Re: [PATCH 10/10] drivers: PL011: add support for the ARM SBSA generic UART

[Andre Przywara <andre.przywara@arm.com>]

Philip,

sorry for the late reply, that was stuck in my Drafts folder :-(

On 02/17/2015 04:16 PM, Dave P Martin wrote:
> On Tue, Feb 17, 2015 at 10:55:35AM -0500, Philip Elcan wrote:
>> On 01/16/2015 12:23 PM, Andre Przywara wrote:
>>> The ARM Server Base System Architecture[1] document describes a
>>> generic UART which is a subset of the PL011 UART.
>>> It lacks DMA support, baud rate control and modem status line
>>> control, among other things.
>>> The idea is to move the UART initialization and setup into the
>>> firmware (which does this job today already) and let the kernel just
>>> use the UART for sending and receiving characters.
>>> We use the recent refactoring the build a new struct uart_ops
>>> variable which points to some new functions avoiding access to the
>>> missing registers. We reuse as much existing PL011 code as possible.
>>>
>>> In contrast to the PL011 the SBSA UART does not define any AMBA or
>>> PrimeCell relations, so we go a pretty generic probe function
>>> which only uses platform device functions.
>>> A DT binding is provided, but other systems can easily attach to it,
>>> too (hint, hint!).
>>>
>>> Signed-off-by: Andre Przywara <andre.przywara@arm.com>
>>> ---
>>
>> <snip>
>>
>> Andre,
>>
>> I'm a little late to address this patchset, but the SBSA defines all
>> the Generic UART registers 32-bit wide. However, the amba-pl011 driver
>> uses 16-bit accessors. How will you be handling that? Can the ARM PL011
>> hardware handle 32-bit access?
> 
> Interesting question.  The PL011 TRM [1] specifies only a 16-bit-wide
> APB bus interface, but does not say that 32-bit accesses won't work.
> 
> I suspect that 32-bit accesses will work on all or most PL011s -- if
> that looks too risky or we can't find enough test platforms to be sure
> of this, then we could maybe abstract the register access size as a
> quirk.
> 
> Andre may already have an answer on this.

I am not sure if the 32-bit register _width_ mentioned in the spec
really mandates 32-bit accesses, also the width may be just a spec bug.
Since the SBSA states that an ARM PL011r1p5 is a valid SBSA-UART
implementation, I wonder how this goes together. Also the highest
non-reserved bit in the registers is bit 15 in PL011 and bit 11 in the
SBSA subset.

I fear the actual bus connection is an implementation detail. Given the
fact that all existing PL011 hardware so far works with the 16bit
accesses, I don't dare to change this.

So I'd suggest to keep it as readw/writew for now and the revisit this
topic if some SBSA UART users complain. That makes it easier to justify
the rather invasive change to all MMIO accessors (which I already tried
on one machine for now without problems, btw).

Cheers,
Andre.