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;
Received-SPF: Pass (protection.outlook.com: domain of amd.com designates
 165.204.84.17 as permitted sender) receiver=protection.outlook.com;
 client-ip=165.204.84.17; helo=SATLEXMB04.amd.com; pr=C
Subject: [PATCH v8 01/13] x86/cpufeatures: Add Slow Memory Bandwidth
 Allocation feature flag
From:   Babu Moger <babu.moger@amd.com>
To:     <corbet@lwn.net>, <reinette.chatre@intel.com>,
        <tglx@linutronix.de>, <mingo@redhat.com>, <bp@alien8.de>
CC:     <fenghua.yu@intel.com>, <dave.hansen@linux.intel.com>,
        <x86@kernel.org>, <hpa@zytor.com>, <paulmck@kernel.org>,
        <akpm@linux-foundation.org>, <quic_neeraju@quicinc.com>,
        <rdunlap@infradead.org>, <damien.lemoal@opensource.wdc.com>,
        <songmuchun@bytedance.com>, <peterz@infradead.org>,
        <jpoimboe@kernel.org>, <pbonzini@redhat.com>, <babu.moger@amd.com>,
        <chang.seok.bae@intel.com>, <pawan.kumar.gupta@linux.intel.com>,
        <jmattson@google.com>, <daniel.sneddon@linux.intel.com>,
        <sandipan.das@amd.com>, <tony.luck@intel.com>,
        <james.morse@arm.com>, <linux-doc@vger.kernel.org>,
        <linux-kernel@vger.kernel.org>, <bagasdotme@gmail.com>,
        <eranian@google.com>
Date:   Fri, 4 Nov 2022 14:59:56 -0500
Message-ID: <166759199632.3281208.5709918293488898752.stgit@bmoger-ubuntu>
In-Reply-To: <166759188265.3281208.11769277079826754455.stgit@bmoger-ubuntu>
References: <166759188265.3281208.11769277079826754455.stgit@bmoger-ubuntu>
User-Agent: StGit/1.1.dev103+g5369f4c
MIME-Version: 1.0
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: quoted-printable
X-MS-Exchange-CrossTenant-OriginalArrivalTime: 04 Nov 2022 19:59:58.7387
 (UTC)
X-MS-Exchange-CrossTenant-Network-Message-Id: 6c8905ac-a3e7-4dda-1911-08dabe9f26e8
X-MS-Exchange-CrossTenant-Id: 3dd8961f-e488-4e60-8e11-a82d994e183d
X-MS-Exchange-CrossTenant-OriginalAttributedTenantConnectingIp: TenantId=3dd8961f-e488-4e60-8e11-a82d994e183d;Ip=[165.204.84.17];Helo=[SATLEXMB04.amd.com]
X-MS-Exchange-CrossTenant-AuthSource: DM6NAM11FT050.eop-nam11.prod.protection.outlook.com
X-MS-Exchange-CrossTenant-AuthAs: Anonymous
X-MS-Exchange-CrossTenant-FromEntityHeader: HybridOnPrem
X-MS-Exchange-Transport-CrossTenantHeadersStamped: DS7PR12MB5720
X-Spam-Status: No, score=-2.1 required=5.0 tests=BAYES_00,DKIM_SIGNED,
        DKIM_VALID,DKIM_VALID_AU,DKIM_VALID_EF,RCVD_IN_DNSWL_NONE,
        RCVD_IN_MSPIKE_H2,SPF_HELO_PASS,SPF_PASS 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

Add the new AMD feature X86_FEATURE_SMBA. With this feature, the QOS=0A=
enforcement policies can be applied to external slow memory connected=0A=
to the host. QOS enforcement is accomplished by assigning a Class Of=0A=
Service (COS) to a processor and specifying allocations or limits for=0A=
that COS for each resource to be allocated.=0A=
=0A=
This feature is identified by the CPUID Function 8000_0020_EBX_x0.=0A=
=0A=
CPUID Fn8000_0020_EBX_x0 AMD Bandwidth Enforcement Feature Identifiers=0A=
(ECX=3D0)=0A=
=0A=
Bits    Field Name      Description=0A=
2       L3SBE           L3 external slow memory bandwidth enforcement=0A=
=0A=
Currently, CXL.memory is the only supported "slow" memory device. With=0A=
the support of SMBA feature, the hardware enables bandwidth allocation=0A=
on the slow memory devices. If there are multiple slow memory devices=0A=
in the system, then the throttling logic groups all the slow sources=0A=
together and applies the limit on them as a whole.=0A=
=0A=
The presence of the SMBA feature(with CXL.memory) is independent of=0A=
whether slow memory device is actually present in the system. If there=0A=
is no slow memory in the system, then setting a SMBA limit will have no=0A=
impact on the performance of the system.=0A=
=0A=
Presence of CXL memory can be identified by numactl command.=0A=
=0A=
$numactl -H=0A=
available: 2 nodes (0-1)=0A=
node 0 cpus: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16=0A=
node 0 size: 63678 MB node 0 free: 59542 MB=0A=
node 1 cpus:=0A=
node 1 size: 16122 MB=0A=
node 1 free: 15627 MB=0A=
node distances:=0A=
node   0   1=0A=
   0:  10  50=0A=
   1:  50  10=0A=
=0A=
CPU list for CXL memory will be empty. The cpu-cxl node distance is=0A=
greater than cpu-to-cpu distances. Node 1 has the CXL memory in this=0A=
case. CXL memory can also be identified using ACPI SRAT table and=0A=
memory maps.=0A=
=0A=
Feature description is available in the specification, "AMD64=0A=
Technology Platform Quality of Service Extensions, Revision: 1.03=0A=
Publication # 56375 Revision: 1.03 Issue Date: February 2022".=0A=
=0A=
Link: https://www.amd.com/en/support/tech-docs/amd64-technology-platform-qu=
ality-service-extensions=0A=
Link: https://bugzilla.kernel.org/show_bug.cgi?id=3D206537=0A=
Signed-off-by: Babu Moger <babu.moger@amd.com>=0A=
---=0A=
 arch/x86/include/asm/cpufeatures.h |    1 +=0A=
 arch/x86/kernel/cpu/scattered.c    |    1 +=0A=
 2 files changed, 2 insertions(+)=0A=
=0A=
diff --git a/arch/x86/include/asm/cpufeatures.h b/arch/x86/include/asm/cpuf=
eatures.h=0A=
index aefd0816a333..d68b4c9c181d 100644=0A=
--- a/arch/x86/include/asm/cpufeatures.h=0A=
+++ b/arch/x86/include/asm/cpufeatures.h=0A=
@@ -305,6 +305,7 @@=0A=
 #define X86_FEATURE_USE_IBPB_FW		(11*32+16) /* "" Use IBPB during runtime =
firmware calls */=0A=
 #define X86_FEATURE_RSB_VMEXIT_LITE	(11*32+17) /* "" Fill RSB on VM exit w=
hen EIBRS is enabled */=0A=
 #define X86_FEATURE_CALL_DEPTH		(11*32+18) /* "" Call depth tracking for R=
SB stuffing */=0A=
+#define X86_FEATURE_SMBA		(11*32+19) /* Slow Memory Bandwidth Allocation *=
/=0A=
 =0A=
 /* Intel-defined CPU features, CPUID level 0x00000007:1 (EAX), word 12 */=
=0A=
 #define X86_FEATURE_AVX_VNNI		(12*32+ 4) /* AVX VNNI instructions */=0A=
diff --git a/arch/x86/kernel/cpu/scattered.c b/arch/x86/kernel/cpu/scattere=
d.c=0A=
index fc01f81f6e2a..5a5f17ed69a2 100644=0A=
--- a/arch/x86/kernel/cpu/scattered.c=0A=
+++ b/arch/x86/kernel/cpu/scattered.c=0A=
@@ -44,6 +44,7 @@ static const struct cpuid_bit cpuid_bits[] =3D {=0A=
 	{ X86_FEATURE_CPB,		CPUID_EDX,  9, 0x80000007, 0 },=0A=
 	{ X86_FEATURE_PROC_FEEDBACK,    CPUID_EDX, 11, 0x80000007, 0 },=0A=
 	{ X86_FEATURE_MBA,		CPUID_EBX,  6, 0x80000008, 0 },=0A=
+	{ X86_FEATURE_SMBA,		CPUID_EBX,  2, 0x80000020, 0 },=0A=
 	{ X86_FEATURE_PERFMON_V2,	CPUID_EAX,  0, 0x80000022, 0 },=0A=
 	{ X86_FEATURE_AMD_LBR_V2,	CPUID_EAX,  1, 0x80000022, 0 },=0A=
 	{ 0, 0, 0, 0, 0 }=0A=
=0A=