Skip to contentSkip to conversation

Implement vdpasim suspend operation

PATCH Superseded Eugenio Pérez 3 participants 13 messages Aug 10, 2022 – Aug 11, 2022

Subsystems (5)

v7 of 8 Superseded Latest is v8 →

lore
Share & tools
lore

Eugenio Pérez

author maintainer
#0 lore raw

v7 cover Implement vdpasim suspend operation

Implement suspend operation for vdpa_sim devices, so vhost-vdpa will offer that backend feature and userspace can effectively suspend the device.

This is a must before getting virtqueue indexes (base) for live migration, since the device could modify them after userland gets them. There are individual ways to perform that action for some devices (VHOST_NET_SET_BACKEND, VHOST_VSOCK_SET_RUNNING, ...) but there was no way to perform it for any vhost device (and, in particular, vhost-vdpa).

After a successful return of ioctl the device must not process more virtqueue descriptors. The device can answer to read or writes of config fields as if it were not suspended. In particular, writing to "queue_enable" with a value of 1 will not make the device start processing virtqueue buffers.

In the future, we will provide features similar to VHOST_USER_GET_INFLIGHT_FD so the device can save pending operations.

Applied on top of [1] branch after removing the old commits.

Comments are welcome.

v7:
* Remove ioctl leftover argument and update doc accordingly.

v6:
* Remove the resume operation, making the ioctl simpler. We can always add another ioctl for VM_STOP/VM_RESUME operation later.
* s/stop/suspend/ to differentiate more from reset.
* Clarify scope of the suspend operation.

v5:
* s/not stop/resume/ in doc.

v4:
* Replace VHOST_STOP to VHOST_VDPA_STOP in vhost ioctl switch case too.

v3:
* s/VHOST_STOP/VHOST_VDPA_STOP/
* Add documentation and requirements of the ioctl above its definition.

v2:
* Replace raw _F_STOP with BIT_ULL(_F_STOP).
* Fix obtaining of stop ioctl arg (it was not obtained but written).
* Add stop to vdpa_sim_blk.

[1] git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost.git

Eugenio Pérez (4):
vdpa: Add suspend operation
vhost-vdpa: introduce SUSPEND backend feature bit
vhost-vdpa: uAPI to suspend the device
vdpa_sim: Implement suspend vdpa op

 drivers/vdpa/vdpa_sim/vdpa_sim.c     | 14 +++++++++++
 drivers/vdpa/vdpa_sim/vdpa_sim.h     |  1 +
 drivers/vdpa/vdpa_sim/vdpa_sim_blk.c |  3 +++
 drivers/vdpa/vdpa_sim/vdpa_sim_net.c |  3 +++
 drivers/vhost/vdpa.c                 | 35 +++++++++++++++++++++++++++-
 include/linux/vdpa.h                 |  4 ++++
 include/uapi/linux/vhost.h           |  9 +++++++
 include/uapi/linux/vhost_types.h     |  2 ++
 8 files changed, 70 insertions(+), 1 deletion(-)
Signature · 1 line
2.31.1
author maintainer
#1 lore raw

PATCH v7 1/4 vdpa: Add suspend operation

This operation is optional: It it's not implemented, backend feature bit will not be exposed.

Message-Id: <[email protected]>

---
 include/linux/vdpa.h | 4 ++++
 1 file changed, 4 insertions(+)
include/linux/vdpa.h +4 -0
index 7b4a13d3bd91..d282f464d2f1 100644
@@ -218,6 +218,9 @@ struct vdpa_map_file {
  * @reset:			Reset device
  *				@vdev: vdpa device
  *				Returns integer: success (0) or error (< 0)
+ * @suspend:			Suspend or resume the device (optional)
+ *				@vdev: vdpa device
+ *				Returns integer: success (0) or error (< 0)
  * @get_config_size:		Get the size of the configuration space includes
  *				fields that are conditional on feature bits.
  *				@vdev: vdpa device
@@ -319,6 +322,7 @@ struct vdpa_config_ops {
 	u8 (*get_status)(struct vdpa_device *vdev);
 	void (*set_status)(struct vdpa_device *vdev, u8 status);
 	int (*reset)(struct vdpa_device *vdev);
+	int (*suspend)(struct vdpa_device *vdev);
 	size_t (*get_config_size)(struct vdpa_device *vdev);
 	void (*get_config)(struct vdpa_device *vdev, unsigned int offset,
 			   void *buf, unsigned int len);
Signature · 1 line
2.31.1
#2 lore raw

Re: [PATCH v7 1/4] vdpa: Add suspend operation

Quoted 6 lines · Eugenio Pérez

This operation is optional: It it's not implemented, backend feature bit will not be exposed.

Message-Id: <[email protected]>

What is this message id doing here?

Quoted 28 lines
---
 include/linux/vdpa.h | 4 ++++
 1 file changed, 4 insertions(+)
include/linux/vdpa.h +4 -0
index 7b4a13d3bd91..d282f464d2f1 100644
@@ -218,6 +218,9 @@ struct vdpa_map_file {
  * @reset:			Reset device
  *				@vdev: vdpa device
  *				Returns integer: success (0) or error (< 0)
+ * @suspend:			Suspend or resume the device (optional)
+ *				@vdev: vdpa device
+ *				Returns integer: success (0) or error (< 0)
  * @get_config_size:		Get the size of the configuration space includes
  *				fields that are conditional on feature bits.
  *				@vdev: vdpa device
@@ -319,6 +322,7 @@ struct vdpa_config_ops {
 	u8 (*get_status)(struct vdpa_device *vdev);
 	void (*set_status)(struct vdpa_device *vdev, u8 status);
 	int (*reset)(struct vdpa_device *vdev);
+	int (*suspend)(struct vdpa_device *vdev);
 	size_t (*get_config_size)(struct vdpa_device *vdev);
 	void (*get_config)(struct vdpa_device *vdev, unsigned int offset,
 			   void *buf, unsigned int len);
Signature · 1 line
2.31.1
#3 lore raw

Re: [PATCH v7 1/4] vdpa: Add suspend operation

Quoted 10 lines · Michael S. Tsirkin
Quoted 6 lines · Eugenio Pérez

This operation is optional: It it's not implemented, backend feature bit will not be exposed.

Message-Id: <[email protected]>

What is this message id doing here?

I like the Message-Id tag. It means you can b4 mbox <mesg-id> and get the thread.

Linus has complained (rough remembering) that everyone is using the
Link: tag for links to the patch itself. It's supposed to be for Links to bugzilla or to the spec or whatever. Extra information, too much to put in the commit message. Now the Link tag is useless because it either points to the patch or to a bugzilla. Depend on what you want it to do, it *always* points to the opposite thing.

But I can't remember what people settled on as the alternative to use to link to lore...

In theory, we should be able to figure out the link to lore automatically and there have been a couple projects which tried to do this but they can't make it work 100%. Maintainers massage and reformat the patches too much before applying them.

regards,
dan carpenter

_______________________________________________
Virtualization mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

#4 lore raw

Re: [PATCH v7 1/4] vdpa: Add suspend operation

Quoted 9 lines · Dan Carpenter
Quoted 8 lines · Michael S. Tsirkin
Quoted 6 lines · Eugenio Pérez

This operation is optional: It it's not implemented, backend feature bit will not be exposed.

Message-Id: <[email protected]>

Quoted 5 lines
Quoted 2 lines

What is this message id doing here?

I like the Message-Id tag. It means you can b4 mbox <mesg-id> and get the thread.

Yes it makes sense in git. But I don't see what it does in this patch posted to the list. It seems to refer to the previous version of the patch here. Which is ok I guess but better called out e.g.

Previous-version: <[email protected]>

Quoted 17 lines

Linus has complained (rough remembering) that everyone is using the
Link: tag for links to the patch itself. It's supposed to be for Links to bugzilla or to the spec or whatever. Extra information, too much to put in the commit message. Now the Link tag is useless because it either points to the patch or to a bugzilla. Depend on what you want it to do, it *always* points to the opposite thing.

But I can't remember what people settled on as the alternative to use to link to lore...

In theory, we should be able to figure out the link to lore automatically and there have been a couple projects which tried to do this but they can't make it work 100%. Maintainers massage and reformat the patches too much before applying them.

regards,
dan carpenter

#5 lore raw

Re: [PATCH v7 1/4] vdpa: Add suspend operation

Quoted 10 lines · Michael S. Tsirkin
Quoted 9 lines · Dan Carpenter
Quoted 8 lines · Michael S. Tsirkin
Quoted 6 lines · Eugenio Pérez

This operation is optional: It it's not implemented, backend feature bit will not be exposed.

Message-Id: <[email protected]>

Quoted 8 lines
Quoted 5 lines
Quoted 2 lines

What is this message id doing here?

I like the Message-Id tag. It means you can b4 mbox <mesg-id> and get the thread.

Yes it makes sense in git. But I don't see what it does in this patch posted to the list.

Ah. Okay. I misunderstood.

Quoted 5 lines

It seems to refer to the previous version of the patch here. Which is ok I guess but better called out e.g.

Previous-version: <[email protected]>

Let's not go over board with storing previous threads. That seems like a headache...

regards,
dan carpenter

_______________________________________________
Virtualization mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

author maintainer
#6 lore raw

PATCH v7 2/4 vhost-vdpa: introduce SUSPEND backend feature bit

In reply to Eugenio Pérez

Userland knows if it can suspend the device or not by checking this feature bit.

It's only offered if the vdpa driver backend implements the suspend() operation callback, and to offer it or userland to ack it if the backend does not offer that callback is an error.

Message-Id: <[email protected]>

---
 drivers/vhost/vdpa.c             | 16 +++++++++++++++-
 include/uapi/linux/vhost_types.h |  2 ++
 2 files changed, 17 insertions(+), 1 deletion(-)
drivers/vhost/vdpa.c +15 -1
index 23dcbfdfa13b..3d636e192061 100644
@@ -347,6 +347,14 @@ static long vhost_vdpa_set_config(struct vhost_vdpa *v,
 	return 0;
 }
 
+static bool vhost_vdpa_can_suspend(const struct vhost_vdpa *v)
+{
+	struct vdpa_device *vdpa = v->vdpa;
+	const struct vdpa_config_ops *ops = vdpa->config;
+
+	return ops->suspend;
+}
+
 static long vhost_vdpa_get_features(struct vhost_vdpa *v, u64 __user *featurep)
 {
 	struct vdpa_device *vdpa = v->vdpa;
@@ -577,7 +585,11 @@ static long vhost_vdpa_unlocked_ioctl(struct file *filep,
 	if (cmd == VHOST_SET_BACKEND_FEATURES) {
 		if (copy_from_user(&features, featurep, sizeof(features)))
 			return -EFAULT;
-		if (features & ~VHOST_VDPA_BACKEND_FEATURES)
+		if (features & ~(VHOST_VDPA_BACKEND_FEATURES |
+				 BIT_ULL(VHOST_BACKEND_F_SUSPEND)))
+			return -EOPNOTSUPP;
+		if ((features & BIT_ULL(VHOST_BACKEND_F_SUSPEND)) &&
+		     !vhost_vdpa_can_suspend(v))
 			return -EOPNOTSUPP;
 		vhost_set_backend_features(&v->vdev, features);
 		return 0;
@@ -628,6 +640,8 @@ static long vhost_vdpa_unlocked_ioctl(struct file *filep,
 		break;
 	case VHOST_GET_BACKEND_FEATURES:
 		features = VHOST_VDPA_BACKEND_FEATURES;
+		if (vhost_vdpa_can_suspend(v))
+			features |= BIT_ULL(VHOST_BACKEND_F_SUSPEND);
 		if (copy_to_user(featurep, &features, sizeof(features)))
 			r = -EFAULT;
 		break;
include/uapi/linux/vhost_types.h +2 -0
index 634cee485abb..1bdd6e363f4c 100644
@@ -161,5 +161,7 @@ struct vhost_vdpa_iova_range {
  * message
  */
 #define VHOST_BACKEND_F_IOTLB_ASID  0x3
+/* Device can be suspended */
+#define VHOST_BACKEND_F_SUSPEND  0x4
 
 #endif
Signature · 1 line
2.31.1
author maintainer
#7 lore raw

PATCH v7 3/4 vhost-vdpa: uAPI to suspend the device

In reply to Eugenio Pérez

The ioctl adds support for suspending the device from userspace.

This is a must before getting virtqueue indexes (base) for live migration, since the device could modify them after userland gets them. There are individual ways to perform that action for some devices (VHOST_NET_SET_BACKEND, VHOST_VSOCK_SET_RUNNING, ...) but there was no way to perform it for any vhost device (and, in particular, vhost-vdpa).

After a successful return of the ioctl call the device must not process more virtqueue descriptors. The device can answer to read or writes of config fields as if it were not suspended. In particular, writing to "queue_enable" with a value of 1 will not make the device start processing buffers of the virtqueue.

Message-Id: <[email protected]>

---
v7: Delete argument to ioctl, unused

---
 drivers/vhost/vdpa.c       | 19 +++++++++++++++++++
 include/uapi/linux/vhost.h |  9 +++++++++
 2 files changed, 28 insertions(+)
drivers/vhost/vdpa.c +19 -0
index 3d636e192061..7fa671ac4bdf 100644
@@ -478,6 +478,22 @@ static long vhost_vdpa_get_vqs_count(struct vhost_vdpa *v, u32 __user *argp)
 	return 0;
 }
 
+/* After a successful return of ioctl the device must not process more
+ * virtqueue descriptors. The device can answer to read or writes of config
+ * fields as if it were not suspended. In particular, writing to "queue_enable"
+ * with a value of 1 will not make the device start processing buffers.
+ */
+static long vhost_vdpa_suspend(struct vhost_vdpa *v)
+{
+	struct vdpa_device *vdpa = v->vdpa;
+	const struct vdpa_config_ops *ops = vdpa->config;
+
+	if (!ops->suspend)
+		return -EOPNOTSUPP;
+
+	return ops->suspend(vdpa);
+}
+
 static long vhost_vdpa_vring_ioctl(struct vhost_vdpa *v, unsigned int cmd,
 				   void __user *argp)
 {
@@ -654,6 +670,9 @@ static long vhost_vdpa_unlocked_ioctl(struct file *filep,
 	case VHOST_VDPA_GET_VQS_COUNT:
 		r = vhost_vdpa_get_vqs_count(v, argp);
 		break;
+	case VHOST_VDPA_SUSPEND:
+		r = vhost_vdpa_suspend(v);
+		break;
 	default:
 		r = vhost_dev_ioctl(&v->vdev, cmd, argp);
 		if (r == -ENOIOCTLCMD)
include/uapi/linux/vhost.h +9 -0
index cab645d4a645..f9f115a7c75b 100644
@@ -171,4 +171,13 @@
 #define VHOST_VDPA_SET_GROUP_ASID	_IOW(VHOST_VIRTIO, 0x7C, \
 					     struct vhost_vring_state)
 
+/* Suspend a device so it does not process virtqueue requests anymore
+ *
+ * After the return of ioctl the device must preserve all the necessary state
+ * (the virtqueue vring base plus the possible device specific states) that is
+ * required for restoring in the future. The device must not change its
+ * configuration after that point.
+ */
+#define VHOST_VDPA_SUSPEND		_IO(VHOST_VIRTIO, 0x7D)
+
 #endif
Signature · 1 line
2.31.1
#8 lore raw

Re: [PATCH v7 3/4] vhost-vdpa: uAPI to suspend the device

Quoted 14 lines · Eugenio Pérez

The ioctl adds support for suspending the device from userspace.

This is a must before getting virtqueue indexes (base) for live migration, since the device could modify them after userland gets them. There are individual ways to perform that action for some devices (VHOST_NET_SET_BACKEND, VHOST_VSOCK_SET_RUNNING, ...) but there was no way to perform it for any vhost device (and, in particular, vhost-vdpa).

After a successful return of the ioctl call the device must not process more virtqueue descriptors. The device can answer to read or writes of config fields as if it were not suspended. In particular, writing to "queue_enable" with a value of 1 will not make the device start processing buffers of the virtqueue.

Quoted 3 lines

Message-Id: <[email protected]>

You are not supposed to include upstream maintainer's signoff like this.

Quoted 64 lines

---
v7: Delete argument to ioctl, unused

---
 drivers/vhost/vdpa.c       | 19 +++++++++++++++++++
 include/uapi/linux/vhost.h |  9 +++++++++
 2 files changed, 28 insertions(+)
drivers/vhost/vdpa.c +19 -0
index 3d636e192061..7fa671ac4bdf 100644
@@ -478,6 +478,22 @@ static long vhost_vdpa_get_vqs_count(struct vhost_vdpa *v, u32 __user *argp)
 	return 0;
 }
 
+/* After a successful return of ioctl the device must not process more
+ * virtqueue descriptors. The device can answer to read or writes of config
+ * fields as if it were not suspended. In particular, writing to "queue_enable"
+ * with a value of 1 will not make the device start processing buffers.
+ */
+static long vhost_vdpa_suspend(struct vhost_vdpa *v)
+{
+	struct vdpa_device *vdpa = v->vdpa;
+	const struct vdpa_config_ops *ops = vdpa->config;
+
+	if (!ops->suspend)
+		return -EOPNOTSUPP;
+
+	return ops->suspend(vdpa);
+}
+
 static long vhost_vdpa_vring_ioctl(struct vhost_vdpa *v, unsigned int cmd,
 				   void __user *argp)
 {
@@ -654,6 +670,9 @@ static long vhost_vdpa_unlocked_ioctl(struct file *filep,
 	case VHOST_VDPA_GET_VQS_COUNT:
 		r = vhost_vdpa_get_vqs_count(v, argp);
 		break;
+	case VHOST_VDPA_SUSPEND:
+		r = vhost_vdpa_suspend(v);
+		break;
 	default:
 		r = vhost_dev_ioctl(&v->vdev, cmd, argp);
 		if (r == -ENOIOCTLCMD)
include/uapi/linux/vhost.h +9 -0
index cab645d4a645..f9f115a7c75b 100644
@@ -171,4 +171,13 @@
 #define VHOST_VDPA_SET_GROUP_ASID	_IOW(VHOST_VIRTIO, 0x7C, \
 					     struct vhost_vring_state)
 
+/* Suspend a device so it does not process virtqueue requests anymore
+ *
+ * After the return of ioctl the device must preserve all the necessary state
+ * (the virtqueue vring base plus the possible device specific states) that is
+ * required for restoring in the future. The device must not change its
+ * configuration after that point.
+ */
+#define VHOST_VDPA_SUSPEND		_IO(VHOST_VIRTIO, 0x7D)
+
 #endif
Signature · 1 line
2.31.1
author maintainer
#9 lore raw

Re: [PATCH v7 3/4] vhost-vdpa: uAPI to suspend the device

Quoted 16 lines · Michael S. Tsirkin
Quoted 14 lines · Eugenio Pérez

The ioctl adds support for suspending the device from userspace.

This is a must before getting virtqueue indexes (base) for live migration, since the device could modify them after userland gets them. There are individual ways to perform that action for some devices (VHOST_NET_SET_BACKEND, VHOST_VSOCK_SET_RUNNING, ...) but there was no way to perform it for any vhost device (and, in particular, vhost-vdpa).

After a successful return of the ioctl call the device must not process more virtqueue descriptors. The device can answer to read or writes of config fields as if it were not suspended. In particular, writing to "queue_enable" with a value of 1 will not make the device start processing buffers of the virtqueue.

Quoted 7 lines
Quoted 3 lines

Message-Id: <[email protected]>

You are not supposed to include upstream maintainer's signoff like this.

I'm very sorry, I modified the commits in your vhost branch and I left the signoff (and message-id) lines by mistake.

Quoted 65 lines
Quoted 64 lines

---
v7: Delete argument to ioctl, unused

---
 drivers/vhost/vdpa.c       | 19 +++++++++++++++++++
 include/uapi/linux/vhost.h |  9 +++++++++
 2 files changed, 28 insertions(+)
drivers/vhost/vdpa.c +19 -0
index 3d636e192061..7fa671ac4bdf 100644
@@ -478,6 +478,22 @@ static long vhost_vdpa_get_vqs_count(struct vhost_vdpa *v, u32 __user *argp)
      return 0;
 }
 
+/* After a successful return of ioctl the device must not process more
+ * virtqueue descriptors. The device can answer to read or writes of config
+ * fields as if it were not suspended. In particular, writing to "queue_enable"
+ * with a value of 1 will not make the device start processing buffers.
+ */
+static long vhost_vdpa_suspend(struct vhost_vdpa *v)
+{
+     struct vdpa_device *vdpa = v->vdpa;
+     const struct vdpa_config_ops *ops = vdpa->config;
+
+     if (!ops->suspend)
+             return -EOPNOTSUPP;
+
+     return ops->suspend(vdpa);
+}
+
 static long vhost_vdpa_vring_ioctl(struct vhost_vdpa *v, unsigned int cmd,
                                 void __user *argp)
 {
@@ -654,6 +670,9 @@ static long vhost_vdpa_unlocked_ioctl(struct file *filep,
      case VHOST_VDPA_GET_VQS_COUNT:
              r = vhost_vdpa_get_vqs_count(v, argp);
              break;
+     case VHOST_VDPA_SUSPEND:
+             r = vhost_vdpa_suspend(v);
+             break;
      default:
              r = vhost_dev_ioctl(&v->vdev, cmd, argp);
              if (r == -ENOIOCTLCMD)
include/uapi/linux/vhost.h +9 -1
index cab645d4a645..f9f115a7c75b 100644
@@ -171,4 +171,13 @@
 #define VHOST_VDPA_SET_GROUP_ASID    _IOW(VHOST_VIRTIO, 0x7C, \
                                           struct vhost_vring_state)
 
+/* Suspend a device so it does not process virtqueue requests anymore
+ *
+ * After the return of ioctl the device must preserve all the necessary state
+ * (the virtqueue vring base plus the possible device specific states) that is
+ * required for restoring in the future. The device must not change its
+ * configuration after that point.
+ */
+#define VHOST_VDPA_SUSPEND           _IO(VHOST_VIRTIO, 0x7D)
+
 #endif
--

2.31.1

author maintainer
#10 lore raw

PATCH v7 4/4 vdpa_sim: Implement suspend vdpa op

In reply to Eugenio Pérez

Implement suspend operation for vdpa_sim devices, so vhost-vdpa will offer that backend feature and userspace can effectively suspend the device.

This is a must before get virtqueue indexes (base) for live migration, since the device could modify them after userland gets them. There are individual ways to perform that action for some devices (VHOST_NET_SET_BACKEND, VHOST_VSOCK_SET_RUNNING, ...) but there was no way to perform it for any vhost device (and, in particular, vhost-vdpa).

Message-Id: <[email protected]>

---
v7: Delete leftover resume code

---
 drivers/vdpa/vdpa_sim/vdpa_sim.c     | 14 ++++++++++++++
 drivers/vdpa/vdpa_sim/vdpa_sim.h     |  1 +
 drivers/vdpa/vdpa_sim/vdpa_sim_blk.c |  3 +++
 drivers/vdpa/vdpa_sim/vdpa_sim_net.c |  3 +++
 4 files changed, 21 insertions(+)
drivers/vdpa/vdpa_sim/vdpa_sim.c +14 -0
index 0f2865899647..79a50edf8998 100644
@@ -107,6 +107,7 @@ static void vdpasim_do_reset(struct vdpasim *vdpasim)
 	for (i = 0; i < vdpasim->dev_attr.nas; i++)
 		vhost_iotlb_reset(&vdpasim->iommu[i]);
 
+	vdpasim->running = true;
 	spin_unlock(&vdpasim->iommu_lock);
 
 	vdpasim->features = 0;
@@ -505,6 +506,17 @@ static int vdpasim_reset(struct vdpa_device *vdpa)
 	return 0;
 }
 
+static int vdpasim_suspend(struct vdpa_device *vdpa)
+{
+	struct vdpasim *vdpasim = vdpa_to_sim(vdpa);
+
+	spin_lock(&vdpasim->lock);
+	vdpasim->running = false;
+	spin_unlock(&vdpasim->lock);
+
+	return 0;
+}
+
 static size_t vdpasim_get_config_size(struct vdpa_device *vdpa)
 {
 	struct vdpasim *vdpasim = vdpa_to_sim(vdpa);
@@ -694,6 +706,7 @@ static const struct vdpa_config_ops vdpasim_config_ops = {
 	.get_status             = vdpasim_get_status,
 	.set_status             = vdpasim_set_status,
 	.reset			= vdpasim_reset,
+	.suspend		= vdpasim_suspend,
 	.get_config_size        = vdpasim_get_config_size,
 	.get_config             = vdpasim_get_config,
 	.set_config             = vdpasim_set_config,
@@ -726,6 +739,7 @@ static const struct vdpa_config_ops vdpasim_batch_config_ops = {
 	.get_status             = vdpasim_get_status,
 	.set_status             = vdpasim_set_status,
 	.reset			= vdpasim_reset,
+	.suspend		= vdpasim_suspend,
 	.get_config_size        = vdpasim_get_config_size,
 	.get_config             = vdpasim_get_config,
 	.set_config             = vdpasim_set_config,
drivers/vdpa/vdpa_sim/vdpa_sim.h +1 -0
index 622782e92239..061986f30911 100644
@@ -66,6 +66,7 @@ struct vdpasim {
 	u32 generation;
 	u64 features;
 	u32 groups;
+	bool running;
 	/* spinlock to synchronize iommu table */
 	spinlock_t iommu_lock;
 };
drivers/vdpa/vdpa_sim/vdpa_sim_blk.c +3 -0
index 42d401d43911..bcdb1982c378 100644
@@ -204,6 +204,9 @@ static void vdpasim_blk_work(struct work_struct *work)
 	if (!(vdpasim->status & VIRTIO_CONFIG_S_DRIVER_OK))
 		goto out;
 
+	if (!vdpasim->running)
+		goto out;
+
 	for (i = 0; i < VDPASIM_BLK_VQ_NUM; i++) {
 		struct vdpasim_virtqueue *vq = &vdpasim->vqs[i];
 
drivers/vdpa/vdpa_sim/vdpa_sim_net.c +3 -0
index 5125976a4df8..886449e88502 100644
@@ -154,6 +154,9 @@ static void vdpasim_net_work(struct work_struct *work)
 
 	spin_lock(&vdpasim->lock);
 
+	if (!vdpasim->running)
+		goto out;
+
 	if (!(vdpasim->status & VIRTIO_CONFIG_S_DRIVER_OK))
 		goto out;
 
Signature · 1 line
2.31.1
#11 lore raw
In reply to Eugenio Pérez
Quoted 14 lines · Eugenio Pérez

Implement suspend operation for vdpa_sim devices, so vhost-vdpa will offer that backend feature and userspace can effectively suspend the device.

This is a must before getting virtqueue indexes (base) for live migration, since the device could modify them after userland gets them. There are individual ways to perform that action for some devices (VHOST_NET_SET_BACKEND, VHOST_VSOCK_SET_RUNNING, ...) but there was no way to perform it for any vhost device (and, in particular, vhost-vdpa).

After a successful return of ioctl the device must not process more virtqueue descriptors. The device can answer to read or writes of config fields as if it were not suspended. In particular, writing to "queue_enable" with a value of 1 will not make the device start processing virtqueue buffers.

Quoted 4 lines

In the future, we will provide features similar to VHOST_USER_GET_INFLIGHT_FD so the device can save pending operations.

Applied on top of [1] branch after removing the old commits.

Except, I can't really do this without invaliding all testing. Can't you post an incremental patch?

Quoted 47 lines

Comments are welcome.

v7:
* Remove ioctl leftover argument and update doc accordingly.

v6:
* Remove the resume operation, making the ioctl simpler. We can always add another ioctl for VM_STOP/VM_RESUME operation later.
* s/stop/suspend/ to differentiate more from reset.
* Clarify scope of the suspend operation.

v5:
* s/not stop/resume/ in doc.

v4:
* Replace VHOST_STOP to VHOST_VDPA_STOP in vhost ioctl switch case too.

v3:
* s/VHOST_STOP/VHOST_VDPA_STOP/
* Add documentation and requirements of the ioctl above its definition.

v2:
* Replace raw _F_STOP with BIT_ULL(_F_STOP).
* Fix obtaining of stop ioctl arg (it was not obtained but written).
* Add stop to vdpa_sim_blk.

[1] git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost.git

Eugenio Pérez (4):
vdpa: Add suspend operation
vhost-vdpa: introduce SUSPEND backend feature bit
vhost-vdpa: uAPI to suspend the device
vdpa_sim: Implement suspend vdpa op

 drivers/vdpa/vdpa_sim/vdpa_sim.c     | 14 +++++++++++
 drivers/vdpa/vdpa_sim/vdpa_sim.h     |  1 +
 drivers/vdpa/vdpa_sim/vdpa_sim_blk.c |  3 +++
 drivers/vdpa/vdpa_sim/vdpa_sim_net.c |  3 +++
 drivers/vhost/vdpa.c                 | 35 +++++++++++++++++++++++++++-
 include/linux/vdpa.h                 |  4 ++++
 include/uapi/linux/vhost.h           |  9 +++++++
 include/uapi/linux/vhost_types.h     |  2 ++
 8 files changed, 70 insertions(+), 1 deletion(-)
Signature · 1 line
2.31.1
author maintainer
#12 lore raw
Quoted 16 lines · Michael S. Tsirkin
Quoted 14 lines · Eugenio Pérez

Implement suspend operation for vdpa_sim devices, so vhost-vdpa will offer that backend feature and userspace can effectively suspend the device.

This is a must before getting virtqueue indexes (base) for live migration, since the device could modify them after userland gets them. There are individual ways to perform that action for some devices (VHOST_NET_SET_BACKEND, VHOST_VSOCK_SET_RUNNING, ...) but there was no way to perform it for any vhost device (and, in particular, vhost-vdpa).

After a successful return of ioctl the device must not process more virtqueue descriptors. The device can answer to read or writes of config fields as if it were not suspended. In particular, writing to "queue_enable" with a value of 1 will not make the device start processing virtqueue buffers.

Quoted 8 lines
Quoted 4 lines

In the future, we will provide features similar to VHOST_USER_GET_INFLIGHT_FD so the device can save pending operations.

Applied on top of [1] branch after removing the old commits.

Except, I can't really do this without invaliding all testing. Can't you post an incremental patch?

Oops, sorry.

I can send it for doc and internal code. But is it ok to remove an ioctl arg with incremental patches?

Thanks!

Quoted 48 lines
Quoted 47 lines

Comments are welcome.

v7:
* Remove ioctl leftover argument and update doc accordingly.

v6:
* Remove the resume operation, making the ioctl simpler. We can always add another ioctl for VM_STOP/VM_RESUME operation later.
* s/stop/suspend/ to differentiate more from reset.
* Clarify scope of the suspend operation.

v5:
* s/not stop/resume/ in doc.

v4:
* Replace VHOST_STOP to VHOST_VDPA_STOP in vhost ioctl switch case too.

v3:
* s/VHOST_STOP/VHOST_VDPA_STOP/
* Add documentation and requirements of the ioctl above its definition.

v2:
* Replace raw _F_STOP with BIT_ULL(_F_STOP).
* Fix obtaining of stop ioctl arg (it was not obtained but written).
* Add stop to vdpa_sim_blk.

[1] git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost.git

Eugenio Pérez (4):
vdpa: Add suspend operation
vhost-vdpa: introduce SUSPEND backend feature bit
vhost-vdpa: uAPI to suspend the device
vdpa_sim: Implement suspend vdpa op

 drivers/vdpa/vdpa_sim/vdpa_sim.c     | 14 +++++++++++
 drivers/vdpa/vdpa_sim/vdpa_sim.h     |  1 +
 drivers/vdpa/vdpa_sim/vdpa_sim_blk.c |  3 +++
 drivers/vdpa/vdpa_sim/vdpa_sim_net.c |  3 +++
 drivers/vhost/vdpa.c                 | 35 +++++++++++++++++++++++++++-
 include/linux/vdpa.h                 |  4 ++++
 include/uapi/linux/vhost.h           |  9 +++++++
 include/uapi/linux/vhost_types.h     |  2 ++
 8 files changed, 70 insertions(+), 1 deletion(-)

--
2.31.1