06:08 Lynne: zzoon: also, h264 streams with 16 refs are broken on intel - https://files.lynne.ee/testsamples/tractor_main_100gop.mp4
06:09 Lynne: I suggest running through the entire https://files.lynne.ee/testsamples/ set to find issues, these are all the files we used while writing radv
06:10 zzoon[m]: ok thanks.
07:49 glehmann: dj-death: does intel have dynamic state that doesn't map to hw? or api stages that use different hw shader stages depending on the other active stages?
08:12 dj-death: glehmann: yeah there are dependencies between shader stages, some can be avoided
08:13 dj-death: glehmann: Wa_18019110168 on DG2 is one of the item that might make this unimplementable
09:48 dolphin: airlied, sima: drm-intel-gt-next PR in your inbox
10:11 mripard: sima: airlied: could you give a look at https://lore.kernel.org/all/20241204134410.1161769-1-dev@lankhorst.se/ ?
10:12 mripard: It looks like we're at the point where we're discussing merging it, so having your thoughts on this would be great
10:12 sima: mripard, I did look, seemed reasonable
10:13 sima: mripard, I think if you get the acks for the amgpu/xe/ttm parts it should be more than enough to land it all
10:22 mripard: sima: awesome :)
10:25 sima: mripard, imo doesn't need an explicit ack from airlied/me beyond that we'll forward it in the -next pr train
13:11 mripard: sima: ack, thanks :)
17:20 Lynne: why is terminate_invocation fragment shader only?
17:23 Lynne: scratch that, I don't want to know, its either that they forgot about compute because no one uses it, or some vendor's 20 year old card doesn't support it in that one niche case no one cares about (twist: its supported, just that no one checked)
17:24 Lynne: at the end of the day it's a useless extension by design
17:27 glehmann: I don't get what you are asking for, in compute you can just return?
17:28 glehmann: unlike fragment shaders, there is nothing that happen at the end that can only be prevented with terminate/demote/discard
17:35 againwithdr: I am with no relation with liars other than getting to stop the lies spread.. What fecalists like you do with their life is your own personal thing, and even if i was wrong which definitely was not the case with my theories of POC, you are not allowed to choose none the less " random victims" to be bombed and accused for your own other faults, so my initiative is only not allowing you
17:35 againwithdr: to lie, because i know where it leads, your parasite feel like they have rights to disturb me everywhere, and it's double no in all ways. I never get into other arguments with you, as you are not my allies. The methods which i promised some continuous mistake fixing but still some mistakes anf fuzzings - are in text whitelogger loged which is great logger and has been available for
17:35 againwithdr: couple days reading, it's cloud backing of mine. The engineer was me, and that does not permit you to lie anymore to cause other trouble to people such as me and my friends. you can say your opinions into the toilet sink about schizophrenia soon enough for the rest of your life, apart from the freaks here, no one will listen to you, and even more since i determined no mental illness, no
17:35 againwithdr: voices and did not stay in mental facility volunteerily they took me with court of law order which was just garbage they assembled together, the problems in my poor performance in life came due to abuse to my body by doctors however, and it affects the mentality few but physical health alot. Those have always been my words, and never changed, and no one doubts me in it, it's diagnosis was
17:35 againwithdr: wrong, however the social program settlement for me was done correctly. After such terror they need to pay me money too ro things in the world would not work out with such banana laws and people get abused like this all the time, so there is no harm done by changing the diagnosis to really correct version, which was depression due to physical injuries caused by doctors deliberately.
23:58 HdkR: Canonical actually never backported that libllvm18 fix from Oracular to Noble? gah.