Skip to contentSkip to conversation

vhost_vdpa: assign irq bypass producer token correctly

RFC PATCH In review Jason Wang 3 participants 10 messages Aug 8, 2024 – Aug 14, 2024

Subsystems (5)
lore
Share & tools
lore

Jason Wang

author
#0 lore raw

RFC PATCH 1/1 vhost_vdpa: assign irq bypass producer token correctly

We used to call irq_bypass_unregister_producer() in vhost_vdpa_setup_vq_irq() which is problematic as we don't know if the token pointer is still valid or not.

Actually, we use the eventfd_ctx as the token so the life cycle of the token should be bound to the VHOST_SET_VRING_CALL instead of vhost_vdpa_setup_vq_irq() which could be called by set_status().

Fixing this by setting up irq bypass producer's token when handling VHOST_SET_VRING_CALL and un-registering the producer before calling vhost_vring_ioctl() to prevent a possible use after free as eventfd could have been released in vhost_vring_ioctl().

Fixes 2cf1ba9a4d15 ("vhost_vdpa: implement IRQ offloading in vhost_vdpa")

---
Note for Dragos: Please check whether this fixes your issue. I slightly test it with vp_vdpa in L2.

---
 drivers/vhost/vdpa.c | 12 +++++++++---
 1 file changed, 9 insertions(+), 3 deletions(-)
drivers/vhost/vdpa.c +9 -3
index e31ec9ebc4ce..388226a48bcc 100644
@@ -209,11 +209,9 @@ static void vhost_vdpa_setup_vq_irq(struct vhost_vdpa *v, u16 qid)
 	if (irq < 0)
 		return;
 
-	irq_bypass_unregister_producer(&vq->call_ctx.producer);
 	if (!vq->call_ctx.ctx)
 		return;
 
-	vq->call_ctx.producer.token = vq->call_ctx.ctx;
 	vq->call_ctx.producer.irq = irq;
 	ret = irq_bypass_register_producer(&vq->call_ctx.producer);
 	if (unlikely(ret))
@@ -709,6 +707,12 @@ static long vhost_vdpa_vring_ioctl(struct vhost_vdpa *v, unsigned int cmd,
 			vq->last_avail_idx = vq_state.split.avail_index;
 		}
 		break;
+	case VHOST_SET_VRING_CALL:
+		if (vq->call_ctx.ctx) {
+			vhost_vdpa_unsetup_vq_irq(v, idx);
+			vq->call_ctx.producer.token = NULL;
+		}
+		break;
 	}
 
 	r = vhost_vring_ioctl(&v->vdev, cmd, argp);
@@ -747,13 +751,14 @@ static long vhost_vdpa_vring_ioctl(struct vhost_vdpa *v, unsigned int cmd,
 			cb.callback = vhost_vdpa_virtqueue_cb;
 			cb.private = vq;
 			cb.trigger = vq->call_ctx.ctx;
+			vq->call_ctx.producer.token = vq->call_ctx.ctx;
+			vhost_vdpa_setup_vq_irq(v, idx);
 		} else {
 			cb.callback = NULL;
 			cb.private = NULL;
 			cb.trigger = NULL;
 		}
 		ops->set_vq_cb(vdpa, idx, &cb);
-		vhost_vdpa_setup_vq_irq(v, idx);
 		break;
 
 	case VHOST_SET_VRING_NUM:
@@ -1419,6 +1424,7 @@ static int vhost_vdpa_open(struct inode *inode, struct file *filep)
 	for (i = 0; i < nvqs; i++) {
 		vqs[i] = &v->vqs[i];
 		vqs[i]->handle_kick = handle_vq_kick;
+		vqs[i]->call_ctx.ctx = NULL;
 	}
 	vhost_dev_init(dev, vqs, nvqs, 0, 0, 0, false,
 		       vhost_vdpa_process_iotlb_msg);
Signature · 1 line
2.31.1
#1 lore raw
Quoted 68 lines · Jason Wang

We used to call irq_bypass_unregister_producer() in vhost_vdpa_setup_vq_irq() which is problematic as we don't know if the token pointer is still valid or not.

Actually, we use the eventfd_ctx as the token so the life cycle of the token should be bound to the VHOST_SET_VRING_CALL instead of vhost_vdpa_setup_vq_irq() which could be called by set_status().

Fixing this by setting up irq bypass producer's token when handling VHOST_SET_VRING_CALL and un-registering the producer before calling vhost_vring_ioctl() to prevent a possible use after free as eventfd could have been released in vhost_vring_ioctl().

Fixes 2cf1ba9a4d15 ("vhost_vdpa: implement IRQ offloading in vhost_vdpa")

---
Note for Dragos: Please check whether this fixes your issue. I slightly test it with vp_vdpa in L2.

---
 drivers/vhost/vdpa.c | 12 +++++++++---
 1 file changed, 9 insertions(+), 3 deletions(-)
drivers/vhost/vdpa.c +8 -3
index e31ec9ebc4ce..388226a48bcc 100644
@@ -209,11 +209,9 @@ static void vhost_vdpa_setup_vq_irq(struct vhost_vdpa *v, u16 qid)
 	if (irq < 0)
 		return;
 
-	irq_bypass_unregister_producer(&vq->call_ctx.producer);
 	if (!vq->call_ctx.ctx)
 		return;
 
-	vq->call_ctx.producer.token = vq->call_ctx.ctx;
 	vq->call_ctx.producer.irq = irq;
 	ret = irq_bypass_register_producer(&vq->call_ctx.producer);
 	if (unlikely(ret))
@@ -709,6 +707,12 @@ static long vhost_vdpa_vring_ioctl(struct vhost_vdpa *v, unsigned int cmd,
 			vq->last_avail_idx = vq_state.split.avail_index;
 		}
 		break;
+	case VHOST_SET_VRING_CALL:
+		if (vq->call_ctx.ctx) {
+			vhost_vdpa_unsetup_vq_irq(v, idx);
+			vq->call_ctx.producer.token = NULL;
+		}
+		break;
 	}
 
 	r = vhost_vring_ioctl(&v->vdev, cmd, argp);
@@ -747,13 +751,14 @@ static long vhost_vdpa_vring_ioctl(struct vhost_vdpa *v, unsigned int cmd,
 			cb.callback = vhost_vdpa_virtqueue_cb;
 			cb.private = vq;
 			cb.trigger = vq->call_ctx.ctx;
+			vq->call_ctx.producer.token = vq->call_ctx.ctx;
+			vhost_vdpa_setup_vq_irq(v, idx);
 		} else {
 			cb.callback = NULL;
 			cb.private = NULL;
 			cb.trigger = NULL;
 		}
 		ops->set_vq_cb(vdpa, idx, &cb);
-		vhost_vdpa_setup_vq_irq(v, idx);
 		break;
 
 	case VHOST_SET_VRING_NUM:

drivers/vhost/vdpa.c static int vhost_vdpa_open(struct inode *inode, struct file *filep) · line 1430

Quoted 7 lines
drivers/vhost/vdpa.c +1 -0
@@ -1419,6 +1424,7 @@ static int vhost_vdpa_open(struct inode *inode, struct file *filep)
 	for (i = 0; i < nvqs; i++) {
 		vqs[i] = &v->vqs[i];
 		vqs[i]->handle_kick = handle_vq_kick;
+		vqs[i]->call_ctx.ctx = NULL;
 	}
 	vhost_dev_init(dev, vqs, nvqs, 0, 0, 0, false,
 		       vhost_vdpa_process_iotlb_msg);

No more crashes, but now getting a lot of:
vhost-vdpa-X: vq Y, irq bypass producer (token 00000000a66e28ab) registration fails, ret = -16

... seems like the irq_bypass_unregister_producer() that was removed might still be needed somewhere?

Thanks,
Dragos

author
#2 lore raw
Quoted 83 lines · Dragos Tatulea
Quoted 68 lines · Jason Wang

We used to call irq_bypass_unregister_producer() in vhost_vdpa_setup_vq_irq() which is problematic as we don't know if the token pointer is still valid or not.

Actually, we use the eventfd_ctx as the token so the life cycle of the token should be bound to the VHOST_SET_VRING_CALL instead of vhost_vdpa_setup_vq_irq() which could be called by set_status().

Fixing this by setting up irq bypass producer's token when handling VHOST_SET_VRING_CALL and un-registering the producer before calling vhost_vring_ioctl() to prevent a possible use after free as eventfd could have been released in vhost_vring_ioctl().

Fixes 2cf1ba9a4d15 ("vhost_vdpa: implement IRQ offloading in vhost_vdpa")

---
Note for Dragos: Please check whether this fixes your issue. I slightly test it with vp_vdpa in L2.

---
 drivers/vhost/vdpa.c | 12 +++++++++---
 1 file changed, 9 insertions(+), 3 deletions(-)
drivers/vhost/vdpa.c +8 -3
index e31ec9ebc4ce..388226a48bcc 100644
@@ -209,11 +209,9 @@ static void vhost_vdpa_setup_vq_irq(struct vhost_vdpa *v, u16 qid)
      if (irq < 0)
              return;
 
-     irq_bypass_unregister_producer(&vq->call_ctx.producer);
      if (!vq->call_ctx.ctx)
              return;
 
-     vq->call_ctx.producer.token = vq->call_ctx.ctx;
      vq->call_ctx.producer.irq = irq;
      ret = irq_bypass_register_producer(&vq->call_ctx.producer);
      if (unlikely(ret))
@@ -709,6 +707,12 @@ static long vhost_vdpa_vring_ioctl(struct vhost_vdpa *v, unsigned int cmd,
                      vq->last_avail_idx = vq_state.split.avail_index;
              }
              break;
+     case VHOST_SET_VRING_CALL:
+             if (vq->call_ctx.ctx) {
+                     vhost_vdpa_unsetup_vq_irq(v, idx);
+                     vq->call_ctx.producer.token = NULL;
+             }
+             break;
      }
 
      r = vhost_vring_ioctl(&v->vdev, cmd, argp);
@@ -747,13 +751,14 @@ static long vhost_vdpa_vring_ioctl(struct vhost_vdpa *v, unsigned int cmd,
                      cb.callback = vhost_vdpa_virtqueue_cb;
                      cb.private = vq;
                      cb.trigger = vq->call_ctx.ctx;
+                     vq->call_ctx.producer.token = vq->call_ctx.ctx;
+                     vhost_vdpa_setup_vq_irq(v, idx);
              } else {
                      cb.callback = NULL;
                      cb.private = NULL;
                      cb.trigger = NULL;
              }
              ops->set_vq_cb(vdpa, idx, &cb);
-             vhost_vdpa_setup_vq_irq(v, idx);
              break;
 
      case VHOST_SET_VRING_NUM:

drivers/vhost/vdpa.c static int vhost_vdpa_open(struct inode *inode, struct file *filep) · line 1430

Quoted 7 lines
drivers/vhost/vdpa.c +1 -0
@@ -1419,6 +1424,7 @@ static int vhost_vdpa_open(struct inode *inode, struct file *filep)
      for (i = 0; i < nvqs; i++) {
              vqs[i] = &v->vqs[i];
              vqs[i]->handle_kick = handle_vq_kick;
+             vqs[i]->call_ctx.ctx = NULL;
      }
      vhost_dev_init(dev, vqs, nvqs, 0, 0, 0, false,
                     vhost_vdpa_process_iotlb_msg);

No more crashes, but now getting a lot of:
vhost-vdpa-X: vq Y, irq bypass producer (token 00000000a66e28ab) registration fails, ret = -16

Quoted 2 lines

... seems like the irq_bypass_unregister_producer() that was removed might still be needed somewhere?

Probably, but I didn't see this when testing vp_vdpa.

When did you meet those warnings? Is it during the boot or migration?

Thanks

Quoted 5 lines

Thanks,
Dragos

author
#3 lore raw
Quoted 85 lines · Jason Wang
Quoted 83 lines · Dragos Tatulea
Quoted 68 lines · Jason Wang

We used to call irq_bypass_unregister_producer() in vhost_vdpa_setup_vq_irq() which is problematic as we don't know if the token pointer is still valid or not.

Actually, we use the eventfd_ctx as the token so the life cycle of the token should be bound to the VHOST_SET_VRING_CALL instead of vhost_vdpa_setup_vq_irq() which could be called by set_status().

Fixing this by setting up irq bypass producer's token when handling VHOST_SET_VRING_CALL and un-registering the producer before calling vhost_vring_ioctl() to prevent a possible use after free as eventfd could have been released in vhost_vring_ioctl().

Fixes 2cf1ba9a4d15 ("vhost_vdpa: implement IRQ offloading in vhost_vdpa")

---
Note for Dragos: Please check whether this fixes your issue. I slightly test it with vp_vdpa in L2.

---
 drivers/vhost/vdpa.c | 12 +++++++++---
 1 file changed, 9 insertions(+), 3 deletions(-)
drivers/vhost/vdpa.c +8 -3
index e31ec9ebc4ce..388226a48bcc 100644
@@ -209,11 +209,9 @@ static void vhost_vdpa_setup_vq_irq(struct vhost_vdpa *v, u16 qid)
      if (irq < 0)
              return;
 
-     irq_bypass_unregister_producer(&vq->call_ctx.producer);
      if (!vq->call_ctx.ctx)
              return;
 
-     vq->call_ctx.producer.token = vq->call_ctx.ctx;
      vq->call_ctx.producer.irq = irq;
      ret = irq_bypass_register_producer(&vq->call_ctx.producer);
      if (unlikely(ret))
@@ -709,6 +707,12 @@ static long vhost_vdpa_vring_ioctl(struct vhost_vdpa *v, unsigned int cmd,
                      vq->last_avail_idx = vq_state.split.avail_index;
              }
              break;
+     case VHOST_SET_VRING_CALL:
+             if (vq->call_ctx.ctx) {
+                     vhost_vdpa_unsetup_vq_irq(v, idx);
+                     vq->call_ctx.producer.token = NULL;
+             }
+             break;
      }
 
      r = vhost_vring_ioctl(&v->vdev, cmd, argp);
@@ -747,13 +751,14 @@ static long vhost_vdpa_vring_ioctl(struct vhost_vdpa *v, unsigned int cmd,
                      cb.callback = vhost_vdpa_virtqueue_cb;
                      cb.private = vq;
                      cb.trigger = vq->call_ctx.ctx;
+                     vq->call_ctx.producer.token = vq->call_ctx.ctx;
+                     vhost_vdpa_setup_vq_irq(v, idx);
              } else {
                      cb.callback = NULL;
                      cb.private = NULL;
                      cb.trigger = NULL;
              }
              ops->set_vq_cb(vdpa, idx, &cb);
-             vhost_vdpa_setup_vq_irq(v, idx);
              break;
 
      case VHOST_SET_VRING_NUM:

drivers/vhost/vdpa.c static int vhost_vdpa_open(struct inode *inode, struct file *filep) · line 1430

Quoted 7 lines
drivers/vhost/vdpa.c +1 -0
@@ -1419,6 +1424,7 @@ static int vhost_vdpa_open(struct inode *inode, struct file *filep)
      for (i = 0; i < nvqs; i++) {
              vqs[i] = &v->vqs[i];
              vqs[i]->handle_kick = handle_vq_kick;
+             vqs[i]->call_ctx.ctx = NULL;
      }
      vhost_dev_init(dev, vqs, nvqs, 0, 0, 0, false,
                     vhost_vdpa_process_iotlb_msg);

No more crashes, but now getting a lot of:
vhost-vdpa-X: vq Y, irq bypass producer (token 00000000a66e28ab) registration fails, ret = -16

Quoted 6 lines
Quoted 2 lines

... seems like the irq_bypass_unregister_producer() that was removed might still be needed somewhere?

Probably, but I didn't see this when testing vp_vdpa.

When did you meet those warnings? Is it during the boot or migration?

Btw, it would be helpful to check if mlx5_get_vq_irq() works correctly. I believe it should return an error if the virtqueue interrupt is not allocated. After a glance at the code, it seems not straightforward to me.

Thanks

Quoted 8 lines

Thanks

Quoted 5 lines

Thanks,
Dragos

#4 lore raw
Quoted 86 lines · Jason Wang
Quoted 85 lines · Jason Wang
Quoted 83 lines · Dragos Tatulea
Quoted 68 lines · Jason Wang

We used to call irq_bypass_unregister_producer() in vhost_vdpa_setup_vq_irq() which is problematic as we don't know if the token pointer is still valid or not.

Actually, we use the eventfd_ctx as the token so the life cycle of the token should be bound to the VHOST_SET_VRING_CALL instead of vhost_vdpa_setup_vq_irq() which could be called by set_status().

Fixing this by setting up irq bypass producer's token when handling VHOST_SET_VRING_CALL and un-registering the producer before calling vhost_vring_ioctl() to prevent a possible use after free as eventfd could have been released in vhost_vring_ioctl().

Fixes 2cf1ba9a4d15 ("vhost_vdpa: implement IRQ offloading in vhost_vdpa")

---
Note for Dragos: Please check whether this fixes your issue. I slightly test it with vp_vdpa in L2.

---
 drivers/vhost/vdpa.c | 12 +++++++++---
 1 file changed, 9 insertions(+), 3 deletions(-)
drivers/vhost/vdpa.c +8 -3
index e31ec9ebc4ce..388226a48bcc 100644
@@ -209,11 +209,9 @@ static void vhost_vdpa_setup_vq_irq(struct vhost_vdpa *v, u16 qid)
      if (irq < 0)
              return;
 
-     irq_bypass_unregister_producer(&vq->call_ctx.producer);
      if (!vq->call_ctx.ctx)
              return;
 
-     vq->call_ctx.producer.token = vq->call_ctx.ctx;
      vq->call_ctx.producer.irq = irq;
      ret = irq_bypass_register_producer(&vq->call_ctx.producer);
      if (unlikely(ret))
@@ -709,6 +707,12 @@ static long vhost_vdpa_vring_ioctl(struct vhost_vdpa *v, unsigned int cmd,
                      vq->last_avail_idx = vq_state.split.avail_index;
              }
              break;
+     case VHOST_SET_VRING_CALL:
+             if (vq->call_ctx.ctx) {
+                     vhost_vdpa_unsetup_vq_irq(v, idx);
+                     vq->call_ctx.producer.token = NULL;
+             }
+             break;
      }
 
      r = vhost_vring_ioctl(&v->vdev, cmd, argp);
@@ -747,13 +751,14 @@ static long vhost_vdpa_vring_ioctl(struct vhost_vdpa *v, unsigned int cmd,
                      cb.callback = vhost_vdpa_virtqueue_cb;
                      cb.private = vq;
                      cb.trigger = vq->call_ctx.ctx;
+                     vq->call_ctx.producer.token = vq->call_ctx.ctx;
+                     vhost_vdpa_setup_vq_irq(v, idx);
              } else {
                      cb.callback = NULL;
                      cb.private = NULL;
                      cb.trigger = NULL;
              }
              ops->set_vq_cb(vdpa, idx, &cb);
-             vhost_vdpa_setup_vq_irq(v, idx);
              break;
 
      case VHOST_SET_VRING_NUM:

drivers/vhost/vdpa.c static int vhost_vdpa_open(struct inode *inode, struct file *filep) · line 1430

Quoted 7 lines
drivers/vhost/vdpa.c +1 -0
@@ -1419,6 +1424,7 @@ static int vhost_vdpa_open(struct inode *inode, struct file *filep)
      for (i = 0; i < nvqs; i++) {
              vqs[i] = &v->vqs[i];
              vqs[i]->handle_kick = handle_vq_kick;
+             vqs[i]->call_ctx.ctx = NULL;
      }
      vhost_dev_init(dev, vqs, nvqs, 0, 0, 0, false,
                     vhost_vdpa_process_iotlb_msg);

No more crashes, but now getting a lot of:
vhost-vdpa-X: vq Y, irq bypass producer (token 00000000a66e28ab) registration fails, ret = -16

Quoted 3 lines
Quoted 3 lines
Quoted 2 lines

... seems like the irq_bypass_unregister_producer() that was removed might still be needed somewhere?

My statement above was not quite correct. The error comes from the VQ irq being registered twice:

1) VHOST_SET_VRING_CALL ioctl gets called for vq 0. VQ irq is unregistered (vhost_vdpa_unsetup_vq_irq() and re-registered (vhost_vdpa_setup_vq_irq()) successfully. So far so good.

2) set status !DRIVER_OK -> DRIVER_OK happens. VQ irq setup is done once again (vhost_vdpa_setup_vq_irq()). As the producer unregister was removed in this patch, the register will complain because the producer token already exists.

Quoted 3 lines
Quoted 3 lines

Probably, but I didn't see this when testing vp_vdpa.

When did you meet those warnings? Is it during the boot or migration?

During boot, on the first 2 VQs only (so before the QPs are resized). Traffic does work though when the VM is booted.

Quoted 6 lines

Btw, it would be helpful to check if mlx5_get_vq_irq() works correctly. I believe it should return an error if the virtqueue interrupt is not allocated. After a glance at the code, it seems not straightforward to me.

I think we're good on that front:
mlx5_get_vq_irq() returns EOPNOTSUPP if the vq irq is not allocated.

Thanks,
Dragos

author
#5 lore raw
Quoted 96 lines · Dragos Tatulea
Quoted 86 lines · Jason Wang
Quoted 85 lines · Jason Wang
Quoted 83 lines · Dragos Tatulea
Quoted 68 lines · Jason Wang

We used to call irq_bypass_unregister_producer() in vhost_vdpa_setup_vq_irq() which is problematic as we don't know if the token pointer is still valid or not.

Actually, we use the eventfd_ctx as the token so the life cycle of the token should be bound to the VHOST_SET_VRING_CALL instead of vhost_vdpa_setup_vq_irq() which could be called by set_status().

Fixing this by setting up irq bypass producer's token when handling VHOST_SET_VRING_CALL and un-registering the producer before calling vhost_vring_ioctl() to prevent a possible use after free as eventfd could have been released in vhost_vring_ioctl().

Fixes 2cf1ba9a4d15 ("vhost_vdpa: implement IRQ offloading in vhost_vdpa")

---
Note for Dragos: Please check whether this fixes your issue. I slightly test it with vp_vdpa in L2.

---
 drivers/vhost/vdpa.c | 12 +++++++++---
 1 file changed, 9 insertions(+), 3 deletions(-)
drivers/vhost/vdpa.c +8 -3
index e31ec9ebc4ce..388226a48bcc 100644
@@ -209,11 +209,9 @@ static void vhost_vdpa_setup_vq_irq(struct vhost_vdpa *v, u16 qid)
      if (irq < 0)
              return;
 
-     irq_bypass_unregister_producer(&vq->call_ctx.producer);
      if (!vq->call_ctx.ctx)
              return;
 
-     vq->call_ctx.producer.token = vq->call_ctx.ctx;
      vq->call_ctx.producer.irq = irq;
      ret = irq_bypass_register_producer(&vq->call_ctx.producer);
      if (unlikely(ret))
@@ -709,6 +707,12 @@ static long vhost_vdpa_vring_ioctl(struct vhost_vdpa *v, unsigned int cmd,
                      vq->last_avail_idx = vq_state.split.avail_index;
              }
              break;
+     case VHOST_SET_VRING_CALL:
+             if (vq->call_ctx.ctx) {
+                     vhost_vdpa_unsetup_vq_irq(v, idx);
+                     vq->call_ctx.producer.token = NULL;
+             }
+             break;
      }
 
      r = vhost_vring_ioctl(&v->vdev, cmd, argp);
@@ -747,13 +751,14 @@ static long vhost_vdpa_vring_ioctl(struct vhost_vdpa *v, unsigned int cmd,
                      cb.callback = vhost_vdpa_virtqueue_cb;
                      cb.private = vq;
                      cb.trigger = vq->call_ctx.ctx;
+                     vq->call_ctx.producer.token = vq->call_ctx.ctx;
+                     vhost_vdpa_setup_vq_irq(v, idx);
              } else {
                      cb.callback = NULL;
                      cb.private = NULL;
                      cb.trigger = NULL;
              }
              ops->set_vq_cb(vdpa, idx, &cb);
-             vhost_vdpa_setup_vq_irq(v, idx);
              break;
 
      case VHOST_SET_VRING_NUM:

drivers/vhost/vdpa.c static int vhost_vdpa_open(struct inode *inode, struct file *filep) · line 1430

Quoted 7 lines
drivers/vhost/vdpa.c +1 -0
@@ -1419,6 +1424,7 @@ static int vhost_vdpa_open(struct inode *inode, struct file *filep)
      for (i = 0; i < nvqs; i++) {
              vqs[i] = &v->vqs[i];
              vqs[i]->handle_kick = handle_vq_kick;
+             vqs[i]->call_ctx.ctx = NULL;
      }
      vhost_dev_init(dev, vqs, nvqs, 0, 0, 0, false,
                     vhost_vdpa_process_iotlb_msg);

No more crashes, but now getting a lot of:
vhost-vdpa-X: vq Y, irq bypass producer (token 00000000a66e28ab) registration fails, ret = -16

Quoted 3 lines
Quoted 3 lines
Quoted 2 lines

... seems like the irq_bypass_unregister_producer() that was removed might still be needed somewhere?

My statement above was not quite correct. The error comes from the VQ irq being registered twice:

Quoted 8 lines

1) VHOST_SET_VRING_CALL ioctl gets called for vq 0. VQ irq is unregistered (vhost_vdpa_unsetup_vq_irq() and re-registered (vhost_vdpa_setup_vq_irq()) successfully. So far so good.

2) set status !DRIVER_OK -> DRIVER_OK happens. VQ irq setup is done once again (vhost_vdpa_setup_vq_irq()). As the producer unregister was removed in this patch, the register will complain because the producer token already exists.

I see. I think it's probably too tricky to try to register/unregister a producer during set_vring_call if DRIVER_OK is not set.

Does it work if we only do vhost_vdpa_unsetup/setup_vq_irq() if DRIVER_OK is set in vhost_vdpa_vring_ioctl() (on top of this patch)?

drivers/vhost/vdpa.c +0 -0
index 388226a48bcc..ab441b8ccd2e 100644
@@ -709,7 +709,9 @@ static long vhost_vdpa_vring_ioctl(struct

vhost_vdpa *v, unsigned int cmd,

                break;
        case VHOST_SET_VRING_CALL:
                if (vq->call_ctx.ctx) {
-                       vhost_vdpa_unsetup_vq_irq(v, idx);
+                       if (ops->get_status(vdpa) &
+                           VIRTIO_CONFIG_S_DRIVER_OK)
+                               vhost_vdpa_unsetup_vq_irq(v, idx);
                        vq->call_ctx.producer.token = NULL;
                }
                break;
@@ -752,7 +754,9 @@ static long vhost_vdpa_vring_ioctl(struct
vhost_vdpa *v, unsigned int cmd,
                        cb.private = vq;
                        cb.trigger = vq->call_ctx.ctx;
                        vq->call_ctx.producer.token = vq->call_ctx.ctx;
-                       vhost_vdpa_setup_vq_irq(v, idx);
+                       if (ops->get_status(vdpa) &
+                           VIRTIO_CONFIG_S_DRIVER_OK)
+                               vhost_vdpa_setup_vq_irq(v, idx);
                } else {
                        cb.callback = NULL;
                        cb.private = NULL;
Quoted 7 lines
Quoted 5 lines
Quoted 5 lines

Probably, but I didn't see this when testing vp_vdpa.

When did you meet those warnings? Is it during the boot or migration?

During boot, on the first 2 VQs only (so before the QPs are resized). Traffic does work though when the VM is booted.

Right.

Quoted 9 lines
Quoted 7 lines

Btw, it would be helpful to check if mlx5_get_vq_irq() works correctly. I believe it should return an error if the virtqueue interrupt is not allocated. After a glance at the code, it seems not straightforward to me.

I think we're good on that front:
mlx5_get_vq_irq() returns EOPNOTSUPP if the vq irq is not allocated.

Good to know that.

Thanks

Quoted 5 lines

Thanks,
Dragos

Tested-by
#6 lore raw
Quoted 117 lines · Jason Wang
Quoted 96 lines · Dragos Tatulea
Quoted 86 lines · Jason Wang
Quoted 85 lines · Jason Wang
Quoted 83 lines · Dragos Tatulea
Quoted 68 lines · Jason Wang

We used to call irq_bypass_unregister_producer() in vhost_vdpa_setup_vq_irq() which is problematic as we don't know if the token pointer is still valid or not.

Actually, we use the eventfd_ctx as the token so the life cycle of the token should be bound to the VHOST_SET_VRING_CALL instead of vhost_vdpa_setup_vq_irq() which could be called by set_status().

Fixing this by setting up irq bypass producer's token when handling VHOST_SET_VRING_CALL and un-registering the producer before calling vhost_vring_ioctl() to prevent a possible use after free as eventfd could have been released in vhost_vring_ioctl().

Fixes 2cf1ba9a4d15 ("vhost_vdpa: implement IRQ offloading in vhost_vdpa")

---
Note for Dragos: Please check whether this fixes your issue. I slightly test it with vp_vdpa in L2.

---
 drivers/vhost/vdpa.c | 12 +++++++++---
 1 file changed, 9 insertions(+), 3 deletions(-)
drivers/vhost/vdpa.c +8 -3
index e31ec9ebc4ce..388226a48bcc 100644
@@ -209,11 +209,9 @@ static void vhost_vdpa_setup_vq_irq(struct vhost_vdpa *v, u16 qid)
      if (irq < 0)
              return;
 
-     irq_bypass_unregister_producer(&vq->call_ctx.producer);
      if (!vq->call_ctx.ctx)
              return;
 
-     vq->call_ctx.producer.token = vq->call_ctx.ctx;
      vq->call_ctx.producer.irq = irq;
      ret = irq_bypass_register_producer(&vq->call_ctx.producer);
      if (unlikely(ret))
@@ -709,6 +707,12 @@ static long vhost_vdpa_vring_ioctl(struct vhost_vdpa *v, unsigned int cmd,
                      vq->last_avail_idx = vq_state.split.avail_index;
              }
              break;
+     case VHOST_SET_VRING_CALL:
+             if (vq->call_ctx.ctx) {
+                     vhost_vdpa_unsetup_vq_irq(v, idx);
+                     vq->call_ctx.producer.token = NULL;
+             }
+             break;
      }
 
      r = vhost_vring_ioctl(&v->vdev, cmd, argp);
@@ -747,13 +751,14 @@ static long vhost_vdpa_vring_ioctl(struct vhost_vdpa *v, unsigned int cmd,
                      cb.callback = vhost_vdpa_virtqueue_cb;
                      cb.private = vq;
                      cb.trigger = vq->call_ctx.ctx;
+                     vq->call_ctx.producer.token = vq->call_ctx.ctx;
+                     vhost_vdpa_setup_vq_irq(v, idx);
              } else {
                      cb.callback = NULL;
                      cb.private = NULL;
                      cb.trigger = NULL;
              }
              ops->set_vq_cb(vdpa, idx, &cb);
-             vhost_vdpa_setup_vq_irq(v, idx);
              break;
 
      case VHOST_SET_VRING_NUM:

drivers/vhost/vdpa.c static int vhost_vdpa_open(struct inode *inode, struct file *filep) · line 1430

Quoted 7 lines
drivers/vhost/vdpa.c +1 -0
@@ -1419,6 +1424,7 @@ static int vhost_vdpa_open(struct inode *inode, struct file *filep)
      for (i = 0; i < nvqs; i++) {
              vqs[i] = &v->vqs[i];
              vqs[i]->handle_kick = handle_vq_kick;
+             vqs[i]->call_ctx.ctx = NULL;
      }
      vhost_dev_init(dev, vqs, nvqs, 0, 0, 0, false,
                     vhost_vdpa_process_iotlb_msg);

No more crashes, but now getting a lot of:
vhost-vdpa-X: vq Y, irq bypass producer (token 00000000a66e28ab) registration fails, ret = -16

Quoted 3 lines
Quoted 3 lines
Quoted 2 lines

... seems like the irq_bypass_unregister_producer() that was removed might still be needed somewhere?

My statement above was not quite correct. The error comes from the VQ irq being registered twice:

Quoted 8 lines

1) VHOST_SET_VRING_CALL ioctl gets called for vq 0. VQ irq is unregistered (vhost_vdpa_unsetup_vq_irq() and re-registered (vhost_vdpa_setup_vq_irq()) successfully. So far so good.

2) set status !DRIVER_OK -> DRIVER_OK happens. VQ irq setup is done once again (vhost_vdpa_setup_vq_irq()). As the producer unregister was removed in this patch, the register will complain because the producer token already exists.

I see. I think it's probably too tricky to try to register/unregister a producer during set_vring_call if DRIVER_OK is not set.

Does it work if we only do vhost_vdpa_unsetup/setup_vq_irq() if DRIVER_OK is set in vhost_vdpa_vring_ioctl() (on top of this patch)?

drivers/vhost/vdpa.c +0 -0
index 388226a48bcc..ab441b8ccd2e 100644

drivers/vhost/vdpa.c static long vhost_vdpa_vring_ioctl(struct · line 716

Quoted 9 lines
drivers/vhost/vdpa.c +3 -1
@@ -709,7 +709,9 @@ static long vhost_vdpa_vring_ioctl(struct
 vhost_vdpa *v, unsigned int cmd,
                break;
        case VHOST_SET_VRING_CALL:
                if (vq->call_ctx.ctx) {
-                       vhost_vdpa_unsetup_vq_irq(v, idx);
+                       if (ops->get_status(vdpa) &
+                           VIRTIO_CONFIG_S_DRIVER_OK)
+                               vhost_vdpa_unsetup_vq_irq(v, idx);
                        vq->call_ctx.producer.token = NULL;

I was wondering if it's safe to set NULL also for !DRIVER_OK case, but then I noticed that the !DRIVER_OK transition doesn't set .token to NULL so this is actually beneficial. Did I get it right?

Quoted 13 lines
drivers/vhost/vdpa.c +0 -0
                }
                break;
drivers/vhost/vdpa.c +3 -1
@@ -752,7 +754,9 @@ static long vhost_vdpa_vring_ioctl(struct
 vhost_vdpa *v, unsigned int cmd,
                        cb.private = vq;
                        cb.trigger = vq->call_ctx.ctx;
                        vq->call_ctx.producer.token = vq->call_ctx.ctx;
-                       vhost_vdpa_setup_vq_irq(v, idx);
+                       if (ops->get_status(vdpa) &
+                           VIRTIO_CONFIG_S_DRIVER_OK)
+                               vhost_vdpa_setup_vq_irq(v, idx);
                } else {
                        cb.callback = NULL;
Quoted 2 lines
                        cb.private = NULL;

Yup, this works.

I do understand the fix, but I don't fully understand why this is better than setting the .token to NULL in vhost_vdpa_unsetup_vq_irq() and keeping the token logic inside the vhost_vdpa_setup/unsetup_vq_irq() API.

In any case, if you send this fix:

Thanks,
Dragos

Quoted 30 lines
Quoted 7 lines
Quoted 5 lines
Quoted 5 lines

Probably, but I didn't see this when testing vp_vdpa.

When did you meet those warnings? Is it during the boot or migration?

During boot, on the first 2 VQs only (so before the QPs are resized). Traffic does work though when the VM is booted.

Right.

Quoted 9 lines
Quoted 7 lines

Btw, it would be helpful to check if mlx5_get_vq_irq() works correctly. I believe it should return an error if the virtqueue interrupt is not allocated. After a glance at the code, it seems not straightforward to me.

I think we're good on that front:
mlx5_get_vq_irq() returns EOPNOTSUPP if the vq irq is not allocated.

Good to know that.

Thanks

Quoted 5 lines

Thanks,
Dragos

author
#7 lore raw
Quoted 125 lines · Dragos Tatulea
Quoted 121 lines · Jason Wang
Quoted 96 lines · Dragos Tatulea
Quoted 86 lines · Jason Wang
Quoted 85 lines · Jason Wang
Quoted 83 lines · Dragos Tatulea
Quoted 68 lines · Jason Wang

We used to call irq_bypass_unregister_producer() in vhost_vdpa_setup_vq_irq() which is problematic as we don't know if the token pointer is still valid or not.

Actually, we use the eventfd_ctx as the token so the life cycle of the token should be bound to the VHOST_SET_VRING_CALL instead of vhost_vdpa_setup_vq_irq() which could be called by set_status().

Fixing this by setting up irq bypass producer's token when handling VHOST_SET_VRING_CALL and un-registering the producer before calling vhost_vring_ioctl() to prevent a possible use after free as eventfd could have been released in vhost_vring_ioctl().

Fixes 2cf1ba9a4d15 ("vhost_vdpa: implement IRQ offloading in vhost_vdpa")

---
Note for Dragos: Please check whether this fixes your issue. I slightly test it with vp_vdpa in L2.

---
 drivers/vhost/vdpa.c | 12 +++++++++---
 1 file changed, 9 insertions(+), 3 deletions(-)
drivers/vhost/vdpa.c +8 -3
index e31ec9ebc4ce..388226a48bcc 100644
@@ -209,11 +209,9 @@ static void vhost_vdpa_setup_vq_irq(struct vhost_vdpa *v, u16 qid)
      if (irq < 0)
              return;
 
-     irq_bypass_unregister_producer(&vq->call_ctx.producer);
      if (!vq->call_ctx.ctx)
              return;
 
-     vq->call_ctx.producer.token = vq->call_ctx.ctx;
      vq->call_ctx.producer.irq = irq;
      ret = irq_bypass_register_producer(&vq->call_ctx.producer);
      if (unlikely(ret))
@@ -709,6 +707,12 @@ static long vhost_vdpa_vring_ioctl(struct vhost_vdpa *v, unsigned int cmd,
                      vq->last_avail_idx = vq_state.split.avail_index;
              }
              break;
+     case VHOST_SET_VRING_CALL:
+             if (vq->call_ctx.ctx) {
+                     vhost_vdpa_unsetup_vq_irq(v, idx);
+                     vq->call_ctx.producer.token = NULL;
+             }
+             break;
      }
 
      r = vhost_vring_ioctl(&v->vdev, cmd, argp);
@@ -747,13 +751,14 @@ static long vhost_vdpa_vring_ioctl(struct vhost_vdpa *v, unsigned int cmd,
                      cb.callback = vhost_vdpa_virtqueue_cb;
                      cb.private = vq;
                      cb.trigger = vq->call_ctx.ctx;
+                     vq->call_ctx.producer.token = vq->call_ctx.ctx;
+                     vhost_vdpa_setup_vq_irq(v, idx);
              } else {
                      cb.callback = NULL;
                      cb.private = NULL;
                      cb.trigger = NULL;
              }
              ops->set_vq_cb(vdpa, idx, &cb);
-             vhost_vdpa_setup_vq_irq(v, idx);
              break;
 
      case VHOST_SET_VRING_NUM:

drivers/vhost/vdpa.c static int vhost_vdpa_open(struct inode *inode, struct file *filep) · line 1430

Quoted 7 lines
drivers/vhost/vdpa.c +1 -0
@@ -1419,6 +1424,7 @@ static int vhost_vdpa_open(struct inode *inode, struct file *filep)
      for (i = 0; i < nvqs; i++) {
              vqs[i] = &v->vqs[i];
              vqs[i]->handle_kick = handle_vq_kick;
+             vqs[i]->call_ctx.ctx = NULL;
      }
      vhost_dev_init(dev, vqs, nvqs, 0, 0, 0, false,
                     vhost_vdpa_process_iotlb_msg);

No more crashes, but now getting a lot of:
vhost-vdpa-X: vq Y, irq bypass producer (token 00000000a66e28ab) registration fails, ret = -16

Quoted 3 lines
Quoted 3 lines
Quoted 2 lines

... seems like the irq_bypass_unregister_producer() that was removed might still be needed somewhere?

My statement above was not quite correct. The error comes from the VQ irq being registered twice:

Quoted 8 lines

1) VHOST_SET_VRING_CALL ioctl gets called for vq 0. VQ irq is unregistered (vhost_vdpa_unsetup_vq_irq() and re-registered (vhost_vdpa_setup_vq_irq()) successfully. So far so good.

2) set status !DRIVER_OK -> DRIVER_OK happens. VQ irq setup is done once again (vhost_vdpa_setup_vq_irq()). As the producer unregister was removed in this patch, the register will complain because the producer token already exists.

I see. I think it's probably too tricky to try to register/unregister a producer during set_vring_call if DRIVER_OK is not set.

Does it work if we only do vhost_vdpa_unsetup/setup_vq_irq() if DRIVER_OK is set in vhost_vdpa_vring_ioctl() (on top of this patch)?

drivers/vhost/vdpa.c +0 -0
index 388226a48bcc..ab441b8ccd2e 100644
@@ -709,7 +709,9 @@ static long vhost_vdpa_vring_ioctl(struct
 vhost_vdpa *v, unsigned int cmd,
                break;
        case VHOST_SET_VRING_CALL:
                if (vq->call_ctx.ctx) {
Quoted 8 lines

drivers/vhost/vdpa.c static long vhost_vdpa_vring_ioctl(struct · line 716

Quoted 5 lines
drivers/vhost/vdpa.c +3 -1
-                       vhost_vdpa_unsetup_vq_irq(v, idx);
+                       if (ops->get_status(vdpa) &
+                           VIRTIO_CONFIG_S_DRIVER_OK)
+                               vhost_vdpa_unsetup_vq_irq(v, idx);
                        vq->call_ctx.producer.token = NULL;

I was wondering if it's safe to set NULL also for !DRIVER_OK case, but then I noticed that the !DRIVER_OK transition doesn't set .token to NULL so this is actually beneficial. Did I get it right?

Yes, actually the reason is that we use eventfd as the token, so the life cycle of the token is tied to eventfd itself. If we don't set the token to NULL here the vhost_vring_ioctl() may just release it which may lead to a use-after-free. So this patch doesn't set a token during DRIVER_OK transition but during SET_VRING_CALL.

Quoted 18 lines
Quoted 14 lines
drivers/vhost/vdpa.c +0 -0
                }
                break;
drivers/vhost/vdpa.c +3 -1
@@ -752,7 +754,9 @@ static long vhost_vdpa_vring_ioctl(struct
 vhost_vdpa *v, unsigned int cmd,
                        cb.private = vq;
                        cb.trigger = vq->call_ctx.ctx;
                        vq->call_ctx.producer.token = vq->call_ctx.ctx;
-                       vhost_vdpa_setup_vq_irq(v, idx);
+                       if (ops->get_status(vdpa) &
+                           VIRTIO_CONFIG_S_DRIVER_OK)
+                               vhost_vdpa_setup_vq_irq(v, idx);
                } else {
                        cb.callback = NULL;
Quoted 2 lines
                        cb.private = NULL;

Yup, this works.

Quoted 4 lines

I do understand the fix, but I don't fully understand why this is better than setting the .token to NULL in vhost_vdpa_unsetup_vq_irq() and keeping the token logic inside the vhost_vdpa_setup/unsetup_vq_irq() API.

See the above explanation, hope it clarifies things.

Quoted 3 lines

In any case, if you send this fix:

Thanks

Quoted 34 lines

Thanks,
Dragos

Quoted 30 lines
Quoted 7 lines
Quoted 5 lines
Quoted 5 lines

Probably, but I didn't see this when testing vp_vdpa.

When did you meet those warnings? Is it during the boot or migration?

During boot, on the first 2 VQs only (so before the QPs are resized). Traffic does work though when the VM is booted.

Right.

Quoted 9 lines
Quoted 7 lines

Btw, it would be helpful to check if mlx5_get_vq_irq() works correctly. I believe it should return an error if the virtqueue interrupt is not allocated. After a glance at the code, it seems not straightforward to me.

I think we're good on that front:
mlx5_get_vq_irq() returns EOPNOTSUPP if the vq irq is not allocated.

Good to know that.

Thanks

Quoted 5 lines

Thanks,
Dragos

Reviewed-by
#8 lore raw
Quoted 135 lines · Jason Wang
Quoted 125 lines · Dragos Tatulea
Quoted 121 lines · Jason Wang
Quoted 96 lines · Dragos Tatulea
Quoted 86 lines · Jason Wang
Quoted 85 lines · Jason Wang
Quoted 83 lines · Dragos Tatulea
Quoted 68 lines · Jason Wang

We used to call irq_bypass_unregister_producer() in vhost_vdpa_setup_vq_irq() which is problematic as we don't know if the token pointer is still valid or not.

Actually, we use the eventfd_ctx as the token so the life cycle of the token should be bound to the VHOST_SET_VRING_CALL instead of vhost_vdpa_setup_vq_irq() which could be called by set_status().

Fixing this by setting up irq bypass producer's token when handling VHOST_SET_VRING_CALL and un-registering the producer before calling vhost_vring_ioctl() to prevent a possible use after free as eventfd could have been released in vhost_vring_ioctl().

Fixes 2cf1ba9a4d15 ("vhost_vdpa: implement IRQ offloading in vhost_vdpa")

---
Note for Dragos: Please check whether this fixes your issue. I slightly test it with vp_vdpa in L2.

---
 drivers/vhost/vdpa.c | 12 +++++++++---
 1 file changed, 9 insertions(+), 3 deletions(-)
drivers/vhost/vdpa.c +8 -3
index e31ec9ebc4ce..388226a48bcc 100644
@@ -209,11 +209,9 @@ static void vhost_vdpa_setup_vq_irq(struct vhost_vdpa *v, u16 qid)
      if (irq < 0)
              return;
 
-     irq_bypass_unregister_producer(&vq->call_ctx.producer);
      if (!vq->call_ctx.ctx)
              return;
 
-     vq->call_ctx.producer.token = vq->call_ctx.ctx;
      vq->call_ctx.producer.irq = irq;
      ret = irq_bypass_register_producer(&vq->call_ctx.producer);
      if (unlikely(ret))
@@ -709,6 +707,12 @@ static long vhost_vdpa_vring_ioctl(struct vhost_vdpa *v, unsigned int cmd,
                      vq->last_avail_idx = vq_state.split.avail_index;
              }
              break;
+     case VHOST_SET_VRING_CALL:
+             if (vq->call_ctx.ctx) {
+                     vhost_vdpa_unsetup_vq_irq(v, idx);
+                     vq->call_ctx.producer.token = NULL;
+             }
+             break;
      }
 
      r = vhost_vring_ioctl(&v->vdev, cmd, argp);
@@ -747,13 +751,14 @@ static long vhost_vdpa_vring_ioctl(struct vhost_vdpa *v, unsigned int cmd,
                      cb.callback = vhost_vdpa_virtqueue_cb;
                      cb.private = vq;
                      cb.trigger = vq->call_ctx.ctx;
+                     vq->call_ctx.producer.token = vq->call_ctx.ctx;
+                     vhost_vdpa_setup_vq_irq(v, idx);
              } else {
                      cb.callback = NULL;
                      cb.private = NULL;
                      cb.trigger = NULL;
              }
              ops->set_vq_cb(vdpa, idx, &cb);
-             vhost_vdpa_setup_vq_irq(v, idx);
              break;
 
      case VHOST_SET_VRING_NUM:

drivers/vhost/vdpa.c static int vhost_vdpa_open(struct inode *inode, struct file *filep) · line 1430

Quoted 7 lines
drivers/vhost/vdpa.c +1 -0
@@ -1419,6 +1424,7 @@ static int vhost_vdpa_open(struct inode *inode, struct file *filep)
      for (i = 0; i < nvqs; i++) {
              vqs[i] = &v->vqs[i];
              vqs[i]->handle_kick = handle_vq_kick;
+             vqs[i]->call_ctx.ctx = NULL;
      }
      vhost_dev_init(dev, vqs, nvqs, 0, 0, 0, false,
                     vhost_vdpa_process_iotlb_msg);

No more crashes, but now getting a lot of:
vhost-vdpa-X: vq Y, irq bypass producer (token 00000000a66e28ab) registration fails, ret = -16

Quoted 3 lines
Quoted 3 lines
Quoted 2 lines

... seems like the irq_bypass_unregister_producer() that was removed might still be needed somewhere?

My statement above was not quite correct. The error comes from the VQ irq being registered twice:

Quoted 8 lines

1) VHOST_SET_VRING_CALL ioctl gets called for vq 0. VQ irq is unregistered (vhost_vdpa_unsetup_vq_irq() and re-registered (vhost_vdpa_setup_vq_irq()) successfully. So far so good.

2) set status !DRIVER_OK -> DRIVER_OK happens. VQ irq setup is done once again (vhost_vdpa_setup_vq_irq()). As the producer unregister was removed in this patch, the register will complain because the producer token already exists.

I see. I think it's probably too tricky to try to register/unregister a producer during set_vring_call if DRIVER_OK is not set.

Does it work if we only do vhost_vdpa_unsetup/setup_vq_irq() if DRIVER_OK is set in vhost_vdpa_vring_ioctl() (on top of this patch)?

drivers/vhost/vdpa.c +0 -0
index 388226a48bcc..ab441b8ccd2e 100644
@@ -709,7 +709,9 @@ static long vhost_vdpa_vring_ioctl(struct
 vhost_vdpa *v, unsigned int cmd,
                break;
        case VHOST_SET_VRING_CALL:
                if (vq->call_ctx.ctx) {
Quoted 8 lines

drivers/vhost/vdpa.c static long vhost_vdpa_vring_ioctl(struct · line 716

Quoted 5 lines
drivers/vhost/vdpa.c +3 -1
-                       vhost_vdpa_unsetup_vq_irq(v, idx);
+                       if (ops->get_status(vdpa) &
+                           VIRTIO_CONFIG_S_DRIVER_OK)
+                               vhost_vdpa_unsetup_vq_irq(v, idx);
                        vq->call_ctx.producer.token = NULL;

I was wondering if it's safe to set NULL also for !DRIVER_OK case, but then I noticed that the !DRIVER_OK transition doesn't set .token to NULL so this is actually beneficial. Did I get it right?

Quoted 6 lines

Yes, actually the reason is that we use eventfd as the token, so the life cycle of the token is tied to eventfd itself. If we don't set the token to NULL here the vhost_vring_ioctl() may just release it which may lead to a use-after-free. So this patch doesn't set a token during DRIVER_OK transition but during SET_VRING_CALL.

Ack. Thanks for the explanation.

Quoted 18 lines
Quoted 18 lines
Quoted 14 lines
drivers/vhost/vdpa.c +0 -0
                }
                break;
drivers/vhost/vdpa.c +3 -1
@@ -752,7 +754,9 @@ static long vhost_vdpa_vring_ioctl(struct
 vhost_vdpa *v, unsigned int cmd,
                        cb.private = vq;
                        cb.trigger = vq->call_ctx.ctx;
                        vq->call_ctx.producer.token = vq->call_ctx.ctx;
-                       vhost_vdpa_setup_vq_irq(v, idx);
+                       if (ops->get_status(vdpa) &
+                           VIRTIO_CONFIG_S_DRIVER_OK)
+                               vhost_vdpa_setup_vq_irq(v, idx);
                } else {
                        cb.callback = NULL;
Quoted 2 lines
                        cb.private = NULL;

Yup, this works.

Quoted 7 lines
Quoted 4 lines

I do understand the fix, but I don't fully understand why this is better than setting the .token to NULL in vhost_vdpa_unsetup_vq_irq() and keeping the token logic inside the vhost_vdpa_setup/unsetup_vq_irq() API.

See the above explanation, hope it clarifies things.

Yes it does.

Quoted 3 lines
Quoted 3 lines

In any case, if you send this fix:

FWIW:

Quoted 38 lines

Thanks

Quoted 34 lines

Thanks,
Dragos

Quoted 30 lines
Quoted 7 lines
Quoted 5 lines
Quoted 5 lines

Probably, but I didn't see this when testing vp_vdpa.

When did you meet those warnings? Is it during the boot or migration?

During boot, on the first 2 VQs only (so before the QPs are resized). Traffic does work though when the VM is booted.

Right.

Quoted 9 lines
Quoted 7 lines

Btw, it would be helpful to check if mlx5_get_vq_irq() works correctly. I believe it should return an error if the virtqueue interrupt is not allocated. After a glance at the code, it seems not straightforward to me.

I think we're good on that front:
mlx5_get_vq_irq() returns EOPNOTSUPP if the vq irq is not allocated.

Good to know that.

Thanks

Quoted 5 lines

Thanks,
Dragos

#9 lore raw
In reply to Jason Wang
Quoted 8 lines · Jason Wang

We used to call irq_bypass_unregister_producer() in vhost_vdpa_setup_vq_irq() which is problematic as we don't know if the token pointer is still valid or not.

Actually, we use the eventfd_ctx as the token so the life cycle of the token should be bound to the VHOST_SET_VRING_CALL instead of vhost_vdpa_setup_vq_irq() which could be called by set_status().

Quoted 7 lines

Fixing this by setting up irq bypass producer's token when handling VHOST_SET_VRING_CALL and un-registering the producer before calling vhost_vring_ioctl() to prevent a possible use after free as eventfd could have been released in vhost_vring_ioctl().

Fixes 2cf1ba9a4d15 ("vhost_vdpa: implement IRQ offloading in vhost_vdpa")

Want to post a non-RFC version?

Quoted 62 lines

---
Note for Dragos: Please check whether this fixes your issue. I slightly test it with vp_vdpa in L2.

---
 drivers/vhost/vdpa.c | 12 +++++++++---
 1 file changed, 9 insertions(+), 3 deletions(-)
drivers/vhost/vdpa.c +9 -3
index e31ec9ebc4ce..388226a48bcc 100644
@@ -209,11 +209,9 @@ static void vhost_vdpa_setup_vq_irq(struct vhost_vdpa *v, u16 qid)
 	if (irq < 0)
 		return;
 
-	irq_bypass_unregister_producer(&vq->call_ctx.producer);
 	if (!vq->call_ctx.ctx)
 		return;
 
-	vq->call_ctx.producer.token = vq->call_ctx.ctx;
 	vq->call_ctx.producer.irq = irq;
 	ret = irq_bypass_register_producer(&vq->call_ctx.producer);
 	if (unlikely(ret))
@@ -709,6 +707,12 @@ static long vhost_vdpa_vring_ioctl(struct vhost_vdpa *v, unsigned int cmd,
 			vq->last_avail_idx = vq_state.split.avail_index;
 		}
 		break;
+	case VHOST_SET_VRING_CALL:
+		if (vq->call_ctx.ctx) {
+			vhost_vdpa_unsetup_vq_irq(v, idx);
+			vq->call_ctx.producer.token = NULL;
+		}
+		break;
 	}
 
 	r = vhost_vring_ioctl(&v->vdev, cmd, argp);
@@ -747,13 +751,14 @@ static long vhost_vdpa_vring_ioctl(struct vhost_vdpa *v, unsigned int cmd,
 			cb.callback = vhost_vdpa_virtqueue_cb;
 			cb.private = vq;
 			cb.trigger = vq->call_ctx.ctx;
+			vq->call_ctx.producer.token = vq->call_ctx.ctx;
+			vhost_vdpa_setup_vq_irq(v, idx);
 		} else {
 			cb.callback = NULL;
 			cb.private = NULL;
 			cb.trigger = NULL;
 		}
 		ops->set_vq_cb(vdpa, idx, &cb);
-		vhost_vdpa_setup_vq_irq(v, idx);
 		break;
 
 	case VHOST_SET_VRING_NUM:
@@ -1419,6 +1424,7 @@ static int vhost_vdpa_open(struct inode *inode, struct file *filep)
 	for (i = 0; i < nvqs; i++) {
 		vqs[i] = &v->vqs[i];
 		vqs[i]->handle_kick = handle_vq_kick;
+		vqs[i]->call_ctx.ctx = NULL;
 	}
 	vhost_dev_init(dev, vqs, nvqs, 0, 0, 0, false,
 		       vhost_vdpa_process_iotlb_msg);
Signature · 1 line
2.31.1