<macc24>
ankhhikes: minetest was running at some point of time on duet
<macc24>
ugh, aformat
<macc24>
ankh\hikes: minetest was running at some point of time on duet
<macc24>
ankh_hikes: ^
<ankh_hikes>
oh
<macc24>
that's a regression
alpernebbi has quit [Remote host closed the connection]
hexdump0815 has joined ##panfrost-offtopic
ankh_hikes has quit [Quit: Connection closed]
<hexdump0815>
macc24: not sure if you know that already - i found what breaks the display on duet for v5.10.4+ by bisecting between v5.10.3 and v5.10.4 and ended up with https://github.com/gregkh/linux/commit/414562b0ef36ce658f0ffec00e7039c7911e4cdc - i.e. as a result the corresponding aliases in the mt8183.dtsi need to be adjusted
<macc24>
thanks for taking a look at it, i was having a lot of trouble
<macc24>
hexdump0815: thank you, i'll steal, err, borrow those changes when i'm done with adding user interface selection to cadmium
<hexdump0815>
i'm also trying to build a kernel which will work for both krane and elm, but the kernel working on krane has a dark screen on elm ... there are two kernel options still different between my two working versions for elm and krane: CONFIG_MTK_INFRACFG set for elm and removed by the krane patches and CONFIG_DRM_VGEM used by krane on not on elm - so
<hexdump0815>
either one of them is a problem or the krane patches i guess
<macc24>
CONFIG_DRM_VGEM is just there, i'm too afraid to remove it lol
<hexdump0815>
there is no pressure - i just wanted to share this as you seemed to fight with the non working display on newer kernels
<hexdump0815>
i would not call it simple to have a 9 level kernel bisect, but the result is indeed rather trivial - i'm not sure if it will help with 5.11+ - not yet tested ... also the usb patches do not apply cleanly anymore for later v5.10.x kernels, but at least the duet is working with a current lts kernel somehow now :)
ankh_hikes has joined ##panfrost-offtopic
<macc24>
hexdump0815: usb patches aren't needed for newer kernels
<macc24>
they were because mediatek messed up bandwidth calculation
<hexdump0815>
macc24: but i'm getting a few "xhci: Not enough bandwidth!" messages now, but they do not seem to harm
<macc24>
they do harm
<macc24>
usb sound card or usb mice don't work for example
<macc24>
hexdump0815: can you make a pull request?
<macc24>
for cadmium with newer kernel
<hexdump0815>
i need to clean up stuff more first - its too much of a mess still and to be honest i'm still not understanding what most of the patches and options are doing
<macc24>
it's 2 lines
<hexdump0815>
btw. is there any communication on the mediatek channel going on? i checked a few times but it was silent all the time ... are you aware of anyone working actively on mt8183 support?
<hexdump0815>
ah for those two line i can do a pull request - sure
<macc24>
hexdump0815: eballetbo is doing mt8183
<hexdump0815>
macc24: yes - i also tried his mainline following tree, but never got it to work on the duet ... maybe i should check the dts in there too :)
<macc24>
hexdump0815: his tree works fine on topic/chromeos/somewhat-stable-5.11 branch or something
<macc24>
maybe i should make a channel for cadmium to not spam ##panfrost-offtopic hmm
hexdump0815 has quit [Quit: Connection closed]
enty has joined ##panfrost-offtopic
ente has quit [Read error: Connection reset by peer]