Re: [PATCHv2] perf kvm: add kvm-stat for arm64

[Leo Yan]

On Tue, Sep 15, 2020 at 08:15:41PM +0900, Sergey Senozhatsky wrote:
> On (20/09/15 19:57), Sergey Senozhatsky wrote:
> > On (20/09/15 18:36), Leo Yan wrote:
> > > > +#define HVC_STUB_ERR		  0xbadca11
> > > > +
> > > > +/* Per asm/kvm_asm.h */
> > > > +#define ARM_EXCEPTION_IRQ		0
> > > > +#define ARM_EXCEPTION_EL1_SERROR	1
> > > > +#define ARM_EXCEPTION_TRAP		2
> > > > +#define ARM_EXCEPTION_IL		3
> > >
> > > Nitpick: from completeness, we also can give out KVM exiting reason
> > > for 'ARM_EXCEPTION_IL'.
> >
> > OK, let me take a look.
> 
> I think ARM_EXCEPTION_IL are reported as HYP_GONE. According to
> arch/arm64/include/asm/kvm_asm.h
> 
> ---
> #define ARM_EXCEPTION_IRQ	  0
> #define ARM_EXCEPTION_EL1_SERROR  1
> #define ARM_EXCEPTION_TRAP	  2
> #define ARM_EXCEPTION_IL	  3
> 
> #define kvm_arm_exception_type					\
> 	{ARM_EXCEPTION_IRQ,		"IRQ"		},	\
> 	{ARM_EXCEPTION_EL1_SERROR, 	"SERROR"	},	\
> 	{ARM_EXCEPTION_TRAP, 		"TRAP"		},	\
> 	{ARM_EXCEPTION_HYP_GONE,	"HYP_GONE"	}
> ---
> 
> So it should be reported already.

Thanks for double checking.  But I still think this is incorrect,
ARM_EXCEPTION_HYP_GONE is defined as HVC_STUB_ERR (0xbadca11), which
is different from ARM_EXCEPTION_IL (3).  So I don't understand why you
have the conclusion that 'ARM_EXCEPTION_IL are reported as HYP_GONE'.

Sorry if I miss anything for this.

Thanks,
Leo