[PATCH V4 0/5] virtio-scsi multiqueue

[PATCH V4 0/5] virtio-scsi multiqueue

[Wanlong Gao]

This series implements virtio-scsi queue steering, which gives
performance improvements of up to 50% (measured both with QEMU and
tcm_vhost backends).

This version rebased on Rusty's virtio ring rework patches.
We hope this can go into virtio-next together with the virtio ring
rework pathes.


V4: rebase on virtio ring rework patches (rusty's pending-rebases branch)

V3 and be found http://marc.info/?l=linux-virtualization&m=136067440717154&w=2


It would probably be easier to get it in via Rusty's tree
because of the prerequisites.  James, can we get your Acked-by?

Paolo Bonzini (4):
  virtio-scsi: redo allocation of target data
  virtio-scsi: pass struct virtio_scsi to virtqueue completion function
  virtio-scsi: push vq lock/unlock into virtscsi_vq_done
  virtio-scsi: introduce multiqueue support

Wanlong Gao (1):
  virtio-scsi: reset virtqueue affinity when doing cpu hotplug

 drivers/scsi/virtio_scsi.c | 364 ++++++++++++++++++++++++++++++++++++---------
 1 file changed, 291 insertions(+), 73 deletions(-)

-- 
1.8.2.rc2

[PATCH V4 1/5] virtio-scsi: redo allocation of target data

[Wanlong Gao]

From: Paolo Bonzini <pbonzini@redhat.com>

virtio_scsi_target_state is now empty, but we will find new uses
for it in the next few patches.  However, dropping the sglist lets
us turn the array-of-pointers into a simple array, which simplifies
the allocation.

However, we do not leave the virtio_scsi_target_state structs in the
flexible array member at the end of struct virtio_scsi, because
we will place the virtqueues there in the next patches.

Cc: linux-scsi@vger.kernel.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>
---
 drivers/scsi/virtio_scsi.c | 45 ++++++++++++---------------------------------
 1 file changed, 12 insertions(+), 33 deletions(-)

diff --git a/drivers/scsi/virtio_scsi.c b/drivers/scsi/virtio_scsi.c
index 6307099..75ccde1 100644
--- a/drivers/scsi/virtio_scsi.c
+++ b/drivers/scsi/virtio_scsi.c
@@ -76,7 +76,7 @@ struct virtio_scsi {
 	/* Get some buffers ready for event vq */
 	struct virtio_scsi_event_node event_list[VIRTIO_SCSI_EVENT_LEN];
 
-	struct virtio_scsi_target_state *tgt[];
+	struct virtio_scsi_target_state *tgt;
 };
 
 static struct kmem_cache *virtscsi_cmd_cache;
@@ -568,18 +568,9 @@ static void virtscsi_init_vq(struct virtio_scsi_vq *virtscsi_vq,
 	virtscsi_vq->vq = vq;
 }
 
-static struct virtio_scsi_target_state *virtscsi_alloc_tgt(
-	struct virtio_device *vdev, int sg_elems)
+static void virtscsi_init_tgt(struct virtio_scsi_target_state *tgt)
 {
-	struct virtio_scsi_target_state *tgt;
-	gfp_t gfp_mask = GFP_KERNEL;
-
-	tgt = kmalloc(sizeof(*tgt), gfp_mask);
-	if (!tgt)
-		return NULL;
-
 	spin_lock_init(&tgt->tgt_lock);
-	return tgt;
 }
 
 static void virtscsi_scan(struct virtio_device *vdev)
@@ -593,17 +584,10 @@ static void virtscsi_remove_vqs(struct virtio_device *vdev)
 {
 	struct Scsi_Host *sh = virtio_scsi_host(vdev);
 	struct virtio_scsi *vscsi = shost_priv(sh);
-	u32 i, num_targets;
 
 	/* Stop all the virtqueues. */
 	vdev->config->reset(vdev);
-
-	num_targets = sh->max_id;
-	for (i = 0; i < num_targets; i++) {
-		kfree(vscsi->tgt[i]);
-		vscsi->tgt[i] = NULL;
-	}
-
+	kfree(vscsi->tgt);
 	vdev->config->del_vqs(vdev);
 }
 
@@ -612,7 +596,7 @@ static int virtscsi_init(struct virtio_device *vdev,
 {
 	int err;
 	struct virtqueue *vqs[3];
-	u32 i, sg_elems;
+	u32 i;
 
 	vq_callback_t *callbacks[] = {
 		virtscsi_ctrl_done,
@@ -640,16 +624,14 @@ static int virtscsi_init(struct virtio_device *vdev,
 	if (virtio_has_feature(vdev, VIRTIO_SCSI_F_HOTPLUG))
 		virtscsi_kick_event_all(vscsi);
 
-	/* We need to know how many segments before we allocate.  */
-	sg_elems = virtscsi_config_get(vdev, seg_max) ?: 1;
-
-	for (i = 0; i < num_targets; i++) {
-		vscsi->tgt[i] = virtscsi_alloc_tgt(vdev, sg_elems);
-		if (!vscsi->tgt[i]) {
-			err = -ENOMEM;
-			goto out;
-		}
+	vscsi->tgt = kmalloc(num_targets * sizeof(vscsi->tgt[0]), GFP_KERNEL);
+	if (!vscsi->tgt) {
+		err = -ENOMEM;
+		goto out;
 	}
+	for (i = 0; i < num_targets; i++)
+		virtscsi_init_tgt(&vscsi->tgt[i]);
+
 	err = 0;
 
 out:
@@ -668,10 +650,7 @@ static int virtscsi_probe(struct virtio_device *vdev)
 
 	/* Allocate memory and link the structs together.  */
 	num_targets = virtscsi_config_get(vdev, max_target) + 1;
-	shost = scsi_host_alloc(&virtscsi_host_template,
-		sizeof(*vscsi)
-		+ num_targets * sizeof(struct virtio_scsi_target_state));
-
+	shost = scsi_host_alloc(&virtscsi_host_template, sizeof(*vscsi));
 	if (!shost)
 		return -ENOMEM;
 
-- 
1.8.2.rc2

Re: [PATCH V4 1/5] virtio-scsi: redo allocation of target data

[Paolo Bonzini]

Il 11/03/2013 03:43, Wanlong Gao ha scritto:
> virtio_scsi_target_state is now empty, but we will find new uses
> for it in the next few patches.  However, dropping the sglist lets
> us turn the array-of-pointers into a simple array, which simplifies
> the allocation.
> 
> However, we do not leave the virtio_scsi_target_state structs in the
> flexible array member at the end of struct virtio_scsi, because
> we will place the virtqueues there in the next patches.

If you have to respin, consider improving the grammar:

virtio_scsi_target_state is now empty.  We will find new uses for it in
the next few patches, so this patch does not drop it completely.
However, having dropped the sglist flexible array member, we can turn
the tgt array-of-pointers into a simple array.  This simplifies the
allocation.

Even simpler would be to place the virtio_scsi_target_state structs in a
flexible array member at the end of struct virtio_scsi.  But we do not
do that, because we will place the virtqueues there in the next patches.

Paolo

> Cc: linux-scsi@vger.kernel.org
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>

Re: [PATCH V4 1/5] virtio-scsi: redo allocation of target data

[Asias He]

On Mon, Mar 11, 2013 at 10:43:58AM +0800, Wanlong Gao wrote:
> From: Paolo Bonzini <pbonzini@redhat.com>
> 
> virtio_scsi_target_state is now empty, but we will find new uses
> for it in the next few patches.  However, dropping the sglist lets
> us turn the array-of-pointers into a simple array, which simplifies
> the allocation.
> 
> However, we do not leave the virtio_scsi_target_state structs in the
> flexible array member at the end of struct virtio_scsi, because
> we will place the virtqueues there in the next patches.
> 
> Cc: linux-scsi@vger.kernel.org
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>
> ---
>  drivers/scsi/virtio_scsi.c | 45 ++++++++++++---------------------------------
>  1 file changed, 12 insertions(+), 33 deletions(-)
> 
> diff --git a/drivers/scsi/virtio_scsi.c b/drivers/scsi/virtio_scsi.c
> index 6307099..75ccde1 100644
> --- a/drivers/scsi/virtio_scsi.c
> +++ b/drivers/scsi/virtio_scsi.c
> @@ -76,7 +76,7 @@ struct virtio_scsi {
>  	/* Get some buffers ready for event vq */
>  	struct virtio_scsi_event_node event_list[VIRTIO_SCSI_EVENT_LEN];
>  
> -	struct virtio_scsi_target_state *tgt[];
> +	struct virtio_scsi_target_state *tgt;
>  };
>  
>  static struct kmem_cache *virtscsi_cmd_cache;
> @@ -568,18 +568,9 @@ static void virtscsi_init_vq(struct virtio_scsi_vq *virtscsi_vq,
>  	virtscsi_vq->vq = vq;
>  }
>  
> -static struct virtio_scsi_target_state *virtscsi_alloc_tgt(
> -	struct virtio_device *vdev, int sg_elems)
> +static void virtscsi_init_tgt(struct virtio_scsi_target_state *tgt)
>  {
> -	struct virtio_scsi_target_state *tgt;
> -	gfp_t gfp_mask = GFP_KERNEL;
> -
> -	tgt = kmalloc(sizeof(*tgt), gfp_mask);
> -	if (!tgt)
> -		return NULL;
> -
>  	spin_lock_init(&tgt->tgt_lock);
> -	return tgt;
>  }
>  
>  static void virtscsi_scan(struct virtio_device *vdev)
> @@ -593,17 +584,10 @@ static void virtscsi_remove_vqs(struct virtio_device *vdev)
>  {
>  	struct Scsi_Host *sh = virtio_scsi_host(vdev);
>  	struct virtio_scsi *vscsi = shost_priv(sh);
> -	u32 i, num_targets;
>  
>  	/* Stop all the virtqueues. */
>  	vdev->config->reset(vdev);
> -
> -	num_targets = sh->max_id;
> -	for (i = 0; i < num_targets; i++) {
> -		kfree(vscsi->tgt[i]);
> -		vscsi->tgt[i] = NULL;
> -	}
> -
> +	kfree(vscsi->tgt);
>  	vdev->config->del_vqs(vdev);
>  }
>  
> @@ -612,7 +596,7 @@ static int virtscsi_init(struct virtio_device *vdev,
>  {
>  	int err;
>  	struct virtqueue *vqs[3];
> -	u32 i, sg_elems;
> +	u32 i;
>  
>  	vq_callback_t *callbacks[] = {
>  		virtscsi_ctrl_done,
> @@ -640,16 +624,14 @@ static int virtscsi_init(struct virtio_device *vdev,
>  	if (virtio_has_feature(vdev, VIRTIO_SCSI_F_HOTPLUG))
>  		virtscsi_kick_event_all(vscsi);
>  
> -	/* We need to know how many segments before we allocate.  */
> -	sg_elems = virtscsi_config_get(vdev, seg_max) ?: 1;

The drop of sg_elems does not belong to this patch. It should be done in
'virtio-scsi: use virtqueue_add_sgs for command buffers'.

> -	for (i = 0; i < num_targets; i++) {
> -		vscsi->tgt[i] = virtscsi_alloc_tgt(vdev, sg_elems);
> -		if (!vscsi->tgt[i]) {
> -			err = -ENOMEM;
> -			goto out;
> -		}
> +	vscsi->tgt = kmalloc(num_targets * sizeof(vscsi->tgt[0]), GFP_KERNEL);
> +	if (!vscsi->tgt) {
> +		err = -ENOMEM;
> +		goto out;
>  	}
> +	for (i = 0; i < num_targets; i++)
> +		virtscsi_init_tgt(&vscsi->tgt[i]);
> +
>  	err = 0;
>  
>  out:
> @@ -668,10 +650,7 @@ static int virtscsi_probe(struct virtio_device *vdev)
>  
>  	/* Allocate memory and link the structs together.  */
>  	num_targets = virtscsi_config_get(vdev, max_target) + 1;
> -	shost = scsi_host_alloc(&virtscsi_host_template,
> -		sizeof(*vscsi)
> -		+ num_targets * sizeof(struct virtio_scsi_target_state));
> -
> +	shost = scsi_host_alloc(&virtscsi_host_template, sizeof(*vscsi));
>  	if (!shost)
>  		return -ENOMEM;
>  
> -- 
> 1.8.2.rc2
> 

-- 
Asias

[PATCH V4 2/5] virtio-scsi: pass struct virtio_scsi to virtqueue completion function

[Wanlong Gao]

From: Paolo Bonzini <pbonzini@redhat.com>

This will be needed soon in order to retrieve the per-target
struct.

Cc: linux-scsi@vger.kernel.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>
---
 drivers/scsi/virtio_scsi.c | 17 +++++++++--------
 1 file changed, 9 insertions(+), 8 deletions(-)

diff --git a/drivers/scsi/virtio_scsi.c b/drivers/scsi/virtio_scsi.c
index 75ccde1..fdb0132 100644
--- a/drivers/scsi/virtio_scsi.c
+++ b/drivers/scsi/virtio_scsi.c
@@ -106,7 +106,7 @@ static void virtscsi_compute_resid(struct scsi_cmnd *sc, u32 resid)
  *
  * Called with vq_lock held.
  */
-static void virtscsi_complete_cmd(void *buf)
+static void virtscsi_complete_cmd(struct virtio_scsi *vscsi, void *buf)
 {
 	struct virtio_scsi_cmd *cmd = buf;
 	struct scsi_cmnd *sc = cmd->sc;
@@ -167,7 +167,8 @@ static void virtscsi_complete_cmd(void *buf)
 	sc->scsi_done(sc);
 }
 
-static void virtscsi_vq_done(struct virtqueue *vq, void (*fn)(void *buf))
+static void virtscsi_vq_done(struct virtio_scsi *vscsi, struct virtqueue *vq,
+			     void (*fn)(struct virtio_scsi *vscsi, void *buf))
 {
 	void *buf;
 	unsigned int len;
@@ -175,7 +176,7 @@ static void virtscsi_vq_done(struct virtqueue *vq, void (*fn)(void *buf))
 	do {
 		virtqueue_disable_cb(vq);
 		while ((buf = virtqueue_get_buf(vq, &len)) != NULL)
-			fn(buf);
+			fn(vscsi, buf);
 	} while (!virtqueue_enable_cb(vq));
 }
 
@@ -186,11 +187,11 @@ static void virtscsi_req_done(struct virtqueue *vq)
 	unsigned long flags;
 
 	spin_lock_irqsave(&vscsi->req_vq.vq_lock, flags);
-	virtscsi_vq_done(vq, virtscsi_complete_cmd);
+	virtscsi_vq_done(vscsi, vq, virtscsi_complete_cmd);
 	spin_unlock_irqrestore(&vscsi->req_vq.vq_lock, flags);
 };
 
-static void virtscsi_complete_free(void *buf)
+static void virtscsi_complete_free(struct virtio_scsi *vscsi, void *buf)
 {
 	struct virtio_scsi_cmd *cmd = buf;
 
@@ -207,7 +208,7 @@ static void virtscsi_ctrl_done(struct virtqueue *vq)
 	unsigned long flags;
 
 	spin_lock_irqsave(&vscsi->ctrl_vq.vq_lock, flags);
-	virtscsi_vq_done(vq, virtscsi_complete_free);
+	virtscsi_vq_done(vscsi, vq, virtscsi_complete_free);
 	spin_unlock_irqrestore(&vscsi->ctrl_vq.vq_lock, flags);
 };
 
@@ -331,7 +332,7 @@ static void virtscsi_handle_event(struct work_struct *work)
 	virtscsi_kick_event(vscsi, event_node);
 }
 
-static void virtscsi_complete_event(void *buf)
+static void virtscsi_complete_event(struct virtio_scsi *vscsi, void *buf)
 {
 	struct virtio_scsi_event_node *event_node = buf;
 
@@ -346,7 +347,7 @@ static void virtscsi_event_done(struct virtqueue *vq)
 	unsigned long flags;
 
 	spin_lock_irqsave(&vscsi->event_vq.vq_lock, flags);
-	virtscsi_vq_done(vq, virtscsi_complete_event);
+	virtscsi_vq_done(vscsi, vq, virtscsi_complete_event);
 	spin_unlock_irqrestore(&vscsi->event_vq.vq_lock, flags);
 };
 
-- 
1.8.2.rc2

[PATCH V4 3/5] virtio-scsi: push vq lock/unlock into virtscsi_vq_done

[Wanlong Gao]

From: Paolo Bonzini <pbonzini@redhat.com>

Avoid duplicated code in all of the callers.

Cc: linux-scsi@vger.kernel.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>
---
 drivers/scsi/virtio_scsi.c | 22 +++++++++-------------
 1 file changed, 9 insertions(+), 13 deletions(-)

diff --git a/drivers/scsi/virtio_scsi.c b/drivers/scsi/virtio_scsi.c
index fdb0132..3573704 100644
--- a/drivers/scsi/virtio_scsi.c
+++ b/drivers/scsi/virtio_scsi.c
@@ -167,28 +167,30 @@ static void virtscsi_complete_cmd(struct virtio_scsi *vscsi, void *buf)
 	sc->scsi_done(sc);
 }
 
-static void virtscsi_vq_done(struct virtio_scsi *vscsi, struct virtqueue *vq,
+static void virtscsi_vq_done(struct virtio_scsi *vscsi,
+			     struct virtio_scsi_vq *virtscsi_vq,
 			     void (*fn)(struct virtio_scsi *vscsi, void *buf))
 {
 	void *buf;
 	unsigned int len;
+	unsigned long flags;
+	struct virtqueue *vq = virtscsi_vq->vq;
 
+	spin_lock_irqsave(&virtscsi_vq->vq_lock, flags);
 	do {
 		virtqueue_disable_cb(vq);
 		while ((buf = virtqueue_get_buf(vq, &len)) != NULL)
 			fn(vscsi, buf);
 	} while (!virtqueue_enable_cb(vq));
+	spin_unlock_irqrestore(&virtscsi_vq->vq_lock, flags);
 }
 
 static void virtscsi_req_done(struct virtqueue *vq)
 {
 	struct Scsi_Host *sh = virtio_scsi_host(vq->vdev);
 	struct virtio_scsi *vscsi = shost_priv(sh);
-	unsigned long flags;
 
-	spin_lock_irqsave(&vscsi->req_vq.vq_lock, flags);
-	virtscsi_vq_done(vscsi, vq, virtscsi_complete_cmd);
-	spin_unlock_irqrestore(&vscsi->req_vq.vq_lock, flags);
+	virtscsi_vq_done(vscsi, &vscsi->req_vq, virtscsi_complete_cmd);
 };
 
 static void virtscsi_complete_free(struct virtio_scsi *vscsi, void *buf)
@@ -205,11 +207,8 @@ static void virtscsi_ctrl_done(struct virtqueue *vq)
 {
 	struct Scsi_Host *sh = virtio_scsi_host(vq->vdev);
 	struct virtio_scsi *vscsi = shost_priv(sh);
-	unsigned long flags;
 
-	spin_lock_irqsave(&vscsi->ctrl_vq.vq_lock, flags);
-	virtscsi_vq_done(vscsi, vq, virtscsi_complete_free);
-	spin_unlock_irqrestore(&vscsi->ctrl_vq.vq_lock, flags);
+	virtscsi_vq_done(vscsi, &vscsi->ctrl_vq, virtscsi_complete_free);
 };
 
 static int virtscsi_kick_event(struct virtio_scsi *vscsi,
@@ -344,11 +343,8 @@ static void virtscsi_event_done(struct virtqueue *vq)
 {
 	struct Scsi_Host *sh = virtio_scsi_host(vq->vdev);
 	struct virtio_scsi *vscsi = shost_priv(sh);
-	unsigned long flags;
 
-	spin_lock_irqsave(&vscsi->event_vq.vq_lock, flags);
-	virtscsi_vq_done(vscsi, vq, virtscsi_complete_event);
-	spin_unlock_irqrestore(&vscsi->event_vq.vq_lock, flags);
+	virtscsi_vq_done(vscsi, &vscsi->event_vq, virtscsi_complete_event);
 };
 
 /**
-- 
1.8.2.rc2

[PATCH V4 4/5] virtio-scsi: introduce multiqueue support

[Wanlong Gao]

From: Paolo Bonzini <pbonzini@redhat.com>

This patch adds queue steering to virtio-scsi.  When a target is sent
multiple requests, we always drive them to the same queue so that FIFO
processing order is kept.  However, if a target was idle, we can choose
a queue arbitrarily.  In this case the queue is chosen according to the
current VCPU, so the driver expects the number of request queues to be
equal to the number of VCPUs.  This makes it easy and fast to select
the queue, and also lets the driver optimize the IRQ affinity for the
virtqueues (each virtqueue's affinity is set to the CPU that "owns"
the queue).

The speedup comes from improving cache locality and giving CPU affinity
to the virtqueues, which is why this scheme was selected.  Assuming that
the thread that is sending requests to the device is I/O-bound, it is
likely to be sleeping at the time the ISR is executed, and thus executing
the ISR on the same processor that sent the requests is cheap.

However, the kernel will not execute the ISR on the "best" processor
unless you explicitly set the affinity.  This is because in practice
you will have many such I/O-bound processes and thus many otherwise
idle processors.  Then the kernel will execute the ISR on a random
processor, rather than the one that is sending requests to the device.

The alternative to per-CPU virtqueues is per-target virtqueues.  To
achieve the same locality, we could dynamically choose the virtqueue's
affinity based on the CPU of the last task that sent a request.  This
is less appealing because we do not set the affinity directly---we only
provide a hint to the irqbalanced running in userspace.  Dynamically
changing the affinity only works if the userspace applies the hint
fast enough.

Cc: linux-scsi@vger.kernel.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>
---
 drivers/scsi/virtio_scsi.c | 269 ++++++++++++++++++++++++++++++++++++++++-----
 1 file changed, 241 insertions(+), 28 deletions(-)

diff --git a/drivers/scsi/virtio_scsi.c b/drivers/scsi/virtio_scsi.c
index 3573704..7a54293 100644
--- a/drivers/scsi/virtio_scsi.c
+++ b/drivers/scsi/virtio_scsi.c
@@ -22,12 +22,14 @@
 #include <linux/virtio_ids.h>
 #include <linux/virtio_config.h>
 #include <linux/virtio_scsi.h>
+#include <linux/cpu.h>
 #include <scsi/scsi_host.h>
 #include <scsi/scsi_device.h>
 #include <scsi/scsi_cmnd.h>
 
 #define VIRTIO_SCSI_MEMPOOL_SZ 64
 #define VIRTIO_SCSI_EVENT_LEN 8
+#define VIRTIO_SCSI_VQ_BASE 2
 
 /* Command queue element */
 struct virtio_scsi_cmd {
@@ -59,24 +61,60 @@ struct virtio_scsi_vq {
 	struct virtqueue *vq;
 };
 
-/* Per-target queue state */
+/*
+ * Per-target queue state.
+ *
+ * This struct holds the data needed by the queue steering policy.  When a
+ * target is sent multiple requests, we need to drive them to the same queue so
+ * that FIFO processing order is kept.  However, if a target was idle, we can
+ * choose a queue arbitrarily.  In this case the queue is chosen according to
+ * the current VCPU, so the driver expects the number of request queues to be
+ * equal to the number of VCPUs.  This makes it easy and fast to select the
+ * queue, and also lets the driver optimize the IRQ affinity for the virtqueues
+ * (each virtqueue's affinity is set to the CPU that "owns" the queue).
+ *
+ * An interesting effect of this policy is that only writes to req_vq need to
+ * take the tgt_lock.  Read can be done outside the lock because:
+ *
+ * - writes of req_vq only occur when atomic_inc_return(&tgt->reqs) returns 1.
+ *   In that case, no other CPU is reading req_vq: even if they were in
+ *   virtscsi_queuecommand_multi, they would be spinning on tgt_lock.
+ *
+ * - reads of req_vq only occur when the target is not idle (reqs != 0).
+ *   A CPU that enters virtscsi_queuecommand_multi will not modify req_vq.
+ *
+ * Similarly, decrements of reqs are never concurrent with writes of req_vq.
+ * Thus they can happen outside the tgt_lock, provided of course we make reqs
+ * an atomic_t.
+ */
 struct virtio_scsi_target_state {
-	/* Never held at the same time as vq_lock.  */
+	/* This spinlock never help at the same time as vq_lock. */
 	spinlock_t tgt_lock;
+
+	/* Count of outstanding requests. */
+	atomic_t reqs;
+
+	/* Currently active virtqueue for requests sent to this target. */
+	struct virtio_scsi_vq *req_vq;
 };
 
 /* Driver instance state */
 struct virtio_scsi {
 	struct virtio_device *vdev;
 
-	struct virtio_scsi_vq ctrl_vq;
-	struct virtio_scsi_vq event_vq;
-	struct virtio_scsi_vq req_vq;
-
 	/* Get some buffers ready for event vq */
 	struct virtio_scsi_event_node event_list[VIRTIO_SCSI_EVENT_LEN];
 
 	struct virtio_scsi_target_state *tgt;
+
+	u32 num_queues;
+
+	/* Does the affinity hint is set for virtqueues? */
+	bool affinity_hint_set;
+
+	struct virtio_scsi_vq ctrl_vq;
+	struct virtio_scsi_vq event_vq;
+	struct virtio_scsi_vq req_vqs[];
 };
 
 static struct kmem_cache *virtscsi_cmd_cache;
@@ -111,6 +149,7 @@ static void virtscsi_complete_cmd(struct virtio_scsi *vscsi, void *buf)
 	struct virtio_scsi_cmd *cmd = buf;
 	struct scsi_cmnd *sc = cmd->sc;
 	struct virtio_scsi_cmd_resp *resp = &cmd->resp.cmd;
+	struct virtio_scsi_target_state *tgt = &vscsi->tgt[sc->device->id];
 
 	dev_dbg(&sc->device->sdev_gendev,
 		"cmd %p response %u status %#02x sense_len %u\n",
@@ -165,6 +204,8 @@ static void virtscsi_complete_cmd(struct virtio_scsi *vscsi, void *buf)
 
 	mempool_free(cmd, virtscsi_cmd_pool);
 	sc->scsi_done(sc);
+
+	atomic_dec(&tgt->reqs);
 }
 
 static void virtscsi_vq_done(struct virtio_scsi *vscsi,
@@ -189,8 +230,42 @@ static void virtscsi_req_done(struct virtqueue *vq)
 {
 	struct Scsi_Host *sh = virtio_scsi_host(vq->vdev);
 	struct virtio_scsi *vscsi = shost_priv(sh);
+	int index = vq->index - VIRTIO_SCSI_VQ_BASE;
+	struct virtio_scsi_vq *req_vq = &vscsi->req_vqs[index];
+
+	/*
+	 * Read req_vq before decrementing the reqs field in
+	 * virtscsi_complete_cmd.
+	 *
+	 * With barriers:
+	 *
+	 * 	CPU #0			virtscsi_queuecommand_multi (CPU #1)
+	 * 	------------------------------------------------------------
+	 * 	lock vq_lock
+	 * 	read req_vq
+	 * 	read reqs (reqs = 1)
+	 * 	write reqs (reqs = 0)
+	 * 				increment reqs (reqs = 1)
+	 * 				write req_vq
+	 *
+	 * Possible reordering without barriers:
+	 *
+	 * 	CPU #0			virtscsi_queuecommand_multi (CPU #1)
+	 * 	------------------------------------------------------------
+	 * 	lock vq_lock
+	 * 	read reqs (reqs = 1)
+	 * 	write reqs (reqs = 0)
+	 * 				increment reqs (reqs = 1)
+	 * 				write req_vq
+	 * 	read (wrong) req_vq
+	 *
+	 * We do not need a full smp_rmb, because req_vq is required to get
+	 * to tgt->reqs: tgt is &vscsi->tgt[sc->device->id], where sc is stored
+	 * in the virtqueue as the user token.
+	 */
+	smp_read_barrier_depends();
 
-	virtscsi_vq_done(vscsi, &vscsi->req_vq, virtscsi_complete_cmd);
+	virtscsi_vq_done(vscsi, req_vq, virtscsi_complete_cmd);
 };
 
 static void virtscsi_complete_free(struct virtio_scsi *vscsi, void *buf)
@@ -412,9 +487,10 @@ static int virtscsi_kick_cmd(struct virtio_scsi_vq *vq,
 	return err;
 }
 
-static int virtscsi_queuecommand(struct Scsi_Host *sh, struct scsi_cmnd *sc)
+static int virtscsi_queuecommand(struct virtio_scsi *vscsi,
+				 struct virtio_scsi_vq *req_vq,
+				 struct scsi_cmnd *sc)
 {
-	struct virtio_scsi *vscsi = shost_priv(sh);
 	struct virtio_scsi_cmd *cmd;
 	int ret;
 
@@ -448,7 +524,7 @@ static int virtscsi_queuecommand(struct Scsi_Host *sh, struct scsi_cmnd *sc)
 	BUG_ON(sc->cmd_len > VIRTIO_SCSI_CDB_SIZE);
 	memcpy(cmd->req.cmd.cdb, sc->cmnd, sc->cmd_len);
 
-	if (virtscsi_kick_cmd(&vscsi->req_vq, cmd,
+	if (virtscsi_kick_cmd(req_vq, cmd,
 			      sizeof cmd->req.cmd, sizeof cmd->resp.cmd,
 			      GFP_ATOMIC) == 0)
 		ret = 0;
@@ -459,6 +535,53 @@ out:
 	return ret;
 }
 
+static int virtscsi_queuecommand_single(struct Scsi_Host *sh,
+					struct scsi_cmnd *sc)
+{
+	struct virtio_scsi *vscsi = shost_priv(sh);
+	struct virtio_scsi_target_state *tgt = &vscsi->tgt[sc->device->id];
+
+	atomic_inc(&tgt->reqs);
+	return virtscsi_queuecommand(vscsi, &vscsi->req_vqs[0], sc);
+}
+
+static struct virtio_scsi_vq *virtscsi_pick_vq(struct virtio_scsi *vscsi,
+					       struct virtio_scsi_target_state *tgt)
+{
+	struct virtio_scsi_vq *vq;
+	unsigned long flags;
+	u32 queue_num;
+
+	spin_lock_irqsave(&tgt->tgt_lock, flags);
+
+	/*
+	 * The memory barrier after atomic_inc_return matches
+	 * the smp_read_barrier_depends() in virtscsi_req_done.
+	 */
+	if (atomic_inc_return(&tgt->reqs) > 1)
+		vq = ACCESS_ONCE(tgt->req_vq);
+	else {
+		queue_num = smp_processor_id();
+		while (unlikely(queue_num >= vscsi->num_queues))
+			queue_num -= vscsi->num_queues;
+
+		tgt->req_vq = vq = &vscsi->req_vqs[queue_num];
+	}
+
+	spin_unlock_irqrestore(&tgt->tgt_lock, flags);
+	return vq;
+}
+
+static int virtscsi_queuecommand_multi(struct Scsi_Host *sh,
+				       struct scsi_cmnd *sc)
+{
+	struct virtio_scsi *vscsi = shost_priv(sh);
+	struct virtio_scsi_target_state *tgt = &vscsi->tgt[sc->device->id];
+	struct virtio_scsi_vq *req_vq = virtscsi_pick_vq(vscsi, tgt);
+
+	return virtscsi_queuecommand(vscsi, req_vq, sc);
+}
+
 static int virtscsi_tmf(struct virtio_scsi *vscsi, struct virtio_scsi_cmd *cmd)
 {
 	DECLARE_COMPLETION_ONSTACK(comp);
@@ -527,12 +650,26 @@ static int virtscsi_abort(struct scsi_cmnd *sc)
 	return virtscsi_tmf(vscsi, cmd);
 }
 
-static struct scsi_host_template virtscsi_host_template = {
+static struct scsi_host_template virtscsi_host_template_single = {
 	.module = THIS_MODULE,
 	.name = "Virtio SCSI HBA",
 	.proc_name = "virtio_scsi",
-	.queuecommand = virtscsi_queuecommand,
 	.this_id = -1,
+	.queuecommand = virtscsi_queuecommand_single,
+	.eh_abort_handler = virtscsi_abort,
+	.eh_device_reset_handler = virtscsi_device_reset,
+
+	.can_queue = 1024,
+	.dma_boundary = UINT_MAX,
+	.use_clustering = ENABLE_CLUSTERING,
+};
+
+static struct scsi_host_template virtscsi_host_template_multi = {
+	.module = THIS_MODULE,
+	.name = "Virtio SCSI HBA",
+	.proc_name = "virtio_scsi",
+	.this_id = -1,
+	.queuecommand = virtscsi_queuecommand_multi,
 	.eh_abort_handler = virtscsi_abort,
 	.eh_device_reset_handler = virtscsi_device_reset,
 
@@ -558,6 +695,47 @@ static struct scsi_host_template virtscsi_host_template = {
 				  &__val, sizeof(__val)); \
 	})
 
+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 - VIRTIO_SCSI_VQ_BASE; i++)
+			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 void virtscsi_init_vq(struct virtio_scsi_vq *virtscsi_vq,
 			     struct virtqueue *vq)
 {
@@ -565,9 +743,11 @@ static void virtscsi_init_vq(struct virtio_scsi_vq *virtscsi_vq,
 	virtscsi_vq->vq = vq;
 }
 
-static void virtscsi_init_tgt(struct virtio_scsi_target_state *tgt)
+static void virtscsi_init_tgt(struct virtio_scsi *vscsi, int i)
 {
+	struct virtio_scsi_target_state *tgt = &vscsi->tgt[i];
 	spin_lock_init(&tgt->tgt_lock);
+	atomic_set(&tgt->reqs, 0);
 }
 
 static void virtscsi_scan(struct virtio_device *vdev)
@@ -582,6 +762,8 @@ 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);
 	kfree(vscsi->tgt);
@@ -592,28 +774,43 @@ static int virtscsi_init(struct virtio_device *vdev,
 			 struct virtio_scsi *vscsi, int num_targets)
 {
 	int err;
-	struct virtqueue *vqs[3];
 	u32 i;
+	u32 num_vqs;
+	vq_callback_t **callbacks;
+	const char **names;
+	struct virtqueue **vqs;
 
-	vq_callback_t *callbacks[] = {
-		virtscsi_ctrl_done,
-		virtscsi_event_done,
-		virtscsi_req_done
-	};
-	const char *names[] = {
-		"control",
-		"event",
-		"request"
-	};
+	num_vqs = vscsi->num_queues + VIRTIO_SCSI_VQ_BASE;
+	vqs = kmalloc(num_vqs * sizeof(struct virtqueue *), GFP_KERNEL);
+	callbacks = kmalloc(num_vqs * sizeof(vq_callback_t *), GFP_KERNEL);
+	names = kmalloc(num_vqs * sizeof(char *), GFP_KERNEL);
+
+	if (!callbacks || !vqs || !names) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	callbacks[0] = virtscsi_ctrl_done;
+	callbacks[1] = virtscsi_event_done;
+	names[0] = "control";
+	names[1] = "event";
+	for (i = VIRTIO_SCSI_VQ_BASE; i < num_vqs; i++) {
+		callbacks[i] = virtscsi_req_done;
+		names[i] = "request";
+	}
 
 	/* Discover virtqueues and write information to configuration.  */
-	err = vdev->config->find_vqs(vdev, 3, vqs, callbacks, names);
+	err = vdev->config->find_vqs(vdev, num_vqs, vqs, callbacks, names);
 	if (err)
 		return err;
 
 	virtscsi_init_vq(&vscsi->ctrl_vq, vqs[0]);
 	virtscsi_init_vq(&vscsi->event_vq, vqs[1]);
-	virtscsi_init_vq(&vscsi->req_vq, vqs[2]);
+	for (i = VIRTIO_SCSI_VQ_BASE; i < num_vqs; i++)
+		virtscsi_init_vq(&vscsi->req_vqs[i - VIRTIO_SCSI_VQ_BASE],
+				 vqs[i]);
+
+	virtscsi_set_affinity(vscsi, true);
 
 	virtscsi_config_set(vdev, cdb_size, VIRTIO_SCSI_CDB_SIZE);
 	virtscsi_config_set(vdev, sense_size, VIRTIO_SCSI_SENSE_SIZE);
@@ -627,11 +824,14 @@ static int virtscsi_init(struct virtio_device *vdev,
 		goto out;
 	}
 	for (i = 0; i < num_targets; i++)
-		virtscsi_init_tgt(&vscsi->tgt[i]);
+		virtscsi_init_tgt(vscsi, i);
 
 	err = 0;
 
 out:
+	kfree(names);
+	kfree(callbacks);
+	kfree(vqs);
 	if (err)
 		virtscsi_remove_vqs(vdev);
 	return err;
@@ -644,10 +844,22 @@ static int virtscsi_probe(struct virtio_device *vdev)
 	int err;
 	u32 sg_elems, num_targets;
 	u32 cmd_per_lun;
+	u32 num_queues;
+	struct scsi_host_template *hostt;
+
+	/* We need to know how many queues before we allocate. */
+	num_queues = virtscsi_config_get(vdev, num_queues) ? : 1;
 
 	/* Allocate memory and link the structs together.  */
 	num_targets = virtscsi_config_get(vdev, max_target) + 1;
-	shost = scsi_host_alloc(&virtscsi_host_template, sizeof(*vscsi));
+
+	if (num_queues == 1)
+		hostt = &virtscsi_host_template_single;
+	else
+		hostt = &virtscsi_host_template_multi;
+
+	shost = scsi_host_alloc(hostt,
+		sizeof(*vscsi) + sizeof(vscsi->req_vqs[0]) * num_queues);
 	if (!shost)
 		return -ENOMEM;
 
@@ -655,6 +867,7 @@ static int virtscsi_probe(struct virtio_device *vdev)
 	shost->sg_tablesize = sg_elems;
 	vscsi = shost_priv(shost);
 	vscsi->vdev = vdev;
+	vscsi->num_queues = num_queues;
 	vdev->priv = shost;
 
 	err = virtscsi_init(vdev, vscsi, num_targets);
-- 
1.8.2.rc2

[PATCH V4 5/5] virtio-scsi: reset virtqueue affinity when doing cpu hotplug

[Wanlong Gao]

Add hot cpu notifier to reset the request virtqueue affinity
when doing cpu hotplug.

Cc: linux-scsi@vger.kernel.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>
---
 drivers/scsi/virtio_scsi.c | 29 +++++++++++++++++++++++++++++
 1 file changed, 29 insertions(+)

diff --git a/drivers/scsi/virtio_scsi.c b/drivers/scsi/virtio_scsi.c
index 7a54293..3bb2bf4 100644
--- a/drivers/scsi/virtio_scsi.c
+++ b/drivers/scsi/virtio_scsi.c
@@ -112,6 +112,9 @@ struct virtio_scsi {
 	/* Does the affinity hint is set for virtqueues? */
 	bool affinity_hint_set;
 
+	/* CPU hotplug notifier */
+	struct notifier_block nb;
+
 	struct virtio_scsi_vq ctrl_vq;
 	struct virtio_scsi_vq event_vq;
 	struct virtio_scsi_vq req_vqs[];
@@ -736,6 +739,23 @@ static void virtscsi_set_affinity(struct virtio_scsi *vscsi, bool affinity)
 	put_online_cpus();
 }
 
+static int virtscsi_cpu_callback(struct notifier_block *nfb,
+				 unsigned long action, void *hcpu)
+{
+	struct virtio_scsi *vscsi = container_of(nfb, struct virtio_scsi, nb);
+	switch(action) {
+	case CPU_ONLINE:
+	case CPU_ONLINE_FROZEN:
+	case CPU_DEAD:
+	case CPU_DEAD_FROZEN:
+		__virtscsi_set_affinity(vscsi, true);
+		break;
+	default:
+		break;
+	}
+	return NOTIFY_OK;
+}
+
 static void virtscsi_init_vq(struct virtio_scsi_vq *virtscsi_vq,
 			     struct virtqueue *vq)
 {
@@ -874,6 +894,13 @@ static int virtscsi_probe(struct virtio_device *vdev)
 	if (err)
 		goto virtscsi_init_failed;
 
+	vscsi->nb.notifier_call = &virtscsi_cpu_callback;
+	err = register_hotcpu_notifier(&vscsi->nb);
+	if (err) {
+		pr_err("registering cpu notifier failed\n");
+		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;
@@ -911,6 +938,8 @@ static void virtscsi_remove(struct virtio_device *vdev)
 
 	scsi_remove_host(shost);
 
+	unregister_hotcpu_notifier(&vscsi->nb);
+
 	virtscsi_remove_vqs(vdev);
 	scsi_host_put(shost);
 }
-- 
1.8.2.rc2

Re: [PATCH V4 0/5] virtio-scsi multiqueue

[Asias He]

On Mon, Mar 11, 2013 at 10:43:57AM +0800, Wanlong Gao wrote:
> This series implements virtio-scsi queue steering, which gives
> performance improvements of up to 50% (measured both with QEMU and
> tcm_vhost backends).
> 
> This version rebased on Rusty's virtio ring rework patches.
> We hope this can go into virtio-next together with the virtio ring
> rework pathes.
> 
> 
> V4: rebase on virtio ring rework patches (rusty's pending-rebases branch)
> 
> V3 and be found http://marc.info/?l=linux-virtualization&m=136067440717154&w=2
> 
> 
> It would probably be easier to get it in via Rusty's tree
> because of the prerequisites.  James, can we get your Acked-by?
> 
> Paolo Bonzini (4):
>   virtio-scsi: redo allocation of target data
>   virtio-scsi: pass struct virtio_scsi to virtqueue completion function
>   virtio-scsi: push vq lock/unlock into virtscsi_vq_done
>   virtio-scsi: introduce multiqueue support
> 
> Wanlong Gao (1):
>   virtio-scsi: reset virtqueue affinity when doing cpu hotplug
> 
>  drivers/scsi/virtio_scsi.c | 364 ++++++++++++++++++++++++++++++++++++---------
>  1 file changed, 291 insertions(+), 73 deletions(-)
> 
> -- 
> 1.8.2.rc2
> 

Patch 2-5
Reviewed-by: Asias He <asias@redhat.com>

-- 
Asias