Jump to content

dolphs

Members
  • Posts

    275
  • Joined

  • Last visited

Posts posted by dolphs

  1. Hi,

     

    seems upgrading to 30 May release 20.05.01 I lost "bbr" as congestion control setting.

    Can this be put back in please? - TiA! -

    root@wgvpn01:/etc/sysctl.d# uname -a
    Linux wgvpn01 5.4.43-rockchip64 #20.05.1 SMP PREEMPT Sat May 30 13:30:37 CEST 2020 aarch64 GNU/Linux
    
    root@wgvpn01:/etc/sysctl.d# sysctl net.ipv4.tcp_available_congestion_control
    net.ipv4.tcp_available_congestion_control = reno cubic
    
    root@wgvpn01:/etc/sysctl.d# sysctl -p wg.conf
    net.core.default_qdisc = fq
    net.core.rmem_max = 16777216
    net.core.wmem_max = 16777216
    net.ipv4.tcp_rmem = 4096 87380 16777216
    net.ipv4.tcp_wmem = 4096 65535 16777216
    sysctl: setting key "net.ipv4.tcp_congestion_control": No such file or directory
    net.ipv4.tcp_congestion_control = bbr
    net.ipv4.tcp_fastopen = 3
    net.ipv4.tcp_no_metrics_save = 1
    net.ipv4.tcp_slow_start_after_idle = 0
    net.ipv4.tcp_mtu_probing = 1

     

  2. Hi - Was wondering about the throttling issue encountered, also on H3/ H5 boards?

    Is there a fix meanwhile, also for the H6 boards or do we need to wait till kernel 5.8 has been completed ( DVFS aka Dynamic voltage and frequency scaling) ?

    Suppose @5kft knows more about this one?

     

    It is just a question and dont push my OPiOnePlus board to the limit, 

    also because of conservative governor setting - works fine also on this H6 board - for wireguard ( just like my fav boards NEO2 Black ) purposes

     

     

  3. 1 hour ago, guidol said:

    Yes - I have to "older" 512MB versions and only one the the "LTS" 1GB version
    I think the 1GB version will be a very rare item in the future

     

    I do envy you, as I have a 256Mb (1.0) / 512Mb (LTS 1.1 which does 1.3v thus 1300 cpu ) and just ordered most likely NEO2 black editions from a last batch.

    H6 OrangePiOnePlus also does good job for my usage, HOWEVER it is the size/performance that makes this NEO2 board so interesting. 

    we'll see what happens next 

  4. @Igor straight from my source :  " Sorry, the original factory no longer produces h5 boards, we have to stop production " so mainlining effort done just in time I suppose and we will have soon collector items

     

    @guidol - Only time can tell, but indeed it is quite strange what happened around the NEO2 H5, many revisions and perhaps a small similar board will be recreated based on one of the new H boards. as Igor said most likely H6   

     

  5. good news , I can concur 

    ii  linux-u-boot-orangepioneplus-dev      20.05.0-trunk                       arm64        Uboot loader 2020.04
    ii  u-boot-tools                          2019.01+dfsg-7                      arm64        companion tools for Das U-Boot bootloader

     

    And OrangePiOnePlus is happy so far:

    root@orangepioneplus's password:
      ___  ____  _    ___
     / _ \|  _ \(_)  / _ \ _ __   ___   _
    | | | | |_) | | | | | | '_ \ / _ \_| |_
    | |_| |  __/| | | |_| | | | |  __/_   _|
     \___/|_|   |_|  \___/|_| |_|\___| |_|
    
    Welcome to Armbian buster with Linux 5.6.5-sunxi64
    
    No end-user support: built from trunk

     

    Paste fericit tuturor :-) [ yes for some among us Easter is now ]

    easter.jpg

  6. The answer coming from FriendlyArm is not promising either:  " Sorry, we can't buy the h5 chip now. It will be difficult for us to continue to produce NEO2 Black boards.  "

    This comes close to what you described @Werner

    Luckily I just ordered a few NEO2 blacks from Ali ( at least it seems they were in stock ) - interesting ...

  7. I put a heatsink as listed here at Pine, left OrangePiOnePlus running for three days now ( idle ) with Kernel 5.6.2 and its temperature was 50-51.

    However I just put it now in a more open space and that saves immediately 10 degrees.

     

    I added screenshot in which the OpiOnePlus is receiving data , " iperf3 -s " (  10 parr. threads )  and  once sending data ( also 10 parr. threads ).

    Now things start to move towards 70 degrees once " armbianmonitor -z " is being executed

     

    As as side note it seems the H5 chip runs in idle mode at least 5 degrees lower as both boards are in same location

    test_opioplus.jpg

     

    On overall it might be interesting using this board as a VPN/ piHole server 

  8. ok firing up a new build, wiped armbian completely and looks promising so far - will flash OPiOnePlus shortly:

    [ o.k. ] * [l][c] xxx-sun50i-h6-enable-higher-clock.patch

     

    And indeed looking much better now so no need for Jira  story anymore ;-), danke-gracias-thanks-bedankt-multumesc-hvala etc etc

     

    @orangepioneplus:~# cpufreq-info |tail -1
      cpufreq stats: 480 MHz:83.61%, 720 MHz:0.92%, 816 MHz:0.20%, 888 MHz:0.02%, 1.08 GHz:0.09%, 1.32 GHz:0.11%, 1.49 GHz:8.99%, 1.64 GHz:0.09%, 1.80 GHz:5.98%  (153)

     

    Now let's see how things will hold as H5's show in "idle mode" ( conservative)  a CPU temp of +/- 35°C ( current CPU frequency is 480 MHz )

    while H6 shows CPU temp of 49°C already with passive cooling  . The 1,8Ghz would be needed only when I synch my NAS with remote site for a couple of minutes.

    For regular TV streaming CPU stays at mostly at 480Mhz (  cpufreq stats: 408 MHz:85.35% ) so I hope active cooling won't be needed because of the 15 degrees temp difference in idle mode already

     

  9. 1 hour ago, Werner said:

    Fixed.

     

    @Wernermoin moin I had to put it in sunx-dev of course,

    me bad however still NOK " only garbage was found in the patch input " - seems I am overlooking something else now

    ouput.log
    ---
    [ warn ] * [u][c] xxx-sun50i-h6-enable-higher-clock.patch [ failed ]
    
    patching.log
    ---
    Processing file /home/armbian/userpatches/kernel/sunxi-dev/xxx-sun50i-h6-enable-higher-clock.patch
    patch: **** Only garbage was found in the patch input.


     

  10. @Werner Err check my previous post, NOK

    I did this previously ( exact steps below ) and even the OPiOnePlus got a nice reboot

     

    Just to be certain this is what my build says : " Linux orangepioneplus1 5.6.2-sunxi64 #trunk SMP Tue Apr 7 09:22:09 CEST 2020 aarch64 GNU/Linux "

    Next attempt when 5.6.3 is incorporated in to dev build

    cd ~/armbian/userpatches     # neither in ~armbian/userpatches/kernel/sunxi-current
    wget https://github.com/armbian/build/blob/master/patch/kernel/sunxi-current/xxx-sun50i-h6-enable-higher-clock.patch
    
    cd..
    ./compile.sh BRANCH=dev RELEASE=buster BUILD_MINIMAL=yes BUILD_DESKTOP=no KERNEL_ONLY=no KERNEL_CONFIGURE=no BUILD_EXTERNAL=no EXTRAWIFI=no

     

  11. @Werner would you mind allowing my the QnD way by adding your " sun50i-h6-orangepi-one-plus.dtb " to your post so I should not need to recompile and just replace current file in " /boot/dtb/allwinner " ...:rolleyes:

     

    [ current state meanwhile after compiling :]

      CC [M]  drivers/staging/rtl8723cs-new/core/rtw_mp.o
      LD [M]  drivers/staging/rtl8723cs-new/8723cs.o
    Makefile:1685: recipe for target 'drivers' failed
    [ error ] ERROR in function compile_kernel [ compilation.sh:378 ]
    [ error ] Kernel was not built [ @host ]
    [ o.k. ] Process terminated

     

    [ this is meanwhile resolved with

    ./compile.sh BRANCH=dev RELEASE=buster BUILD_MINIMAL=yes BUILD_DESKTOP=no KERNEL_ONLY=no KERNEL_CONFIGURE=no BUILD_EXTERNAL=no WIREGUARD=no EXTRAWIFI=no AUFS=no ]

  12. unfortunately not on my OpiOnePlus, at least with this config

    @orangepioneplus1:~$ cat /etc/default/cpufrequtils
    ENABLE=true
    MIN_SPEED=480000
    MAX_SPEED=1810000
    GOVERNOR=conservative

     

    Board shows after "reboot" :

    root@orangepioneplus1:~# cpufreq-info |grep cpufreq
    cpufrequtils 008: cpufreq-info (C) Dominik Brodowski 2004-2009
    Report errors and bugs to cpufreq@vger.kernel.org, please.
      driver: cpufreq-dt
      available cpufreq governors: conservative, userspace, powersave, ondemand, performance, schedutil
      cpufreq stats: 480 MHz:85.35%, 720 MHz:4.05%, 816 MHz:0.51%, 888 MHz:0.95%, 1.08 GHz:1.39%, 1.32 GHz:1.31%, 1.49 GHz:6.45%  (179)
      driver: cpufreq-dt
      available cpufreq governors: conservative, userspace, powersave, ondemand, performance, schedutil
      cpufreq stats: 480 MHz:85.35%, 720 MHz:4.05%, 816 MHz:0.51%, 888 MHz:0.95%, 1.08 GHz:1.39%, 1.32 GHz:1.31%, 1.49 GHz:6.45%  (179)
      driver: cpufreq-dt
      available cpufreq governors: conservative, userspace, powersave, ondemand, performance, schedutil
      cpufreq stats: 480 MHz:85.35%, 720 MHz:4.05%, 816 MHz:0.51%, 888 MHz:0.95%, 1.08 GHz:1.39%, 1.32 GHz:1.31%, 1.49 GHz:6.45%  (179)
      driver: cpufreq-dt
      available cpufreq governors: conservative, userspace, powersave, ondemand, performance, schedutil
      cpufreq stats: 480 MHz:85.36%, 720 MHz:4.05%, 816 MHz:0.51%, 888 MHz:0.95%, 1.08 GHz:1.39%, 1.32 GHz:1.31%, 1.49 GHz:6.45%  (179)

     

    As well armbian-config not showing 1,8GHz , did not find an overlay setting ( like H5 ) so I guess this is WIP and will await a permanent solution.

    for now I am OK as it boots in to the new kernel and therefore I can access and play with both boards  

  13. 13 hours ago, martinayotte said:

    Switch is not committed

     

    Err I assume it has been committed after all, juist built first 5.6 image :  " Armbian_20.05.0-trunk_Orangepioneplus_buster_dev_5.6.2_minimal.img " using BRANCH=dev RELEASE=buster BUILD_MINIMAL=yes BUILD_DESKTOP=no KERNEL_ONLY=no KERNEL_CONFIGURE=no. Now after flashing I noiced

     

    1/ no 1.8 GHz (yet) or this needs to be added specifically to overlay first, eg " cpu-clock-X-Yv "

    available frequency steps: 480 MHz, 720 MHz, 816 MHz, 888 MHz, 1.08 GHz, 1.32 GHz, 1.49 GHz

     

    2/ wireguard is now incorporated in to kernel 5.6 so I assume flag " WIREGUARD=no " has no value anymore

     

    etc etc, but yes this looks promising :-), thanks

      ___  ____  _    ___
     / _ \|  _ \(_)  / _ \ _ __   ___   _
    | | | | |_) | | | | | | '_ \ / _ \_| |_
    | |_| |  __/| | | |_| | | | |  __/_   _|
     \___/|_|   |_|  \___/|_| |_|\___| |_|
    
    Welcome to Armbian buster with Linux 5.6.2-sunxi64
    
    No end-user support: built from trunk
    
    System load:   0.01 0.07 0.06   Up time:       6 min            Local users:   2
    Memory usage:  9 % of 988MB     IP:            192.168.10.235
    CPU temp:      43°C
    Usage of /:    5% of 15G

     

     

  14. Wow - Would love to see my H6 and H5 boards performing with this new dev build,

    due to mainlining efforts code being merged in 5.6 as well official kernel support for wg.

     

    Oh yea H6 orange pi one plus now runs on 1.8 again or similar "hack" needed as back then for neo2 v1.1 to " overclock " while 1,5 is default ( 1 GHz on h5 if I remember correctly )

     

    BTW I found something strange on 28 March update ( neo2 ):

     

    /var/log.hdd/apt/term.log {
      rotate 12
      monthly
      compress
      missingok
      notifempty
    }
    /var/log.hdd/apt/history.log {
      rotate 12
      monthly
      compress
      missingok
      notifempty
    }
     

    Where is this coming from in the armbianEnv.txt OR my podgies #fingers ?

     

×
×
  • Create New...

Important Information

Terms of Use - Privacy Policy - Guidelines