ChanServ changed the topic of #wayland to: https://wayland.freedesktop.org | Discussion about the Wayland protocol and its implementations, plus libinput
<Ermine>
dviola: sure
<dviola>
thanks
<Ermine>
will test on my hw tomorrow
bjorkintosh has quit [Quit: "Every day, computers are making people easier to use." David Temkin]
<Ermine>
Seems like virtio-gpu/virgl still need a lot of work
tokyo4j has quit [Read error: Connection reset by peer]
tokyo4j has joined #wayland
melnary has quit [Remote host closed the connection]
melnary has joined #wayland
mceier has quit [Quit: Reconnecting]
mceier has joined #wayland
garnacho has quit [Ping timeout: 480 seconds]
glennk has joined #wayland
yaslam_ has joined #wayland
garnacho has joined #wayland
yaslam has quit [Ping timeout: 480 seconds]
kts has joined #wayland
kts has quit [Quit: Leaving]
bookworm_ has joined #wayland
bookworm has quit [Read error: Connection reset by peer]
vyivel has quit [Remote host closed the connection]
vyivel has joined #wayland
garnacho has quit [Ping timeout: 480 seconds]
dviola has quit [Quit: WeeChat 4.4.3]
kabaka101 has quit [Remote host closed the connection]
kabaka101 has joined #wayland
kts has joined #wayland
Gilfoyl3 has joined #wayland
Gilfoyl3 has quit [Remote host closed the connection]
Gilfoyl3 has joined #wayland
Gilfoyl3 has quit [Quit: Gilfoyl3]
kts has quit [Quit: Leaving]
fmuellner has quit [Ping timeout: 480 seconds]
kts has joined #wayland
rgallaispou has quit [Read error: Connection reset by peer]
rgallaispou has joined #wayland
mxz_ has joined #wayland
mxz has quit [Ping timeout: 480 seconds]
mxz__ has quit [Ping timeout: 480 seconds]
mxz_ is now known as mxz
kts has quit [Quit: Leaving]
dviola has joined #wayland
<dviola>
found more new things...
<dviola>
*if* I reboot my VM (fresh reboot) and start weston, there is no cursor issue in the VM whatsoever
<dviola>
*if* I do a fresh reboot and start sway, exit sway, then start weston, there is a cursor issue
<dviola>
let me confirm if it happens with river too
<danieldg>
that sounds like weston is assuming something is in an initial state and the kernel isn't guaranteeing that
<danieldg>
sway would be leaving it in some different state
<danieldg>
might be a kernel bug (not properly resetting) or a weston bug (assuming a state when it needs to have set it)
<dviola>
it happens with river too
<dviola>
hyprland doesn't seem to destroy the cursor
<danieldg>
try running one with software cursors turned on
<danieldg>
I don't recall the environment variable, wlroots docs should have it
<dviola>
WLR_NO_HARDWARE_CURSORS=1 ?
<danieldg>
sounds right
<dviola>
well, I'm glad we have some clue now... there's something that wlroots is doing that wrecks the cursor
<dviola>
looks like soreau was spot on on that
<dviola>
one sec I'll try that in a bit
<dviola>
danieldg: yes, starting sway with WLR_NO_HARDWARE_CURSORS=1 prevents the cursor from being broken with the other compositors (weston, hyprland)
<llyyr>
dviola: what does sway do if you recompile wlroots/sway after setting line 117 in backend/drm/drm.c to #if 0
<dviola>
llyyr: just tried that and that fixes the issue
<llyyr>
are you sure it's not just using a software cursor or something
<dviola>
*shrug*
<llyyr>
should say in logs, or check if direct scanout is working
<dviola>
it works OOTB, that's the beauty of it!
<dviola>
let me check that though
<llyyr>
well it's no different than setting WLR_NO_HARDWARE_CURSORS, I'm guessing the reason why your bisect ended up at that commit is because that's when the hotspot_x/y props became available
<dviola>
understood
<dviola>
but yeah, I don't need hardware cursors on the VM, I want things to just work OOTB, otherwise folks go like "wayland sucks" and go back to X
Gilfoyl3 has quit [Quit: Gilfoyl3]
<llyyr>
open an issue on wlroots tracker so folks more familiar with this stuff can take a look at it. the reason why weston/hyprland worked is because they don't set this client cap, and the reason why river worked is because it's using the latest wlroots release which doesn't have this yet
Gilfoyl3 has joined #wayland
<dviola>
will do, thanks
CuteMath has joined #wayland
Gilfoyl3 has quit []
Gilfoyl3 has joined #wayland
Gilfoyl3 has quit [Remote host closed the connection]
<llyyr>
it's using software cursor... because your cursor is too big?
<llyyr>
>[types/output/cursor.c:211] Cursor texture too large (24x24), exceeds hardware limitations
Gilfoyl3 has joined #wayland
<llyyr>
nope, it just means no cursor plane was found. weird that it's printed like that
<llyyr>
so yeah, it is using software cursor. You also never actually provided a log file with wlroots/sway master, maybe it says something useful too. that's why I suggested opening an issue
<dviola>
right, apologies for being late on that
glennk has quit [Ping timeout: 480 seconds]
<dviola>
went through a lot of things, my mind was set on the kernel and mesa for about two days straight
Gilfoyl3 has quit [Remote host closed the connection]
Gilfoyl3 has joined #wayland
Gilfoyl3 has quit [Quit: Gilfoyl3]
Gilfoyl3 has joined #wayland
Gilfoyl3 has quit [Remote host closed the connection]
<dviola>
not sure what to call the issue, "wlroots cursor regression" maybe?
Gilfoyl3 has joined #wayland
<dviola>
yeah, I think that will do
Gilfoyl3 has quit [Quit: Gilfoyl3]
Gilfoyl3 has joined #wayland
Gilfoyl3 has quit []
<soreau>
<danieldg> might be a kernel bug (not properly resetting) or a weston bug (assuming a state when it needs to have set it) <-- (or a wlroots bug) it seems like the most likely culprit is the kernel, because after exiting one and starting the other, the state should be 'clean'
Gilfoyl3 has joined #wayland
Gilfoyl3 has quit []
<psykose>
the same happens for enabling hdr in a compositor then swapping to another
<psykose>
last i asked i think there was no way to really do that yet
<soreau>
the two should have nothing to do with each other, but if you can reliably reproduce this.. the problem variable/state/whatever must lie in the kernel
<soreau>
dviola: have you looked through kernel logs for these cases?
<soreau>
maybe something obvious pops up
<dviola>
kernel logs are fine
<dviola>
nothing interesting there
<soreau>
dviola: even after starting sway, shutting it down and running weston?
<soreau>
am I blind, or is this a case of ub due to uninitialized variable?
bookworm_ is now known as bookworm
<llyyr>
dviola: please add logs to your issue
<dviola>
done
<dviola>
I was doing that
Gilfoyl3 has joined #wayland
Gilfoyl3 has quit [Remote host closed the connection]
<soreau>
even if struct drm_file was calloc'd, there's nothing that changes it
Gilfoyl3 has joined #wayland
Gilfoyl3 has quit [Remote host closed the connection]
<dviola>
feel free to request more info on the issue or here
Gilfoyl3 has joined #wayland
Gilfoyl3 has quit [Remote host closed the connection]
<dviola>
soreau: yes, tested that, there's nothing on dmesg
<soreau>
dviola: what is the output of `grep -Rn supports_virtualized_cursor_plane linux-git/drivers/gpu/drm linux-git/include/drm`? (where linux-git is your linux git repo)
Gilfoyl3 has joined #wayland
Gilfoyl3 has quit [Remote host closed the connection]
<dviola>
give me some time, have to re-clone
<dviola>
I'm back to running stock arch kernel
kts has joined #wayland
Gilfoyl3 has joined #wayland
<soreau>
I see it's drivers/gpu/drm/drm_ioctl.c:374:file_priv->supports_virtualized_cursor_plane = req->value;
crazybyte has quit [Read error: Connection reset by peer]
crazybyte6 is now known as crazybyte
sima has joined #wayland
JakeSays has joined #wayland
JakeSays1 has quit [Ping timeout: 480 seconds]
leon-anavi has joined #wayland
<dviola>
hey folks, I understand now that sway is trying to do the right thing, so please don't feel you have to change anything to accomodate my needs with the VM, I'm all for "just work" purposes because I think it helps with usability but it sounds like sway prefers to do that right
<dviola>
hopefully the investigation is still useful in some way
<dviola>
accommodate*
<MrCooper>
soreau: "after exiting one and starting the other, the state should be 'clean'" would be nice, wouldn't it? Unfortunately, the current reality is that KMS state remains unchanged in that case
glennk has joined #wayland
<soreau>
MrCooper: are you saying that the compositor has to ensure something happens so that subsequent instances don't exhibit the problem?
<MrCooper>
more like: every DRM master is responsible for ensuring the state they want; nice ones disable "unusual" state when giving up control
<soreau>
hm, what if it crashes?
<MrCooper>
hilarity may ensue
<soreau>
it seems like the kernel would be the bigger entity and track who made the call
<MrCooper>
especially since not even fbcon fully re-initializes state
<MrCooper>
it's been a known issue since (atomic?) KMS exists, we've all just been kicking the can down the street though
rasterman has joined #wayland
<JEEB>
yea my understanding is that it just keeps last configured state, and then whatever application takes over KMS master, that inherits whatever the initial state is?
JakeSays1 has joined #wayland
<MrCooper>
whatever the current (not initial) state is
<JEEB>
yes, I then understood the word "initial" is ambiguous
<JEEB>
I meant "whatever the state was as that point"
<JEEB>
not "some nice fuzzy initial state"
JakeSays has quit [Ping timeout: 480 seconds]
<dviola>
sorry for polluting the channel the last couple of days, I didn't realize the channel was for protocol discussion :|
<dviola>
I mean, I forgot
MrCooper_ has joined #wayland
garnacho has joined #wayland
rgallaispou has quit [Read error: Connection reset by peer]
rgallaispou has joined #wayland
iomari891 has joined #wayland
MrCooper has quit [Ping timeout: 480 seconds]
slim has quit [Quit: slim]
bindu has quit [Remote host closed the connection]
bindu has joined #wayland
adamtajti_ has joined #wayland
adamtajti has quit [Read error: Connection reset by peer]
MrCooper_ is now known as MrCooper
anarsoul|2 has quit [Remote host closed the connection]
anarsoul has joined #wayland
feaneron has joined #wayland
fmuellner has joined #wayland
Moprius has joined #wayland
Gilfoyl3 has joined #wayland
rasterman has quit [Quit: Gettin' stinky!]
Gilfoyl3 has quit []
kts has joined #wayland
yshui_ has joined #wayland
FreeFull has quit []
abhimanyu1 has joined #wayland
rasterman has joined #wayland
yshui has quit [Ping timeout: 480 seconds]
rasterman has quit []
cmichael has joined #wayland
yshui has joined #wayland
yshui_ has quit [Ping timeout: 480 seconds]
kts has quit [Quit: Leaving]
Moprius has quit [Remote host closed the connection]
yshui has quit [Remote host closed the connection]
yshui has joined #wayland
yshui has quit [Remote host closed the connection]
yshui has joined #wayland
FreeFull has joined #wayland
rasterman has joined #wayland
bjorkintosh has joined #wayland
glennk has quit [Remote host closed the connection]
glennk has joined #wayland
kts has joined #wayland
kts has quit [Ping timeout: 480 seconds]
d_ed[m] has joined #wayland
<d_ed[m]>
jadahl: I have a start of a Qt implementation of Xdg Session Manager.
<d_ed[m]>
If I run this against mutter with the modified xx_ prefix, what should I be able to see in terms of end results?
bl4ckb0ne has joined #wayland
pochu has quit [Remote host closed the connection]
rgallaispou has quit [Read error: Connection reset by peer]
cmichael has quit [Quit: Leaving]
abhimanyu1 has quit [Remote host closed the connection]
abhimanyu1 has joined #wayland
rasterman has quit [Ping timeout: 480 seconds]
rasterman has joined #wayland
rasterman has quit [Quit: Gettin' stinky!]
tzimmermann has quit [Quit: Leaving]
adamtajti_ has quit [Ping timeout: 480 seconds]
rasterman has joined #wayland
feaneron has quit [Quit: feaneron]
nysach has joined #wayland
nysach has quit [Remote host closed the connection]
rasterman has quit [Quit: Gettin' stinky!]
yshui has quit [Read error: Connection reset by peer]
abhimanyu1 has quit [Read error: No route to host]
abhimanyu1 has joined #wayland
abhimanyu1 has quit [Read error: Connection reset by peer]
abhimanyu1 has joined #wayland
abhimanyu1 has quit [Read error: No route to host]
abhimanyu1 has joined #wayland
abhimanyu1 has quit []
feaneron has joined #wayland
<d_ed[m]>
<jadahl> "d_ed: iirc window position" <- Got everything working. Can we try and do a meeting sometime to work on the higher level part making it useful.
kabaka101 has joined #wayland
CME_ has joined #wayland
CME has quit [Ping timeout: 480 seconds]
yshui has joined #wayland
yshui has quit [Read error: Connection reset by peer]
yshui has joined #wayland
kabaka101 has quit [Ping timeout: 480 seconds]
yaslam has joined #wayland
kabaka101 has joined #wayland
<jadahl>
d_ed[m]: higher level as in higher than Wayland, or the higher level protocol bits?
<d_ed[m]>
Higher than wayland. I assume we'll need some .desktop file group where we can say
<d_ed[m]>
Exec=myApp -session %s
<d_ed[m]>
So that we can restore them on a given session.
karenw has joined #wayland
Calandracas has joined #wayland
Calandracas has quit [Remote host closed the connection]
Calandracas has joined #wayland
Calandracas_ has quit [Ping timeout: 480 seconds]
<jadahl>
d_ed[m]: there has been work in glib land to try to come up with a method for the "higher level" bits, but can't remember what. probably need more than a desktop file, i.e. a session manager API one registers too, or something. there was some discussion in https://gitlab.freedesktop.org/xdg/xdg-specs/-/merge_requests/79
<jadahl>
or it seems the idea is other way around, o.f.Restartable
Moprius has joined #wayland
melnary has quit [Remote host closed the connection]
melnary has joined #wayland
melnary has quit [Remote host closed the connection]