07:04 fdobridge: <d​jdeath3483> Interesting, I'm finishing some push constant rework too
11:38 Ilgaz: karolherbst: I bet you guessed it was me reporting the bug :-) https://lore.kernel.org/all/20240517110853.8481-1-tiwai@suse.de/
11:39 Ilgaz: Happy weekend.
16:01 fdobridge: <M​isyl with Max-Q Design> The default mode on my display does nto work with nouveau still
16:01 fdobridge: <M​isyl with Max-Q Design> and I have to launch with nomodeset or I can't even use tty
16:10 fdobridge: <r​edsheep> Does that default mode have reduced blanking?
16:11 fdobridge: <M​isyl with Max-Q Design> I have no idea
16:11 fdobridge: <M​isyl with Max-Q Design> https://cdn.discordapp.com/attachments/1034184951790305330/1241422906446053456/message.txt?ex=664a24ad&is=6648d32d&hm=7865dd4b1a024ecd838aeddf305c91bda40091a6d34d3e6907f879ff6230ecde&
16:12 fdobridge: <b​abblebones> Does the kernel complain anything if you ssh in?
16:13 fdobridge: <M​isyl with Max-Q Design> no
16:14 fdobridge: <b​abblebones> ahhh, that's not great
16:14 fdobridge: <b​abblebones>
16:14 fdobridge: <b​abblebones> No other lesser displays on hand to float until bug addressed?
16:14 fdobridge: <M​isyl with Max-Q Design> i have plenty of displays
16:14 fdobridge: <M​isyl with Max-Q Design> not a blocker to try stuff for me
17:00 fdobridge: <r​edsheep> The best mode here would require DSC, especially if it actually runs at 10 bit which I espect it wouldn't
17:00 fdobridge: <M​isyl with Max-Q Design> That isn't the default mode though
17:01 fdobridge: <M​isyl with Max-Q Design> The preferred mode is 3840x2160 @ 59.99Hz
17:01 fdobridge: <M​isyl with Max-Q Design> which shouldn't need DSC at all
17:07 fdobridge: <r​edsheep> Yeah for sure 60 should not need anything special. Am I missing something saying it is default? Sorry this is not how I normally work with these, much more used to this on Windows.
17:08 fdobridge: <r​edsheep> If I am parsing this right it does strangely seem like that 60hz mode is doing really unnecessarily tight timings, it would be closest to a cvt-rb mode afaict, so yes reduced blanking
17:10 fdobridge: <r​edsheep> I don't think nouveau does reduced blanking right now at all, though it's still not really clear to me why as the code doesn't really seem like it should produce that result. Not one of the kernel developers though.
17:11 fdobridge: <r​edsheep> At least that display isn't going anywhere near as tight as a cvt-rbv2 mode. Still, very weird. Seems like a mistake with the EDID to be honest.
17:18 fdobridge: <r​edsheep> I see now where it says that DTD-1 is default. Yeah that is almost exactly a CVT-RB mode, aside from shifting some from vback to vfront
17:18 fdobridge: <M​isyl with Max-Q Design> what is reduced blanking and why would a device need to know anything? I am not aware of any specific handling like that on other devices?
17:19 fdobridge: <r​edsheep> Reduced blanking literally just means having the timings on the mode tighter than CVT standard
17:20 fdobridge: <r​edsheep> Having tightened timings for a 60hz mode is a bizarre choice as the only purpose it could serve is really aggressively trying to lower the pixel clock
17:21 fdobridge: <M​isyl with Max-Q Design> Many OLED displays only operate at a fixed pixel clock and just grow the vfp
17:22 fdobridge: <r​edsheep> This isn't oled though, right?
17:22 fdobridge: <M​isyl with Max-Q Design> Correct, but I don't think it's weird that it isn't CVT
17:23 fdobridge: <r​edsheep> In any case something needs to get fixed for more modes to work, that's for sure. My displays don't work at their best modes.
17:24 fdobridge: <r​edsheep> Weird question but if you force through the default mode from inside an x11 session with xrandr does it work?
17:24 fdobridge: <r​edsheep> I can make my cvt-rbv2 monitor work doing that, which, again seems really weird and I still haven't figured out why that works yet.
17:25 fdobridge: <r​edsheep> You've got more insight into the kernel code of course, if you found the bug that would be amazing
17:39 fdobridge: <k​arolherbst🐧🦀> anyway, HBR3 is supported and it might just be a nouveau bug
17:40 fdobridge: <k​arolherbst🐧🦀> and the kernel log should tell you what's wrong
17:42 fdobridge: <k​arolherbst🐧🦀> @prop_energy_ball would have to fetch kernel logs to know what's going on there. But it's also hard to parse what's not working in your case anyway. Is it the mode not showing up and userspace doing silly things, or the mode is there but just not working, or....
17:42 fdobridge: <r​edsheep> You were saying there's a kernel parameter to get more debug for this right?
17:43 fdobridge: <k​arolherbst🐧🦀> it depends on what's going on
17:45 fdobridge: <r​edsheep> Have you had a chance to look at DSC or do you know if anyone else has tried?
17:49 fdobridge: <k​arolherbst🐧🦀> I don't know
18:08 fdobridge: <M​isyl with Max-Q Design> Not even userspace, just tty doesnt even work
18:09 fdobridge: <M​isyl with Max-Q Design> Not even userspace, tty doesnt even work (edited)
18:09 fdobridge: <M​isyl with Max-Q Design> I can get logs if you give me the right params
18:09 fdobridge: <M​isyl with Max-Q Design> The symptom is just a black screen
18:25 fdobridge: <k​arolherbst🐧🦀> just dmesg from the boot would be enough at this point