On Mon, Nov 26, 2018 at 10:17:45AM -0700, Keith Busch wrote:
> The code had asserted an interrupt every time it was requested to check
> for new completion queue entries.This can result in spurious interrupts
> seen by the guest OS.
>
> Fix this by asserting an interrupt only if there are un-acknowledged
> completion queue entries available.
>
> Reported-by: Guenter Roeck <linux@roeck-us.net>
> Signed-off-by: Keith Busch <keith.busch@intel.com>
Tested-by: Guenter Roeck <linux@roeck-us.net>
> ---
> hw/block/nvme.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/hw/block/nvme.c b/hw/block/nvme.c
> index 9fbe5673cb..7c8c63e8f5 100644
> --- a/hw/block/nvme.c
> +++ b/hw/block/nvme.c
> @@ -272,7 +272,9 @@ static void nvme_post_cqes(void *opaque)
> sizeof(req->cqe));
> QTAILQ_INSERT_TAIL(&sq->req_list, req, entry);
> }
> - nvme_irq_assert(n, cq);
> + if (cq->tail != cq->head) {
> + nvme_irq_assert(n, cq);
> + }
> }
>
> static void nvme_enqueue_req_completion(NvmeCQueue *cq, NvmeRequest *req)
> --
> 2.14.4
>