Add back-end's internal state migration support

[germag]

Summary of the PR

A new set of messages have been introduced in the vhost-user protocol to support the migration of devices with back-end's internal state[0].

These are the additions to the protocol:

• New vhost-user protocol feature VHOST_USER_PROTOCOL_F_DEVICE_STATE
• SET_DEVICE_STATE_FD function: Front-end and back-end negotiate a file descriptor over which to transfer the state. It also includes establishing the direction of transfer and migration phase.
• CHECK_DEVICE_STATE: After the state has been transferred through the file descriptor, the front-end invokes this function to verify success. There is no in-band way (through the file descriptor) to indicate failure, so we need to check explicitly.

[0] https://github.com/qemu/qemu/commit/019233096c03b826e0e677115b6e3c550a54a48d

This code is part of the effort to support live migration in virtiofsd and is currently being used in a version of virtiofsd that supports migration: https://gitlab.com/virtiofsd-live-migration

Note to reviewers

~Please do not merge this PR yet, this PR depends on https://github.com/rust-vmm/vhost/pull/202, because by adding FrontendReq::SET_DEVICE_STATE_FD a "hole" was created in value 41, this makes the APi unsound, since VhostUserMsgHeader::get_code() call transmute_copy() which assumes that the values are continuous and have no "holes" between them.~

~I'll rebase and adjust the code after PR 202 gets merged.~

[germag]

/cc @XanClic

[XanClic]

The compiled version of the documentation linked in patch 1 is here: https://qemu-project.gitlab.io/qemu/interop/vhost-user.html

[XanClic]

Looks good to me, thanks!

[stefano-garzarella]

@XanClic thanks!

Looks good also to me!

[germag]

v2:

• Add comments in message definition

[germag]

Rebase