<javierm>
tzimmermann_, daniels: for trivial changes in different drivers, do you want the patches to be spit per driver or a single one changing all is OK ?
<javierm>
*split
<javierm>
tzimmermann_, danvet: I also realized that currently just using nomodeset is enough to have the wanted behaviour, because the simpledrm driver don't check for that while other drivers do
<danvet>
javierm, yeah that was my point from yesterday :-)
<danvet>
javierm, split up helps to get acks and stuff, but also tends to be more of a drag
<danvet>
if it's just renaming/changing the included header to get at the check function and nothing else then I think one patch is ok
<danvet>
but if you also massage the checks in drivers, then better to split
<javierm>
danvet: It's a simple rename although it inverts the logic check
<javierm>
danvet: yes, I didn't get your point because hadn't realized that. I also learned yesterday that no all drivers honour nomodeset, I plan to fix that as a follow-up
<MrCooper>
javierm: nomodeset is enough except for the amdgpu issue we discussed yesterday?
<javierm>
MrCooper: yes, because drivers need to opt-in and only a subset of them do. The simpledrm driver is one that doesn't
<javierm>
so using nomodeset currently will allow simpledrm to load but will prevent for example virtio-gpu
<MrCooper>
and will prevent loading amdgpu without a reboot
<javierm>
MrCooper: yes, but that's a separate issue. I think that either amdgpu needs to also have a modeset module param or we add a drm.modeset that could override nomodeset value
<javierm>
danvet: I think that will keep as a single patch for now since will be v1 and I expect some bikeshedding on the function names, header used for declaration, etc
<MrCooper>
k, requiring drivers to have a modeset parameter just to override nomodeset seems a bit silly :)
<javierm>
MrCooper: agreed. That's why I mentioned a drm.modeset param. Maybe even the per-driver ones could be dropped and people just use that
<MrCooper>
yep, that sounds good
<javierm>
I can't think of a user case where someone would want to only override for a driver but no for other
<javierm>
*use
<MrCooper>
there's always modprobe.blacklist for disabling individual drivers
<javierm>
MrCooper: yeah
<danvet>
MrCooper, yeah for debugging that's the one I use
<danvet>
for users nomodeset seems perfectly good enough
<MrCooper>
converting nomodeset to a drm parameter should make it easier to provide generic troubleshooting instructions
<javierm>
danvet: problem I think is that it's hard to understand what it means
camus1 has joined #dri-devel
camus has quit [Read error: Connection reset by peer]
mclasen has joined #dri-devel
jkrzyszt has quit [Remote host closed the connection]
jkrzyszt has joined #dri-devel
<pq>
Wouldn't making nomodeset runtime-changeable via sysfs avoid the need of adding another option to override it?
shadeslayer has joined #dri-devel
mclasen has quit [Ping timeout: 480 seconds]
<javierm>
pq: but that's precisely the issue, that global (non-module) parameters can't be changed via sysfs
<javierm>
so we need something like a /sys/module/drm/modeset
mclasen has joined #dri-devel
tzimmermann_ has quit []
tzimmermann has joined #dri-devel
shadeslayer has quit [Quit: Konversation terminated!]
jkrzyszt has quit [Remote host closed the connection]
<pq>
javierm, I never knew that about global parameters. Seems a bit odd.
<javierm>
pq: yeah, I also thought that there was possible to switch those through sysfs
<javierm>
and of course I forgot to include dri-devel in the Cc list, sigh...
jkrzyszt has joined #dri-devel
<MrCooper>
pq: right now nomodeset only exists on the kernel command line, which cannot be modified
aravind has quit [Read error: Connection reset by peer]
aravind has joined #dri-devel
<pq>
not even as a bool or int variable, huh
nroberts has joined #dri-devel
<javierm>
pq: yeah, __setup() and __setup_param() macros don't support type like module_param() and module_param_named()
<javierm>
I was surprised as well
vsyrjala has quit [Ping timeout: 480 seconds]
pendingchaos has quit [Quit: Quit]
<tzimmermann>
javierm, what's the usecase?
<javierm>
tzimmermann: what MrCooper mentioned, you boot with nomodeset but want to load the module after booting (if for example want to debug why is failing)
<javierm>
the only way right now is to reboot without nomodeset in the cmdline
<tzimmermann>
modprobe.blacklist?
itoral has quit [Remote host closed the connection]
itoral has joined #dri-devel
<javierm>
tzimmermann: right, there are workarounds for this but I was mentioning the general fact that non-module parameters can't be changed from sysfs entries
<javierm>
I would had expected there was a way to do that from user-space
camus has joined #dri-devel
camus1 has quit [Ping timeout: 480 seconds]
thelounge51 is now known as alatiera
itoral has quit [Remote host closed the connection]
itoral has joined #dri-devel
vsyrjala has joined #dri-devel
vivijim has joined #dri-devel
itoral has quit [Remote host closed the connection]
<javierm>
an, found why I missed some addresses the in the Cc list. The ./scripts/get_maintainer.pl script doesn't work correctly when 'git config diff.noprefix true'
itoral has joined #dri-devel
itoral has quit [Remote host closed the connection]
itoral has joined #dri-devel
itoral has quit [Remote host closed the connection]
itoral has joined #dri-devel
Akari has quit [Ping timeout: 480 seconds]
Company has joined #dri-devel
pendingchaos has joined #dri-devel
sdutt has joined #dri-devel
rando25892 has quit [Quit: No windows for this server]
JohnnyonFlame has quit [Read error: Connection reset by peer]
jewins has joined #dri-devel
nchery has joined #dri-devel
f11f13 has quit [Ping timeout: 480 seconds]
itoral has quit [Remote host closed the connection]
<zmike>
anholt: I'm looking at the deqp docker image and it seems like there's no runner for core gl on it? only es?
Akari has joined #dri-devel
Ryback_[WORK] has quit []
Ryback_ has joined #dri-devel
<zmike>
oh nvm
Duke`` has joined #dri-devel
fxkamd has quit []
pendingchaos has quit [Ping timeout: 480 seconds]
The_Company has joined #dri-devel
Company has quit [Ping timeout: 480 seconds]
Thymo has quit [Ping timeout: 480 seconds]
Thymo has joined #dri-devel
mbrost has joined #dri-devel
sravn has joined #dri-devel
ybogdano has joined #dri-devel
gouchi has joined #dri-devel
<Venemo>
what happened to the vulkan layers? did these get removed? I get this error: build64/meson-logs/meson-log.txt:WARNING: Unknown options: "vulkan-device-select-layer, vulkan-overlay-layer"
<imirkin>
zmike: that's been the case since the dawn of the "stable" process...
<zmike>
great, but probably not everyone has been here for the full 20 years, so having an announcement when new features are added to it would help bring everyone up to speed too
<imirkin>
zmike: new features to what? it's not a new feature of the stable process
<imirkin>
anyways, i'm 99% sure it was documented somewhere