Merge tag 'kvm-s390-next-4.6-2' of...

Merge tag 'kvm-s390-next-4.6-2' of git://git.kernel.org/pub/scm/linux/kernel/git/kvms390/linux into HEAD

KVM: s390: Fixes and features for kvm/next (4.6) part 2

- add watchdog diagnose to trace event decoder
- better handle the cpu timer when not inside the guest
- only provide STFLE if the CPU model has STFLE
- reduce DMA page usage

arch/s390/include/asm/kvm_host.h

@@ -20,6 +20,7 @@
 #include <linux/kvm_types.h>
 #include <linux/kvm_host.h>
 #include <linux/kvm.h>
+#include <linux/seqlock.h>
 #include <asm/debug.h>
 #include <asm/cpu.h>
 #include <asm/fpu/api.h>
@@ -552,6 +553,15 @@ struct kvm_vcpu_arch {
 	unsigned long pfault_token;
 	unsigned long pfault_select;
 	unsigned long pfault_compare;
+	bool cputm_enabled;
+	/*
+	 * The seqcount protects updates to cputm_start and sie_block.cputm,
+	 * this way we can have non-blocking reads with consistent values.
+	 * Only the owning VCPU thread (vcpu->cpu) is allowed to change these
+	 * values and to start/stop/enable/disable cpu timer accounting.
+	 */
+	seqcount_t cputm_seqcount;
+	__u64 cputm_start;
 };
 
 struct kvm_vm_stat {
@@ -590,15 +600,11 @@ struct s390_io_adapter {
 #define S390_ARCH_FAC_MASK_SIZE_U64 \
 	(S390_ARCH_FAC_MASK_SIZE_BYTE / sizeof(u64))
 
-struct kvm_s390_fac {
-	/* facility list requested by guest */
-	__u64 list[S390_ARCH_FAC_LIST_SIZE_U64];
-	/* facility mask supported by kvm & hosting machine */
-	__u64 mask[S390_ARCH_FAC_LIST_SIZE_U64];
-};
-
 struct kvm_s390_cpu_model {
-	struct kvm_s390_fac *fac;
+	/* facility mask supported by kvm & hosting machine */
+	__u64 fac_mask[S390_ARCH_FAC_LIST_SIZE_U64];
+	/* facility list requested by guest (in dma page) */
+	__u64 *fac_list;
 	struct cpuid cpu_id;
 	unsigned short ibc;
 };
@@ -617,6 +623,16 @@ struct kvm_s390_crypto_cb {
 	__u8    reserved80[128];                /* 0x0080 */
 };
 
+/*
+ * sie_page2 has to be allocated as DMA because fac_list and crycb need
+ * 31bit addresses in the sie control block.
+ */
+struct sie_page2 {
+	__u64 fac_list[S390_ARCH_FAC_LIST_SIZE_U64];	/* 0x0000 */
+	struct kvm_s390_crypto_cb crycb;		/* 0x0800 */
+	u8 reserved900[0x1000 - 0x900];			/* 0x0900 */
+} __packed;
+
 struct kvm_arch{
 	void *sca;
 	int use_esca;
@@ -637,6 +653,7 @@ struct kvm_arch{
 	int ipte_lock_count;
 	struct mutex ipte_mutex;
 	spinlock_t start_stop_lock;
+	struct sie_page2 *sie_page2;
 	struct kvm_s390_cpu_model model;
 	struct kvm_s390_crypto crypto;
 	u64 epoch;

arch/s390/include/uapi/asm/sie.h

@@ -7,6 +7,7 @@
 	{ 0x9c, "DIAG (0x9c) time slice end directed" },	\
 	{ 0x204, "DIAG (0x204) logical-cpu utilization" },	\
 	{ 0x258, "DIAG (0x258) page-reference services" },	\
+	{ 0x288, "DIAG (0x288) watchdog functions" },		\
 	{ 0x308, "DIAG (0x308) ipl functions" },		\
 	{ 0x500, "DIAG (0x500) KVM virtio functions" },		\
 	{ 0x501, "DIAG (0x501) KVM breakpoint" }

arch/s390/kvm/interrupt.c

@@ -182,8 +182,9 @@ static int cpu_timer_interrupts_enabled(struct kvm_vcpu *vcpu)
 
 static int cpu_timer_irq_pending(struct kvm_vcpu *vcpu)
 {
-	return (vcpu->arch.sie_block->cputm >> 63) &&
-	       cpu_timer_interrupts_enabled(vcpu);
+	if (!cpu_timer_interrupts_enabled(vcpu))
+		return 0;
+	return kvm_s390_get_cpu_timer(vcpu) >> 63;
 }
 
 static inline int is_ioirq(unsigned long irq_type)
@@ -908,9 +909,35 @@ int kvm_cpu_has_pending_timer(struct kvm_vcpu *vcpu)
 	return ckc_irq_pending(vcpu) || cpu_timer_irq_pending(vcpu);
 }
 
+static u64 __calculate_sltime(struct kvm_vcpu *vcpu)
+{
+	u64 now, cputm, sltime = 0;
+
+	if (ckc_interrupts_enabled(vcpu)) {
+		now = kvm_s390_get_tod_clock_fast(vcpu->kvm);
+		sltime = tod_to_ns(vcpu->arch.sie_block->ckc - now);
+		/* already expired or overflow? */
+		if (!sltime || vcpu->arch.sie_block->ckc <= now)
+			return 0;
+		if (cpu_timer_interrupts_enabled(vcpu)) {
+			cputm = kvm_s390_get_cpu_timer(vcpu);
+			/* already expired? */
+			if (cputm >> 63)
+				return 0;
+			return min(sltime, tod_to_ns(cputm));
+		}
+	} else if (cpu_timer_interrupts_enabled(vcpu)) {
+		sltime = kvm_s390_get_cpu_timer(vcpu);
+		/* already expired? */
+		if (sltime >> 63)
+			return 0;
+	}
+	return sltime;
+}
+
 int kvm_s390_handle_wait(struct kvm_vcpu *vcpu)
 {
-	u64 now, sltime;
+	u64 sltime;
 
 	vcpu->stat.exit_wait_state++;
 
@@ -923,22 +950,20 @@ int kvm_s390_handle_wait(struct kvm_vcpu *vcpu)
 		return -EOPNOTSUPP; /* disabled wait */
 	}
 
-	if (!ckc_interrupts_enabled(vcpu)) {
+	if (!ckc_interrupts_enabled(vcpu) &&
+	    !cpu_timer_interrupts_enabled(vcpu)) {
 		VCPU_EVENT(vcpu, 3, "%s", "enabled wait w/o timer");
 		__set_cpu_idle(vcpu);
 		goto no_timer;
 	}
 
-	now = kvm_s390_get_tod_clock_fast(vcpu->kvm);
-	sltime = tod_to_ns(vcpu->arch.sie_block->ckc - now);
-
-	/* underflow */
-	if (vcpu->arch.sie_block->ckc < now)
+	sltime = __calculate_sltime(vcpu);
+	if (!sltime)
 		return 0;
 
 	__set_cpu_idle(vcpu);
 	hrtimer_start(&vcpu->arch.ckc_timer, ktime_set (0, sltime) , HRTIMER_MODE_REL);
-	VCPU_EVENT(vcpu, 4, "enabled wait via clock comparator: %llu ns", sltime);
+	VCPU_EVENT(vcpu, 4, "enabled wait: %llu ns", sltime);
 no_timer:
 	srcu_read_unlock(&vcpu->kvm->srcu, vcpu->srcu_idx);
 	kvm_vcpu_block(vcpu);
@@ -965,18 +990,16 @@ void kvm_s390_vcpu_wakeup(struct kvm_vcpu *vcpu)
 enum hrtimer_restart kvm_s390_idle_wakeup(struct hrtimer *timer)
 {
 	struct kvm_vcpu *vcpu;
-	u64 now, sltime;
+	u64 sltime;
 
 	vcpu = container_of(timer, struct kvm_vcpu, arch.ckc_timer);
-	now = kvm_s390_get_tod_clock_fast(vcpu->kvm);
-	sltime = tod_to_ns(vcpu->arch.sie_block->ckc - now);
+	sltime = __calculate_sltime(vcpu);
 
 	/*
 	 * If the monotonic clock runs faster than the tod clock we might be
 	 * woken up too early and have to go back to sleep to avoid deadlocks.
 	 */
-	if (vcpu->arch.sie_block->ckc > now &&
-	    hrtimer_forward_now(timer, ns_to_ktime(sltime)))
+	if (sltime && hrtimer_forward_now(timer, ns_to_ktime(sltime)))
 		return HRTIMER_RESTART;
 	kvm_s390_vcpu_wakeup(vcpu);
 	return HRTIMER_NORESTART;

arch/s390/kvm/kvm-s390.h

@@ -54,6 +54,11 @@ static inline int is_vcpu_stopped(struct kvm_vcpu *vcpu)
 	return atomic_read(&vcpu->arch.sie_block->cpuflags) & CPUSTAT_STOPPED;
 }
 
+static inline int is_vcpu_idle(struct kvm_vcpu *vcpu)
+{
+	return atomic_read(&vcpu->arch.sie_block->cpuflags) & CPUSTAT_WAIT;
+}
+
 static inline int kvm_is_ucontrol(struct kvm *kvm)
 {
 #ifdef CONFIG_KVM_S390_UCONTROL
@@ -155,8 +160,8 @@ static inline void kvm_s390_set_psw_cc(struct kvm_vcpu *vcpu, unsigned long cc)
 /* test availability of facility in a kvm instance */
 static inline int test_kvm_facility(struct kvm *kvm, unsigned long nr)
 {
-	return __test_facility(nr, kvm->arch.model.fac->mask) &&
-		__test_facility(nr, kvm->arch.model.fac->list);
+	return __test_facility(nr, kvm->arch.model.fac_mask) &&
+		__test_facility(nr, kvm->arch.model.fac_list);
 }
 
 static inline int set_kvm_facility(u64 *fac_list, unsigned long nr)
@@ -263,6 +268,8 @@ int kvm_s390_vcpu_setup_cmma(struct kvm_vcpu *vcpu);
 void kvm_s390_vcpu_unsetup_cmma(struct kvm_vcpu *vcpu);
 unsigned long kvm_s390_fac_list_mask_size(void);
 extern unsigned long kvm_s390_fac_list_mask[];
+void kvm_s390_set_cpu_timer(struct kvm_vcpu *vcpu, __u64 cputm);
+__u64 kvm_s390_get_cpu_timer(struct kvm_vcpu *vcpu);
 
 /* implemented in diag.c */
 int kvm_s390_handle_diag(struct kvm_vcpu *vcpu);

arch/s390/kvm/priv.c

@@ -354,7 +354,7 @@ static int handle_stfl(struct kvm_vcpu *vcpu)
 	 * We need to shift the lower 32 facility bits (bit 0-31) from a u64
 	 * into a u32 memory representation. They will remain bits 0-31.
 	 */
-	fac = *vcpu->kvm->arch.model.fac->list >> 32;
+	fac = *vcpu->kvm->arch.model.fac_list >> 32;
 	rc = write_guest_lc(vcpu, offsetof(struct lowcore, stfl_fac_list),
 			    &fac, sizeof(fac));
 	if (rc)