balbes150

andrew@m9cmax4k:~/Benchmarking/Linpack/hpl-2.1/bin/Aarch64$ uptime 13:43:47 up 39 days, 14:22,  1 user,  load average: 0.00, 0.01, 0.00 andrew@m9cmax4k:~/Benchmarking/Linpack/hpl-2.1/bin/Aarch64$ uname -a Linux m9cmax4k 4.19.7-aml-s9xxx #5.67 SMP PREEMPT Fri Dec 28 10:16:25 MSK 2018 aarch64 aarch64 aarch64 GNU/Linux 39 days uptime.
Ubuntu 18.04 running from eMMC. I practically never have to reboot this box (used as a file server with a USB hard disk), it is that good.
As always, I am grateful to Oleg and the entire Armbian development team for their amazingly good work.  
 

@balbes150 Thanks for the .deb packages from 5.76 to 5.77
 
for "normal" armbian-releases I did add me a package-information at the end of the file /etc/update-motd.d/10-armbian-header
ARMBIAN_bsp=$(more /etc/armbian-release|grep VERSION|cut -f 2 -d '=') ARMBIAN_kernel=$(dpkg -s linux-image-$BRANCH-$LINUXFAMILY|grep Version|cut -f 2 -d ' ') ARMBIAN_uboot=$(dpkg -s linux-u-boot-$BOARD-$BRANCH|grep Version|cut -f 2 -d ' ') ARMBIAN_dtb=$(dpkg -s linux-dtb-$BRANCH-$LINUXFAMILY|grep Version|cut -f 2 -d ' ') ARMBIAN_firmware=$(dpkg -s armbian-firmware|grep Version|cut -f 2 -d ' ') ARMBIAN_config=$(dpkg -s armbian-config|grep Version|cut -f 2 -d ' ') printf 'package bsp-kernel[\e[0;91m%s\x1B[0m] u-boot[\e[0;91m%s\x1B[0m] dtb[\e[0;91m%s\x1B[0m] firmware[\e[0;91m%s\x1B[0m] config[\e[0;91m%s\x1B[0m]\n' "$ARMBIAN_kernel" "$ARMBIAN_uboot" "$ARMBIAN_dtb" "$ARMBIAN_firmware" "$ARMBIAN_config" printf '\n' for your armbian-version I had to change the "source" because you didnt use as much environment-variables like the "real" one
 
ARMBIAN_bsp=$(more /etc/armbian-release|grep VERSION|cut -f 2 -d '=') ARMBIAN_kernel=$(dpkg -s linux-image-aml-s912|grep Version|cut -f 2 -d ' ') ARMBIAN_uboot=$(dpkg -s linux-u-boot-aml-s912-default|grep Version|cut -f 2 -d ' ') ARMBIAN_dtb=$(dpkg -s linux-dtb-aml-s912|grep Version|cut -f 2 -d ' ') ARMBIAN_firmware=$(dpkg -s firmware-armbian-full|grep Version|cut -f 2 -d ' ') ARMBIAN_config=$(dpkg -s armbian-config|grep Version|cut -f 2 -d ' ') printf 'package bsp-kernel[\e[0;91m%s\x1B[0m] u-boot[\e[0;91m%s\x1B[0m] dtb[\e[0;91m%s\x1B[0m] firmware[\e[0;91m%s\x1B[0m] config[\e[0;91m%s\x1B[0m]\n' "$ARMBIAN_kernel" "$ARMBIAN_uboot" "$ARMBIAN_dtb" "$ARMBIAN_firmware" "$ARMBIAN_config" printf '\n' but now its works too
 
armbian-aml-S912 shows 5.73 because of the package
 
linux-stretch-root-rockpro64    5.73   
Login with the new released puTTY 0.71  ==> https://www.chiark.greenend.org.uk/~sgtatham/putty/changes.html
 

use commit
https://github.com/150balbes/Build-Armbian/commit/2ec509ae341cdd8508a2ca245ee861dafa276f54
 
 
to all
New image 5.77

S905X was released three years ago. The main WiFi is rtl8189. 9082C WiFi is an alternative. Now most S905X/S905W are using SSV6051WiFi.
I'm a manufacturer of Android TV box and I'm currently developing armbian and learning how to compile it.

Yes, I was going to say that too! Specially for open-source, the Amlogic S9XXX TV boxes and SBC's are vastly superior to the RPi's.
Check this FAQ: http://wiki.loverpi.com/faq:sbc:libre-aml-s805x-faq#toc6

Now RPI strongly loses in all respects to s9xxx and rk33xx TV boxes. The only place where the use of RPI is justified, systems that use a variety of external sensors and actuators, of which a large number are produced and there are ready-made systems for their use.
 
 
 
 LE did not use patches, but a ready-made closed binary module (driver), which was taken from the finished Android firmware. This binary module was built exclusively for one version of the 3.14 kernel and cannot work with any other versions of the kernel. The manufacturer of the chip does not give anyone the source code from this driver, so no one will spend their time disassembling the binary code and studying it to write a new driver for the kernel 4\5. If you want, you can do the job yourself.

The point, in case I wasn't clear enough, is that there is zero documentation and zero support from the S9082 vendor, and the development effort to reverse engineer a linux driver for such a chip is ridiculously high. So, no, I don't think there will ever be a driver for the 4.xx/5.xx linux kernels for a totally unsupported chip such as the S9082. Again, if you need Wifi on a box that uses such a chip, use a USB WiFi dongle.
 
Check this thread: https://forum.libreelec.tv/thread/5457-s905x-unknown-wlan-chip/

That is correct yes, once the system boots into Armbian the disk is accessible without issue.
 
I've also managed to enable the built-in SATA port by writing the GPIO registers from u-boot, which is also detected correctly in u-boot (that's a JMicron 2329). I'll make a follow-up post about that.

If you have NAND, you will not be able to install new versions with kernel 4 and 5 in NAND. These kernels only work with eMMC.

I did this with things I had at home, I had a night of 12 leftovers, put to 5v and nothing is heard. And so I cooled an SSD which is where I have Armbian installed.
Cut with the dremel and I put a dust filter.
And this Q912 does not get too hot anymore (I think the metal housing acts as a heatsink), so it's cool.

Pay attention, many S905 is used with legacy u-boot. Therefore, do not try to install the latest versions in eMMC. The regular u-boot will be destroyed and you will need to restore the firmware via a USB cable or from a bootable SD card.

New image 5.76  kernel 5.0 .
 

New image 5.76  kernel 5.0 .
 

We create our own future, if you firmly believe that you will be all right-then everything will be fine.

Just wanted to report on the status of my three Mecool Km8-P TV boxes.
- All three boxes have now had their firmware flashed with an older Android version which includes a u-boot that correctly boots 4.xx kernels. The u-boot version in the shipped factory firmware on this recent batch of Km8-P boxes would boot kernel 3.14.29 fine but would absolutely refuse to boot any 4.xx/5.xx kernel.
- All three boxes are now running Armbian_5.75_Aml-s912_Ubuntu_bionic_default_4.20.5_20190211.img.xz with meson-gxm-q201.dtb. This release has proved to be exceptionally stable on these Mecool Km8-P TV boxes.
- Two boxes are running Armbian from a Sandisk A1 SD card. On one box I have installed Armbian on the eMMC, simply executing the install.sh script did the trick without any problems.
- I use these boxes in headless mode (no monitor, keyboard or mouse, access through ssh) as distcc compile nodes and they perform admirably. I can compile the Linux kernel in under 15 minutes on my compile farm.
- These boxes run at a cool 38~42C (depending on room temperature) at idle and 72~76C under heavy load and do not require any special cooling. I have not observed any thermal throttling even under heavy load.
- I have benchmarked the Km8-P using both sbc-bench and the Phoronix Benchmark Suite, see the thread I started on the subject of benchmarking an inexpensive S912 TV box. In summary, these Km8-P boxes place in the middle of the pack for single-threaded benchmarks and at the top of the pack for multithreaded benchmarks, and are the best in terms of performance per dollar compared to any other SBC, as of February 2019.
- I have not tested any of the multimedia capabilities: audio, video decoding, 3D graphics acceleration, HDMI, etc.
 

Blender test. Checking the relevance of SETTINGS instead of focusing on irrelevant hardware details like DDR3 vs. DDR4:
RockPro64, Ubuntu Bionic with LXDE, 2.0/1.5GHz, CONFIG_HZ=1000: 1:15:31 RockPro64, Ubuntu Bionic with LXDE, 2.0/1.5GHz, CONFIG_HZ=250: 1:06:59 RockPro64, Ubuntu Cosmic with LXDE, 2.0/1.5GHz, CONFIG_HZ=250: 1:01:11  
That's 8:30 minutes difference due to switching from CONFIG_HZ=1000 to CONFIG_HZ=250. Check %sys vs. %user below (iostat 60 output). And why is Cosmic faster than Bionic if it's exactly same Blender version? Since Cosmic (18.10) uses GCC 8.2 while Bionic (18.04) uses GCC 7.3 to build the packages. So by switching from default SoC vendor kernel settings to something better and by letting modern compilers do their job we get almost 25% performance 'for free'.
 

There is no such thing as the "right" one, because each person has a different use case, some buy a TV box to use as a server, others for multimedia, etc...
However, if people want to buy an Amlogic S912 SBC to experiment, I strongly recommend a Khadas Vim2. 

I'm not accusing you of forgery. This was an example of how many different factors (both often implicit and difficult to track) affect events. 
 
 
 
Maybe I'm not looking hard enough, N2 is WoL ?

Completed tests:
 
1. Previous u-boot (from file s905x-uboot.img in Armbian 5.75 image) : Image Armbian_5.76_Aml-s905_Ubuntu_bionic_next_5.0.0-rc7-next-20190219-g43dc36c94-dirty.img.xz boots fine on Km8-P (S912) with 1Gb RAM. Ethernet (wired) works erratically. USB does NOT work (kernel or dtb issue?)
 
(Edit)
2. u-boot.bin from deb package for S905: Image Armbian_5.76_Aml-s905_Ubuntu_bionic_next_5.0.0-rc7-next-20190219-g43dc36c94-dirty.img.xz boots BUT there is a kernel panic before prompt is reached (perhaps too tight memory timings?).
 
(Edit)
3. u-boot.bin from deb package for S912: does not boot. Error is in u-boot RAM detection, as before:
BL2 Built : 14:59:36, Aug  9 2017. gxl ge8c6a83 - xiaobo.gu@droid12 set vdd cpu_a to 1120 mv set vdd cpu_b to 1050 mv set vddee to 1000 mv Board ID = 6 CPU clk: 1200MHz DQS-corr enabled DDR scramble enabled DDR3 chl: Rank0+1 @ 912MHz - FAIL DDR4 chl: Rank0+1 @ 1008MHz - FAIL DDR init failed... Reset... I am going to use Armbian Image 5.75 with kernel 4.20 with u-boot from file s905x-uboot.img in Armbian 5.75 image for a few days to test for stability and if I don't find any problems, I think I am going to install it in eMMC on my three Km8-P boxes, which I use mainly for kernel compiling.

I tryed some ways of cooling that device passive, but only this one keeps the temperature constant over time. Current came first from a modded usb-cable, but connecting to the 3,3v of the uart port does the job too and keep the noise down. Not pretty, but it improves the lifetime of the cpu 😎

For mpeg-2 on rk3288/rk3328 there are some other patches needed, see my rockchip-5.x-vpu branch for working rk3288/rk3328 mpeg-2 decoding on v5.0-rc6. clk, drm and dts patches will be sent upstream any day now.
Also check out https://github.com/mpv-player/mpv/pull/6461 and the linked ffmpeg hwaccel if you want to use mpv or kodi-gbm for testing, I recently pushed dynamic selection of media/video device to hwaccel so should work without forcing decoder to /dev/video0.
 
I pushed two libreelec test images to http://kwiboo.libreelec.tv/test/ for tinker board and rock64 if you want to test mpeg-2 decoder, it includes patches from my rockchip-5.x-rebase and rockchip-5.x-vpu linux branch.

Still no real specs.
TL told me that it in the beginning of Februari would be sold(2nd week). But still nothing has changed on the website.
Here a announcement with picture of the H64.
https://liliputing.com/2019/02/pine64s-single-board-computers-are-getting-2019-upgrades-prices-still-start-at-25.html
I also made a video about it and the other new products they'll bring out.

You can buy it here on Amazon.de.
https://www.amazon.de/Pine-H64-3GB-Board-Modell/dp/B07BYXB173/ref=cm_cr_arp_d_product_top?ie=UTF8

The board does look good, I like the model b a lot more than model a. I hope there's quickly going to be good software for it so I can test it well.

I bought on aliexpress. You can use any options, but the larger the radiator, the better it removes heat.
 
https://ru.aliexpress.com/item/20x20x6mm-Aluminum-IC-LED-Cooling-Cooler-Heat-Sink-Heatsink-Radiator-10pcs-Free-Freight/32251373935.html?spm=a2g0s.9042311.0.0.274233edpb0Hqw

Variants of alteration of the cooling system. For example, three models MVR9 (rk3328), MX10 (rk3328) x96mini (s905w).
In the body, above the radiator of the processor, a large hole is made and on the sides of the body a large hole is made for the free passage of air.
On x96mini had to add a new radiator.
 
Note need necessarily two holes, one for air intake, the second for exit. This is a very common mistake TV box manufacturers, they sometimes make only one type of holes, which is not enough for the normal operation of the cooling system. The air must pass freely through the body. The direction of air supply for TV boxes is optimal from the hole on the side surface, to the exit through the hole above the radiator. This allows the most heated air from the radiator to be immediately removed from the body. Then it doesn't heat the rest of the parts. If you do the opposite, to blow air down on the radiator and then output from the side vents, the heated (from the radiator), the air will simultaneously heat the remaining elements in the housing in its path. After this alteration and the addition of a USB fan, the system began to work stably in all modes for many hours. For anyone who uses their TV boxes as a full desktop, I strongly recommend to redesign the cooling system. Pay attention to the model x96mini, it has a default motherboard and radiator located at the bottom (upside down from the correct position). When using x96mini (and other TV boxes) be sure to turn them in the correct position - radiator up. This nonsense was invented by bad engineers who do not know the basic laws of physics. Such "engineers" think that the water can flow up the mountain. 
 
foto
 
https://yadi.sk/d/vQw9J81koxpuHg

New version 5.75 of the image and script.  dir s905_mali
 
Simplified use of the script. Once your image is running nominally for Armbian need to download, unzip and run the script. After the script finishes, reboot the system. If after reboot there will be no sound, you need to perform, on behalf of the ROOT user, one command
 
update-initramfs -v -c -k `uname -r`
 
and reboot the system. No other additional steps or commands are required.  Please note that the new version of the script works only with this version of the image (which is with him on the site in the same directory).