Re: [PATCH] serial: Fix not set tty->port race condition

[Jiri Slaby <jirislaby@kernel.org>]

On 22. 01. 26, 18:11, Krzysztof Kozlowski wrote:
> On 22/01/2026 18:00, Krzysztof Kozlowski wrote:
>> Revert commit bfc467db60b7 ("serial: remove redundant
>> tty_port_link_device()") because the tty_port_link_device() is not
> 
> And grumpy side note because I was looking at this for more than a day
> blaming my new hardware:
> 
> I really wish commits (e.g. bfc467db60b7) calling something redundant
> had that much of message written why something is redundant as the
> commit (fb2b90014d78) which introduced that part of code.

It was clear enough: because tty_port_register_device_attr_serdev() 
links the port few lines below.

But it/I somehow didn't take the hidden uart_console() in 
uart_configure_port() into account.

> If someone wrote one page of text why foo is needed, we should write not
> less why it is not needed :)

I think I could generate a bloat of text. But you will still have a 
broken kernel the same way :)?

thanks,
-- 
js
suse labs