Re: [PATCH 2/4] perf/x86: Fix data source decoding for Skylake

[Thomas Gleixner]

On Wed, 21 Jun 2017, Andi Kleen wrote:
> I didn't add a generic flag for hops because it's not needed
> for Skylake.

When is that going to come? 

> diff --git a/include/uapi/linux/perf_event.h b/include/uapi/linux/perf_event.h
> index b1c0b187acfe..7cfeb54e0b5a 100644
> --- a/include/uapi/linux/perf_event.h
> +++ b/include/uapi/linux/perf_event.h
> @@ -931,14 +931,20 @@ union perf_mem_data_src {
>  			mem_snoop:5,	/* snoop mode */
>  			mem_lock:2,	/* lock instr */
>  			mem_dtlb:7,	/* tlb access */
> -			mem_rsvd:31;
> +			mem_lvl_num:4,	/* memory hierarchy level number */
> +			mem_remote:1,   /* remote */
> +			mem_snoopx:2,	/* snoop mode, ext */

Is that extending snoopx another time?

The question really is, whether the 2 bits are going to be sufficient for a
while or are we going to add snoopx2 in a few month from now.

Thanks,

	tglx