Efforts to develop firmware for X96 X6 RK3566 (8G/64G)

[loi xin]

Hi everybody,
I just bought an X96 X6 TV box (8G/64G) to run Home Assistant. However it is not popular so there is no Armina for it. Home Assistant requires Debian 11/12 (linux kernel, no desktop version) to operate. Hope everyone can help me. Thank you very much. I only need 1 file *.img file to install Home Assistant, no need: USB, wifi, Bluetooth... because it can be difficult.

Thank you very much!

[Chris4arm]

I try to run a Armbian server on this nice RK3566 box.
I use the Station M2 image Armbian_community_24.11.0-trunk.167_Station-m2_bookworm_current_6.6.52_minimal.img.xz
with the rk3566-box-demo.dtb dtb and this boots 🙂
But there is only one of the three USB ports working (the middle one).
Is this the best image to use ?
Is there a dtb more accurate to this box ?

Thanks for any help

[Chris4arm]

I have extracted the Android dtb but, of course, it doesn't work as is.

Has somebody had full success with this box ?

 

 

[Chris4arm]

Quote

I have extracted the Android dtb but, of course, it doesn't work as is.

Has somebody had full success with this box ?

 

At least with USB ?

 

[Vincenzoernst1]

do you have any pics of the pcb?!


also this seems to be be similar to the h96-max-v56-rk3566 which also has a 8GB RAM version

[Chris4arm]

Hi,

I have no pics of the pcb but it seems @maka has done a dtb for this box, but I don't see such a dtb in  /boot/dtb/rockchip 🤔

[Chris4arm]

I retrieved the DTB rk3566-box-demo_x96_x6.dtb in the thread dedicated to the H96 Max and... all the USB ports are ok 🙂
But on the other hand the ethernet port does not work anymore : it is up and the negotiation is done well (1000baseT) but, if what the router emits is well received by the box, the latter does not see what it emits.
Strange: if we start with the dtb rk3566-box-demo.dtb and we restart with rk3566-box-demo_x96_x6.dtb (without power cut) ethernet continues to work!
It is as if an initialization was missing, present in rk3566-box-demo.dtb, absent in rk3566-box-demo_x96_x6.dtb.
I'm trying to find the differences between rk3566-box-demo and rk3566-box-demo_x96_x6 but there are hundreds of them and it's hard to identify the real semantic differences 😞
Any help appreciated...

 

Kernel log :

Quote

rk_gmac-dwmac fe010000.ethernet: IRQ eth_lpi not found
rk_gmac-dwmac fe010000.ethernet: IRQ sfty not found
rk_gmac-dwmac fe010000.ethernet: supply phy not found, using dummy regulator
rk_gmac-dwmac fe010000.ethernet: clock input or output? (input).
rk_gmac-dwmac fe010000.ethernet: TX delay(0x4f).
rk_gmac-dwmac fe010000.ethernet: RX delay(0x2d).
rk_gmac-dwmac fe010000.ethernet: integrated PHY? (no).
rk_gmac-dwmac fe010000.ethernet: clock input from PHY
rk_gmac-dwmac fe010000.ethernet: init for RGMII
rk_gmac-dwmac fe010000.ethernet: User ID: 0x30, Synopsys ID: 0x51
rk_gmac-dwmac fe010000.ethernet:        DWMAC4/5
rk_gmac-dwmac fe010000.ethernet: DMA HW capability register supported
rk_gmac-dwmac fe010000.ethernet: RX Checksum Offload Engine supported
rk_gmac-dwmac fe010000.ethernet: TX Checksum insertion supported
rk_gmac-dwmac fe010000.ethernet: Wake-Up On Lan supported
rk_gmac-dwmac fe010000.ethernet: TSO supported
rk_gmac-dwmac fe010000.ethernet: Enable RX Mitigation via HW Watchdog Timer
rk_gmac-dwmac fe010000.ethernet: Enabled RFS Flow TC (entries=10)
rk_gmac-dwmac fe010000.ethernet: TSO feature enabled
rk_gmac-dwmac fe010000.ethernet: Using 32/32 bits DMA host/device width
rk_gmac-dwmac fe010000.ethernet end0: renamed from eth0
rk_gmac-dwmac fe010000.ethernet end0: Register MEM_TYPE_PAGE_POOL RxQ-0
rk_gmac-dwmac fe010000.ethernet end0: PHY [stmmac-0:01] driver [RTL8211F Gigabit Ethernet] (irq=POLL)
dwmac4: Master AXI performs any burst length
rk_gmac-dwmac fe010000.ethernet end0: No Safety Features support found
rk_gmac-dwmac fe010000.ethernet end0: IEEE 1588-2008 Advanced Timestamp supported
rk_gmac-dwmac fe010000.ethernet end0: registered PTP clock
rk_gmac-dwmac fe010000.ethernet end0: configuring for phy/rgmii link mode
rk_gmac-dwmac fe010000.ethernet end0: Link is Up - 1Gbps/Full - flow control rx/tx
rk_gmac-dwmac fe010000.ethernet end0: Link is Down
rk_gmac-dwmac fe010000.ethernet end0: Link is Up - 1Gbps/Full - flow control rx/tx
rk_gmac-dwmac fe010000.ethernet end0: Link is Down
rk_gmac-dwmac fe010000.ethernet end0: Link is Up - 1Gbps/Full - flow control rx/tx

 

[maka]

Tx and Rx delay maybe wrong, compare with the android dtb if It worked or change to the box- demo values.

 

Have you tried the icplus dtb?

 

Edited November 25 by maka

[Chris4arm]

HI,

TX and RX delays are the same between rk3566-box-demo which works and rk3566-box-demo_x96_x6 which don't.
I tried with Android  values, slightly different (2b/4b vs 2d/4f), without success.

I tried rk3566-box-demo_x96_ip1001m.dtb with same results : the box receive packets but hers are not received/emitted.
Here are the differences of dwmac logs between rk3566-box-demo_x96_x6 and rk3566-box-demo_x96_ip1001m :
 

Quote

# diff dwmac-box-demo-x96-x6.txt dwmac-rk3566-box-demo_x96_ip1001m.txt
4c4
< rk_gmac-dwmac fe010000.ethernet: clock input or output? (input).
---
> rk_gmac-dwmac fe010000.ethernet: clock input or output? (output).
8,9c8
< rk_gmac-dwmac fe010000.ethernet: clock input from PHY
< rk_gmac-dwmac fe010000.ethernet: init for RGMII
---
> rk_gmac-dwmac fe010000.ethernet: init for RGMII_TXID
23c22
< rk_gmac-dwmac fe010000.ethernet end0: PHY [stmmac-0:01] driver [RTL8211F Gigabit Ethernet] (irq=POLL)
---
> rk_gmac-dwmac fe010000.ethernet end0: PHY [stmmac-0:01] driver [ICPlus IP1001] (irq=POLL)
28c27
< rk_gmac-dwmac fe010000.ethernet end0: configuring for phy/rgmii link mode
---
> rk_gmac-dwmac fe010000.ethernet end0: configuring for phy/rgmii-txid link mode

mii-tool box-demo :
 

Quote

Using SIOCGMIIPHY=0x8947
end0: negotiated 1000baseT-FD flow-control, link ok
  registers for MII PHY 1:
    1040 79ad 001c c916 0de1 cde1 006f 0000
    0000 0200 4800 0000 0000 0000 0000 2000
    0000 0000 0000 0000 0000 0400 0f00 0f00
    2118 0863 39ee 0002 0000 0000 0000 0000
  product info: vendor 00:e0:4c or 00:07:32, model 17 rev 6
  basic mode:   autonegotiation enabled
  basic status: autonegotiation complete, link ok
  capabilities: 1000baseT-FD 100baseTx-FD 100baseTx-HD 10baseT-FD 10baseT-HD
  advertising:  1000baseT-FD 100baseTx-FD 100baseTx-HD 10baseT-FD 10baseT-HD flow-control
  link partner: 1000baseT-FD 100baseTx-FD 100baseTx-HD 10baseT-FD 10baseT-HD flow-control

mii-tool box-demo-x96-x6 :
 

Quote

end0: negotiated 1000baseT-FD flow-control, link ok
  registers for MII PHY 1:
    1040 79ad 001c c916 0de1 cde1 006d 0000
    0000 0200 0800 0000 0000 0000 0000 2000
    0000 0000 0000 0000 0000 0400 0f00 0f00
    2118 0863 31ee 0002 6295 0000 0000 0000
  product info: vendor 00:e0:4c or 00:07:32, model 17 rev 6
  basic mode:   autonegotiation enabled
  basic status: autonegotiation complete, link ok
  capabilities: 1000baseT-FD 100baseTx-FD 100baseTx-HD 10baseT-FD 10baseT-HD
  advertising:  1000baseT-FD 100baseTx-FD 100baseTx-HD 10baseT-FD 10baseT-HD flow-control
  link partner: 1000baseT-FD 100baseTx-FD 100baseTx-HD 10baseT-FD 10baseT-HD flow-control

Only the registers are different.

Interestingly if one connect the box (with box-demo-x96-x6 dtb) on a 100baseT interface, the behavior is the same (one way communication) while the delays are much less important.

What could be missing in the box-demo-x96-x6 dtb to prevent output  communication? a power supply ?

 

[maka]

You almost get the same result with rtl or icplus drivers so It seems that you have another kind of chip for ethernet.

[Chris4arm]

ok but why does it work with rk3566-box-demo and not with rk3566-box-demo_x96_x6  ? 🤔

[maka]

Maybe the kernel has made the old dtbs broken. Must be that because rockchip get patches every day.

 

You have one USB otg. Its not that enough?

Edited November 27 by maka

[Chris4arm]

I bought this box because it has an external antenna (good wifi performance), a fast ethernet port  and 3 USB ports for a SSD, a  SDR and a monitoring system (without external hub 😉)...

Ethernet works on your X96 X6 box ?

[maka]

Yes but is ICPLUS chip

 

You can try this, maybe the chip is awaked

 

mdio {
            compatible = "snps,dwmac-mdio";
            #address-cells = < 0x01 >;
            #size-cells = < 0x00 >;

            ethernet-phy@1 {
                compatible = "ethernet-phy-ieee802.3-c22"; ====> compatible = "ethernet-phy-id001c.c916", "ethernet-phy-ieee802.3-c22";

 

Edited November 27 by maka

[Chris4arm]

3 hours ago, maka said:

You can try this, maybe the chip is awaked

 

mdio {
            compatible = "snps,dwmac-mdio";
            #address-cells = < 0x01 >;
            #size-cells = < 0x00 >;

            ethernet-phy@1 {
                compatible = "ethernet-phy-ieee802.3-c22"; ====> compatible = "ethernet-phy-id001c.c916", "ethernet-phy-ieee802.3-c22";

 

Thanks for your advice but there is no difference: the box receives packets but don't emit.

The logs are identical as those of box-demo and if we boot with box-demo then reboot (without power off) with box-demo_x96_x6 modified, ethernet continues to work !

It seems box-demo does something, which survives reboot, which is missing in box-demo_x96_x6.

Any idea to find that ?

 

[maka]

I have made your work and mend the file that I think you are reporting to work well (I am not sure because your reports are not traceable)

 

I hope you are lucky.

rk3566-box-demo_x96x6_24.dtb

Edited November 28 by maka

[Chris4arm]

I'm lucky and you are great !
ethernet works two way now (936 Mbits/sec) 🙂
what was the trick ?
Now I'll look what is needed for bluetooth and the led display...
Many thanks @maka

[maka]

Maybe I just deserve one like?

[Chris4arm]

Of course ! I ignored there was a like button : it's almost invisible on my screen 🫢

[Chris4arm]

For bluetooth it's just a matter of

 

Quote

<             compatible = "brcm,bcm43438-bt";
---
>             compatible = "realtek,rtl8822cs-bt";