[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: New flavour combined orion5x and kirkwood



On Tue, 2016-02-09 at 00:48 +0900, Roger Shimizu wrote:
> On Mon, Feb 8, 2016 at 10:36 PM, Ben Hutchings <ben@decadent.org.uk> wrote:
> > On Mon, 2016-02-08 at 13:12 +0900, Roger Shimizu wrote:
> > > [CC: debian-kernel@lists.debian.org and debian-arm@lists.debian.org]
> > > 
> > > Dear Ben, Ian, Martin,
> > > 
> > > I noticed orion5x and kirkwood are combined into new flavour called
> > > marvell in both kernel and flash-kernel git repo [0][1].
> > > 
> > > Maybe it's too late,
> > 
> > It's not; that name isn't even used in experimental yet.
> > 
> > > but I still hope the flavour can be named
> > > "mvebu", which will be consistent with upstream.
> > > AFAIK, all orion5x and kirkwood changes go into mvebu repo [2] first,
> > > and then merge into arm-soc repo [3], and finally reach linus repo
> > > [4].
> > [...]
> > 
> > I did consider using 'mvebu', but I thought that people might assume
> > that it was supposed to support all the ARMv5 SoCs included in that
> > family.  That isn't the case and we're unlikely to add support for more
> > SoCs at this stage.

I'm sorry, I misremembered what my reasoning was.  The only ARMv5 SoC
included in 'mvebu' is Kirkwood.  Orion is not part of 'mvebu', even
though they have many functional blocks and drivers in common.

> Thanks for your feedback!
> Now I know why not to choose "mvebu", but I still cannot understand why to
> choose "marvell", which seems more generic/wide scope.

It actually does have wider scope.

> I checked
> - arch/arm/configs/mvebu_v5_defconfig
> and compared with
> - arch/arm/configs/multi_v5_defconfig
> 
> I find mvebu_v5 actually is less than multi_v5 for a few flavour, such
> as iMX, U300.
> And in mvebu_v5, there's only orion5x and kirkwood related items.
> 
> AFAIK mvebu_v7 belongs to armhf, so we can safely call upstream's mvebu_v5
> as armel-mvebu.

Ben.

-- 
Ben Hutchings
Nothing is ever a complete failure; it can always serve as a bad example.

Attachment: signature.asc
Description: This is a digitally signed message part


Reply to: