[PATCH v5 33/33] riscv: Add Sipeed Maix support

[Bin Meng]

Hi Rick,

On Wed, Mar 4, 2020 at 11:11 PM Sean Anderson <seanga2 at gmail.com> wrote:
>
> On 3/4/20 2:47 AM, Rick Chen wrote:
> > Hi Sean
> >
> > This patch series become larger and larger from v1 with 11 patches to
> > v5 with 33 patches.
>
> I didn't intend for it to balloon so large. My original goal for this
> revision was just to get the SPI and MMC slots working. However, I
> discovered that I would need to implement a pinmux slot and GPIOs for
> the MMC slot to work (since I think that is the primary motivating
> factor for using U-Boot on this board, as opposed to the native
> bootloader).
>
> > You shall just fix the suggestions from the previous version in the
> > next version.
>
> I will try to do that :)
>
> > Additional extra features and subsystem drivers that you want to
> > support, you shall send them individually instead of mixing them
> > together.
>
> In the past, I have been told to keep this sort of thing in one series
> (e.g. by Bin Meng). I'm really not sure what things I should split off
> and which I should keep together.
>

If separate patches are sent via different tree, we only get a
complete new RISC-V board support after all trees are merged. Also
there might be inter-dependencies between patches. My practice is to
wait for sub-domain maintainers to give Ack or RB tags, then pull all
series via one tree.

> >
> > If you can separate them into different series(spi, gpio,led ,
> > pinctrl, watchdog, those drivers were not present in v1) and they will
> > not block the reviewing schedule each other.
>
> Hm, ok. Perhaps split those off into series which depend on this one?
>
> > Narrow down the dependency, it can help to speed up the patch work.
> > And I am happy to help pulling the RISC-V relative patchs via riscv
> > tree ASAP.
>
> Ok, I will try and get another version out which fixes the feedback I
> have gotten on those patches.

Regards,
Bin