[dpdk-dev] [PATCH 0/3] Extend vhost to support VFIO based accelerator

Tiwei Bie tiwei.bie at intel.com
Tue Mar 6 11:43:24 CET 2018


This patch set introduces the VFIO based accelerator support
for vhost. This is a new vhost user protocol feature to better
support the vDPA device at the vhost user backend. It allows
interrupts/notifications being delivered between the driver
in guest and the device in host directly.

Dependencies:

1. This patch set depends on the below patch set for QEMU:

http://lists.nongnu.org/archive/html/qemu-devel/2018-01/msg06028.html

Some of the enum definitions in this patch set have been
updated for the latest QEMU. A new patch set for QEMU will
be sent out later.

2. This patch set depends on Zhihong's "selective datapath"
   patch set:

http://dpdk.org/ml/archives/dev/2018-March/091858.html

This patch set is generated on the latest master branch of
dpdk-next-virtio with Zhihong's patches applied.

Best regards,
Tiwei Bie

Tiwei Bie (3):
  vhost: do not generate signal when sendmsg fails
  vhost: support sending fds via send_vhost_message()
  vhost: support VFIO based accelerator

 lib/librte_vhost/rte_vhost.h           |  28 ++++++
 lib/librte_vhost/rte_vhost_version.map |   1 +
 lib/librte_vhost/socket.c              |   2 +-
 lib/librte_vhost/vhost_user.c          | 174 ++++++++++++++++++++++++++++++++-
 lib/librte_vhost/vhost_user.h          |   9 ++
 5 files changed, 209 insertions(+), 5 deletions(-)

-- 
2.11.0



More information about the dev mailing list