<jenatali>
I suspect daniels will need to take a look (tomorrow probably)
camus1 has quit [Remote host closed the connection]
camus has joined #dri-devel
mbrost has quit [Remote host closed the connection]
ella-0_ has joined #dri-devel
ella-0 has quit [Read error: Connection reset by peer]
vivijim has joined #dri-devel
slattann has joined #dri-devel
macromorgan has joined #dri-devel
<airlied>
mareko: in your optimising draw paths how do you feel about indirect pointer vs compare and branch?
boistordu_ex has joined #dri-devel
boistordu has quit [Ping timeout: 480 seconds]
camus1 has joined #dri-devel
gpuman has joined #dri-devel
camus has quit [Ping timeout: 480 seconds]
aravind has joined #dri-devel
gpuman has quit [Remote host closed the connection]
gpuman has joined #dri-devel
nikitalita4 has quit [Ping timeout: 480 seconds]
camus1 has quit [Remote host closed the connection]
camus has joined #dri-devel
lemonzest has joined #dri-devel
mclasen has quit [Ping timeout: 480 seconds]
nikitalita4 has joined #dri-devel
nikitalita48 has joined #dri-devel
nikitalita4 has quit [Remote host closed the connection]
sneil_ has joined #dri-devel
Duke`` has joined #dri-devel
sneil has quit [Ping timeout: 480 seconds]
danvet has joined #dri-devel
tzimmermann has joined #dri-devel
FireBurn has quit [Remote host closed the connection]
FireBurn has joined #dri-devel
sneil_ has quit [Ping timeout: 480 seconds]
gpuman has quit [Remote host closed the connection]
sneil has joined #dri-devel
gpuman has joined #dri-devel
aravind has quit [Ping timeout: 480 seconds]
jewins has quit [Read error: Connection reset by peer]
mattrope has quit [Read error: Connection reset by peer]
aravind has joined #dri-devel
mlankhorst has joined #dri-devel
alanc has quit [Remote host closed the connection]
alanc has joined #dri-devel
OftenTimeConsuming has quit [Remote host closed the connection]
pnowack has joined #dri-devel
tursulin has joined #dri-devel
rasterman has joined #dri-devel
boistordu_ex has quit []
anujp has quit [Ping timeout: 480 seconds]
<mripard>
danvet: the merge resolution is not entirely right, how do we fix it?
<daniels>
jenatali: you’ve got more access to that machine than I do :\
anujp has joined #dri-devel
camus1 has joined #dri-devel
camus has quit [Remote host closed the connection]
<javierm>
tzimmermann: another option is for drivers that have a custom modeset module parameter, to just keep open coding it
<javierm>
hopefully we should be able to drop those too when introducing a drm.foobar one
<tzimmermann>
javierm, sure. but it's so much nicer if we use the existing infrastructure :)
<javierm>
tzimmermann: agreed, I can give a try and then danvet decide if is too complicated or the change worth it
sdutt has quit [Ping timeout: 480 seconds]
<javierm>
but for drivers that just check for if (drm_firmware_drivers_only()), custom drm_ registering macros seems the way to go a you said
<javierm>
and also could be used in the future for other logic that's needed in all the drivers
somkls^ has joined #dri-devel
OftenTimeConsuming has joined #dri-devel
boistordu has joined #dri-devel
gawin has joined #dri-devel
<javierm>
tzimmermann: I never thought that proposing dropping the fbdev drivers in Fedora in favour of simpledrm would drag me into this rabbit hole :)
<jani>
javierm: from the department of bikeshedding, should we add a drm.modeset *module* parameter and start promoting that (drm.modeset=0) over the nomodeset kernel parameter?
<tzimmermann>
javierm, lol! simpledrm started as a little side-project...
<tzimmermann>
well here we are
<tzimmermann>
jani, maybe next year
<javierm>
jani: yes, we already discussed that with tzimmermann as the next step
<jani>
tzimmermann: heh
<tzimmermann>
javierm, my company gave us some time to experiment. i wanted to see if i can run DRM on a very simple framebuffer, just like efifb. but we have a lot of problems with fbdev, and it's also outdated. so i moved forward with simpledrm.
<javierm>
tzimmermann: yes, simpledrm is awesome. Thanks a lot for that
pcercuei has joined #dri-devel
<tzimmermann>
i did almost nothing TBH. the work is in all the DRM helpers. i did little of that
<mripard>
tzimmermann: while you're there, I'm not sure what the issue was with vc4 yesterday?
<mripard>
drm-tip had the proper merge afaik, so I'm a bit confused about why that caused any issue?
slattann has quit []
<tzimmermann>
mripard, i'm not sure. i only know that vc4 required a fixup in drm-tip because to patches wrt gem_create_object conflicted
<tzimmermann>
danvet ^
<mareko>
airlied: inlining is the best
<mareko>
airlied: for DrawGallium, the best thing to do is to inline it in all draw functions; additionally, I've been meaning to inline _mesa_Draw into _mesa_unmarshal_Draw, but that's a lot more complicated
<danvet>
mripard, the drm-tip merge has put one of the if () checks in commit_tail in the wrong spot
<danvet>
I moved it to where it should be
<danvet>
from a quick look you did that
<danvet>
mripard, the one that has the clk_set_min_rate(hvs->core_clk, core_rate);
<danvet>
call
<danvet>
it was before the for loop over HVS_NUM_CHANNELS
<danvet>
you somehow put it after in the git rerere merge resolution
<danvet>
and since I couldn't figure out why it should move I undid that part
<danvet>
I also added the right Fixes: line to point at my merge
<mripard>
yes
<danvet>
thx
<danvet>
tzimmermann, airlied should be resolved soon
<tzimmermann>
ok
<danvet>
apologies to everyone, I generally try to do better, no idea why I didn't spot that commit which moved stuff yesterday
<danvet>
it's clearly there now even with the gitk command from shell history from yesterday
alatiera has quit [Remote host closed the connection]
alatiera has joined #dri-devel
ramaling has quit []
ramaling has joined #dri-devel
hikiko_ has joined #dri-devel
hikiko has quit [Ping timeout: 480 seconds]
Company has joined #dri-devel
hikiko_ has quit [Remote host closed the connection]
hikiko has joined #dri-devel
<danvet>
tzimmermann, scrip got stuck again, but drm-next is pushed now in case you still want to backmerge
<tzimmermann>
danvet, there's no hurry
camus1 has joined #dri-devel
camus has quit [Ping timeout: 480 seconds]
<mripard>
danvet: don't worry, it happens :)
mclasen has joined #dri-devel
flacks has quit [Quit: Quitter]
flacks has joined #dri-devel
OftenTimeConsuming has quit [Remote host closed the connection]
OftenTimeConsuming has joined #dri-devel
<HdkR>
Gah. Someone slipped in to my DMs and told me that XDC 2022 was cancelled. Freaked me out so I had to look it up and found a website that I took at face value. Wrong XDC. whew, whole gamut of emotions.
camus has joined #dri-devel
camus1 has quit [Ping timeout: 480 seconds]
<marex>
danvet: tzimmermann: is there nobody maintainering drm-misc anymore ?
<marex>
I posted various patches, panel-simple ones, bridge patches, but there is nobody to apply them
<danvet>
marex, you don't have commit rights?
<marex>
danvet: nope (also, applying my own patches sounds ... wrong)
<danvet>
marex, but in general we try to list maintainers for specific sections, so ping them
<danvet>
marex, you just need a reviewed-by or acked-by from someone else
<danvet>
which the scripts check for
<marex>
danvet: sravn: asked me to ping him in a few days if there is no feedback on this one bridge patch ... that was three pings and two months ago :)
jkrzyszt has joined #dri-devel
<danvet>
also everywhere else pushing your own patches is totally standard for regular contributors
<danvet>
marex, bridge is a bit abandoned land in general
<marex>
danvet: when did I become some sort of regular contributor ?
<marex>
I am a total drive-by
<danvet>
I'm good at volunteering people
<danvet>
or at least trying :-)
<marex>
danvet: I am good at volunteering myself, but I know where my skills are insufficient yet
<marex>
well anyway, has Sam been around recently ?
<marex>
sravn: ping ?
OftenTimeConsuming has quit [Remote host closed the connection]
<cmarcelo>
I have a separate MR that fix the actual issue with nir_serialize, that I intend to land first, but with the MR above you can confirm the error in debian-vulkan is more reasonable.
<cmarcelo>
daniels: ^
ybogdano has joined #dri-devel
<daniels>
cmarcelo: nice, thanks for the quick fix! you can have a very lazy A-b on that
slattann has joined #dri-devel
camus1 has quit [Ping timeout: 480 seconds]
camus has joined #dri-devel
jljusten has joined #dri-devel
fxkamd has joined #dri-devel
fxkamd has quit []
ybogdano has quit [Ping timeout: 480 seconds]
vivijim has quit [Quit: leaving]
fxkamd has joined #dri-devel
mlankhorst has quit [Ping timeout: 480 seconds]
<airlied>
mareko: the problem is ctx->Driver.DrawGallium selection, if I get rid of it I have to introduce a if unlikely(ctx->RenderMode != GL_RENDER) in the draw path
<idr>
airlied: It will be hard to argue with data. :)
<idr>
(But not impossible.)
tarnsformers has joined #dri-devel
camus1 has joined #dri-devel
nwv1eeb has joined #dri-devel
tarnsformers has quit []
camus has quit [Ping timeout: 480 seconds]
nwv1eeb has quit []
<jekstrand>
anholt: Running shader-db on HSW now
<jekstrand>
anholt: there you go
lemonzest has quit [Quit: WeeChat 3.3]
<sravn>
marex, danvet, tzimmermann, busy with real life stuff. Moving to an older house that needs renovating takes all available time, and more. No promises but hope to have time and energy for some linux stuff again sometime
<anholt>
jekstrand: was there a trick I was missing?
<jekstrand>
anholt: Having an actual Haswell?
<anholt>
:/
<anholt>
we used to be able to do this without that
<jekstrand>
I'm sure it's possible somehow but I don't know how off-hand
<jekstrand>
idr knows
<jekstrand>
or at leat has a script somewhere
<idr>
The bit6 thing is recent, I think.
<idr>
I had that problem with something else recently.
<idr>
It seems like shader-db worked, though.
Daanct12 has joined #dri-devel
<idr>
That's odd... my script always specifies -o, and it won't load iris now. But it will load crocus.
Danct12 has quit [Ping timeout: 480 seconds]
Danct12 has joined #dri-devel
Daanct12 has quit [Ping timeout: 480 seconds]
f11f12 has joined #dri-devel
rasterman has quit [Quit: Gettin' stinky!]
<idr>
Ugh... my bisect seems to be narrowing in on my intel-stub commits from !14132. :(
<idr>
Actually, no... something after that. Whew!
<marex>
sravn: well that sucks, you were one of my fav maintainers, good luck with the house
thellstrom has joined #dri-devel
slattann has quit []
ngcortes has joined #dri-devel
<idr>
anholt, jekstrand: The culprit is 88e97d75d06 (refs/bisect/bad) intel/dev: Add gtt_size to devinfo
<idr>
(At least on builds without assertions.)
danvet has quit [Ping timeout: 480 seconds]
OftenTimeConsuming has joined #dri-devel
ybogdano has joined #dri-devel
<jekstrand>
idr: That makes sense
<jekstrand>
We're fetching the GTT size a lot earlier now
<idr>
Do we just need more stuff in the shim?
<idr>
I have some other patches in that area already.
jcline_ has left #dri-devel [#dri-devel]
camus has joined #dri-devel
f11f12 has quit [Quit: Leaving]
mszyprow has joined #dri-devel
camus1 has quit [Ping timeout: 480 seconds]
Duke`` has quit [Ping timeout: 480 seconds]
JohnnyonFlame has joined #dri-devel
gouchi has quit [Remote host closed the connection]
<idr>
Hm... I see what the problem is. My script doesn't use drm-shim via intel_stub_gpu. Instead it uses INTEL_DEVID_OVERRIDE and LD_PRELOAD.
<idr>
That causes no_hw, and the prevents a bunch of the initialization.
<idr>
That's also why I don't hit the bit6_swizzle assertion.
<idr>
I guess that means that any attempt to use INTEL_DEVID_OVERRIDE or INTEL_NO_HW will fail similarly.
* idr
lunches, then fixes that problem.
Haaninjo has quit [Quit: Ex-Chat]
mszyprow has quit [Ping timeout: 480 seconds]
<anholt>
heads up: I'm fixing up ASan support for vulkan in CI. we should have leak checks soon.