00:01 <valpackett> or, well, use DSC (what's missing for it?? https://lore.kernel.org/all/20250918-v6-16-rc2-quad-pipe-upstream-4-v16-10-ff6232e3472f@linaro.org/ this?)

00:02 <angeloverlain[m]> how did you find the monitor?

00:07 <valpackett> was it supposed to be hard to find? ebay..

00:21 <anthony25> it doesn't support DSC already?

00:22 <anthony25> with 4 lanes, 4K@120Hz works

01:19 <valpackett> oh oops i do have `data-lanes = <0 1>;` left over in my dts..

01:19 <HdkR> No one really needed those additional lanes anyway :D

01:21 <valpackett> oh no https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=458de584248a5630878ed11ea23188f6007036b2 went in *before* the thena dtsi, which is why it couldn't touch it

01:27 <valpackett> linux's merge-based development is so ass

02:36 <valpackett> oh welp 4 lanes + USB-2 doesn't work right now, i remember that..

02:40 <HdkR> Good to know that limitation.

02:56 <steev> steveej[m]: freekurt[m]: i just built the qcom-laptops branch (which should be the default) with a make defconfig; make qcom_laptops.config; make bindeb-pkg, then rebooted into it and bluetooth is working here; 6.17.0-rc7

02:57 <steev> bluetooth speaker connected and played audio, grabbing a mouse now

02:58 <steev> mouse and keyboard aren't working though

03:11 <valpackett> @konradybcio:matrix.org : re https://lore.kernel.org/all/f21b7d52-4c3f-4e5b-bee7-f8b2945b5b02@oss.qualcomm.com/ how would i go about fixing usb2 with 4-lane dp? got the perfect device to test now

03:12 <valpackett> also yeah right now USB2 dies completely on the port that was used for 4-lane DP, does not seem to recover

03:14 <valpackett> was interesting to learn that the entire e.g. keyboard→typec-hub→monitor→laptop chain is native usb2 all the way through and nothing ever muxes it onto the superspeed bus

03:25 <clover[m]> bluez tagged 5.84 like a week ago but havent "released" it yet. weird

03:26 <clover[m]> maybe they forgot? lol

03:53 <steev> it's not bluez

03:53 <steev> 6.16.9 with the same bluez setup works - i'm assuming it's either a kernel config option or a patch of some sort to 6.17

04:11 <valpackett> interestingly, on resume currently the qmp phy enables pipe_clk unconditionally..

04:38 <valpackett> i think i might have sort of fixed it? well, umm, my usb2 devices are working and dp_debug says num_lanes is 4. but gnome gives me 50 hz.. 0.o

04:38 <valpackett> * i think i might have sort of fixed it? well, umm, my usb2 devices are working and `dp_debug` says `num_lanes` is 4. but gnome gives me 50 hz.. 0.o

04:40 <valpackett> not sure what konrad meant by "the hardware disagrees" with keeping the usb pipe pll on in dp only mode, i couldn't find any actual hw registers related to the pll pipe clksrc which is defined as a fixed rate in phy_pipe_clk_register

04:40 <valpackett> but moving the enables for pipe_clk from the usb part to the common part got me this far xD

04:41 <valpackett> * rate in `phy_pipe_clk_register, * phy_pipe_clk_register`

05:14 <valpackett> rebooting without unplugging the cable though, and usb2 is dead.. nah so that's still a lot of weirdness in any case

05:39 <valpackett> sooooo the 50 Hz is because the crtc rejects 60 Hz due to `max_core_clk_rate 575000000 < adjusted_mode_clk 663337000`

05:44 <valpackett> https://lore.kernel.org/all/20250922-modeclk-fix-v1-1-a9b8faec74f8@oss.qualcomm.com/ heeeeere we go, i see

05:45 <valpackett> or rather, https://lore.kernel.org/all/20250923-modeclk-fix-v2-1-01fcd0b2465a@oss.qualcomm.com/ is the v2

06:03 <valpackett> also i'm seeing a very interesting bug when trying the docked lid-closed mode.. while the lid is closed, the main plane of the external monitor is frozen, but the cursor is fine and keeps updating

06:45 <jdb[m]> What is 3D Merge in this context? I've never heard of it

07:23 <valpackett> very internal display pipeline stuff

09:46 <valpackett> PATCH for 4-lane DP + USB 2.0: https://lore.kernel.org/all/20250927093915.45124-2-val@packett.cool/ Please test on your external monitor setups :)

09:47 <valpackett> Particularly, if you have a 4-lane monitor, check if connecting it kills USB 2.0 functionality on that Type-C port (unplug monitor, plug in legacy device like a keyboard), and if that patch resolves the issue

11:43 <angeryboi[m]> Hmmm, is 5ghz being broken a known issue? WPA3 works if I manually set it as preferred per network, but 5ghz doesn't work at all.

12:17 <angeryboi[m]> Apparently gpu acceleration is completely broken in snap browser packages(brave and firefox) as even using --use-angle=gles doesn't resolve it.

12:17 <angeryboi[m]> But troubleshooting is super annoying on ubuntu when it overrides apt commands as snaps for no reason 🫠

12:31 <alexVinarskis[m]> both 5ghz and 6ghz work for well for me, x1p with wcn6855 and x1e with wcn7850. x1p sometimes has trouble connecting to 6ghz WPA3, connect&reconnect fixes it.

12:34 <alexVinarskis[m]> for snap browser packages, why use snap? :)

12:34 <alexVinarskis[m]> firefox, zen-browser from Flatpak work oob

12:35 <alexVinarskis[m]> Haven't ever tried `--use-angle=gles` actually, im running vscode and others with `--enable-features=UseOzonePlatform,WaylandWindowDecorations --ozone-platform=wayland` as that fixes blurry electron apps when using wayland+fractional scaling. Appears to also fix the artifacts due to broken gpu acceleration.

12:37 <alexVinarskis[m]> iirc there were some funny things with wifi firmware, perhaps try switching? in my case both laptops are on linux next from 8th September, x1p has wifi firmware from 2024-10-17, x1e from 2023-12-06.

12:57 <angeryboi[m]> <alexVinarskis[m]> "Haven't ever tried `--use-angle..." <- Just tried these flags, text doesn't render in brave://gpu page and scrollbar also draws continuously over itself

13:02 <angeryboi[m]> I get a spam of following error

13:02 <angeryboi[m]> 4361:4384:0927/155659.989304:ERROR:gpu/command_buffer/service/shared_image/shared_image_manager.cc:521] SharedImageManager::ProduceMemory: Trying to Produce a Memory representation from a non-existent mailbox.

13:02 <angeryboi[m]> Regardless of the launch flags used so this could ne either brave issue or snap issue overall

13:09 <\[m]> maybe I should cut down on all the electron bloatware but comfort is nice -_- I don't know if mine makes more fan noise or that I'm overly sensitive to the noise of fans but it seems that you can't and will never be able to change fan curves on arm? it's firmware controlled?

13:24 <\[m]> it's in the dtb?

13:47 <anthony25> it seems EC controlled

17:45 <\[m]> seems tb dock doesn't work, usb c fallback fails

18:41 <alexVinarskis[m]> Are you trying to test display via dock? Not all thunderbolt docks can do dp alt mode fallback, some are thunderbolt only. If it can (like eg. Dell wd19tb/wd22tb4) it will not fallback automatically - you need to go to laptops bios and disable thunderbolt support, if its a thing. On Asus Zenbook A14 thats not possible, so universal tb/dp alt mode dock wont work until thunderbolt support lands. On Dell you can do it, i

18:41 <alexVinarskis[m]> disabled tb support on my XPS 13, and it now works with my WD19TB in dp alt mode.

19:46 <steev> angeryboi[m]: try passing force_gl_vendor=angleisbroken

19:46 <steev> well, i dunno about firefox, but brave is a chromium package and angle is broken

19:47 <steev> with freedreno/msm

19:48 <steev> environment variable, not a -- option

20:03 <steveej[m]> steev: thanks for trying it out. i'll downgrade to rc5 tomorrow to see if it's better there

21:15 <angeryboi[m]> <steev> "angeryboi: try passing force_gl_..." <- GPU still crashes on launch

21:16 <angeryboi[m]> It complains about rgba_8888 formats for wayland it seems

21:16 <steev> weird, i'm gnome on wayland here without issues with firefox

21:19 <angeryboi[m]> Firefox is also software only rn

21:19 <angeryboi[m]> And windows only has msedge with working gpu accel, so not rly better on that side either.

21:29 <steev> what system?