LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
* [PATCH 1/4] KVM: x86: Introduce .pcpu_is_idle() stub infrastructure
@ 2021-08-31  1:59 Tianqiang Xu
  2021-08-31  1:59 ` [PATCH 2/4] Scheduler changes Tianqiang Xu
                   ` (4 more replies)
  0 siblings, 5 replies; 10+ messages in thread
From: Tianqiang Xu @ 2021-08-31  1:59 UTC (permalink / raw)
  To: x86
  Cc: pbonzini, seanjc, vkuznets, wanpengli, jmattson, joro, tglx,
	mingo, bp, kvm, hpa, jarkko, dave.hansen, linux-kernel,
	linux-sgx, Tianqiang Xu

This patch series aims to fix performance issue caused by current
para-virtualized scheduling design.

The current para-virtualized scheduling design uses 'preempted' field of
kvm_steal_time to avoid scheduling task on the preempted vCPU.
However, when the pCPU where the preempted vCPU most recently run is idle,
it will result in low cpu utilization, and consequently poor performance.

The new field: 'is_idle' of kvm_steal_time can precisely reveal
the status of pCPU where preempted vCPU most recently run, and
then improve cpu utilization.

pcpu_is_idle() is used to get the value of 'is_idle' of kvm_steal_time.

Experiments on a VM with 16 vCPUs show that the patch can reduce around
50% to 80% execution time for most PARSEC benchmarks. 
This also holds true for a VM with 112 vCPUs.

Experiments on 2 VMs with 112 vCPUs show that the patch can reduce around
20% to 80% execution time for most PARSEC benchmarks. 

Test environment:
-- PowerEdge R740
-- 56C-112T CPU Intel(R) Xeon(R) Gold 6238R CPU
-- Host 190G DRAM
-- QEMU 5.0.0
-- PARSEC 3.0 Native Inputs
-- Host is idle during the test
-- Host and Guest kernel are both kernel-5.14.0

Results:
1. 1 VM, 16 VCPU, 16 THREAD.
   Host Topology: sockets=2 cores=28 threads=2
   VM Topology:   sockets=1 cores=16 threads=1
   Command: <path to parsec>/bin/parsecmgmt -a run -p <benchmark> -i native -n 16
   Statistics below are the real time of running each benchmark.(lower is better)

			before patch    after patch	improvements
bodytrack		52.866s		22.619s		57.21%
fluidanimate		84.009s		38.148s		54.59%
streamcluster		270.17s		42.726s		84.19%
splash2x.ocean_cp	31.932s		9.539s		70.13%
splash2x.ocean_ncp	36.063s		14.189s		60.65%
splash2x.volrend	134.587s	21.79s		83.81%

2. 1VM, 112 VCPU. Some benchmarks require the number of threads to be the power of 2,
so we run them with 64 threads and 128 threads.
   Host Topology: sockets=2 cores=28 threads=2
   VM Topology:   sockets=1 cores=112 threads=1
   Command: <path to parsec>/bin/parsecmgmt -a run -p <benchmark> -i native -n <64,112,128>
   Statistics below are the real time of running each benchmark.(lower is better)

                        		before patch    after patch     improvements
fluidanimate(64 thread)			124.235s	27.924s		77.52%
fluidanimate(128 thread)		169.127s	64.541s		61.84%
streamcluster(112 thread)		861.879s	496.66s		42.37%
splash2x.ocean_cp(64 thread)		46.415s		18.527s		60.08%
splash2x.ocean_cp(128 thread)		53.647s		28.929s		46.08%
splash2x.ocean_ncp(64 thread)		47.613s		19.576s		58.89%
splash2x.ocean_ncp(128 thread)		54.94s		29.199s		46.85%
splash2x.volrend(112 thread)		801.384s	144.824s	81.93%

3. 2VM, each VM: 112 VCPU. Some benchmarks require the number of threads to
be the power of 2, so we run them with 64 threads and 128 threads.
   Host Topology: sockets=2 cores=28 threads=2
   VM Topology:   sockets=1 cores=112 threads=1
   Command: <path to parsec>/bin/parsecmgmt -a run -p <benchmark> -i native -n <64,112,128>
   Statistics below are the average real time of running each benchmark in 2 VMs.(lower is better)

                                        before patch    after patch	improvements
fluidanimate(64 thread)			135.2125s	49.827s		63.15%
fluidanimate(128 thread)		178.309s	86.964s		51.23%
splash2x.ocean_cp(64 thread)		47.4505s	20.314s		57.19%
splash2x.ocean_cp(128 thread)		55.5645s	30.6515s	44.84%
splash2x.ocean_ncp(64 thread)		49.9775s	23.489s		53.00%
splash2x.ocean_ncp(128 thread)		56.847s		28.545s		49.79%
splash2x.volrend(112 thread)		838.939s	239.632s	71.44%

For space limit, we list representative statistics here.

--
Authors: Tianqiang Xu, Dingji Li, Zeyu Mi
	 Shanghai Jiao Tong University

Signed-off-by: Tianqiang Xu <skyele@sjtu.edu.cn>
---
 arch/x86/hyperv/hv_spinlock.c         |  7 +++++++
 arch/x86/include/asm/cpufeatures.h    |  1 +
 arch/x86/include/asm/kvm_host.h       |  1 +
 arch/x86/include/asm/paravirt.h       |  8 ++++++++
 arch/x86/include/asm/paravirt_types.h |  1 +
 arch/x86/include/asm/qspinlock.h      |  7 +++++++
 arch/x86/include/uapi/asm/kvm_para.h  |  4 +++-
 arch/x86/kernel/asm-offsets_64.c      |  1 +
 arch/x86/kernel/kvm.c                 | 21 +++++++++++++++++++++
 arch/x86/kernel/paravirt-spinlocks.c  | 15 +++++++++++++++
 arch/x86/kernel/paravirt.c            |  2 ++
 11 files changed, 67 insertions(+), 1 deletion(-)

diff --git a/arch/x86/hyperv/hv_spinlock.c b/arch/x86/hyperv/hv_spinlock.c
index 91cfe698bde0..9782e188904a 100644
--- a/arch/x86/hyperv/hv_spinlock.c
+++ b/arch/x86/hyperv/hv_spinlock.c
@@ -60,6 +60,12 @@ static void hv_qlock_wait(u8 *byte, u8 val)
 /*
  * Hyper-V does not support this so far.
  */
+__visible bool hv_pcpu_is_idle(int vcpu)
+{
+	return false;
+}
+PV_CALLEE_SAVE_REGS_THUNK(hv_pcpu_is_idle);
+
 __visible bool hv_vcpu_is_preempted(int vcpu)
 {
 	return false;
@@ -82,6 +88,7 @@ void __init hv_init_spinlocks(void)
 	pv_ops.lock.wait = hv_qlock_wait;
 	pv_ops.lock.kick = hv_qlock_kick;
 	pv_ops.lock.vcpu_is_preempted = PV_CALLEE_SAVE(hv_vcpu_is_preempted);
+	pv_ops.lock.pcpu_is_idle = PV_CALLEE_SAVE(hv_pcpu_is_idle);
 }
 
 static __init int hv_parse_nopvspin(char *arg)
diff --git a/arch/x86/include/asm/cpufeatures.h b/arch/x86/include/asm/cpufeatures.h
index d0ce5cfd3ac1..efda9b9a4cad 100644
--- a/arch/x86/include/asm/cpufeatures.h
+++ b/arch/x86/include/asm/cpufeatures.h
@@ -238,6 +238,7 @@
 #define X86_FEATURE_VMW_VMMCALL		( 8*32+19) /* "" VMware prefers VMMCALL hypercall instruction */
 #define X86_FEATURE_PVUNLOCK		( 8*32+20) /* "" PV unlock function */
 #define X86_FEATURE_VCPUPREEMPT		( 8*32+21) /* "" PV vcpu_is_preempted function */
+#define X86_FEATURE_PCPUISIDLE          ( 8*32+22) /* "" PV pcpu_is_idle function */
 
 /* Intel-defined CPU features, CPUID level 0x00000007:0 (EBX), word 9 */
 #define X86_FEATURE_FSGSBASE		( 9*32+ 0) /* RDFSBASE, WRFSBASE, RDGSBASE, WRGSBASE instructions*/
diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
index af6ce8d4c86a..705c55be0eed 100644
--- a/arch/x86/include/asm/kvm_host.h
+++ b/arch/x86/include/asm/kvm_host.h
@@ -742,6 +742,7 @@ struct kvm_vcpu_arch {
 
 	struct {
 		u8 preempted;
+		u8 is_idle;
 		u64 msr_val;
 		u64 last_steal;
 		struct gfn_to_pfn_cache cache;
diff --git a/arch/x86/include/asm/paravirt.h b/arch/x86/include/asm/paravirt.h
index da3a1ac82be5..f34dec6eb515 100644
--- a/arch/x86/include/asm/paravirt.h
+++ b/arch/x86/include/asm/paravirt.h
@@ -609,8 +609,16 @@ static __always_inline bool pv_vcpu_is_preempted(long cpu)
 				ALT_NOT(X86_FEATURE_VCPUPREEMPT));
 }
 
+static __always_inline bool pv_pcpu_is_idle(long cpu)
+{
+	return PVOP_ALT_CALLEE1(bool, lock.pcpu_is_idle, cpu,
+				"xor %%" _ASM_AX ", %%" _ASM_AX ";",
+				ALT_NOT(X86_FEATURE_PCPUISIDLE));
+}
+
 void __raw_callee_save___native_queued_spin_unlock(struct qspinlock *lock);
 bool __raw_callee_save___native_vcpu_is_preempted(long cpu);
+bool __raw_callee_save___native_pcpu_is_idle(long cpu);
 
 #endif /* SMP && PARAVIRT_SPINLOCKS */
 
diff --git a/arch/x86/include/asm/paravirt_types.h b/arch/x86/include/asm/paravirt_types.h
index d9d6b0203ec4..7d9b5906580c 100644
--- a/arch/x86/include/asm/paravirt_types.h
+++ b/arch/x86/include/asm/paravirt_types.h
@@ -257,6 +257,7 @@ struct pv_lock_ops {
 	void (*kick)(int cpu);
 
 	struct paravirt_callee_save vcpu_is_preempted;
+	struct paravirt_callee_save pcpu_is_idle;
 } __no_randomize_layout;
 
 /* This contains all the paravirt structures: we get a convenient
diff --git a/arch/x86/include/asm/qspinlock.h b/arch/x86/include/asm/qspinlock.h
index d86ab942219c..c32f2eb6186c 100644
--- a/arch/x86/include/asm/qspinlock.h
+++ b/arch/x86/include/asm/qspinlock.h
@@ -63,6 +63,13 @@ static inline bool vcpu_is_preempted(long cpu)
 }
 #endif
 
+#define pcpu_is_idle pcpu_is_idle
+static inline bool pcpu_is_idle(long cpu)
+{
+	return pv_pcpu_is_idle(cpu);
+}
+#endif
+
 #ifdef CONFIG_PARAVIRT
 /*
  * virt_spin_lock_key - enables (by default) the virt_spin_lock() hijack.
diff --git a/arch/x86/include/uapi/asm/kvm_para.h b/arch/x86/include/uapi/asm/kvm_para.h
index 5146bbab84d4..82940e4b76d5 100644
--- a/arch/x86/include/uapi/asm/kvm_para.h
+++ b/arch/x86/include/uapi/asm/kvm_para.h
@@ -63,12 +63,14 @@ struct kvm_steal_time {
 	__u32 version;
 	__u32 flags;
 	__u8  preempted;
-	__u8  u8_pad[3];
+	__u8  is_idle;
+	__u8  u8_pad[2];
 	__u32 pad[11];
 };
 
 #define KVM_VCPU_PREEMPTED          (1 << 0)
 #define KVM_VCPU_FLUSH_TLB          (1 << 1)
+#define KVM_PCPU_IS_IDLE            (1 << 0)
 
 #define KVM_CLOCK_PAIRING_WALLCLOCK 0
 struct kvm_clock_pairing {
diff --git a/arch/x86/kernel/asm-offsets_64.c b/arch/x86/kernel/asm-offsets_64.c
index b14533af7676..b587bbe44470 100644
--- a/arch/x86/kernel/asm-offsets_64.c
+++ b/arch/x86/kernel/asm-offsets_64.c
@@ -22,6 +22,7 @@ int main(void)
 
 #if defined(CONFIG_KVM_GUEST) && defined(CONFIG_PARAVIRT_SPINLOCKS)
 	OFFSET(KVM_STEAL_TIME_preempted, kvm_steal_time, preempted);
+	OFFSET(KVM_STEAL_TIME_is_idle, kvm_steal_time, is_idle);
 	BLANK();
 #endif
 
diff --git a/arch/x86/kernel/kvm.c b/arch/x86/kernel/kvm.c
index a26643dc6bd6..b167589fffbc 100644
--- a/arch/x86/kernel/kvm.c
+++ b/arch/x86/kernel/kvm.c
@@ -900,11 +900,19 @@ __visible bool __kvm_vcpu_is_preempted(long cpu)
 }
 PV_CALLEE_SAVE_REGS_THUNK(__kvm_vcpu_is_preempted);
 
+__visible bool __kvm_pcpu_is_idle(long cpu)
+{
+	struct kvm_steal_time *src = &per_cpu(steal_time, cpu);
+
+	return !!(src->preempted & KVM_PCPU_IS_IDLE);
+}
+PV_CALLEE_SAVE_REGS_THUNK(__kvm_pcpu_is_idle);
 #else
 
 #include <asm/asm-offsets.h>
 
 extern bool __raw_callee_save___kvm_vcpu_is_preempted(long);
+extern bool __raw_callee_save___kvm_pcpu_is_idle(long);
 
 /*
  * Hand-optimize version for x86-64 to avoid 8 64-bit register saving and
@@ -922,6 +930,17 @@ asm(
 ".size __raw_callee_save___kvm_vcpu_is_preempted, .-__raw_callee_save___kvm_vcpu_is_preempted;"
 ".popsection");
 
+asm(
+".pushsection .text;"
+".global __raw_callee_save___kvm_pcpu_is_idle;"
+".type __raw_callee_save___kvm_pcpu_is_idle, @function;"
+"__raw_callee_save___kvm_pcpu_is_idle:"
+"movq	__per_cpu_offset(,%rdi,8), %rax;"
+"cmpb	$0, " __stringify(KVM_STEAL_TIME_is_idle) "+steal_time(%rax);"
+"setne	%al;"
+"ret;"
+".size __raw_callee_save___kvm_pcpu_is_idle, .-__raw_callee_save___kvm_pcpu_is_idle;"
+".popsection");
 #endif
 
 /*
@@ -970,6 +989,8 @@ void __init kvm_spinlock_init(void)
 	if (kvm_para_has_feature(KVM_FEATURE_STEAL_TIME)) {
 		pv_ops.lock.vcpu_is_preempted =
 			PV_CALLEE_SAVE(__kvm_vcpu_is_preempted);
+		pv_ops.lock.pcpu_is_idle =
+			PV_CALLEE_SAVE(__kvm_pcpu_is_idle);
 	}
 	/*
 	 * When PV spinlock is enabled which is preferred over
diff --git a/arch/x86/kernel/paravirt-spinlocks.c b/arch/x86/kernel/paravirt-spinlocks.c
index 9e1ea99ad9df..d7f6a461d0a5 100644
--- a/arch/x86/kernel/paravirt-spinlocks.c
+++ b/arch/x86/kernel/paravirt-spinlocks.c
@@ -27,12 +27,24 @@ __visible bool __native_vcpu_is_preempted(long cpu)
 }
 PV_CALLEE_SAVE_REGS_THUNK(__native_vcpu_is_preempted);
 
+__visible bool __native_pcpu_is_idle(long cpu)
+{
+	return false;
+}
+PV_CALLEE_SAVE_REGS_THUNK(__native_pcpu_is_idle);
+
 bool pv_is_native_vcpu_is_preempted(void)
 {
 	return pv_ops.lock.vcpu_is_preempted.func ==
 		__raw_callee_save___native_vcpu_is_preempted;
 }
 
+bool pv_is_native_pcpu_is_idle(void)
+{
+	return pv_ops.lock.pcpu_is_idle.func ==
+		__raw_callee_save___native_pcpu_is_idle;
+}
+
 void __init paravirt_set_cap(void)
 {
 	if (!pv_is_native_spin_unlock())
@@ -40,4 +52,7 @@ void __init paravirt_set_cap(void)
 
 	if (!pv_is_native_vcpu_is_preempted())
 		setup_force_cpu_cap(X86_FEATURE_VCPUPREEMPT);
+
+	if (!pv_is_native_pcpu_is_idle())
+		setup_force_cpu_cap(X86_FEATURE_PCPUISIDLE);
 }
diff --git a/arch/x86/kernel/paravirt.c b/arch/x86/kernel/paravirt.c
index 04cafc057bed..543d89856161 100644
--- a/arch/x86/kernel/paravirt.c
+++ b/arch/x86/kernel/paravirt.c
@@ -366,6 +366,8 @@ struct paravirt_patch_template pv_ops = {
 	.lock.kick			= paravirt_nop,
 	.lock.vcpu_is_preempted		=
 				PV_CALLEE_SAVE(__native_vcpu_is_preempted),
+	.lock.pcpu_is_idle		=
+				PV_CALLEE_SAVE(__native_pcpu_is_idle),
 #endif /* SMP */
 #endif
 };
-- 
2.26.0


^ permalink raw reply	[flat|nested] 10+ messages in thread

* [PATCH 2/4] Scheduler changes
  2021-08-31  1:59 [PATCH 1/4] KVM: x86: Introduce .pcpu_is_idle() stub infrastructure Tianqiang Xu
@ 2021-08-31  1:59 ` Tianqiang Xu
  2021-08-31  7:14   ` Peter Zijlstra
  2021-08-31  1:59 ` [PATCH 3/4] KVM host implementation Tianqiang Xu
                   ` (3 subsequent siblings)
  4 siblings, 1 reply; 10+ messages in thread
From: Tianqiang Xu @ 2021-08-31  1:59 UTC (permalink / raw)
  To: x86
  Cc: pbonzini, seanjc, vkuznets, wanpengli, jmattson, joro, tglx,
	mingo, bp, kvm, hpa, jarkko, dave.hansen, linux-kernel,
	linux-sgx, Tianqiang Xu

available_idle_cpu_sched() invokes pcpu_is_idle() to detect whether
a pCPU where the preempted vCPU most recently run is idle or not, which
is used by the guest OS.

get_cpu_nr_running() gets the run queue of current cpu, which is used
by the host KVM to know whether a pCPU where the preempted vCPU mostly
recently run is idle or not.

--
Authors: Tianqiang Xu, Dingji Li, Zeyu Mi
	 Shanghai Jiao Tong University

Signed-off-by: Tianqiang Xu <skyele@sjtu.edu.cn>

---
 include/linux/sched.h |  1 +
 kernel/sched/core.c   | 17 +++++++++++++++++
 2 files changed, 18 insertions(+)

diff --git a/include/linux/sched.h b/include/linux/sched.h
index ec8d07d88641..dd4c41d2d8d3 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -1736,6 +1736,7 @@ extern int can_nice(const struct task_struct *p, const int nice);
 extern int task_curr(const struct task_struct *p);
 extern int idle_cpu(int cpu);
 extern int available_idle_cpu(int cpu);
+extern int available_idle_cpu_sched(int cpu);
 extern int sched_setscheduler(struct task_struct *, int, const struct sched_param *);
 extern int sched_setscheduler_nocheck(struct task_struct *, int, const struct sched_param *);
 extern void sched_set_fifo(struct task_struct *p);
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index f3b27c6c5153..c777dbcbeb9c 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -6752,6 +6752,17 @@ int available_idle_cpu(int cpu)
 	return 1;
 }
 
+int available_idle_cpu_sched(int cpu)
+{
+	if (!idle_cpu(cpu))
+		return 0;
+
+	if (!pcpu_is_idle(cpu))
+		return 0;
+
+	return 1;
+}
+
 /**
  * idle_task - return the idle task for a given CPU.
  * @cpu: the processor in question.
@@ -10504,3 +10515,9 @@ void call_trace_sched_update_nr_running(struct rq *rq, int count)
 {
         trace_sched_update_nr_running_tp(rq, count);
 }
+
+int get_cpu_nr_running(int cpu)
+{
+	return cpu_rq(cpu)->nr_running;
+}
+EXPORT_SYMBOL_GPL(get_cpu_nr_running);
-- 
2.26.0


^ permalink raw reply	[flat|nested] 10+ messages in thread

* [PATCH 3/4] KVM host implementation
  2021-08-31  1:59 [PATCH 1/4] KVM: x86: Introduce .pcpu_is_idle() stub infrastructure Tianqiang Xu
  2021-08-31  1:59 ` [PATCH 2/4] Scheduler changes Tianqiang Xu
@ 2021-08-31  1:59 ` Tianqiang Xu
  2021-08-31  7:16   ` Peter Zijlstra
  2021-08-31  1:59 ` [PATCH 4/4] KVM guest implementation Tianqiang Xu
                   ` (2 subsequent siblings)
  4 siblings, 1 reply; 10+ messages in thread
From: Tianqiang Xu @ 2021-08-31  1:59 UTC (permalink / raw)
  To: x86
  Cc: pbonzini, seanjc, vkuznets, wanpengli, jmattson, joro, tglx,
	mingo, bp, kvm, hpa, jarkko, dave.hansen, linux-kernel,
	linux-sgx, Tianqiang Xu

Host OS sets 'is_idle' field of kvm_steal_time to 1 if
cpu_rq(this_cpu)->nr_running is 1 before a vCPU being scheduled out.
On this condition, there is no other task on this pCPU to run.
Thus, is_idle == 1 means the pCPU where the preempted vCPU most
recently run is idle.

Host OS invokes get_cpu_nr_running() to get the value of
cpu_rq(this_cpu)->nr_running.

--
Authors: Tianqiang Xu, Dingji Li, Zeyu Mi
	 Shanghai Jiao Tong University

Signed-off-by: Tianqiang Xu <skyele@sjtu.edu.cn>
---
 arch/x86/include/asm/qspinlock.h |  1 -
 arch/x86/kvm/x86.c               | 88 +++++++++++++++++++++++++++++++-
 2 files changed, 87 insertions(+), 2 deletions(-)

diff --git a/arch/x86/include/asm/qspinlock.h b/arch/x86/include/asm/qspinlock.h
index c32f2eb6186c..1832dd8308ca 100644
--- a/arch/x86/include/asm/qspinlock.h
+++ b/arch/x86/include/asm/qspinlock.h
@@ -61,7 +61,6 @@ static inline bool vcpu_is_preempted(long cpu)
 {
 	return pv_vcpu_is_preempted(cpu);
 }
-#endif
 
 #define pcpu_is_idle pcpu_is_idle
 static inline bool pcpu_is_idle(long cpu)
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index e5d5c5ed7dd4..1fb1ab3d6fca 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -3181,6 +3181,72 @@ static void kvm_vcpu_flush_tlb_guest(struct kvm_vcpu *vcpu)
 	static_call(kvm_x86_tlb_flush_guest)(vcpu);
 }
 
+static void kvm_steal_time_set_is_idle(struct kvm_vcpu *vcpu)
+{
+	struct kvm_host_map map;
+	struct kvm_steal_time *st;
+
+	if (!(vcpu->arch.st.msr_val & KVM_MSR_ENABLED))
+		return;
+
+	if (vcpu->arch.st.is_idle)
+		return;
+
+	if (kvm_map_gfn(vcpu, vcpu->arch.st.msr_val >> PAGE_SHIFT, &map,
+			&vcpu->arch.st.cache, true))
+		return;
+
+	st = map.hva +
+		offset_in_page(vcpu->arch.st.msr_val & KVM_STEAL_VALID_BITS);
+
+	st->is_idle = vcpu->arch.st.is_idle = KVM_PCPU_IS_IDLE;
+
+	kvm_unmap_gfn(vcpu, &map, &vcpu->arch.st.cache, true, true);
+}
+
+static void kvm_steal_time_clear_is_idle(struct kvm_vcpu *vcpu)
+{
+	struct kvm_host_map map;
+	struct kvm_steal_time *st;
+
+	if (!(vcpu->arch.st.msr_val & KVM_MSR_ENABLED))
+		return;
+
+	if (vcpu->arch.st.is_idle)
+		return;
+
+	if (kvm_map_gfn(vcpu, vcpu->arch.st.msr_val >> PAGE_SHIFT, &map,
+			&vcpu->arch.st.cache, false))
+		return;
+
+	st = map.hva +
+		offset_in_page(vcpu->arch.st.msr_val & KVM_STEAL_VALID_BITS);
+
+	if (guest_pv_has(vcpu, KVM_FEATURE_PV_TLB_FLUSH))
+		xchg(&st->is_idle, 0);
+	else
+		st->is_idle = 0;
+
+	vcpu->arch.st.is_idle = 0;
+
+	kvm_unmap_gfn(vcpu, &map, &vcpu->arch.st.cache, true, false);
+}
+
+
+static DEFINE_PER_CPU(struct kvm_vcpu *, this_cpu_pre_run_vcpu);
+
+static void vcpu_load_update_pre_vcpu_callback(struct kvm_vcpu *new_vcpu, struct kvm_steal_time *st)
+{
+	struct kvm_vcpu *old_vcpu = __this_cpu_read(this_cpu_pre_run_vcpu);
+
+	if (!old_vcpu)
+		return;
+	if (old_vcpu != new_vcpu)
+		kvm_steal_time_clear_is_idle(old_vcpu);
+	else
+		st->is_idle = new_vcpu->arch.st.is_idle = KVM_PCPU_IS_IDLE;
+}
+
 static void record_steal_time(struct kvm_vcpu *vcpu)
 {
 	struct kvm_host_map map;
@@ -3219,6 +3285,8 @@ static void record_steal_time(struct kvm_vcpu *vcpu)
 
 	vcpu->arch.st.preempted = 0;
 
+	vcpu_load_update_pre_vcpu_callback(vcpu, st);
+
 	if (st->version & 1)
 		st->version += 1;  /* first time write, random junk */
 
@@ -4290,6 +4358,8 @@ static void kvm_steal_time_set_preempted(struct kvm_vcpu *vcpu)
 	kvm_unmap_gfn(vcpu, &map, &vcpu->arch.st.cache, true, true);
 }
 
+extern int get_cpu_nr_running(int cpu);
+
 void kvm_arch_vcpu_put(struct kvm_vcpu *vcpu)
 {
 	int idx;
@@ -4304,8 +4374,14 @@ void kvm_arch_vcpu_put(struct kvm_vcpu *vcpu)
 	idx = srcu_read_lock(&vcpu->kvm->srcu);
 	if (kvm_xen_msr_enabled(vcpu->kvm))
 		kvm_xen_runstate_set_preempted(vcpu);
-	else
+	else {
 		kvm_steal_time_set_preempted(vcpu);
+
+		if (get_cpu_nr_running(smp_processor_id()) <= 1)
+			kvm_steal_time_set_is_idle(vcpu);
+		else
+			kvm_steal_time_clear_is_idle(vcpu);
+	}
 	srcu_read_unlock(&vcpu->kvm->srcu, idx);
 
 	static_call(kvm_x86_vcpu_put)(vcpu);
@@ -9693,6 +9769,8 @@ static int vcpu_enter_guest(struct kvm_vcpu *vcpu)
 	local_irq_enable();
 	preempt_enable();
 
+	__this_cpu_write(this_cpu_pre_run_vcpu, vcpu);
+
 	vcpu->srcu_idx = srcu_read_lock(&vcpu->kvm->srcu);
 
 	/*
@@ -11253,6 +11331,14 @@ void kvm_arch_pre_destroy_vm(struct kvm *kvm)
 
 void kvm_arch_destroy_vm(struct kvm *kvm)
 {
+	int cpu;
+	struct kvm_vcpu *vcpu;
+
+	for_each_possible_cpu(cpu) {
+		vcpu = per_cpu(this_cpu_pre_run_vcpu, cpu);
+		if (vcpu && vcpu->kvm == kvm)
+			per_cpu(this_cpu_pre_run_vcpu, cpu) = NULL;
+	}
+
 	if (current->mm == kvm->mm) {
 		/*
 		 * Free memory regions allocated on behalf of userspace,
-- 
2.26.0


^ permalink raw reply	[flat|nested] 10+ messages in thread

* [PATCH 4/4] KVM guest implementation
  2021-08-31  1:59 [PATCH 1/4] KVM: x86: Introduce .pcpu_is_idle() stub infrastructure Tianqiang Xu
  2021-08-31  1:59 ` [PATCH 2/4] Scheduler changes Tianqiang Xu
  2021-08-31  1:59 ` [PATCH 3/4] KVM host implementation Tianqiang Xu
@ 2021-08-31  1:59 ` Tianqiang Xu
  2021-08-31  7:21   ` Peter Zijlstra
  2021-09-01  4:39 ` [PATCH 1/4] KVM: x86: Introduce .pcpu_is_idle() stub infrastructure Jarkko Sakkinen
  2021-12-17  9:39 ` Jinrong Liang
  4 siblings, 1 reply; 10+ messages in thread
From: Tianqiang Xu @ 2021-08-31  1:59 UTC (permalink / raw)
  To: x86
  Cc: pbonzini, seanjc, vkuznets, wanpengli, jmattson, joro, tglx,
	mingo, bp, kvm, hpa, jarkko, dave.hansen, linux-kernel,
	linux-sgx, Tianqiang Xu

Guest OS uses 'is_idle' field of kvm_steal_time to know if a pCPU
is idle and decides whether to schedule a task to a preempted vCPU
or not. If the pCPU is idle, scheduling a task to this pCPU will
improve cpu utilization. If not, avoiding scheduling a task to this
preempted vCPU can avoid host/guest switch, hence improving performance.

Guest OS invokes available_idle_cpu_sched() to get the value of
'is_idle' field of kvm_steal_time.

Other modules in kernel except kernel/sched/fair.c which invokes
available_idle_cpu() is left unchanged, because other modules in
kernel need the semantic provided by 'preempted' field of kvm_steal_time.

--
Authors: Tianqiang Xu, Dingji Li, Zeyu Mi
	 Shanghai Jiao Tong University

Signed-off-by: Tianqiang Xu <skyele@sjtu.edu.cn>

---
 kernel/sched/fair.c | 24 ++++++++++++------------
 1 file changed, 12 insertions(+), 12 deletions(-)

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 44c452072a1b..f69f0a8d2abe 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -5871,13 +5871,13 @@ wake_affine_idle(int this_cpu, int prev_cpu, int sync)
 	 * a cpufreq perspective, it's better to have higher utilisation
 	 * on one CPU.
 	 */
-	if (available_idle_cpu(this_cpu) && cpus_share_cache(this_cpu, prev_cpu))
-		return available_idle_cpu(prev_cpu) ? prev_cpu : this_cpu;
+	if (available_idle_cpu_sched(this_cpu) && cpus_share_cache(this_cpu, prev_cpu))
+		return available_idle_cpu_sched(prev_cpu) ? prev_cpu : this_cpu;
 
 	if (sync && cpu_rq(this_cpu)->nr_running == 1)
 		return this_cpu;
 
-	if (available_idle_cpu(prev_cpu))
+	if (available_idle_cpu_sched(prev_cpu))
 		return prev_cpu;
 
 	return nr_cpumask_bits;
@@ -5976,7 +5976,7 @@ find_idlest_group_cpu(struct sched_group *group, struct task_struct *p, int this
 		if (sched_idle_cpu(i))
 			return i;
 
-		if (available_idle_cpu(i)) {
+		if (available_idle_cpu_sched(i)) {
 			struct cpuidle_state *idle = idle_get_state(rq);
 			if (idle && idle->exit_latency < min_exit_latency) {
 				/*
@@ -6064,7 +6064,7 @@ static inline int find_idlest_cpu(struct sched_domain *sd, struct task_struct *p
 
 static inline int __select_idle_cpu(int cpu, struct task_struct *p)
 {
-	if ((available_idle_cpu(cpu) || sched_idle_cpu(cpu)) &&
+	if ((available_idle_cpu_sched(cpu) || sched_idle_cpu(cpu)) &&
 	    sched_cpu_cookie_match(cpu_rq(cpu), p))
 		return cpu;
 
@@ -6115,7 +6115,7 @@ void __update_idle_core(struct rq *rq)
 		if (cpu == core)
 			continue;
 
-		if (!available_idle_cpu(cpu))
+		if (!available_idle_cpu_sched(cpu))
 			goto unlock;
 	}
 
@@ -6138,7 +6138,7 @@ static int select_idle_core(struct task_struct *p, int core, struct cpumask *cpu
 		return __select_idle_cpu(core, p);
 
 	for_each_cpu(cpu, cpu_smt_mask(core)) {
-		if (!available_idle_cpu(cpu)) {
+		if (!available_idle_cpu_sched(cpu)) {
 			idle = false;
 			if (*idle_cpu == -1) {
 				if (sched_idle_cpu(cpu) && cpumask_test_cpu(cpu, p->cpus_ptr)) {
@@ -6171,7 +6171,7 @@ static int select_idle_smt(struct task_struct *p, struct sched_domain *sd, int t
 		if (!cpumask_test_cpu(cpu, p->cpus_ptr) ||
 		    !cpumask_test_cpu(cpu, sched_domain_span(sd)))
 			continue;
-		if (available_idle_cpu(cpu) || sched_idle_cpu(cpu))
+		if (available_idle_cpu_sched(cpu) || sched_idle_cpu(cpu))
 			return cpu;
 	}
 
@@ -6302,7 +6302,7 @@ select_idle_capacity(struct task_struct *p, struct sched_domain *sd, int target)
 	for_each_cpu_wrap(cpu, cpus, target) {
 		unsigned long cpu_cap = capacity_of(cpu);
 
-		if (!available_idle_cpu(cpu) && !sched_idle_cpu(cpu))
+		if (!available_idle_cpu_sched(cpu) && !sched_idle_cpu(cpu))
 			continue;
 		if (fits_capacity(task_util, cpu_cap))
 			return cpu;
@@ -6348,7 +6348,7 @@ static int select_idle_sibling(struct task_struct *p, int prev, int target)
 	 */
 	lockdep_assert_irqs_disabled();
 
-	if ((available_idle_cpu(target) || sched_idle_cpu(target)) &&
+	if ((available_idle_cpu_sched(target) || sched_idle_cpu(target)) &&
 	    asym_fits_capacity(task_util, target))
 		return target;
 
@@ -6356,7 +6356,7 @@ static int select_idle_sibling(struct task_struct *p, int prev, int target)
 	 * If the previous CPU is cache affine and idle, don't be stupid:
 	 */
 	if (prev != target && cpus_share_cache(prev, target) &&
-	    (available_idle_cpu(prev) || sched_idle_cpu(prev)) &&
+	    (available_idle_cpu_sched(prev) || sched_idle_cpu(prev)) &&
 	    asym_fits_capacity(task_util, prev))
 		return prev;
 
@@ -6379,7 +6379,7 @@ static int select_idle_sibling(struct task_struct *p, int prev, int target)
 	if (recent_used_cpu != prev &&
 	    recent_used_cpu != target &&
 	    cpus_share_cache(recent_used_cpu, target) &&
-	    (available_idle_cpu(recent_used_cpu) || sched_idle_cpu(recent_used_cpu)) &&
+	    (available_idle_cpu_sched(recent_used_cpu) || sched_idle_cpu(recent_used_cpu)) &&
 	    cpumask_test_cpu(p->recent_used_cpu, p->cpus_ptr) &&
 	    asym_fits_capacity(task_util, recent_used_cpu)) {
 		/*
-- 
2.26.0


^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: [PATCH 2/4] Scheduler changes
  2021-08-31  1:59 ` [PATCH 2/4] Scheduler changes Tianqiang Xu
@ 2021-08-31  7:14   ` Peter Zijlstra
  0 siblings, 0 replies; 10+ messages in thread
From: Peter Zijlstra @ 2021-08-31  7:14 UTC (permalink / raw)
  To: Tianqiang Xu
  Cc: x86, pbonzini, seanjc, vkuznets, wanpengli, jmattson, joro, tglx,
	mingo, bp, kvm, hpa, jarkko, dave.hansen, linux-kernel,
	linux-sgx

On Tue, Aug 31, 2021 at 09:59:17AM +0800, Tianqiang Xu wrote:
> Authors: Tianqiang Xu, Dingji Li, Zeyu Mi
> 	 Shanghai Jiao Tong University
> 
> Signed-off-by: Tianqiang Xu <skyele@sjtu.edu.cn>

Authors is not a valid tag, please see our Documentation on submitting
patches.

> @@ -10504,3 +10515,9 @@ void call_trace_sched_update_nr_running(struct rq *rq, int count)
>  {
>          trace_sched_update_nr_running_tp(rq, count);
>  }
> +
> +int get_cpu_nr_running(int cpu)
> +{
> +	return cpu_rq(cpu)->nr_running;
> +}
> +EXPORT_SYMBOL_GPL(get_cpu_nr_running);

Yeah, not going to happen.

^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: [PATCH 3/4] KVM host implementation
  2021-08-31  1:59 ` [PATCH 3/4] KVM host implementation Tianqiang Xu
@ 2021-08-31  7:16   ` Peter Zijlstra
  2021-08-31  7:17     ` Peter Zijlstra
  0 siblings, 1 reply; 10+ messages in thread
From: Peter Zijlstra @ 2021-08-31  7:16 UTC (permalink / raw)
  To: Tianqiang Xu
  Cc: x86, pbonzini, seanjc, vkuznets, wanpengli, jmattson, joro, tglx,
	mingo, bp, kvm, hpa, jarkko, dave.hansen, linux-kernel,
	linux-sgx

On Tue, Aug 31, 2021 at 09:59:18AM +0800, Tianqiang Xu wrote:
> @@ -4304,8 +4374,14 @@ void kvm_arch_vcpu_put(struct kvm_vcpu *vcpu)
>  	idx = srcu_read_lock(&vcpu->kvm->srcu);
>  	if (kvm_xen_msr_enabled(vcpu->kvm))
>  		kvm_xen_runstate_set_preempted(vcpu);
> -	else
> +	else {
>  		kvm_steal_time_set_preempted(vcpu);
> +
> +		if (get_cpu_nr_running(smp_processor_id()) <= 1)
> +			kvm_steal_time_set_is_idle(vcpu);
> +		else
> +			kvm_steal_time_clear_is_idle(vcpu);
> +	}
>  	srcu_read_unlock(&vcpu->kvm->srcu, idx);


This cannot be right. The CPU could long since be running tasks again,
but as long as this vCPU crud doesn't run, the guest keeps thinking it's
physically idle.

^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: [PATCH 3/4] KVM host implementation
  2021-08-31  7:16   ` Peter Zijlstra
@ 2021-08-31  7:17     ` Peter Zijlstra
  0 siblings, 0 replies; 10+ messages in thread
From: Peter Zijlstra @ 2021-08-31  7:17 UTC (permalink / raw)
  To: Tianqiang Xu
  Cc: x86, pbonzini, seanjc, vkuznets, wanpengli, jmattson, joro, tglx,
	mingo, bp, kvm, hpa, jarkko, dave.hansen, linux-kernel,
	linux-sgx

On Tue, Aug 31, 2021 at 09:16:09AM +0200, Peter Zijlstra wrote:
> On Tue, Aug 31, 2021 at 09:59:18AM +0800, Tianqiang Xu wrote:
> > @@ -4304,8 +4374,14 @@ void kvm_arch_vcpu_put(struct kvm_vcpu *vcpu)
> >  	idx = srcu_read_lock(&vcpu->kvm->srcu);
> >  	if (kvm_xen_msr_enabled(vcpu->kvm))
> >  		kvm_xen_runstate_set_preempted(vcpu);
> > -	else
> > +	else {
> >  		kvm_steal_time_set_preempted(vcpu);
> > +
> > +		if (get_cpu_nr_running(smp_processor_id()) <= 1)
> > +			kvm_steal_time_set_is_idle(vcpu);
> > +		else
> > +			kvm_steal_time_clear_is_idle(vcpu);
> > +	}
> >  	srcu_read_unlock(&vcpu->kvm->srcu, idx);
> 
> 
> This cannot be right. The CPU could long since be running tasks again,
> but as long as this vCPU crud doesn't run, the guest keeps thinking it's
> physically idle.

More fundamentally, a blocked task doesn't have a CPU. So unless you've
pinned your vCPU threads to physical CPUs, the whole thing is bonkers.

^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: [PATCH 4/4] KVM guest implementation
  2021-08-31  1:59 ` [PATCH 4/4] KVM guest implementation Tianqiang Xu
@ 2021-08-31  7:21   ` Peter Zijlstra
  0 siblings, 0 replies; 10+ messages in thread
From: Peter Zijlstra @ 2021-08-31  7:21 UTC (permalink / raw)
  To: Tianqiang Xu
  Cc: x86, pbonzini, seanjc, vkuznets, wanpengli, jmattson, joro, tglx,
	mingo, bp, kvm, hpa, jarkko, dave.hansen, linux-kernel,
	linux-sgx

On Tue, Aug 31, 2021 at 09:59:19AM +0800, Tianqiang Xu wrote:
> Guest OS uses 'is_idle' field of kvm_steal_time to know if a pCPU
> is idle and decides whether to schedule a task to a preempted vCPU
> or not. If the pCPU is idle, scheduling a task to this pCPU will
> improve cpu utilization. If not, avoiding scheduling a task to this
> preempted vCPU can avoid host/guest switch, hence improving performance.
> 
> Guest OS invokes available_idle_cpu_sched() to get the value of
> 'is_idle' field of kvm_steal_time.
> 
> Other modules in kernel except kernel/sched/fair.c which invokes
> available_idle_cpu() is left unchanged, because other modules in
> kernel need the semantic provided by 'preempted' field of kvm_steal_time.

> ---
>  kernel/sched/fair.c | 24 ++++++++++++------------
>  1 file changed, 12 insertions(+), 12 deletions(-)

Goes and replaces every single available_idle_cpu() in fair with the new
function that doesn't consider vCPU preemption.

So what do you reckon now happens in the oversubscribed virt scenario
where each CPU has multiple vCPUs?

^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: [PATCH 1/4] KVM: x86: Introduce .pcpu_is_idle() stub infrastructure
  2021-08-31  1:59 [PATCH 1/4] KVM: x86: Introduce .pcpu_is_idle() stub infrastructure Tianqiang Xu
                   ` (2 preceding siblings ...)
  2021-08-31  1:59 ` [PATCH 4/4] KVM guest implementation Tianqiang Xu
@ 2021-09-01  4:39 ` Jarkko Sakkinen
  2021-12-17  9:39 ` Jinrong Liang
  4 siblings, 0 replies; 10+ messages in thread
From: Jarkko Sakkinen @ 2021-09-01  4:39 UTC (permalink / raw)
  To: Tianqiang Xu, x86
  Cc: pbonzini, seanjc, vkuznets, wanpengli, jmattson, joro, tglx,
	mingo, bp, kvm, hpa, dave.hansen, linux-kernel, linux-sgx

On Tue, 2021-08-31 at 09:59 +0800, Tianqiang Xu wrote:
> This patch series aims to fix performance issue caused by current
> para-virtualized scheduling design.

Series?

This looks to me like a patch, not a cover letter.

If you want a cover letter, please make one.

/Jarkko

^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: [PATCH 1/4] KVM: x86: Introduce .pcpu_is_idle() stub infrastructure
  2021-08-31  1:59 [PATCH 1/4] KVM: x86: Introduce .pcpu_is_idle() stub infrastructure Tianqiang Xu
                   ` (3 preceding siblings ...)
  2021-09-01  4:39 ` [PATCH 1/4] KVM: x86: Introduce .pcpu_is_idle() stub infrastructure Jarkko Sakkinen
@ 2021-12-17  9:39 ` Jinrong Liang
  4 siblings, 0 replies; 10+ messages in thread
From: Jinrong Liang @ 2021-12-17  9:39 UTC (permalink / raw)
  To: Tianqiang Xu
  Cc: x86, pbonzini, seanjc, vkuznets, wanpengli, jmattson, joro, tglx,
	mingo, bp, kvm, hpa, jarkko, dave.hansen, linux-kernel,
	linux-sgx

Hi Tianqiang,
Tianqiang Xu <skyele@sjtu.edu.cn> 于2021年12月17日周五 15:55写道:
>
> This patch series aims to fix performance issue caused by current
> para-virtualized scheduling design.
>
> The current para-virtualized scheduling design uses 'preempted' field of
> kvm_steal_time to avoid scheduling task on the preempted vCPU.
> However, when the pCPU where the preempted vCPU most recently run is idle,
> it will result in low cpu utilization, and consequently poor performance.
>
> The new field: 'is_idle' of kvm_steal_time can precisely reveal
> the status of pCPU where preempted vCPU most recently run, and
> then improve cpu utilization.
>
> pcpu_is_idle() is used to get the value of 'is_idle' of kvm_steal_time.
>
> Experiments on a VM with 16 vCPUs show that the patch can reduce around
> 50% to 80% execution time for most PARSEC benchmarks.
> This also holds true for a VM with 112 vCPUs.
>
> Experiments on 2 VMs with 112 vCPUs show that the patch can reduce around
> 20% to 80% execution time for most PARSEC benchmarks.
>
> Test environment:
> -- PowerEdge R740
> -- 56C-112T CPU Intel(R) Xeon(R) Gold 6238R CPU
> -- Host 190G DRAM
> -- QEMU 5.0.0
> -- PARSEC 3.0 Native Inputs
> -- Host is idle during the test
> -- Host and Guest kernel are both kernel-5.14.0
>
> Results:
> 1. 1 VM, 16 VCPU, 16 THREAD.
>    Host Topology: sockets=2 cores=28 threads=2
>    VM Topology:   sockets=1 cores=16 threads=1
>    Command: <path to parsec>/bin/parsecmgmt -a run -p <benchmark> -i native -n 16
>    Statistics below are the real time of running each benchmark.(lower is better)
>
>                         before patch    after patch     improvements
> bodytrack               52.866s         22.619s         57.21%
> fluidanimate            84.009s         38.148s         54.59%
> streamcluster           270.17s         42.726s         84.19%
> splash2x.ocean_cp       31.932s         9.539s          70.13%
> splash2x.ocean_ncp      36.063s         14.189s         60.65%
> splash2x.volrend        134.587s        21.79s          83.81%
>
> 2. 1VM, 112 VCPU. Some benchmarks require the number of threads to be the power of 2,
> so we run them with 64 threads and 128 threads.
>    Host Topology: sockets=2 cores=28 threads=2
>    VM Topology:   sockets=1 cores=112 threads=1
>    Command: <path to parsec>/bin/parsecmgmt -a run -p <benchmark> -i native -n <64,112,128>
>    Statistics below are the real time of running each benchmark.(lower is better)
>
>                                         before patch    after patch     improvements
> fluidanimate(64 thread)                 124.235s        27.924s         77.52%
> fluidanimate(128 thread)                169.127s        64.541s         61.84%
> streamcluster(112 thread)               861.879s        496.66s         42.37%
> splash2x.ocean_cp(64 thread)            46.415s         18.527s         60.08%
> splash2x.ocean_cp(128 thread)           53.647s         28.929s         46.08%
> splash2x.ocean_ncp(64 thread)           47.613s         19.576s         58.89%
> splash2x.ocean_ncp(128 thread)          54.94s          29.199s         46.85%
> splash2x.volrend(112 thread)            801.384s        144.824s        81.93%
>
> 3. 2VM, each VM: 112 VCPU. Some benchmarks require the number of threads to
> be the power of 2, so we run them with 64 threads and 128 threads.
>    Host Topology: sockets=2 cores=28 threads=2
>    VM Topology:   sockets=1 cores=112 threads=1
>    Command: <path to parsec>/bin/parsecmgmt -a run -p <benchmark> -i native -n <64,112,128>
>    Statistics below are the average real time of running each benchmark in 2 VMs.(lower is better)
>
>                                         before patch    after patch     improvements
> fluidanimate(64 thread)                 135.2125s       49.827s         63.15%
> fluidanimate(128 thread)                178.309s        86.964s         51.23%
> splash2x.ocean_cp(64 thread)            47.4505s        20.314s         57.19%
> splash2x.ocean_cp(128 thread)           55.5645s        30.6515s        44.84%
> splash2x.ocean_ncp(64 thread)           49.9775s        23.489s         53.00%
> splash2x.ocean_ncp(128 thread)          56.847s         28.545s         49.79%
> splash2x.volrend(112 thread)            838.939s        239.632s        71.44%
>
> For space limit, we list representative statistics here.

I did a performance test according to the description in the patch,
but did not get the performance improvement described in the description.

I suspect that the big difference between my kernel configuration
and yours has caused this problem. Can you please provide more detailed
test information, such as kernel configuration that must be turned on or off ?

Regards,
Jinrong Liang

^ permalink raw reply	[flat|nested] 10+ messages in thread

end of thread, other threads:[~2021-12-17  9:39 UTC | newest]

Thread overview: 10+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-08-31  1:59 [PATCH 1/4] KVM: x86: Introduce .pcpu_is_idle() stub infrastructure Tianqiang Xu
2021-08-31  1:59 ` [PATCH 2/4] Scheduler changes Tianqiang Xu
2021-08-31  7:14   ` Peter Zijlstra
2021-08-31  1:59 ` [PATCH 3/4] KVM host implementation Tianqiang Xu
2021-08-31  7:16   ` Peter Zijlstra
2021-08-31  7:17     ` Peter Zijlstra
2021-08-31  1:59 ` [PATCH 4/4] KVM guest implementation Tianqiang Xu
2021-08-31  7:21   ` Peter Zijlstra
2021-09-01  4:39 ` [PATCH 1/4] KVM: x86: Introduce .pcpu_is_idle() stub infrastructure Jarkko Sakkinen
2021-12-17  9:39 ` Jinrong Liang

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).