Re: [PATCH] serial: 8250_bcm7271: improve bcm7271 8250 port

[Justin Chen <justin.chen@broadcom.com>]

[-- Attachment #1: Type: text/plain, Size: 1317 bytes --]



On 8/14/23 8:12 AM, Andy Shevchenko wrote:
> On Sat, Aug 12, 2023 at 09:24:21PM -0700, Justin Chen wrote:
>> On Sat, Aug 12, 2023 at 3:50 AM Greg Kroah-Hartman
>> <gregkh@linuxfoundation.org> wrote:
>>> On Fri, Aug 11, 2023 at 03:14:01PM -0700, Justin Chen wrote:
> 
>>>> +     [PORT_BCM7271] = {
>>>> +             .name           = "bcm7271_uart",
> 
> This is badly named port type.
> 

Would "Brcmstb 7271 UART" suffice?

>>>> +             .fifo_size      = 32,
>>>> +             .tx_loadsz      = 32,
>>>> +             .fcr            = UART_FCR_ENABLE_FIFO | UART_FCR_R_TRIG_01,
>>>> +             .rxtrig_bytes   = {1, 8, 16, 30},
>>>> +             .flags          = UART_CAP_FIFO | UART_CAP_AFE
>>>> +     },
>>>>   };
> 
> This is almost a dup of PORT_ALTR_16550_F32. Use it if you wish.
> You can always rename it if it feels the right thing to do.
> 

There is some other PORT_ALTR logic that I would like to avoid. I would 
also like to avoid future changes to PORT_ALTR that wouldn't be 
applicable to us.

> But why 8 and not 16 is the default rxtrig?
> 

We were seeing some latency issues on our chips where 16 would cause 
overflows. Trying to kill 2 birds with one stone. If creating another 
port type is avoidable then alternatively I can change the default in 
userspace.

Thanks,
Justin

[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4206 bytes --]