KVM: PPC: Allow book3s_hv guests to use SMT processor modes

This lifts the restriction that book3s_hv guests can only run one
hardware thread per core, and allows them to use up to 4 threads
per core on POWER7.  The host still has to run single-threaded.

This capability is advertised to qemu through a new KVM_CAP_PPC_SMT
capability.  The return value of the ioctl querying this capability
is the number of vcpus per virtual CPU core (vcore), currently 4.

To use this, the host kernel should be booted with all threads
active, and then all the secondary threads should be offlined.
This will put the secondary threads into nap mode.  KVM will then
wake them from nap mode and use them for running guest code (while
they are still offline).  To wake the secondary threads, we send
them an IPI using a new xics_wake_cpu() function, implemented in
arch/powerpc/sysdev/xics/icp-native.c.  In other words, at this stage
we assume that the platform has a XICS interrupt controller and
we are using icp-native.c to drive it.  Since the woken thread will
need to acknowledge and clear the IPI, we also export the base
physical address of the XICS registers using kvmppc_set_xics_phys()
for use in the low-level KVM book3s code.

When a vcpu is created, it is assigned to a virtual CPU core.
The vcore number is obtained by dividing the vcpu number by the
number of threads per core in the host.  This number is exported
to userspace via the KVM_CAP_PPC_SMT capability.  If qemu wishes
to run the guest in single-threaded mode, it should make all vcpu
numbers be multiples of the number of threads per core.

We distinguish three states of a vcpu: runnable (i.e., ready to execute
the guest), blocked (that is, idle), and busy in host.  We currently
implement a policy that the vcore can run only when all its threads
are runnable or blocked.  This way, if a vcpu needs to execute elsewhere
in the kernel or in qemu, it can do so without being starved of CPU
by the other vcpus.

When a vcore starts to run, it executes in the context of one of the
vcpu threads.  The other vcpu threads all go to sleep and stay asleep
until something happens requiring the vcpu thread to return to qemu,
or to wake up to run the vcore (this can happen when another vcpu
thread goes from busy in host state to blocked).

It can happen that a vcpu goes from blocked to runnable state (e.g.
because of an interrupt), and the vcore it belongs to is already
running.  In that case it can start to run immediately as long as
the none of the vcpus in the vcore have started to exit the guest.
We send the next free thread in the vcore an IPI to get it to start
to execute the guest.  It synchronizes with the other threads via
the vcore->entry_exit_count field to make sure that it doesn't go
into the guest if the other vcpus are exiting by the time that it
is ready to actually enter the guest.

Note that there is no fixed relationship between the hardware thread
number and the vcpu number.  Hardware threads are assigned to vcpus
as they become runnable, so we will always use the lower-numbered
hardware threads in preference to higher-numbered threads if not all
the vcpus in the vcore are runnable, regardless of which vcpus are
runnable.

Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>

1 files changed, 289 insertions, 27 deletions

diff --git a/arch/powerpc/kvm/book3s_hv.c b/arch/powerpc/kvm/book3s_hv.c
index 6fe469eabce8..36b6d98f1197 100644
--- a/arch/powerpc/kvm/book3s_hv.c
+++ b/arch/powerpc/kvm/book3s_hv.c
@@ -39,6 +39,7 @@
 #include <asm/mmu_context.h>
 #include <asm/lppaca.h>
 #include <asm/processor.h>
+#include <asm/cputhreads.h>
 #include <linux/gfp.h>
 #include <linux/sched.h>
 #include <linux/vmalloc.h>
@@ -51,12 +52,16 @@
 void kvmppc_core_vcpu_load(struct kvm_vcpu *vcpu, int cpu)
 {
 	local_paca->kvm_hstate.kvm_vcpu = vcpu;
+	local_paca->kvm_hstate.kvm_vcore = vcpu->arch.vcore;
 }
 
 void kvmppc_core_vcpu_put(struct kvm_vcpu *vcpu)
 {
 }
 
+static void kvmppc_vcpu_blocked(struct kvm_vcpu *vcpu);
+static void kvmppc_vcpu_unblocked(struct kvm_vcpu *vcpu);
+
 void kvmppc_vcpu_block(struct kvm_vcpu *vcpu)
 {
 	u64 now;
@@ -74,11 +79,15 @@ void kvmppc_vcpu_block(struct kvm_vcpu *vcpu)
 			      HRTIMER_MODE_REL);
 	}
 
+	kvmppc_vcpu_blocked(vcpu);
+
 	kvm_vcpu_block(vcpu);
 	vcpu->stat.halt_wakeup++;
 
 	if (vcpu->arch.dec_expires != ~(u64)0)
 		hrtimer_try_to_cancel(&vcpu->arch.dec_timer);
+
+	kvmppc_vcpu_unblocked(vcpu);
 }
 
 void kvmppc_set_msr(struct kvm_vcpu *vcpu, u64 msr)
@@ -429,9 +438,16 @@ int kvmppc_core_check_processor_compat(void)
 struct kvm_vcpu *kvmppc_core_vcpu_create(struct kvm *kvm, unsigned int id)
 {
 	struct kvm_vcpu *vcpu;
-	int err = -ENOMEM;
+	int err = -EINVAL;
+	int core;
+	struct kvmppc_vcore *vcore;
 	unsigned long lpcr;
 
+	core = id / threads_per_core;
+	if (core >= KVM_MAX_VCORES)
+		goto out;
+
+	err = -ENOMEM;
 	vcpu = kzalloc(sizeof(struct kvm_vcpu), GFP_KERNEL);
 	if (!vcpu)
 		goto out;
@@ -454,6 +470,38 @@ struct kvm_vcpu *kvmppc_core_vcpu_create(struct kvm *kvm, unsigned int id)
 
 	kvmppc_mmu_book3s_hv_init(vcpu);
 
+	/*
+	 * Some vcpus may start out in stopped state.  If we initialize
+	 * them to busy-in-host state they will stop other vcpus in the
+	 * vcore from running.  Instead we initialize them to blocked
+	 * state, effectively considering them to be stopped until we
+	 * see the first run ioctl for them.
+	 */
+	vcpu->arch.state = KVMPPC_VCPU_BLOCKED;
+
+	init_waitqueue_head(&vcpu->arch.cpu_run);
+
+	mutex_lock(&kvm->lock);
+	vcore = kvm->arch.vcores[core];
+	if (!vcore) {
+		vcore = kzalloc(sizeof(struct kvmppc_vcore), GFP_KERNEL);
+		if (vcore) {
+			INIT_LIST_HEAD(&vcore->runnable_threads);
+			spin_lock_init(&vcore->lock);
+		}
+		kvm->arch.vcores[core] = vcore;
+	}
+	mutex_unlock(&kvm->lock);
+
+	if (!vcore)
+		goto free_vcpu;
+
+	spin_lock(&vcore->lock);
+	++vcore->num_threads;
+	++vcore->n_blocked;
+	spin_unlock(&vcore->lock);
+	vcpu->arch.vcore = vcore;
+
 	return vcpu;
 
 free_vcpu:
@@ -468,21 +516,121 @@ void kvmppc_core_vcpu_free(struct kvm_vcpu *vcpu)
 	kfree(vcpu);
 }
 
+static void kvmppc_vcpu_blocked(struct kvm_vcpu *vcpu)
+{
+	struct kvmppc_vcore *vc = vcpu->arch.vcore;
+
+	spin_lock(&vc->lock);
+	vcpu->arch.state = KVMPPC_VCPU_BLOCKED;
+	++vc->n_blocked;
+	if (vc->n_runnable > 0 &&
+	    vc->n_runnable + vc->n_blocked == vc->num_threads) {
+		vcpu = list_first_entry(&vc->runnable_threads, struct kvm_vcpu,
+					arch.run_list);
+		wake_up(&vcpu->arch.cpu_run);
+	}
+	spin_unlock(&vc->lock);
+}
+
+static void kvmppc_vcpu_unblocked(struct kvm_vcpu *vcpu)
+{
+	struct kvmppc_vcore *vc = vcpu->arch.vcore;
+
+	spin_lock(&vc->lock);
+	vcpu->arch.state = KVMPPC_VCPU_BUSY_IN_HOST;
+	--vc->n_blocked;
+	spin_unlock(&vc->lock);
+}
+
 extern int __kvmppc_vcore_entry(struct kvm_run *kvm_run, struct kvm_vcpu *vcpu);
+extern void xics_wake_cpu(int cpu);
 
-static int kvmppc_run_vcpu(struct kvm_run *run, struct kvm_vcpu *vcpu)
+static void kvmppc_remove_runnable(struct kvmppc_vcore *vc,
+				   struct kvm_vcpu *vcpu)
 {
-	u64 now;
+	struct kvm_vcpu *v;
 
-	if (signal_pending(current)) {
-		run->exit_reason = KVM_EXIT_INTR;
-		return -EINTR;
+	if (vcpu->arch.state != KVMPPC_VCPU_RUNNABLE)
+		return;
+	vcpu->arch.state = KVMPPC_VCPU_BUSY_IN_HOST;
+	--vc->n_runnable;
+	/* decrement the physical thread id of each following vcpu */
+	v = vcpu;
+	list_for_each_entry_continue(v, &vc->runnable_threads, arch.run_list)
+		--v->arch.ptid;
+	list_del(&vcpu->arch.run_list);
+}
+
+static void kvmppc_start_thread(struct kvm_vcpu *vcpu)
+{
+	int cpu;
+	struct paca_struct *tpaca;
+	struct kvmppc_vcore *vc = vcpu->arch.vcore;
+
+	cpu = vc->pcpu + vcpu->arch.ptid;
+	tpaca = &paca[cpu];
+	tpaca->kvm_hstate.kvm_vcpu = vcpu;
+	tpaca->kvm_hstate.kvm_vcore = vc;
+	smp_wmb();
+#ifdef CONFIG_PPC_ICP_NATIVE
+	if (vcpu->arch.ptid) {
+		tpaca->cpu_start = 0x80;
+		tpaca->kvm_hstate.in_guest = KVM_GUEST_MODE_GUEST;
+		wmb();
+		xics_wake_cpu(cpu);
+		++vc->n_woken;
 	}
+#endif
+}
 
-	flush_fp_to_thread(current);
-	flush_altivec_to_thread(current);
-	flush_vsx_to_thread(current);
-	preempt_disable();
+static void kvmppc_wait_for_nap(struct kvmppc_vcore *vc)
+{
+	int i;
+
+	HMT_low();
+	i = 0;
+	while (vc->nap_count < vc->n_woken) {
+		if (++i >= 1000000) {
+			pr_err("kvmppc_wait_for_nap timeout %d %d\n",
+			       vc->nap_count, vc->n_woken);
+			break;
+		}
+		cpu_relax();
+	}
+	HMT_medium();
+}
+
+/*
+ * Check that we are on thread 0 and that any other threads in
+ * this core are off-line.
+ */
+static int on_primary_thread(void)
+{
+	int cpu = smp_processor_id();
+	int thr = cpu_thread_in_core(cpu);
+
+	if (thr)
+		return 0;
+	while (++thr < threads_per_core)
+		if (cpu_online(cpu + thr))
+			return 0;
+	return 1;
+}
+
+/*
+ * Run a set of guest threads on a physical core.
+ * Called with vc->lock held.
+ */
+static int kvmppc_run_core(struct kvmppc_vcore *vc)
+{
+	struct kvm_vcpu *vcpu, *vnext;
+	long ret;
+	u64 now;
+
+	/* don't start if any threads have a signal pending */
+	list_for_each_entry(vcpu, &vc->runnable_threads, arch.run_list)
+		if (signal_pending(vcpu->arch.run_task))
+			return 0;
 
 	/*
 	 * Make sure we are running on thread 0, and that
@@ -490,36 +638,150 @@ static int kvmppc_run_vcpu(struct kvm_run *run, struct kvm_vcpu *vcpu)
 	 * XXX we should also block attempts to bring any
 	 * secondary threads online.
 	 */
-	if (threads_per_core > 1) {
-		int cpu = smp_processor_id();
-		int thr = cpu_thread_in_core(cpu);
-
-		if (thr)
-			goto out;
-		while (++thr < threads_per_core)
-			if (cpu_online(cpu + thr))
-				goto out;
+	if (threads_per_core > 1 && !on_primary_thread()) {
+		list_for_each_entry(vcpu, &vc->runnable_threads, arch.run_list)
+			vcpu->arch.ret = -EBUSY;
+		goto out;
 	}
 
-	kvm_guest_enter();
+	vc->n_woken = 0;
+	vc->nap_count = 0;
+	vc->entry_exit_count = 0;
+	vc->vcore_running = 1;
+	vc->in_guest = 0;
+	vc->pcpu = smp_processor_id();
+	list_for_each_entry(vcpu, &vc->runnable_threads, arch.run_list)
+		kvmppc_start_thread(vcpu);
+	vcpu = list_first_entry(&vc->runnable_threads, struct kvm_vcpu,
+				arch.run_list);
+
+	spin_unlock(&vc->lock);
 
+	preempt_disable();
+	kvm_guest_enter();
 	__kvmppc_vcore_entry(NULL, vcpu);
 
+	/* wait for secondary threads to finish writing their state to memory */
+	spin_lock(&vc->lock);
+	if (vc->nap_count < vc->n_woken)
+		kvmppc_wait_for_nap(vc);
+	/* prevent other vcpu threads from doing kvmppc_start_thread() now */
+	vc->vcore_running = 2;
+	spin_unlock(&vc->lock);
+
+	/* make sure updates to secondary vcpu structs are visible now */
+	smp_mb();
 	kvm_guest_exit();
 
 	preempt_enable();
 	kvm_resched(vcpu);
 
 	now = get_tb();
-	/* cancel pending dec exception if dec is positive */
-	if (now < vcpu->arch.dec_expires && kvmppc_core_pending_dec(vcpu))
-		kvmppc_core_dequeue_dec(vcpu);
-
-	return kvmppc_handle_exit(run, vcpu, current);
+	list_for_each_entry(vcpu, &vc->runnable_threads, arch.run_list) {
+		/* cancel pending dec exception if dec is positive */
+		if (now < vcpu->arch.dec_expires &&
+		    kvmppc_core_pending_dec(vcpu))
+			kvmppc_core_dequeue_dec(vcpu);
+		if (!vcpu->arch.trap) {
+			if (signal_pending(vcpu->arch.run_task)) {
+				vcpu->arch.kvm_run->exit_reason = KVM_EXIT_INTR;
+				vcpu->arch.ret = -EINTR;
+			}
+			continue;		/* didn't get to run */
+		}
+		ret = kvmppc_handle_exit(vcpu->arch.kvm_run, vcpu,
+					 vcpu->arch.run_task);
+		vcpu->arch.ret = ret;
+		vcpu->arch.trap = 0;
+	}
 
+	spin_lock(&vc->lock);
  out:
-	preempt_enable();
-	return -EBUSY;
+	vc->vcore_running = 0;
+	list_for_each_entry_safe(vcpu, vnext, &vc->runnable_threads,
+				 arch.run_list) {
+		if (vcpu->arch.ret != RESUME_GUEST) {
+			kvmppc_remove_runnable(vc, vcpu);
+			wake_up(&vcpu->arch.cpu_run);
+		}
+	}
+
+	return 1;
+}
+
+static int kvmppc_run_vcpu(struct kvm_run *kvm_run, struct kvm_vcpu *vcpu)
+{
+	int ptid;
+	int wait_state;
+	struct kvmppc_vcore *vc;
+	DEFINE_WAIT(wait);
+
+	/* No need to go into the guest when all we do is going out */
+	if (signal_pending(current)) {
+		kvm_run->exit_reason = KVM_EXIT_INTR;
+		return -EINTR;
+	}
+
+	kvm_run->exit_reason = 0;
+	vcpu->arch.ret = RESUME_GUEST;
+	vcpu->arch.trap = 0;
+
+	flush_fp_to_thread(current);
+	flush_altivec_to_thread(current);
+	flush_vsx_to_thread(current);
+
+	/*
+	 * Synchronize with other threads in this virtual core
+	 */
+	vc = vcpu->arch.vcore;
+	spin_lock(&vc->lock);
+	/* This happens the first time this is called for a vcpu */
+	if (vcpu->arch.state == KVMPPC_VCPU_BLOCKED)
+		--vc->n_blocked;
+	vcpu->arch.state = KVMPPC_VCPU_RUNNABLE;
+	ptid = vc->n_runnable;
+	vcpu->arch.run_task = current;
+	vcpu->arch.kvm_run = kvm_run;
+	vcpu->arch.ptid = ptid;
+	list_add_tail(&vcpu->arch.run_list, &vc->runnable_threads);
+	++vc->n_runnable;
+
+	wait_state = TASK_INTERRUPTIBLE;
+	while (vcpu->arch.state == KVMPPC_VCPU_RUNNABLE) {
+		if (signal_pending(current)) {
+			if (!vc->vcore_running) {
+				kvm_run->exit_reason = KVM_EXIT_INTR;
+				vcpu->arch.ret = -EINTR;
+				break;
+			}
+			/* have to wait for vcore to stop executing guest */
+			wait_state = TASK_UNINTERRUPTIBLE;
+			smp_send_reschedule(vc->pcpu);
+		}
+
+		if (!vc->vcore_running &&
+		    vc->n_runnable + vc->n_blocked == vc->num_threads) {
+			/* we can run now */
+			if (kvmppc_run_core(vc))
+				continue;
+		}
+
+		if (vc->vcore_running == 1 && VCORE_EXIT_COUNT(vc) == 0)
+			kvmppc_start_thread(vcpu);
+
+		/* wait for other threads to come in, or wait for vcore */
+		prepare_to_wait(&vcpu->arch.cpu_run, &wait, wait_state);
+		spin_unlock(&vc->lock);
+		schedule();
+		finish_wait(&vcpu->arch.cpu_run, &wait);
+		spin_lock(&vc->lock);
+	}
+
+	if (vcpu->arch.state == KVMPPC_VCPU_RUNNABLE)
+		kvmppc_remove_runnable(vc, vcpu);
+	spin_unlock(&vc->lock);
+
+	return vcpu->arch.ret;
 }
 
 int kvmppc_vcpu_run(struct kvm_run *run, struct kvm_vcpu *vcpu)