On Tue, Dec 06, 2016 at 11:55:08AM +0000, Will Deacon wrote:
> Hi Andy,
>
> On Tue, Nov 29, 2016 at 01:44:22AM -0600, Andy Gross wrote:
> > This patch adds a quirk parameter to the arm_smccc_smc call. The quirk
> > structure allows for specialized SMC operations due to SoC specific
> > requirements.
> >
> > This patch also fixes up all the current users of the arm_smccc_smc API.
> >
> > This patch and partial implementation was suggested by Will Deacon.
> >
> > Signed-off-by: Andy Gross <andy.gr...@linaro.org>
> > ---
> > arch/arm/kernel/smccc-call.S | 3 ++-
> > arch/arm/mach-artpec/board-artpec6.c | 2 +-
> > arch/arm64/kernel/asm-offsets.c | 7 +++++--
> > arch/arm64/kernel/smccc-call.S | 3 ++-
> > drivers/clk/rockchip/clk-ddr.c | 6 +++---
> > drivers/devfreq/rk3399_dmc.c | 6 +++---
> > drivers/firmware/meson/meson_sm.c | 2 +-
> > drivers/firmware/psci.c | 2 +-
> > drivers/firmware/qcom_scm-64.c | 4 ++--
> > drivers/gpu/drm/mediatek/mtk_hdmi.c | 2 +-
> > include/linux/arm-smccc.h | 18 ++++++++++++++++--
> > 11 files changed, 37 insertions(+), 18 deletions(-)
>
> Thanks for respinning this; I'd forgotten about it!
>
> > diff --git a/arch/arm/kernel/smccc-call.S b/arch/arm/kernel/smccc-call.S
> > index 37669e7..e77950a 100644
> > --- a/arch/arm/kernel/smccc-call.S
> > +++ b/arch/arm/kernel/smccc-call.S
> > @@ -47,7 +47,8 @@ UNWIND( .fnend)
> > /*
> > * void smccc_smc(unsigned long a0, unsigned long a1, unsigned long a2,
> > * unsigned long a3, unsigned long a4, unsigned long a5,
> > - * unsigned long a6, unsigned long a7, struct arm_smccc_res *res)
> > + * unsigned long a6, unsigned long a7, struct arm_smccc_res *res,
> > + * struct arm_smccc_quirk *quirk)
>
> I'd not thought of doing it like this -- I envisaged embedding the quirk
> structure into arm_smccc_res, but this works too. I wonder if we could avoid
> having to pass NULL everywhere if we renamed arm_smccc_{hvc.smc} and added
> a default wrapper around them?
>
> For example, rename arm_smccc_smc to __arm_smccc_smc, add a macro called
> arm_smccc_smc that passes a NULL argument for the quirk, then finally add
> a macro arm_smccc_smc_quirk that takes the additional parameter?
This would work pretty well. I'll send another version with this
implementation.
Regards,
Andy
