pocosparc

Where do we stand right now with the mainline kernels. What works and what doesn't? I was a bit busy lately with everything and have set aside this project. I still need to fix that openVFD driver. I the latest trunk already stable enough to reflash it to eMMC?

Thx. I am a few days off for a small vecation. Will take a look

I made OpenVFD work. There are some driver changes needed to get it to work. I commited everyting into ArmBoardBringUp git repo with the tutorial on how to get it to work README.txt. There are still some minor tweaks needed as currently the time is upside down (conf file changes) and also the wrong icons are on (also inversed). I will play with these after a few days, but you can try them out BTW Nice work on the video acceleration. IR remote codes are also there in order to be committed and compiled. Maybe KODI is next?

@Deoptim@maka I tried it your way and that is installing chromium via the snap - actually installed version 123 as it GitHub thread states that since version 121 all changes are mainstream. I also added these rules to 70-snap.chromium.rules: SUBSYSTEM=="video4linux", ATTR{name}=="rockchip,rk3568-vpu-dec", SYMLINK+="video-dec%n" SUBSYSTEM=="media", ATTR{model}=="hantro-vpu", SYMLINK+="media-dec%n" I tried to run chromium with a flag --disable_webgpu_shared_images=1 and the chrome://gpu still reports WebGPU and Vulkan as disabled. I am not completely sure if this is ok or not, but video playback was quite ok (once the page loads) and full-screen playback (1440p) is between 35-50% on all cores. To double check everything I only recently enabled the vpu node in the device tree and I did not spend any time investigating this at that time, is this enough for the mainline drivers or do we need to add different codecs as well? &vpu { status = "okay"; }; about-gpu-2024-04-27T12-05-58-533Z.txt

I tried the install package from this PPA , but it is a rather old version (114.x). As I can tell I think there is HW acceleration active. I am enclosing the whole dump of chrome://gpu. The system only reports that WebGPU and video_encode are disabled - the latter I think is clear as there is no DTB tree pointing to the hw codecs (yet). Lets see what you version will bring as you need to recompile the whole thing. I will try with firefox now. chromeReport.log

I started with the command written above, but for some reason it does not except the right path. I am using the ZSH and also exported path to /run/user/1000 in both .zshrc and .bashrc, but it does not accept it. h96-tvbox-3566:~:% sudo glmark2-es2-wayland error: XDG_RUNTIME_DIR not set in the environment. Error: main: Could not initialize canvas I then tried it out a bit differently, but this time I got a nice 3D horse on a 1440p rotating quite fluidly. h96-tvbox-3566:~:% glmark2-es2-wayland --fullscreen ======================================================= glmark2 2021.02 ======================================================= OpenGL Information GL_VENDOR: Mesa GL_RENDERER: Mali-G52 r1 (Panfrost) GL_VERSION: OpenGL ES 3.1 Mesa 23.2.1-1ubuntu3.1~22.04.2 ======================================================= [build] use-vbo=false: FPS: 196 FrameTime: 5.102 ms [build] use-vbo=true: FPS: 205 FrameTime: 4.878 ms ======================================================= glmark2 Score: 200 ======================================================= Edit: It was the sudo fault. I now executed the whole test, while connected via ssh - every core was like 25% during the whole test. I will try with the Chrome and Firefox now. h96-tvbox-3566:~:% glmark2-es2-wayland ======================================================= glmark2 2021.02 ======================================================= OpenGL Information GL_VENDOR: Mesa GL_RENDERER: Mali-G52 r1 (Panfrost) GL_VERSION: OpenGL ES 3.1 Mesa 23.2.1-1ubuntu3.1~22.04.2 ======================================================= [build] use-vbo=false: FPS: 461 FrameTime: 2.169 ms [build] use-vbo=true: FPS: 496 FrameTime: 2.016 ms [texture] texture-filter=nearest: FPS: 1000 FrameTime: 1.000 ms [texture] texture-filter=linear: FPS: 1005 FrameTime: 0.995 ms [texture] texture-filter=mipmap: FPS: 1020 FrameTime: 0.980 ms [shading] shading=gouraud: FPS: 322 FrameTime: 3.106 ms [shading] shading=blinn-phong-inf: FPS: 280 FrameTime: 3.571 ms [shading] shading=phong: FPS: 311 FrameTime: 3.215 ms [shading] shading=cel: FPS: 315 FrameTime: 3.175 ms [bump] bump-render=high-poly: FPS: 109 FrameTime: 9.174 ms [bump] bump-render=normals: FPS: 1043 FrameTime: 0.959 ms [bump] bump-render=height: FPS: 1043 FrameTime: 0.959 ms [effect2d] kernel=0,1,0;1,-4,1;0,1,0;: FPS: 570 FrameTime: 1.754 ms [effect2d] kernel=1,1,1,1,1;1,1,1,1,1;1,1,1,1,1;: FPS: 200 FrameTime: 5.000 ms [pulsar] light=false:quads=5:texture=false: FPS: 999 FrameTime: 1.001 ms [desktop] blur-radius=5:effect=blur:passes=1:separable=true:windows=4: FPS: 210 FrameTime: 4.762 ms [desktop] effect=shadow:windows=4: FPS: 579 FrameTime: 1.727 ms [buffer] columns=200:interleave=false:update-dispersion=0.9:update-fraction=0.5:update-method=map: FPS: 59 FrameTime: 16.949> [buffer] columns=200:interleave=false:update-dispersion=0.9:update-fraction=0.5:update-method=subdata: FPS: 59 FrameTime: 16> [buffer] columns=200:interleave=true:update-dispersion=0.9:update-fraction=0.5:update-method=map: FPS: 70 FrameTime: 14.286 > [ideas] speed=duration: FPS: 308 FrameTime: 3.247 ms [jellyfish] <default>: FPS: 388 FrameTime: 2.577 ms [terrain] <default>: FPS: 25 FrameTime: 40.000 ms [shadow] <default>: FPS: 214 FrameTime: 4.673 ms [refract] <default>: FPS: 26 FrameTime: 38.462 ms [conditionals] fragment-steps=0:vertex-steps=0: FPS: 788 FrameTime: 1.269 ms [conditionals] fragment-steps=5:vertex-steps=0: FPS: 786 FrameTime: 1.272 ms [conditionals] fragment-steps=0:vertex-steps=5: FPS: 784 FrameTime: 1.276 ms [function] fragment-complexity=low:fragment-steps=5: FPS: 780 FrameTime: 1.282 ms [function] fragment-complexity=medium:fragment-steps=5: FPS: 769 FrameTime: 1.300 ms [loop] fragment-loop=false:fragment-steps=5:vertex-steps=5: FPS: 771 FrameTime: 1.297 ms [loop] fragment-steps=5:fragment-uniform=false:vertex-steps=5: FPS: 775 FrameTime: 1.290 ms [loop] fragment-steps=5:fragment-uniform=true:vertex-steps=5: FPS: 758 FrameTime: 1.319 ms ======================================================= glmark2 Score: 524 =======================================================

I have installed the PPA and stuff from this thread: When I run it it produces the following trace. When playing videos all 4 cores are at 100% so I´d say it is not HW accelerated. Maybe there are also missing settings, but I have not seen any proper tutorial on how to set it up. h96-tvbox-3566:~:% chromium-browser [35467:35467:0426/133812.562432:ERROR:chrome_browser_cloud_management_controller.cc(162)] Cloud management controller initialization aborted as CBCM is not enabled. [35467:35467:0426/133813.132334:ERROR:object_proxy.cc(590)] Failed to call method: org.freedesktop.portal.Settings.Read: object_path= /org/freedesktop/portal/desktop: org.freedesktop.DBus.Error.ServiceUnknown: The name org.freedesktop.portal.Desktop was not provided by any .service files [35528:35528:0426/133813.233668:ERROR:gl_angle_util_vulkan.cc(189)] Failed to retrieve vkGetInstanceProcAddr [35528:35528:0426/133813.234447:ERROR:vulkan_instance.cc(91)] Failed to get vkGetInstanceProcAddr pointer from ANGLE.

Currently as a desktop. Wanted to try it out, but the kernel update over apt upgrade bricked my device. So I need to reflash everything. I am still doing everything on the current 6.6 kernel. IR works I just need to create a keymap and I will test KODI out - just need to recompile it in a VM. I am still waiting for my sound bar that has SPDIF to test it out. I think currently it is not working as I saw that spdif@fe460000 is currently disabled. I am also looking for the current Linux-headers file 6.6.28 so that I can try with openvfd. I was unable to find them with the previous kernel version 6.6.27. Once I install everything back I will retry this. What is funny about this box - every reinstallation brings a new MAC address. I am also trying to set up a CasaOS - currently in YouTube mode on how to set up Cloudflare tunnels.

@Hqnicolas I created a pull request for a working HDMI sound: https://github.com/hqnicolas/ArmBoardBringUp/pull/1/files

@lucky62 Change your line 736 to pin->pin = of_get_named_gpio(pdev->dev.of_node, name, 0);

Nice! You are using ubuntu GNOME image from the armbian repo or did you build it now yourself with the right dts file?

@Hqnicolas With this dts I was able to start a wifi with a simple symlink sudo ln -s fw_bcm4335b0_ag.bin brcmfmac4335-sdio.h96-TVbox,rk3566.bin I need to check a bit more into the chip IDs as in my honest opinion this AP6335 should be recognized as 4339 or that I try to dig up the clm_blob file that is currently missing? Bluetooth also works as hci1, because hci0 is recognized as the SDIO device - which means there are some settings in that node not completely valid. On the other hand, the wifi module has both communication lines. rk3566-h96-tvbox.dts Sadly I am not allowed to push to your git repo. If you add me to a collaboration list, then I will make the changes as a PR (my github username is kulica). Anyway, I am attaching the corrected file to this post.

Yes it needs to be low and then the broadcom dongle host driver needs to search for any hw on the sdio, which fires the init seaquence and enables the wifi module. This dhd is for some reason not started during boot.

It is a twofold problem. First is a base definition compiled with kernel itself, which has to simply put a LUT which ID loads which driver bin file. If it is not there, you will see a dmesg error message. So if this is not recompiled with kernel itself it won't even try to load a bin file nor will it throw any error messages. Can you check if the kernel sees any new hw device IDs?

@Hqnicolas Do you specifically select Broadcom 43xx SDIO kernel drivers as they were not enabled per default on my side. Lets see what will come out of this build process.

Good work! Will try to compile it and test it out on my side.

How are you compiling this dts? Is there a shorter way or you are recompiling the whole kernel?

https://elixir.bootlin.com/linux/latest/source/arch/arm64/boot/dts/rockchip/rk3566-box-demo.dts This one looks promising. Maybe we only need to change some pins.

Hi, yes this was also my conclusion. The pins were defined OK, but there was no voltage on that VCCIO4 domain - so the whole GPIOx was kept inactive or in a reset. It was confusing me, why BT is running, but there is no WiFi - the chip has to be powered somehow. I also see that you already defined the backlights on this PCBA. This one is also confusing me - how is this one PD7? It should normally be PB7 on the GPIO2. Maybe can you share your current dtb file and I will also test it out?

Thx right on the spot. I made me a coffee to start to crack this wifi stuff

Yes you are right. It was working on the previous 1.1 image. There was the 6.1 kernel being active.

Yes it was working before as the kernel itself reported that the bin files are missing. Now I am not sure how the kernel was compiled. So it will automatically parse from the config and driver bin files and then create the wlan0 interface?

What do you want to say with this? This is how I added the tx/rx DMA IRQ names, yes. I don't know why the overlay (if it already exists), was not active on my Rockshit device.

This is how @ning enable the LED and InfraRed https://github.com/zhangn1985/linux-stable/commit/eb97929f95b1af3257b74528159a5d55a6409bba I will try it out. In his repo I have also seen that he implemented the right keystrokes for it. Will need to check this one. For the dts I edited it right on the machine - the file that Armbian in settingsEnv is pointing at. Can you try dmesg at your side if you see any errors that uart1 DMAs have no name and therefore it wont be active? I am not sure how this overlays work, but if I remember correctly - I read somewhere that Rockchip does not support overlays. Do you know how to turn them on? Do they need to be compiled against the "main" dtb file? Why these patches need to be external. There are already lots of dtb files in that brcm folder or in a worst case they can be setup in a special deb package, whereas the dependancies to it could be managable. I think this is the RPi way of working and if there is a driver update, you get a new "firmware" debian package. @Hqnicolas To start fiddling with the WiFi I need to recompile the whole kernel once again or is there a quicker route?

@Hqnicolas Nice Job! Sorry I was away for the weekend and I already see that I need to reflash the board. Well I needed to start again with the version 1.2 - I started with Bluetooth and made it work - Although still not sure what is physically written on the chip and how it behaves. I have hcy6335, which in turn should be AP6335 containing the BCM4335 and for Bluetooth BCM4339. Somehow the chip gets represented as hci0: BCM4335A0 (002.001.006) build 0000. I am attaching the hcd file to this post that needs to be copied to the /lib/firmware/brcm. I also changed the dtb for the uart1 as there is no UART1 DMA mode possible otherwise. One just needs to add the dma-names to serial@fe650000 node: serial@fe650000 { compatible = "rockchip,rk3568-uart\0snps,dw-apb-uart"; reg = <0x00 0xfe650000 0x00 0x100>; interrupts = <0x00 0x75 0x04>; clocks = <0x0e 0x11f 0x0e 0x11c>; clock-names = "baudclk\0apb_pclk"; dmas = <0x24 0x02 0x24 0x03>; dma-names = "tx\0rx"; pinctrl-0 = <0x91 0x92 0x93>; pinctrl-names = "default"; reg-io-width = <0x04>; reg-shift = <0x02>; status = "okay"; phandle = <0x10a>; bluetooth { compatible = "brcm,bcm43438-bt"; clocks = <0x94 0x01>; clock-names = "lpo"; device-wakeup-gpios = <0x95 0x11 0x00>; host-wakeup-gpios = <0x95 0x10 0x00>; shutdown-gpios = <0x95 0x0f 0x00>; max-speed = <0x16e360>; pinctrl-names = "default"; pinctrl-0 = <0x96 0x97 0x98>; vbat-supply = <0x23>; vddio-supply = <0x63>; }; }; These are now the dmesg logs that I am getting [ 16.268929] Bluetooth: hci0: BCM: chip id 82 [ 16.269548] Bluetooth: hci0: BCM: features 0x2f [ 16.272162] Bluetooth: hci0: BCM4335A0 [ 16.272173] Bluetooth: hci0: BCM4335A0 (002.001.006) build 0000 [ 16.274413] Bluetooth: hci0: BCM4335A0 'brcm/BCM4335A0.hcd' Patch [ 17.114840] systemd[1]: Finished Armbian memory supported logging. [ 17.152605] systemd[1]: Starting Journal Service... [ 17.299179] systemd[1]: Started Journal Service. [ 17.339995] systemd-journald[632]: Received client request to flush runtime journal. [ 17.369055] Bluetooth: hci0: BCM: features 0x2f [ 17.372008] Bluetooth: hci0: BCM4335B0 JF-LTE MurataXJ AFH_LimitPwr_EDR 2STOPBIT-0343 [ 17.372025] Bluetooth: hci0: BCM4335A0 (002.001.006) build 0353 [ 17.482709] RPC: Registered named UNIX socket transport module. [ 17.482729] RPC: Registered udp transport module. [ 17.482733] RPC: Registered tcp transport module. [ 17.482736] RPC: Registered tcp-with-tls transport module. [ 17.482740] RPC: Registered tcp NFSv4.1 backchannel transport module. [ 18.604874] Bluetooth: BNEP (Ethernet Emulation) ver 1.3 [ 18.604896] Bluetooth: BNEP filters: protocol multicast [ 18.604914] Bluetooth: BNEP socket layer initialized [ 18.647499] Bluetooth: MGMT ver 1.22 ... [ 23.677339] Bluetooth: RFCOMM TTY layer initialized [ 23.677399] Bluetooth: RFCOMM socket layer initialized [ 23.677430] Bluetooth: RFCOMM ver 1.11 h96-tvbox-3566-wifi:~:% hciconfig -a hci0: Type: Primary Bus: UART BD Address: 43:35:B0:07:1F:AC ACL MTU: 1021:8 SCO MTU: 64:1 UP RUNNING RX bytes:4846 acl:0 sco:0 events:556 errors:0 TX bytes:71684 acl:0 sco:0 commands:556 errors:0 Features: 0xbf 0xfe 0xcf 0xff 0xdf 0xff 0x7b 0x87 Packet type: DM1 DM3 DM5 DH1 DH3 DH5 HV1 HV2 HV3 Link policy: RSWITCH SNIFF Link mode: PERIPHERAL ACCEPT Name: 'h96-tvbox-3566-wifi' Class: 0x6c0000 Service Classes: Rendering, Capturing, Audio, Telephony Device Class: Miscellaneous, HCI Version: 4.0 (0x6) Revision: 0x161 LMP Version: 4.0 (0x6) Subversion: 0x4106 Manufacturer: Broadcom Corporation (15) h96-tvbox-3566-wifi:~:% bluetoothctl Agent registered [CHG] Controller 43:35:B0:07:1F:AC Pairable: yes [bluetooth]# show Controller 43:35:B0:07:1F:AC (public) Name: h96-tvbox-3566-wifi Alias: h96-tvbox-3566-wifi Class: 0x006c0000 Powered: yes Discoverable: no DiscoverableTimeout: 0x000000b4 Pairable: yes UUID: A/V Remote Control (0000110e-0000-1000-8000-00805f9b34fb) UUID: Handsfree Audio Gateway (0000111f-0000-1000-8000-00805f9b34fb) UUID: PnP Information (00001200-0000-1000-8000-00805f9b34fb) UUID: Audio Sink (0000110b-0000-1000-8000-00805f9b34fb) UUID: Headset (00001108-0000-1000-8000-00805f9b34fb) UUID: A/V Remote Control Target (0000110c-0000-1000-8000-00805f9b34fb) UUID: Generic Access Profile (00001800-0000-1000-8000-00805f9b34fb) UUID: Audio Source (0000110a-0000-1000-8000-00805f9b34fb) UUID: Generic Attribute Profile (00001801-0000-1000-8000-00805f9b34fb) UUID: Device Information (0000180a-0000-1000-8000-00805f9b34fb) Modalias: usb:v1D6Bp0246d0540 Discovering: no Roles: central Roles: peripheral Advertising Features: ActiveInstances: 0x00 (0) SupportedInstances: 0x05 (5) SupportedIncludes: tx-power SupportedIncludes: appearance SupportedIncludes: local-name [bluetooth]# scan on Discovery started [CHG] Controller 43:35:B0:07:1F:AC Discovering: yes [NEW] Device 4C:BA:D7:02:F5:B7 4C-BA-D7-02-F5-B7 @ning @Hqnicolas UART1 speed is set to max-speed = <1500000>; but according to the datasheet (still no clue if it is the right one) we can go up to 4 MBit/s. Another thing is the schematics that I found somewhere on the internet, but still they are using some AP6xxx wifi/BT chip. Need to dive into this, but hey, now we see the missing components needed for SD-Card slot. Heartbeat is a nice thing, but I will try to enable the backlight blue LEDs, WiFi runniung, and then try to get some HDMI audio out of this thing. BCM4335A0.hcd p562297-AP6335 datasheet_V1.3_02102014.pdf ROC-3566-PC-V10-20210419.pdf