Re: [Qemu-devel] [PATCH v5 0/6] vhost-user-blk: Add support for backend reconnecting

[Jason Wang]

On 2019/1/22 下午4:31, address@hidden wrote:
> From: Xie Yongji <address@hidden>
>
> This patchset is aimed at supporting qemu to reconnect
> vhost-user-blk backend after vhost-user-blk backend crash or
> restart.
>
> The patch 1 introduces two new messages VHOST_USER_GET_INFLIGHT_FD
> and VHOST_USER_SET_INFLIGHT_FD to support support transferring shared
> buffer between qemu and backend.
>
> The patch 2,3 are the corresponding libvhost-user patches of
> patch 1. Make libvhost-user support VHOST_USER_GET_INFLIGHT_FD
> and VHOST_USER_SET_INFLIGHT_FD.
>
> The patch 4 allows vhost-user-blk to use the two new messages
> to get/set inflight buffer from/to backend.
>
> The patch 5 supports vhost-user-blk to reconnect backend when
> connection closed.
>
> The patch 6 introduces VHOST_USER_PROTOCOL_F_SLAVE_SHMFD
> to vhost-user-blk backend which is used to tell qemu that
> we support reconnecting now.
>
> This series is based on Daniel P. Berrangé's patchset:
>
> https://lists.gnu.org/archive/html/qemu-devel/2019-01/msg03344.html
>
> To use it, we could start qemu with:
>
> qemu-system-x86_64 \
>          -chardev socket,id=char0,path=/path/vhost.socket,reconnect=1, \
>          -device vhost-user-blk-pci,chardev=char0 \
>
> and start vhost-user-blk backend with:
>
> vhost-user-blk -b /path/file -s /path/vhost.socket
>
> Then we can restart vhost-user-blk at any time during VM running.
>
> V4 to V5:
> - Drop patch that enables "nowait" option on client sockets
> - Support resubmitting inflight I/O in order
> - Make inflight I/O tracking more robust
> - Remove align field and add queue size field in VhostUserInflight
> - Document more details in vhost-user.txt


I'm still not convinced about this approach. If the maintainer decide to 
merge, at least two things needs to be added besides the correctness of 
the code:

- you need prove that this approach can work for packed ring

-  an unit-test to test the crash during logging in-flight descriptor.

Thanks


> V3 to V4:
> - Drop messages VHOST_USER_GET_SHM_SIZE and VHOST_USER_SET_SHM_FD
> - Introduce two new messages VHOST_USER_GET_INFLIGHT_FD
>    and VHOST_USER_SET_INFLIGHT_FD
> - Allocate inflight buffer in backend rather than in qemu
> - Document a recommended format for inflight buffer
>
> V2 to V3:
> - Using exisiting wait/nowait options to control connection on
>    client sockets instead of introducing "disconnected" option.
> - Support the case that vhost-user backend restart during initialzation
>    of vhost-user-blk device.
>
> V1 to V2:
> - Introduce "disconnected" option for chardev instead of reuse "wait"
>    option
> - Support the case that QEMU starts before vhost-user backend
> - Drop message VHOST_USER_SET_VRING_INFLIGHT
> - Introduce two new messages VHOST_USER_GET_SHM_SIZE
>    and VHOST_USER_SET_SHM_FD
>
> Xie Yongji (6):
>    vhost-user: Support transferring inflight buffer between qemu and
>      backend
>    libvhost-user: Introduce vu_queue_map_desc()
>    libvhost-user: Support tracking inflight I/O in shared memory
>    vhost-user-blk: Add support to get/set inflight buffer
>    vhost-user-blk: Add support to reconnect backend
>    contrib/vhost-user-blk: enable inflight I/O tracking
>
>   Makefile                                |   2 +-
>   contrib/libvhost-user/libvhost-user.c   | 402 ++++++++++++++++++++----
>   contrib/libvhost-user/libvhost-user.h   |  40 +++
>   contrib/vhost-user-blk/vhost-user-blk.c |   3 +-
>   docs/interop/vhost-user.txt             | 101 ++++++
>   hw/block/vhost-user-blk.c               | 227 ++++++++++---
>   hw/virtio/vhost-user.c                  | 110 +++++++
>   hw/virtio/vhost.c                       | 105 +++++++
>   include/hw/virtio/vhost-backend.h       |  10 +
>   include/hw/virtio/vhost-user-blk.h      |   5 +
>   include/hw/virtio/vhost.h               |  19 ++
>   11 files changed, 925 insertions(+), 99 deletions(-)