RK3328 Kernel

[TonyMac32]

OK, So for the boot issue, I just booted against every display in my home, including 2 televisions, no issues with the 4.14 kernel (4.14.80).  That, in my opinion, rules out u-boot to the extent that I can test.  Given my experience with the 4.4 kernel, I am of the opinion that it can't get any worse than it is, especially since we've been mangling the shite out of it to try to keep up with Rockchip and made one hell of a mess.  @JMCC if you have tested the 4.4 you've frozen so far, we can get that up to feature parity and move forward.  We can't keep dodging Rockchip updates, and we can't go back to the old MQMaker kernel because it's old enough to have moss growing on it.  I'll take a look at the patches and see which ones go where and how much work we're in for, things like the reboot I think need a complete rework, some of the later Rockchip work fixed it in a more correct fashion if I'm reading properly.

[Tido]

Hi Tony,

 

Thank you for checking and if JMCC's (a working fork of RK BSP kernel) is working like a charm, is for sure better than random reboot's :-)

 

To avoid misunderstanding I never had HDMI issues, mine wouldn't do a cold start, just the red LED. No matter HDMI connected or not.
As I wrote here: https://forum.armbian.com/topic/3327-asus-tinkerboard/?do=findComment&comment=65557 dtb -> dtb-4.14.79-rockchip/  if you now run a 4.14.80 this issue maybe fixed by now.

 

[TonyMac32]

2 hours ago, Tido said:

To avoid misunderstanding I never had HDMI issues, mine wouldn't do a cold start, just the red LED. No matter HDMI connected or not.

No worries, I was testing to try to remove u-boot from the list of things that might be wrong. 

 

[dragonlost]

19 hours ago, Tido said:

Hi Tony,

 

Thank you for checking and if JMCC's (a working fork of RK BSP kernel) is working like a charm, is for sure better than random reboot's :-)

 

To avoid misunderstanding I never had HDMI issues, mine wouldn't do a cold start, just the red LED. No matter HDMI connected or not.
As I wrote here: https://forum.armbian.com/topic/3327-asus-tinkerboard/?do=findComment&comment=65557 dtb -> dtb-4.14.79-rockchip/  if you now run a 4.14.80 this issue maybe fixed by now.

 

Hello. I have the problem of hdmi that prevents me from booting with a screen that serves me for my tests.
I am currently on the kernel 4.14.76.
How to upgrade to this 4.14.80 kernel to test it?

[Igor]

5 minutes ago, dragonlost said:

How to upgrade to this 4.14.80 kernel to test it?

Build on your own https://github.com/armbian/build or wait.

[dragonlost]

6 minutes ago, Igor said:

Build on your own https://github.com/armbian/build or wait.

Ok  i waited

[TonyMac32]

It would seem booting it without the HDMI attached and adding it later works. Then, if you build your own kernel, install the dtb's. Now remember my testing was not exhaustive.
@Igor has there been any issue with Allwinner HDMI recently? This board and the Amlogic SoC's use the DW-HDMI, and Amlogic is a mess with HDMI right now as well. Probably unlikely, but seems odd.

Sent from my Pixel using Tapatalk

[dragonlost]

I tested to intaller Mate desktop on next armbian ubuntu desktop 4.14.76. For that I had to purge the package "linux-bionic-root-next-tinkerboard". The installation is pass without much problem and I now have an armbian version on Mate.

So I would like to know what this package consists of: linux-bionic-root-next-tinkerboard ? Thanks

[Igor]

2 hours ago, dragonlost said:

I tested to intaller Mate desktop on next armbian ubuntu desktop

 

How exactly? It seems we have some bug if you had to remove our BSP package ... which is needed for proper functioning. 

BTW. Our stock XFCE is faster (and nicer) than Mate desktop.

[Tido]

I give a shit about Mate on the deskotp I need the device to work and programms running. Everything else is putting lipstick on a pig. Such discussion are waste of precious time.

 

Question @Igor, if I restore again and run the update, will I get now 4.14.80 ?

[TonyMac32]

@Tido I would suggest to take a step back. Relating one's experience and results is perfectly fine, as well as replying to such information reasonably. I also caution that my experience with 4.14 is of course with my hardware, not with your hardware.

Now then, I will have some time to review the bsp, at least the rockchip-specific part, directly. We can then reasonably and *with civil discourse* discuss the path forward. I would very much like the dmesg output you get when plugging your monitors in, and of course any serial debug output you can provide on failed boot events.

Sent from my Pixel using Tapatalk

[Igor]

17 minutes ago, Tido said:

if I restore again and run the update, will I get now 4.14.80 ?

https://beta.armbian.com/pool/main/l/ Nope, not yet, perhaps in the morning.

[Tido]

1 hour ago, TonyMac32 said:

@Tido I would suggest to take a step back. Relating one's experience

I run Mate on all my Desktop's Tower and Notebook, but I touch the Operating System or the Desktop Environment about 1% of my work, for the rest I use programs.

So I arranged myself with armbian's XFCE because it is not worth spending time for 1%.  That written I fully stand to what I wrote above - it is not worth spending time on such a topic.

However having no display or not booting at all is a really bad experience, in compare. So adding dragonlost's MATE on top of already serious topics, I wrote what I think about.

 

If you are running it on a tinker board, chances are quite small we have different HW.

 

UART, on which connectors/output will it sit with 4.14   5.54 ?

[mpmc]

The following issue may have already been reported, so apologies if it has.

 

Fresh install of both desktop 18.04 and Stretch next (Armbian_5.59_Tinkerboard_Debian_stretch_next_4.14.67) on the Tinkerboard Ethernet comes up but does not connect without manual intervention., on the desktop I had to login and connect manually, with the "server" edition I had to modify the interfaces file to get an address.

 

Intentional or a bug?

[TonyMac32]

7 hours ago, mpmc said:

on the Tinkerboard Ethernet comes up but does not connect without manual intervention.

 

I can't imagine that being on purpose, sounds like an environment issue rather than a BSP package issue though, So it's not my department. 

 

[edit]  My 4.14.80 build does not have any issues with bringing up LAN on boot. 

 

 

[dragonlost]

21 hours ago, Igor said:

 

How exactly? It seems we have some bug if you had to remove our BSP package ... which is needed for proper functioning. 

BTW. Our stock XFCE is faster (and nicer) than Mate desktop.

# purge xfce

sudo apt-get -y purge xfce*

 

# purge next tinkerboard

sudo dpkg --purge linux-bionic-root-next-tinkerboard

 

# install lightdm

sudo apt-get -y install lightdm

sudo apt-get -y install xserver-xorg

sudo apt-get -y install lightdm-greeter

 

# select NoDm for test

 

# install Mate

sudo apt-get -y install mate

sudo apt-get -y install ubuntu-mate-core ubuntu-mate-desktop

sudo apt-get -y install ubuntu-mate-default-settings ubuntu-mate-icon-themes

sudo apt-get -y install ubuntu-mate-live-settings ubuntu-mate-guide

sudo apt-get -y install mate-desktop-environment-extras mate-indicator-applet

sudo apt-get -y install synaptic

sudo apt-get -y install engrapa

sudo apt-get -y install caja-actions

sudo apt-get -y install caja-wallpaper

sudo apt-get -y install caja-open-terminal

sudo apt-get -y install mozo

sudo apt-get -y install pluma

sudo apt-get -y install mate-tweak

sudo apt-get -y install mate-themes

sudo apt-get -y install mate-applets gnome-media

sudo apt-get -y install mate-panel

sudo apt-get -y install mate-system-monitor

sudo apt-get -y install blueman

sudo apt-get -y install firefox

sudo apt-get -y install ubuntu-mate-themes

sudo apt-get -y install pulseaudio indicator-sound indicator-sound-gtk2 libcanberra-pulse paprefs

sudo apt-get -y install pulseaudio-module-bluetooth pulse-audio-module-gconf pulseaudio-module-zeroconf

sudo alsa force-reload

 

it's not optimized. probably too many packages

[Igor]

15 minutes ago, dragonlost said:

# purge next tinkerboard

sudo dpkg --purge linux-bionic-root-next-tinkerboard

 

This package contains some important services and optimisations. Just look into it. I don't understand why you think you need to remove that? MATE works with it just fine.

[dragonlost]

16 hours ago, Igor said:

 

This package contains some important services and optimisations. Just look into it. I don't understand why you think you need to remove that? MATE works with it just fine. 

 

I do not want to delete it. It's just that if I keep it I can not install Mate because there is conflict:

 

Preparing to unpack .../39-ubuntu-mate-default-settings_18.04.17_armhf.deb ... Unpacking ubuntu-mate-default-settings (18.04.17) ... dpkg: error processing archive /tmp/apt-dpkg-install-E9v1ON/39-ubuntu-mate-default-settings_18.04.17_armhf.deb (--unpack): trying to overwrite '/usr/lib/chromium-browser/master_preferences', which is also in package linux-bionic-root-next-tinkerboard 5.60

he uses all 2 the same directory :  /usr/lib/chromium-browser/master_preferences

[Igor]

26 minutes ago, dragonlost said:

I do not want to delete it. It's just that if I keep it I can not install Mate because there is conflict:

Aha, ok. I'll fix that.

[dragonlost]

2 minutes ago, Igor said:

Aha, ok. I'll fix that.

thanks !

[TonyMac32]

Split cold non-booting issue to own topic, still need to review HDMI topic. 

 

In other news, I noticed my 4.14.80 install mentioned I had an SDR104 SDHC and that it had successfully tuned the phase, not something I remember seeing before (haven't looked in a long time)

 

So with the typical iozone parameters I got  > 66 MB/s random read, and  > 58 MB/s random write.  Not too shabby.     (SD Card:  SanDisk Extreme 32 GB U3)

        File size set to 102400 kB
        Record Size 4 kB
        Record Size 16 kB
        Record Size 512 kB
        Record Size 1024 kB
        Record Size 16384 kB
        Command line used: iozone -e -I -a -s 100M -r 4k -r 16k -r 512k -r 1024k -r 16384k -i 0 -i 1 -i 2
        Output is in kBytes/sec
        Time Resolution = 0.000001 seconds.
        Processor cache size set to 1024 kBytes.
        Processor cache line size set to 32 bytes.
        File stride size set to 17 * record size.
                                                              random    random     bkwd    record    stride
              kB  reclen    write  rewrite    read    reread    read     write     read   rewrite      read   fwrite frewrite    fread  freread
          102400       4     3228     3295    12302    12802    11459     5036
          102400      16    10544    10785    30695    30729    30434    15796
          102400     512    51482    50931    62070    61959    61823    48392
          102400    1024    58030    58118    63218    63221    63333    53391
          102400   16384    57062    60069    68049    67967    68066    59411

 

[dragonlost]

I just tested the latest version of the kernel for Tinkerboard ( 4.14.81-rockchip #488).

Mate's installation is worry-free and conflict free.
I have no problem connecting ethernet networks but without IP at startup.
But I still have the problem of HDMI port which prevents the boot with some screen (HP EliteDisplay E242)

[TonyMac32]

I'm wondering if CEC is part of the problem, I just disabled DW_HDMI_CEC on Amlogic to solve a similar (looking) problem.

[dragonlost]

16 hours ago, TonyMac32 said:

I'm wondering if CEC is part of the problem, I just disabled DW_HDMI_CEC on Amlogic to solve a similar (looking) problem.

How do we do that? please

[valant]

Hi, rk3288 has 4 SD/eMMC controllers, they are called SDMMC, SDIO (2 instances) and EMMC. The latter seems to be a newer IP, supporting more modes and newer versions of the specifications. Looks like SDMMC can't do SDR104, whereas EMMC can. The question is what of them the SD card slot is connetcted to? On Tinker Board. And on Tinker Board S, because unlike the former, the latter has eMMC module, so it may make difference (eMMC module's HS400 vs. SD card's SDR104). I know, it is written somewhere in the DT, but I forgot where in the linux/uboot dark forests those trees grow.  Thank you for your replies.

[TonyMac32]

SDR104 is enabled for the SDMMC interface, and if you look at the iozone testing I did above I was getting well over 50 MB/s, exceeding the SDR50, maximum.

Another note:

I have disabled video in U-boot on this architecture for now, there have been changes to the HDMI drivers within the Rockchip family that the old patches may not be properly addressing, specifically the RK3288 has been split from the rest of the Rockchip code.

Sent from my Pixel using Tapatalk

[valant]

26 minutes ago, TonyMac32 said:

SDR104 is enabled for the SDMMC interface, and if you look at the iozone testing I did above I was getting well over 50 MB/s, exceeding the SDR50, maximum.

Thank you! Oh my god, I feel so stupid - the post about exactly this is just on this very page!

[dragonlost]

16 hours ago, TonyMac32 said:

SDR104 is enabled for the SDMMC interface, and if you look at the iozone testing I did above I was getting well over 50 MB/s, exceeding the SDR50, maximum.

Another note:

I have disabled video in U-boot on this architecture for now, there have been changes to the HDMI drivers within the Rockchip family that the old patches may not be properly addressing, specifically the RK3288 has been split from the rest of the Rockchip code.

Sent from my Pixel using Tapatalk
 

I do not understand. Have you disabled CEC support in the new version?

[TonyMac32]

The entire DW-HDMI setup in U-boot appears questionable on RK3288. We had a patch from ASUS designed to "fix some issues" with HDMI, that included a lot of data channel "tweaks". I disabled that patch and the problem got worse (I went from not seeing this issue to having an issue on 1 monitor. So what I've done is disabled the entire video subsystem on U-boot, the kernel brings up the interface just fine.

Sent from my Pixel using Tapatalk

[dragonlost]

1 hour ago, TonyMac32 said:

The entire DW-HDMI setup in U-boot appears questionable on RK3288. We had a patch from ASUS designed to "fix some issues" with HDMI, that included a lot of data channel "tweaks". I disabled that patch and the problem got worse (I went from not seeing this issue to having an issue on 1 monitor. So what I've done is disabled the entire video subsystem on U-boot, the kernel brings up the interface just fine.

Sent from my Pixel using Tapatalk
 

ok