KVM: PPC: Accelerate H_PUT_TCE by implementing it in real mode

This improves I/O performance for guests using the PAPR
paravirtualization interface by making the H_PUT_TCE hcall faster, by
implementing it in real mode.  H_PUT_TCE is used for updating virtual
IOMMU tables, and is used both for virtual I/O and for real I/O in the
PAPR interface.

Since this moves the IOMMU tables into the kernel, we define a new
KVM_CREATE_SPAPR_TCE ioctl to allow qemu to create the tables.  The
ioctl returns a file descriptor which can be used to mmap the newly
created table.  The qemu driver models use them in the same way as
userspace managed tables, but they can be updated directly by the
guest with a real-mode H_PUT_TCE implementation, reducing the number
of host/guest context switches during guest IO.

There are certain circumstances where it is useful for userland qemu
to write to the TCE table even if the kernel H_PUT_TCE path is used
most of the time.  Specifically, allowing this will avoid awkwardness
when we need to reset the table.  More importantly, we will in the
future need to write the table in order to restore its state after a
checkpoint resume or migration.

Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>

1 files changed, 115 insertions, 1 deletions

diff --git a/arch/powerpc/kvm/book3s_hv.c b/arch/powerpc/kvm/book3s_hv.c
index af862c30b70e..6fe469eabce8 100644
--- a/arch/powerpc/kvm/book3s_hv.c
+++ b/arch/powerpc/kvm/book3s_hv.c
@@ -538,6 +538,116 @@ int kvmppc_vcpu_run(struct kvm_run *run, struct kvm_vcpu *vcpu)
 	return r;
 }
 
+static long kvmppc_stt_npages(unsigned long window_size)
+{
+	return ALIGN((window_size >> SPAPR_TCE_SHIFT)
+		     * sizeof(u64), PAGE_SIZE) / PAGE_SIZE;
+}
+
+static void release_spapr_tce_table(struct kvmppc_spapr_tce_table *stt)
+{
+	struct kvm *kvm = stt->kvm;
+	int i;
+
+	mutex_lock(&kvm->lock);
+	list_del(&stt->list);
+	for (i = 0; i < kvmppc_stt_npages(stt->window_size); i++)
+		__free_page(stt->pages[i]);
+	kfree(stt);
+	mutex_unlock(&kvm->lock);
+
+	kvm_put_kvm(kvm);
+}
+
+static int kvm_spapr_tce_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
+{
+	struct kvmppc_spapr_tce_table *stt = vma->vm_file->private_data;
+	struct page *page;
+
+	if (vmf->pgoff >= kvmppc_stt_npages(stt->window_size))
+		return VM_FAULT_SIGBUS;
+
+	page = stt->pages[vmf->pgoff];
+	get_page(page);
+	vmf->page = page;
+	return 0;
+}
+
+static const struct vm_operations_struct kvm_spapr_tce_vm_ops = {
+	.fault = kvm_spapr_tce_fault,
+};
+
+static int kvm_spapr_tce_mmap(struct file *file, struct vm_area_struct *vma)
+{
+	vma->vm_ops = &kvm_spapr_tce_vm_ops;
+	return 0;
+}
+
+static int kvm_spapr_tce_release(struct inode *inode, struct file *filp)
+{
+	struct kvmppc_spapr_tce_table *stt = filp->private_data;
+
+	release_spapr_tce_table(stt);
+	return 0;
+}
+
+static struct file_operations kvm_spapr_tce_fops = {
+	.mmap           = kvm_spapr_tce_mmap,
+	.release	= kvm_spapr_tce_release,
+};
+
+long kvm_vm_ioctl_create_spapr_tce(struct kvm *kvm,
+				   struct kvm_create_spapr_tce *args)
+{
+	struct kvmppc_spapr_tce_table *stt = NULL;
+	long npages;
+	int ret = -ENOMEM;
+	int i;
+
+	/* Check this LIOBN hasn't been previously allocated */
+	list_for_each_entry(stt, &kvm->arch.spapr_tce_tables, list) {
+		if (stt->liobn == args->liobn)
+			return -EBUSY;
+	}
+
+	npages = kvmppc_stt_npages(args->window_size);
+
+	stt = kzalloc(sizeof(*stt) + npages* sizeof(struct page *),
+		      GFP_KERNEL);
+	if (!stt)
+		goto fail;
+
+	stt->liobn = args->liobn;
+	stt->window_size = args->window_size;
+	stt->kvm = kvm;
+
+	for (i = 0; i < npages; i++) {
+		stt->pages[i] = alloc_page(GFP_KERNEL | __GFP_ZERO);
+		if (!stt->pages[i])
+			goto fail;
+	}
+
+	kvm_get_kvm(kvm);
+
+	mutex_lock(&kvm->lock);
+	list_add(&stt->list, &kvm->arch.spapr_tce_tables);
+
+	mutex_unlock(&kvm->lock);
+
+	return anon_inode_getfd("kvm-spapr-tce", &kvm_spapr_tce_fops,
+				stt, O_RDWR);
+
+fail:
+	if (stt) {
+		for (i = 0; i < npages; i++)
+			if (stt->pages[i])
+				__free_page(stt->pages[i]);
+
+		kfree(stt);
+	}
+	return ret;
+}
+
 int kvmppc_core_prepare_memory_region(struct kvm *kvm,
 				struct kvm_userspace_memory_region *mem)
 {
@@ -559,13 +669,17 @@ int kvmppc_core_init_vm(struct kvm *kvm)
 
 	/* Allocate hashed page table */
 	r = kvmppc_alloc_hpt(kvm);
+	if (r)
+		return r;
 
-	return r;
+	INIT_LIST_HEAD(&kvm->arch.spapr_tce_tables);
+	return 0;
 }
 
 void kvmppc_core_destroy_vm(struct kvm *kvm)
 {
 	kvmppc_free_hpt(kvm);
+	WARN_ON(!list_empty(&kvm->arch.spapr_tce_tables));
 }
 
 /* These are stubs for now */