Merge tag 'loongarch-kvm-6.20' of git://git.kernel.org/pub/scm/linux/kernel/git/chenhuacai/linux-loongson into HEAD

LoongArch KVM changes for v6.20

1. Add more CPUCFG mask bits.
2. Improve feature detection.
3. Add FPU/LBT delay load support.
4. Set default return value in KVM IO bus ops.
5. Add paravirt preempt feature support.
6. Add KVM steal time test case for tools/selftests.

Author: bonzini

arch/loongarch/include/asm/kvm_host.h

@@ -37,6 +37,7 @@
 #define KVM_REQ_TLB_FLUSH_GPA		KVM_ARCH_REQ(0)
 #define KVM_REQ_STEAL_UPDATE		KVM_ARCH_REQ(1)
 #define KVM_REQ_PMU			KVM_ARCH_REQ(2)
+#define KVM_REQ_AUX_LOAD		KVM_ARCH_REQ(3)
 
 #define KVM_GUESTDBG_SW_BP_MASK		\
 	(KVM_GUESTDBG_ENABLE | KVM_GUESTDBG_USE_SW_BP)
@@ -164,6 +165,7 @@ enum emulation_result {
 
 #define LOONGARCH_PV_FEAT_UPDATED	BIT_ULL(63)
 #define LOONGARCH_PV_FEAT_MASK		(BIT(KVM_FEATURE_IPI) |		\
+					 BIT(KVM_FEATURE_PREEMPT) |	\
 					 BIT(KVM_FEATURE_STEAL_TIME) |	\
 					 BIT(KVM_FEATURE_USER_HCALL) |	\
 					 BIT(KVM_FEATURE_VIRT_EXTIOI))
@@ -200,6 +202,7 @@ struct kvm_vcpu_arch {
 
 	/* Which auxiliary state is loaded (KVM_LARCH_*) */
 	unsigned int aux_inuse;
+	unsigned int aux_ldtype;
 
 	/* FPU state */
 	struct loongarch_fpu fpu FPU_ALIGN;
@@ -252,6 +255,7 @@ struct kvm_vcpu_arch {
 		u64 guest_addr;
 		u64 last_steal;
 		struct gfn_to_hva_cache cache;
+		u8  preempted;
 	} st;
 };
 
@@ -265,6 +269,11 @@ static inline void writel_sw_gcsr(struct loongarch_csrs *csr, int reg, unsigned
 	csr->csrs[reg] = val;
 }
 
+static inline bool kvm_guest_has_msgint(struct kvm_vcpu_arch *arch)
+{
+	return arch->cpucfg[1] & CPUCFG1_MSGINT;
+}
+
 static inline bool kvm_guest_has_fpu(struct kvm_vcpu_arch *arch)
 {
 	return arch->cpucfg[2] & CPUCFG2_FP;

arch/loongarch/include/asm/kvm_para.h

@@ -37,8 +37,10 @@ struct kvm_steal_time {
 	__u64 steal;
 	__u32 version;
 	__u32 flags;
-	__u32 pad[12];
+	__u8  preempted;
+	__u8  pad[47];
 };
+#define KVM_VCPU_PREEMPTED		(1 << 0)
 
 /*
  * Hypercall interface for KVM hypervisor

arch/loongarch/include/asm/loongarch.h

@@ -690,6 +690,7 @@
 #define LOONGARCH_CSR_ISR3		0xa3
 
 #define LOONGARCH_CSR_IRR		0xa4
+#define LOONGARCH_CSR_IPR		0xa5
 
 #define LOONGARCH_CSR_PRID		0xc0
 

arch/loongarch/include/asm/qspinlock.h

@@ -34,6 +34,10 @@ static inline bool virt_spin_lock(struct qspinlock *lock)
 	return true;
 }
 
+#define vcpu_is_preempted vcpu_is_preempted
+
+bool vcpu_is_preempted(int cpu);
+
 #endif /* CONFIG_PARAVIRT */
 
 #include <asm-generic/qspinlock.h>

arch/loongarch/include/uapi/asm/kvm.h

@@ -105,6 +105,7 @@ struct kvm_fpu {
 #define  KVM_LOONGARCH_VM_FEAT_PV_STEALTIME	7
 #define  KVM_LOONGARCH_VM_FEAT_PTW		8
 #define  KVM_LOONGARCH_VM_FEAT_MSGINT		9
+#define  KVM_LOONGARCH_VM_FEAT_PV_PREEMPT	10
 
 /* Device Control API on vcpu fd */
 #define KVM_LOONGARCH_VCPU_CPUCFG	0

arch/loongarch/include/uapi/asm/kvm_para.h

@@ -15,6 +15,7 @@
 #define CPUCFG_KVM_FEATURE		(CPUCFG_KVM_BASE + 4)
 #define  KVM_FEATURE_IPI		1
 #define  KVM_FEATURE_STEAL_TIME		2
+#define  KVM_FEATURE_PREEMPT		3
 /* BIT 24 - 31 are features configurable by user space vmm */
 #define  KVM_FEATURE_VIRT_EXTIOI	24
 #define  KVM_FEATURE_USER_HCALL		25

arch/loongarch/kernel/paravirt.c

@@ -12,6 +12,7 @@ static int has_steal_clock;
 struct static_key paravirt_steal_enabled;
 struct static_key paravirt_steal_rq_enabled;
 static DEFINE_PER_CPU(struct kvm_steal_time, steal_time) __aligned(64);
+static DEFINE_STATIC_KEY_FALSE(virt_preempt_key);
 DEFINE_STATIC_KEY_FALSE(virt_spin_lock_key);
 
 static u64 native_steal_clock(int cpu)
@@ -267,6 +268,18 @@ static int pv_time_cpu_down_prepare(unsigned int cpu)
 
 	return 0;
 }
+
+bool vcpu_is_preempted(int cpu)
+{
+	struct kvm_steal_time *src;
+
+	if (!static_branch_unlikely(&virt_preempt_key))
+		return false;
+
+	src = &per_cpu(steal_time, cpu);
+	return !!(src->preempted & KVM_VCPU_PREEMPTED);
+}
+EXPORT_SYMBOL(vcpu_is_preempted);
 #endif
 
 static void pv_cpu_reboot(void *unused)
@@ -308,6 +321,9 @@ int __init pv_time_init(void)
 		pr_err("Failed to install cpu hotplug callbacks\n");
 		return r;
 	}
+
+	if (kvm_para_has_feature(KVM_FEATURE_PREEMPT))
+		static_branch_enable(&virt_preempt_key);
 #endif
 
 	static_call_update(pv_steal_clock, paravt_steal_clock);
@@ -318,7 +334,10 @@ int __init pv_time_init(void)
 		static_key_slow_inc(&paravirt_steal_rq_enabled);
 #endif
 
-	pr_info("Using paravirt steal-time\n");
+	if (static_key_enabled(&virt_preempt_key))
+		pr_info("Using paravirt steal-time with preempt enabled\n");
+	else
+		pr_info("Using paravirt steal-time with preempt disabled\n");
 
 	return 0;
 }

arch/loongarch/kvm/exit.c

@@ -754,7 +754,8 @@ static int kvm_handle_fpu_disabled(struct kvm_vcpu *vcpu, int ecode)
 		return RESUME_HOST;
 	}
 
-	kvm_own_fpu(vcpu);
+	vcpu->arch.aux_ldtype = KVM_LARCH_FPU;
+	kvm_make_request(KVM_REQ_AUX_LOAD, vcpu);
 
 	return RESUME_GUEST;
 }
@@ -792,8 +793,12 @@ static long kvm_save_notify(struct kvm_vcpu *vcpu)
  */
 static int kvm_handle_lsx_disabled(struct kvm_vcpu *vcpu, int ecode)
 {
-	if (kvm_own_lsx(vcpu))
+	if (!kvm_guest_has_lsx(&vcpu->arch))
 		kvm_queue_exception(vcpu, EXCCODE_INE, 0);
+	else {
+		vcpu->arch.aux_ldtype = KVM_LARCH_LSX;
+		kvm_make_request(KVM_REQ_AUX_LOAD, vcpu);
+	}
 
 	return RESUME_GUEST;
 }
@@ -808,16 +813,24 @@ static int kvm_handle_lsx_disabled(struct kvm_vcpu *vcpu, int ecode)
  */
 static int kvm_handle_lasx_disabled(struct kvm_vcpu *vcpu, int ecode)
 {
-	if (kvm_own_lasx(vcpu))
+	if (!kvm_guest_has_lasx(&vcpu->arch))
 		kvm_queue_exception(vcpu, EXCCODE_INE, 0);
+	else {
+		vcpu->arch.aux_ldtype = KVM_LARCH_LASX;
+		kvm_make_request(KVM_REQ_AUX_LOAD, vcpu);
+	}
 
 	return RESUME_GUEST;
 }
 
 static int kvm_handle_lbt_disabled(struct kvm_vcpu *vcpu, int ecode)
 {
-	if (kvm_own_lbt(vcpu))
+	if (!kvm_guest_has_lbt(&vcpu->arch))
 		kvm_queue_exception(vcpu, EXCCODE_INE, 0);
+	else {
+		vcpu->arch.aux_ldtype = KVM_LARCH_LBT;
+		kvm_make_request(KVM_REQ_AUX_LOAD, vcpu);
+	}
 
 	return RESUME_GUEST;
 }

arch/loongarch/kvm/intc/eiointc.c

@@ -119,7 +119,7 @@ void eiointc_set_irq(struct loongarch_eiointc *s, int irq, int level)
 static int loongarch_eiointc_read(struct kvm_vcpu *vcpu, struct loongarch_eiointc *s,
 				gpa_t addr, unsigned long *val)
 {
-	int index, ret = 0;
+	int index;
 	u64 data = 0;
 	gpa_t offset;
 
@@ -150,40 +150,36 @@ static int loongarch_eiointc_read(struct kvm_vcpu *vcpu, struct loongarch_eioint
 		data = s->coremap[index];
 		break;
 	default:
-		ret = -EINVAL;
 		break;
 	}
 	*val = data;
 
-	return ret;
+	return 0;
 }
 
 static int kvm_eiointc_read(struct kvm_vcpu *vcpu,
 			struct kvm_io_device *dev,
 			gpa_t addr, int len, void *val)
 {
-	int ret = -EINVAL;
 	unsigned long flags, data, offset;
 	struct loongarch_eiointc *eiointc = vcpu->kvm->arch.eiointc;
 
 	if (!eiointc) {
 		kvm_err("%s: eiointc irqchip not valid!\n", __func__);
-		return -EINVAL;
+		return 0;
 	}
 
 	if (addr & (len - 1)) {
 		kvm_err("%s: eiointc not aligned addr %llx len %d\n", __func__, addr, len);
-		return -EINVAL;
+		return 0;
 	}
 
 	offset = addr & 0x7;
 	addr -= offset;
 	vcpu->stat.eiointc_read_exits++;
 	spin_lock_irqsave(&eiointc->lock, flags);
-	ret = loongarch_eiointc_read(vcpu, eiointc, addr, &data);
+	loongarch_eiointc_read(vcpu, eiointc, addr, &data);
 	spin_unlock_irqrestore(&eiointc->lock, flags);
-	if (ret)
-		return ret;
 
 	data = data >> (offset * 8);
 	switch (len) {
@@ -208,7 +204,7 @@ static int loongarch_eiointc_write(struct kvm_vcpu *vcpu,
 				struct loongarch_eiointc *s,
 				gpa_t addr, u64 value, u64 field_mask)
 {
-	int index, irq, ret = 0;
+	int index, irq;
 	u8 cpu;
 	u64 data, old, mask;
 	gpa_t offset;
@@ -287,54 +283,52 @@ static int loongarch_eiointc_write(struct kvm_vcpu *vcpu,
 		eiointc_update_sw_coremap(s, index * 8, data, sizeof(data), true);
 		break;
 	default:
-		ret = -EINVAL;
 		break;
 	}
 
-	return ret;
+	return 0;
 }
 
 static int kvm_eiointc_write(struct kvm_vcpu *vcpu,
 			struct kvm_io_device *dev,
 			gpa_t addr, int len, const void *val)
 {
-	int ret = -EINVAL;
 	unsigned long flags, value;
 	struct loongarch_eiointc *eiointc = vcpu->kvm->arch.eiointc;
 
 	if (!eiointc) {
 		kvm_err("%s: eiointc irqchip not valid!\n", __func__);
-		return -EINVAL;
+		return 0;
 	}
 
 	if (addr & (len - 1)) {
 		kvm_err("%s: eiointc not aligned addr %llx len %d\n", __func__, addr, len);
-		return -EINVAL;
+		return 0;
 	}
 
 	vcpu->stat.eiointc_write_exits++;
 	spin_lock_irqsave(&eiointc->lock, flags);
 	switch (len) {
 	case 1:
 		value = *(unsigned char *)val;
-		ret = loongarch_eiointc_write(vcpu, eiointc, addr, value, 0xFF);
+		loongarch_eiointc_write(vcpu, eiointc, addr, value, 0xFF);
 		break;
 	case 2:
 		value = *(unsigned short *)val;
-		ret = loongarch_eiointc_write(vcpu, eiointc, addr, value, USHRT_MAX);
+		loongarch_eiointc_write(vcpu, eiointc, addr, value, USHRT_MAX);
 		break;
 	case 4:
 		value = *(unsigned int *)val;
-		ret = loongarch_eiointc_write(vcpu, eiointc, addr, value, UINT_MAX);
+		loongarch_eiointc_write(vcpu, eiointc, addr, value, UINT_MAX);
 		break;
 	default:
 		value = *(unsigned long *)val;
-		ret = loongarch_eiointc_write(vcpu, eiointc, addr, value, ULONG_MAX);
+		loongarch_eiointc_write(vcpu, eiointc, addr, value, ULONG_MAX);
 		break;
 	}
 	spin_unlock_irqrestore(&eiointc->lock, flags);
 
-	return ret;
+	return 0;
 }
 
 static const struct kvm_io_device_ops kvm_eiointc_ops = {
@@ -352,7 +346,7 @@ static int kvm_eiointc_virt_read(struct kvm_vcpu *vcpu,
 
 	if (!eiointc) {
 		kvm_err("%s: eiointc irqchip not valid!\n", __func__);
-		return -EINVAL;
+		return 0;
 	}
 
 	addr -= EIOINTC_VIRT_BASE;
@@ -376,28 +370,25 @@ static int kvm_eiointc_virt_write(struct kvm_vcpu *vcpu,
 				struct kvm_io_device *dev,
 				gpa_t addr, int len, const void *val)
 {
-	int ret = 0;
 	unsigned long flags;
 	u32 value = *(u32 *)val;
 	struct loongarch_eiointc *eiointc = vcpu->kvm->arch.eiointc;
 
 	if (!eiointc) {
 		kvm_err("%s: eiointc irqchip not valid!\n", __func__);
-		return -EINVAL;
+		return 0;
 	}
 
 	addr -= EIOINTC_VIRT_BASE;
 	spin_lock_irqsave(&eiointc->lock, flags);
 	switch (addr) {
 	case EIOINTC_VIRT_FEATURES:
-		ret = -EPERM;
 		break;
 	case EIOINTC_VIRT_CONFIG:
 		/*
 		 * eiointc features can only be set at disabled status
 		 */
 		if ((eiointc->status & BIT(EIOINTC_ENABLE)) && value) {
-			ret = -EPERM;
 			break;
 		}
 		eiointc->status = value & eiointc->features;
@@ -407,7 +398,7 @@ static int kvm_eiointc_virt_write(struct kvm_vcpu *vcpu,
 	}
 	spin_unlock_irqrestore(&eiointc->lock, flags);
 
-	return ret;
+	return 0;
 }
 
 static const struct kvm_io_device_ops kvm_eiointc_virt_ops = {