component/bt: allow setting a higher baudrate in UART HCI mode (IDFGH-2779)

[redchenjs]

The ESP32 datasheet specifies that the internal Bluetooth UART HCI baudrate can reach up to 4 Mbps, but the baudrate range is limited to 921600 bps in Kconfig.

I tried to use a FT232R usb serial converter to communicate with the Bluetooth controller at 3 Mbps baudrate (the maximum baudrate supported by the FT232R chip), and it worked as expected.

When using 921600 bps baudrate the SPP speed is limited to 90 KB/s, after changing the baudrate to 3 Mbps the SPP speed is increased to 260 KB/s, so just allow users to set a higher baudrate as the hardware does support it.

Example: controller_hci_uart

[Alvin1Zhang]

@redchenjs Thanks for the contribution.

[AbnerFederer]

Hi, @redchenjs

I have to do a bunch of tests to verify the 3M baud rate is proper for the controller to handle.
If all tests pass, I will merge this PR, so please wait for my update.

Thanks
Abner

[redchenjs]

An update:

I tried another serial converter FT232H (not FT232R), which supports the baud rate of 4 Mbps, the controller can handle it as well.

[AbnerFederer]

Hi @redchenjs

Have you done some throughput actions to test?
Any update for this PR is welcomed.

Thanks
Abner

[redchenjs]

Hi @redchenjs

Have you done some throughput actions to test?
Any update for this PR is welcomed.

Thanks
Abner

Hi @AbnerFederer

Here's my test result:

Baudrate  Speed(SPP)

921600 bps: 90 KB/s
1 Mbps: 97 KB/s
1.5 Mbps: 147 KB/s
2 Mbps: 194 KB/s
2.5 Mbps: 241 KB/s
3 Mbps: 247 Kb/s - 255 KB/s
3.5 Mbps: 247 Kb/s - 255 KB/s
4 Mbps: 247 Kb/s - 255 KB/s

It looks like there is almost no difference between 3M and 4M baud rates :)

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[redchenjs]

@AbnerFederer @Alvin1Zhang Is there any update?

[ghost]

Hi

I'm afraid this will cause some errors. So, could you please help add some comments in Help part?
And what about adding several levels of baud rate selections, like 115200, 921600, 200000, 300000, 400000?

Thanks

[redchenjs]

Hi

I'm afraid this will cause some errors. So, could you please help add some comments in Help part?
And what about adding several levels of baud rate selections, like 115200, 921600, 200000, 300000, 400000?

Thanks

Hi @Wth-Esp,

How does it look now? Now we can select the baud rate in menuconfig, and all these baud rates are from linux/drivers/tty/tty_baudrate.c, they should be the standard baud rate.

[ghost]

Hi @redchenjs

LGTM now.
I will do the PR handling procedure as soon as possible.

Thanks for your contribution.

[Alvin1Zhang]

Thanks for your contribution, and sorry for late reply, we have put the changes into our internal review queue, and the PR will be updated once the changes are available on GitHub. Thanks.

[maciejopalinski]

Hi @redchenjs
Have you done some throughput actions to test?
Any update for this PR is welcomed.
Thanks
Abner

Hi @AbnerFederer

Here's my test result:

Baudrate  Speed(SPP)

921600 bps: 90 KB/s
1 Mbps: 97 KB/s
1.5 Mbps: 147 KB/s
2 Mbps: 194 KB/s
2.5 Mbps: 241 KB/s
3 Mbps: 247 Kb/s - 255 KB/s
3.5 Mbps: 247 Kb/s - 255 KB/s
4 Mbps: 247 Kb/s - 255 KB/s

It looks like there is almost no difference between 3M and 4M baud rates :)

It is very interesting that there is no difference between 3M, 3.5M and 4M baud rates. I checked the communication with a logic analyzer and 3.5M/4M baud is giving the same exact signal as 3M. I calculated it being exactly 3M baud. Is this a linux bug or a limitation of my CH340 chip?