On Aug 25 15:47, Jinhao Fan wrote:
> When the new option 'irq-eventfd' is turned on, the IO emulation code
> signals an eventfd when it want to (de)assert an irq. The main loop
> eventfd handler does the actual irq (de)assertion. This paves the way
> for iothread support since QEMU's interrupt emulation is not thread
> safe.
>
> Asserting and deasseting irq with eventfd has some performance
> implications. For small queue depth it increases request latency but
> for large queue depth it effectively coalesces irqs.
>
> Comparision (KIOPS):
>
> QD 1 4 16 64
> QEMU 38 123 210 329
> irq-eventfd 32 106 240 364
>
> Signed-off-by: Jinhao Fan <fanjinhao...@ict.ac.cn>
> ---
> hw/nvme/ctrl.c | 136 ++++++++++++++++++++++++++++++++++++++++++-------
> hw/nvme/nvme.h | 4 ++
> 2 files changed, 123 insertions(+), 17 deletions(-)
>
> diff --git a/hw/nvme/ctrl.c b/hw/nvme/ctrl.c
> index 87aeba0564..6ecf6fafd9 100644
> --- a/hw/nvme/ctrl.c
> +++ b/hw/nvme/ctrl.c
> @@ -1377,8 +1448,25 @@ static void nvme_post_cqes(void *opaque)
> QTAILQ_INSERT_TAIL(&sq->req_list, req, entry);
> }
> if (cq->tail != cq->head) {
> - if (cq->irq_enabled && !pending) {
> - n->cq_pending++;
> + if (cq->irq_enabled) {
> + if (!pending) {
> + n->cq_pending++;
> + }
> +
> + if (unlikely(cq->first_io_cqe)) {
> + /*
> + * Initilize event notifier when first cqe is posted. For
> irqfd
> + * support we need to register the MSI message in KVM. We
> + * can not do this registration at CQ creation time because
> + * Linux's NVMe driver changes the MSI message after CQ
> creation.
> + */
> + cq->first_io_cqe = false;
> +
> + if (n->params.irq_eventfd) {
> + nvme_init_irq_notifier(n, cq);
> + }
> + }I'm still a bit perplexed by this issue, so I just tried moving nvme_init_irq_notifier() to the end of nvme_init_cq() and removing this first_io_cqe thing. I did not observe any particular issues? What bad behavior did you encounter, it seems to work fine to me?
signature.asc
Description: PGP signature
