On 19-07-17 10:45:27, Benjamin Herrenschmidt wrote:
> Based on reverse engineering and original patch by
>
> Paul Pawlowski <p...@mrarm.io>
>
> This adds support for Apple weird implementation of NVME in their
> 2018 or later machines. It accounts for the twice-as-big SQ entries
> for the IO queues, and the fact that only interrupt vector 0 appears
> to function properly.
>
> Signed-off-by: Benjamin Herrenschmidt <b...@kernel.crashing.org>
>
> # Conflicts:
> # drivers/nvme/host/core.c
> ---
> drivers/nvme/host/nvme.h | 10 ++++++++++
> drivers/nvme/host/pci.c | 21 ++++++++++++++++++++-
> 2 files changed, 30 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/nvme/host/nvme.h b/drivers/nvme/host/nvme.h
> index 716a876119c8..ced0e0a7e039 100644
> --- a/drivers/nvme/host/nvme.h
> +++ b/drivers/nvme/host/nvme.h
> @@ -92,6 +92,16 @@ enum nvme_quirks {
> * Broken Write Zeroes.
> */
> NVME_QUIRK_DISABLE_WRITE_ZEROES = (1 << 9),
> +
> + /*
> + * Use only one interrupt vector for all queues
> + */
> + NVME_QUIRK_SINGLE_VECTOR = (1 << 10),
> +
> + /*
> + * Use non-standard 128 bytes SQEs.
> + */
> + NVME_QUIRK_128_BYTES_SQES = (1 << 11),
> };
>
> /*
> diff --git a/drivers/nvme/host/pci.c b/drivers/nvme/host/pci.c
> index 1637677afb78..7088971d4c42 100644
> --- a/drivers/nvme/host/pci.c
> +++ b/drivers/nvme/host/pci.c
> @@ -2081,6 +2081,13 @@ static int nvme_setup_irqs(struct nvme_dev *dev,
> unsigned int nr_io_queues)
> dev->io_queues[HCTX_TYPE_DEFAULT] = 1;
> dev->io_queues[HCTX_TYPE_READ] = 0;
>
> + /*
> + * Some Apple controllers require all queues to use the
> + * first vector.
> + */
> + if (dev->ctrl.quirks & NVME_QUIRK_SINGLE_VECTOR)
> + irq_queues = 1;
> +
> return pci_alloc_irq_vectors_affinity(pdev, 1, irq_queues,
> PCI_IRQ_ALL_TYPES | PCI_IRQ_AFFINITY, &affd);
> }
> @@ -2322,7 +2329,16 @@ static int nvme_pci_enable(struct nvme_dev *dev)
> io_queue_depth);
> dev->db_stride = 1 << NVME_CAP_STRIDE(dev->ctrl.cap);
> dev->dbs = dev->bar + 4096;
> - dev->io_sqes = NVME_NVM_IOSQES;
> +
> + /*
> + * Some Apple controllers require a non-standard SQE size.
> + * Interestingly they also seem to ignore the CC:IOSQES register
> + * so we don't bother updating it here.
> + */
That is really interesting.
This also looks good to me.
Reviewed-by: Minwoo Im <minwoo.im....@gmail.com>
Thanks,
