Jump to content

Next LTS kernel 4.19.y Allwinner A10, A20, A64, H2+, H3, H5, H6 debugging party

Igor

By Igor
in Advanced users - Development

 Share

Recommended Posts

belfastraven

belfastraven

Posted
Posted

Re Pine64 WiFi.  Using a USB Wifi adapter, WiFi is fine.  The Wifi/bluetooth module told by Piine64 is not working.  I can load the driver,  but the SDIO device is not being enumerated.     Still looking into the issue...

vlad59

vlad59

Posted
Posted
  On 11/23/2018 at 3:35 PM, martinayotte said:

It is probably something missing in the Main DT.

Expand  

 

Strange wifi is working on my Pine64 but I have an old SDIO adaptor :

  

root@pine64:~# dmesg | grep 8723
[    5.240274] r8723bs: module is from the staging directory, the quality is unknown, you have been warned.
[    5.287430] RTL8723BS: module init start
[    5.287439] RTL8723BS: rtl8723bs v4.3.5.5_12290.20140916_BTCOEX20140507-4E40
[    5.287441] RTL8723BS: rtl8723bs BT-Coex version = BTCOEX20140507-4E40
[    5.397757] RTL8723BS: rtw_ndev_init(wlan0)
[    5.406514] RTL8723BS: module init ret =0
[   16.270611] rtl8723bs: acquire FW from file:rtlwifi/rtl8723bs_nic.bin

 

belfastraven

belfastraven

Posted
Posted

Yes,  I just tried another pine64 sdio adapter and device is now working.   I don't know how I managed to kill the first one..

 

 

 5.433718] r8723bs: module is from the staging directory, the quality is unknown, you have been warned.
[    5.481983] RTL8723BS: module init start
[    5.481994] RTL8723BS: rtl8723bs v4.3.5.5_12290.20140916_BTCOEX20140507-4E40
[    5.481997] RTL8723BS: rtl8723bs BT-Coex version = BTCOEX20140507-4E40
[    5.522202] RTL8723BS: rtw_ndev_init(wlan0)
[    5.524817] RTL8723BS: module init ret =0
[   10.488627] rtl8723bs: acquire FW from file:rtlwifi/rtl8723bs_nic.bin
[   14.397509] RTL8723BS: rtw_set_802_11_connect(wlan0)  fw_state = 0x00000008
[   14.535050] RTL8723BS: start auth
[   14.537783] RTL8723BS: auth success, start assoc
[   14.548427] RTL8723BS: rtw_cfg80211_indicate_connect(wlan0) BSS not found !!
[   14.548446] RTL8723BS: assoc success
[   14.558911] RTL8723BS: send eapol packet
[   14.580272] RTL8723BS: send eapol packet
[   14.618036] RTL8723BS: set pairwise key camid:4, addr:92:3b:ad:ad:a7:34, kid:0, type:AES
[   14.618588] RTL8723BS: set group key camid:5, addr:92:3b:ad:ad:a7:34, kid:1, type:AES
 

 

 I  haven't checked bluetooth yet.   It isn't defined in the megous linux DTS,  but from Anarsouls--

 

Defined here in Anarsouls DTS

/* On Wifi/BT connector, with RTS/CTS */
&uart1 {
    pinctrl-names = "default";
    pinctrl-0 = <&uart1_pins>, <&uart1_rts_cts_pins>;
    status = "okay";

    bluetooth {
        compatible = "realtek,rtl8723bs-bt";
        reset-gpios = <&r_pio 0 4 GPIO_ACTIVE_LOW>; /* PL4 */
        device-wake-gpios = <&r_pio 0 5 GPIO_ACTIVE_HIGH>; /* PL5 */
        host-wake-gpios = <&r_pio 0 6 GPIO_ACTIVE_HIGH>; /* PL6 */
        firmware-postfix = "pine64";
    };
};

 

 

 

vlad59

vlad59

Posted
Posted

@belfastraven

 

It's great you also got it working. I bought 3 pine64 + 3 SDIO Wifi / BT addons and I got a faulty one, fortunately I tested all 3 three and I got another one for free ;).

 

About bluetooth I also saw Anarsoul kernel, In case you didn't saw you may have to tweak the firmware blobs :

 * https://forum.manjaro.org/t/bluetooth-not-working-with-rtl8723bs-bt/51906/10

 * https://github.com/lwfinger/rtl8723bs_bt/issues/28#issuecomment-432806835

 

If nobody beats me to it, I'll probably try to fix bluetooth next week.

lomady

lomady

Posted
Posted
  On 11/18/2018 at 7:54 PM, Igor said:


Mine hanged with current stock settings already at first boot ... then I tried orangepipc2 DT and ... it looks fine. I notice DT is missing regulator bits and after adding them ... at least that problem is gone. I did several reboots and cold starts. At one cold start it hanged ... Not sure if everything is fine yet.

Expand  

@IgorShould this update be available via apt-get update && apt-get upgrade?

Look like daily builds stopped on Nov 18th :huh:

martinayotte

martinayotte

Posted
Posted
  On 11/23/2018 at 6:49 PM, vlad59 said:
  On 11/23/2018 at 4:38 PM, martinayotte said:

Using which kernel ?

I don't see any SDIO node in Mainline 4.19.y DT for Pine64 ...

Expand  

Latest build from Armbian dev 4.19.2

Expand  

Still strange ... Are you using pinebook image on a pine64 board ?

Because the SDIO node doesn't exist on Pine64 DT, but it is there for Pinebook ...

belfastraven

belfastraven

Posted
Posted

FWIW,  with my working  hardware module,  this is where the device enumerates. 

 

/sys/bus/sdio/devices/mmc1:0001:1

 

wlan0 show up in that directory tree.  

 

and there is this entry in the device tree:

 

&mmc1 {
    pinctrl-names = "default";
    pinctrl-0 = <&mmc1_pins>;
    vmmc-supply = <&reg_dldo4>;
    vqmmc-supply = <&reg_eldo1>;
    mmc-pwrseq = <&wifi_pwrseq>;
    non-removable;
    bus-width = <4>;
    status = "okay";
};

 

I just compiled a dtb with the bluetooth node,   and will see what else it takes to make bluetooth work.  vlad59  gave a lot of helpful information in his last post, also.   I know that the wifi driver and firmware were loaded automatically at boot, once the hardware module was recognized.  I notice that Bluetooth enabling can be done in armbian-config,  but am not sure what has to be in place for that to work.  

 

 

Perier

Perier

Posted
Posted

Orange Pi Zero Plus (H5) 

http://ix.io/1ujE

 

I can not understand what kind of error or what is it connected with? (present since version 4.14.y) 

[    6.888226] 29 
[    6.888237] 81 
[    6.888239] 03 
[    6.888241] CC 
[    6.888243] 00 
[    6.888245] 00 
[    6.888247] 50 
[    6.888249] 00 
[    6.888251] 00 
[    6.888253] 00 
[    6.888255] 04 
[    6.888257] CC 
[    6.888259] 0A 
[    6.888262] 0C 
[    6.888264] 00 
[    6.888266] 00
[    6.888268] FF

and other errors:

  Reveal hidden contents

platform regulatory.0 (error) - appeared in the dev branch, there is none in the mainline branch

 

CPU temp, without heatsink:

  Reveal hidden contents

 

TonyMac32

TonyMac32

Posted
Posted

@Evgeny would you be able to put the code blocks into the Spoiler containers so they don't hog an entire page?

 

As to the question, I assume the Opi zero plus has a "dumb" regulator setup with no voltage switching? (OK, found it, it will do 1.1 and 1.3V).  Basically you have no voltage to support those higher operating points, the processor will theoretically fall on it's face if you tried. 

 

As for the temp and random hex showing up, I do not know.

Perier

Perier

Posted
Posted
  On 11/25/2018 at 6:00 AM, TonyMac32 said:

@Evgeny would you be able to put the code blocks into the Spoiler containers so they don't hog an entire page?

 

As to the question, I assume the Opi zero plus has a "dumb" regulator setup with no voltage switching? (OK, found it, it will do 1.1 and 1.3V).  Basically you have no voltage to support those higher operating points, the processor will theoretically fall on it's face if you tried. 

 

As for the temp and random hex showing up, I do not know.

Expand  

Thank you for a hint about a spoiler (corrected)

 

HEX-error migrated from the mainline 4.14.y

 

 

Igor

Igor

Posted
  • Author
Posted
  On 11/25/2018 at 6:00 AM, TonyMac32 said:

As for the temp and random hex showing up, I do not know.

Expand  


It's related to wireless drivers, probably showing firmware loading progress. In any case safe to ignore.

Perier

Perier

Posted
Posted
  On 11/25/2018 at 7:44 AM, Igor said:


It's related to wireless drivers, probably showing firmware loading progress. In any case safe to ignore.

Expand  

But how did you determine that the problem in loading wireless drivers? (for own development)

guidol

guidol

Posted
Posted

Yesterday I did compile 2 variants:

Armbian_5.67_Orangepizeroplus_Debian_stretch_dev_4.19.4

and
Armbian_5.67_Orangepizeroplus2-h5_Debian_stretch_dev_4.19.4

 

The OPi Zero Plus (a H5) does boot and I did create a report.

 

The OPi Zero Plus2 H5 doesnt seem to boot (right)
I got no serial TTL-Port connected, because I do use the original black Cube-Case,

but I can see the green Power LED and a few seconds a red Heartbeat while powering up.

But after the 1.-2 seconds the Heartbeat will go out and no HDMI, Ethernet (USB activity - which normally works as eth0 DHCP)

 

So from me no report for the OPi Zero Plus2 H5 at this time

PS: I have to check the flashing of the card, because etcher version did jump from 1.4.5 to 1.4.8
Named now BalenaEtcher with a new Icon and Adverts :(

Balena_Etcher_1_48_with_adverts.jpg

Igor

Igor

Posted
  • Author
Posted
  On 11/26/2018 at 4:45 PM, guidol said:

The OPi Zero Plus2 H5 doesnt seem to boot (right)

Expand  


Yes, there is something wrong ... with/around powering wireless. I haven't resolved yet.
 

  On 11/26/2018 at 4:45 PM, guidol said:

Named now BalenaEtcher with a new Icon and Adverts

Expand  

 

Is there any libretcher? :)

guidol

guidol

Posted
Posted
  On 11/26/2018 at 6:56 PM, Igor said:

Is there any libretcher? :)

Expand  

Nope :(

1.4.7 renamed etcher to balena-etcher and

1.4.8 does include now the "featured project" advert

etcher changelog:
 

  Reveal hidden contents

 

guidol

guidol

Posted
Posted
  On 11/16/2018 at 5:48 AM, TonyMac32 said:

I see NanoPi K1+ is done already, so my work here is done.  :-)

Expand  

I did create a report for NanoPi K1 Plus-dev, but using the HDMI-console image (not Desktop).
I'll use Desktop on SBCs when I got one with 4Gb of Ram - 2GB is not much enough for GUI and Web-Browser in my opinion ;) 

TonyMac32

TonyMac32

Posted
Posted

For 64-bit you are probably correct. On the other side, the Tinker board web browses quite well with it's 2GB of RAM.

Sent from my Pixel using Tapatalk

lomady

lomady

Posted
Posted (edited)
  On 11/18/2018 at 7:54 PM, Igor said:

Mine hanged with current stock settings already at first boot ... then I tried orangepipc2 DT and ... it looks fine. I notice DT is missing regulator bits and after adding them ... at least that problem is gone. I did several reboots and cold starts. At one cold start it hanged ... Not sure if everything is fine yet.

Expand  

Did another fresh install and switch to daily dev 27 Nov. Cold boot hangs at various stages with HDMI image distortions:

  Reveal hidden contents

I am sure that my Orange PC Prime is not defective because I have stress tested it with official Arch linux image (kernel 3.10)

Since I got two of these boards - I tried another one. It boots but with kernel oops'es at random:

  Reveal hidden contents

Can I suggest that it is something with DDR settings?

Edited by lomady
Tried second board
vlad59

vlad59

Posted
Posted

In case that's useful, I reinstalled my Orange Pi +2E with latest DEV image. Everything seems to work (I can send a report if needed but the current one is quite recent).

 

This morning I checked `dmesg` output for something else and I saw these three red lines (full armbianmonitor output is http://ix.io/1uAk) 

[    6.087868] sun4i-codec 1c22c00.codec: ASoC: codec-analog@1f015c0 not registered
[    6.087882] sun4i-codec 1c22c00.codec: Failed to register our card
[    6.140691] sun8i_ths 1c25000.thermal-sensor: no memory resources defined

I have no way to check if audio is working for now (I'm at work). But I checked DVFS / Thermal protection and it's works perfectly fine.

 

hojnikb

hojnikb

Posted
Posted

Any easy way to modify DVFS tables ? Or perhaps set a set frequency for testing purposes.

 

I'd like to test out each frequency point (and do some undervolting) to see how power consumption responds. This used to be doable with a single command, not sure if it's possible now on newer kernel.

 

Thanks

Igor

Igor

Posted
  • Author
Posted
  On 11/27/2018 at 7:23 PM, hojnikb said:

Any easy way to modify DVFS tables ? Or perhaps set a set frequency for testing purposes.

 

I'd like to test out each frequency point (and do some undervolting) to see how power consumption responds. This used to be doable with a single command, not sure if it's possible now on newer kernel.

 

Thanks

Expand  

 

For H6?
https://github.com/armbian/build/blob/master/patch/kernel/sunxi-dev/board-h6-clock-higher.patch
arch/arm64/boot/dts/allwinner/sun50i-h6.dtsi

You can decompile DTB, edit and compile back. No need to recompile kernel.

Igor

Igor

Posted
  • Author
Posted
  On 11/28/2018 at 7:02 AM, hojnikb said:

Sorry, i meant for H5. I assume it's the same procedure ?

Expand  

 

Yes. just location is elsewhere if you look from kernel sources perspective.  IIRC its inside common sunxi-h3-h5 include file which is under arch/arm/boot/dts not arm64 ... which is a bit confusing.

Guest
This topic is now closed to further replies.
 Share
Go to topic listing
×
×
  • Create New...

Important Information

Terms of Use - Privacy Policy - Guidelines

  I accept