02:08 <brads> narmstrong: on Odroid N2 still having the same issues after disabling the power regulator using GPIOH_6, also tried disabling GPIOH_5. The board was booting a few weeks ago on amlogic/v5.1/g12b branch and looks like pinctrl might have changed a bit?

02:11 <brads> I will try to go back and see what patch causes it to break

03:29 <brads> Its working on the amlogic/v5.1/g12b branch for me here is the log https://pastebin.com/C7u4qAJR I think the MDIO needs to be added on top of defconfig for ethernet

03:31 <brads> something in g12b-odroid-n2 seems to be the cause, using g12b branch the next line in the log after the problem is [ 1.652063] g12a_ao_cec

05:52 <brads> narmstrong: It seems to be either a gitlab sync issue or me ;) The g12b branch is working fine with emmc, network & USB3.

05:58 <memeka> brads: cool

05:58 <memeka> is hdmi working?

06:28 <brads> memeka: yes hdmi console is, I was about to start adding some display packages

06:35 <brads> glmark complains initialising and glmark-drm does too then seqfaults. Will do some more testing

06:48 <memeka> brads: can you show me a modetest output?

07:11 <brads> memeka: gnome works, firefox & youtube. What is the easiest way to get a modetest?

07:24 <memeka> brads: modetest is a binary in libdrm package

07:24 <memeka> Gnome on wayland or X11 no accel?

07:30 <brads> no accel

07:33 <narmstrong> brads: interesting, i’ll retest the branch next week

07:46 <brads> memeka, there is no modetest on ubuntu bionic with libdrm[-dev] installed I can find, I can try to compile it from source.

07:53 <memeka> brads: I know you gotta compile source :(

07:58 <brads> memeka: gnome in wayland mode is slower than gnome on X from what I can tell but there is no accel on either. Hardkernel mentioned if DRM gets running then they/ amlogic will release a wayland driver.

07:59 <brads> looks to be running now for me :)

07:59 <memeka> brads: yeah slow wo mali driver :)

08:00 <memeka> brads: also you won’t get this https://i.imgur.com/sl1qwmA.png

08:00 <brads> its much better than c2 but yes slow

08:00 <memeka> (Screenshot on xu4)

08:07 <memeka> brads: what’s major and missing atm? :)

08:18 <brads> memeka: not much if you want the basics, emmc, usb, ethernet, HDMI display. Will try an sd card and some usb storage next.

16:36 <ndufresne> maybe someone here knows, but for IR keytables, should they be submitted to udev project ?

16:37 <ndufresne> ah, apparently they mostly come from v4l2-utils

16:37 <ndufresne> but they seem to be in toml

17:14 <xdarklight> ndufresne: some also go here: https://github.com/torvalds/linux/blob/master/drivers/media/rc/keymaps/rc-geekbox.c (I submitted that one like 2 years ago and it was accepted). I'm not sure about the "correct place" though

18:17 <ndufresne> hmm, I see, and any rules on wheter they should be files or code ?

18:18 <ndufresne> at the same time, you are not locked to one remote control, it just happen that there is one shipped with these devices

18:18 <ndufresne> \o/

18:19 <ndufresne> another thing I didn't find is how to properly map the power and the tv_power

18:20 <xdarklight> ndufresne: I believe the .toml files in v4l-utils are generated based on the in-kernel keymaps. that's also what this commit message suggests: https://git.linuxtv.org/v4l-utils.git/commit/utils/keytable/rc_keymaps?id=f3d77d6df975b6fb8fbf9f9f8fe2c4a809136b86

18:21 <ndufresne> interesting, there must be a reason to maintain both ... or simply that this allow using newer maps on older kernel ?

18:21 <xdarklight> ndufresne: you can select the in-kernel keymap in the .dtb using a property (see https://github.com/torvalds/linux/blob/master/Documentation/devicetree/bindings/media/meson-ir.txt)

18:23 <xdarklight> I don't know why they're maintained in two places though

18:23 <ndufresne> ok, so now I got a motivation to submit a dtb or this alfawise h96proplus box

18:23 <xdarklight> (maybe re-mapping of buttons easily is a use-case that people want?)

18:24 <ndufresne> I must admit it was handy to have a large set of example mapping, comparing against similar devices to get mine right

18:24 <ndufresne> what I didn't find though is a complete list of KEY_ value that would be accepted

18:24 <ndufresne> maybe that's in the kernel in the end

18:25 <ndufresne> thanks for all the pointer btw

18:25 <xdarklight> I used this when I wrote the geekbox keymap: https://elixir.bootlin.com/linux/v5.1-rc1/source/include/uapi/linux/input-event-codes.h

18:27 <ndufresne> lol, some key names are ... interesting, KEY_ZENKAKUHANKAKU

18:30 <xdarklight> :D