Subject: Re: [PATCH v2 2/4] perf/x86: Fix data source decoding for Skylake
To: Peter Zijlstra <peterz@infradead.org>,
        Andi Kleen <andi@firstfloor.org>
Cc: acme@kernel.org, linux-kernel@vger.kernel.org, jolsa@kernel.org,
        eranian@google.com, Andi Kleen <ak@linux.intel.com>,
        sukadev@linux.vnet.ibm.com, Michael Ellerman <mpe@ellerman.id.au>
References: <20170607232226.26365-1-andi@firstfloor.org>
 <20170607232226.26365-3-andi@firstfloor.org>
 <20170608081531.iv27xaghntniwmm6@hirez.programming.kicks-ass.net>
From: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Date: Fri, 9 Jun 2017 13:34:35 +0530
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101
 Thunderbird/52.1.1
MIME-Version: 1.0
In-Reply-To: <20170608081531.iv27xaghntniwmm6@hirez.programming.kicks-ass.net>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 7bit
Content-Language: en-US
Message-Id: <1d43622e-e1f3-de5a-ae49-9b4926db0ad4@linux.vnet.ibm.com>
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 2696
Lines: 77


On Thursday 08 June 2017 01:45 PM, Peter Zijlstra wrote:
> On Wed, Jun 07, 2017 at 04:22:24PM -0700, Andi Kleen wrote:
>
>> diff --git a/include/uapi/linux/perf_event.h b/include/uapi/linux/perf_event.h
>> index b1c0b187acfe..95daade294d7 100644
>> --- a/include/uapi/linux/perf_event.h
>> +++ b/include/uapi/linux/perf_event.h
>> @@ -931,14 +931,18 @@ union perf_mem_data_src {
>>   			mem_snoop:5,	/* snoop mode */
>>   			mem_lock:2,	/* lock instr */
>>   			mem_dtlb:7,	/* tlb access */
>> -			mem_rsvd:31;
>> +			mem_lvlx:8,	/* memory hierarchy level, ext */
>> +			mem_snoopx:2,	/* snoop mode, ext */
>> +			mem_rsvd:21;
>>   	};
>>   };
>>   #elif defined(__BIG_ENDIAN_BITFIELD)
>>   union perf_mem_data_src {
>>   	__u64 val;
>>   	struct {
>> -		__u64	mem_rsvd:31,
>> +		__u64	mem_rsvd:21,
>> +			mem_snoopx:2,	/* snoop mode, ext */
>> +			mem_lvlx:8,	/* memory hierarchy level, ext */
>>   			mem_dtlb:7,	/* tlb access */
>>   			mem_lock:2,	/* lock instr */
>>   			mem_snoop:5,	/* snoop mode */
> So one thing we could do is add a mem_hops field and always set that,
> even for the old stuff. The old stuff will not know about that field and
> ignore the bits, but new stuff will then not need as many LVL bits.
>
> Of course, we then get into the problem of how many bits of hops we
> need.. Power guys ?

Currently we support 3 hops (local, remote and distant) and
in future we may have another for capi. So 4 levels of hops
might do. 8 would be nice future proof.

Maddy

>
>> @@ -975,6 +979,16 @@ union perf_mem_data_src {
>>   #define PERF_MEM_LVL_UNC	0x2000 /* Uncached memory */
>>   #define PERF_MEM_LVL_SHIFT	5
>>   
>> +#define PERF_MEM_LVLX_L4	0x01 /* L4 */
>> +#define PERF_MEM_LVLX_REM_L4    0x02 /* Remote L4 */
>> +#define PERF_MEM_LVLX_REM_RAM	0x04 /* Remote Ram, unknown hops */
>> +#define PERF_MEM_LVLX_PMEM	0x08 /* Persistent Memory */
>> +#define PERF_MEM_LVLX_REM_PMEM	0x10 /* Remote Persistent Memory */
>> +#define PERF_MEM_LVLX_REM_NA	0x20 /* Remote N/A level */
> Still wondering what the point of REM_NA is.. can you explain?
>
>> +/* 2 free */
>> +
>> +#define PERF_MEM_LVLX_SHIFT	33
>> +
>>   /* snoop mode */
>>   #define PERF_MEM_SNOOP_NA	0x01 /* not available */
>>   #define PERF_MEM_SNOOP_NONE	0x02 /* no snoop */
>> @@ -983,6 +997,10 @@ union perf_mem_data_src {
>>   #define PERF_MEM_SNOOP_HITM	0x10 /* snoop hit modified */
>>   #define PERF_MEM_SNOOP_SHIFT	19
>>   
>> +#define PERF_MEM_SNOOPX_FWD	0x01 /* forward */
>> +/* 1 free */
>> +#define PERF_MEM_SNOOPX_SHIFT	41
>> +
>>   /* locked instruction */
>>   #define PERF_MEM_LOCK_NA	0x01 /* not available */
>>   #define PERF_MEM_LOCK_LOCKED	0x02 /* locked transaction */
>> -- 
>> 2.9.4
>>