On Mon, Nov 09, 2020 at 06:10:05PM +0000, Marc Zyngier wrote:
> On 2020-11-09 16:59, Quentin Perret wrote:
> > Hey David,
> >
> > On Monday 09 Nov 2020 at 11:32:26 (+0000), David Brazdil wrote:
> > > Add helper macro for computing the PA of a kernel symbol in nVHE hyp
> > > code. This will be useful for computing the PA of a PSCI CPU_ON entry
> > > point.
> > >
> > > Signed-off-by: David Brazdil <dbraz...@google.com>
> > > ---
> > > arch/arm64/kvm/hyp/nvhe/psci.c | 10 ++++++++++
> > > 1 file changed, 10 insertions(+)
> > >
> > > diff --git a/arch/arm64/kvm/hyp/nvhe/psci.c
> > > b/arch/arm64/kvm/hyp/nvhe/psci.c
> > > index b0b5df590ba5..7510b9e174e9 100644
> > > --- a/arch/arm64/kvm/hyp/nvhe/psci.c
> > > +++ b/arch/arm64/kvm/hyp/nvhe/psci.c
> > > @@ -20,6 +20,16 @@ s64 hyp_physvirt_offset;
> > >
> > > #define __hyp_pa(x) ((phys_addr_t)((x)) + hyp_physvirt_offset)
> > >
> > > +#define __hyp_pa_symbol(sym) \
> > > + ({ \
> > > + extern char sym[]; \
> > > + unsigned long kern_va; \
> > > + \
> > > + asm volatile("ldr %0, =%1" : "=r" (kern_va) \
> > > + : "S" (sym)); \
> > > + kern_va - kimage_voffset; \
> > > + })
> > > +
> >
> > Could this be simplified to __hyp_pa(hyp_symbol_addr(sym))? That would
> > avoid the dependency on kimage_voffset.
Ah, didn't see that one. Ok, removing this patch.
>
> I'm going to move away from evaluating kimage_voffset at runtime anyway,
> see [1].
Awesome! One more dependency gone.
>
> Thanks,
>
> M.
>
> [1] https://lore.kernel.org/r/20201109175923.445945-1-...@kernel.org
> --
> Jazz is not dead. It just smells funny...
