Re: [RFC PATCHv5 1/7] HSI: hsi: Introducing HSI framework

[Linus Walleij]

On Wed, Jun 22, 2011 at 9:11 PM, Sjur Brændeland
<sjur.brandeland@xxxxxxxxxxxxxx> wrote:
>>+/*
>>+ * API for HSI clients
>>+ */
>>+int hsi_async(struct hsi_client *cl, struct hsi_msg *msg);
>>+
>>+/**
>
> I'm pleased to see scatter list support. But is this supported by all HW?
> What is the behavior if HW doesn't support this, will hsi_async fail, or
> is the scatter-list handled 'under the hood'?

I think it's pretty straight-forward even if you have to use PIO (IRQs and
CPU-filling of FIFO). You simply use a memory iterator (sg_miter), so
for example the MMC layer can only handle SGlists and in mmci.c we
use sg_miter_start/next/stop to do this in the PIO case.

If that kind of hardware would ever be fast enough for anyone using
a HSI link is a good question though :-)

Yours,
Linus Walleij
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/