[GIT PULL] ARM: hisi: soc changes for v4.8

[olof@lixom.net (Olof Johansson)]

On Wed, Jul 06, 2016 at 11:15:24AM +0100, Wei Xu wrote:
> Hi Olof,
> 
> On 06/07/2016 05:33, Olof Johansson wrote:
> > Hi,
> > 
> > On Tue, Jun 28, 2016 at 06:43:48PM +0100, Wei Xu wrote:
> >> Hi Arnd, Hi Olof, Hi Kevin,
> >>
> >> Please help to pull the following changes.
> >> Thanks!
> >>
> >> Best Regards,
> >> Wei
> >>
> >> ---
> >>
> >> The following changes since commit 1a695a905c18548062509178b98bc91e67510864:
> >>
> >>   Linux 4.7-rc1 (2016-05-29 09:29:24 -0700)
> >>
> >> are available in the git repository at:
> >>
> >>   git://github.com/hisilicon/linux-hisi.git tags/hisi-armv7-soc-for-4.8
> >>
> >> for you to fetch changes up to c90047616d4411ff6e7a8a40ed87731dda0753b9:
> >>
> >>   ARM: hisi: add compatible string for Hi3519 soc (2016-06-28 16:24:33 +0100)
> >>
> >> ----------------------------------------------------------------
> >> ARM: mach-hisi: Hisilicon SoC updates for 4.8
> >>
> >> - Add initial support for the hi3519 SoC based on armv7
> >> - Avoid the compiling warning by making unexported symbols static
> >>
> >> ----------------------------------------------------------------
> >> Ben Dooks (1):
> >>       ARM: hisi: make unexported symbols static
> >>
> >> Jiancheng Xue (1):
> >>       ARM: hisi: add compatible string for Hi3519 soc
> > 
> > This patch does more than this, it consolidates a bunch of machine entries.
> > 
> > Since none of them are needed (the generic machine will match instead), just
> > remove all of them instead, please.
> > 
> > I'd recommend a future cleanup to do this instead, as I did for Qualcomm, but
> > since you're already churning this code it's better to do it now.
> >
> > So, please respin this pull request with the above addressed.
> 
> There is a specific "map_io" for hi3620 so that we kept 2 machine entries.
> Do you mean we need to merge all of them in one machine entry?
> In that case, I will add a new "hisilicon_map_io" function to invoke
> "debug_ll_io_init" and check the board compatible string.
> Is it OK?

I mean that you only need one machine entry: The one that needs map_io. The
others don't need a machine entry at all; they'll be handled by the generic
entry in arch/arm/kernel/devtree.c.


-Olof