Hi, Patch 1 of this RFC series adds virtio-fs-specific operations to vhost for transferring a binary blob of back-end-internal state, and implements those for vhost-user.
Patch 2 uses those operations to implement stateful migration for vhost-user-fs devices, assuming the back-end (virtiofsd) supports it. This is an RFC for multiple reasons, most notably: - Patch 1 proposes yet undiscussed changes to the vhost protocol, which makes it RFC by default. - Without much experience in the fields of migration or vhost (on the qemu side), I hope marking this as an RFC leads to extra scrutiny on the reviewerâs side. O:) Hanna Czenczek (2): vhost-user: Add interface for virtio-fs migration vhost-user-fs: Implement stateful migration include/hw/virtio/vhost-backend.h | 9 ++ include/hw/virtio/vhost.h | 68 ++++++++++++ hw/virtio/vhost-user-fs.c | 171 +++++++++++++++++++++++++++++- hw/virtio/vhost-user.c | 138 ++++++++++++++++++++++++ hw/virtio/vhost.c | 29 +++++ 5 files changed, 414 insertions(+), 1 deletion(-) -- 2.39.1
