[Bug 207843] New: Some ACPI discovered BRCM HCI UART chip write clock failed.

[Bug 207843] New: Some ACPI discovered BRCM HCI UART chip write clock failed.

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=207843

            Bug ID: 207843
           Summary: Some ACPI discovered BRCM HCI UART chip write clock
                    failed.
           Product: Drivers
           Version: 2.5
    Kernel Version: 5.6.13
          Hardware: All
                OS: Linux
              Tree: Mainline
            Status: NEW
          Severity: low
          Priority: P1
         Component: Bluetooth
          Assignee: linux-bluetooth@vger.kernel.org
          Reporter: zechfox@gmail.com
        Regression: No

Created attachment 289213
  --> https://bugzilla.kernel.org/attachment.cgi?id=289213&action=edit
dmesg output

>Overview:
>>My tablet uses AP6255 wifi/bluetooth module with BCM43455 chip.
>>The bluetooth controller worked well except bluetooth headset play sound
>>glitch.It's not configuration problem in bluez or pulseaudio.
>>Some log in dmesg indicated that baud rate was not correctly configured. 
>>dmesg:
>>>Bluetooth: BCM: failed to write clock (-56)
>>>Bluetooth: Failed to set baudrate.

>Steps to Reproduce:
>>1. Power on bluetooth headset.
>>2. If it was not automatically connect, connect to headset by bluetoothctl.
>>3. play sound.
>Actual Results:
>>sound glitches.
>Expect Results:
>>sound well.
>Additional Information: 
>>I suspect the baud rate was too low to transmit audio.
>>It seems the chip can't support write clock HCI command. Then bluetooth
>>driver used default 115200 baud rate.
>>I tried to force configure 4Mbps baud rate to my chip after write clock
>>failed, but reset it failed.  Then I hardcode baud rate to 3Mbps, it solved
>>the glitch problem. 
>>And I found the commit that introduced the 4Mbsp to ACPI device:
>>https://github.com/torvalds/linux/commit/74183a1c50a3c61e62d2e585107ea3c0b942a3ff
>>It seems not all ACPI discovered devices can support write clock.
>>So maybe the driver should fallback baud rate to 3Mbps if write clock failed.

-- 
Smiles ^__________^ zech

-- 
You are receiving this mail because:
You are the assignee for the bug.