c0rnelius | tuxd3v__: Yeah I don't know. I'd be happy you got the wifi working. | 00:08 |
---|---|---|
c0rnelius | Except for mild experimenting on the NEO Plus2 I rarely bother really trying to make it work anymore. | 00:10 |
c0rnelius | Personally I think something needs to be patched kernel side. | 00:10 |
c0rnelius | Example on the NEO plus2 - https://github.com/pyavitz/debian-image-builder/blob/feature/patches/allwinner/nanopi/nano_pi_neo_plus2_bluetooth.disable | 00:12 |
c0rnelius | But it makes the board run unstable | 00:12 |
zmoment | Tenkawa: when unzipped, the devuan_beowulf_3.1.0_arm64_rpi3.img.zip created a 32010928128 bytes file. I unzipped it again and it generates an image of 1.9GB. Don't know what happened. | 00:18 |
zmoment | c0rnelius: the armv8 that you support are the ones in https://github.com/pyavitz/rpi-img-builder/releases/download/images/ ? Both the devuan and the others? | 00:24 |
tuxd3v__ | this should be in the minumum 1.1 volts | 00:24 |
tuxd3v__ | its for supply power to cpu | 00:24 |
tuxd3v__ | the max 1.3v | 00:24 |
tuxd3v__ | but there are a patch that has a rampdelay of 50 uv/yS | 00:25 |
tuxd3v__ | but there are a patch that has a rampdelay of 50 uv/uS | 00:25 |
c0rnelius | zmoment: yeah armv8 and armv6. I don't support the v7's kernel wise as I'm only one person :) | 00:25 |
tuxd3v__ | but looking at it it gives me in 4ms only o.2volts | 00:25 |
c0rnelius | At that would mean compiling things twice and blah blah and I just don't wanna deal with it. People can make their own though. | 00:26 |
zmoment | I'll use that armv8 for my RPi3B+. Thank you . | 00:31 |
tuxd3v__ | yeah it gives a lot of trouble to support all arm32, aarch32, then armv8, and some day who knows armv9.. I don't expect that soon.. | 00:32 |
zmoment | I'm used to Linux, not to debian/devuan. Last time I used debian was maybe sixteen or seventeen years ago. | 00:32 |
tuxd3v__ | zmoment, that is a long time :) | 00:32 |
tuxd3v__ | c0rnelius, the patch includes this: https://paste.debian.net/hidden/68ce7fcd/ | 00:33 |
tuxd3v__ | I am thinking in trying something there to see if I can get bluetooth power sequence right.. | 00:34 |
c0rnelius | tuxd3v__: https://github.com/armbian/build/blob/master/patch/kernel/archive/sunxi-5.10/board-h3-nanopi-neo-air-add-camera-wifi-bluetooth-otg.patch | 00:40 |
tuxd3v__ | c0rnelius, thanks :) | 00:44 |
c0rnelius | yep | 00:44 |
tuxd3v__ | I am using majority of things from there.. | 00:44 |
tuxd3v__ | but I am including in the uart3 the bluetooth inside, but by some strange factor it doesn't upload the patch and put bluetooth up | 00:45 |
tuxd3v__ | armbian on the otherside describe uart3, but doesn't put bluetooth inside.. II think they are doing later when system comep up: | 00:46 |
tuxd3v__ | hciattach /dev/serial3 bcm43xx 921600 noflow - $BDADDR | 00:47 |
tuxd3v__ | its another way.. | 00:47 |
c0rnelius | Yes this is like how it needs to be done on the a64 module | 00:47 |
tuxd3v__ | and probably one that works.. I need to cjeck their image | 00:47 |
c0rnelius | you need a helper script | 00:47 |
c0rnelius | What it would be for that board I don't know though? | 00:47 |
c0rnelius | for the a64 module its something like this - https://github.com/pyavitz/debian-image-builder/blob/feature/lib/function/allwinner#L581 | 00:48 |
tuxd3v__ | For bananapi m2 zero and for nanaopi neo Air, I am doing it in aother way: | 01:18 |
tuxd3v__ | this case if for nanopi nei Air: https://paste.debian.net/hidden/f26c5ea2/ | 01:18 |
tuxd3v__ | it in theory should load the module automatically without any helper | 01:18 |
tuxd3v__ | it works fine in bananapi m2 zero | 01:18 |
tuxd3v__ | but not in neo Air, only maybe 1 work in 100 tries.. | 01:19 |
tuxd3v__ | The other wau to do it, is describing only uart3, and then do like you say, I need yet to experiment with that version to see if I get bluetooth reliably :) | 01:20 |
tuxd3v__ | its a option indeed :) | 01:20 |
tuxd3v__ | c0rnelius, I am with a kernel that doesn't have bluetooth defines in DT, only partially for rfkill | 05:21 |
tuxd3v__ | but the think is.. | 05:22 |
tuxd3v__ | ttyS3 does indeed apear now under /dev/ttyS3 | 05:22 |
tuxd3v__ | but the device gives timout.. | 05:22 |
tuxd3v__ | its also impossible to reset the device | 05:23 |
tuxd3v__ | 2 diferent ways: | 05:23 |
tuxd3v__ | https://paste.debian.net/hidden/e5e95419/ | 05:44 |
tuxd3v__ | There seems to be "something in the water".. | 05:44 |
tuxd3v__ | do you remember that patch... yeah tomorow I will try it in the bluetooth tools | 05:45 |
tuxd3v__ | https://gitea.devuan.org/tuxd3v/bluez/src/branch/master/debian/patches/cypress_bluetooth_broadcomm.patch | 05:45 |
tuxd3v__ | but in this case I need to better look at it.. | 05:46 |
tuxd3v__ | because the device in case has diferent vendor id.. | 05:46 |
tuxd3v__ | for what I inderstand the protocol used should be 'HCI_UART_H4', but I don't know if the system knows how to reset the device.. | 05:50 |
tuxd3v__ | the kconfig is 'CONFIG_BT_HCIUART_BCM' | 05:55 |
tuxd3v__ | ShorTie, hold yourself tight :) | 06:34 |
tuxd3v | something in the water during the night.. | 18:43 |
tuxd3v | left my system compiling a new kernel, and wen I wake up system have rebooted :S | 18:43 |
tuxd3v | I loged in, and I was in a diferent operating system..crazy thing.. | 18:44 |
tuxd3v | probably my cat during the night took my keyboard as a piano | 18:44 |
tuxd3v | :S | 18:44 |
Tenkawa | fmha | 18:47 |
Tenkawa | er ha | 18:47 |
Tenkawa | sounds like my cats | 18:47 |
tuxd3v | yeah, almost for sure he made his kungfu on my keyboard.. | 18:47 |
tuxd3v | Tenkawa, c0rnelius I believe bluetooth is being activated before fylesystem.. | 19:56 |
tuxd3v | filesystem* | 19:56 |
tuxd3v | and at that phase there are no patch to load to bluetooth device.. | 19:57 |
tuxd3v | there are not yet a filesystem to load patch from.. | 19:57 |
tuxd3v | now, I can sort that out, maybe loading the firmware for bluetooth in the kernel | 19:58 |
tuxd3v | in that way it got loaded after uboot.. | 19:58 |
Tenkawa | tuxd3v: doesn't work that way | 20:00 |
Tenkawa | it either all gets loaded at kernel load time, at module load time, r at initrd load time | 20:01 |
Tenkawa | s/r/or | 20:01 |
tuxd3v | kernel is loaded in memory first in the uboot process.. | 20:01 |
Tenkawa | not the userland | 20:01 |
tuxd3v | it then will parse the device tree | 20:01 |
Tenkawa | no | 20:01 |
tuxd3v | but at that time there are not yet a fylesystem in place to load the firmware patch | 20:02 |
Tenkawa | userland for the bt is only activated once needd | 20:02 |
Tenkawa | and you havent made a call to it yet | 20:02 |
tuxd3v | in my case is automatically loaded via device tree:https://paste.debian.net/hidden/f14b66ee/ | 20:05 |
tuxd3v | I could be wrong and it has nothing todo with it, but its a possibility | 20:05 |
tuxd3v | I retired a lot of bluetooth drivers from kernel | 20:06 |
tuxd3v | it has giving support to virtually any tipe of device availlable, and I only have one fixed device, so I checked dependencies for my device and disabled other kernel modules.. | 20:06 |
tuxd3v | next step, load the firware in kernel | 20:07 |
tuxd3v | firmware* | 20:07 |
tuxd3v | to try to validate my theory | 20:07 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!