2023-02-21 07:13:44

by Christian Hewitt

[permalink] [raw]
Subject: Bug/Problem with rtl_bt/rtl8822cs_config.bin

I’m working with the RTL8822CS WiFi/BT mezzanine for BananaPi M2S/M5 boards.

SDIO WiFi works fine. The BT side fails to load firmware:

[ 7.413297] Bluetooth: hci0: RTL: examining hci_ver=0a hci_rev=000c lmp_ver=0a lmp_subver=8822
[ 7.416873] Bluetooth: hci0: RTL: rom_version status=0 version=3
[ 7.416892] Bluetooth: hci0: RTL: loading rtl_bt/rtl8822cs_fw.bin
[ 7.418200] Bluetooth: hci0: RTL: loading rtl_bt/rtl8822cs_config.bin
[ 7.453778] Bluetooth: hci0: RTL: cfg_sz 33, total sz 36529
[ 9.469947] Bluetooth: hci0: command 0xfc20 tx timeout
[ 17.629661] Bluetooth: hci0: RTL: download fw command failed (-110)

The same is also observed on an Amlogic Android box with 08 silicon:

[ 3.950116] Bluetooth: hci0: RTL: examining hci_ver=08 hci_rev=000c lmp_ver=08 lmp_subver=8822
[ 3.963304] Bluetooth: hci0: RTL: rom_version status=0 version=3
[ 3.969522] Bluetooth: hci0: RTL: loading rtl_bt/rtl8822cs_fw.bin
[ 3.969556] Bluetooth: hci0: RTL: loading rtl_bt/rtl8822cs_config.bin
[ 4.026336] Bluetooth: hci0: RTL: cfg_sz 33, total sz 36529
[ 6.086301] Bluetooth: hci0: command 0xfc20 tx timeout
[ 14.566290] Bluetooth: hci0: RTL: download fw command failed (-110)

The device is working in Armbian who bundle an older firmware:

[ 7.319820] Bluetooth: hci0: RTL: examining hci_ver=0a hci_rev=000c lmp_ver=0a lmp_subver=8822
[ 7.323114] Bluetooth: hci0: RTL: rom_version status=0 version=3
[ 7.323150] Bluetooth: hci0: RTL: loading rtl_bt/rtl8822cs_fw.bin
[ 7.325103] Bluetooth: hci0: RTL: loading rtl_bt/rtl8822cs_config.bin
[ 7.348972] Bluetooth: hci0: RTL: cfg_sz 73, total sz 40777
[ 7.699139] Bluetooth: hci0: RTL: fw version 0x05a91a4a

However it’s not a regression in firmware. Further testing pinpointed the
rtl8822cs_config.bin file. I can use the latest firmware (FFB8_ABD6) with
the config file bundled with Armbian:

[ 7.504405] Bluetooth: hci0: RTL: examining hci_ver=0a hci_rev=000c lmp_ver=0a lmp_subver=8822
[ 7.507827] Bluetooth: hci0: RTL: rom_version status=0 version=3
[ 7.507850] Bluetooth: hci0: RTL: loading rtl_bt/rtl8822cs_fw.bin
[ 7.509158] Bluetooth: hci0: RTL: loading rtl_bt/rtl8822cs_config.bin
[ 7.529486] Bluetooth: hci0: RTL: cfg_sz 73, total sz 36569
[ 7.850914] Bluetooth: hci0: RTL: fw version 0xffb8abd6

I can also use the oldest upstream firmware with the Armbian config:

[ 7.315358] Bluetooth: hci0: RTL: examining hci_ver=0a hci_rev=000c lmp_ver=0a lmp_subver=8822
[ 7.318754] Bluetooth: hci0: RTL: rom_version status=0 version=3
[ 7.318772] Bluetooth: hci0: RTL: loading rtl_bt/rtl8822cs_fw.bin
[ 7.319680] Bluetooth: hci0: RTL: loading rtl_bt/rtl8822cs_config.bin
[ 7.341340] Bluetooth: hci0: RTL: cfg_sz 73, total sz 30817
[ 7.612533] Bluetooth: hci0: RTL: fw version 0x05990d54

The config file Armbian are using [0] is sourced from an Amlogic (Android) BSP
repo shared to manufacturer Jethome (Vyacheslav on CC). There is no obvious
version info. There are newer (more recently committed) config files available
in updates to the Amlogic repo (alongside older 0599_0D54 firmware) and all work
with any upstream firmware file.

I’m happy to apply test patches to get more output and I can provide remote
access to the board or arrange hardware to be shipped if that helps at all.

Hopefully RTL folks can spot whether this is a driver issue or something not
present in the upstream config?

Christian

[0] https://github.com/armbian/firmware/blob/master/rtl_bt/rtl8822cs_config.bin



2023-02-21 21:23:11

by Martin Blumenstingl

[permalink] [raw]
Subject: Re: Bug/Problem with rtl_bt/rtl8822cs_config.bin

Hi Christian,

+ Max Chou as the author of commit 848fc6164158 ("Bluetooth: hci_h5:
btrtl: Add support for RTL8822C")

On Tue, Feb 21, 2023 at 8:13 AM Christian Hewitt
<[email protected]> wrote:
[...]
> I can also use the oldest upstream firmware with the Armbian config:
>
> [ 7.315358] Bluetooth: hci0: RTL: examining hci_ver=0a hci_rev=000c lmp_ver=0a lmp_subver=8822
> [ 7.318754] Bluetooth: hci0: RTL: rom_version status=0 version=3
> [ 7.318772] Bluetooth: hci0: RTL: loading rtl_bt/rtl8822cs_fw.bin
> [ 7.319680] Bluetooth: hci0: RTL: loading rtl_bt/rtl8822cs_config.bin
> [ 7.341340] Bluetooth: hci0: RTL: cfg_sz 73, total sz 30817
> [ 7.612533] Bluetooth: hci0: RTL: fw version 0x05990d54
>
> The config file Armbian are using [0] is sourced from an Amlogic (Android) BSP
> repo shared to manufacturer Jethome (Vyacheslav on CC). There is no obvious
> version info. There are newer (more recently committed) config files available
> in updates to the Amlogic repo (alongside older 0599_0D54 firmware) and all work
> with any upstream firmware file.
I did compare the two config files and they differ in certain areas.
- upstream config: 1500000 baud, flow control is disabled
- Armbian config: 2000000 baud, flow control is enabled

Seeing that the Armbian config works with a higher baud rate my guess
is that there's some issue with flow control instead.

For those who are interested, here's the raw data provided by the
rtlfw tool from [0].
Armbian (= working) config:
Signature: 0x8723ab55
Data len: 67
len=16 offset=000c,{ 02 50 00 00 50 c5 ea 19 e1 1b fd af 5f 01 a4 0b
},UART_CONFIG
len=2 offset=01be,{ 1a 19 }
len=1 offset=01bb,{ 3d }
len=4 offset=01c0,{ 10 00 60 01 }
len=11 offset=1000,{ 00 ff 04 5d 00 03 00 ff ff ff ff }
len=11 offset=1040,{ 28 08 30 30 00 ff ff ff ff ff ff }
len=1 offset=01c4,{ 02 }

upstream (= non-working) config:
Signature: 0x8723ab55
Data len: 27
len=16 offset=000c,{ 02 80 92 04 50 c5 ea 19 e1 1b fd af 5b 01 a4 0b
},UART_CONFIG
len=1 offset=007a,{ 37 }
len=1 offset=00d9,{ 0f }

The baud rate is encoded in the first four bytes and can be decoded
with the help of btrtl_convert_baudrate() from Linux'
drivers/bluetooth/btrtl.c
Flow control is encoded as BIT(2) in data[12].

I am hoping that Hilda and/or Max can provide some ideas on how to
solve this issue.


Best regards,
Martin


[0] https://git.kernel.org/pub/scm/bluetooth/bluez.git/tree/tools/rtlfw.c?id=5d347b54714e0f2d750253be09b68b0c3119dd0a

2023-02-23 03:25:07

by Max Chou

[permalink] [raw]
Subject: RE: Bug/Problem with rtl_bt/rtl8822cs_config.bin

Hi! Christian,
I just try to test the config file as linux-firmware on one model of Chromebook.
It's able to initialize RTL8822CS Bluetooth controller.
The issue does not present as your mention on my side.

I compare the two config files that you listed.
The key difference might to impact the initialization are UART baud rate and UART flow control as Martin's mention.
Recommend you to edit the config file to check what is the root cause for this issue on your Host.
Here're the three cases for the config changed, please use hexadecimal editor for the three changes. Thanks.

1. Based on the config file of linux-firmware, to edit UART baudrate as below.
Original: 1500000 baud, flow control disabled
00000000 55 ab 23 87 1b 00 0c 00 10 02 80 92 04 50 c5 ea
00000010 19 e1 1b fd af 5b 01 a4 0b 7a 00 01 37 d9 00 01
00000020 0f

After: 2000000 baud, flow control disabled
00000000 55 ab 23 87 1b 00 0c 00 10 02 50 00 00 50 c5 ea
^^^^^^^^^
00000010 19 e1 1b fd af 5b 01 a4 0b 7a 00 01 37 d9 00 01
00000020 0f

2. Based on the config file of linux-firmware, to edit UART baudrate as below.
Original: 1500000 baud, flow control disabled
00000000 55 ab 23 87 1b 00 0c 00 10 02 80 92 04 50 c5 ea
00000010 19 e1 1b fd af 5b 01 a4 0b 7a 00 01 37 d9 00 01
00000020 0f

After: 1500000 baud, flow control enabled
00000000 55 ab 23 87 1b 00 0c 00 10 02 80 92 04 50 c5 ea
00000010 19 e1 1b fd af 5f 01 a4 0b 7a 00 01 37 d9 00 01
^^^
00000020 0f

3. Based on the config file of linux-firmware, to edit UART baudrate as below.
Original: 1500000 baud, flow control disabled
00000000 55 ab 23 87 1b 00 0c 00 10 02 80 92 04 50 c5 ea
00000010 19 e1 1b fd af 5b 01 a4 0b 7a 00 01 37 d9 00 01
00000020 0f

After: 2000000 baud, flow control disabled
00000000 55 ab 23 87 1b 00 0c 00 10 02 50 00 00 50 c5 ea
^^^^^^^^^
00000010 19 e1 1b fd af 5f 01 a4 0b 7a 00 01 37 d9 00 01
^^^
00000020 0f

BRs,
Max


> -----Original Message-----
> From: Martin Blumenstingl <[email protected]>
> Sent: Wednesday, February 22, 2023 5:23 AM
> To: Christian Hewitt <[email protected]>
> Cc: Hilda Wu <[email protected]>; Josh Boyer <[email protected]>;
> Linux Firmware <[email protected]>; Bluez mailing list
> <[email protected]>; Vyacheslav <[email protected]>; Max Chou
> <[email protected]>
> Subject: Re: Bug/Problem with rtl_bt/rtl8822cs_config.bin
>
>
> External mail.
>
>
>
> Hi Christian,
>
> + Max Chou as the author of commit 848fc6164158 ("Bluetooth: hci_h5:
> btrtl: Add support for RTL8822C")
>
> On Tue, Feb 21, 2023 at 8:13 AM Christian Hewitt
> <[email protected]> wrote:
> [...]
> > I can also use the oldest upstream firmware with the Armbian config:
> >
> > [ 7.315358] Bluetooth: hci0: RTL: examining hci_ver=0a hci_rev=000c
> lmp_ver=0a lmp_subver=8822
> > [ 7.318754] Bluetooth: hci0: RTL: rom_version status=0 version=3
> > [ 7.318772] Bluetooth: hci0: RTL: loading rtl_bt/rtl8822cs_fw.bin
> > [ 7.319680] Bluetooth: hci0: RTL: loading rtl_bt/rtl8822cs_config.bin
> > [ 7.341340] Bluetooth: hci0: RTL: cfg_sz 73, total sz 30817
> > [ 7.612533] Bluetooth: hci0: RTL: fw version 0x05990d54
> >
> > The config file Armbian are using [0] is sourced from an Amlogic
> > (Android) BSP repo shared to manufacturer Jethome (Vyacheslav on CC).
> > There is no obvious version info. There are newer (more recently
> > committed) config files available in updates to the Amlogic repo
> > (alongside older 0599_0D54 firmware) and all work with any upstream
> firmware file.
> I did compare the two config files and they differ in certain areas.
> - upstream config: 1500000 baud, flow control is disabled
> - Armbian config: 2000000 baud, flow control is enabled
>
> Seeing that the Armbian config works with a higher baud rate my guess is that
> there's some issue with flow control instead.
>
> For those who are interested, here's the raw data provided by the rtlfw tool
> from [0].
> Armbian (= working) config:
> Signature: 0x8723ab55
> Data len: 67
> len=16 offset=000c,{ 02 50 00 00 50 c5 ea 19 e1 1b fd af 5f 01 a4
> 0b },UART_CONFIG
> len=2 offset=01be,{ 1a 19 }
> len=1 offset=01bb,{ 3d }
> len=4 offset=01c0,{ 10 00 60 01 }
> len=11 offset=1000,{ 00 ff 04 5d 00 03 00 ff ff ff ff }
> len=11 offset=1040,{ 28 08 30 30 00 ff ff ff ff ff ff }
> len=1 offset=01c4,{ 02 }
>
> upstream (= non-working) config:
> Signature: 0x8723ab55
> Data len: 27
> len=16 offset=000c,{ 02 80 92 04 50 c5 ea 19 e1 1b fd af 5b 01 a4
> 0b },UART_CONFIG
> len=1 offset=007a,{ 37 }
> len=1 offset=00d9,{ 0f }
>
> The baud rate is encoded in the first four bytes and can be decoded with the
> help of btrtl_convert_baudrate() from Linux'
> drivers/bluetooth/btrtl.c
> Flow control is encoded as BIT(2) in data[12].
>
> I am hoping that Hilda and/or Max can provide some ideas on how to solve
> this issue.
>
>
> Best regards,
> Martin
>
>
> [0]
> https://git.kernel.org/pub/scm/bluetooth/bluez.git/tree/tools/rtlfw.c?id=5d347
> b54714e0f2d750253be09b68b0c3119dd0a
>
> ------Please consider the environment before printing this e-mail.

2023-02-23 07:51:34

by Christian Hewitt

[permalink] [raw]
Subject: Re: Bug/Problem with rtl_bt/rtl8822cs_config.bin

Hello Max,

The bug report stimulated Martin’s maintainer memory over unfinished improvements
to the Amlogic UART used in the g12a and newer SoC families (we are testing with
g12b and sm1 devices) resulting in this series being submitted last night:

https://patchwork.kernel.org/project/linux-amlogic/list/?series=724172

With those patches applied the boards support stable 1500000 baud rate and BT is
working with the latest fw/config files from linux-firmware:

LibreELEC:/ # dmesg | grep Blue
[ 6.839950] Bluetooth: Core ver 2.22
[ 6.840128] Bluetooth: HCI device and connection manager initialized
[ 6.840419] Bluetooth: HCI socket layer initialized
[ 6.840437] Bluetooth: L2CAP socket layer initialized
[ 6.840740] Bluetooth: SCO socket layer initialized
[ 6.874518] Bluetooth: HCI UART driver ver 2.3
[ 6.874533] Bluetooth: HCI UART protocol H4 registered
[ 6.874619] Bluetooth: HCI UART protocol Three-wire (H5) registered
[ 6.874809] Bluetooth: HCI UART protocol Broadcom registered
[ 6.874842] Bluetooth: HCI UART protocol QCA registered
[ 7.636373] Bluetooth: hci0: RTL: examining hci_ver=0a hci_rev=000c lmp_ver=0a lmp_subver=8822
[ 7.639863] Bluetooth: hci0: RTL: rom_version status=0 version=3
[ 7.639959] Bluetooth: hci0: RTL: loading rtl_bt/rtl8822cs_fw.bin
[ 7.655585] Bluetooth: hci0: RTL: loading rtl_bt/rtl8822cs_config.bin
[ 7.697202] Bluetooth: hci0: RTL: cfg_sz 33, total sz 36529
[ 8.095507] Bluetooth: hci0: RTL: fw version 0xffb8abd6
[ 8.149476] Bluetooth: MGMT ver 1.22

[bluetooth]# show
Controller 68:4E:05:F2:1E:42 (public)
Name: LibreELEC
Alias: LibreELEC
Class: 0x000c0000
Powered: yes
Discoverable: no
DiscoverableTimeout: 0x000000b4
Pairable: yes
UUID: A/V Remote Control (0000110e-0000-1000-8000-00805f9b34fb)
UUID: Audio Source (0000110a-0000-1000-8000-00805f9b34fb)
UUID: PnP Information (00001200-0000-1000-8000-00805f9b34fb)
UUID: Audio Sink (0000110b-0000-1000-8000-00805f9b34fb)
UUID: A/V Remote Control Target (0000110c-0000-1000-8000-00805f9b34fb)
UUID: Generic Access Profile (00001800-0000-1000-8000-00805f9b34fb)
UUID: Generic Attribute Profile (00001801-0000-1000-8000-00805f9b34fb)
UUID: Device Information (0000180a-0000-1000-8000-00805f9b34fb)
Modalias: usb:v1D6Bp0246d0542
Discovering: no
Roles: central
Roles: peripheral
Advertising Features:
ActiveInstances: 0x00 (0)
SupportedInstances: 0x04 (4)
SupportedIncludes: tx-power
SupportedIncludes: appearance
SupportedIncludes: local-name
SupportedSecondaryChannels: 1M
SupportedSecondaryChannels: 2M
SupportedSecondaryChannels: Coded
[bluetooth]# scan on
Discovery started
[CHG] Controller 68:4E:05:F2:1E:42 Discovering: yes
[NEW] Device 4C:87:5D:89:E8:D6 4C-87-5D-89-E8-D6
[DEL] Device 4C:87:5D:89:E8:D6 4C-87-5D-89-E8-D6
[NEW] Device 5B:34:59:3E:EE:27 5B-34-59-3E-EE-27
[NEW] Device 67:0D:92:0F:99:DB 67-0D-92-0F-99-DB
[DEL] Device 5B:34:59:3E:EE:27 5B-34-59-3E-EE-27
[DEL] Device 67:0D:92:0F:99:DB 67-0D-92-0F-99-DB
[NEW] Device 4F:3B:07:B7:13:9D 4F-3B-07-B7-13-9D
[NEW] Device 43:87:04:F2:86:69 43-87-04-F2-86-69
[CHG] Device 4F:3B:07:B7:13:9D RSSI: -89
[CHG] Device 4F:3B:07:B7:13:9D RSSI: -101
[DEL] Device 4F:3B:07:B7:13:9D 4F-3B-07-B7-13-9D
[DEL] Device 43:87:04:F2:86:69 43-87-04-F2-86-69
[NEW] Device 49:17:FA:07:A1:F9 49-17-FA-07-A1-F9
[NEW] Device 7F:7B:B2:FF:43:27 7F-7B-B2-FF-43-27
[NEW] Device 69:0E:9F:24:06:96 69-0E-9F-24-06-96
[NEW] Device 51:87:CE:DA:B4:41 51-87-CE-DA-B4-41
[DEL] Device 49:17:FA:07:A1:F9 49-17-FA-07-A1-F9
[DEL] Device 7F:7B:B2:FF:43:27 7F-7B-B2-FF-43-27
[DEL] Device 51:87:CE:DA:B4:41 51-87-CE-DA-B4-41
[DEL] Device 69:0E:9F:24:06:96 69-0E-9F-24-06-96

So all appears to be working now.

Thanks to all..

Christian


> On 23 Feb 2023, at 7:24 am, Max Chou <[email protected]> wrote:
>
> Hi! Christian,
> I just try to test the config file as linux-firmware on one model of Chromebook.
> It's able to initialize RTL8822CS Bluetooth controller.
> The issue does not present as your mention on my side.
>
> I compare the two config files that you listed.
> The key difference might to impact the initialization are UART baud rate and UART flow control as Martin's mention.
> Recommend you to edit the config file to check what is the root cause for this issue on your Host.
> Here're the three cases for the config changed, please use hexadecimal editor for the three changes. Thanks.
>
> 1. Based on the config file of linux-firmware, to edit UART baudrate as below.
> Original: 1500000 baud, flow control disabled
> 00000000 55 ab 23 87 1b 00 0c 00 10 02 80 92 04 50 c5 ea
> 00000010 19 e1 1b fd af 5b 01 a4 0b 7a 00 01 37 d9 00 01
> 00000020 0f
>
> After: 2000000 baud, flow control disabled
> 00000000 55 ab 23 87 1b 00 0c 00 10 02 50 00 00 50 c5 ea
> ^^^^^^^^^
> 00000010 19 e1 1b fd af 5b 01 a4 0b 7a 00 01 37 d9 00 01
> 00000020 0f
>
> 2. Based on the config file of linux-firmware, to edit UART baudrate as below.
> Original: 1500000 baud, flow control disabled
> 00000000 55 ab 23 87 1b 00 0c 00 10 02 80 92 04 50 c5 ea
> 00000010 19 e1 1b fd af 5b 01 a4 0b 7a 00 01 37 d9 00 01
> 00000020 0f
>
> After: 1500000 baud, flow control enabled
> 00000000 55 ab 23 87 1b 00 0c 00 10 02 80 92 04 50 c5 ea
> 00000010 19 e1 1b fd af 5f 01 a4 0b 7a 00 01 37 d9 00 01
> ^^^
> 00000020 0f
>
> 3. Based on the config file of linux-firmware, to edit UART baudrate as below.
> Original: 1500000 baud, flow control disabled
> 00000000 55 ab 23 87 1b 00 0c 00 10 02 80 92 04 50 c5 ea
> 00000010 19 e1 1b fd af 5b 01 a4 0b 7a 00 01 37 d9 00 01
> 00000020 0f
>
> After: 2000000 baud, flow control disabled
> 00000000 55 ab 23 87 1b 00 0c 00 10 02 50 00 00 50 c5 ea
> ^^^^^^^^^
> 00000010 19 e1 1b fd af 5f 01 a4 0b 7a 00 01 37 d9 00 01
> ^^^
> 00000020 0f
>
> BRs,
> Max
>
>
>> -----Original Message-----
>> From: Martin Blumenstingl <[email protected]>
>> Sent: Wednesday, February 22, 2023 5:23 AM
>> To: Christian Hewitt <[email protected]>
>> Cc: Hilda Wu <[email protected]>; Josh Boyer <[email protected]>;
>> Linux Firmware <[email protected]>; Bluez mailing list
>> <[email protected]>; Vyacheslav <[email protected]>; Max Chou
>> <[email protected]>
>> Subject: Re: Bug/Problem with rtl_bt/rtl8822cs_config.bin
>>
>>
>> External mail.
>>
>>
>>
>> Hi Christian,
>>
>> + Max Chou as the author of commit 848fc6164158 ("Bluetooth: hci_h5:
>> btrtl: Add support for RTL8822C")
>>
>> On Tue, Feb 21, 2023 at 8:13 AM Christian Hewitt
>> <[email protected]> wrote:
>> [...]
>>> I can also use the oldest upstream firmware with the Armbian config:
>>>
>>> [ 7.315358] Bluetooth: hci0: RTL: examining hci_ver=0a hci_rev=000c
>> lmp_ver=0a lmp_subver=8822
>>> [ 7.318754] Bluetooth: hci0: RTL: rom_version status=0 version=3
>>> [ 7.318772] Bluetooth: hci0: RTL: loading rtl_bt/rtl8822cs_fw.bin
>>> [ 7.319680] Bluetooth: hci0: RTL: loading rtl_bt/rtl8822cs_config.bin
>>> [ 7.341340] Bluetooth: hci0: RTL: cfg_sz 73, total sz 30817
>>> [ 7.612533] Bluetooth: hci0: RTL: fw version 0x05990d54
>>>
>>> The config file Armbian are using [0] is sourced from an Amlogic
>>> (Android) BSP repo shared to manufacturer Jethome (Vyacheslav on CC).
>>> There is no obvious version info. There are newer (more recently
>>> committed) config files available in updates to the Amlogic repo
>>> (alongside older 0599_0D54 firmware) and all work with any upstream
>> firmware file.
>> I did compare the two config files and they differ in certain areas.
>> - upstream config: 1500000 baud, flow control is disabled
>> - Armbian config: 2000000 baud, flow control is enabled
>>
>> Seeing that the Armbian config works with a higher baud rate my guess is that
>> there's some issue with flow control instead.
>>
>> For those who are interested, here's the raw data provided by the rtlfw tool
>> from [0].
>> Armbian (= working) config:
>> Signature: 0x8723ab55
>> Data len: 67
>> len=16 offset=000c,{ 02 50 00 00 50 c5 ea 19 e1 1b fd af 5f 01 a4
>> 0b },UART_CONFIG
>> len=2 offset=01be,{ 1a 19 }
>> len=1 offset=01bb,{ 3d }
>> len=4 offset=01c0,{ 10 00 60 01 }
>> len=11 offset=1000,{ 00 ff 04 5d 00 03 00 ff ff ff ff }
>> len=11 offset=1040,{ 28 08 30 30 00 ff ff ff ff ff ff }
>> len=1 offset=01c4,{ 02 }
>>
>> upstream (= non-working) config:
>> Signature: 0x8723ab55
>> Data len: 27
>> len=16 offset=000c,{ 02 80 92 04 50 c5 ea 19 e1 1b fd af 5b 01 a4
>> 0b },UART_CONFIG
>> len=1 offset=007a,{ 37 }
>> len=1 offset=00d9,{ 0f }
>>
>> The baud rate is encoded in the first four bytes and can be decoded with the
>> help of btrtl_convert_baudrate() from Linux'
>> drivers/bluetooth/btrtl.c
>> Flow control is encoded as BIT(2) in data[12].
>>
>> I am hoping that Hilda and/or Max can provide some ideas on how to solve
>> this issue.
>>
>>
>> Best regards,
>> Martin
>>
>>
>> [0]
>> https://git.kernel.org/pub/scm/bluetooth/bluez.git/tree/tools/rtlfw.c?id=5d347
>> b54714e0f2d750253be09b68b0c3119dd0a
>>
>> ------Please consider the environment before printing this e-mail.


2023-02-23 14:14:37

by Viacheslav

[permalink] [raw]
Subject: Re: Bug/Problem with rtl_bt/rtl8822cs_config.bin



On 23/02/2023 10.51, Christian Hewitt wrote:
> Hello Max,
>
> The bug report stimulated Martin’s maintainer memory over unfinished improvements
> to the Amlogic UART used in the g12a and newer SoC families (we are testing with
> g12b and sm1 devices) resulting in this series being submitted last night:
>
> https://patchwork.kernel.org/project/linux-amlogic/list/?series=724172

However, this does not fix the problem with rtl8822cs BT on gxl/axg chipsets

>
> With those patches applied the boards support stable 1500000 baud rate and BT is
> working with the latest fw/config files from linux-firmware:
>
> LibreELEC:/ # dmesg | grep Blue
> [ 6.839950] Bluetooth: Core ver 2.22
> [ 6.840128] Bluetooth: HCI device and connection manager initialized
> [ 6.840419] Bluetooth: HCI socket layer initialized
> [ 6.840437] Bluetooth: L2CAP socket layer initialized
> [ 6.840740] Bluetooth: SCO socket layer initialized
> [ 6.874518] Bluetooth: HCI UART driver ver 2.3
> [ 6.874533] Bluetooth: HCI UART protocol H4 registered
> [ 6.874619] Bluetooth: HCI UART protocol Three-wire (H5) registered
> [ 6.874809] Bluetooth: HCI UART protocol Broadcom registered
> [ 6.874842] Bluetooth: HCI UART protocol QCA registered
> [ 7.636373] Bluetooth: hci0: RTL: examining hci_ver=0a hci_rev=000c lmp_ver=0a lmp_subver=8822
> [ 7.639863] Bluetooth: hci0: RTL: rom_version status=0 version=3
> [ 7.639959] Bluetooth: hci0: RTL: loading rtl_bt/rtl8822cs_fw.bin
> [ 7.655585] Bluetooth: hci0: RTL: loading rtl_bt/rtl8822cs_config.bin
> [ 7.697202] Bluetooth: hci0: RTL: cfg_sz 33, total sz 36529
> [ 8.095507] Bluetooth: hci0: RTL: fw version 0xffb8abd6
> [ 8.149476] Bluetooth: MGMT ver 1.22
>
> [bluetooth]# show
> Controller 68:4E:05:F2:1E:42 (public)
> Name: LibreELEC
> Alias: LibreELEC
> Class: 0x000c0000
> Powered: yes
> Discoverable: no
> DiscoverableTimeout: 0x000000b4
> Pairable: yes
> UUID: A/V Remote Control (0000110e-0000-1000-8000-00805f9b34fb)
> UUID: Audio Source (0000110a-0000-1000-8000-00805f9b34fb)
> UUID: PnP Information (00001200-0000-1000-8000-00805f9b34fb)
> UUID: Audio Sink (0000110b-0000-1000-8000-00805f9b34fb)
> UUID: A/V Remote Control Target (0000110c-0000-1000-8000-00805f9b34fb)
> UUID: Generic Access Profile (00001800-0000-1000-8000-00805f9b34fb)
> UUID: Generic Attribute Profile (00001801-0000-1000-8000-00805f9b34fb)
> UUID: Device Information (0000180a-0000-1000-8000-00805f9b34fb)
> Modalias: usb:v1D6Bp0246d0542
> Discovering: no
> Roles: central
> Roles: peripheral
> Advertising Features:
> ActiveInstances: 0x00 (0)
> SupportedInstances: 0x04 (4)
> SupportedIncludes: tx-power
> SupportedIncludes: appearance
> SupportedIncludes: local-name
> SupportedSecondaryChannels: 1M
> SupportedSecondaryChannels: 2M
> SupportedSecondaryChannels: Coded
> [bluetooth]# scan on
> Discovery started
> [CHG] Controller 68:4E:05:F2:1E:42 Discovering: yes
> [NEW] Device 4C:87:5D:89:E8:D6 4C-87-5D-89-E8-D6
> [DEL] Device 4C:87:5D:89:E8:D6 4C-87-5D-89-E8-D6
> [NEW] Device 5B:34:59:3E:EE:27 5B-34-59-3E-EE-27
> [NEW] Device 67:0D:92:0F:99:DB 67-0D-92-0F-99-DB
> [DEL] Device 5B:34:59:3E:EE:27 5B-34-59-3E-EE-27
> [DEL] Device 67:0D:92:0F:99:DB 67-0D-92-0F-99-DB
> [NEW] Device 4F:3B:07:B7:13:9D 4F-3B-07-B7-13-9D
> [NEW] Device 43:87:04:F2:86:69 43-87-04-F2-86-69
> [CHG] Device 4F:3B:07:B7:13:9D RSSI: -89
> [CHG] Device 4F:3B:07:B7:13:9D RSSI: -101
> [DEL] Device 4F:3B:07:B7:13:9D 4F-3B-07-B7-13-9D
> [DEL] Device 43:87:04:F2:86:69 43-87-04-F2-86-69
> [NEW] Device 49:17:FA:07:A1:F9 49-17-FA-07-A1-F9
> [NEW] Device 7F:7B:B2:FF:43:27 7F-7B-B2-FF-43-27
> [NEW] Device 69:0E:9F:24:06:96 69-0E-9F-24-06-96
> [NEW] Device 51:87:CE:DA:B4:41 51-87-CE-DA-B4-41
> [DEL] Device 49:17:FA:07:A1:F9 49-17-FA-07-A1-F9
> [DEL] Device 7F:7B:B2:FF:43:27 7F-7B-B2-FF-43-27
> [DEL] Device 51:87:CE:DA:B4:41 51-87-CE-DA-B4-41
> [DEL] Device 69:0E:9F:24:06:96 69-0E-9F-24-06-96
>
> So all appears to be working now.
>
> Thanks to all..
>
> Christian
>
>
>> On 23 Feb 2023, at 7:24 am, Max Chou <[email protected]> wrote:
>>
>> Hi! Christian,
>> I just try to test the config file as linux-firmware on one model of Chromebook.
>> It's able to initialize RTL8822CS Bluetooth controller.
>> The issue does not present as your mention on my side.
>>
>> I compare the two config files that you listed.
>> The key difference might to impact the initialization are UART baud rate and UART flow control as Martin's mention.
>> Recommend you to edit the config file to check what is the root cause for this issue on your Host.
>> Here're the three cases for the config changed, please use hexadecimal editor for the three changes. Thanks.
>>
>> 1. Based on the config file of linux-firmware, to edit UART baudrate as below.
>> Original: 1500000 baud, flow control disabled
>> 00000000 55 ab 23 87 1b 00 0c 00 10 02 80 92 04 50 c5 ea
>> 00000010 19 e1 1b fd af 5b 01 a4 0b 7a 00 01 37 d9 00 01
>> 00000020 0f
>>
>> After: 2000000 baud, flow control disabled
>> 00000000 55 ab 23 87 1b 00 0c 00 10 02 50 00 00 50 c5 ea
>> ^^^^^^^^^
>> 00000010 19 e1 1b fd af 5b 01 a4 0b 7a 00 01 37 d9 00 01
>> 00000020 0f
>>
>> 2. Based on the config file of linux-firmware, to edit UART baudrate as below.
>> Original: 1500000 baud, flow control disabled
>> 00000000 55 ab 23 87 1b 00 0c 00 10 02 80 92 04 50 c5 ea
>> 00000010 19 e1 1b fd af 5b 01 a4 0b 7a 00 01 37 d9 00 01
>> 00000020 0f
>>
>> After: 1500000 baud, flow control enabled
>> 00000000 55 ab 23 87 1b 00 0c 00 10 02 80 92 04 50 c5 ea
>> 00000010 19 e1 1b fd af 5f 01 a4 0b 7a 00 01 37 d9 00 01
>> ^^^
>> 00000020 0f
>>
>> 3. Based on the config file of linux-firmware, to edit UART baudrate as below.
>> Original: 1500000 baud, flow control disabled
>> 00000000 55 ab 23 87 1b 00 0c 00 10 02 80 92 04 50 c5 ea
>> 00000010 19 e1 1b fd af 5b 01 a4 0b 7a 00 01 37 d9 00 01
>> 00000020 0f
>>
>> After: 2000000 baud, flow control disabled
>> 00000000 55 ab 23 87 1b 00 0c 00 10 02 50 00 00 50 c5 ea
>> ^^^^^^^^^
>> 00000010 19 e1 1b fd af 5f 01 a4 0b 7a 00 01 37 d9 00 01
>> ^^^
>> 00000020 0f
>>
>> BRs,
>> Max
>>
>>
>>> -----Original Message-----
>>> From: Martin Blumenstingl <[email protected]>
>>> Sent: Wednesday, February 22, 2023 5:23 AM
>>> To: Christian Hewitt <[email protected]>
>>> Cc: Hilda Wu <[email protected]>; Josh Boyer <[email protected]>;
>>> Linux Firmware <[email protected]>; Bluez mailing list
>>> <[email protected]>; Vyacheslav <[email protected]>; Max Chou
>>> <[email protected]>
>>> Subject: Re: Bug/Problem with rtl_bt/rtl8822cs_config.bin
>>>
>>>
>>> External mail.
>>>
>>>
>>>
>>> Hi Christian,
>>>
>>> + Max Chou as the author of commit 848fc6164158 ("Bluetooth: hci_h5:
>>> btrtl: Add support for RTL8822C")
>>>
>>> On Tue, Feb 21, 2023 at 8:13 AM Christian Hewitt
>>> <[email protected]> wrote:
>>> [...]
>>>> I can also use the oldest upstream firmware with the Armbian config:
>>>>
>>>> [ 7.315358] Bluetooth: hci0: RTL: examining hci_ver=0a hci_rev=000c
>>> lmp_ver=0a lmp_subver=8822
>>>> [ 7.318754] Bluetooth: hci0: RTL: rom_version status=0 version=3
>>>> [ 7.318772] Bluetooth: hci0: RTL: loading rtl_bt/rtl8822cs_fw.bin
>>>> [ 7.319680] Bluetooth: hci0: RTL: loading rtl_bt/rtl8822cs_config.bin
>>>> [ 7.341340] Bluetooth: hci0: RTL: cfg_sz 73, total sz 30817
>>>> [ 7.612533] Bluetooth: hci0: RTL: fw version 0x05990d54
>>>>
>>>> The config file Armbian are using [0] is sourced from an Amlogic
>>>> (Android) BSP repo shared to manufacturer Jethome (Vyacheslav on CC).
>>>> There is no obvious version info. There are newer (more recently
>>>> committed) config files available in updates to the Amlogic repo
>>>> (alongside older 0599_0D54 firmware) and all work with any upstream
>>> firmware file.
>>> I did compare the two config files and they differ in certain areas.
>>> - upstream config: 1500000 baud, flow control is disabled
>>> - Armbian config: 2000000 baud, flow control is enabled
>>>
>>> Seeing that the Armbian config works with a higher baud rate my guess is that
>>> there's some issue with flow control instead.
>>>
>>> For those who are interested, here's the raw data provided by the rtlfw tool
>>> from [0].
>>> Armbian (= working) config:
>>> Signature: 0x8723ab55
>>> Data len: 67
>>> len=16 offset=000c,{ 02 50 00 00 50 c5 ea 19 e1 1b fd af 5f 01 a4
>>> 0b },UART_CONFIG
>>> len=2 offset=01be,{ 1a 19 }
>>> len=1 offset=01bb,{ 3d }
>>> len=4 offset=01c0,{ 10 00 60 01 }
>>> len=11 offset=1000,{ 00 ff 04 5d 00 03 00 ff ff ff ff }
>>> len=11 offset=1040,{ 28 08 30 30 00 ff ff ff ff ff ff }
>>> len=1 offset=01c4,{ 02 }
>>>
>>> upstream (= non-working) config:
>>> Signature: 0x8723ab55
>>> Data len: 27
>>> len=16 offset=000c,{ 02 80 92 04 50 c5 ea 19 e1 1b fd af 5b 01 a4
>>> 0b },UART_CONFIG
>>> len=1 offset=007a,{ 37 }
>>> len=1 offset=00d9,{ 0f }
>>>
>>> The baud rate is encoded in the first four bytes and can be decoded with the
>>> help of btrtl_convert_baudrate() from Linux'
>>> drivers/bluetooth/btrtl.c
>>> Flow control is encoded as BIT(2) in data[12].
>>>
>>> I am hoping that Hilda and/or Max can provide some ideas on how to solve
>>> this issue.
>>>
>>>
>>> Best regards,
>>> Martin
>>>
>>>
>>> [0]
>>> https://git.kernel.org/pub/scm/bluetooth/bluez.git/tree/tools/rtlfw.c?id=5d347
>>> b54714e0f2d750253be09b68b0c3119dd0a
>>>
>>> ------Please consider the environment before printing this e-mail.
>

2023-02-23 21:00:39

by Martin Blumenstingl

[permalink] [raw]
Subject: Re: Bug/Problem with rtl_bt/rtl8822cs_config.bin

Hello Vyacheslav,

On Thu, Feb 23, 2023 at 3:14 PM Vyacheslav <[email protected]> wrote:
[...]
> > The bug report stimulated Martin’s maintainer memory over unfinished improvements
> > to the Amlogic UART used in the g12a and newer SoC families (we are testing with
> > g12b and sm1 devices) resulting in this series being submitted last night:
> >
> > https://patchwork.kernel.org/project/linux-amlogic/list/?series=724172
>
> However, this does not fix the problem with rtl8822cs BT on gxl/axg chipsets
My understanding is that Meson GXBB and newer can also use XTAL along
with the UART controller's internal divider - without any additional
pre-divider like "XTAL divided by 2" or "XTAL divided by 3".
Using XTAL directly (24MHz) also makes it possible to achieve 1500000
baud without any jitter.

The problem with the Meson UART driver however is that using XTAL
directly is not implemented.
Yu Tu from Amlogic worked on patches last year. v6 of his series made
it into -next briefly, but got reverted due to issues. v7 then
stalled: [0] and only the bare minimum (enabling "XTAL divided by 2")
for the S4 SoC made it upstream in the end.

My suggestion is:
- let's get my small patchset upstreamed which at least fixes the
1500000 baud rate on G12A (and newer) SoCs
- then work on resurrecting Yu Tu's patchset and implement "XTAL
without any pre-dividers" support on top of that. Common Clock
Framework will then choose the correct parent depending on the SoC
generation, hopefully allowing arbitrary baud rates in the end - with
no jitter/as little jitter as possible.

What do you think?


Best regards,
Martin


[0] https://lore.kernel.org/all/[email protected]/