[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <ZMDccXK4N4ZEwsPn@surfacebook>
Date: Wed, 26 Jul 2023 11:42:25 +0300
From: andy.shevchenko@...il.com
To: Rodríguez Barbarin, José Javier
<JoseJavier.Rodriguez@...gon.com>
Cc: "gregkh@...uxfoundation.org" <gregkh@...uxfoundation.org>,
"jirislaby@...nel.org" <jirislaby@...nel.org>,
"morbidrsa@...il.com" <morbidrsa@...il.com>,
"linux-serial@...r.kernel.org" <linux-serial@...r.kernel.org>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"jth@...nel.org" <jth@...nel.org>,
Sanjuán García, Jorge
<Jorge.SanjuanGarcia@...gon.com>
Subject: Re: [PATCH 0/3] 8250_men_mcb: Make UART port autoconfigurable
Wed, Jul 05, 2023 at 01:14:51PM +0000, Rodríguez Barbarin, José Javier kirjoitti:
> Make configuration be handled by the 8250 UART subsystem
Actually this is not the best idea.
> to avoid weird behaviours
The opposite.
8250 detection is full of quirks and was developed to handle tons of different
UARTs when the driver was in a single file. Since you have a separate 8250_*.c
module for your UART and you _know_ the type beforehand, why on earth you need
to rely on the old and maybe not very suitable code? Have you thought about
corner cases with IRQ detection, for example?
> and for better maintainability.
The opposite.
I don't know if it affects your hardware to the date, but it may be different
for the future models, or models that you hadn't tested.
That said, I highly recommend to reconsider.
--
With Best Regards,
Andy Shevchenko
Powered by blists - more mailing lists