<mareko>
zmike: one of my regrets of accepting PIPE_CAP_SUPPORTED_PRIM_MODES_WITH_RESTART is that 1) it breaks tessellation if drivers report 0 for patches 2) if drivers don't report PIPE_CAP_EMULATE_NONFIXED_PRIMITIVE_RESTART, it takes effect only if there are user buffers, which have nothing to do with it, it's a mess
<zmike>
mareko: hm
mbrost has quit [Ping timeout: 480 seconds]
<zmike>
could always adjust PIPE_CAP_SUPPORTED_PRIM_MODES_WITH_RESTART to always report 1 for patches (i.e., add it when the value is read)
JGue has quit [Remote host closed the connection]
<zmike>
not sure I get what you mean with #2
<zmike>
I made a ticket, will do MRs once we get to solutions
aravind has joined #dri-devel
Duke`` has joined #dri-devel
ddavenport has joined #dri-devel
tzimmermann has joined #dri-devel
jewins has quit [Quit: jewins]
ddavenport has quit [Ping timeout: 480 seconds]
aravind has quit [Remote host closed the connection]
luzipher_ has joined #dri-devel
camus1 has quit [Remote host closed the connection]
mattrope has quit [Remote host closed the connection]
camus has joined #dri-devel
YuGiOhJCJ has quit [Quit: YuGiOhJCJ]
luzipher__ has quit [Ping timeout: 480 seconds]
camus has quit [Remote host closed the connection]
camus has joined #dri-devel
luzipher__ has joined #dri-devel
sdutt has quit [Remote host closed the connection]
luzipher_ has quit [Ping timeout: 480 seconds]
mbrost has joined #dri-devel
martin19 has joined #dri-devel
mbrost has quit [Remote host closed the connection]
randomher0 has quit []
mlankhorst has joined #dri-devel
sneil_ has joined #dri-devel
sneil has quit [Ping timeout: 480 seconds]
Duke`` has quit [Ping timeout: 480 seconds]
luzipher_ has joined #dri-devel
RobertC has joined #dri-devel
pnowack has joined #dri-devel
luzipher__ has quit [Ping timeout: 480 seconds]
alanc has quit [Remote host closed the connection]
aravind has joined #dri-devel
alanc has joined #dri-devel
mbrost has joined #dri-devel
tobiasjakobi has joined #dri-devel
tobiasjakobi has quit [Remote host closed the connection]
luzipher__ has joined #dri-devel
luzipher_ has quit [Ping timeout: 480 seconds]
danvet has joined #dri-devel
gouchi has joined #dri-devel
bcarvalho has joined #dri-devel
bcarvalho has quit []
NiksDev has joined #dri-devel
yevhenii has quit [Remote host closed the connection]
gouchi has quit [Quit: Quitte]
andrey-konovalov has joined #dri-devel
<MrCooper>
mdnavare: conceptually, there's no reason why async flips couldn't work with VRR; if the flip becomes ready during vertical blank, it's the same as non-async, otherwise there may be tearing
<MrCooper>
AFAIK this is working with amdgpu DC
bcarvalho has joined #dri-devel
luzipher_ has joined #dri-devel
luzipher__ has quit [Ping timeout: 480 seconds]
pekkari has joined #dri-devel
lynxeye has joined #dri-devel
rasterman has joined #dri-devel
Lucretia has joined #dri-devel
pcercuei has joined #dri-devel
ddavenport has joined #dri-devel
mbrost has quit [Remote host closed the connection]
ddavenport has quit [Remote host closed the connection]
reb0rn_ has joined #dri-devel
ddavenport has joined #dri-devel
ddavenport has quit [Remote host closed the connection]
mbrost has joined #dri-devel
camus1 has joined #dri-devel
camus has quit [Ping timeout: 480 seconds]
mbrost_ has joined #dri-devel
mbrost__ has joined #dri-devel
mbrost has quit [Remote host closed the connection]
mbrost_ has quit [Remote host closed the connection]
ddavenport has joined #dri-devel
ddavenport has quit [Remote host closed the connection]
mbrost__ has quit [Remote host closed the connection]
<sravn>
pinchartl: I have started converting all bridges to the atomic_{enable,disable} functions. Looks mostly trivial and the win is some nice reduction in loc in drm_bridge.h and drm_bridge.c
<sravn>
pinchartl: Just is case you took the bait and had planned to update the documentation - no need as I am on it
<pinchartl>
sravn: thank you :-)
ppascher has quit [Ping timeout: 480 seconds]
camus1 has quit [Remote host closed the connection]
camus has joined #dri-devel
<martin19>
what could be the reason the framebuffer id changes every few seconds for i915 on linux x11 and wayland?
tobiasjakobi has joined #dri-devel
tobiasjakobi has quit [Remote host closed the connection]
khfeng has quit [Ping timeout: 480 seconds]
camus has quit []
gpoo has joined #dri-devel
ppascher has joined #dri-devel
Peste_Bubonica has joined #dri-devel
idr has quit [Ping timeout: 480 seconds]
<mareko>
zmike: #2 is implied the fact that fallback_always ignores PIPE_CAP_SUPPORTED_PRIM_MODES_WITH_RESTART
<mareko>
zmike: on a different note, have you ever considered taking a vacation, which is the time when normal people don't do anything productive? ;)
<MrCooper>
martin19: I'd expect it to change every frame actually, due to page flipping
<martin19>
MrCooper: do i need to discover the current fbid every time i'll call drm drmModeGetFb2? how could i detect a mode change?
<MrCooper>
KMS isn't suitable for reliable generic frame grabbing, if that's what you're after
<martin19>
nomachine is doing it i suppose
<MrCooper>
it's better to get the contents from the display server
iive has joined #dri-devel
aravind has quit [Remote host closed the connection]
<martin19>
do you know if there is a zero copy approach for x11?
<MrCooper>
there can't be with any method, without risking that buffer getting overwritten prematurely for screen updates
<MrCooper>
for X11, I recommend copying from the root window to a pixmap, then that pixmap can be used zero copy via DRI3
<MrCooper>
or, if available, use the XDG screencast portal, which works the same with X11 or Wayland
<martin19>
makes all sense to use these apis if they are stable and performance is good.
<martin19>
i'll try these approaches you mention and compare. Thanks for your help!
<MrCooper>
np, good luck
<zmike>
mareko: huh, not doing anything productive, you say? is such a thing even possible?
blue__penquin has quit []
<martin19>
just one thing..: i've been looking for some time for low latency screen sharing on linux. I'm happy with experimental solutions I dont want to sell anything. I want to see what is possible. cheesy vnc clients I think there are enough..
pcercuei has quit [Quit: brb]
pcercuei has joined #dri-devel
gouchi has joined #dri-devel
sdutt has joined #dri-devel
NiksDev has quit [Ping timeout: 480 seconds]
idr has joined #dri-devel
sneil_ has quit []
sneil has joined #dri-devel
gouchi has quit [Remote host closed the connection]
RobertC has quit [Ping timeout: 480 seconds]
gouchi has joined #dri-devel
mattrope has joined #dri-devel
sneil has quit []
sneil has joined #dri-devel
reb0rn_ has quit []
<mareko>
zmike: trust me it is, I just did it today
<zmike>
mareko: well maybe I'll give it a shot then 🤔
* jekstrand
feels entirely unmotivated to do anything today
<jekstrand>
Might have something to do with not getting to sleep until 3:00 AM and then waking up with the sun. :-(
<mareko>
jekstrand: I haven̈́'t slept today, but will in about 2 hours
sarnex has quit [Quit: Quit]
<jekstrand>
Maybe I'll read zmike's latest ANV MR. That sounds slightly productive.
<jekstrand>
And then zmike will be forced to be productive today too. :P
pekkari has quit [Quit: Konversation terminated!]
<zmike>
I've already been productive
<zmike>
it's too late for me
<zmike>
save yourself
<jekstrand>
Too late. Now you're abliged to make 3 tiny changes and land multi-draw. :P
gouchi has quit [Remote host closed the connection]
gouchi has joined #dri-devel
<zmike>
that sounds like a job for post-gym me
<zmike>
danylo: wow, I think 2.5% is the record
<danylo>
In which direction?
<zmike>
smallest improvement
<zmike>
even ANV got about 40%
<danylo>
wow, on turnip when there is no state change between draw calls - there is nothing to emit but a draw itself.
<karolherbst>
same on nv hardware
<zmike>
do you have dynamic state hooked up?
GloriousEggroll has joined #dri-devel
<danylo>
yes
<zmike>
huh
<zmike>
nice efficiency :)
Duke`` has joined #dri-devel
gouchi has quit [Ping timeout: 480 seconds]
<danylo>
on which tests ANV got 40% improvement?
<zmike>
drawoverhead
<zmike>
base tests
Peste_Bubonica has quit [Quit: Leaving]
sdutt has quit []
gouchi has joined #dri-devel
NiksDev has joined #dri-devel
gouchi has quit [Ping timeout: 480 seconds]
hanetzer has quit [Ping timeout: 480 seconds]
gouchi has joined #dri-devel
alyssa has joined #dri-devel
<alyssa>
this might be the single worst shader-db stat i've seen in my career to date
<alyssa>
ldst in affected programs: 2 -> 2120 (105900.00%)
<karolherbst>
alyssa: what did you do
<karolherbst>
spilling?
<alyssa>
yep.
<karolherbst>
sounds familiar :D
gouchi has quit [Remote host closed the connection]
gouchi has joined #dri-devel
tzimmermann has quit [Quit: Leaving]
ngcortes has joined #dri-devel
mbrost has joined #dri-devel
gouchi has quit [Remote host closed the connection]
sarnex has joined #dri-devel
gpoo has quit [Quit: Leaving]
gpoo has joined #dri-devel
gouchi has joined #dri-devel
yk has joined #dri-devel
pcercuei has quit [Quit: bbl]
gouchi has quit [Ping timeout: 480 seconds]
gouchi has joined #dri-devel
hanetzer has joined #dri-devel
mlankhorst has quit [Ping timeout: 480 seconds]
Ryback_ has quit [Quit: leaving]
ngcortes has quit [Remote host closed the connection]
sdutt has joined #dri-devel
user12521 has joined #dri-devel
user12521 has left #dri-devel [#dri-devel]
user12521 has joined #dri-devel
<user12521>
agd5f: kernel 5.12.13 broke vega20. Fans roar loud, radeontop shows 100% gpu load when idle. reverting to 5.12.12 fixes the situation
sarnex has quit [Ping timeout: 480 seconds]
lynxeye has quit [Quit: Leaving.]
alyssa has left #dri-devel [#dri-devel]
Ryback_ has joined #dri-devel
Ryback_ has quit []
bcarvalho has quit [Remote host closed the connection]
gouchi has quit [Read error: Connection reset by peer]
gouchi_ has quit [Remote host closed the connection]
GloriousEggroll has quit [Quit: Death is life's way of telling you you've been fired.]
GloriousEggroll has joined #dri-devel
pnowack has quit [Quit: pnowack]
Duke`` has quit [Ping timeout: 480 seconds]
<cmarcelo>
jekstrand: should we treat sampler & sampled_image also as nir_var_mem_image?
andrey-konovalov has quit [Ping timeout: 480 seconds]
danvet has quit [Ping timeout: 480 seconds]
user12521 has quit [Quit: Konversation terminated!]
craftyguy has quit [Quit: craftyguy]
craftyguy has joined #dri-devel
<cmarcelo>
alyssa: jekstrand: wondering if for nir_var_mem_image MR, I should change the passes/drivers to support images in both nir_var_uniform | nir_var_mem_image, then flip the producers, and in another patch remove nir_var_uniform image handling...
gpoo has quit [Ping timeout: 480 seconds]
gpoo has joined #dri-devel
bcarvalho has joined #dri-devel
bcarvalho has quit []
<jekstrand>
cmarcelo: I don't think you care about sampled images
<jekstrand>
cmarcelo: which I think actually means sampled images still go in uniform and not image
<jekstrand>
But it shouldn't matter, effectively
<jenatali>
Caio Oliveira: I was under the impression that the image itself was still going to be uniform, but any pointers into the image data were going to be mem_image
mbrost has quit [Remote host closed the connection]
zf has joined #dri-devel
gpoo has joined #dri-devel
iive has quit []
reinist12 has joined #dri-devel
<jenatali>
jekstrand: We're seeing a get_ssbo_size() call that takes vulkan_resource_index's result as a source... unless we set use_deref_buffer_array_length, in which case after lower_explicit_io, get_ssbo_size() takes load_vulkan_descriptor as a source
<jenatali>
Is that expected? Seems... kinda weird
pcercuei has quit [Quit: dodo]
<jekstrand>
SSBO lengths are weird
<jenatali>
Somehow I thought there was going to be more to that thought :)
jewins has joined #dri-devel
<zf>
so I understand it's a well-known fact that nouveau has multithreading problems. is it understood what specifically these threading problems are? is there any advice one could hypothetically give to a new contributor interested in fixing them?