Re: [PATCH v3 0/6] USB: serial: mxuport: add MUX50U support and updates

[Johan Hovold <johan@kernel.org>]

On Tue, Aug 11, 2026 at 02:04:55PM +0800, Crescent Hsieh wrote:
> This series updates the mxuport driver to support MUX50U-based devices
> and fixes device-specific runtime behaviours.
> 
> The first three patches clean up the USB device ID table and firmware
> version handling, and add support for UPort G2 and Platform UART
> devices. The device family is encoded in the USB device ID table so
> that the table provides both the port count and the information needed
> for firmware handling.
> 
> The fourth patch adds optional callbacks to the generic USB-serial
> write path so that drivers can stop write submission based on
> device-specific state and restore that state if URB submission fails.
> 
> The fifth patch uses these callbacks to implement the per-port
> SEND_NEXT transmit flow-control protocol without duplicating the
> generic write implementation.
> 
> The sixth patch adds RS485 mode configuration through the standard
> TIOCSRS485/TIOCGRS485 interface.
> 
> ---
> Changes in v3:
>   - [PATCH v3 1/6]:
>     Rename the existing PID definitions after their corresponding
>     products and introduce a helper to keep each USB device ID table
>     entry on one line.
>   - [PATCH v3 2/6]:
>     Rename the firmware-version composition helper and retain the
>     existing firmware image-size validation.
>   - [PATCH v3 3/6]:
>     Name the new PID definitions after their corresponding products,
>     list all supported products by device family, rename the family
>     definitions to describe device rather than firmware families, and
>     use BIT() for the nonzero family values.
>   - [PATCH v3 4/6] and [PATCH v3 5/6]:
>     Replace the mxuport-specific copies of the generic write path with
>     optional write_ready() and write_rollback() callbacks. Use the
>     callbacks to pause SEND_NEXT writes and restore the per-port state
>     if URB submission fails. Expand the SEND_NEXT commit message to
>     describe the per-port protocol.

Again, sorry about the late feedback on this one.

Now applied for 7.4 with two whitespace amendments.

It was a joy to review, excellent work.

Johan