this work but I am in Mesa, and your solution of using virgl is OK for non-demanding apps, but adding another process in the command submission path is bound to lower performance and, worst of all, increase jitter : the very reason why modern CPUs come with ginormous caches.

    On top of this, Virgl is barely maintained and has been a thorn in our side for a while now.

    Making the entire flatpak ecosystem depend on it is bound to piss off a lot of people, basically everyone but flatpak devs... even if I know that this is architecturally the right solution until gpu drivers become more embeddable.

    💬 2🔄 0⭐ 1
  • Nov 15, 2026, 1:41 PM

    @mupuf @swick @TTimo it's been a while since i last used virglrenderer... does flatpak use virgl or venus? the article mixes both... is only virgl unmaintained, or also venus? regarding the extra latency/jitter inherent to virtualization, I'd hope this isn't too impactful since it's used in production by asahi on games (although I can't remember if they use virglrenderer, gfxstream or something else) (edit: virglrenderer with drm forwarding)

    💬 0🔄 0⭐ 0
  • Nov 15, 2026, 2:40 PM

    @mupuf @TTimo Virgl specifically was an exploration on how to provide a fallback when the drivers from the runtime become too old. We could also add libcapsule support, but like I explained in the blog post (and is very conveniently being ignored by Valve) this creates a dependency on the host, specifically on i386. The only reason why Valve gets away with it is because no one wants to drop support for Steam which means no one drops i386 on the host. Except immutable distros where flatpak then provides i386 but there the drivers do come from the Flatpak runtime.

    💬 0🔄 0⭐ 0
  • Nov 15, 2026, 2:46 PM

    @mupuf @TTimo also, what is the maintained solution then? The native context stuff requires matching code on both sides which completely defeats the purpose here.

    💬 0🔄 0⭐ 0
  • Nov 16, 2026, 4:05 AM

    @swick @TTimo The more modern solutions are gfxstream or SRIOV, as far as I know.

    The former shares the same performance impact as Virgl due to the necessary context switch in the command submission path. The latter requires modern hardware and virtual machines AFAIK.

    The big issue with Virgl is the lack of users and the reliance on TGSI, which we definitely are trying hard to get rid of, as it has been superseded by NIR. Gfxstream helps with that part, but still lack wide adoption...

    💬 0🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • 💬 1🔄 0⭐ 0
  • Nov 15, 2026, 2:18 PM

    @siosm we'll need to keep 32 bit graphics drivers at least for that high performance, fully native path. Of course for a lot of older, less demanding titles bits of virtualization and emulation may work just fine.

    💬 2🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • Nov 15, 2026, 2:43 PM

    @TTimo @siosm That high performance fully native path could simply come from the runtime. You only need a fallback path for when you stop updating drivers for a runtime.

    💬 0🔄 0⭐ 0
  • 💬 1🔄 0⭐ 0
  • Nov 16, 2026, 10:43 AM

    @mupuf @TTimo @siosm you are aware that a huge chunk of people use that flatpak version of steam, yes? There are escape hatches for all the things you mentioned. We tried to get valve to collaborate on things like input, but instead they essentially blocked any kind of improvement in that area.

    💬 1🔄 0⭐ 1
  • Nov 16, 2026, 10:45 AM

    @mupuf @TTimo @siosm and we got bullied into supporting DRM leasing via Wayland by valve again, so now any app can DoS the compositor trivially because there is no access control. All because they refuse to use dbus.

    💬 1🔄 0⭐ 1
  • Nov 16, 2026, 12:37 PM

    @swick @mupuf @TTimo @siosm access control is up to the compositor to handle (and is needed for other WSI stuff anyways), it just so happens they've decided not to implement it

    💬 1🔄 0⭐ 0
  • Nov 16, 2026, 12:43 PM

    @emersion @mupuf @TTimo @siosm nothing else that we support in wayland needs access control except drm leasing. it's absolutely dishonest to pretend that one can do access control with wayland. the only thing that practically works is statically pre-define access to an entire protocol.

    💬 1🔄 0⭐ 0
  • Nov 16, 2026, 1:16 PM

    @swick @mupuf @TTimo @siosm keyboard-shortcuts-inhibit, idle-inhibit, pointer-constraints also benefit from access controls. all of the protocols support denying access on a case-by-case basis.

    💬 1🔄 0⭐ 0
  • 💬 1🔄 0⭐ 0
  • Nov 16, 2026, 1:24 PM

    @swick @mupuf @TTimo @siosm Mutter has its own permission dialog for keyboard-shortcuts-inhibit (src/wayland/meta-wayland-inhibit-shortcuts-dialog.c). Sway denies keyboard-shortcuts-inhibit for sandboxed apps.

    idle-inhibit is supported by Mutter (src/wayland/meta-wayland-idle-inhibit.c).

    💬 1🔄 0⭐ 0
  • Nov 16, 2026, 1:34 PM

    @emersion @mupuf @TTimo @siosm ah yes, the very sophisticated model of popping up a yes/no dialog every time someone tries to use it and is completely disconnected from all the other permissions. and static configuration. we're truly living in the 1900s.

    (then it was one of the other inhibit related protocols; all of them are a mistake anyway...)

    💬 1🔄 0⭐ 0
  • Nov 16, 2026, 1:36 PM

    @swick @mupuf @TTimo @siosm I'm not saying Mutter's implementation of keyboard-shortcuts-inhibit is good. I'm saying it should be improved by tying it to Flatpak's permission system, and generalized for other protocols that need permission gating.

    💬 0🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • Nov 16, 2026, 2:35 PM

    @swick @TTimo @siosm If I understand you correctly, you are saying that you've tried to get Valve to modify Steam so that it can run better in a flatpak? And they were not receptive to your proposals, which made the flatpak version not as good as the native one? Or what else do you mean by "they essentially blocked any kind of improvement in that area".

    From what I heard, the Steam developers just don't want to support being run in flatpak... so they don't do what you want from them.

    💬 0🔄 0⭐ 0