Re: [PATCH v5.99.1 RFC 1/4] xen/arm: Duplicate gic-v2.c file to support hip04 platform version

[Ian Campbell <ian.campbell@citrix.com>]

On Wed, 2015-02-25 at 15:34 +0000, Julien Grall wrote:
> Hi Frediano,
> 
> On 25/02/15 15:28, Frediano Ziglio wrote:
> > HiSilison Hip04 platform use a slightly different version
> 
> I honestly don't like the idea to copy the whole GIC-v2 drivers. It will
> require more maintenance for us.

Will it? This puts the hip driver in its own file where I presume
Frediano et al will take care of it.

> Is there any way to introduce callbacks in the GICv2 code where the
> hisilicon hardware differs?

That's even worse than the first approach (if (hip()) in gic-v2.c) IMHO.

The if-s (or hooks) would be a maintenance burden in the gic-v2.c
driver, making maintenance harder for all the other platforms which
actually followed the spec.

> 
> > Signed-off-by: Frediano Ziglio <frediano.ziglio@huawei.com>
> > ---
> >  xen/arch/arm/Makefile    |   2 +-
> >  xen/arch/arm/gic-hip04.c | 788 +++++++++++++++++++++++++++++++++++++++++++++++
> >  2 files changed, 789 insertions(+), 1 deletion(-)
> >  create mode 100644 xen/arch/arm/gic-hip04.c
> > 
> > diff --git a/xen/arch/arm/Makefile b/xen/arch/arm/Makefile
> > index 41aba2e..6fb8ba9 100644
> > --- a/xen/arch/arm/Makefile
> > +++ b/xen/arch/arm/Makefile
> > @@ -11,7 +11,7 @@ obj-y += vpsci.o
> >  obj-y += domctl.o
> >  obj-y += sysctl.o
> >  obj-y += domain_build.o
> > -obj-y += gic.o gic-v2.o
> > +obj-y += gic.o gic-v2.o gic-hip04.o
> 
> As you do a verbatim copy, you should enable to compilation on the next
> patch.
> 
> Regards,
>