To protect against CPU over utilization when guest performs unneeded
busy waiting loop on an empty CQ.
Signed-off-by: Yuval Shaia
---
hw/rdma/rdma_backend.c | 11 +++
hw/rdma/rdma_backend.h | 2 +-
hw/rdma/rdma_rm.c | 1 +
hw/rdma/rdma_rm_defs.h | 6 +-
hw
Hi Yuval,
On 1/31/19 3:08 PM, Yuval Shaia wrote:
To protect against CPU over utilization when guest performs unneeded
busy waiting loop on an empty CQ.
Signed-off-by: Yuval Shaia
---
hw/rdma/rdma_backend.c | 11 +++
hw/rdma/rdma_backend.h | 2 +-
hw/rdma/rdma_rm.c
On Wed, Feb 06, 2019 at 12:14:24PM +0200, Marcel Apfelbaum wrote:
> Hi Yuval,
>
> On 1/31/19 3:08 PM, Yuval Shaia wrote:
> > To protect against CPU over utilization when guest performs unneeded
> > busy waiting loop on an empty CQ.
> >
> > Signed-off-by: Yuval Shaia
> > ---
> > hw/rdma/rdma_ba
> > @@ -60,6 +60,8 @@ static int pvrdma_post_cqe(PVRDMADev *dev, uint32_t
> > cq_handle,
> > return -EINVAL;
> > }
> > +atomic_dec(&cq->missing_cqe);
> > +
>
> We should set it to 0 here? (If we are counting cq-empty hits)
No, this counter just count the number of missing CQE