[PATCH linux dev-4.13] serial: 8250: Add Nuvoton NPCM UART

[Joel Stanley]

On Tue, Feb 6, 2018 at 10:54 PM, Tomer Maimon <tmaimon77 at gmail.com> wrote:
> On 6 February 2018 at 12:37, Avi Fishman <avifishman70 at gmail.com> wrote:
>> On Tue, Feb 6, 2018 at 7:50 AM, Joel Stanley <joel at jms.id.au> wrote:
>>> On Mon, Feb 5, 2018 at 7:39 PM, Tomer Maimon <tmaimon77 at gmail.com> wrote:
>>>> On 2 February 2018 at 06:27, Joel Stanley <joel at jms.id.au> wrote:
>>>
>>>>>  /* Uart divisor latch read */
>>>>> @@ -2143,6 +2153,15 @@ int serial8250_do_startup(struct uart_port *port)
>>>>>                                 UART_DA830_PWREMU_MGMT_FREE);
>>>>>         }
>>>>>
>>>>> +       if (port->type == PORT_NPCM) {
>>>>> +               /*
>>>>> +                * Nuvoton calls the scratch register 'UART_TOR' (timeout
>>>>> +                * register). Enable it, and set TIOC (timeout interrupt
>>>>> +                * comparator) to be 0x20 for correct operation.
>>>>> +                */
>>>>> +               serial_port_out(port, UART_NPCM_TOR, UART_NPCM_TOIE | 0x20);
>>>>> +       }
>>>>
>>>> Can we set the TOR register dynamically (adding special IOCTL)?
>>>
>>> We would have to have a good reason to do so.
>>>
>>> Under what circumstances would a user need to change this dynamically?
>>
>> Tomer, I agree it is not important to change at run time.
> I agree Avi,
> but the only problem I see with it that if we will need in the future
> to TOR register dynamically (new SOC, client demand, etc.)
> Could it be done in the 8250 files?

If we have a requirement in the future I suggest we decide how to
solve that problem then. Cross that bridge when we come to it.


>>
>>>
>>>>
>>>>> +
>>>>>  #ifdef CONFIG_SERIAL_8250_RSA
>>>>>         /*
>>>>>          * If this is an RSA port, see if we can kick it up to the
>>>>> @@ -2465,6 +2484,15 @@ static unsigned int xr17v35x_get_divisor(struct uart_8250_port *up,
>>>>>         return quot_16 >> 4;
>>>>>  }
>>>>>
>>>>> +/* Nuvoton NPCM UARTs have a custom divisor calculation */
>>>>> +static unsigned int npcm_get_divisor(struct uart_8250_port *up,
>>>>> +               unsigned int baud)
>>>>> +{
>>>>> +       struct uart_port *port = &up->port;
>>>>> +
>>>>> +       return DIV_ROUND_CLOSEST(port->uartclk, 16 * baud + 2) - 2;
>>>>> +}
>>>
>>> Did you test this version of the driver on your hardware?
> Yes Thanks!!! it working fine.

Great!

>>>
>>> I think I understand why you had the extra code, but does it affect
>>> operation in any of your tests?
>>>
>>> From what I understand the baud rates are often calculated to be "near enough".
>>
>> Joel, I made the original calculation but I agree that this one is
>> good enough (I confused Tomer to think that your calculation will not
>> fit :)
>>
> I see you add the TOR to include/uapi/linux/serial_reg.h will the
> maintainer accept it?

It's in a private include, include/uapi/linux/serial_reg.h.

Initially I didn't do this, but if you examine the end of that file
there are a number of other 8250 ports that do this.

If you would like, I will send this patch upstream now.

Cheers,

Joel