From: Zengruan Ye yezengruan@huawei.com

virt inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8WMFU CVE: NA

--------------------------------

Introduce a paravirtualization interface for KVM/arm64 to PV-sched.

A hypercall interface is provided for the guest to interrogate the hypervisor's support for this interface and the location of the shared memory structures.

Signed-off-by: Zengruan Ye yezengruan@huawei.com Reviewed-by: Zhanghailiang zhang.zhanghailiang@huawei.com Signed-off-by: Zheng Zengkai zhengzengkai@huawei.com

Signed-off-by: lishusen lishusen2@huawei.com --- Documentation/virt/kvm/arm/pvsched.rst | 58 ++++++++++++++++++++++++++ 1 file changed, 58 insertions(+) create mode 100644 Documentation/virt/kvm/arm/pvsched.rst

diff --git a/Documentation/virt/kvm/arm/pvsched.rst b/Documentation/virt/kvm/arm/pvsched.rst new file mode 100644 index 000000000000..8f7112a8a9cd --- /dev/null +++ b/Documentation/virt/kvm/arm/pvsched.rst @@ -0,0 +1,58 @@ +.. SPDX-License-Identifier: GPL-2.0 + +Paravirtualized sched support for arm64 +======================================= + +KVM/arm64 provides some hypervisor service calls to support a paravirtualized +sched. + +Some SMCCC compatible hypercalls are defined: + +* PV_SCHED_FEATURES: 0xC5000090 +* PV_SCHED_IPA_INIT: 0xC5000091 +* PV_SCHED_IPA_RELEASE: 0xC5000092 + +The existence of the PV_SCHED hypercall should be probed using the SMCCC 1.1 +ARCH_FEATURES mechanism before calling it. + +PV_SCHED_FEATURES + ============= ======== ========== + Function ID: (uint32) 0xC5000090 + PV_call_id: (uint32) The function to query for support. + Return value: (int64) NOT_SUPPORTED (-1) or SUCCESS (0) if the relevant + PV-sched feature is supported by the hypervisor. + ============= ======== ========== + +PV_SCHED_IPA_INIT + ============= ======== ========== + Function ID: (uint32) 0xC5000091 + Return value: (int64) NOT_SUPPORTED (-1) or SUCCESS (0) if the IPA of + this vCPU's PV data structure is shared to the + hypervisor. + ============= ======== ========== + +PV_SCHED_IPA_RELEASE + ============= ======== ========== + Function ID: (uint32) 0xC5000092 + Return value: (int64) NOT_SUPPORTED (-1) or SUCCESS (0) if the IPA of + this vCPU's PV data structure is released. + ============= ======== ========== + +PV sched state +-------------- + +The structure pointed to by the PV_SCHED_IPA hypercall is as follows: + ++-----------+-------------+-------------+-----------------------------------+ +| Field | Byte Length | Byte Offset | Description | ++===========+=============+=============+===================================+ +| preempted | 4 | 0 | Indicates that the vCPU that owns | +| | | | this struct is running or not. | +| | | | Non-zero values mean the vCPU has | +| | | | been preempted. Zero means the | +| | | | vCPU is not preempted. | ++-----------+-------------+-------------+-----------------------------------+ + +The preempted field will be updated to 0 by the hypervisor prior to scheduling +a vCPU. When the vCPU is scheduled out, the preempted field will be updated +to 1 by the hypervisor.

From: Zengruan Ye yezengruan@huawei.com

virt inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8WMFU CVE: NA

--------------------------------

Implement the service call for configuring a shared structure between a vCPU and the hypervisor in which the hypervisor can tell the vCPU that is running or not.

Signed-off-by: Zengruan Ye yezengruan@huawei.com Reviewed-by: Zhanghailiang zhang.zhanghailiang@huawei.com Signed-off-by: Zheng Zengkai zhengzengkai@huawei.com

Signed-off-by: lishusen lishusen2@huawei.com --- C | 0 arch/arm64/include/asm/kvm_host.h | 16 ++++++++++++++++ arch/arm64/kvm/arm.c | 9 +++++++++ arch/arm64/kvm/hypercalls.c | 11 +++++++++++ arch/arm64/kvm/pvsched.c | 28 ++++++++++++++++++++++++++++ 5 files changed, 64 insertions(+) create mode 100644 C

diff --git a/C b/C new file mode 100644 index 000000000000..e69de29bb2d1 diff --git a/arch/arm64/include/asm/kvm_host.h b/arch/arm64/include/asm/kvm_host.h index 50fcc92257cf..e07cb6ff542c 100644 --- a/arch/arm64/include/asm/kvm_host.h +++ b/arch/arm64/include/asm/kvm_host.h @@ -596,6 +596,11 @@ struct kvm_vcpu_arch { gpa_t base; } steal;

+ /* Guest PV sched state */ + struct { + gpa_t base; + } pvsched; + /* Per-vcpu CCSIDR override or NULL */ u32 *ccsidr;

@@ -1052,6 +1057,17 @@ static inline bool kvm_arm_is_pvtime_enabled(struct kvm_vcpu_arch *vcpu_arch) }

long kvm_hypercall_pvsched_features(struct kvm_vcpu *vcpu); +void kvm_update_pvsched_preempted(struct kvm_vcpu *vcpu, u32 preempted); + +static inline void kvm_arm_pvsched_vcpu_init(struct kvm_vcpu_arch *vcpu_arch) +{ + vcpu_arch->pvsched.base = GPA_INVALID; +} + +static inline bool kvm_arm_is_pvsched_enabled(struct kvm_vcpu_arch *vcpu_arch) +{ + return (vcpu_arch->pvsched.base != GPA_INVALID); +}

void kvm_set_sei_esr(struct kvm_vcpu *vcpu, u64 syndrome);

diff --git a/arch/arm64/kvm/arm.c b/arch/arm64/kvm/arm.c index dce5804b5df2..837ced945691 100644 --- a/arch/arm64/kvm/arm.c +++ b/arch/arm64/kvm/arm.c @@ -415,6 +415,8 @@ int kvm_arch_vcpu_create(struct kvm_vcpu *vcpu)

kvm_arm_pvtime_vcpu_init(&vcpu->arch);

+ kvm_arm_pvsched_vcpu_init(&vcpu->arch); + vcpu->arch.hw_mmu = &vcpu->kvm->arch.mmu;

err = kvm_vgic_vcpu_init(vcpu); @@ -500,11 +502,15 @@ void kvm_arch_vcpu_load(struct kvm_vcpu *vcpu, int cpu)

if (vcpu_has_ptrauth(vcpu)) vcpu_ptrauth_disable(vcpu); + kvm_arch_vcpu_load_debug_state_flags(vcpu);

if (!cpumask_test_cpu(cpu, vcpu->kvm->arch.supported_cpus)) vcpu_set_on_unsupported_cpu(vcpu);

+ if (kvm_arm_is_pvsched_enabled(&vcpu->arch)) + kvm_update_pvsched_preempted(vcpu, 0); + #ifdef CONFIG_KVM_HISI_VIRT kvm_hisi_dvmbm_load(vcpu); #endif @@ -524,6 +530,9 @@ void kvm_arch_vcpu_put(struct kvm_vcpu *vcpu) vcpu_clear_on_unsupported_cpu(vcpu); vcpu->cpu = -1;

+ if (kvm_arm_is_pvsched_enabled(&vcpu->arch)) + kvm_update_pvsched_preempted(vcpu, 1); + #ifdef CONFIG_KVM_HISI_VIRT kvm_hisi_dvmbm_put(vcpu); #endif diff --git a/arch/arm64/kvm/hypercalls.c b/arch/arm64/kvm/hypercalls.c index 190670d8dd3f..b72224609c3e 100644 --- a/arch/arm64/kvm/hypercalls.c +++ b/arch/arm64/kvm/hypercalls.c @@ -366,6 +366,17 @@ int kvm_smccc_call_handler(struct kvm_vcpu *vcpu) case ARM_SMCCC_HV_PV_SCHED_FEATURES: val[0] = kvm_hypercall_pvsched_features(vcpu); break; + case ARM_SMCCC_HV_PV_SCHED_IPA_INIT: + gpa = smccc_get_arg1(vcpu); + if (gpa != GPA_INVALID) { + vcpu->arch.pvsched.base = gpa; + val = SMCCC_RET_SUCCESS; + } + break; + case ARM_SMCCC_HV_PV_SCHED_IPA_RELEASE: + vcpu->arch.pvsched.base = GPA_INVALID; + val = SMCCC_RET_SUCCESS; + break; default: return kvm_psci_call(vcpu); } diff --git a/arch/arm64/kvm/pvsched.c b/arch/arm64/kvm/pvsched.c index 3d96122fcf9e..b923c4b6c52e 100644 --- a/arch/arm64/kvm/pvsched.c +++ b/arch/arm64/kvm/pvsched.c @@ -5,9 +5,35 @@ */

#include <linux/arm-smccc.h> +#include <linux/kvm_host.h> + +#include <asm/pvsched-abi.h>

#include <kvm/arm_hypercalls.h>

+void kvm_update_pvsched_preempted(struct kvm_vcpu *vcpu, u32 preempted) +{ + struct kvm *kvm = vcpu->kvm; + u64 base = vcpu->arch.pvsched.base; + u64 offset = offsetof(struct pvsched_vcpu_state, preempted); + int idx; + + if (base == GPA_INVALID) + return; + + /* + * This function is called from atomic context, so we need to + * disable page faults. + */ + pagefault_disable(); + + idx = srcu_read_lock(&kvm->srcu); + kvm_put_guest(kvm, base + offset, cpu_to_le32(preempted)); + srcu_read_unlock(&kvm->srcu, idx); + + pagefault_enable(); +} + long kvm_hypercall_pvsched_features(struct kvm_vcpu *vcpu) { u32 feature = smccc_get_arg1(vcpu); @@ -15,6 +41,8 @@ long kvm_hypercall_pvsched_features(struct kvm_vcpu *vcpu)

switch (feature) { case ARM_SMCCC_HV_PV_SCHED_FEATURES: + case ARM_SMCCC_HV_PV_SCHED_IPA_INIT: + case ARM_SMCCC_HV_PV_SCHED_IPA_RELEASE: val = SMCCC_RET_SUCCESS; break; }

From: Zengruan Ye yezengruan@huawei.com

virt inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8WMFU CVE: NA

--------------------------------

Support the vcpu_is_preempted() functionality under KVM/arm64. This will enhance lock performance on overcommitted hosts (more runnable vCPUs than physical CPUs in the system) as doing busy waits for preempted vCPUs will hurt system performance far worse than early yielding.

unix benchmark result: host: kernel 4.19.87, HiSilicon Kunpeng920, 8 CPUs guest: kernel 4.19.87, 16 vCPUs

test-case | after-patch | before-patch ----------------------------------------+-------------------+------------------ Dhrystone 2 using register variables | 338955728.5 lps | 339266319.5 lps Double-Precision Whetstone | 30634.9 MWIPS | 30884.4 MWIPS Execl Throughput | 6753.2 lps | 3580.1 lps File Copy 1024 bufsize 2000 maxblocks | 490048.0 KBps | 313282.3 KBps File Copy 256 bufsize 500 maxblocks | 129662.5 KBps | 83550.7 KBps File Copy 4096 bufsize 8000 maxblocks | 1552551.5 KBps | 814327.0 KBps Pipe Throughput | 8976422.5 lps | 9048628.4 lps Pipe-based Context Switching | 258641.7 lps | 252925.9 lps Process Creation | 5312.2 lps | 4507.9 lps Shell Scripts (1 concurrent) | 8704.2 lpm | 6720.9 lpm Shell Scripts (8 concurrent) | 1708.8 lpm | 607.2 lpm System Call Overhead | 3714444.7 lps | 3746386.8 lps ----------------------------------------+-------------------+------------------ System Benchmarks Index Score | 2270.6 | 1679.2

Signed-off-by: Zengruan Ye yezengruan@huawei.com Reviewed-by: Zhanghailiang zhang.zhanghailiang@huawei.com Signed-off-by: Zheng Zengkai zhengzengkai@huawei.com

Signed-off-by: lishusen lishusen2@huawei.com --- arch/arm64/include/asm/kvm_host.h | 4 +- arch/arm64/include/asm/paravirt.h | 3 + arch/arm64/kernel/paravirt.c | 106 ++++++++++++++++++++++++++++++ arch/arm64/kernel/setup.c | 2 + arch/arm64/kvm/hypercalls.c | 8 +-- arch/arm64/kvm/pvsched.c | 2 +- include/linux/cpuhotplug.h | 1 + 7 files changed, 119 insertions(+), 7 deletions(-)

diff --git a/arch/arm64/include/asm/kvm_host.h b/arch/arm64/include/asm/kvm_host.h index e07cb6ff542c..0537099a0bd0 100644 --- a/arch/arm64/include/asm/kvm_host.h +++ b/arch/arm64/include/asm/kvm_host.h @@ -1061,12 +1061,12 @@ void kvm_update_pvsched_preempted(struct kvm_vcpu *vcpu, u32 preempted);

static inline void kvm_arm_pvsched_vcpu_init(struct kvm_vcpu_arch *vcpu_arch) { - vcpu_arch->pvsched.base = GPA_INVALID; + vcpu_arch->pvsched.base = INVALID_GPA; }

static inline bool kvm_arm_is_pvsched_enabled(struct kvm_vcpu_arch *vcpu_arch) { - return (vcpu_arch->pvsched.base != GPA_INVALID); + return (vcpu_arch->pvsched.base != INVALID_GPA); }

void kvm_set_sei_esr(struct kvm_vcpu *vcpu, u64 syndrome); diff --git a/arch/arm64/include/asm/paravirt.h b/arch/arm64/include/asm/paravirt.h index 53282dc429f8..4e35851bcecf 100644 --- a/arch/arm64/include/asm/paravirt.h +++ b/arch/arm64/include/asm/paravirt.h @@ -20,6 +20,8 @@ static inline u64 paravirt_steal_clock(int cpu)

int __init pv_time_init(void);

+int __init pv_sched_init(void); + __visible bool __native_vcpu_is_preempted(int cpu); DECLARE_STATIC_CALL(pv_vcpu_preempted, __native_vcpu_is_preempted);

@@ -31,6 +33,7 @@ static inline bool pv_vcpu_is_preempted(int cpu) #else

#define pv_time_init() do {} while (0) +#define pv_sched_init() do {} while (0)

#endif // CONFIG_PARAVIRT

diff --git a/arch/arm64/kernel/paravirt.c b/arch/arm64/kernel/paravirt.c index aa718d6a9274..844622baf6bf 100644 --- a/arch/arm64/kernel/paravirt.c +++ b/arch/arm64/kernel/paravirt.c @@ -22,6 +22,7 @@

#include <asm/paravirt.h> #include <asm/pvclock-abi.h> +#include <asm/pvsched-abi.h> #include <asm/smp_plat.h>

struct static_key paravirt_steal_enabled; @@ -174,3 +175,108 @@ int __init pv_time_init(void)

return 0; } + + +DEFINE_PER_CPU(struct pvsched_vcpu_state, pvsched_vcpu_region) __aligned(64); +EXPORT_PER_CPU_SYMBOL(pvsched_vcpu_region); + +static bool kvm_vcpu_is_preempted(int cpu) +{ + struct pvsched_vcpu_state *reg; + u32 preempted; + + reg = &per_cpu(pvsched_vcpu_region, cpu); + if (!reg) { + pr_warn_once("PV sched enabled but not configured for cpu %d\n", + cpu); + return false; + } + + preempted = le32_to_cpu(READ_ONCE(reg->preempted)); + + return !!preempted; +} + +static int pvsched_vcpu_state_dying_cpu(unsigned int cpu) +{ + struct pvsched_vcpu_state *reg; + struct arm_smccc_res res; + + reg = this_cpu_ptr(&pvsched_vcpu_region); + if (!reg) + return -EFAULT; + + arm_smccc_1_1_invoke(ARM_SMCCC_HV_PV_SCHED_IPA_RELEASE, &res); + memset(reg, 0, sizeof(*reg)); + + return 0; +} + +static int init_pvsched_vcpu_state(unsigned int cpu) +{ + struct pvsched_vcpu_state *reg; + struct arm_smccc_res res; + + reg = this_cpu_ptr(&pvsched_vcpu_region); + if (!reg) + return -EFAULT; + + /* Pass the memory address to host via hypercall */ + arm_smccc_1_1_invoke(ARM_SMCCC_HV_PV_SCHED_IPA_INIT, + virt_to_phys(reg), &res); + + return 0; +} + +static int kvm_arm_init_pvsched(void) +{ + int ret; + + ret = cpuhp_setup_state(CPUHP_AP_ARM_KVM_PVSCHED_STARTING, + "hypervisor/arm/pvsched:starting", + init_pvsched_vcpu_state, + pvsched_vcpu_state_dying_cpu); + + if (ret < 0) { + pr_warn("PV sched init failed\n"); + return ret; + } + + return 0; +} + +static bool has_kvm_pvsched(void) +{ + struct arm_smccc_res res; + + /* To detect the presence of PV sched support we require SMCCC 1.1+ */ + if (arm_smccc_1_1_get_conduit() == SMCCC_CONDUIT_NONE) + return false; + + arm_smccc_1_1_invoke(ARM_SMCCC_ARCH_FEATURES_FUNC_ID, + ARM_SMCCC_HV_PV_SCHED_FEATURES, &res); + + return (res.a0 == SMCCC_RET_SUCCESS); +} + +int __init pv_sched_init(void) +{ + int ret; + + if (is_hyp_mode_available()) + return 0; + + if (!has_kvm_pvsched()) { + pr_warn("PV sched is not available\n"); + return 0; + } + + ret = kvm_arm_init_pvsched(); + if (ret) + return ret; + + static_call_update(pv_vcpu_preempted, kvm_vcpu_is_preempted); + pr_info("using PV sched preempted\n"); + + return 0; +} diff --git a/arch/arm64/kernel/setup.c b/arch/arm64/kernel/setup.c index 645034e52496..0255788cb9c4 100644 --- a/arch/arm64/kernel/setup.c +++ b/arch/arm64/kernel/setup.c @@ -416,6 +416,8 @@ void __init __no_sanitize_address setup_arch(char **cmdline_p) smp_init_cpus(); smp_build_mpidr_hash();

+ pv_sched_init(); + /* Init percpu seeds for random tags after cpus are set up. */ kasan_init_sw_tags();

diff --git a/arch/arm64/kvm/hypercalls.c b/arch/arm64/kvm/hypercalls.c index b72224609c3e..a96b88d90cf8 100644 --- a/arch/arm64/kvm/hypercalls.c +++ b/arch/arm64/kvm/hypercalls.c @@ -368,14 +368,14 @@ int kvm_smccc_call_handler(struct kvm_vcpu *vcpu) break; case ARM_SMCCC_HV_PV_SCHED_IPA_INIT: gpa = smccc_get_arg1(vcpu); - if (gpa != GPA_INVALID) { + if (gpa != INVALID_GPA) { vcpu->arch.pvsched.base = gpa; - val = SMCCC_RET_SUCCESS; + val[0] = SMCCC_RET_SUCCESS; } break; case ARM_SMCCC_HV_PV_SCHED_IPA_RELEASE: - vcpu->arch.pvsched.base = GPA_INVALID; - val = SMCCC_RET_SUCCESS; + vcpu->arch.pvsched.base = INVALID_GPA; + val[0] = SMCCC_RET_SUCCESS; break; default: return kvm_psci_call(vcpu); diff --git a/arch/arm64/kvm/pvsched.c b/arch/arm64/kvm/pvsched.c index b923c4b6c52e..06290c831101 100644 --- a/arch/arm64/kvm/pvsched.c +++ b/arch/arm64/kvm/pvsched.c @@ -18,7 +18,7 @@ void kvm_update_pvsched_preempted(struct kvm_vcpu *vcpu, u32 preempted) u64 offset = offsetof(struct pvsched_vcpu_state, preempted); int idx;

- if (base == GPA_INVALID) + if (base == INVALID_GPA) return;

/* diff --git a/include/linux/cpuhotplug.h b/include/linux/cpuhotplug.h index 624d4a38c358..f94a1b8e34e0 100644 --- a/include/linux/cpuhotplug.h +++ b/include/linux/cpuhotplug.h @@ -190,6 +190,7 @@ enum cpuhp_state { CPUHP_AP_DUMMY_TIMER_STARTING, CPUHP_AP_ARM_XEN_STARTING, CPUHP_AP_ARM_XEN_RUNSTATE_STARTING, + CPUHP_AP_ARM_KVM_PVSCHED_STARTING, CPUHP_AP_ARM_CORESIGHT_STARTING, CPUHP_AP_ARM_CORESIGHT_CTI_STARTING, CPUHP_AP_ARM64_ISNDEP_STARTING,

From: Zengruan Ye yezengruan@huawei.com

virt inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8WMFU CVE: NA

--------------------------------

Implement the service call for waking up a WFI state vCPU.

Signed-off-by: Zengruan Ye yezengruan@huawei.com Reviewed-by: Zhanghailiang zhang.zhanghailiang@huawei.com Signed-off-by: Zheng Zengkai zhengzengkai@huawei.com

Signed-off-by: lishusen lishusen2@huawei.com --- arch/arm64/include/asm/kvm_host.h | 2 ++ arch/arm64/kvm/arm.c | 4 +++- arch/arm64/kvm/handle_exit.c | 1 + arch/arm64/kvm/hypercalls.c | 3 +++ arch/arm64/kvm/pvsched.c | 25 +++++++++++++++++++++++++ 5 files changed, 34 insertions(+), 1 deletion(-)

diff --git a/arch/arm64/include/asm/kvm_host.h b/arch/arm64/include/asm/kvm_host.h index 0537099a0bd0..e837b5fe0a97 100644 --- a/arch/arm64/include/asm/kvm_host.h +++ b/arch/arm64/include/asm/kvm_host.h @@ -598,6 +598,7 @@ struct kvm_vcpu_arch {

/* Guest PV sched state */ struct { + bool pv_unhalted; gpa_t base; } pvsched;

@@ -1058,6 +1059,7 @@ static inline bool kvm_arm_is_pvtime_enabled(struct kvm_vcpu_arch *vcpu_arch)

long kvm_hypercall_pvsched_features(struct kvm_vcpu *vcpu); void kvm_update_pvsched_preempted(struct kvm_vcpu *vcpu, u32 preempted); +long kvm_pvsched_kick_vcpu(struct kvm_vcpu *vcpu);

static inline void kvm_arm_pvsched_vcpu_init(struct kvm_vcpu_arch *vcpu_arch) { diff --git a/arch/arm64/kvm/arm.c b/arch/arm64/kvm/arm.c index 837ced945691..8c355fc14e9d 100644 --- a/arch/arm64/kvm/arm.c +++ b/arch/arm64/kvm/arm.c @@ -613,7 +613,9 @@ int kvm_arch_vcpu_ioctl_set_mpstate(struct kvm_vcpu *vcpu, int kvm_arch_vcpu_runnable(struct kvm_vcpu *v) { bool irq_lines = *vcpu_hcr(v) & (HCR_VI | HCR_VF); - return ((irq_lines || kvm_vgic_vcpu_pending_irq(v)) + bool pv_unhalted = v->arch.pvsched.pv_unhalted; + + return ((irq_lines || kvm_vgic_vcpu_pending_irq(v) || pv_unhalted) && !kvm_arm_vcpu_stopped(v) && !v->arch.pause); }

diff --git a/arch/arm64/kvm/handle_exit.c b/arch/arm64/kvm/handle_exit.c index 617ae6dea5d5..b9a44c3bebb7 100644 --- a/arch/arm64/kvm/handle_exit.c +++ b/arch/arm64/kvm/handle_exit.c @@ -121,6 +121,7 @@ static int kvm_handle_wfx(struct kvm_vcpu *vcpu) } else { trace_kvm_wfx_arm64(*vcpu_pc(vcpu), false); vcpu->stat.wfi_exit_stat++; + vcpu->arch.pvsched.pv_unhalted = false; }

if (esr & ESR_ELx_WFx_ISS_WFxT) { diff --git a/arch/arm64/kvm/hypercalls.c b/arch/arm64/kvm/hypercalls.c index a96b88d90cf8..46def0572461 100644 --- a/arch/arm64/kvm/hypercalls.c +++ b/arch/arm64/kvm/hypercalls.c @@ -377,6 +377,9 @@ int kvm_smccc_call_handler(struct kvm_vcpu *vcpu) vcpu->arch.pvsched.base = INVALID_GPA; val[0] = SMCCC_RET_SUCCESS; break; + case ARM_SMCCC_HV_PV_SCHED_KICK_CPU: + val[0] = kvm_pvsched_kick_vcpu(vcpu); + break; default: return kvm_psci_call(vcpu); } diff --git a/arch/arm64/kvm/pvsched.c b/arch/arm64/kvm/pvsched.c index 06290c831101..49d03fd6f4ad 100644 --- a/arch/arm64/kvm/pvsched.c +++ b/arch/arm64/kvm/pvsched.c @@ -34,6 +34,30 @@ void kvm_update_pvsched_preempted(struct kvm_vcpu *vcpu, u32 preempted) pagefault_enable(); }

+long kvm_pvsched_kick_vcpu(struct kvm_vcpu *vcpu) +{ + unsigned int vcpu_idx; + long val = SMCCC_RET_NOT_SUPPORTED; + struct kvm *kvm = vcpu->kvm; + struct kvm_vcpu *target = NULL; + + vcpu_idx = smccc_get_arg1(vcpu); + target = kvm_get_vcpu(kvm, vcpu_idx); + if (!target) + goto out; + + target->arch.pvsched.pv_unhalted = true; + kvm_make_request(KVM_REQ_IRQ_PENDING, target); + kvm_vcpu_kick(target); + if (READ_ONCE(target->ready)) + kvm_vcpu_yield_to(target); + + val = SMCCC_RET_SUCCESS; + +out: + return val; +} + long kvm_hypercall_pvsched_features(struct kvm_vcpu *vcpu) { u32 feature = smccc_get_arg1(vcpu); @@ -43,6 +67,7 @@ long kvm_hypercall_pvsched_features(struct kvm_vcpu *vcpu) case ARM_SMCCC_HV_PV_SCHED_FEATURES: case ARM_SMCCC_HV_PV_SCHED_IPA_INIT: case ARM_SMCCC_HV_PV_SCHED_IPA_RELEASE: + case ARM_SMCCC_HV_PV_SCHED_KICK_CPU: val = SMCCC_RET_SUCCESS; break; }

From: Zengruan Ye yezengruan@huawei.com

virt inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8WMFU CVE: NA

--------------------------------

Linux kernel builds were run in KVM guest on HiSilicon Kunpeng920 system. VM guests were set up with 32, 48 and 64 vCPUs on the 32 physical CPUs. The kernel build (make -j<n>) was done in a VM with unpinned vCPUs 3 times with the best time selected and <n> is the number of vCPUs available. The build times of the original linux 4.19.87, pvqspinlock with various number of vCPUs are as follows:

Kernel 32 vCPUs 48 vCPUs 60 vCPUs ---------- -------- -------- -------- 4.19.87 342.336s 602.048s 950.340s pvqsinlock 341.366s 376.135s 437.037s

Signed-off-by: Zengruan Ye yezengruan@huawei.com Reviewed-by: Zhanghailiang zhang.zhanghailiang@huawei.com Signed-off-by: Zheng Zengkai zhengzengkai@huawei.com

Signed-off-by: lishusen lishusen2@huawei.com --- arch/arm64/include/asm/paravirt.h | 5 +++ arch/arm64/kernel/paravirt.c | 60 +++++++++++++++++++++++++++++++ 2 files changed, 65 insertions(+)

diff --git a/arch/arm64/include/asm/paravirt.h b/arch/arm64/include/asm/paravirt.h index 1c9011d31562..f10552839663 100644 --- a/arch/arm64/include/asm/paravirt.h +++ b/arch/arm64/include/asm/paravirt.h @@ -31,6 +31,7 @@ static inline bool pv_vcpu_is_preempted(int cpu) }

#if defined(CONFIG_SMP) && defined(CONFIG_PARAVIRT_SPINLOCKS) +void __init pv_qspinlock_init(void); bool pv_is_native_spin_unlock(void); DECLARE_STATIC_CALL(pv_qspinlock_queued_spin_lock_slowpath, native_queued_spin_lock_slowpath); static inline void pv_queued_spin_lock_slowpath(struct qspinlock *lock, u32 val) @@ -53,6 +54,10 @@ static inline void pv_kick(int cpu) { return; } +#else + +#define pv_qspinlock_init() do {} while (0) + #endif /* SMP && PARAVIRT_SPINLOCKS */

#else diff --git a/arch/arm64/kernel/paravirt.c b/arch/arm64/kernel/paravirt.c index 844622baf6bf..19e264a553a9 100644 --- a/arch/arm64/kernel/paravirt.c +++ b/arch/arm64/kernel/paravirt.c @@ -23,6 +23,7 @@ #include <asm/paravirt.h> #include <asm/pvclock-abi.h> #include <asm/pvsched-abi.h> +#include <asm/qspinlock_paravirt.h> #include <asm/smp_plat.h>

struct static_key paravirt_steal_enabled; @@ -259,6 +260,63 @@ static bool has_kvm_pvsched(void) return (res.a0 == SMCCC_RET_SUCCESS); }

+#ifdef CONFIG_PARAVIRT_SPINLOCKS +static bool arm_pvspin; + +/* Kick a cpu by its cpuid. Used to wake up a halted vcpu */ +static void kvm_kick_cpu(int cpu) +{ + struct arm_smccc_res res; + + arm_smccc_1_1_invoke(ARM_SMCCC_HV_PV_SCHED_KICK_CPU, cpu, &res); +} + +static void kvm_wait(u8 *ptr, u8 val) +{ + unsigned long flags; + + if (in_nmi()) + return; + + local_irq_save(flags); + + if (READ_ONCE(*ptr) != val) + goto out; + + dsb(sy); + wfi(); + +out: + local_irq_restore(flags); +} + +void __init pv_qspinlock_init(void) +{ + /* Don't use the PV qspinlock code if there is only 1 vCPU. */ + if (num_possible_cpus() == 1) + arm_pvspin = false; + + if (!arm_pvspin) { + pr_info("PV qspinlocks disabled\n"); + return; + } + pr_info("PV qspinlocks enabled\n"); + + __pv_init_lock_hash(); + pv_ops.sched.queued_spin_lock_slowpath = __pv_queued_spin_lock_slowpath; + pv_ops.sched.queued_spin_unlock = __pv_queued_spin_unlock; + pv_ops.sched.wait = kvm_wait; + pv_ops.sched.kick = kvm_kick_cpu; +} + +static __init int arm_parse_pvspin(char *arg) +{ + arm_pvspin = true; + return 0; +} +early_param("arm_pvspin", arm_parse_pvspin); +#endif /* CONFIG_PARAVIRT_SPINLOCKS */ + int __init pv_sched_init(void) { int ret; @@ -278,5 +336,7 @@ int __init pv_sched_init(void) static_call_update(pv_vcpu_preempted, kvm_vcpu_is_preempted); pr_info("using PV sched preempted\n");

+ pv_qspinlock_init(); + return 0; }

From: yezengruan yezengruan@huawei.com

virt inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8WMFU CVE: NA

--------------------------------------------

1. Clear kvm_vcpu_arch::pvsched::base on vcpu reset The guest memory will otherwise be corrupted by KVM if we reboot into an old guest kernel which is **not** aware of the pvsched feature.

2. Fix boot cpu pvsched init abnormal The pv_sched_init was called too early in the boot in setup_arch, hence pvsched_vcpu_state was not initializedfo vcpu 0.

Signed-off-by: Zenghui Yu yuzenghui@huawei.com Signed-off-by: yezengruan yezengruan@huawei.com

Signed-off-by: lishusen lishusen2@huawei.com --- arch/arm64/kernel/paravirt.c | 1 + arch/arm64/kernel/setup.c | 2 -- arch/arm64/kvm/arm.c | 2 ++ 3 files changed, 3 insertions(+), 2 deletions(-)

diff --git a/arch/arm64/kernel/paravirt.c b/arch/arm64/kernel/paravirt.c index 6ad463cadd68..c5fc82372e47 100644 --- a/arch/arm64/kernel/paravirt.c +++ b/arch/arm64/kernel/paravirt.c @@ -346,3 +346,4 @@ int __init pv_sched_init(void)

return 0; } +early_initcall(pv_sched_init); diff --git a/arch/arm64/kernel/setup.c b/arch/arm64/kernel/setup.c index 0255788cb9c4..645034e52496 100644 --- a/arch/arm64/kernel/setup.c +++ b/arch/arm64/kernel/setup.c @@ -416,8 +416,6 @@ void __init __no_sanitize_address setup_arch(char **cmdline_p) smp_init_cpus(); smp_build_mpidr_hash();

- pv_sched_init(); - /* Init percpu seeds for random tags after cpus are set up. */ kasan_init_sw_tags();

diff --git a/arch/arm64/kvm/arm.c b/arch/arm64/kvm/arm.c index 8c355fc14e9d..1460c0648813 100644 --- a/arch/arm64/kvm/arm.c +++ b/arch/arm64/kvm/arm.c @@ -1402,6 +1402,8 @@ static int kvm_arch_vcpu_ioctl_vcpu_init(struct kvm_vcpu *vcpu,

spin_unlock(&vcpu->arch.mp_state_lock);

+ kvm_arm_pvsched_vcpu_init(&vcpu->arch); + return 0; }

virt inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8WMFU CVE: NA

Add configuration for feature pv-sched

Signed-off-by: lishusen lishusen2@huawei.com --- arch/arm64/Kconfig | 7 ++++ arch/arm64/include/asm/kvm_host.h | 4 +++ arch/arm64/include/asm/paravirt.h | 2 ++ arch/arm64/kernel/paravirt.c | 54 ++++++++++++++++--------------- arch/arm64/kvm/arm.c | 13 ++++++++ arch/arm64/kvm/handle_exit.c | 2 ++ arch/arm64/kvm/hypercalls.c | 4 +++ arch/arm64/kvm/pvsched.c | 2 ++ 8 files changed, 62 insertions(+), 26 deletions(-)

diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig index 971cf9d54101..145d85fc37fd 100644 --- a/arch/arm64/Kconfig +++ b/arch/arm64/Kconfig @@ -1561,6 +1561,13 @@ config PARAVIRT_SPINLOCKS

If you are unsure how to answer this question, answer Y.

+config PARAVIRT_SCHED + bool "Paravirtualization layer for sched" + depends on PARAVIRT + help + + If you are unsure how to answer this question, answer Y. + config PARAVIRT_TIME_ACCOUNTING bool "Paravirtual steal time accounting" select PARAVIRT diff --git a/arch/arm64/include/asm/kvm_host.h b/arch/arm64/include/asm/kvm_host.h index e837b5fe0a97..b0cb4f71754c 100644 --- a/arch/arm64/include/asm/kvm_host.h +++ b/arch/arm64/include/asm/kvm_host.h @@ -596,11 +596,13 @@ struct kvm_vcpu_arch { gpa_t base; } steal;

+#ifdef CONFIG_PARAVIRT_SCHED /* Guest PV sched state */ struct { bool pv_unhalted; gpa_t base; } pvsched; +#endif

/* Per-vcpu CCSIDR override or NULL */ u32 *ccsidr; @@ -1057,6 +1059,7 @@ static inline bool kvm_arm_is_pvtime_enabled(struct kvm_vcpu_arch *vcpu_arch) return (vcpu_arch->steal.base != INVALID_GPA); }

+#ifdef CONFIG_PARAVIRT_SCHED long kvm_hypercall_pvsched_features(struct kvm_vcpu *vcpu); void kvm_update_pvsched_preempted(struct kvm_vcpu *vcpu, u32 preempted); long kvm_pvsched_kick_vcpu(struct kvm_vcpu *vcpu); @@ -1070,6 +1073,7 @@ static inline bool kvm_arm_is_pvsched_enabled(struct kvm_vcpu_arch *vcpu_arch) { return (vcpu_arch->pvsched.base != INVALID_GPA); } +#endif

void kvm_set_sei_esr(struct kvm_vcpu *vcpu, u64 syndrome);

diff --git a/arch/arm64/include/asm/paravirt.h b/arch/arm64/include/asm/paravirt.h index a29eeffa49aa..1ce7187eed48 100644 --- a/arch/arm64/include/asm/paravirt.h +++ b/arch/arm64/include/asm/paravirt.h @@ -20,7 +20,9 @@ static inline u64 paravirt_steal_clock(int cpu)

int __init pv_time_init(void);

+#ifdef CONFIG_PARAVIRT_SCHED int __init pv_sched_init(void); +#endif

__visible bool __native_vcpu_is_preempted(int cpu); DECLARE_STATIC_CALL(pv_vcpu_preempted, __native_vcpu_is_preempted); diff --git a/arch/arm64/kernel/paravirt.c b/arch/arm64/kernel/paravirt.c index 73a8b9886775..126dff610e07 100644 --- a/arch/arm64/kernel/paravirt.c +++ b/arch/arm64/kernel/paravirt.c @@ -180,7 +180,7 @@ int __init pv_time_init(void) return 0; }

- +#ifdef CONFIG_PARAVIRT_SCHED DEFINE_PER_CPU(struct pvsched_vcpu_state, pvsched_vcpu_region) __aligned(64); EXPORT_PER_CPU_SYMBOL(pvsched_vcpu_region);

@@ -263,6 +263,33 @@ static bool has_kvm_pvsched(void) return (res.a0 == SMCCC_RET_SUCCESS); }

+int __init pv_sched_init(void) +{ + int ret; + + if (is_hyp_mode_available()) + return 0; + + if (!has_kvm_pvsched()) { + pr_warn("PV sched is not available\n"); + return 0; + } + + ret = kvm_arm_init_pvsched(); + if (ret) + return ret; + + static_call_update(pv_vcpu_preempted, kvm_vcpu_is_preempted); + pr_info("using PV sched preempted\n"); + + pv_qspinlock_init(); + + return 0; +} + +early_initcall(pv_sched_init); +#endif /* CONFIG_PARAVIRT_SCHED */ + #ifdef CONFIG_PARAVIRT_SPINLOCKS static bool arm_pvspin;

@@ -336,28 +363,3 @@ static __init int arm_parse_pvspin(char *arg) } early_param("arm_pvspin", arm_parse_pvspin); #endif /* CONFIG_PARAVIRT_SPINLOCKS */ - -int __init pv_sched_init(void) -{ - int ret; - - if (is_hyp_mode_available()) - return 0; - - if (!has_kvm_pvsched()) { - pr_warn("PV sched is not available\n"); - return 0; - } - - ret = kvm_arm_init_pvsched(); - if (ret) - return ret; - - static_call_update(pv_vcpu_preempted, kvm_vcpu_is_preempted); - pr_info("using PV sched preempted\n"); - - pv_qspinlock_init(); - - return 0; -} -early_initcall(pv_sched_init); diff --git a/arch/arm64/kvm/arm.c b/arch/arm64/kvm/arm.c index 1460c0648813..bbb1b7d32cc3 100644 --- a/arch/arm64/kvm/arm.c +++ b/arch/arm64/kvm/arm.c @@ -415,7 +415,9 @@ int kvm_arch_vcpu_create(struct kvm_vcpu *vcpu)

kvm_arm_pvtime_vcpu_init(&vcpu->arch);

+#ifdef CONFIG_PARAVIRT_SCHED kvm_arm_pvsched_vcpu_init(&vcpu->arch); +#endif

vcpu->arch.hw_mmu = &vcpu->kvm->arch.mmu;

@@ -508,8 +510,10 @@ void kvm_arch_vcpu_load(struct kvm_vcpu *vcpu, int cpu) if (!cpumask_test_cpu(cpu, vcpu->kvm->arch.supported_cpus)) vcpu_set_on_unsupported_cpu(vcpu);

+#ifdef CONFIG_PARAVIRT_SCHED if (kvm_arm_is_pvsched_enabled(&vcpu->arch)) kvm_update_pvsched_preempted(vcpu, 0); +#endif

#ifdef CONFIG_KVM_HISI_VIRT kvm_hisi_dvmbm_load(vcpu); @@ -530,8 +534,10 @@ void kvm_arch_vcpu_put(struct kvm_vcpu *vcpu) vcpu_clear_on_unsupported_cpu(vcpu); vcpu->cpu = -1;

+#ifdef CONFIG_PARAVIRT_SCHED if (kvm_arm_is_pvsched_enabled(&vcpu->arch)) kvm_update_pvsched_preempted(vcpu, 1); +#endif

#ifdef CONFIG_KVM_HISI_VIRT kvm_hisi_dvmbm_put(vcpu); @@ -613,10 +619,15 @@ int kvm_arch_vcpu_ioctl_set_mpstate(struct kvm_vcpu *vcpu, int kvm_arch_vcpu_runnable(struct kvm_vcpu *v) { bool irq_lines = *vcpu_hcr(v) & (HCR_VI | HCR_VF); +#ifdef CONFIG_PARAVIRT_SCHED bool pv_unhalted = v->arch.pvsched.pv_unhalted;

return ((irq_lines || kvm_vgic_vcpu_pending_irq(v) || pv_unhalted) && !kvm_arm_vcpu_stopped(v) && !v->arch.pause); +#else + return ((irq_lines || kvm_vgic_vcpu_pending_irq(v)) + && !kvm_arm_vcpu_stopped(v) && !v->arch.pause); +#endif }

bool kvm_arch_vcpu_in_kernel(struct kvm_vcpu *vcpu) @@ -1402,7 +1413,9 @@ static int kvm_arch_vcpu_ioctl_vcpu_init(struct kvm_vcpu *vcpu,

spin_unlock(&vcpu->arch.mp_state_lock);

+#ifdef CONFIG_PARAVIRT_SCHED kvm_arm_pvsched_vcpu_init(&vcpu->arch); +#endif

return 0; } diff --git a/arch/arm64/kvm/handle_exit.c b/arch/arm64/kvm/handle_exit.c index b9a44c3bebb7..f4671ae23b1f 100644 --- a/arch/arm64/kvm/handle_exit.c +++ b/arch/arm64/kvm/handle_exit.c @@ -121,7 +121,9 @@ static int kvm_handle_wfx(struct kvm_vcpu *vcpu) } else { trace_kvm_wfx_arm64(*vcpu_pc(vcpu), false); vcpu->stat.wfi_exit_stat++; +#ifdef CONFIG_PARAVIRT_SCHED vcpu->arch.pvsched.pv_unhalted = false; +#endif }

if (esr & ESR_ELx_WFx_ISS_WFxT) { diff --git a/arch/arm64/kvm/hypercalls.c b/arch/arm64/kvm/hypercalls.c index 46def0572461..2e7a1d06e6ff 100644 --- a/arch/arm64/kvm/hypercalls.c +++ b/arch/arm64/kvm/hypercalls.c @@ -332,9 +332,11 @@ int kvm_smccc_call_handler(struct kvm_vcpu *vcpu) &smccc_feat->std_hyp_bmap)) val[0] = SMCCC_RET_SUCCESS; break; +#ifdef CONFIG_PARAVIRT_SCHED case ARM_SMCCC_HV_PV_SCHED_FEATURES: val[0] = SMCCC_RET_SUCCESS; break; +#endif } break; case ARM_SMCCC_HV_PV_TIME_FEATURES: @@ -363,6 +365,7 @@ int kvm_smccc_call_handler(struct kvm_vcpu *vcpu) case ARM_SMCCC_TRNG_RND32: case ARM_SMCCC_TRNG_RND64: return kvm_trng_call(vcpu); +#ifdef CONFIG_PARAVIRT_SCHED case ARM_SMCCC_HV_PV_SCHED_FEATURES: val[0] = kvm_hypercall_pvsched_features(vcpu); break; @@ -380,6 +383,7 @@ int kvm_smccc_call_handler(struct kvm_vcpu *vcpu) case ARM_SMCCC_HV_PV_SCHED_KICK_CPU: val[0] = kvm_pvsched_kick_vcpu(vcpu); break; +#endif default: return kvm_psci_call(vcpu); } diff --git a/arch/arm64/kvm/pvsched.c b/arch/arm64/kvm/pvsched.c index bcfd4760b491..eae6ad7b6c5d 100644 --- a/arch/arm64/kvm/pvsched.c +++ b/arch/arm64/kvm/pvsched.c @@ -4,6 +4,7 @@ * Author: Zengruan Ye yezengruan@huawei.com */

+#ifdef CONFIG_PARAVIRT_SCHED #include <linux/arm-smccc.h> #include <linux/kvm_host.h>

@@ -77,3 +78,4 @@ long kvm_hypercall_pvsched_features(struct kvm_vcpu *vcpu)

return val; } +#endif /* CONFIG_PARAVIRT_SCHED */

From: Zenghui Yu yuzenghui@huawei.com

virt inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8WMFU CVE: NA

----------------------------------------------------

Currently fp_asimd_exit_stat is accumulated for *both* FP/ASIMD and SVE traps so that user can not distinguish between these two via debugfs.

Fix the manipulation for both exception classes.

Signed-off-by: Zenghui Yu yuzenghui@huawei.com Reviewed-by: Keqian Zhu zhukeqian1@huawei.com Signed-off-by: Zheng Zengkai zhengzengkai@huawei.com

Signed-off-by: lishusen lishusen2@huawei.com --- arch/arm64/kvm/hyp/include/hyp/switch.h | 1 + 1 file changed, 1 insertion(+)

diff --git a/arch/arm64/kvm/hyp/include/hyp/switch.h b/arch/arm64/kvm/hyp/include/hyp/switch.h index c91e641346d6..61907949e00f 100644 --- a/arch/arm64/kvm/hyp/include/hyp/switch.h +++ b/arch/arm64/kvm/hyp/include/hyp/switch.h @@ -297,6 +297,7 @@ static bool kvm_hyp_handle_fpsimd(struct kvm_vcpu *vcpu, u64 *exit_code) vcpu->stat.fp_asimd_exit_stat++; break; case ESR_ELx_EC_SVE: + vcpu->stat.sve_exit_stat++; if (!sve_guest) return false; break;