Re: [PATCH 2/2] dmaengine: apple-sio: Add Apple SIO driver

[Martin Povišer]

Sorry I missed this message before.

> On 3. 8. 2023, at 13:25, Vinod Koul <vkoul@xxxxxxxxxx> wrote:
> 
> On 01-08-23, 23:55, Martin Povišer wrote:
> 
>>> can you use virt_dma_chan, that should simplify list handling etc
>> 
>> I looked into that when I wrote the sister driver apple-admac.c, I don’t
>> remember anymore why I decided against it, and I don’t think it came up
>> during review. Now that this driver is done, I hope we can take it as is.
>> 
>> There’s some benefit from the drivers having a similar structure, I sent
>> one or two fixes to apple-admac for things I found out because I was
>> writing this other driver.
> 
> And this would be a chance to covert the other one and get rid of list
> handling code in that driver as well

I guess...

>>>> +};
>>>> +
>>>> +#define SIO_NTAGS		16
>>>> +
>>>> +typedef void (*sio_ack_callback)(struct sio_chan *, void *, bool);
>>> 
>>> any reason not to use dmaengine callbacks?
>> 
>> Not sure what dmaengine callback you mean here. This callback means
>> the coprocessor acked a tag, not sure how we can fit something dmaengine
>> onto it.
> 
> Okay lets understand, how is this one used

This one is used to signal completion of IPC calls to the coprocessor when
that call is made from atomic context. Only user (currently) is issuing of
coproc descriptors. I can provide more detail but not sure in what
direction.

Martin

> -- 
> ~Vinod
>