Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <20211029201402.428284-1-wander@redhat.com> <CAHp75VeZBp4gKvGBDzaD=EpGRDZ1-wTvD8K9Ui6Q59kDjmkXmQ@mail.gmail.com>
 <CAAq0SUmzHqEzNk3aw3SEgYVWRukQeHK1WtcJ3MjXcQKJrbC1Dw@mail.gmail.com>
In-Reply-To: <CAAq0SUmzHqEzNk3aw3SEgYVWRukQeHK1WtcJ3MjXcQKJrbC1Dw@mail.gmail.com>
From:   Andy Shevchenko <andy.shevchenko@gmail.com>
Date:   Mon, 1 Nov 2021 17:32:12 +0200
Message-ID: <CAHp75Vf6DjNcPWpE4Dh3SuzUMJbFQjq1UNCkrCa60uw35SpqKg@mail.gmail.com>
Subject: Re: [PATCH] tty: serial: Use fifo in 8250 console driver
To:     Wander Costa <wcosta@redhat.com>
Cc:     wander@redhat.com, Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        Jiri Slaby <jirislaby@kernel.org>,
        "Maciej W. Rozycki" <macro@orcam.me.uk>,
        Johan Hovold <johan@kernel.org>,
        Andrew Jeffery <andrew@aj.id.au>,
        "open list:SERIAL DRIVERS" <linux-serial@vger.kernel.org>,
        open list <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Mon, Nov 1, 2021 at 5:22 PM Wander Costa <wcosta@redhat.com> wrote:
> Em s=C3=A1b., 30 de out. de 2021 04:41, Andy Shevchenko
> <andy.shevchenko@gmail.com> escreveu:
> > On Friday, October 29, 2021, <wander@redhat.com> wrote:

...

> > I don't see any links.
>
> Oops, sorry about that. I must have accidentally deleted it while
> editing the commit message.
> Here it is https://github.com/walac/serial-console-test.
> I will update the patch with the link.

Thanks!

...

> > On how many different UARTs have you tested this? Have you tested oops =
and NMI contexts?
> >
> I only tested in a half dozen machines that I have available. I tried
> it in panic, warnings, IRQ contexts, etc. Theoretically, this change
> should not be affected by the context. Theoretically...
>
> > What I would like to say here is that the code is being used on zillion=
s of different 8250 implementations here and I would be rather skeptical ab=
out enabling the feature for everyone.
> >
> I did my homework and studied the 16550 datasheets, but yes, there is
> always this risk. Maybe people more experienced with PC serial ports
> than me might think the patch is not worth the risk of breaking some
> unknown number of devices out there, and I am ok with that. It is a
> valid point.

Here is a translation of my comment to a roadmap.

1. Introduce yet another UART quirk or capability (see corresponding
UART_CAP_* or UART_*_QUIRK definitions)
2. Add your patch conditionally based on the above
3. Enable it on UART(s) you _have tested_

--=20
With Best Regards,
Andy Shevchenko