Merge tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost

Pull vhost updates from Michael Tsirkin:
 "virtio, vhost: optimizations, fixes

  Looks like a quiet cycle for vhost/virtio, just a couple of minor
  tweaks. Most notable is automatic interrupt affinity for blk and scsi.
  Hopefully other devices are not far behind"

* tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost:
  virtio-console: avoid DMA from stack
  vhost: introduce O(1) vq metadata cache
  virtio_scsi: use virtio IRQ affinity
  virtio_blk: use virtio IRQ affinity
  blk-mq: provide a default queue mapping for virtio device
  virtio: provide a method to get the IRQ affinity mask for a virtqueue
  virtio: allow drivers to request IRQ affinity when creating VQs
  virtio_pci: simplify MSI-X setup
  virtio_pci: don't duplicate the msix_enable flag in struct pci_dev
  virtio_pci: use shared interrupts for virtqueues
  virtio_pci: remove struct virtio_pci_vq_info
  vhost: try avoiding avail index access when getting descriptor
  virtio_mmio: expose header to userspace

block/Kconfig

@@ -189,4 +189,9 @@ config BLK_MQ_PCI
 	depends on BLOCK && PCI
 	default y
 
+config BLK_MQ_VIRTIO
+	bool
+	depends on BLOCK && VIRTIO
+	default y
+
 source block/Kconfig.iosched

block/Makefile

@@ -25,6 +25,7 @@ obj-$(CONFIG_BLOCK_COMPAT)	+= compat_ioctl.o
 obj-$(CONFIG_BLK_CMDLINE_PARSER)	+= cmdline-parser.o
 obj-$(CONFIG_BLK_DEV_INTEGRITY) += bio-integrity.o blk-integrity.o t10-pi.o
 obj-$(CONFIG_BLK_MQ_PCI)	+= blk-mq-pci.o
+obj-$(CONFIG_BLK_MQ_VIRTIO)	+= blk-mq-virtio.o
 obj-$(CONFIG_BLK_DEV_ZONED)	+= blk-zoned.o
 obj-$(CONFIG_BLK_WBT)		+= blk-wbt.o
 obj-$(CONFIG_BLK_DEBUG_FS)	+= blk-mq-debugfs.o

block/blk-mq-virtio.c

@@ -0,0 +1,54 @@
+/*
+ * Copyright (c) 2016 Christoph Hellwig.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ */
+#include <linux/device.h>
+#include <linux/blk-mq.h>
+#include <linux/blk-mq-virtio.h>
+#include <linux/virtio_config.h>
+#include <linux/module.h>
+#include "blk-mq.h"
+
+/**
+ * blk_mq_virtio_map_queues - provide a default queue mapping for virtio device
+ * @set:	tagset to provide the mapping for
+ * @vdev:	virtio device associated with @set.
+ * @first_vec:	first interrupt vectors to use for queues (usually 0)
+ *
+ * This function assumes the virtio device @vdev has at least as many available
+ * interrupt vetors as @set has queues.  It will then queuery the vector
+ * corresponding to each queue for it's affinity mask and built queue mapping
+ * that maps a queue to the CPUs that have irq affinity for the corresponding
+ * vector.
+ */
+int blk_mq_virtio_map_queues(struct blk_mq_tag_set *set,
+		struct virtio_device *vdev, int first_vec)
+{
+	const struct cpumask *mask;
+	unsigned int queue, cpu;
+
+	if (!vdev->config->get_vq_affinity)
+		goto fallback;
+
+	for (queue = 0; queue < set->nr_hw_queues; queue++) {
+		mask = vdev->config->get_vq_affinity(vdev, first_vec + queue);
+		if (!mask)
+			goto fallback;
+
+		for_each_cpu(cpu, mask)
+			set->mq_map[cpu] = queue;
+	}
+
+	return 0;
+fallback:
+	return blk_mq_map_queues(set);
+}
+EXPORT_SYMBOL_GPL(blk_mq_virtio_map_queues);

drivers/block/virtio_blk.c

@@ -5,6 +5,7 @@
 #include <linux/hdreg.h>
 #include <linux/module.h>
 #include <linux/mutex.h>
+#include <linux/interrupt.h>
 #include <linux/virtio.h>
 #include <linux/virtio_blk.h>
 #include <linux/scatterlist.h>
@@ -12,6 +13,7 @@
 #include <scsi/scsi_cmnd.h>
 #include <linux/idr.h>
 #include <linux/blk-mq.h>
+#include <linux/blk-mq-virtio.h>
 #include <linux/numa.h>
 
 #define PART_BITS 4
@@ -426,6 +428,7 @@ static int init_vq(struct virtio_blk *vblk)
 	struct virtqueue **vqs;
 	unsigned short num_vqs;
 	struct virtio_device *vdev = vblk->vdev;
+	struct irq_affinity desc = { 0, };
 
 	err = virtio_cread_feature(vdev, VIRTIO_BLK_F_MQ,
 				   struct virtio_blk_config, num_queues,
@@ -452,7 +455,8 @@ static int init_vq(struct virtio_blk *vblk)
 	}
 
 	/* Discover virtqueues and write information to configuration.  */
-	err = vdev->config->find_vqs(vdev, num_vqs, vqs, callbacks, names);
+	err = vdev->config->find_vqs(vdev, num_vqs, vqs, callbacks, names,
+			&desc);
 	if (err)
 		goto out;
 
@@ -586,10 +590,18 @@ static int virtblk_init_request(void *data, struct request *rq,
 	return 0;
 }
 
+static int virtblk_map_queues(struct blk_mq_tag_set *set)
+{
+	struct virtio_blk *vblk = set->driver_data;
+
+	return blk_mq_virtio_map_queues(set, vblk->vdev, 0);
+}
+
 static struct blk_mq_ops virtio_mq_ops = {
 	.queue_rq	= virtio_queue_rq,
 	.complete	= virtblk_request_done,
 	.init_request	= virtblk_init_request,
+	.map_queues	= virtblk_map_queues,
 };
 
 static unsigned int virtblk_queue_depth;

drivers/char/virtio_console.c

@@ -1136,6 +1136,8 @@ static int put_chars(u32 vtermno, const char *buf, int count)
 {
 	struct port *port;
 	struct scatterlist sg[1];
+	void *data;
+	int ret;
 
 	if (unlikely(early_put_chars))
 		return early_put_chars(vtermno, buf, count);
@@ -1144,8 +1146,14 @@ static int put_chars(u32 vtermno, const char *buf, int count)
 	if (!port)
 		return -EPIPE;
 
-	sg_init_one(sg, buf, count);
-	return __send_to_port(port, sg, 1, count, (void *)buf, false);
+	data = kmemdup(buf, count, GFP_ATOMIC);
+	if (!data)
+		return -ENOMEM;
+
+	sg_init_one(sg, data, count);
+	ret = __send_to_port(port, sg, 1, count, data, false);
+	kfree(data);
+	return ret;
 }
 
 /*
@@ -1939,7 +1947,7 @@ static int init_vqs(struct ports_device *portdev)
 	/* Find the queues. */
 	err = portdev->vdev->config->find_vqs(portdev->vdev, nr_queues, vqs,
 					      io_callbacks,
-					      (const char **)io_names);
+					      (const char **)io_names, NULL);
 	if (err)
 		goto free;
 

drivers/crypto/virtio/virtio_crypto_core.c

@@ -120,7 +120,7 @@ static int virtcrypto_find_vqs(struct virtio_crypto *vi)
 	}
 
 	ret = vi->vdev->config->find_vqs(vi->vdev, total_vqs, vqs, callbacks,
-					 names);
+					 names, NULL);
 	if (ret)
 		goto err_find;
 

drivers/gpu/drm/virtio/virtgpu_kms.c

@@ -176,7 +176,7 @@ int virtio_gpu_driver_load(struct drm_device *dev, unsigned long flags)
 #endif
 
 	ret = vgdev->vdev->config->find_vqs(vgdev->vdev, 2, vqs,
-					    callbacks, names);
+					    callbacks, names, NULL);
 	if (ret) {
 		DRM_ERROR("failed to find virt queues\n");
 		goto err_vqs;

drivers/misc/mic/vop/vop_main.c

@@ -374,7 +374,7 @@ unmap:
 static int vop_find_vqs(struct virtio_device *dev, unsigned nvqs,
 			struct virtqueue *vqs[],
 			vq_callback_t *callbacks[],
-			const char * const names[])
+			const char * const names[], struct irq_affinity *desc)
 {
 	struct _vop_vdev *vdev = to_vopvdev(dev);
 	struct vop_device *vpdev = vdev->vpdev;

drivers/net/caif/caif_virtio.c

@@ -679,7 +679,8 @@ static int cfv_probe(struct virtio_device *vdev)
 		goto err;
 
 	/* Get the TX virtio ring. This is a "guest side vring". */
-	err = vdev->config->find_vqs(vdev, 1, &cfv->vq_tx, &vq_cbs, &names);
+	err = vdev->config->find_vqs(vdev, 1, &cfv->vq_tx, &vq_cbs, &names,
+			NULL);
 	if (err)
 		goto err;
 

drivers/net/virtio_net.c

@@ -2080,7 +2080,7 @@ static int virtnet_find_vqs(struct virtnet_info *vi)
 	}
 
 	ret = vi->vdev->config->find_vqs(vi->vdev, total_vqs, vqs, callbacks,
-					 names);
+					 names, NULL);
 	if (ret)
 		goto err_find;
 

drivers/remoteproc/remoteproc_virtio.c

@@ -137,7 +137,8 @@ static void rproc_virtio_del_vqs(struct virtio_device *vdev)
 static int rproc_virtio_find_vqs(struct virtio_device *vdev, unsigned int nvqs,
 				 struct virtqueue *vqs[],
 				 vq_callback_t *callbacks[],
-				 const char * const names[])
+				 const char * const names[],
+				 struct irq_affinity *desc)
 {
 	int i, ret;
 

drivers/rpmsg/virtio_rpmsg_bus.c

@@ -869,7 +869,7 @@ static int rpmsg_probe(struct virtio_device *vdev)
 	init_waitqueue_head(&vrp->sendq);
 
 	/* We expect two virtqueues, rx and tx (and in this order) */
-	err = vdev->config->find_vqs(vdev, 2, vqs, vq_cbs, names);
+	err = vdev->config->find_vqs(vdev, 2, vqs, vq_cbs, names, NULL);
 	if (err)
 		goto free_vrp;
 

drivers/s390/virtio/kvm_virtio.c

@@ -255,7 +255,8 @@ static void kvm_del_vqs(struct virtio_device *vdev)
 static int kvm_find_vqs(struct virtio_device *vdev, unsigned nvqs,
 			struct virtqueue *vqs[],
 			vq_callback_t *callbacks[],
-			const char * const names[])
+			const char * const names[],
+			struct irq_affinity *desc)
 {
 	struct kvm_device *kdev = to_kvmdev(vdev);
 	int i;

drivers/s390/virtio/virtio_ccw.c

@@ -628,7 +628,8 @@ out:
 static int virtio_ccw_find_vqs(struct virtio_device *vdev, unsigned nvqs,
 			       struct virtqueue *vqs[],
 			       vq_callback_t *callbacks[],
-			       const char * const names[])
+			       const char * const names[],
+			       struct irq_affinity *desc)
 {
 	struct virtio_ccw_device *vcdev = to_vc_device(vdev);
 	unsigned long *indicatorp = NULL;

drivers/scsi/virtio_scsi.c

@@ -18,6 +18,7 @@
 #include <linux/module.h>
 #include <linux/slab.h>
 #include <linux/mempool.h>
+#include <linux/interrupt.h>
 #include <linux/virtio.h>
 #include <linux/virtio_ids.h>
 #include <linux/virtio_config.h>
@@ -29,6 +30,7 @@
 #include <scsi/scsi_cmnd.h>
 #include <scsi/scsi_tcq.h>
 #include <linux/seqlock.h>
+#include <linux/blk-mq-virtio.h>
 
 #define VIRTIO_SCSI_MEMPOOL_SZ 64
 #define VIRTIO_SCSI_EVENT_LEN 8
@@ -108,7 +110,6 @@ struct virtio_scsi {
 	bool affinity_hint_set;
 
 	struct hlist_node node;
-	struct hlist_node node_dead;
 
 	/* Protected by event_vq lock */
 	bool stop_events;
@@ -118,7 +119,6 @@ struct virtio_scsi {
 	struct virtio_scsi_vq req_vqs[];
 };
 
-static enum cpuhp_state virtioscsi_online;
 static struct kmem_cache *virtscsi_cmd_cache;
 static mempool_t *virtscsi_cmd_pool;
 
@@ -766,6 +766,13 @@ static void virtscsi_target_destroy(struct scsi_target *starget)
 	kfree(tgt);
 }
 
+static int virtscsi_map_queues(struct Scsi_Host *shost)
+{
+	struct virtio_scsi *vscsi = shost_priv(shost);
+
+	return blk_mq_virtio_map_queues(&shost->tag_set, vscsi->vdev, 2);
+}
+
 static struct scsi_host_template virtscsi_host_template_single = {
 	.module = THIS_MODULE,
 	.name = "Virtio SCSI HBA",
@@ -801,6 +808,7 @@ static struct scsi_host_template virtscsi_host_template_multi = {
 	.use_clustering = ENABLE_CLUSTERING,
 	.target_alloc = virtscsi_target_alloc,
 	.target_destroy = virtscsi_target_destroy,
+	.map_queues = virtscsi_map_queues,
 	.track_queue_depth = 1,
 };
 
@@ -817,80 +825,6 @@ static struct scsi_host_template virtscsi_host_template_multi = {
 		virtio_cwrite(vdev, struct virtio_scsi_config, fld, &__val); \
 	} while(0)
 
-static void __virtscsi_set_affinity(struct virtio_scsi *vscsi, bool affinity)
-{
-	int i;
-	int cpu;
-
-	/* In multiqueue mode, when the number of cpu is equal
-	 * to the number of request queues, we let the qeueues
-	 * to be private to one cpu by setting the affinity hint
-	 * to eliminate the contention.
-	 */
-	if ((vscsi->num_queues == 1 ||
-	     vscsi->num_queues != num_online_cpus()) && affinity) {
-		if (vscsi->affinity_hint_set)
-			affinity = false;
-		else
-			return;
-	}
-
-	if (affinity) {
-		i = 0;
-		for_each_online_cpu(cpu) {
-			virtqueue_set_affinity(vscsi->req_vqs[i].vq, cpu);
-			i++;
-		}
-
-		vscsi->affinity_hint_set = true;
-	} else {
-		for (i = 0; i < vscsi->num_queues; i++) {
-			if (!vscsi->req_vqs[i].vq)
-				continue;
-
-			virtqueue_set_affinity(vscsi->req_vqs[i].vq, -1);
-		}
-
-		vscsi->affinity_hint_set = false;
-	}
-}
-
-static void virtscsi_set_affinity(struct virtio_scsi *vscsi, bool affinity)
-{
-	get_online_cpus();
-	__virtscsi_set_affinity(vscsi, affinity);
-	put_online_cpus();
-}
-
-static int virtscsi_cpu_online(unsigned int cpu, struct hlist_node *node)
-{
-	struct virtio_scsi *vscsi = hlist_entry_safe(node, struct virtio_scsi,
-						     node);
-	__virtscsi_set_affinity(vscsi, true);
-	return 0;
-}
-
-static int virtscsi_cpu_notif_add(struct virtio_scsi *vi)
-{
-	int ret;
-
-	ret = cpuhp_state_add_instance(virtioscsi_online, &vi->node);
-	if (ret)
-		return ret;
-
-	ret = cpuhp_state_add_instance(CPUHP_VIRT_SCSI_DEAD, &vi->node_dead);
-	if (ret)
-		cpuhp_state_remove_instance(virtioscsi_online, &vi->node);
-	return ret;
-}
-
-static void virtscsi_cpu_notif_remove(struct virtio_scsi *vi)
-{
-	cpuhp_state_remove_instance_nocalls(virtioscsi_online, &vi->node);
-	cpuhp_state_remove_instance_nocalls(CPUHP_VIRT_SCSI_DEAD,
-					    &vi->node_dead);
-}
-
 static void virtscsi_init_vq(struct virtio_scsi_vq *virtscsi_vq,
 			     struct virtqueue *vq)
 {
@@ -900,14 +834,8 @@ static void virtscsi_init_vq(struct virtio_scsi_vq *virtscsi_vq,
 
 static void virtscsi_remove_vqs(struct virtio_device *vdev)
 {
-	struct Scsi_Host *sh = virtio_scsi_host(vdev);
-	struct virtio_scsi *vscsi = shost_priv(sh);
-
-	virtscsi_set_affinity(vscsi, false);
-
 	/* Stop all the virtqueues. */
 	vdev->config->reset(vdev);
-
 	vdev->config->del_vqs(vdev);
 }
 
@@ -920,6 +848,7 @@ static int virtscsi_init(struct virtio_device *vdev,
 	vq_callback_t **callbacks;
 	const char **names;
 	struct virtqueue **vqs;
+	struct irq_affinity desc = { .pre_vectors = 2 };
 
 	num_vqs = vscsi->num_queues + VIRTIO_SCSI_VQ_BASE;
 	vqs = kmalloc(num_vqs * sizeof(struct virtqueue *), GFP_KERNEL);
@@ -941,7 +870,8 @@ static int virtscsi_init(struct virtio_device *vdev,
 	}
 
 	/* Discover virtqueues and write information to configuration.  */
-	err = vdev->config->find_vqs(vdev, num_vqs, vqs, callbacks, names);
+	err = vdev->config->find_vqs(vdev, num_vqs, vqs, callbacks, names,
+			&desc);
 	if (err)
 		goto out;
 
@@ -1007,10 +937,6 @@ static int virtscsi_probe(struct virtio_device *vdev)
 	if (err)
 		goto virtscsi_init_failed;
 
-	err = virtscsi_cpu_notif_add(vscsi);
-	if (err)
-		goto scsi_add_host_failed;
-
 	cmd_per_lun = virtscsi_config_get(vdev, cmd_per_lun) ?: 1;
 	shost->cmd_per_lun = min_t(u32, cmd_per_lun, shost->can_queue);
 	shost->max_sectors = virtscsi_config_get(vdev, max_sectors) ?: 0xFFFF;
@@ -1065,9 +991,6 @@ static void virtscsi_remove(struct virtio_device *vdev)
 		virtscsi_cancel_event_work(vscsi);
 
 	scsi_remove_host(shost);
-
-	virtscsi_cpu_notif_remove(vscsi);
-
 	virtscsi_remove_vqs(vdev);
 	scsi_host_put(shost);
 }
@@ -1075,10 +998,6 @@ static void virtscsi_remove(struct virtio_device *vdev)
 #ifdef CONFIG_PM_SLEEP
 static int virtscsi_freeze(struct virtio_device *vdev)
 {
-	struct Scsi_Host *sh = virtio_scsi_host(vdev);
-	struct virtio_scsi *vscsi = shost_priv(sh);
-
-	virtscsi_cpu_notif_remove(vscsi);
 	virtscsi_remove_vqs(vdev);
 	return 0;
 }
@@ -1093,11 +1012,6 @@ static int virtscsi_restore(struct virtio_device *vdev)
 	if (err)
 		return err;
 
-	err = virtscsi_cpu_notif_add(vscsi);
-	if (err) {
-		vdev->config->del_vqs(vdev);
-		return err;
-	}
 	virtio_device_ready(vdev);
 
 	if (virtio_has_feature(vdev, VIRTIO_SCSI_F_HOTPLUG))
@@ -1152,16 +1066,6 @@ static int __init init(void)
 		pr_err("mempool_create() for virtscsi_cmd_pool failed\n");
 		goto error;
 	}
-	ret = cpuhp_setup_state_multi(CPUHP_AP_ONLINE_DYN,
-				      "scsi/virtio:online",
-				      virtscsi_cpu_online, NULL);
-	if (ret < 0)
-		goto error;
-	virtioscsi_online = ret;
-	ret = cpuhp_setup_state_multi(CPUHP_VIRT_SCSI_DEAD, "scsi/virtio:dead",
-				      NULL, virtscsi_cpu_online);
-	if (ret)
-		goto error;
 	ret = register_virtio_driver(&virtio_scsi_driver);
 	if (ret < 0)
 		goto error;
@@ -1177,17 +1081,12 @@ error:
 		kmem_cache_destroy(virtscsi_cmd_cache);
 		virtscsi_cmd_cache = NULL;
 	}
-	if (virtioscsi_online)
-		cpuhp_remove_multi_state(virtioscsi_online);
-	cpuhp_remove_multi_state(CPUHP_VIRT_SCSI_DEAD);
 	return ret;
 }
 
 static void __exit fini(void)
 {
 	unregister_virtio_driver(&virtio_scsi_driver);
-	cpuhp_remove_multi_state(virtioscsi_online);
-	cpuhp_remove_multi_state(CPUHP_VIRT_SCSI_DEAD);
 	mempool_destroy(virtscsi_cmd_pool);
 	kmem_cache_destroy(virtscsi_cmd_cache);
 }

drivers/vhost/vhost.c

@@ -282,6 +282,22 @@ void vhost_poll_queue(struct vhost_poll *poll)
 }
 EXPORT_SYMBOL_GPL(vhost_poll_queue);
 
+static void __vhost_vq_meta_reset(struct vhost_virtqueue *vq)
+{
+	int j;
+
+	for (j = 0; j < VHOST_NUM_ADDRS; j++)
+		vq->meta_iotlb[j] = NULL;
+}
+
+static void vhost_vq_meta_reset(struct vhost_dev *d)
+{
+	int i;
+
+	for (i = 0; i < d->nvqs; ++i)
+		__vhost_vq_meta_reset(d->vqs[i]);
+}
+
 static void vhost_vq_reset(struct vhost_dev *dev,
 			   struct vhost_virtqueue *vq)
 {
@@ -312,6 +328,7 @@ static void vhost_vq_reset(struct vhost_dev *dev,
 	vq->busyloop_timeout = 0;
 	vq->umem = NULL;
 	vq->iotlb = NULL;
+	__vhost_vq_meta_reset(vq);
 }
 
 static int vhost_worker(void *data)
@@ -691,6 +708,18 @@ static int vq_memory_access_ok(void __user *log_base, struct vhost_umem *umem,
 	return 1;
 }
 
+static inline void __user *vhost_vq_meta_fetch(struct vhost_virtqueue *vq,
+					       u64 addr, unsigned int size,
+					       int type)
+{
+	const struct vhost_umem_node *node = vq->meta_iotlb[type];
+
+	if (!node)
+		return NULL;
+
+	return (void *)(uintptr_t)(node->userspace_addr + addr - node->start);
+}
+
 /* Can we switch to this memory table? */
 /* Caller should have device mutex but not vq mutex */
 static int memory_access_ok(struct vhost_dev *d, struct vhost_umem *umem,
@@ -733,8 +762,14 @@ static int vhost_copy_to_user(struct vhost_virtqueue *vq, void __user *to,
 		 * could be access through iotlb. So -EAGAIN should
 		 * not happen in this case.
 		 */
-		/* TODO: more fast path */
 		struct iov_iter t;
+		void __user *uaddr = vhost_vq_meta_fetch(vq,
+				     (u64)(uintptr_t)to, size,
+				     VHOST_ADDR_DESC);
+
+		if (uaddr)
+			return __copy_to_user(uaddr, from, size);
+
 		ret = translate_desc(vq, (u64)(uintptr_t)to, size, vq->iotlb_iov,
 				     ARRAY_SIZE(vq->iotlb_iov),
 				     VHOST_ACCESS_WO);
@@ -762,8 +797,14 @@ static int vhost_copy_from_user(struct vhost_virtqueue *vq, void *to,
 		 * could be access through iotlb. So -EAGAIN should
 		 * not happen in this case.
 		 */
-		/* TODO: more fast path */
+		void __user *uaddr = vhost_vq_meta_fetch(vq,
+				     (u64)(uintptr_t)from, size,
+				     VHOST_ADDR_DESC);
 		struct iov_iter f;
+
+		if (uaddr)
+			return __copy_from_user(to, uaddr, size);
+
 		ret = translate_desc(vq, (u64)(uintptr_t)from, size, vq->iotlb_iov,
 				     ARRAY_SIZE(vq->iotlb_iov),
 				     VHOST_ACCESS_RO);
@@ -783,17 +824,12 @@ out:
 	return ret;
 }
 
-static void __user *__vhost_get_user(struct vhost_virtqueue *vq,
-				     void __user *addr, unsigned size)
+static void __user *__vhost_get_user_slow(struct vhost_virtqueue *vq,
+					  void __user *addr, unsigned int size,
+					  int type)
 {
 	int ret;
 
-	/* This function should be called after iotlb
-	 * prefetch, which means we're sure that vq
-	 * could be access through iotlb. So -EAGAIN should
-	 * not happen in this case.
-	 */
-	/* TODO: more fast path */
 	ret = translate_desc(vq, (u64)(uintptr_t)addr, size, vq->iotlb_iov,
 			     ARRAY_SIZE(vq->iotlb_iov),
 			     VHOST_ACCESS_RO);
@@ -814,14 +850,32 @@ static void __user *__vhost_get_user(struct vhost_virtqueue *vq,
 	return vq->iotlb_iov[0].iov_base;
 }
 
-#define vhost_put_user(vq, x, ptr) \
+/* This function should be called after iotlb
+ * prefetch, which means we're sure that vq
+ * could be access through iotlb. So -EAGAIN should
+ * not happen in this case.
+ */
+static inline void __user *__vhost_get_user(struct vhost_virtqueue *vq,
+					    void *addr, unsigned int size,
+					    int type)
+{
+	void __user *uaddr = vhost_vq_meta_fetch(vq,
+			     (u64)(uintptr_t)addr, size, type);
+	if (uaddr)
+		return uaddr;
+
+	return __vhost_get_user_slow(vq, addr, size, type);
+}
+
+#define vhost_put_user(vq, x, ptr)		\
 ({ \
 	int ret = -EFAULT; \
 	if (!vq->iotlb) { \
 		ret = __put_user(x, ptr); \
 	} else { \
 		__typeof__(ptr) to = \
-			(__typeof__(ptr)) __vhost_get_user(vq, ptr, sizeof(*ptr)); \
+			(__typeof__(ptr)) __vhost_get_user(vq, ptr,	\
+					  sizeof(*ptr), VHOST_ADDR_USED); \
 		if (to != NULL) \
 			ret = __put_user(x, to); \
 		else \
@@ -830,14 +884,16 @@ static void __user *__vhost_get_user(struct vhost_virtqueue *vq,
 	ret; \
 })
 
-#define vhost_get_user(vq, x, ptr) \
+#define vhost_get_user(vq, x, ptr, type)		\
 ({ \
 	int ret; \
 	if (!vq->iotlb) { \
 		ret = __get_user(x, ptr); \
 	} else { \
 		__typeof__(ptr) from = \
-			(__typeof__(ptr)) __vhost_get_user(vq, ptr, sizeof(*ptr)); \
+			(__typeof__(ptr)) __vhost_get_user(vq, ptr, \
+							   sizeof(*ptr), \
+							   type); \
 		if (from != NULL) \
 			ret = __get_user(x, from); \
 		else \
@@ -846,6 +902,12 @@ static void __user *__vhost_get_user(struct vhost_virtqueue *vq,
 	ret; \
 })
 
+#define vhost_get_avail(vq, x, ptr) \
+	vhost_get_user(vq, x, ptr, VHOST_ADDR_AVAIL)
+
+#define vhost_get_used(vq, x, ptr) \
+	vhost_get_user(vq, x, ptr, VHOST_ADDR_USED)
+
 static void vhost_dev_lock_vqs(struct vhost_dev *d)
 {
 	int i = 0;
@@ -951,6 +1013,7 @@ static int vhost_process_iotlb_msg(struct vhost_dev *dev,
 			ret = -EFAULT;
 			break;
 		}
+		vhost_vq_meta_reset(dev);
 		if (vhost_new_umem_range(dev->iotlb, msg->iova, msg->size,
 					 msg->iova + msg->size - 1,
 					 msg->uaddr, msg->perm)) {
@@ -960,6 +1023,7 @@ static int vhost_process_iotlb_msg(struct vhost_dev *dev,
 		vhost_iotlb_notify_vq(dev, msg);
 		break;
 	case VHOST_IOTLB_INVALIDATE:
+		vhost_vq_meta_reset(dev);
 		vhost_del_umem_range(dev->iotlb, msg->iova,
 				     msg->iova + msg->size - 1);
 		break;
@@ -1103,12 +1167,26 @@ static int vq_access_ok(struct vhost_virtqueue *vq, unsigned int num,
 			sizeof *used + num * sizeof *used->ring + s);
 }
 
+static void vhost_vq_meta_update(struct vhost_virtqueue *vq,
+				 const struct vhost_umem_node *node,
+				 int type)
+{
+	int access = (type == VHOST_ADDR_USED) ?
+		     VHOST_ACCESS_WO : VHOST_ACCESS_RO;
+
+	if (likely(node->perm & access))
+		vq->meta_iotlb[type] = node;
+}
+
 static int iotlb_access_ok(struct vhost_virtqueue *vq,
-			   int access, u64 addr, u64 len)
+			   int access, u64 addr, u64 len, int type)
 {
 	const struct vhost_umem_node *node;
 	struct vhost_umem *umem = vq->iotlb;
-	u64 s = 0, size;
+	u64 s = 0, size, orig_addr = addr;
+
+	if (vhost_vq_meta_fetch(vq, addr, len, type))
+		return true;
 
 	while (len > s) {
 		node = vhost_umem_interval_tree_iter_first(&umem->umem_tree,
@@ -1125,6 +1203,10 @@ static int iotlb_access_ok(struct vhost_virtqueue *vq,
 		}
 
 		size = node->size - addr + node->start;
+
+		if (orig_addr == addr && size >= len)
+			vhost_vq_meta_update(vq, node, type);
+
 		s += size;
 		addr += size;
 	}
@@ -1141,13 +1223,15 @@ int vq_iotlb_prefetch(struct vhost_virtqueue *vq)
 		return 1;
 
 	return iotlb_access_ok(vq, VHOST_ACCESS_RO, (u64)(uintptr_t)vq->desc,
-			       num * sizeof *vq->desc) &&
+			       num * sizeof(*vq->desc), VHOST_ADDR_DESC) &&
 	       iotlb_access_ok(vq, VHOST_ACCESS_RO, (u64)(uintptr_t)vq->avail,
 			       sizeof *vq->avail +
-			       num * sizeof *vq->avail->ring + s) &&
+			       num * sizeof(*vq->avail->ring) + s,
+			       VHOST_ADDR_AVAIL) &&
 	       iotlb_access_ok(vq, VHOST_ACCESS_WO, (u64)(uintptr_t)vq->used,
 			       sizeof *vq->used +
-			       num * sizeof *vq->used->ring + s);
+			       num * sizeof(*vq->used->ring) + s,
+			       VHOST_ADDR_USED);
 }
 EXPORT_SYMBOL_GPL(vq_iotlb_prefetch);
 
@@ -1728,7 +1812,7 @@ int vhost_vq_init_access(struct vhost_virtqueue *vq)
 		r = -EFAULT;
 		goto err;
 	}
-	r = vhost_get_user(vq, last_used_idx, &vq->used->idx);
+	r = vhost_get_used(vq, last_used_idx, &vq->used->idx);
 	if (r) {
 		vq_err(vq, "Can't access used idx at %p\n",
 		       &vq->used->idx);
@@ -1930,29 +2014,36 @@ int vhost_get_vq_desc(struct vhost_virtqueue *vq,
 
 	/* Check it isn't doing very strange things with descriptor numbers. */
 	last_avail_idx = vq->last_avail_idx;
-	if (unlikely(vhost_get_user(vq, avail_idx, &vq->avail->idx))) {
-		vq_err(vq, "Failed to access avail idx at %p\n",
-		       &vq->avail->idx);
-		return -EFAULT;
-	}
-	vq->avail_idx = vhost16_to_cpu(vq, avail_idx);
 
-	if (unlikely((u16)(vq->avail_idx - last_avail_idx) > vq->num)) {
-		vq_err(vq, "Guest moved used index from %u to %u",
-		       last_avail_idx, vq->avail_idx);
-		return -EFAULT;
-	}
+	if (vq->avail_idx == vq->last_avail_idx) {
+		if (unlikely(vhost_get_avail(vq, avail_idx, &vq->avail->idx))) {
+			vq_err(vq, "Failed to access avail idx at %p\n",
+				&vq->avail->idx);
+			return -EFAULT;
+		}
+		vq->avail_idx = vhost16_to_cpu(vq, avail_idx);
 
-	/* If there's nothing new since last we looked, return invalid. */
-	if (vq->avail_idx == last_avail_idx)
-		return vq->num;
+		if (unlikely((u16)(vq->avail_idx - last_avail_idx) > vq->num)) {
+			vq_err(vq, "Guest moved used index from %u to %u",
+				last_avail_idx, vq->avail_idx);
+			return -EFAULT;
+		}
+
+		/* If there's nothing new since last we looked, return
+		 * invalid.
+		 */
+		if (vq->avail_idx == last_avail_idx)
+			return vq->num;
 
-	/* Only get avail ring entries after they have been exposed by guest. */
-	smp_rmb();
+		/* Only get avail ring entries after they have been
+		 * exposed by guest.
+		 */
+		smp_rmb();
+	}
 
 	/* Grab the next descriptor number they're advertising, and increment
 	 * the index we've seen. */
-	if (unlikely(vhost_get_user(vq, ring_head,
+	if (unlikely(vhost_get_avail(vq, ring_head,
 		     &vq->avail->ring[last_avail_idx & (vq->num - 1)]))) {
 		vq_err(vq, "Failed to read head: idx %d address %p\n",
 		       last_avail_idx,
@@ -2168,7 +2259,7 @@ static bool vhost_notify(struct vhost_dev *dev, struct vhost_virtqueue *vq)
 		 * with the barrier that the Guest executes when enabling
 		 * interrupts. */
 		smp_mb();
-		if (vhost_get_user(vq, flags, &vq->avail->flags)) {
+		if (vhost_get_avail(vq, flags, &vq->avail->flags)) {
 			vq_err(vq, "Failed to get flags");
 			return true;
 		}
@@ -2195,7 +2286,7 @@ static bool vhost_notify(struct vhost_dev *dev, struct vhost_virtqueue *vq)
 	 * interrupts. */
 	smp_mb();
 
-	if (vhost_get_user(vq, event, vhost_used_event(vq))) {
+	if (vhost_get_avail(vq, event, vhost_used_event(vq))) {
 		vq_err(vq, "Failed to get used event idx");
 		return true;
 	}
@@ -2242,7 +2333,7 @@ bool vhost_vq_avail_empty(struct vhost_dev *dev, struct vhost_virtqueue *vq)
 	if (vq->avail_idx != vq->last_avail_idx)
 		return false;
 
-	r = vhost_get_user(vq, avail_idx, &vq->avail->idx);
+	r = vhost_get_avail(vq, avail_idx, &vq->avail->idx);
 	if (unlikely(r))
 		return false;
 	vq->avail_idx = vhost16_to_cpu(vq, avail_idx);
@@ -2278,7 +2369,7 @@ bool vhost_enable_notify(struct vhost_dev *dev, struct vhost_virtqueue *vq)
 	/* They could have slipped one in as we were doing that: make
 	 * sure it's written, then check again. */
 	smp_mb();
-	r = vhost_get_user(vq, avail_idx, &vq->avail->idx);
+	r = vhost_get_avail(vq, avail_idx, &vq->avail->idx);
 	if (r) {
 		vq_err(vq, "Failed to check avail idx at %p: %d\n",
 		       &vq->avail->idx, r);

drivers/vhost/vhost.h

@@ -76,6 +76,13 @@ struct vhost_umem {
 	int numem;
 };
 
+enum vhost_uaddr_type {
+	VHOST_ADDR_DESC = 0,
+	VHOST_ADDR_AVAIL = 1,
+	VHOST_ADDR_USED = 2,
+	VHOST_NUM_ADDRS = 3,
+};
+
 /* The virtqueue structure describes a queue attached to a device. */
 struct vhost_virtqueue {
 	struct vhost_dev *dev;
@@ -86,6 +93,7 @@ struct vhost_virtqueue {
 	struct vring_desc __user *desc;
 	struct vring_avail __user *avail;
 	struct vring_used __user *used;
+	const struct vhost_umem_node *meta_iotlb[VHOST_NUM_ADDRS];
 	struct file *kick;
 	struct file *call;
 	struct file *error;

drivers/virtio/virtio_balloon.c

@@ -413,7 +413,8 @@ static int init_vqs(struct virtio_balloon *vb)
 	 * optionally stat.
 	 */
 	nvqs = virtio_has_feature(vb->vdev, VIRTIO_BALLOON_F_STATS_VQ) ? 3 : 2;
-	err = vb->vdev->config->find_vqs(vb->vdev, nvqs, vqs, callbacks, names);
+	err = vb->vdev->config->find_vqs(vb->vdev, nvqs, vqs, callbacks, names,
+			NULL);
 	if (err)
 		return err;
 

drivers/virtio/virtio_input.c

@@ -173,7 +173,8 @@ static int virtinput_init_vqs(struct virtio_input *vi)
 	static const char * const names[] = { "events", "status" };
 	int err;
 
-	err = vi->vdev->config->find_vqs(vi->vdev, 2, vqs, cbs, names);
+	err = vi->vdev->config->find_vqs(vi->vdev, 2, vqs, cbs, names,
+			NULL);
 	if (err)
 		return err;
 	vi->evt = vqs[0];

drivers/virtio/virtio_mmio.c

@@ -70,7 +70,7 @@
 #include <linux/spinlock.h>
 #include <linux/virtio.h>
 #include <linux/virtio_config.h>
-#include <linux/virtio_mmio.h>
+#include <uapi/linux/virtio_mmio.h>
 #include <linux/virtio_ring.h>
 
 
@@ -446,7 +446,8 @@ error_available:
 static int vm_find_vqs(struct virtio_device *vdev, unsigned nvqs,
 		       struct virtqueue *vqs[],
 		       vq_callback_t *callbacks[],
-		       const char * const names[])
+		       const char * const names[],
+		       struct irq_affinity *desc)
 {
 	struct virtio_mmio_device *vm_dev = to_virtio_mmio_device(vdev);
 	unsigned int irq = platform_get_irq(vm_dev->pdev, 0);

drivers/virtio/virtio_pci_common.c

@@ -33,10 +33,8 @@ void vp_synchronize_vectors(struct virtio_device *vdev)
 	struct virtio_pci_device *vp_dev = to_vp_device(vdev);
 	int i;
 
-	if (vp_dev->intx_enabled)
-		synchronize_irq(vp_dev->pci_dev->irq);
-
-	for (i = 0; i < vp_dev->msix_vectors; ++i)
+	synchronize_irq(pci_irq_vector(vp_dev->pci_dev, 0));
+	for (i = 1; i < vp_dev->msix_vectors; i++)
 		synchronize_irq(pci_irq_vector(vp_dev->pci_dev, i));
 }
 
@@ -62,16 +60,13 @@ static irqreturn_t vp_config_changed(int irq, void *opaque)
 static irqreturn_t vp_vring_interrupt(int irq, void *opaque)
 {
 	struct virtio_pci_device *vp_dev = opaque;
-	struct virtio_pci_vq_info *info;
 	irqreturn_t ret = IRQ_NONE;
-	unsigned long flags;
+	struct virtqueue *vq;
 
-	spin_lock_irqsave(&vp_dev->lock, flags);
-	list_for_each_entry(info, &vp_dev->virtqueues, node) {
-		if (vring_interrupt(irq, info->vq) == IRQ_HANDLED)
+	list_for_each_entry(vq, &vp_dev->vdev.vqs, list) {
+		if (vq->callback && vring_interrupt(irq, vq) == IRQ_HANDLED)
 			ret = IRQ_HANDLED;
 	}
-	spin_unlock_irqrestore(&vp_dev->lock, flags);
 
 	return ret;
 }
@@ -102,237 +97,185 @@ static irqreturn_t vp_interrupt(int irq, void *opaque)
 	return vp_vring_interrupt(irq, opaque);
 }
 
-static int vp_request_msix_vectors(struct virtio_device *vdev, int nvectors,
-				   bool per_vq_vectors)
+static void vp_remove_vqs(struct virtio_device *vdev)
 {
 	struct virtio_pci_device *vp_dev = to_vp_device(vdev);
-	const char *name = dev_name(&vp_dev->vdev.dev);
-	unsigned i, v;
-	int err = -ENOMEM;
-
-	vp_dev->msix_vectors = nvectors;
-
-	vp_dev->msix_names = kmalloc(nvectors * sizeof *vp_dev->msix_names,
-				     GFP_KERNEL);
-	if (!vp_dev->msix_names)
-		goto error;
-	vp_dev->msix_affinity_masks
-		= kzalloc(nvectors * sizeof *vp_dev->msix_affinity_masks,
-			  GFP_KERNEL);
-	if (!vp_dev->msix_affinity_masks)
-		goto error;
-	for (i = 0; i < nvectors; ++i)
-		if (!alloc_cpumask_var(&vp_dev->msix_affinity_masks[i],
-					GFP_KERNEL))
-			goto error;
-
-	err = pci_alloc_irq_vectors(vp_dev->pci_dev, nvectors, nvectors,
-			PCI_IRQ_MSIX);
-	if (err < 0)
-		goto error;
-	vp_dev->msix_enabled = 1;
-
-	/* Set the vector used for configuration */
-	v = vp_dev->msix_used_vectors;
-	snprintf(vp_dev->msix_names[v], sizeof *vp_dev->msix_names,
-		 "%s-config", name);
-	err = request_irq(pci_irq_vector(vp_dev->pci_dev, v),
-			  vp_config_changed, 0, vp_dev->msix_names[v],
-			  vp_dev);
-	if (err)
-		goto error;
-	++vp_dev->msix_used_vectors;
-
-	v = vp_dev->config_vector(vp_dev, v);
-	/* Verify we had enough resources to assign the vector */
-	if (v == VIRTIO_MSI_NO_VECTOR) {
-		err = -EBUSY;
-		goto error;
-	}
-
-	if (!per_vq_vectors) {
-		/* Shared vector for all VQs */
-		v = vp_dev->msix_used_vectors;
-		snprintf(vp_dev->msix_names[v], sizeof *vp_dev->msix_names,
-			 "%s-virtqueues", name);
-		err = request_irq(pci_irq_vector(vp_dev->pci_dev, v),
-				  vp_vring_interrupt, 0, vp_dev->msix_names[v],
-				  vp_dev);
-		if (err)
-			goto error;
-		++vp_dev->msix_used_vectors;
-	}
-	return 0;
-error:
-	return err;
-}
-
-static struct virtqueue *vp_setup_vq(struct virtio_device *vdev, unsigned index,
-				     void (*callback)(struct virtqueue *vq),
-				     const char *name,
-				     u16 msix_vec)
-{
-	struct virtio_pci_device *vp_dev = to_vp_device(vdev);
-	struct virtio_pci_vq_info *info = kmalloc(sizeof *info, GFP_KERNEL);
-	struct virtqueue *vq;
-	unsigned long flags;
-
-	/* fill out our structure that represents an active queue */
-	if (!info)
-		return ERR_PTR(-ENOMEM);
+	struct virtqueue *vq, *n;
 
-	vq = vp_dev->setup_vq(vp_dev, info, index, callback, name, msix_vec);
-	if (IS_ERR(vq))
-		goto out_info;
+	list_for_each_entry_safe(vq, n, &vdev->vqs, list) {
+		if (vp_dev->msix_vector_map) {
+			int v = vp_dev->msix_vector_map[vq->index];
 
-	info->vq = vq;
-	if (callback) {
-		spin_lock_irqsave(&vp_dev->lock, flags);
-		list_add(&info->node, &vp_dev->virtqueues);
-		spin_unlock_irqrestore(&vp_dev->lock, flags);
-	} else {
-		INIT_LIST_HEAD(&info->node);
+			if (v != VIRTIO_MSI_NO_VECTOR)
+				free_irq(pci_irq_vector(vp_dev->pci_dev, v),
+					vq);
+		}
+		vp_dev->del_vq(vq);
 	}
-
-	vp_dev->vqs[index] = info;
-	return vq;
-
-out_info:
-	kfree(info);
-	return vq;
-}
-
-static void vp_del_vq(struct virtqueue *vq)
-{
-	struct virtio_pci_device *vp_dev = to_vp_device(vq->vdev);
-	struct virtio_pci_vq_info *info = vp_dev->vqs[vq->index];
-	unsigned long flags;
-
-	spin_lock_irqsave(&vp_dev->lock, flags);
-	list_del(&info->node);
-	spin_unlock_irqrestore(&vp_dev->lock, flags);
-
-	vp_dev->del_vq(info);
-	kfree(info);
 }
 
 /* the config->del_vqs() implementation */
 void vp_del_vqs(struct virtio_device *vdev)
 {
 	struct virtio_pci_device *vp_dev = to_vp_device(vdev);
-	struct virtqueue *vq, *n;
 	int i;
 
-	list_for_each_entry_safe(vq, n, &vdev->vqs, list) {
-		if (vp_dev->per_vq_vectors) {
-			int v = vp_dev->vqs[vq->index]->msix_vector;
-
-			if (v != VIRTIO_MSI_NO_VECTOR)
-				free_irq(pci_irq_vector(vp_dev->pci_dev, v),
-					vq);
-		}
-		vp_del_vq(vq);
-	}
-	vp_dev->per_vq_vectors = false;
-
-	if (vp_dev->intx_enabled) {
-		free_irq(vp_dev->pci_dev->irq, vp_dev);
-		vp_dev->intx_enabled = 0;
-	}
+	if (WARN_ON_ONCE(list_empty_careful(&vdev->vqs)))
+		return;
 
-	for (i = 0; i < vp_dev->msix_used_vectors; ++i)
-		free_irq(pci_irq_vector(vp_dev->pci_dev, i), vp_dev);
+	vp_remove_vqs(vdev);
 
-	for (i = 0; i < vp_dev->msix_vectors; i++)
-		if (vp_dev->msix_affinity_masks[i])
+	if (vp_dev->pci_dev->msix_enabled) {
+		for (i = 0; i < vp_dev->msix_vectors; i++)
 			free_cpumask_var(vp_dev->msix_affinity_masks[i]);
 
-	if (vp_dev->msix_enabled) {
 		/* Disable the vector used for configuration */
 		vp_dev->config_vector(vp_dev, VIRTIO_MSI_NO_VECTOR);
 
-		pci_free_irq_vectors(vp_dev->pci_dev);
-		vp_dev->msix_enabled = 0;
+		kfree(vp_dev->msix_affinity_masks);
+		kfree(vp_dev->msix_names);
+		kfree(vp_dev->msix_vector_map);
 	}
 
-	vp_dev->msix_vectors = 0;
-	vp_dev->msix_used_vectors = 0;
-	kfree(vp_dev->msix_names);
-	vp_dev->msix_names = NULL;
-	kfree(vp_dev->msix_affinity_masks);
-	vp_dev->msix_affinity_masks = NULL;
-	kfree(vp_dev->vqs);
-	vp_dev->vqs = NULL;
+	free_irq(pci_irq_vector(vp_dev->pci_dev, 0), vp_dev);
+	pci_free_irq_vectors(vp_dev->pci_dev);
 }
 
 static int vp_find_vqs_msix(struct virtio_device *vdev, unsigned nvqs,
-			      struct virtqueue *vqs[],
-			      vq_callback_t *callbacks[],
-			      const char * const names[],
-			      bool per_vq_vectors)
+		struct virtqueue *vqs[], vq_callback_t *callbacks[],
+		const char * const names[], struct irq_affinity *desc)
 {
 	struct virtio_pci_device *vp_dev = to_vp_device(vdev);
+	const char *name = dev_name(&vp_dev->vdev.dev);
+	int i, err = -ENOMEM, allocated_vectors, nvectors;
+	unsigned flags = PCI_IRQ_MSIX;
+	bool shared = false;
 	u16 msix_vec;
-	int i, err, nvectors, allocated_vectors;
 
-	vp_dev->vqs = kcalloc(nvqs, sizeof(*vp_dev->vqs), GFP_KERNEL);
-	if (!vp_dev->vqs)
-		return -ENOMEM;
+	if (desc) {
+		flags |= PCI_IRQ_AFFINITY;
+		desc->pre_vectors++; /* virtio config vector */
+	}
 
-	if (per_vq_vectors) {
-		/* Best option: one for change interrupt, one per vq. */
-		nvectors = 1;
-		for (i = 0; i < nvqs; ++i)
-			if (callbacks[i])
-				++nvectors;
-	} else {
-		/* Second best: one for change, shared for all vqs. */
-		nvectors = 2;
+	nvectors = 1;
+	for (i = 0; i < nvqs; i++)
+		if (callbacks[i])
+			nvectors++;
+
+	/* Try one vector per queue first. */
+	err = pci_alloc_irq_vectors_affinity(vp_dev->pci_dev, nvectors,
+			nvectors, flags, desc);
+	if (err < 0) {
+		/* Fallback to one vector for config, one shared for queues. */
+		shared = true;
+		err = pci_alloc_irq_vectors(vp_dev->pci_dev, 2, 2,
+				PCI_IRQ_MSIX);
+		if (err < 0)
+			return err;
+	}
+	if (err < 0)
+		return err;
+
+	vp_dev->msix_vectors = nvectors;
+	vp_dev->msix_names = kmalloc_array(nvectors,
+			sizeof(*vp_dev->msix_names), GFP_KERNEL);
+	if (!vp_dev->msix_names)
+		goto out_free_irq_vectors;
+
+	vp_dev->msix_affinity_masks = kcalloc(nvectors,
+			sizeof(*vp_dev->msix_affinity_masks), GFP_KERNEL);
+	if (!vp_dev->msix_affinity_masks)
+		goto out_free_msix_names;
+
+	for (i = 0; i < nvectors; ++i) {
+		if (!alloc_cpumask_var(&vp_dev->msix_affinity_masks[i],
+				GFP_KERNEL))
+			goto out_free_msix_affinity_masks;
 	}
 
-	err = vp_request_msix_vectors(vdev, nvectors, per_vq_vectors);
+	/* Set the vector used for configuration */
+	snprintf(vp_dev->msix_names[0], sizeof(*vp_dev->msix_names),
+		 "%s-config", name);
+	err = request_irq(pci_irq_vector(vp_dev->pci_dev, 0), vp_config_changed,
+			0, vp_dev->msix_names[0], vp_dev);
 	if (err)
-		goto error_find;
+		goto out_free_msix_affinity_masks;
+
+	/* Verify we had enough resources to assign the vector */
+	if (vp_dev->config_vector(vp_dev, 0) == VIRTIO_MSI_NO_VECTOR) {
+		err = -EBUSY;
+		goto out_free_config_irq;
+	}
+
+	vp_dev->msix_vector_map = kmalloc_array(nvqs,
+			sizeof(*vp_dev->msix_vector_map), GFP_KERNEL);
+	if (!vp_dev->msix_vector_map)
+		goto out_disable_config_irq;
 
-	vp_dev->per_vq_vectors = per_vq_vectors;
-	allocated_vectors = vp_dev->msix_used_vectors;
+	allocated_vectors = 1; /* vector 0 is the config interrupt */
 	for (i = 0; i < nvqs; ++i) {
 		if (!names[i]) {
 			vqs[i] = NULL;
 			continue;
 		}
 
-		if (!callbacks[i])
-			msix_vec = VIRTIO_MSI_NO_VECTOR;
-		else if (vp_dev->per_vq_vectors)
-			msix_vec = allocated_vectors++;
+		if (callbacks[i])
+			msix_vec = allocated_vectors;
 		else
-			msix_vec = VP_MSIX_VQ_VECTOR;
-		vqs[i] = vp_setup_vq(vdev, i, callbacks[i], names[i], msix_vec);
+			msix_vec = VIRTIO_MSI_NO_VECTOR;
+
+		vqs[i] = vp_dev->setup_vq(vp_dev, i, callbacks[i], names[i],
+				msix_vec);
 		if (IS_ERR(vqs[i])) {
 			err = PTR_ERR(vqs[i]);
-			goto error_find;
+			goto out_remove_vqs;
 		}
 
-		if (!vp_dev->per_vq_vectors || msix_vec == VIRTIO_MSI_NO_VECTOR)
+		if (msix_vec == VIRTIO_MSI_NO_VECTOR) {
+			vp_dev->msix_vector_map[i] = VIRTIO_MSI_NO_VECTOR;
 			continue;
+		}
 
-		/* allocate per-vq irq if available and necessary */
-		snprintf(vp_dev->msix_names[msix_vec],
-			 sizeof *vp_dev->msix_names,
-			 "%s-%s",
+		snprintf(vp_dev->msix_names[i + 1],
+			 sizeof(*vp_dev->msix_names), "%s-%s",
 			 dev_name(&vp_dev->vdev.dev), names[i]);
 		err = request_irq(pci_irq_vector(vp_dev->pci_dev, msix_vec),
-				  vring_interrupt, 0,
-				  vp_dev->msix_names[msix_vec],
-				  vqs[i]);
-		if (err)
-			goto error_find;
+				  vring_interrupt, IRQF_SHARED,
+				  vp_dev->msix_names[i + 1], vqs[i]);
+		if (err) {
+			/* don't free this irq on error */
+			vp_dev->msix_vector_map[i] = VIRTIO_MSI_NO_VECTOR;
+			goto out_remove_vqs;
+		}
+		vp_dev->msix_vector_map[i] = msix_vec;
+
+		/*
+		 * Use a different vector for each queue if they are available,
+		 * else share the same vector for all VQs.
+		 */
+		if (!shared)
+			allocated_vectors++;
 	}
+
 	return 0;
 
-error_find:
-	vp_del_vqs(vdev);
+out_remove_vqs:
+	vp_remove_vqs(vdev);
+	kfree(vp_dev->msix_vector_map);
+out_disable_config_irq:
+	vp_dev->config_vector(vp_dev, VIRTIO_MSI_NO_VECTOR);
+out_free_config_irq:
+	free_irq(pci_irq_vector(vp_dev->pci_dev, 0), vp_dev);
+out_free_msix_affinity_masks:
+	for (i = 0; i < nvectors; i++) {
+		if (vp_dev->msix_affinity_masks[i])
+			free_cpumask_var(vp_dev->msix_affinity_masks[i]);
+	}
+	kfree(vp_dev->msix_affinity_masks);
+out_free_msix_names:
+	kfree(vp_dev->msix_names);
+out_free_irq_vectors:
+	pci_free_irq_vectors(vp_dev->pci_dev);
 	return err;
 }
 
@@ -343,53 +286,42 @@ static int vp_find_vqs_intx(struct virtio_device *vdev, unsigned nvqs,
 	struct virtio_pci_device *vp_dev = to_vp_device(vdev);
 	int i, err;
 
-	vp_dev->vqs = kcalloc(nvqs, sizeof(*vp_dev->vqs), GFP_KERNEL);
-	if (!vp_dev->vqs)
-		return -ENOMEM;
-
 	err = request_irq(vp_dev->pci_dev->irq, vp_interrupt, IRQF_SHARED,
 			dev_name(&vdev->dev), vp_dev);
 	if (err)
-		goto out_del_vqs;
+		return err;
 
-	vp_dev->intx_enabled = 1;
-	vp_dev->per_vq_vectors = false;
 	for (i = 0; i < nvqs; ++i) {
 		if (!names[i]) {
 			vqs[i] = NULL;
 			continue;
 		}
-		vqs[i] = vp_setup_vq(vdev, i, callbacks[i], names[i],
+		vqs[i] = vp_dev->setup_vq(vp_dev, i, callbacks[i], names[i],
 				VIRTIO_MSI_NO_VECTOR);
 		if (IS_ERR(vqs[i])) {
 			err = PTR_ERR(vqs[i]);
-			goto out_del_vqs;
+			goto out_remove_vqs;
 		}
 	}
 
 	return 0;
-out_del_vqs:
-	vp_del_vqs(vdev);
+
+out_remove_vqs:
+	vp_remove_vqs(vdev);
+	free_irq(pci_irq_vector(vp_dev->pci_dev, 0), vp_dev);
 	return err;
 }
 
 /* the config->find_vqs() implementation */
 int vp_find_vqs(struct virtio_device *vdev, unsigned nvqs,
-		struct virtqueue *vqs[],
-		vq_callback_t *callbacks[],
-		const char * const names[])
+		struct virtqueue *vqs[], vq_callback_t *callbacks[],
+		const char * const names[], struct irq_affinity *desc)
 {
 	int err;
 
-	/* Try MSI-X with one vector per queue. */
-	err = vp_find_vqs_msix(vdev, nvqs, vqs, callbacks, names, true);
+	err = vp_find_vqs_msix(vdev, nvqs, vqs, callbacks, names, desc);
 	if (!err)
 		return 0;
-	/* Fallback: MSI-X with one vector for config, one shared for queues. */
-	err = vp_find_vqs_msix(vdev, nvqs, vqs, callbacks, names, false);
-	if (!err)
-		return 0;
-	/* Finally fall back to regular interrupts. */
 	return vp_find_vqs_intx(vdev, nvqs, vqs, callbacks, names);
 }
 
@@ -409,16 +341,15 @@ int vp_set_vq_affinity(struct virtqueue *vq, int cpu)
 {
 	struct virtio_device *vdev = vq->vdev;
 	struct virtio_pci_device *vp_dev = to_vp_device(vdev);
-	struct virtio_pci_vq_info *info = vp_dev->vqs[vq->index];
-	struct cpumask *mask;
-	unsigned int irq;
 
 	if (!vq->callback)
 		return -EINVAL;
 
-	if (vp_dev->msix_enabled) {
-		mask = vp_dev->msix_affinity_masks[info->msix_vector];
-		irq = pci_irq_vector(vp_dev->pci_dev, info->msix_vector);
+	if (vp_dev->pci_dev->msix_enabled) {
+		int vec = vp_dev->msix_vector_map[vq->index];
+		struct cpumask *mask = vp_dev->msix_affinity_masks[vec];
+		unsigned int irq = pci_irq_vector(vp_dev->pci_dev, vec);
+
 		if (cpu == -1)
 			irq_set_affinity_hint(irq, NULL);
 		else {
@@ -430,6 +361,17 @@ int vp_set_vq_affinity(struct virtqueue *vq, int cpu)
 	return 0;
 }
 
+const struct cpumask *vp_get_vq_affinity(struct virtio_device *vdev, int index)
+{
+	struct virtio_pci_device *vp_dev = to_vp_device(vdev);
+	unsigned int *map = vp_dev->msix_vector_map;
+
+	if (!map || map[index] == VIRTIO_MSI_NO_VECTOR)
+		return NULL;
+
+	return pci_irq_get_affinity(vp_dev->pci_dev, map[index]);
+}
+
 #ifdef CONFIG_PM_SLEEP
 static int virtio_pci_freeze(struct device *dev)
 {
@@ -498,8 +440,6 @@ static int virtio_pci_probe(struct pci_dev *pci_dev,
 	vp_dev->vdev.dev.parent = &pci_dev->dev;
 	vp_dev->vdev.dev.release = virtio_pci_release_dev;
 	vp_dev->pci_dev = pci_dev;
-	INIT_LIST_HEAD(&vp_dev->virtqueues);
-	spin_lock_init(&vp_dev->lock);
 
 	/* enable the device */
 	rc = pci_enable_device(pci_dev);

drivers/virtio/virtio_pci_common.h

@@ -31,17 +31,6 @@
 #include <linux/highmem.h>
 #include <linux/spinlock.h>
 
-struct virtio_pci_vq_info {
-	/* the actual virtqueue */
-	struct virtqueue *vq;
-
-	/* the list node for the virtqueues list */
-	struct list_head node;
-
-	/* MSI-X vector (or none) */
-	unsigned msix_vector;
-};
-
 /* Our device structure */
 struct virtio_pci_device {
 	struct virtio_device vdev;
@@ -75,47 +64,25 @@ struct virtio_pci_device {
 	/* the IO mapping for the PCI config space */
 	void __iomem *ioaddr;
 
-	/* a list of queues so we can dispatch IRQs */
-	spinlock_t lock;
-	struct list_head virtqueues;
-
-	/* array of all queues for house-keeping */
-	struct virtio_pci_vq_info **vqs;
-
-	/* MSI-X support */
-	int msix_enabled;
-	int intx_enabled;
 	cpumask_var_t *msix_affinity_masks;
 	/* Name strings for interrupts. This size should be enough,
 	 * and I'm too lazy to allocate each name separately. */
 	char (*msix_names)[256];
-	/* Number of available vectors */
-	unsigned msix_vectors;
-	/* Vectors allocated, excluding per-vq vectors if any */
-	unsigned msix_used_vectors;
-
-	/* Whether we have vector per vq */
-	bool per_vq_vectors;
+	/* Total Number of MSI-X vectors (including per-VQ ones). */
+	int msix_vectors;
+	/* Map of per-VQ MSI-X vectors, may be NULL */
+	unsigned *msix_vector_map;
 
 	struct virtqueue *(*setup_vq)(struct virtio_pci_device *vp_dev,
-				      struct virtio_pci_vq_info *info,
 				      unsigned idx,
 				      void (*callback)(struct virtqueue *vq),
 				      const char *name,
 				      u16 msix_vec);
-	void (*del_vq)(struct virtio_pci_vq_info *info);
+	void (*del_vq)(struct virtqueue *vq);
 
 	u16 (*config_vector)(struct virtio_pci_device *vp_dev, u16 vector);
 };
 
-/* Constants for MSI-X */
-/* Use first vector for configuration changes, second and the rest for
- * virtqueues Thus, we need at least 2 vectors for MSI. */
-enum {
-	VP_MSIX_CONFIG_VECTOR = 0,
-	VP_MSIX_VQ_VECTOR = 1,
-};
-
 /* Convert a generic virtio device to our structure */
 static struct virtio_pci_device *to_vp_device(struct virtio_device *vdev)
 {
@@ -130,9 +97,8 @@ bool vp_notify(struct virtqueue *vq);
 void vp_del_vqs(struct virtio_device *vdev);
 /* the config->find_vqs() implementation */
 int vp_find_vqs(struct virtio_device *vdev, unsigned nvqs,
-		       struct virtqueue *vqs[],
-		       vq_callback_t *callbacks[],
-		       const char * const names[]);
+		struct virtqueue *vqs[], vq_callback_t *callbacks[],
+		const char * const names[], struct irq_affinity *desc);
 const char *vp_bus_name(struct virtio_device *vdev);
 
 /* Setup the affinity for a virtqueue:
@@ -142,6 +108,8 @@ const char *vp_bus_name(struct virtio_device *vdev);
  */
 int vp_set_vq_affinity(struct virtqueue *vq, int cpu);
 
+const struct cpumask *vp_get_vq_affinity(struct virtio_device *vdev, int index);
+
 #if IS_ENABLED(CONFIG_VIRTIO_PCI_LEGACY)
 int virtio_pci_legacy_probe(struct virtio_pci_device *);
 void virtio_pci_legacy_remove(struct virtio_pci_device *);

drivers/virtio/virtio_pci_legacy.c

@@ -112,7 +112,6 @@ static u16 vp_config_vector(struct virtio_pci_device *vp_dev, u16 vector)
 }
 
 static struct virtqueue *setup_vq(struct virtio_pci_device *vp_dev,
-				  struct virtio_pci_vq_info *info,
 				  unsigned index,
 				  void (*callback)(struct virtqueue *vq),
 				  const char *name,
@@ -130,8 +129,6 @@ static struct virtqueue *setup_vq(struct virtio_pci_device *vp_dev,
 	if (!num || ioread32(vp_dev->ioaddr + VIRTIO_PCI_QUEUE_PFN))
 		return ERR_PTR(-ENOENT);
 
-	info->msix_vector = msix_vec;
-
 	/* create the vring */
 	vq = vring_create_virtqueue(index, num,
 				    VIRTIO_PCI_VRING_ALIGN, &vp_dev->vdev,
@@ -162,14 +159,13 @@ out_deactivate:
 	return ERR_PTR(err);
 }
 
-static void del_vq(struct virtio_pci_vq_info *info)
+static void del_vq(struct virtqueue *vq)
 {
-	struct virtqueue *vq = info->vq;
 	struct virtio_pci_device *vp_dev = to_vp_device(vq->vdev);
 
 	iowrite16(vq->index, vp_dev->ioaddr + VIRTIO_PCI_QUEUE_SEL);
 
-	if (vp_dev->msix_enabled) {
+	if (vp_dev->pci_dev->msix_enabled) {
 		iowrite16(VIRTIO_MSI_NO_VECTOR,
 			  vp_dev->ioaddr + VIRTIO_MSI_QUEUE_VECTOR);
 		/* Flush the write out to device */
@@ -194,6 +190,7 @@ static const struct virtio_config_ops virtio_pci_config_ops = {
 	.finalize_features = vp_finalize_features,
 	.bus_name	= vp_bus_name,
 	.set_vq_affinity = vp_set_vq_affinity,
+	.get_vq_affinity = vp_get_vq_affinity,
 };
 
 /* the PCI probing function */

drivers/virtio/virtio_pci_modern.c

@@ -293,7 +293,6 @@ static u16 vp_config_vector(struct virtio_pci_device *vp_dev, u16 vector)
 }
 
 static struct virtqueue *setup_vq(struct virtio_pci_device *vp_dev,
-				  struct virtio_pci_vq_info *info,
 				  unsigned index,
 				  void (*callback)(struct virtqueue *vq),
 				  const char *name,
@@ -323,8 +322,6 @@ static struct virtqueue *setup_vq(struct virtio_pci_device *vp_dev,
 	/* get offset of notification word for this vq */
 	off = vp_ioread16(&cfg->queue_notify_off);
 
-	info->msix_vector = msix_vec;
-
 	/* create the vring */
 	vq = vring_create_virtqueue(index, num,
 				    SMP_CACHE_BYTES, &vp_dev->vdev,
@@ -387,13 +384,12 @@ err_map_notify:
 }
 
 static int vp_modern_find_vqs(struct virtio_device *vdev, unsigned nvqs,
-			      struct virtqueue *vqs[],
-			      vq_callback_t *callbacks[],
-			      const char * const names[])
+		struct virtqueue *vqs[], vq_callback_t *callbacks[],
+		const char * const names[], struct irq_affinity *desc)
 {
 	struct virtio_pci_device *vp_dev = to_vp_device(vdev);
 	struct virtqueue *vq;
-	int rc = vp_find_vqs(vdev, nvqs, vqs, callbacks, names);
+	int rc = vp_find_vqs(vdev, nvqs, vqs, callbacks, names, desc);
 
 	if (rc)
 		return rc;
@@ -409,14 +405,13 @@ static int vp_modern_find_vqs(struct virtio_device *vdev, unsigned nvqs,
 	return 0;
 }
 
-static void del_vq(struct virtio_pci_vq_info *info)
+static void del_vq(struct virtqueue *vq)
 {
-	struct virtqueue *vq = info->vq;
 	struct virtio_pci_device *vp_dev = to_vp_device(vq->vdev);
 
 	vp_iowrite16(vq->index, &vp_dev->common->queue_select);
 
-	if (vp_dev->msix_enabled) {
+	if (vp_dev->pci_dev->msix_enabled) {
 		vp_iowrite16(VIRTIO_MSI_NO_VECTOR,
 			     &vp_dev->common->queue_msix_vector);
 		/* Flush the write out to device */
@@ -442,6 +437,7 @@ static const struct virtio_config_ops virtio_pci_config_nodev_ops = {
 	.finalize_features = vp_finalize_features,
 	.bus_name	= vp_bus_name,
 	.set_vq_affinity = vp_set_vq_affinity,
+	.get_vq_affinity = vp_get_vq_affinity,
 };
 
 static const struct virtio_config_ops virtio_pci_config_ops = {
@@ -457,6 +453,7 @@ static const struct virtio_config_ops virtio_pci_config_ops = {
 	.finalize_features = vp_finalize_features,
 	.bus_name	= vp_bus_name,
 	.set_vq_affinity = vp_set_vq_affinity,
+	.get_vq_affinity = vp_get_vq_affinity,
 };
 
 /**

include/linux/blk-mq-virtio.h

@@ -0,0 +1,10 @@
+#ifndef _LINUX_BLK_MQ_VIRTIO_H
+#define _LINUX_BLK_MQ_VIRTIO_H
+
+struct blk_mq_tag_set;
+struct virtio_device;
+
+int blk_mq_virtio_map_queues(struct blk_mq_tag_set *set,
+		struct virtio_device *vdev, int first_vec);
+
+#endif /* _LINUX_BLK_MQ_VIRTIO_H */

include/linux/cpuhotplug.h

@@ -26,7 +26,6 @@ enum cpuhp_state {
 	CPUHP_ARM_OMAP_WAKE_DEAD,
 	CPUHP_IRQ_POLL_DEAD,
 	CPUHP_BLOCK_SOFTIRQ_DEAD,
-	CPUHP_VIRT_SCSI_DEAD,
 	CPUHP_ACPI_CPUDRV_DEAD,
 	CPUHP_S390_PFAULT_DEAD,
 	CPUHP_BLK_MQ_DEAD,

include/linux/virtio_config.h

@@ -7,6 +7,8 @@
 #include <linux/virtio_byteorder.h>
 #include <uapi/linux/virtio_config.h>
 
+struct irq_affinity;
+
 /**
  * virtio_config_ops - operations for configuring a virtio device
  * @get: read the value of a configuration field
@@ -56,6 +58,7 @@
  *      This returns a pointer to the bus name a la pci_name from which
  *      the caller can then copy.
  * @set_vq_affinity: set the affinity for a virtqueue.
+ * @get_vq_affinity: get the affinity for a virtqueue (optional).
  */
 typedef void vq_callback_t(struct virtqueue *);
 struct virtio_config_ops {
@@ -68,14 +71,15 @@ struct virtio_config_ops {
 	void (*set_status)(struct virtio_device *vdev, u8 status);
 	void (*reset)(struct virtio_device *vdev);
 	int (*find_vqs)(struct virtio_device *, unsigned nvqs,
-			struct virtqueue *vqs[],
-			vq_callback_t *callbacks[],
-			const char * const names[]);
+			struct virtqueue *vqs[], vq_callback_t *callbacks[],
+			const char * const names[], struct irq_affinity *desc);
 	void (*del_vqs)(struct virtio_device *);
 	u64 (*get_features)(struct virtio_device *vdev);
 	int (*finalize_features)(struct virtio_device *vdev);
 	const char *(*bus_name)(struct virtio_device *vdev);
 	int (*set_vq_affinity)(struct virtqueue *vq, int cpu);
+	const struct cpumask *(*get_vq_affinity)(struct virtio_device *vdev,
+			int index);
 };
 
 /* If driver didn't advertise the feature, it will never appear. */
@@ -169,7 +173,7 @@ struct virtqueue *virtio_find_single_vq(struct virtio_device *vdev,
 	vq_callback_t *callbacks[] = { c };
 	const char *names[] = { n };
 	struct virtqueue *vq;
-	int err = vdev->config->find_vqs(vdev, 1, &vq, callbacks, names);
+	int err = vdev->config->find_vqs(vdev, 1, &vq, callbacks, names, NULL);
 	if (err < 0)
 		return ERR_PTR(err);
 	return vq;

include/uapi/linux/Kbuild

@@ -466,6 +466,7 @@ header-y += virtio_console.h
 header-y += virtio_gpu.h
 header-y += virtio_ids.h
 header-y += virtio_input.h
+header-y += virtio_mmio.h
 header-y += virtio_net.h
 header-y += virtio_pci.h
 header-y += virtio_ring.h

include/uapi/linux/virtio_pci.h

@@ -79,7 +79,7 @@
  * configuration space */
 #define VIRTIO_PCI_CONFIG_OFF(msix_enabled)	((msix_enabled) ? 24 : 20)
 /* Deprecated: please use VIRTIO_PCI_CONFIG_OFF instead */
-#define VIRTIO_PCI_CONFIG(dev)	VIRTIO_PCI_CONFIG_OFF((dev)->msix_enabled)
+#define VIRTIO_PCI_CONFIG(dev)	VIRTIO_PCI_CONFIG_OFF((dev)->pci_dev->msix_enabled)
 
 /* Virtio ABI version, this must match exactly */
 #define VIRTIO_PCI_ABI_VERSION		0

net/vmw_vsock/virtio_transport.c

@@ -532,7 +532,8 @@ static int virtio_vsock_probe(struct virtio_device *vdev)
 	vsock->vdev = vdev;
 
 	ret = vsock->vdev->config->find_vqs(vsock->vdev, VSOCK_VQ_MAX,
-					    vsock->vqs, callbacks, names);
+					    vsock->vqs, callbacks, names,
+					    NULL);
 	if (ret < 0)
 		goto out;