Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Subject: Re: [PATCH RFC 1/1] perf stat: do not fatal if the leader is errored
To:     "Liang, Kan" <kan.liang@linux.intel.com>,
        Like Xu <like.xu.linux@gmail.com>
Cc:     linux-kernel@vger.kernel.org, peterz@infradead.org,
        mingo@redhat.com, acme@kernel.org, mark.rutland@arm.com,
        alexander.shishkin@linux.intel.com, jolsa@kernel.org,
        namhyung@kernel.org, joe.jin@oracle.com,
        linux-perf-users@vger.kernel.org, kvm list <kvm@vger.kernel.org>
References: <20220922071017.17398-1-dongli.zhang@oracle.com>
 <ad2572d0-06b5-7250-31f2-a5efa1048cc0@gmail.com>
 <27cb9747-8911-b3cc-25d9-9438521db832@linux.intel.com>
From:   Dongli Zhang <dongli.zhang@oracle.com>
Message-ID: <d03d8a07-a05a-f03e-189d-a07c6aecbb8a@oracle.com>
Date:   Thu, 22 Sep 2022 11:00:26 -0700
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101
 Thunderbird/78.6.1
In-Reply-To: <27cb9747-8911-b3cc-25d9-9438521db832@linux.intel.com>
Content-Type: text/plain; charset=utf-8
Content-Language: en-US
Content-Transfer-Encoding: 7bit
MIME-Version: 1.0
X-MS-Exchange-AntiSpam-MessageData-ChunkCount: 1
X-MS-Exchange-AntiSpam-MessageData-0: =?utf-8?B?anFNcWpqeXBDQ0RWZUN4ZTFZeUREMkRBSlhiOW5GOW5aUWU5QitZaS9GQ01p?=
 =?utf-8?B?UFVvM1Zvb2ZuYXo4Q254cTBYUDdHU201b1BHZXlBdThzN3pGbWErZTBOcjVv?=
 =?utf-8?B?dDM1TUc0dlFsWGlReVoxQnYzR2tnMkxGR2tOTUxQUTVZM0NNZnUxaFBVM1Fa?=
 =?utf-8?B?WEJ4Um54dTc0b2k0S3p6eWxaZlRzTVlBektBV3dqc1UxcU1LT1BmbnV6dTRN?=
 =?utf-8?B?SCs0eVMzT1duSXRGNTdsWWd1dmNSNTRac1V0SDJrOEFLY2tWSzVSOXhWZHBq?=
 =?utf-8?B?QncxeTd6VmkzWE1JK1BrWGtocWhIL2swRmJVV2J5c0NTUHlqUEhXZlIxZFdH?=
 =?utf-8?B?UyttTXF1UUk3ZWhRbkQrVmhudjJybXliNWR6YXFkOTR0ZTd1dFVGS3pmTW9Y?=
 =?utf-8?B?TjBnbXVrTjFmSnp3Mmo3NGlTQWRQbzJ3WUVoRVFFS0JaN2p5eENTZmhuVnJl?=
 =?utf-8?B?UFBQNGh2Zk9jMk9EanVMdzR2T2Z0cjVIYWFXdjVMcThrVk4vQkptY0lJS3Qz?=
 =?utf-8?B?anJWSXNFYWFOSjB1N1lGU2VjQ1ZPZWQrZzA4TVpnR1RtQjEzV09VOHRVUEhs?=
 =?utf-8?B?ZkJNejN1ZVZlTXp0R2VXYXQ2Z0pGNnAvNGhkeWtZK0Q4RWtMa21BM2xUb1lF?=
 =?utf-8?B?dEJBaTg1MlljbUxmb280Z3hHUjJUM0hDLzl0YzA1VExxbXY5T0xFa2s5Vm5M?=
 =?utf-8?B?RlRFL0YyaktURE90ZzJXR25JZkR2K2JqVk0zVnVqVVFYYk8zR3ppVU04dEp1?=
 =?utf-8?B?b3k5ZmtSRHFlYlEyM2RKOU5HNXFYVkpVS042Rkx1ZlkyZjVvenZmY1YzNTRW?=
 =?utf-8?B?ajVuQ2ZnWUw5M2xPU1RwSWhLTjN3akYyWnhGT3BBWFBRZVBsbDRqVW9va2xi?=
 =?utf-8?B?NDY5aGdHWEJQcTUyZnlxbXdBWTIyRFNPWjcyK09ZUXBIaExZdVYzcG1ZL2tU?=
 =?utf-8?B?L1BYcXdHMzRsdkFKUHYwVkpLcVdpbDVWaGQyT2JkS0k4K0RlcGc2QmNnR2gy?=
 =?utf-8?B?MEx3RFNCZFlUVDJvRkJMM29LUkRGVmdSQnQ0Z1RhVDVMK0dRbEZYNVhmMVZ4?=
 =?utf-8?B?ZG9yMXczRFNTNFNmb1AySHRVVDVYRjFOSlNtOXpQY3JNa250UmxmbkMvUTNi?=
 =?utf-8?B?UXM1WWxiSnNrUlhkYXRUbzlXb084YkF6UVp5Z0d5SXA2aUY1YXdQZ1lHYzZW?=
 =?utf-8?B?NXcxSHQwWHplanJ2aC9CV2l5ZUQ1ZUYwVXVoSjQ5a1NkVUtTUngzZmF2WnBp?=
 =?utf-8?B?QjhKL3dzbFRPSnhXR1RYaWMvSldMV3hrSFcvRjNKWmIxOE44VVVPYXZ3d29C?=
 =?utf-8?B?N2QrZkpwVWJDMVVsQWM3b0hGV0d2djk1cHp3TGcrcWpOVk5RZWJlVmVkRUFY?=
 =?utf-8?B?OXovcnhkaEpad0R5d3A1cUNUV0VpblZaMmpXeHBJcGVrb3Y5QlRITUt1YmZJ?=
 =?utf-8?B?SXFyUnAyQ25JWUZ3ZDBoZDdMak9tdkRYMmJSRlQvczltZ05CdFZaVDY1NlFi?=
 =?utf-8?B?V2hvWnFsdm9ZVHVuMjBQckdyakFsajBEenV2VGtwWU9WR0YzYy80cTF1NnJ3?=
 =?utf-8?B?c2MzYS9yZ2VMWXE2WXB4VmlUYWpRaEdRNFpEZVNDbnBUTFEvKzhBMXlRUzZV?=
 =?utf-8?B?WkFzbWlYTExReit4NUo2dUMwUDdMazRCQzhDMWFTczZ6UXlXaUtOT21sa0Nn?=
 =?utf-8?B?cFowY0NkNVdyUk1LNzVtQndRT0duVmJlZGVvQkw4UTBjWWFPY0Q4TkVVNzRu?=
 =?utf-8?B?elptWHl1Y3pKL3Z5ZHcwQkx2NURFa1NFUXpwRkRSemFNYy94dXFkKzREZEpV?=
 =?utf-8?B?Y094NkpRN3hnNVJ1clR4QlZVWEp6citERStqRlpOcEd1eit0SGNjVUxGcjhH?=
 =?utf-8?B?MGpBNDZpSTNsQkpUMk1UNVV0Tm5aaUlLOUVsaEtzenRLMjgzS1dQN2taN1VY?=
 =?utf-8?B?OStUQWxHWVkzZzVicmIweElFenFZWmZoWlZ6MGI3U2wvZUVLRXd0azlCV3lH?=
 =?utf-8?B?ZTgwK1ZZT2QzanJOUXhOV3NvaTJmdkJVR0hrWDIwOFJFaU9JMEVqZFRKeUx5?=
 =?utf-8?B?TklyVGhvM3dCZ0Nrdk0xT201ZHoxMWVNMEdJcWZlMnBtd3dCejVQRlEzNDlp?=
 =?utf-8?B?eGdSTGlXOWIwS2ZPTXJjWmd2RUp1bDFtdUNjTDBOWm5CM3dkbTREZTB4SzFu?=
 =?utf-8?Q?AG5mb1+lQ/Br/9A0clD6saY=3D?=
X-OriginatorOrg: oracle.com
X-MS-Exchange-CrossTenant-Network-Message-Id: 34cc2b60-c520-4d04-7f2b-08da9cc457a0
X-MS-Exchange-CrossTenant-AuthSource: BYAPR10MB2663.namprd10.prod.outlook.com
X-MS-Exchange-CrossTenant-AuthAs: Internal
X-MS-Exchange-CrossTenant-OriginalArrivalTime: 22 Sep 2022 18:00:32.5099
 (UTC)
X-MS-Exchange-CrossTenant-FromEntityHeader: Hosted
X-MS-Exchange-CrossTenant-Id: 4e2c6054-71cb-48f1-bd6c-3a9705aca71b
X-MS-Exchange-CrossTenant-MailboxType: HOSTED
X-MS-Exchange-CrossTenant-UserPrincipalName: jjPJgP5WndKq3fxo+v2NMH51sYMySNDKw710yVw+XDzhtMRsHGG5fzuTLrhTyggp2CxybwmgjAf7an7wET53WA==
X-MS-Exchange-Transport-CrossTenantHeadersStamped: MW4PR10MB6395
X-Proofpoint-Virus-Version: vendor=baseguard
 engine=ICAP:2.0.205,Aquarius:18.0.895,Hydra:6.0.528,FMLib:17.11.122.1
 definitions=2022-09-22_12,2022-09-22_01,2022-06-22_01
X-Proofpoint-Spam-Details: rule=notspam policy=default score=0 suspectscore=0 mlxlogscore=999
 spamscore=0 adultscore=0 mlxscore=0 malwarescore=0 bulkscore=0
 phishscore=0 classifier=spam adjust=0 reason=mlx scancount=1
 engine=8.12.0-2209130000 definitions=main-2209220119
X-Proofpoint-GUID: VtAEvO76XeypFOpiNtd2Bi6Wx_G-shPb
X-Proofpoint-ORIG-GUID: VtAEvO76XeypFOpiNtd2Bi6Wx_G-shPb
X-Spam-Status: No, score=-4.7 required=5.0 tests=BAYES_00,DKIM_SIGNED,
        DKIM_VALID,DKIM_VALID_AU,DKIM_VALID_EF,NICE_REPLY_A,RCVD_IN_DNSWL_LOW,
        RCVD_IN_MSPIKE_H2,SPF_HELO_NONE,SPF_NONE autolearn=ham
        autolearn_force=no version=3.4.6
X-Spam-Checker-Version: SpamAssassin 3.4.6 (2021-04-09) on
        lindbergh.monkeyblade.net
Precedence: bulk
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

Hi Kan,

I have tested that the patch works by hiding 'slots' sysfs entries.

# ll /sys/bus/event_source/devices/cpu/events/
total 0
-r--r--r--. 1 root root 4096 Sep 22 17:57 branch-instructions
-r--r--r--. 1 root root 4096 Sep 22 17:57 branch-misses
-r--r--r--. 1 root root 4096 Sep 22 17:57 bus-cycles
-r--r--r--. 1 root root 4096 Sep 22 17:57 cache-misses
-r--r--r--. 1 root root 4096 Sep 22 17:57 cache-references
-r--r--r--. 1 root root 4096 Sep 22 17:57 cpu-cycles
-r--r--r--. 1 root root 4096 Sep 22 17:57 instructions
-r--r--r--. 1 root root 4096 Sep 22 17:57 ref-cycles

# perf stat
^C
 Performance counter stats for 'system wide':

         19,256.20 msec cpu-clock                        #   16.025 CPUs
utilized
               179      context-switches                 #    9.296 /sec

                17      cpu-migrations                   #    0.883 /sec

                 3      page-faults                      #    0.156 /sec

         7,502,294      cycles                           #    0.000 GHz

         2,512,587      instructions                     #    0.33  insn per
cycle
           552,989      branches                         #   28.717 K/sec

            15,999      branch-misses                    #    2.89% of all
branches

       1.201628129 seconds time elapsed


Would you send this patch to fix at the kernel side?

Thank you very much!

Dongli Zhang

On 9/22/22 6:34 AM, Liang, Kan wrote:
> 
> 
> On 2022-09-22 4:07 a.m., Like Xu wrote:
>> On 22/9/2022 3:10 pm, Dongli Zhang wrote:
>>> There are three options to fix the issue.
>>>
>>> 1. Do not expose /sys/bus/event_source/devices/cpu/events/slots to
>>> userspace so that pmu_have_event(pmu->name, "slots") returns false.
>>
>> IMO, the guest PMU driver should be fixed
>> since it misrepresents emulated hardware capabilities in terms of slots.
> 
> Yes, we need to fix the kernel to hide the slots event if it's not
> available.
> 
> The patch as below should fix it. (Not tested yet)
> 
> diff --git a/arch/x86/events/intel/core.c b/arch/x86/events/intel/core.c
> index b20e646c8205..27ee43faba32 100644
> --- a/arch/x86/events/intel/core.c
> +++ b/arch/x86/events/intel/core.c
> @@ -5565,6 +5565,19 @@ static struct attribute *intel_pmu_attrs[] = {
>  	NULL,
>  };
> 
> +static umode_t
> +td_is_visible(struct kobject *kobj, struct attribute *attr, int i)
> +{
> +	/*
> +	 * Hide the perf metrics topdown events
> +	 * if the slots is not in CPUID.
> +	 */
> +	if (x86_pmu.num_topdown_events)
> +		return (x86_pmu.intel_ctrl & INTEL_PMC_MSK_FIXED_SLOTS) ? attr->mode : 0;
> +
> +	return attr->mode;
> +}
> +
>  static umode_t
>  tsx_is_visible(struct kobject *kobj, struct attribute *attr, int i)
>  {
> @@ -5600,6 +5613,7 @@ default_is_visible(struct kobject *kobj, struct
> attribute *attr, int i)
> 
>  static struct attribute_group group_events_td  = {
>  	.name = "events",
> +	.is_visible = td_is_visible,
>  };
> 
>  static struct attribute_group group_events_mem = {
> @@ -5758,6 +5772,23 @@ static inline int
> hybrid_find_supported_cpu(struct x86_hybrid_pmu *pmu)
>  	return (cpu >= nr_cpu_ids) ? -1 : cpu;
>  }
> 
> +static umode_t hybrid_td_is_visible(struct kobject *kobj,
> +					struct attribute *attr, int i)
> +{
> +	struct device *dev = kobj_to_dev(kobj);
> +	struct x86_hybrid_pmu *pmu =
> +		 container_of(dev_get_drvdata(dev), struct x86_hybrid_pmu, pmu);
> +
> +	if (!is_attr_for_this_pmu(kobj, attr))
> +		return 0;
> +
> +	/* Only check the big core which supports perf metrics */
> +	if (pmu->cpu_type == hybrid_big)
> +		return (pmu->intel_ctrl & INTEL_PMC_MSK_FIXED_SLOTS) ? attr->mode : 0;
> +
> +	return attr->mode;
> +}
> +
>  static umode_t hybrid_tsx_is_visible(struct kobject *kobj,
>  				     struct attribute *attr, int i)
>  {
> @@ -5784,7 +5815,7 @@ static umode_t hybrid_format_is_visible(struct
> kobject *kobj,
> 
>  static struct attribute_group hybrid_group_events_td  = {
>  	.name		= "events",
> -	.is_visible	= hybrid_events_is_visible,
> +	.is_visible	= hybrid_td_is_visible,
>  };
> 
>  static struct attribute_group hybrid_group_events_mem = {
> 
> 
> Thanks,
> Kan
>