mbox series

[v14,00/10] implement packed virtqueues

Message ID 20181217213139.19299-1-jfreimann@redhat.com (mailing list archive)
Headers
Series implement packed virtqueues |

Message

Jens Freimann Dec. 17, 2018, 9:31 p.m. UTC
  This is a basic implementation of packed virtqueues as specified in the
Virtio 1.1 draft. A compiled version of the current draft is available
at https://github.com/oasis-tcs/virtio-docs.git (or as .pdf at
https://github.com/oasis-tcs/virtio-docs/blob/master/virtio-v1.1-packed-wd10.pdf

A packed virtqueue is different from a split virtqueue in that it
consists of only a single descriptor ring that replaces available and
used ring, index and descriptor pointers.

Each descriptor is readable and writable and has a flags field. These flags
will mark if a descriptor is available or used.  To detect new available descriptors
even after the ring has wrapped, device and driver each have a
single-bit wrap counter that is flipped from 0 to 1 and vice versa every time
the last descriptor in the ring is used/made available.

With this patch set I see performance equal or slightly better (+2-3%) in a PVP
scenario compared to split virtqueues (v18.11 in host) 

regards,
Jens 

v13->v14:
 * replace (_)set_desc_avail()
 * remove unrelated changes in split vq code
 * remove trailing line
 * remove RING_EVENT_IDX code from virtqueue_enable_intr_packed()
 * set interrupt disable flag unconditionally
 * change error message for disabled control virtqueues

v12->v13:
 * re-order patches 1-3 to address Maximes comment. Move all defines and
   data structures to patch one, all helpers to patch 2.  
 * build-tested and ran checkpatch on all patches again
 * add split/packed versions for virtio_enable_intr()
 * remove redundant changes from split vq code
 * just return -1 when cq is enabled for packed vqs in virtio-user 

v11->v12:
 * add a patch to disable control vq when packed vq is enabled.
   I have a patch to support this but it needs a bit more work
   and I think it shouldn't stop this series from being applied
 * rework mergeable receive buffer code to be more efficient, by
   batching descriptor refill, similar to what Maxime proposed for
   split virtqueues
 * removed unnecessary checks in virtio_recv_mergeable_pkts_packed
   (Maxime) 
 * Did not merge receive functions as Maxime suggested because it seemed
   to cause a small performance regression
 * Move event_flags_shadow from patch 3 to 1 (Maxime)
 * Did not merge xmit functions and call _split/_packed functions from
   there because it seemed to cause small performance drop (-0.5%)
   (Maxime)

v10-v11:
 * this version includes some fixes from Tiwei, so I added his
   Signed-off-by to some of the patches
 * fix hang with mergable rx buffers (Tiwei)
 * clean-up code and simplify buffer handling (Tiwei)
 * rebase to current virtio-next master branch

v9-v10:
 * don't mix index into buffer list and descriptors
 * whitespace and formatting issues
 * remove "VQ:" in dump virtqueue patch
 * add extra packed vring struct to virtqueue and change function
   prototypes and code accordingly
 * move wrap_counters to virtqueue
 * make if-conditions for packed and !packed more clear in
   set_rxtx_funcs()
 * initialize wrap counters in first patch, instead of rx and tx
   implementation patch
 * make virtio-user not supported with packed virtqueues, to
   be fixed in other patch set?

v8-v9:
 * fix virtio_ring_free_chain_packed() to handle descriptors
   correctly in case of out-of-order
 * fix check in virtqueue_xmit_cleanup_packed() to improve performance

v7->v8:
 * move desc_is_used change to correct patch
 * remove trailing newline
 * correct xmit code, flags update and memory barrier
 * move packed desc init to dedicated function, split
   and packed variant



Jens Freimann (9):
  net/virtio: add packed virtqueue defines
  net/virtio: add packed virtqueue helpers
  net/virtio: vring init for packed queues
  net/virtio: dump packed virtqueue data
  net/virtio: implement transmit path for packed queues
  net/virtio: implement receive path for packed queues
  net/virtio: add virtio send command packed queue support
  net/virtio-user: fail if cq used with packed vq
  net/virtio: enable packed virtqueues by default

Yuanhan Liu (1):
  net/virtio-user: add option to use packed queues

 drivers/net/virtio/virtio_ethdev.c            | 228 +++++--
 drivers/net/virtio/virtio_ethdev.h            |   8 +
 drivers/net/virtio/virtio_pci.h               |   7 +
 drivers/net/virtio/virtio_ring.h              |  58 +-
 drivers/net/virtio/virtio_rxtx.c              | 611 +++++++++++++++++-
 .../net/virtio/virtio_user/virtio_user_dev.c  |  27 +-
 .../net/virtio/virtio_user/virtio_user_dev.h  |   2 +-
 drivers/net/virtio/virtio_user_ethdev.c       |  14 +-
 drivers/net/virtio/virtqueue.c                |  43 +-
 drivers/net/virtio/virtqueue.h                | 109 +++-
 10 files changed, 1044 insertions(+), 63 deletions(-)
  

Comments

Maxime Coquelin Dec. 18, 2018, 2:02 p.m. UTC | #1
On 12/17/18 10:31 PM, Jens Freimann wrote:
> This is a basic implementation of packed virtqueues as specified in the
> Virtio 1.1 draft. A compiled version of the current draft is available
> at https://github.com/oasis-tcs/virtio-docs.git (or as .pdf at
> https://github.com/oasis-tcs/virtio-docs/blob/master/virtio-v1.1-packed-wd10.pdf
> 
> A packed virtqueue is different from a split virtqueue in that it
> consists of only a single descriptor ring that replaces available and
> used ring, index and descriptor pointers.
> 
> Each descriptor is readable and writable and has a flags field. These flags
> will mark if a descriptor is available or used.  To detect new available descriptors
> even after the ring has wrapped, device and driver each have a
> single-bit wrap counter that is flipped from 0 to 1 and vice versa every time
> the last descriptor in the ring is used/made available.
> 
> With this patch set I see performance equal or slightly better (+2-3%) in a PVP
> scenario compared to split virtqueues (v18.11 in host)
> 
> regards,
> Jens
> 
> v13->v14:
>   * replace (_)set_desc_avail()
>   * remove unrelated changes in split vq code
>   * remove trailing line
>   * remove RING_EVENT_IDX code from virtqueue_enable_intr_packed()
>   * set interrupt disable flag unconditionally
>   * change error message for disabled control virtqueues
> 
> v12->v13:
>   * re-order patches 1-3 to address Maximes comment. Move all defines and
>     data structures to patch one, all helpers to patch 2.
>   * build-tested and ran checkpatch on all patches again
>   * add split/packed versions for virtio_enable_intr()
>   * remove redundant changes from split vq code
>   * just return -1 when cq is enabled for packed vqs in virtio-user
> 
> v11->v12:
>   * add a patch to disable control vq when packed vq is enabled.
>     I have a patch to support this but it needs a bit more work
>     and I think it shouldn't stop this series from being applied
>   * rework mergeable receive buffer code to be more efficient, by
>     batching descriptor refill, similar to what Maxime proposed for
>     split virtqueues
>   * removed unnecessary checks in virtio_recv_mergeable_pkts_packed
>     (Maxime)
>   * Did not merge receive functions as Maxime suggested because it seemed
>     to cause a small performance regression
>   * Move event_flags_shadow from patch 3 to 1 (Maxime)
>   * Did not merge xmit functions and call _split/_packed functions from
>     there because it seemed to cause small performance drop (-0.5%)
>     (Maxime)
> 
> v10-v11:
>   * this version includes some fixes from Tiwei, so I added his
>     Signed-off-by to some of the patches
>   * fix hang with mergable rx buffers (Tiwei)
>   * clean-up code and simplify buffer handling (Tiwei)
>   * rebase to current virtio-next master branch
> 
> v9-v10:
>   * don't mix index into buffer list and descriptors
>   * whitespace and formatting issues
>   * remove "VQ:" in dump virtqueue patch
>   * add extra packed vring struct to virtqueue and change function
>     prototypes and code accordingly
>   * move wrap_counters to virtqueue
>   * make if-conditions for packed and !packed more clear in
>     set_rxtx_funcs()
>   * initialize wrap counters in first patch, instead of rx and tx
>     implementation patch
>   * make virtio-user not supported with packed virtqueues, to
>     be fixed in other patch set?
> 
> v8-v9:
>   * fix virtio_ring_free_chain_packed() to handle descriptors
>     correctly in case of out-of-order
>   * fix check in virtqueue_xmit_cleanup_packed() to improve performance
> 
> v7->v8:
>   * move desc_is_used change to correct patch
>   * remove trailing newline
>   * correct xmit code, flags update and memory barrier
>   * move packed desc init to dedicated function, split
>     and packed variant
> 
> 
> 
> Jens Freimann (9):
>    net/virtio: add packed virtqueue defines
>    net/virtio: add packed virtqueue helpers
>    net/virtio: vring init for packed queues
>    net/virtio: dump packed virtqueue data
>    net/virtio: implement transmit path for packed queues
>    net/virtio: implement receive path for packed queues
>    net/virtio: add virtio send command packed queue support
>    net/virtio-user: fail if cq used with packed vq
>    net/virtio: enable packed virtqueues by default
> 
> Yuanhan Liu (1):
>    net/virtio-user: add option to use packed queues
> 
>   drivers/net/virtio/virtio_ethdev.c            | 228 +++++--
>   drivers/net/virtio/virtio_ethdev.h            |   8 +
>   drivers/net/virtio/virtio_pci.h               |   7 +
>   drivers/net/virtio/virtio_ring.h              |  58 +-
>   drivers/net/virtio/virtio_rxtx.c              | 611 +++++++++++++++++-
>   .../net/virtio/virtio_user/virtio_user_dev.c  |  27 +-
>   .../net/virtio/virtio_user/virtio_user_dev.h  |   2 +-
>   drivers/net/virtio/virtio_user_ethdev.c       |  14 +-
>   drivers/net/virtio/virtqueue.c                |  43 +-
>   drivers/net/virtio/virtqueue.h                | 109 +++-
>   10 files changed, 1044 insertions(+), 63 deletions(-)
> 


Applied to dpdk-next-virtio

Thanks,
Maxime