[PATCH v6 2/3] vhost/vsock: discard IOTLB when ACCESS_PLATFORM is cleared
COLD45d
From: Jia Jia <hidden>
Date: 2026-08-18 04:26:53
Also in:
kvm, lkml, netdev
Subsystem:
the rest, virtio and vhost vsock driver, virtio host (vhost) · Maintainers:
Linus Torvalds, Stefan Hajnoczi, Stefano Garzarella, "Michael S. Tsirkin", Jason Wang, Eugenio Pérez
Revision v6 of 3 in this series.
Revisions (3)
- v5 [diff vs current]
- v6 current
- v7 [diff vs current]
Clear the device IOTLB when userspace clears VIRTIO_F_ACCESS_PLATFORM.
Otherwise descriptor translation can continue to use mappings installed
before the feature change.
The common helper invalidates cached vring access and applies the
transition even while a backend is attached. The backend remains attached,
but userspace must configure the vring addresses for the new address mode
after a successful live transition.
Fixes: e13a6915a03f ("vhost/vsock: add IOTLB API support")
Suggested-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Jia Jia <redacted>
---
drivers/vhost/vsock.c | 10 +++++++---
1 file changed, 7 insertions(+), 3 deletions(-)
diff --git a/drivers/vhost/vsock.c b/drivers/vhost/vsock.c
index 9aaab6bb8061..77a5a6775c86 100644
--- a/drivers/vhost/vsock.c
+++ b/drivers/vhost/vsock.c@@ -854,6 +854,7 @@ static int vhost_vsock_set_cid(struct vhost_vsock *vsock, u64 guest_cid) static int vhost_vsock_set_features(struct vhost_vsock *vsock, u64 features) { struct vhost_virtqueue *vq; + int ret = -EFAULT; int i; if (features & ~VHOST_VSOCK_FEATURES)
@@ -865,9 +866,12 @@ static int vhost_vsock_set_features(struct vhost_vsock *vsock, u64 features) goto err; } - if ((features & (1ULL << VIRTIO_F_ACCESS_PLATFORM))) { - if (vhost_init_device_iotlb(&vsock->dev)) + if (features & (1ULL << VIRTIO_F_ACCESS_PLATFORM)) { + ret = vhost_init_device_iotlb(&vsock->dev); + if (ret) goto err; + } else { + vhost_clear_device_iotlb(&vsock->dev); } vsock->seqpacket_allow = features & (1ULL << VIRTIO_VSOCK_F_SEQPACKET);
@@ -883,7 +887,7 @@ static int vhost_vsock_set_features(struct vhost_vsock *vsock, u64 features) err: mutex_unlock(&vsock->dev.mutex); - return -EFAULT; + return ret; } static long vhost_vsock_dev_ioctl(struct file *f, unsigned int ioctl,