On Thu, 15 Nov 2018 14:04:00 +0100
Miquel Raynal <miquel.raynal@xxxxxxxxxxx> wrote:
Hi Liang,
Liang Yang <liang.yang@xxxxxxxxxxx> wrote on Thu, 15 Nov 2018 19:25:07
+0800:
Hi Boris,
I have implemented dma access base on these helpers you provided below.
we prepare to send v7 version now, so when will these helpers be pushed?
Thanks for your work. You can send the v7 so we will have a look at the
overall driver; but since we raised the DMA buffers issue we had a
discussion with Boris about how to handle them and I think we are going
to adopt the same solution as Wolfram in the I2C subsystem: manual
flagging. Sadly, this is probably the best we can do to ensure proper
DMA support.
There is nothing set is stone yet but I started a small rework to
handle MTD operations differently (and add a DMA_SAFE flag), you can
have a look there [1]. Don't base your work on it for now as it is just
a preliminary version, subject to big changes.
In order to not block the driver, I'd suggest that you move the helper
I proposed directly into your driver and prefix them with 'meson_'.
.