ChanServ changed the topic of #wayland to: https://wayland.freedesktop.org | Discussion about the Wayland protocol and its implementations, plus libinput | register your nick to speak
immibis has quit [Remote host closed the connection]
immibis has joined #wayland
duxsco has quit [Quit: duxsco]
co1umbarius has joined #wayland
columbarius has quit [Ping timeout: 480 seconds]
<DemiMarieObenour[m]>
What is the problem here exactly?
<DemiMarieObenour[m]>
Also, as someone who only needs to deal with software rendering, I guess I have it easy 🤣
<DemiMarieObenour[m]>
I really hope that stuff like AVX-512 and AMX will take off in the CPU world, so that LLVMpipe can be a bit less miserable. Any chance of that happening?
<bl4ckb0ne>
isnt there major issues with avx512 on the latest intel cpus?
<DemiMarieObenour[m]>
yes
<DemiMarieObenour[m]>
from a security perspective, GPU acceleration is a nightmare, and so every high-assurance solution I am aware of falls back to software rendering despite its poor performance
<DemiMarieObenour[m]>
there is work being done to change that, but it is a giant mess
<DemiMarieObenour[m]>
the only reasonably secure solutions I am aware of right now for GPU-accelerated graphics are (1) PCIe pass-through of an integrated GPU (but that only works if you have a separate display for management) and (2) AMD MxGPU + VMware Virtual Shared Pass-Through Graphics (but that is incredibly expensive)
<DemiMarieObenour[m]>
And since neither of those are really practical for desktop use, the GPU winds up being used for window compositing only.
<i509VCB>
Virtio gpu passthrough isn't what you are looking for either I assume?
<DemiMarieObenour[m]>
The problem is that virtio-gpu winds up having way too much host-side attack surface. You wind up doing proxying at the OpenGL or Vulkan API level, and that isn’t an attack surface solutions like Qubes OS are comfortable exposing. It’s a defensible security boundary, but only in the sense that browsers or WebGL are defensible: you wind up playing vulnerability wack-a-mole.
<DemiMarieObenour[m]>
My understanding is that AMD MxGPU basically partitions the GPU into many little GPUs, and then assigns an entire mini-GPU to each guest. That keeps the host software’s attack surface to a minimum.
<DemiMarieObenour[m]>
The problem is that it is only available via VMware, public clouds, or ancient, end-of-life hardware.
<DemiMarieObenour[m]>
A good way to tell if a solution is truely secure against hostile workloads is to see if public cloud providers are willing to trust it to isolate workloads belonging to different customers. MxGPU, Xen, KVM (usually with a non-QEMU VMM), and Hyper-V are. I am not aware of any cloud providers offering virtio GPUs.
ts has joined #wayland
nerdopolis has quit [Ping timeout: 480 seconds]
codedchaos has joined #wayland
floof58_ has joined #wayland
floof58 is now known as Guest1888
floof58_ is now known as floof58
Guest1888 has quit [Ping timeout: 480 seconds]
Company has quit [Quit: Leaving]
fmuellner has quit [Ping timeout: 480 seconds]
neonking has quit [Ping timeout: 480 seconds]
shankaru has joined #wayland
zebrag has quit [Quit: Konversation terminated!]
rektide has quit [Quit: Lost terminal]
rektide has joined #wayland
naveenk2 has joined #wayland
Administrator has joined #wayland
Administrator has quit [Remote host closed the connection]
rgallaispou1 has joined #wayland
rgallaispou has quit [Ping timeout: 480 seconds]
rgallaispou1 has quit [Read error: Connection reset by peer]
rgallaispou has joined #wayland
shankaru has quit [Quit: Leaving.]
shankaru has joined #wayland
pnowack has joined #wayland
pnowack has quit [Remote host closed the connection]
vyivel has quit [Remote host closed the connection]
bl4ckb0ne has quit [Remote host closed the connection]
emersion has quit [Remote host closed the connection]
bl4ckb0ne has joined #wayland
emersion has joined #wayland
jmdaemon has quit [Ping timeout: 480 seconds]
vyivel has joined #wayland
creich_ has quit []
creich has joined #wayland
txtsd has joined #wayland
pnowack has quit [Quit: pnowack]
ts has quit []
creich has quit [Quit: Leaving]
pnowack has joined #wayland
maxzor has quit [Ping timeout: 480 seconds]
dblsaiko has quit [Remote host closed the connection]
dblsaiko has joined #wayland
flacks has quit [Quit: Quitter]
flacks has joined #wayland
fmuellner has joined #wayland
<swick>
pq: unfortunately not
<swick>
I do have a colorspace property now but als can't verify that it actually works correctly
<swick>
the plan is to write a KMS app and test a few things
<pq>
swick, my problem is more, when the Colorspace property does not exist, what should a compositor do? Refuse HDR completely?
<pq>
huh, 'evince: error while loading shared libraries: libwayland-cursor.so.0: failed to map segment from shared object' when I'm using self-built libwayland, but with Debian's libwayland it works fine.
<emersion>
have you set LD_LIBRARY_PATH?
<pq>
the same even if I build libwayland 1.18.0 that Debian says to carry
<emersion>
any reason tu use these instead of a blank line?
<pq>
blank line causes... a blank line
<pq>
a line with only a . in the beginning would be the closest to a blank line that doesn't cause a blank line
<pq>
It's odd. I tried to find out what a blank line actually means, and it looks like it could be used to separate normal paragraphs, but having a blank line just before a section heading is... odd. It does add vertical space before the heading.
<pq>
emersion, btw. the commit message tried to say this
<daniels>
ManMower: even if you are really bad at using labels, thanks for at least breaking stuff up into small chunked MRs - makes it way easy to review
<ManMower>
daniels: so yeah, when you said "labels labels labels" earlier, I totally missed what you meant, and thought it was something to do with the content of the patch that fixed the surface label crash.
<daniels>
ManMower: haha no, it was light heckling rather than enormous enthusiasm :P but thanks for fixing labels all the same (wooooo yay!)
<ManMower>
hahaha
Arnavion has quit [Remote host closed the connection]
Arnavion has joined #wayland
<kennylevinsen>
I assume a jitsi/zoom/meet/something link will be sent out later?
manuel__ has quit [Remote host closed the connection]
<emersion>
d_ed: ^
manuel1985 has joined #wayland
<emersion>
last time we used KDE's jitsi iir
<emersion>
iirc*
<jadahl>
is it still aiming to be at 22:00 CEST?
<daniels>
I hope so, because I still need to get stuff for dinner on my way home first
<kennylevinsen>
Yeah apologies for being the inflexible one this time around :(
___nick___ has quit []
___nick___ has joined #wayland
___nick___ has quit []
___nick___ has joined #wayland
naveenk2 has quit []
<d_ed>
emersion: I think I put a link in the doodle
<d_ed>
in any case, I'll make sure to paste here just before too