[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <93002681-37f8-cb12-725c-48c1695d29a0@socionext.com>
Date: Wed, 8 May 2019 21:22:05 +0900
From: "Sugaya, Taichi" <sugaya.taichi@...ionext.com>
To: Alan Cox <gnomes@...rguk.ukuu.org.uk>
Cc: Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
Jiri Slaby <jslaby@...e.com>, Arnd Bergmann <arnd@...db.de>,
Takao Orito <orito.takao@...ionext.com>,
Kazuhiro Kasai <kasai.kazuhiro@...ionext.com>,
Shinji Kanematsu <kanematsu.shinji@...ionext.com>,
Jassi Brar <jaswinder.singh@...aro.org>,
Masami Hiramatsu <masami.hiramatsu@...aro.org>,
linux-kernel@...r.kernel.org, linux-serial@...r.kernel.org
Subject: Re: [PATCH v3] serial: Add Milbeaut serial control
Hi,
Thank you for pointing out.
On 2019/04/27 3:15, Alan Cox wrote:
> O
>> +static void mlb_usio_set_termios(struct uart_port *port,
>> + struct ktermios *termios, struct ktermios *old)
>> +{
>> + unsigned int escr, smr = MLB_USIO_SMR_SOE;
>> + unsigned long flags, baud, quot;
>> +
>> + switch (termios->c_cflag & CSIZE) {
>> + case CS5:
>> + escr = MLB_USIO_ESCR_L_5BIT;
>> + break;
>> + case CS6:
>> + escr = MLB_USIO_ESCR_L_6BIT;
>> + break;
>> + case CS7:
>> + escr = MLB_USIO_ESCR_L_7BIT;
>> + break;
>> + case CS8:
>> + default:
>> + escr = MLB_USIO_ESCR_L_8BIT;
>> + break;
>> + }
>> +
>> + if (termios->c_cflag & CSTOPB)
>> + smr |= MLB_USIO_SMR_SBL;
>> +
>> + if (termios->c_cflag & PARENB) {
>> + escr |= MLB_USIO_ESCR_PEN;
>> + if (termios->c_cflag & PARODD)
>> + escr |= MLB_USIO_ESCR_P;
>> + }
>
> If you don't suport CMSPAR then clear that bit in termios as well
>
OK, clear the bit because of not supported.
>> + /* Set hard flow control */
>> + if (of_property_read_bool(port->dev->of_node, "auto-flow-control") ||
>> + (termios->c_cflag & CRTSCTS))
>> + escr |= MLB_USIO_ESCR_FLWEN;
>
> That's just broken. The termios bits are the definitive things for the
> port, and in addition even if they are forced you need to correct the
> termios data.
>
> You might want to control flow control *at boot* with an OF property but
> doing it post boot is just busted.
>
Ah, Yes.
I think OF property should not be here, and it may only be used to determine
the characteristics of the port.
I try to make a fixes patch.
Thanks,
Sugaya Taichi
> Alan
>
Powered by blists - more mailing lists