Re: [virtio-comment] [PATCH] [PATCH v5] virtio-spi: add the device specification

[Cornelia Huck <cohuck@redhat.com>]

On Wed, Nov 29 2023, Haixu Cui <quic_haixcui@quicinc.com> wrote:

> Hi Viresh,
>
> On 11/29/2023 3:30 PM, Viresh Kumar wrote:
>> On 28-11-23, 20:58, Haixu Cui wrote:
>>> On 11/27/2023 6:17 PM, Viresh Kumar wrote:
>>>> On 24-11-23, 15:20, Haixu Cui wrote:
>>>>> +For half-duplex read transfer, \field{rx_buf} is filled by Virtio SPI device and consumed
>>>>> +by Virtio SPI driver. For half-duplex write transfer, \field{tx_buf} is filled by Virtio
>>>>> +SPI driver and consumed by Virtio SPI device. And for full-duplex read and write transfer,
>>>>> +both \field{tx_buf} and \field{rx_buf} are used.
>>>>
>>>> Should the length of both the buffers in full-duplex mode be same ? If yes, then
>>>> this should be mentioned (in case it is not).
>>>>
>>>
>>> No, there is no such limitation. Write and read buffers may be different is
>>> size.
>> 
>> Hmm, I worked with a SPI controller over a decade ago, and I must be forgetting
>> something here I guess. But from whatever little I remember, with full-duplex
>> transfer, data flows on both MOSI and MISO lines as soon as clock signal is
>> applied.  And so amount of data sent is always be equal to amount of data
>> received by both sides.
>> 
>> Also if I see Linux's implementation of the `struct spi_transfer` [1], I see
>> `tx_buf`, `rx_buf` and a single `len` field, which applies to both the buffers.
>> Which I guess is indicating that both buffers are supposed to be of same length.
>> 
>> What am I missing ?
>
> Oh so sorry for that. And I don't make it clear. Yes, tx_buf and rx_buf 
> have the same size, Linux has such restriction. Just as you mention, 
> kernel level spi_transfer has single "len", the same for 
> spi_ioc_transfer passed from the userland.
>
> But I am not sure if this is in the scope of the spec. Because this is 
> ensured by Linux, but Virtio SPI driver won't also can't verify this.
> This is a prerequisite for virtio spi processing requests.
>
> What is your suggestion? How about adding some descriptions here, like 
> "for full-duplex, tx_buf and rx_buf are same in size, this is guaranteed 
> by the kernel"?

We must not really make any assumptions in the spec about concrete
implementations (here, the Linux kernel), as someone implementing it in
a different environment will need to make explicit choices.

So, if tx_buf and rx_buf are required to be of the same size, it needs
to be explicitly stated in the spec, or an implementation might choose
to do it differently.