Below is the Xorg.0.log from stock Ubuntu OS which may help in getting the graphics working in Debian. X.Org X Server 1.7.6 Release Date: 2010-03-17 X Protocol Version 11, Revision 0 Build Operating System: Linux 2.6.32.9-dove-5.3.6-1-armel armv7l Ubuntu Current Operating System: Linux HPf4ce4625c9fa 2.6.32.9-dove-5.3.6-1-armel #1 Tue Oct 9 10:08:35 MDT 2012 armv7l Kernel command lineby Koen - Debian
Hi, I've logged into stock OS and extracted some info. Not sure if it is useful since bodhi already has Debian working however I may as well put it here as a reference. There is no ssh on the device so all commands have been executed locally and the output saved in text files on a USB drive. Koen uname: Linux HPf4ce4625c9fa 2.6.32.9-dove-5.3.6-1-armel #1 Tue Oct 9 10:08:35 MDT 20by Koen - Debian
List of the markings on the larger chips I can decipher: jmf601 BGBA1 flash alcor AU6256 USB controller 4x Samsung k4b2g0846D memory TFP410 TI PanelBus Digital Transmitter hynix memory (number unreadable) marvell ethernet chip (M visible, number unreadable) Koenby Koen - Debian
Unit arrived. Ssh in stock is disabled so will need to hook up a monitor and keyboard. Lots of dust inside making it hard to decipher chip numbers at the moment. Will have a closer look this evening. Koenby Koen - Debian
I'm a bit confused about all the GPU related stuff. If I understand things correctly the armada module is the built in graphics driver which should be able to output simple graphics to the screen. For hardware accelerated graphics there are closed source galcore / vivante and vmeta drivers / modules. The wandboard seems to use the same vivante gpu and there are some instructions over hereby Koen - Debian
Not sure how important this is but Debian reports a much lower CPU speed than in the original Ubuntu. Although this may be due to a change in the algorithm. https://en.wikipedia.org/wiki/BogoMips Ubuntu 10: Calibrating delay loop... 992.87 BogoMIPS (lpj=4964352) Debian: [ 0.018906] Calibrating delay loop (skipped), value calculated using timer frequency.. 333.33 BogoMIPS (lpjby Koen - Debian
Quotebodhi The Armada driver has been in mainline kernel for a few years. I think we are just missing some steps to activate the device (we should be seeing in dmesg but not yet). We should look for the bootlogs of similar boxes that has Armada 510 to see if we can find some hints. The most similar device I have found is the original cubox by solidrun. https://en.wikipedia.org/wiki/CuBoxby Koen - Debian
In the original dmesg I see this part: Initializing ths8200_init dovedcon has been initialized. dovefb_display_init dovefb_display driver init ok. Googling ths8200 comes up with some sort of a video chip. http://www.ti.com/product/THS8200 I haven't seen that chip when I inspected the board before but I'll take the device apart and have another look tomorrow.by Koen - Debian
One thing I've found out is that with the new boot args the front USB ports are not very reliable to plug the rootfs in. The output from the modprobe and drm commands is as follows: root@debian:~# modprobe armada-drm modprobe: FATAL: Module armada-drm not found in directory /lib/modules/4.18.5-mvebu-tld-1 root@debian:~# root@debian:~# modprobe armada-lcd modprobe: FATAL: Module arby Koen - Debian
Hi bodhi, I've added the bootargs as described but removed the ' at the end since I assumed that was a typo. Could you let me know if that ' is needed on not. The device boots and I'm getting the bootlog on the serial console. I've included it below. There is no output to the screen but maybe this new log will give some more info to troubleshoot with. Koenby Koen - Debian
My unit arrives later this week but it seems like I've missed most of the fun :-) Koenby Koen - Debian
Hi, I've recently changed my pogoE02 to a HP T5325Z and I've also started working on getting a Dell Wyse T50 to run Debian. Both the HP T5325Z and T50 have relatively large internal memory which is not being used anymore now that the operating system runs from the USB stick. Would it be OK to use this internal memory as swap space or does this suffer from wear like the USB drive?by Koen - Debian
Hi bodhi, I've installed your latest kernel. It does show an armada module is loaded: root@debian:~# lsmod Module Size Used by arc4 16384 2 rt2800pci 16384 0 rt2800mmio 16384 1 rt2800pci rt2800lib 94208 2 rt2800mmio,rt2800pci rt2x00pci 16384 1 rt2800pci rt2x00mmio 16384 2 rby Koen - Debian
I've had a look through the Ubuntu sources and found this stuff related to the different SPI chips. Not sure if useful since source code is like a foreign language to me. in ubuntu-lucid/arm/arch/mach-dove/dove-db-setup.c it mentions an 8M chip that looks a bit like the one in my dmesg log (s25fl064k ) /***************************************************************************** *by Koen - Debian
spi related results from the latest dtb file root@debian:~# dmesg | grep spi [ 1.184135] m25p80 spi0.0: s25fl064k (8192 Kbytes) root@debian:~#by Koen - Debian
Hi bodhi, do you want me to try these latest dtb files as well? I've had another response from Dell that they will chase the technical team for the uboot GPL sources of the T50. Koenby Koen - Debian
Some more info about mtd. It seems the chip in my box is 8M while that in Bob's device is reported to be only 4M although that may possibly be due to different fw_env.config settings. Koen root@debian:~# mtd_debug info /dev/mtd0 mtd.type = MTD_NORFLASH mtd.flags = MTD_CAP_NORFLASH mtd.size = 8388608 (8M) mtd.erasesize = 4096 (4K) mtd.writesize = 1 mtd.oobsize = 0 regions = 0by Koen - Debian
Just got home and it is great to see that more progress has been made. Below is the dmesg output of booting with the latest dtb file. Ethernet is working and the box powers off properly as well. root@debian:~# dmesg [ 0.000000] Booting Linux on physical CPU 0x0 [ 0.000000] Linux version 4.12.4-mvebu-tld-1 (root@tldDebianVM) (gcc version 5.4.0 20160609 (Ubuntu/Linaro 5.4.0-6ubuntu1~16by Koen - Debian
Hi, really cool to see more progress on this. I'm still away from home so won't be able to test until later next week. I did have the same issue with shutdown not working properly and ended up just pulling the power cable. Koenby Koen - Debian
Hi bodhi, below the dmesg output from booting with the latest dtb file appended. I'm away for the next 2 weeks so won't be able to do any testing until I'm back. Koen root@debian:~# dmesg [ 0.000000] Booting Linux on physical CPU 0x0 [ 0.000000] Linux version 4.12.4-mvebu-tld-1 (root@tldDebianVM) (gcc version 5.4.0 20160609 (Ubuntu/Linaro 5.4.0-6ubuntu1~16.04.4)by Koen - Debian
Hi, I've had a better look at the chips on the board and this is what I can read. It seems my earlier assumption about the lankom chip was wrong. Koen Top of the board: Lankom Elec LG 2405S ENE KB3930 QF JMD330: jmicron sata bridge chip Bottom of the board: samsung k9k8g08u0d scb0: flash memory? 55vd020 60-c-tqwe: nor flash? alc5630: realtek audio? gl850g: usb conby Koen - Debian
Hi, The lankom chip was the closest to the Ethernet port so I assumed that was it. I'll have a look when I get home tonight and will post a list of the different chips on the board. Koenby Koen - Debian
dmesg output booting using the latest dove-wyse-t50.dtb file root@debian:~# dmesg [ 0.000000] Booting Linux on physical CPU 0x0 [ 0.000000] Linux version 4.12.4-mvebu-tld-1 (root@tldDebianVM) (gcc version 5.4.0 20160609 (Ubuntu/Linaro 5.4.0-6ubuntu1~16.04.4) ) #1 SMP PREEMPT Sat Jul 29 21:56:37 PDT 2017 [ 0.000000] CPU: ARMv7 Processor [560f5815] revision 5 (ARMv7), cr=10c5387dby Koen - Debian
bodhi Wrote: ------------------------------------------------------- > > Linux variant of T50 has EOL a long time back > and > > we don’t provide any kind of support. Also > the > > T50 (and T series thinclients) were not using > > u-boot. They had W-boot (Wyse boot), our > > > proprietary bios. > > I think this is wrong. Unless I misrby Koen - Debian
These links may be of interest as well since it contains stuff related to dove http://oem.archive.canonical.com/updates/pool/public/l/linux-mvl-dove/ http://oem.archive.canonical.com/updates/pool/public/l/linux-meta-mvl-dove/ Koenby Koen - Debian
Reply from Dell about GPL sources: QuoteDell legal person Hi Koen, Linux variant of T50 has EOL a long time back and we don’t provide any kind of support. Also the T50 (and T series thinclients) were not using u-boot. They had W-boot (Wyse boot), our proprietary bios. For Linux base OS, canonical maintained a separate repository for the project Triton and Janus. If you openby Koen - Debian
bodhi Wrote: ------------------------------------------------------- > Koen, > > > One thing which shows that the ethernet is not > > configured properly in the dts file yet is that > it > > now shows up as 100Mb/s while it used to be > > 1000Mb/s > > Does this box really have 1Gbs? > > > This reverse engineering is quite addictive buby Koen - Debian
You are right, I just had to install the correct firmware and some other wireless packages to get the wifi card to show up. root@debian:~# ifconfig eth0: flags=4163<UP,BROADCAST,RUNNING,MULTICAST> mtu 1500 inet 192.168.1.233 netmask 255.255.255.0 broadcast 192.168.1.255 inet6 fe80::280:64ff:feaa:7d6c prefixlen 64 scopeid 0x20<link> ether 00:80:64by Koen - Debian
Ok so I've modified the dts again by setting most of the 'disabled' options to 'okay'. When there are multiple nodes / sections with the same name I have only enabled the 1st one. The dmesg output is listed below. There is no video output on the screen and I've not tested audio yet. The log shows the pcie wifi card but I'm not sure if it is actually working sincby Koen - Debian
Attached is the .dts file I used to get ethernet working. This result does make it very tempting to also switch the audio controller, gpu and lcd sections from 'disabled' to 'okay'. Is there much risk in trying this or do wrongly defined parts just not work? Koenby Koen - Debian