diff mbox

[4/6] ARM: rockchip: enable support for RK3288 SoCs

Message ID 4355781.cnmcVo5AVu@diego (mailing list archive)
State New, archived
Headers show

Commit Message

Heiko Stuebner July 15, 2014, 11:01 p.m. UTC
Enable HAVE_ARM_ARCH_TIMER and add a rockchip,rk3288 compatible.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
---
 arch/arm/mach-rockchip/Kconfig    | 1 +
 arch/arm/mach-rockchip/rockchip.c | 1 +
 2 files changed, 2 insertions(+)

Comments

Doug Anderson July 16, 2014, 7:57 p.m. UTC | #1
Heiko,

On Tue, Jul 15, 2014 at 4:01 PM, Heiko Stübner <heiko@sntech.de> wrote:
> Enable HAVE_ARM_ARCH_TIMER and add a rockchip,rk3288 compatible.
>
> Signed-off-by: Heiko Stuebner <heiko@sntech.de>
> ---
>  arch/arm/mach-rockchip/Kconfig    | 1 +
>  arch/arm/mach-rockchip/rockchip.c | 1 +
>  2 files changed, 2 insertions(+)
>
> diff --git a/arch/arm/mach-rockchip/Kconfig b/arch/arm/mach-rockchip/Kconfig
> index e4564c2..d168669 100644
> --- a/arch/arm/mach-rockchip/Kconfig
> +++ b/arch/arm/mach-rockchip/Kconfig
> @@ -6,6 +6,7 @@ config ARCH_ROCKCHIP
>         select ARCH_REQUIRE_GPIOLIB
>         select ARM_GIC
>         select CACHE_L2X0
> +       select HAVE_ARM_ARCH_TIMER

Do we want to think about allowing someone to enable the A9-based
Rockchip SoCs separately than the A12-based ones?  I know it doesn't
hurt to have the arch timer code present on A9 SoCs (it will figure
things out at runtime), but people trying to build an A9-based system
might not want the extra code?

Anyway, I don't feel strongly about it, so:

Reviewed-by: Doug Anderson <dianders@chromium.org>
Tested-by: Doug Anderson <dianders@chromium.org>
Heiko Stuebner July 16, 2014, 8:22 p.m. UTC | #2
Am Mittwoch, 16. Juli 2014, 12:57:21 schrieb Doug Anderson:
> Heiko,
> 
> On Tue, Jul 15, 2014 at 4:01 PM, Heiko Stübner <heiko@sntech.de> wrote:
> > Enable HAVE_ARM_ARCH_TIMER and add a rockchip,rk3288 compatible.
> > 
> > Signed-off-by: Heiko Stuebner <heiko@sntech.de>
> > ---
> > 
> >  arch/arm/mach-rockchip/Kconfig    | 1 +
> >  arch/arm/mach-rockchip/rockchip.c | 1 +
> >  2 files changed, 2 insertions(+)
> > 
> > diff --git a/arch/arm/mach-rockchip/Kconfig
> > b/arch/arm/mach-rockchip/Kconfig index e4564c2..d168669 100644
> > --- a/arch/arm/mach-rockchip/Kconfig
> > +++ b/arch/arm/mach-rockchip/Kconfig
> > @@ -6,6 +6,7 @@ config ARCH_ROCKCHIP
> > 
> >         select ARCH_REQUIRE_GPIOLIB
> >         select ARM_GIC
> >         select CACHE_L2X0
> > 
> > +       select HAVE_ARM_ARCH_TIMER
> 
> Do we want to think about allowing someone to enable the A9-based
> Rockchip SoCs separately than the A12-based ones?  I know it doesn't
> hurt to have the arch timer code present on A9 SoCs (it will figure
> things out at runtime), but people trying to build an A9-based system
> might not want the extra code?
> 
> Anyway, I don't feel strongly about it, so:

I've also thought about this previously. Personally I would want to wait with 
introducing more complexity here until someone comes along with a use case.
Simply because we're talking about 7kb (stripped) for the arch-timer and 
machines with >1GB of memory.
So I'm not adverse to it, but I guess it will make more sense when more soc-
specific code lands - suspend stuff for example.


But I think we should be able to drop the dw_apb_timer altogether, as it stems 
from a time before I found the global-timer informations and all A9 SoCs 
should be able to use this one instead.




> 
> Reviewed-by: Doug Anderson <dianders@chromium.org>
> Tested-by: Doug Anderson <dianders@chromium.org>
Will Deacon July 17, 2014, 10:20 a.m. UTC | #3
On Wed, Jul 16, 2014 at 08:57:21PM +0100, Doug Anderson wrote:
> Heiko,
> 
> On Tue, Jul 15, 2014 at 4:01 PM, Heiko Stübner <heiko@sntech.de> wrote:
> > Enable HAVE_ARM_ARCH_TIMER and add a rockchip,rk3288 compatible.
> >
> > Signed-off-by: Heiko Stuebner <heiko@sntech.de>
> > ---
> >  arch/arm/mach-rockchip/Kconfig    | 1 +
> >  arch/arm/mach-rockchip/rockchip.c | 1 +
> >  2 files changed, 2 insertions(+)
> >
> > diff --git a/arch/arm/mach-rockchip/Kconfig b/arch/arm/mach-rockchip/Kconfig
> > index e4564c2..d168669 100644
> > --- a/arch/arm/mach-rockchip/Kconfig
> > +++ b/arch/arm/mach-rockchip/Kconfig
> > @@ -6,6 +6,7 @@ config ARCH_ROCKCHIP
> >         select ARCH_REQUIRE_GPIOLIB
> >         select ARM_GIC
> >         select CACHE_L2X0
> > +       select HAVE_ARM_ARCH_TIMER
> 
> Do we want to think about allowing someone to enable the A9-based
> Rockchip SoCs separately than the A12-based ones?  I know it doesn't
> hurt to have the arch timer code present on A9 SoCs (it will figure
> things out at runtime), but people trying to build an A9-based system
> might not want the extra code?

More likely, people using the A12-based system won't want the CACHE_L2X0
code (which adds an outer_cache.sync check to wmb()).

Will
diff mbox

Patch

diff --git a/arch/arm/mach-rockchip/Kconfig b/arch/arm/mach-rockchip/Kconfig
index e4564c2..d168669 100644
--- a/arch/arm/mach-rockchip/Kconfig
+++ b/arch/arm/mach-rockchip/Kconfig
@@ -6,6 +6,7 @@  config ARCH_ROCKCHIP
 	select ARCH_REQUIRE_GPIOLIB
 	select ARM_GIC
 	select CACHE_L2X0
+	select HAVE_ARM_ARCH_TIMER
 	select HAVE_ARM_SCU if SMP
 	select HAVE_ARM_TWD if SMP
 	select DW_APB_TIMER_OF
diff --git a/arch/arm/mach-rockchip/rockchip.c b/arch/arm/mach-rockchip/rockchip.c
index 968cc34..8ab9e0e 100644
--- a/arch/arm/mach-rockchip/rockchip.c
+++ b/arch/arm/mach-rockchip/rockchip.c
@@ -29,6 +29,7 @@  static const char * const rockchip_board_dt_compat[] = {
 	"rockchip,rk3066a",
 	"rockchip,rk3066b",
 	"rockchip,rk3188",
+	"rockchip,rk3288",
 	NULL,
 };