00:16 redsheep[d]: skeggsb9778[d]: That branch seems to have very garbled display output. Do I need a different gsp.bin or something?
00:17 redsheep[d]: Oh if I don't use zink it isn't garbled
00:18 redsheep[d]: And still no audio. So there's a difference between ad104 and ad102? That seems really strange
00:19 redsheep[d]: That, or maybe the monitor itself matters?
00:23 redsheep[d]: Or maybe the desktop environment, if that's even possible for userspace stuff to break it? I'm on plasma. For good measure I tested all the display port outputs, same result
00:31 redsheep[d]: Hmm. I wonder if maybe 535 gsp had audio issues? The audio stuff in the kernel doesn't really look complicated enough to justify different dies from the same generation having different results
00:35 redsheep[d]: skeggsb9778[d]: Did you leave the 550 port lying around somewhere I can grab to test? I suppose I could also revert the prop driver on my system all the way back to 535 if not
01:01 redsheep[d]: https://cdn.discordapp.com/attachments/1034184951790305330/1256051842408775841/fail.txt?ex=667f5cef&is=667e0b6f&hm=4704f1674386ec5b26543cb140446f2e2b5fb5b44b98c54cacc81ab8ee3a30e7&
01:01 redsheep[d]: Uggh the 535 open driver doesn't build anymore
01:12 rinlovesyou[d]: yeah no there were some kernel changes so
01:12 rinlovesyou[d]: old drivers are pretty dead
01:13 rinlovesyou[d]: there was a good period where nvidia was just not working at all
04:58 tiredchiku[d]: redsheep[d]: pkgbuild remains unchanged for 6.10-rc5
05:48 asdqueerfromeu[d]: redsheep[d]: 6.9.6
05:50 asdqueerfromeu[d]: tiredchiku[d]: Do you have working HDMI/DP audio?
05:51 redsheep[d]: My HDMI audio does work, just to be clear
06:00 asdqueerfromeu[d]: redsheep[d]: So it's only DP audio that's broken (then I guess my laptop does use DisplayPort alt mode)?
06:02 tiredchiku[d]: asdqueerfromeu[d]: hdmi audio works for me, yes
06:34 redsheep[d]: asdqueerfromeu[d]: I am pretty sure most of the time a type c dock with an hdmi port is still using dp alt mode, yeah
16:16 ahuillet[d]: skeggsb9778[d]: is it known that RTX3080 fails to resume from suspend with Nouveau? I get a black screen (didn't take time to dump logs, needed computer for sth else)
16:36 mohamexiety[d]: not just 3080, there were resume issues throughout the board. but I believe this was fixed recently?
16:46 redsheep[d]: That sounds a lot like what me and a couple friends have been seeing on openrm recently. Strange.
17:34 maba_kalox[d]: redsheep[d]: Do you have context restore enabled? My asrock mobo passes bios in ~5-10 seconds when this setting is enabled.
17:36 redsheep[d]: maba_kalox[d]: ASRock and gigabyte have managed some really good results with memory with am5, Asus... Hasn't. My $700 motherboard cannot do both MCR and high speed memory at all, even with the latest bios.
17:37 maba_kalox[d]: redsheep[d]: Ups, understood, then sorry for useless suggestion.
17:37 redsheep[d]: No worries
17:37 redsheep[d]: It's the right idea but for whatever reason it's just broken
17:39 redsheep[d]: It's a sad state, and makes me question buying asus again.
18:11 RSpliet: Always had mixed feelings about Asus. Their UEFI support tends to be broken in some ways that WMI drivers always had to hack around
18:11 RSpliet: but on the plus side, every time I bought a laptop from theirs in the past 15 years, I could get a refund for the Windows license
18:12 RSpliet: other than that they seem to not engage with the Linux community at all
18:53 kuter: Hello, I wanted to share something I had been working on before the weekend. By carefully fuzzing nvdisasm I am automatically generating ISA encoding specs for Nvidia SM90a. Each instruction goes through several disambiguation and enumeration steps. https://kuterdinel.com/sm90a_wip.html
18:55 kuter: It's also possible to find which registers are being read/written to by analyzing the output of `--print-life-ranges` mode of nvdisasm but I haven't integrated that to the html output yet.
18:59 kuter: the html I sent is missing quite a lot of instructions, my instruction enumeration is not perfect(yet) and also there are some instructions that
18:59 kuter: require special parsing.
20:51 ahuillet[d]: redsheep[d]: resume issues? yes, I've also been getting them with the blob since very recently
21:14 redsheep[d]: ahuillet[d]: Exactly, black screen on resume
21:16 redsheep[d]: I haven't retested on 555.58 but 555.42 had it