On Tue, Mar 27, 2018 at 09:39:08AM -0600, Keith Busch wrote:
> The admin and first IO queues shared the first irq vector, which has an
> affinity mask including cpu0. If a system allows cpu0 to be offlined,
> the admin queue may not be usable if no other CPUs in the affinity mask
> are online. This is a problem since unlike IO queues, there is only
> one admin queue that always needs to be usable.
>
> To fix, this patch allocates one pre_vector for the admin queue that
> is assigned all CPUs, so will always be accessible. The IO queues are
> assigned the remaining managed vectors.
>
> In case a controller has only one interrupt vector available, the admin
> and IO queues will share the pre_vector with all CPUs assigned.
>
> Cc: Jianchao Wang <jianchao.w.wang@xxxxxxxxxx>
> Cc: Ming Lei <ming.lei@xxxxxxxxxx>
> Signed-off-by: Keith Busch <keith.busch@xxxxxxxxx>
> ---
> v1 -> v2:
>
> Update to use new blk-mq API.
>
> Removed unnecessary braces, inline functions, and temp variables.
>
> Amended author (this has evolved significantly from the original).
>
> drivers/nvme/host/pci.c | 23 +++++++++++++++++------
> 1 file changed, 17 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/nvme/host/pci.c b/drivers/nvme/host/pci.c
> index cc47fbe32ea5..50c8eaf51d92 100644
> --- a/drivers/nvme/host/pci.c
> +++ b/drivers/nvme/host/pci.c
> @@ -84,6 +84,7 @@ struct nvme_dev {
> struct dma_pool *prp_small_pool;
> unsigned online_queues;
> unsigned max_qid;
> + unsigned int num_vecs;
> int q_depth;
> u32 db_stride;
> void __iomem *bar;
> @@ -414,7 +415,8 @@ static int nvme_pci_map_queues(struct blk_mq_tag_set *set)
> {
> struct nvme_dev *dev = set->driver_data;
>
> - return blk_mq_pci_map_queues(set, to_pci_dev(dev->dev), 0);
> + return blk_mq_pci_map_queues(set, to_pci_dev(dev->dev),
> + dev->num_vecs > 1);
> }
>
> /**
> @@ -1455,7 +1457,11 @@ static int nvme_create_queue(struct nvme_queue *nvmeq, int qid)
> nvmeq->sq_cmds_io = dev->cmb + offset;
> }
>
> - nvmeq->cq_vector = qid - 1;
> + /*
> + * A queue's vector matches the queue identifier unless the controller
> + * has only one vector available.
> + */
> + nvmeq->cq_vector = dev->num_vecs == 1 ? 0 : qid;
> result = adapter_alloc_cq(dev, qid, nvmeq);
> if (result < 0)
> goto release_vector;
> @@ -1909,6 +1915,10 @@ static int nvme_setup_io_queues(struct nvme_dev *dev)
> int result, nr_io_queues;
> unsigned long size;
>
> + struct irq_affinity affd = {
> + .pre_vectors = 1
> + };
> +
> nr_io_queues = num_present_cpus();
> result = nvme_set_queue_count(&dev->ctrl, &nr_io_queues);
> if (result < 0)
> @@ -1944,11 +1954,12 @@ static int nvme_setup_io_queues(struct nvme_dev *dev)
> * setting up the full range we need.
> */
> pci_free_irq_vectors(pdev);
> - nr_io_queues = pci_alloc_irq_vectors(pdev, 1, nr_io_queues,
> - PCI_IRQ_ALL_TYPES | PCI_IRQ_AFFINITY);
> - if (nr_io_queues <= 0)
> + result = pci_alloc_irq_vectors_affinity(pdev, 1, nr_io_queues + 1,
> + PCI_IRQ_ALL_TYPES | PCI_IRQ_AFFINITY, &affd);
> + if (result <= 0)
> return -EIO;
> - dev->max_qid = nr_io_queues;
> + dev->num_vecs = result;
> + dev->max_qid = max(result - 1, 1);
>
> /*
> * Should investigate if there's a performance win from allocating
> --
> 2.14.3
>
Reviewed-by: Ming Lei <ming.lei@xxxxxxxxxx>
--
Ming
