Hi Eelco,

On 5/17/23 11:08, Eelco Chaudron wrote:
This change will allow the vhost interrupt datapath handling to be split
between two processed without one of them holding an explicit lock.


As I had a tuned PVP benchmarking setup at hand, I ran a 0.02% loss
RFC2544 test with and without this patch to ensure moving to RX locks
would not introduce performance regression.

I can confirm there are no performance regression introduced with this
patch applied:
Tested-by: Maxime Coquelin <maxime.coque...@redhat.com>

And the patch looks good to me:
Reviewed-by: Maxime Coquelin <maxime.coque...@redhat.com>

Thanks,
Maxime


Signed-off-by: Eelco Chaudron <echau...@redhat.com>
---
  lib/eal/include/generic/rte_rwlock.h |   17 ++++++
  lib/vhost/vhost.c                    |   46 +++++++++--------
  lib/vhost/vhost.h                    |    4 +-
  lib/vhost/vhost_user.c               |   14 +++--
  lib/vhost/virtio_net.c               |   90 +++++++++++++++++-----------------
  5 files changed, 94 insertions(+), 77 deletions(-)


Reply via email to