QCA9888: Board file for Kuwfi WB5G08 #56

grosjo · 2024-12-28T10:35:41Z

No description provided.

hauke · 2025-01-04T17:34:53Z

@Ansuel or @robimarko Is this looking fine?

It is related to this PR: openwrt/openwrt#17281

grosjo · 2025-01-04T23:44:44Z

yes

grosjo · 2025-01-05T15:39:50Z

@hauke once this PR is approved, what is the process to use the board file in my PR openwrt/openwrt#17281 ?

testuser7 · 2025-01-05T22:12:20Z

@hauke once this PR is approved, what is the process to use the board file in my PR openwrt/openwrt#17281 ?

You can then add a commit to your PR like this: https://github.com/openwrt/openwrt/pull/17254/commits

grosjo · 2025-01-06T12:56:42Z

I still don't understand. Where should I place the board.bin file in my PR file tree to have it included ? Or are you referring to something else ?

robimarko · 2025-01-06T13:24:10Z

This is not valid, its not even wrapped by ath10k-bdencoder, it just looks like plain BDF

grosjo · 2025-01-07T00:23:15Z

This is not valid, its not even wrapped by ath10k-bdencoder, it just looks like plain BDF

This is the board.bin from the initial firmware

What should I do to make it valid ?

robimarko · 2025-01-08T13:40:03Z

Yes, but its just plain BDF.

For use in OpenWrt you need to wrap it with ath10k-bdencoder

grosjo · 2025-01-09T04:58:14Z

@robimarko Can you explain how ? No docs available here...

testuser7 · 2025-01-09T09:41:15Z

@grosjo ath10k-bdencoder tool: https://github.com/qca/qca-swiss-army-knife/blob/master/tools/scripts/ath10k/ath10k-bdencoder

Rename your BDF to bus=pci,bmi-chip-id=0,bmi-board-id=16,variant=KuWfi-WB5G08.bin (your board-id is 16)

Create board-2.json file:

[
    {
        "names": [
            "bus=pci,bmi-chip-id=0,bmi-board-id=16,variant=KuWfi-WB5G08"
        ],
        "data": "bus=pci,bmi-chip-id=0,bmi-board-id=16,variant=KuWfi-WB5G08.bin"
    }
]

Run ath10k-bdencoder -c board-2.json -o board-kuwfi_wb5g08.qca9888

grosjo · 2025-01-09T10:07:08Z

@robimarko @testuser7
Done

grosjo · 2025-01-13T00:44:11Z

How to move forward ?

grosjo · 2025-01-18T10:37:27Z

any way to move forward ?

robimarko · 2025-01-19T10:04:50Z

Sorry, but this is not correct:

[
    {
     	"names": [
            "bus=pci,vendor=168c,device=0056,subsystem-vendor=0000,subsystem-device=0000"
        ],
	"data": "bus=pci,vendor=168c,device=0056,subsystem-vendor=0000,subsystem-device=0000.bin"
    }
]

How do you plan to match the variant when there is not one

grosjo · 2025-01-20T07:10:12Z

What do you mean ? This is the board related to the PR associated

What do you want to say ?

testuser7 · 2025-01-20T07:21:12Z

@grosjo The filename bus=pci,vendor=168c,device=0056,subsystem-vendor=0000,subsystem-device=0000 is invalid. Why didn't you use the name bus=pci,bmi-chip-id=0,bmi-board-id=16,variant=KuWfi-WB5G08 that I gave you earlier?

The name you used is used for qca9889. But if you must use it, add a variant at the end: bus=pci,vendor=168c,device=0056,subsystem-vendor=0000,subsystem-device=0000,variant=KuWfi-WB5G08

grosjo · 2025-01-20T07:23:21Z

I have not done anything, this is just the bin file from the OEM board

testuser7 · 2025-01-20T09:41:18Z

So you need to change the name inside bard file to bus=pci,vendor=168c,device=0056,subsystem-vendor=0000,subsystem-device=0000,variant=KuWfi-WB5G08.

grosjo · 2025-01-23T05:14:54Z

How can I do that ?

grosjo · 2025-01-23T05:21:48Z

I tried something, I have no real clue what I am doing, but maybe it has worked ?

grosjo · 2025-01-23T13:55:27Z

@robimarko @testuser7 Looks good now ?

grosjo · 2025-01-25T09:40:40Z

What is your feedback ?

testuser7 · 2025-01-25T23:00:02Z

You have now used the name bus=pci,bmi-chip-id=0,bmi-board-id=16,variant=KuWfi-WB5G08. Can you show the system logs from when the BDF file is read? Is the name recognized correctly?

grosjo · 2025-02-04T13:08:32Z

No, as tested already before, it does not

grosjo · 2025-02-05T04:48:07Z

There are only 2 options
1 - raw BDF as board.bin without precal/calibration input in the DTS
2 - "embedded" board-2.bin with precal/cal

Usage of the 2 (precal and cal) is actually what they are doing in the initial OEM firmware (wifi0.cal and wifi1.cal)

Can we now move forward ? What would be the step to have this board-2.bin merged in the compilation tree ?

grosjo · 2025-02-06T11:45:34Z

@testuser7 @robimarko what would be the next step ?

grosjo · 2025-02-08T10:37:50Z

How to move forward ?

testuser7 · 2025-02-08T11:02:49Z

The current implementation cannot be accepted because it is incorrect. Check the comments in the main PR.
What MAC address do you have for this radio interface in OpenWrt?

grosjo · 2025-02-08T13:50:52Z

What is incorrect ?

grosjo · 2025-02-08T14:51:54Z

# hexdump /dev/mtdblock6
0000000 7c27007c 3c5a003c edf400ed ffff00ff ffff00ff ffff00ff ffff00ff ffff00ff
0000010 ffff00ff ffff00ff ffff00ff ffff00ff ffff00ff ffff00ff ffff00ff ffff00ff
*
0001000 ffff00ff 7c27007c 3c5a003c edf500ed ffff00ff ffff00ff ffff00ff ffff00ff
0001010 ffff00ff ffff00ff ffff00ff ffff00ff ffff00ff ffff00ff ffff00ff ffff00ff
*
0005000 202f0020 e80000 1010001 7c27007c 3c5a003c edf600ed 0000 20000020
0005010 15e00015 20000 600000 0000 15770015 2000002 0000 0000
0005020 10000010 20000 33000033 0000 20000 5030005 4d43004d 4e5f004e
0005030 4e4f004e 545f0054 464f0046 554e0055 44000044 0000 0000 0000
0005040 c3e000c 10000 0000 0000 330000 33330033 f4f200f4 f20000f2
0005050 4b14004b c00000c ffff00ff ff3f00ff ffff00ff ff3f00ff ffff00ff ff3f00ff
0005060 3140003 0000 0000 0000 0000 0000 0000 0000
0005070 0000 0000 0000 0000 0000 0000 0000 0000
*
(..)

# ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1000
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host 
       valid_lft forever preferred_lft forever
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel master br-lan state UP qlen 1000
    link/ether 7c:27:3c:5a:ed:f4 brd ff:ff:ff:ff:ff:ff
3: br-lan: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP qlen 1000
    link/ether 7c:27:3c:5a:ed:f4 brd ff:ff:ff:ff:ff:ff
    inet 192.168.1.1/24 brd 192.168.1.255 scope global br-lan
       valid_lft forever preferred_lft forever
4: phy0-ap0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master br-lan state UP qlen 1000
    link/ether 12:34:56:78:90:12 brd ff:ff:ff:ff:ff:ff

grosjo · 2025-02-08T15:24:15Z

To avoid the 12345678901 mac, I need to change my board.json as follows:

[
    {
        "names": [
            "bus=pci,vendor=168c,device=0056,subsystem-vendor=0000,subsystem-device=0000,variant=KuWfi-WB5G08",
            "bus=pci,bmi-chip-id=0,bmi-board-id=31,variant=KuWfi-WB5G08",
            "bus=pci,bmi-chip-id=0,bmi-board-id=16,variant=KuWfi-WB5G08"
        ],
        "data": "boardData_2_0_QCA9888_5G_Y9690.bin"
    }
]

grosjo · 2025-02-08T15:33:33Z

Then I get

# ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1000
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host 
       valid_lft forever preferred_lft forever
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel master br-lan state UP qlen 1000
    link/ether 7c:27:3c:5a:ed:f4 brd ff:ff:ff:ff:ff:ff
3: br-lan: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP qlen 1000
    link/ether 7c:27:3c:5a:ed:f4 brd ff:ff:ff:ff:ff:ff
    inet 192.168.1.1/24 brd 192.168.1.255 scope global br-lan
       valid_lft forever preferred_lft forever
    inet6 fe80::7e27:3cff:fe5a:edf4/64 scope link 
       valid_lft forever preferred_lft forever
4: phy0-ap0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master br-lan state UP qlen 1000
    link/ether 7c:27:3c:5a:ed:f6 brd ff:ff:ff:ff:ff:ff

grosjo · 2025-02-10T14:34:21Z

How to proceed further?

Signed-off-by: Joan Moreau <[email protected]>

grosjo · 2025-02-14T06:08:43Z

Ping

testuser7 · 2025-02-14T10:01:46Z

Share a dump of the mtd6 partition. Sometimes you get this result:

hexdump /dev/mtd6
0000000 7c27007c 3c5a003c edeb00ed ffff00ff ffff00ff ffff00ff ffff00ff ffff00ff
0000010 ffff00ff ffff00ff ffff00ff ffff00ff ffff00ff ffff00ff ffff00ff ffff00ff
*
0001000 ffff00ff 7c27007c 3c5a003c edec00ed ffff00ff ffff00ff ffff00ff ffff00ff
0001010 ffff00ff ffff00ff ffff00ff ffff00ff ffff00ff ffff00ff ffff00ff ffff00ff
*
0005000 ffff00ff 526f0052 ffff00ff 7c27007c 3c5a003c eded00ed ffff00ff ffff00ff
0005010 ffff00ff ffff00ff ffff00ff ffff00ff ffff00ff ffff00ff ffff00ff ffff00ff
*
0010000

and other times this:

# hexdump /dev/mtdblock6
0000000 7c27007c 3c5a003c edf400ed ffff00ff ffff00ff ffff00ff ffff00ff ffff00ff
0000010 ffff00ff ffff00ff ffff00ff ffff00ff ffff00ff ffff00ff ffff00ff ffff00ff
*
0001000 ffff00ff 7c27007c 3c5a003c edf500ed ffff00ff ffff00ff ffff00ff ffff00ff
0001010 ffff00ff ffff00ff ffff00ff ffff00ff ffff00ff ffff00ff ffff00ff ffff00ff
*
0005000 202f0020 e80000 1010001 7c27007c 3c5a003c edf600ed 0000 20000020
0005010 15e00015 20000 600000 0000 15770015 2000002 0000 0000
0005020 10000010 20000 33000033 0000 20000 5030005 4d43004d 4e5f004e
0005030 4e4f004e 545f0054 464f0046 554e0055 44000044 0000 0000 0000
0005040 c3e000c 10000 0000 0000 330000 33330033 f4f200f4 f20000f2
0005050 4b14004b c00000c ffff00ff ff3f00ff ffff00ff ff3f00ff ffff00ff ff3f00ff
0005060 3140003 0000 0000 0000 0000 0000 0000 0000
0005070 0000 0000 0000 0000 0000 0000 0000 0000
*
(..)

In the second case there are correct calibration data in the art partition.

grosjo · 2025-02-14T23:43:02Z

different devices, different data

testuser7 · 2025-02-15T01:09:17Z

The problem is that one of your devices does not have calibration data. Such cases are not supported.
The only thing you can do is copy calibration data from one device to another (you need to change the MAC address in the calibration data).

I have added some comments to the main PR.

This one PR is not needed and can be closed.

grosjo · 2025-02-15T05:58:00Z

Why closed ? Devices are not starting up without the board file

testuser7 · 2025-02-15T07:21:19Z

What is the result of the commands?

ls -la /lib/firmware/ath10k/QCA9888/hw2.0/
hexdump -C -n 224 /lib/firmware/ath10k/QCA9888/hw2.0/board.bin
hexdump -C -n 224 /lib/firmware/ath10k/QCA9888/hw2.0/board-2.bin

grosjo · 2025-02-15T09:51:59Z

I remove those files and put the one from firmware to make the device work

grosjo · 2025-02-15T11:26:01Z

Latest changes in DTS (see very latest comment) allow indeed use of the default baord-2.bin

grosjo mentioned this pull request Dec 28, 2024

ath79: add support for KuWfi WB5G08 openwrt/openwrt#17281

Open

grosjo force-pushed the main branch from 187a4cc to bc80acf Compare December 28, 2024 10:41

grosjo changed the title ~~QCA988x: Board file for Kuwfi WG5G08~~ QCA9888: Board file for Kuwfi WG5G08 Jan 5, 2025

grosjo changed the title ~~QCA9888: Board file for Kuwfi WG5G08~~ QCA9888: Board file for Kuwfi WB5G08 Jan 5, 2025

grosjo force-pushed the main branch from bc80acf to 952177c Compare January 9, 2025 10:06

grosjo force-pushed the main branch 2 times, most recently from 461e4d0 to 4a03a92 Compare January 23, 2025 05:21

grosjo force-pushed the main branch from 447efea to a8defed Compare February 6, 2025 11:45

grosjo force-pushed the main branch from a8defed to 327228c Compare February 8, 2025 15:29

grosjo force-pushed the main branch 3 times, most recently from 505481b to 8efcd92 Compare February 8, 2025 15:55

grosjo closed this Feb 9, 2025

grosjo reopened this Feb 9, 2025

grosjo force-pushed the main branch 2 times, most recently from 703da4f to ae77e2e Compare February 9, 2025 14:42

qca9888 : add BDF for Kuwfi WB5G08

9b53492

Signed-off-by: Joan Moreau <[email protected]>

grosjo force-pushed the main branch from ae77e2e to 9b53492 Compare February 14, 2025 06:08

grosjo closed this Feb 15, 2025

QCA9888: Board file for Kuwfi WB5G08 #56

QCA9888: Board file for Kuwfi WB5G08 #56

Conversation

grosjo commented Dec 28, 2024

hauke commented Jan 4, 2025

grosjo commented Jan 4, 2025

grosjo commented Jan 5, 2025

testuser7 commented Jan 5, 2025

grosjo commented Jan 6, 2025

robimarko commented Jan 6, 2025

grosjo commented Jan 7, 2025

robimarko commented Jan 8, 2025

grosjo commented Jan 9, 2025

testuser7 commented Jan 9, 2025

grosjo commented Jan 9, 2025

grosjo commented Jan 13, 2025

grosjo commented Jan 18, 2025

robimarko commented Jan 19, 2025

grosjo commented Jan 20, 2025

testuser7 commented Jan 20, 2025 • edited Loading

grosjo commented Jan 20, 2025

testuser7 commented Jan 20, 2025

grosjo commented Jan 23, 2025

grosjo commented Jan 23, 2025

grosjo commented Jan 23, 2025

grosjo commented Jan 25, 2025

testuser7 commented Jan 25, 2025

grosjo commented Feb 4, 2025

grosjo commented Feb 5, 2025

grosjo commented Feb 6, 2025

grosjo commented Feb 8, 2025

testuser7 commented Feb 8, 2025

grosjo commented Feb 8, 2025 • edited Loading

grosjo commented Feb 8, 2025

grosjo commented Feb 8, 2025 • edited Loading

grosjo commented Feb 8, 2025

grosjo commented Feb 10, 2025

grosjo commented Feb 14, 2025

testuser7 commented Feb 14, 2025 • edited Loading

grosjo commented Feb 14, 2025

testuser7 commented Feb 15, 2025

grosjo commented Feb 15, 2025

testuser7 commented Feb 15, 2025

grosjo commented Feb 15, 2025

grosjo commented Feb 15, 2025

testuser7 commented Jan 20, 2025 •

edited

Loading

grosjo commented Feb 8, 2025 •

edited

Loading

grosjo commented Feb 8, 2025 •

edited

Loading

testuser7 commented Feb 14, 2025 •

edited

Loading