04:01 <wens> I see patches were sent for H616 Mali

06:09 <jernej> loki666: Your GPU patches look good. Only nitpick would be that compatible should probably be alphabetically ordered.

06:09 <jernej> Note that my test/hack patches are taking shortcuts and are not fully done according to kernel code style :)

06:12 <jernej> which is pretty obvious in H616 DE3.3 series...

06:15 <jernej> apritzel, wens: DE3.3 is again very similar to DE3. Biggest novelty is that now planes are not part of the mixer anymore. There is separate region where reside now and can be associated either with mixer0 or mixer1. This is not yet modeled in latest DE3.3 series. Second big feature is DMA for registers. Idea is to have register values in some memory and at vblank they get copied in HW registers. That's not implemented yet.

07:11 <tokyovigilante> jernej: do you think I should pull out those magic values? I haven't actually tried without them to see if they are required. Any other suggestions to try and get the DE33 work across the line welcome, given that it seems about as minimal as it can be in terms of functionality now.

07:13 <tokyovigilante> sorry I have not had much time to be responsive to feedback on list, will also add the regmap descriptions as suggested by Rob and apritzel

11:37 <junari> I can't figure out how I can combine two cpu clusters with two temperature sensors

11:38 <junari> I did something like this https://elixir.bootlin.com/linux/v6.13.6/source/arch/arm/boot/dts/allwinner/sun8i-a83t.dtsi#L1207

11:39 <junari> But htop does not show the temperature https://drive.google.com/file/d/1pyiLtG8rcm-lKiD4jh3Iq4L1W60iTthW/view?usp=drive_link

11:48 <junari> Also, what are your calibration values for the temperature sensors in the A523 and co? I think mine don't seem to contain any values or I'm looking at the wrong register

16:55 <kuba2k2> hello, are there any known issues with the sunxi MUSB peripheral/gadget driver? I can only get it to work if connected to an USB 1.1 hub

16:56 <kuba2k2> everything else (USB 2.0/3.0, etc) fails to retrieve the hardware IDs (testing on Windows, on two PCs, with different cables, with or without USB 2.0 hubs)

16:57 <kuba2k2> this is on a D1s/F133, kernel configured as per sunxi wiki USBGadget/Ethernet

16:58 <kuba2k2> in DT, ehci0 and ohci0 are disabled, as this is the only way I can get it to switch to peripheral mode at all

17:06 <apritzel> this last thing is a known issue, loki666 sent some patches a few weeks ago: https://lore.kernel.org/linux-sunxi/20250118102207.9339-1-simons.philippe@gmail.com/T/#u

17:09 <kuba2k2> that one is not a big deal, I don't need to switch roles on runtime

17:22 <kuba2k2> oh, nevermind, it seems like I can't even negotiate high speed (480 Mbps) in host mode, not even on BSP... so this must be a hardware issue then :(

17:36 <kuba2k2> oddly enough, USB drives and hubs can't negotiate 480Mb/s, but every USB Wi-Fi card I tested can..

17:37 <jakllsch> odd, i this is bringing back memories of single ended zeros or something to me :-P

17:41 <kuba2k2> granted, this is a custom hardware design, but come on... USB2.0 works on, like, any piece of wire pretty much

17:43 <jakllsch> sometimes the gadget/otg driver needs to be told what speed it will operate at, iirc?

17:44 <kuba2k2> yeah I thought the gadget part was the issue, but if it can't even get 480mbps in *host* mode (device descriptor read/64, error -71), then I think I have bigger problems

17:45 <jakllsch> ah, perhaps

18:04 <kuba2k2> well, the issue was... TVS diodes. Removed them from the PCB, and now it works like a charm

18:05 <jakllsch> figures it'd be something like that

21:11 <jernej> tokyovigilante: yes, those numbers don't belong here. It is just static plane allocation for mixer0. I have a plan to work on "plane" driver to see how it would turn out in the next week or so, if you don't mind.