Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.36 as permitted sender) client-ip=23.128.96.36;
Date:   Sun, 29 Oct 2023 11:35:26 +0100
From:   Borislav Petkov <bp@alien8.de>
To:     Andrew Cooper <andrew.cooper3@citrix.com>
Cc:     Peter Zijlstra <peterz@infradead.org>, X86 ML <x86@kernel.org>,
        Kishon VijayAbraham <Kishon.VijayAbraham@amd.com>,
        LKML <linux-kernel@vger.kernel.org>
Subject: [PATCH -v3] x86/barrier: Do not serialize MSR accesses on AMD
Message-ID: <20231029103526.GAZT41bj8qMjt9+7Ql@fat_crate.local>
References: <20230704074631.GAZKPOV/9BfqP0aU8v@fat_crate.local>
 <20230704090132.GP4253@hirez.programming.kicks-ass.net>
 <20230704092222.GBZKPkzgdM8rbPe7zA@fat_crate.local>
 <20231027153327.GKZTvYR3qslaTUjtCT@fat_crate.local>
 <20231027153458.GMZTvYou1tlK6HD8/Y@fat_crate.local>
 <20231027185641.GE26550@noisy.programming.kicks-ass.net>
 <20231027191633.GRZTwMkaiW1nyvnzzO@fat_crate.local>
 <20231027192907.GSZTwPg8v7NF6+Zn0w@fat_crate.local>
 <3c56e807-945c-4996-9ac1-3205a23248ab@citrix.com>
 <20231027202328.GTZTwcQIh8wFyZUtQd@fat_crate.local>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
In-Reply-To: <20231027202328.GTZTwcQIh8wFyZUtQd@fat_crate.local>
Precedence: bulk

On Fri, Oct 27, 2023 at 10:23:28PM +0200, Borislav Petkov wrote:
> So the feature bit should be named something more specific:
> 
> 	X86_FEATURE_APIC_TSC_MSRS_NEED_FENCING
> 
> or so.

Plain and simple:

---
From: "Borislav Petkov (AMD)" <bp@alien8.de>
Date: Fri, 27 Oct 2023 14:24:16 +0200

AMD does not have the requirement for a synchronization barrier when
acccessing a certain group of MSRs. Do not incur that unnecessary
penalty there.

While at it, move to processor.h to avoid include hell. Untangling that
file properly is a matter for another day.

Some notes on the performance aspect of why this is relevant, courtesy
of Kishon VijayAbraham <Kishon.VijayAbraham@amd.com>:

On a AMD Zen4 system with 96 cores, a modified ipi-bench[1] on a VM
shows x2AVIC IPI rate is 3% to 4% lower than AVIC IPI rate. The
ipi-bench is modified so that the IPIs are sent between two vCPUs in the
same CCX. This also requires to pin the vCPU to a physical core to
prevent any latencies. This simulates the use case of pinning vCPUs to
the thread of a single CCX to avoid interrupt IPI latency.

In order to avoid run-to-run variance (for both x2AVIC and AVIC), the
below configurations are done:

  1) Disable Power States in BIOS (to prevent the system from going to
     lower power state)

  2) Run the system at fixed frequency 2500MHz (to prevent the system
     from increasing the frequency when the load is more)

With the above configuration:

*) Performance measured using ipi-bench for AVIC:
  Average Latency:  1124.98ns [Time to send IPI from one vCPU to another vCPU]

  Cumulative throughput: 42.6759M/s [Total number of IPIs sent in a second from
  				     48 vCPUs simultaneously]

*) Performance measured using ipi-bench for x2AVIC:
  Average Latency:  1172.42ns [Time to send IPI from one vCPU to another vCPU]

  Cumulative throughput: 40.9432M/s [Total number of IPIs sent in a second from
  				     48 vCPUs simultaneously]

From above, x2AVIC latency is ~4% more than AVIC. However, the expectation is
x2AVIC performance to be better or equivalent to AVIC. Upon analyzing
the perf captures, it is observed significant time is spent in
weak_wrmsr_fence() invoked by x2apic_send_IPI().

With the fix to skip weak_wrmsr_fence()

*) Performance measured using ipi-bench for x2AVIC:
  Average Latency:  1117.44ns [Time to send IPI from one vCPU to another vCPU]

  Cumulative throughput: 42.9608M/s [Total number of IPIs sent in a second from
  				     48 vCPUs simultaneously]

Comparing the performance of x2AVIC with and without the fix, it can be seen
the performance improves by ~4%.

Performance captured using an unmodified ipi-bench using the 'mesh-ipi' option
with and without weak_wrmsr_fence() on a Zen4 system also showed significant
performance improvement without weak_wrmsr_fence(). The 'mesh-ipi' option ignores
CCX or CCD and just picks random vCPU.

  Average throughput (10 iterations) with weak_wrmsr_fence(),
        Cumulative throughput: 4933374 IPI/s

  Average throughput (10 iterations) without weak_wrmsr_fence(),
        Cumulative throughput: 6355156 IPI/s

[1] https://github.com/bytedance/kvm-utils/tree/master/microbenchmark/ipi-bench

Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
---
 arch/x86/include/asm/barrier.h     | 18 ------------------
 arch/x86/include/asm/cpufeatures.h |  2 +-
 arch/x86/include/asm/processor.h   | 18 ++++++++++++++++++
 arch/x86/kernel/cpu/amd.c          |  3 +++
 arch/x86/kernel/cpu/common.c       |  7 +++++++
 arch/x86/kernel/cpu/hygon.c        |  3 +++
 6 files changed, 32 insertions(+), 19 deletions(-)

diff --git a/arch/x86/include/asm/barrier.h b/arch/x86/include/asm/barrier.h
index 35389b2af88e..0216f63a366b 100644
--- a/arch/x86/include/asm/barrier.h
+++ b/arch/x86/include/asm/barrier.h
@@ -81,22 +81,4 @@ do {									\
 
 #include <asm-generic/barrier.h>
 
-/*
- * Make previous memory operations globally visible before
- * a WRMSR.
- *
- * MFENCE makes writes visible, but only affects load/store
- * instructions.  WRMSR is unfortunately not a load/store
- * instruction and is unaffected by MFENCE.  The LFENCE ensures
- * that the WRMSR is not reordered.
- *
- * Most WRMSRs are full serializing instructions themselves and
- * do not require this barrier.  This is only required for the
- * IA32_TSC_DEADLINE and X2APIC MSRs.
- */
-static inline void weak_wrmsr_fence(void)
-{
-	asm volatile("mfence; lfence" : : : "memory");
-}
-
 #endif /* _ASM_X86_BARRIER_H */
diff --git a/arch/x86/include/asm/cpufeatures.h b/arch/x86/include/asm/cpufeatures.h
index 58cb9495e40f..0091f1008314 100644
--- a/arch/x86/include/asm/cpufeatures.h
+++ b/arch/x86/include/asm/cpufeatures.h
@@ -308,10 +308,10 @@
 #define X86_FEATURE_SMBA		(11*32+21) /* "" Slow Memory Bandwidth Allocation */
 #define X86_FEATURE_BMEC		(11*32+22) /* "" Bandwidth Monitoring Event Configuration */
 #define X86_FEATURE_USER_SHSTK		(11*32+23) /* Shadow stack support for user mode applications */
-
 #define X86_FEATURE_SRSO		(11*32+24) /* "" AMD BTB untrain RETs */
 #define X86_FEATURE_SRSO_ALIAS		(11*32+25) /* "" AMD BTB untrain RETs through aliasing */
 #define X86_FEATURE_IBPB_ON_VMEXIT	(11*32+26) /* "" Issue an IBPB only on VMEXIT */
+#define X86_FEATURE_APIC_MSRS_FENCE	(11*32+27) /* "" IA32_TSC_DEADLINE and X2APIC MSRs need fencing */
 
 /* Intel-defined CPU features, CPUID level 0x00000007:1 (EAX), word 12 */
 #define X86_FEATURE_AVX_VNNI		(12*32+ 4) /* AVX VNNI instructions */
diff --git a/arch/x86/include/asm/processor.h b/arch/x86/include/asm/processor.h
index 4b130d894cb6..061aa86b4662 100644
--- a/arch/x86/include/asm/processor.h
+++ b/arch/x86/include/asm/processor.h
@@ -752,4 +752,22 @@ enum mds_mitigations {
 
 extern bool gds_ucode_mitigated(void);
 
+/*
+ * Make previous memory operations globally visible before
+ * a WRMSR.
+ *
+ * MFENCE makes writes visible, but only affects load/store
+ * instructions.  WRMSR is unfortunately not a load/store
+ * instruction and is unaffected by MFENCE.  The LFENCE ensures
+ * that the WRMSR is not reordered.
+ *
+ * Most WRMSRs are full serializing instructions themselves and
+ * do not require this barrier.  This is only required for the
+ * IA32_TSC_DEADLINE and X2APIC MSRs.
+ */
+static inline void weak_wrmsr_fence(void)
+{
+	alternative("mfence; lfence", "", ALT_NOT(X86_FEATURE_APIC_MSRS_FENCE));
+}
+
 #endif /* _ASM_X86_PROCESSOR_H */
diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c
index a7eab05e5f29..841e21213668 100644
--- a/arch/x86/kernel/cpu/amd.c
+++ b/arch/x86/kernel/cpu/amd.c
@@ -1162,6 +1162,9 @@ static void init_amd(struct cpuinfo_x86 *c)
 	if (!cpu_has(c, X86_FEATURE_HYPERVISOR) &&
 	     cpu_has_amd_erratum(c, amd_erratum_1485))
 		msr_set_bit(MSR_ZEN4_BP_CFG, MSR_ZEN4_BP_CFG_SHARED_BTB_FIX_BIT);
+
+	/* AMD CPUs don't need fencing after x2APIC/TSC_DEADLINE MSR writes. */
+	clear_cpu_cap(c, X86_FEATURE_APIC_MSRS_FENCE);
 }
 
 #ifdef CONFIG_X86_32
diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c
index 9058da9ae011..4d4b87c6885d 100644
--- a/arch/x86/kernel/cpu/common.c
+++ b/arch/x86/kernel/cpu/common.c
@@ -1856,6 +1856,13 @@ static void identify_cpu(struct cpuinfo_x86 *c)
 	c->topo.apicid = apic->phys_pkg_id(c->topo.initial_apicid, 0);
 #endif
 
+
+	/*
+	 * Set default APIC and TSC_DEADLINE MSR fencing flag. AMD and
+	 * Hygon will clear it in ->c_init() below.
+	 */
+	set_cpu_cap(c, X86_FEATURE_APIC_MSRS_FENCE);
+
 	/*
 	 * Vendor-specific initialization.  In this section we
 	 * canonicalize the feature flags, meaning if there are
diff --git a/arch/x86/kernel/cpu/hygon.c b/arch/x86/kernel/cpu/hygon.c
index 6f247d66758d..f0cd95502faa 100644
--- a/arch/x86/kernel/cpu/hygon.c
+++ b/arch/x86/kernel/cpu/hygon.c
@@ -354,6 +354,9 @@ static void init_hygon(struct cpuinfo_x86 *c)
 		set_cpu_bug(c, X86_BUG_SYSRET_SS_ATTRS);
 
 	check_null_seg_clears_base(c);
+
+	/* Hygon CPUs don't need fencing after x2APIC/TSC_DEADLINE MSR writes. */
+	clear_cpu_cap(c, X86_FEATURE_APIC_MSRS_FENCE);
 }
 
 static void cpu_detect_tlb_hygon(struct cpuinfo_x86 *c)
-- 
2.42.0.rc0.25.ga82fb66fed25

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette