[GIT PULL] ARM: aspeed: soc changes for 5.10

[Joel Stanley]

On Sat, 3 Oct 2020 at 20:31, Olof Johansson <olof at lixom.net> wrote:
>
> Hi Joel,
>
> On Fri, Sep 25, 2020 at 06:31:28AM +0000, Joel Stanley wrote:
> > Hello Soc maintainers,
> >
> > Here are some ASPEED changes for the 5.10 merge window.
> >
> > The following changes since commit e55f541e51b5136fc0ced0bdf2b33ee3cca3bc96:
> >
> >   soc: aspeed: xdma: Add reset ioctl (2020-09-25 15:32:57 +0930)
> >
> > are available in the Git repository at:
> >
> >   git://git.kernel.org/pub/scm/linux/kernel/git/joel/aspeed.git
> > tags/aspeed-5.10-soc
> >
> > for you to fetch changes up to e55f541e51b5136fc0ced0bdf2b33ee3cca3bc96:
> >
> >   soc: aspeed: xdma: Add reset ioctl (2020-09-25 15:32:57 +0930)
> >
> > ----------------------------------------------------------------
> > ASPEED soc driver updates for 5.10
> >
> > New drivers:
> >
> >  - XDMA driver for the BMC to host PCIe DMA device
>
> I don't think this driver belongs in drivers/soc, it's not "soc glue
> logic", and it has a userspace interface. Some of the commits even
> references it as a "misc driver". Mind resubmitting it for drivers/misc
> instead?

(Argh I missed this mail)

We used to have the various little drivers in drivers/misc, but
somewhere along the line we decided to move them to drivers/soc. I
think that was Arnd asking Patrick Venture?

https://lore.kernel.org/linux-arm-kernel/CAK8P3a3CK4o8KnD6M084ULEmm+6_CtNFqYHjSqE5vp+U9YAmkA@mail.gmail.com/

The existing drivers have userspace interfaces for misc parts of the
aspeed chip, mostly relating to the strange (from a traditional
Linux/host perspective) interfaces it exposes to the host. This XDMA
driver should go alongside those other drivers.

With that context, how would you like to proceed?

Cheers,

Joel