Re: [Xen-devel] [PATCH 1/2] arm: smccc: handle SMCs/HVCs according to SMCCC

[Volodymyr Babchuk <volodymyr_babchuk@xxxxxxxx>]

Hello Julien,


On 27.06.17 14:08, Julien Grall wrote:
> On 06/22/2017 05:29 PM, Volodymyr Babchuk wrote:
> > Hi Julien,
> Hi Volodymyr,
>
> > On 15.06.17 13:48, Julien Grall wrote:
> > > > diff --git a/xen/arch/arm/traps.c b/xen/arch/arm/traps.c
> > > > index 6cf9ee7..2d0b058 100644
> > > > --- a/xen/arch/arm/traps.c
> > > > +++ b/xen/arch/arm/traps.c
> > > > @@ -44,6 +44,7 @@
> > > >  #include <asm/cpufeature.h>
> > > >  #include <asm/flushtlb.h>
> > > >  #include <asm/monitor.h>
> > > > +#include <asm/smccc.h>
> > > >
> > > >  #include "decode.h"
> > > >  #include "vtimer.h"
> > > > @@ -2781,11 +2782,16 @@ static void do_trap_smc(struct 
> > > > cpu_user_regs *regs, const union hsr hsr)
> > > >  {
> > > I think it would make sense to push this function in the new file.
> > Unfortunately, I can't do this, because it uses local functions such as
> > inject_undef_exception() or advance_pc().
> The code is not set in stone. It is perfectly fine to rework it if it 
> does not fit.
Okay. Probably, I can put forward declarations of the mentioned 
functions into as-arm/processor.h.
Should I leave implementation of those functions in the traps.c or 
should I move them into processor.c ?
> > > Also, I was expecting some change in the HVC path as you say that 
> > > this will be used for both HVC and SMC.
> > Actually, I plan to use this particular function to handle only SMCs,
> > because it does SMC-specific tasks, such as calling to a monitor.
> >
> > HVCs will be handled in different call path in the next patch,
> > because currently HVC callpath is used by PSCI code.
> But your commit title says: "handle SMCs/HVCs according to SMCCC". So 
> something does not match...
Yes, you are right. I'll rework the patches.

Will you review the V2 series or should I push V3?




_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel