Updated the driver to utilize all the MSI-X interrupt vectors supported by each OCTEON endpoint VF, instead of relying on a single vector. Enabling more interrupts allows packets from multiple rings to be distributed across multiple cores, improving parallelism and performance.
Reviewed-by: Dan Carpenter <dan.carpen...@linaro.org> Acked-by: Jason Wang <jasow...@redhat.com> Signed-off-by: Shijith Thotton <sthot...@marvell.com> --- v1: - https://lore.kernel.org/virtualization/20241120070508.789508-1-sthot...@marvell.com Changes in v2: - Handle reset getting called twice. - Use devm_kcalloc to allocate irq array. - IRQ is never zero. Adjusted code accordingly. Changes in v3: - Made the interrupt handling independent of device type. - Dropped the patch to add crypto device specific interrupt handler. - Moved spec specific macro and define to spec header file. Changes in v4: - Included comments from spec in the spec header file. - Added BUILD_BUG_ON to check vendor data structure size. drivers/vdpa/octeon_ep/octep_vdpa.h | 12 ++-- drivers/vdpa/octeon_ep/octep_vdpa_hw.c | 2 - drivers/vdpa/octeon_ep/octep_vdpa_main.c | 87 +++++++++++++++--------- 3 files changed, 62 insertions(+), 39 deletions(-) diff --git a/drivers/vdpa/octeon_ep/octep_vdpa.h b/drivers/vdpa/octeon_ep/octep_vdpa.h index 046710ec4d42..2cadb878e679 100644 --- a/drivers/vdpa/octeon_ep/octep_vdpa.h +++ b/drivers/vdpa/octeon_ep/octep_vdpa.h @@ -29,12 +29,12 @@ #define OCTEP_EPF_RINFO(x) (0x000209f0 | ((x) << 25)) #define OCTEP_VF_MBOX_DATA(x) (0x00010210 | ((x) << 17)) #define OCTEP_PF_MBOX_DATA(x) (0x00022000 | ((x) << 4)) - -#define OCTEP_EPF_RINFO_RPVF(val) (((val) >> 32) & 0xF) -#define OCTEP_EPF_RINFO_NVFS(val) (((val) >> 48) & 0x7F) +#define OCTEP_VF_IN_CTRL(x) (0x00010000 | ((x) << 17)) +#define OCTEP_VF_IN_CTRL_RPVF(val) (((val) >> 48) & 0xF) #define OCTEP_FW_READY_SIGNATURE0 0xFEEDFEED #define OCTEP_FW_READY_SIGNATURE1 0x3355ffaa +#define OCTEP_MAX_CB_INTR 8 enum octep_vdpa_dev_status { OCTEP_VDPA_DEV_STATUS_INVALID, @@ -48,9 +48,8 @@ enum octep_vdpa_dev_status { struct octep_vring_info { struct vdpa_callback cb; void __iomem *notify_addr; - u32 __iomem *cb_notify_addr; + void __iomem *cb_notify_addr; phys_addr_t notify_pa; - char msix_name[256]; }; struct octep_hw { @@ -68,7 +67,8 @@ struct octep_hw { u64 features; u16 nr_vring; u32 config_size; - int irq; + int nb_irqs; + int *irqs; }; u8 octep_hw_get_status(struct octep_hw *oct_hw); diff --git a/drivers/vdpa/octeon_ep/octep_vdpa_hw.c b/drivers/vdpa/octeon_ep/octep_vdpa_hw.c index 1d4767b33315..d5a599f87e18 100644 --- a/drivers/vdpa/octeon_ep/octep_vdpa_hw.c +++ b/drivers/vdpa/octeon_ep/octep_vdpa_hw.c @@ -495,8 +495,6 @@ int octep_hw_caps_read(struct octep_hw *oct_hw, struct pci_dev *pdev) if (!oct_hw->vqs) return -ENOMEM; - oct_hw->irq = -1; - dev_info(&pdev->dev, "Device features : %llx\n", oct_hw->features); dev_info(&pdev->dev, "Maximum queues : %u\n", oct_hw->nr_vring); diff --git a/drivers/vdpa/octeon_ep/octep_vdpa_main.c b/drivers/vdpa/octeon_ep/octep_vdpa_main.c index cd55b1aac151..e9c3e57b321f 100644 --- a/drivers/vdpa/octeon_ep/octep_vdpa_main.c +++ b/drivers/vdpa/octeon_ep/octep_vdpa_main.c @@ -49,11 +49,25 @@ static irqreturn_t octep_vdpa_intr_handler(int irq, void *data) struct octep_hw *oct_hw = data; int i; - for (i = 0; i < oct_hw->nr_vring; i++) { - if (oct_hw->vqs[i].cb.callback && ioread32(oct_hw->vqs[i].cb_notify_addr)) { - /* Acknowledge the per queue notification to the device */ - iowrite32(0, oct_hw->vqs[i].cb_notify_addr); - oct_hw->vqs[i].cb.callback(oct_hw->vqs[i].cb.private); + /* Each device has multiple interrupts (nb_irqs) shared among rings + * (nr_vring). Device interrupts are mapped to the rings in a + * round-robin fashion. + * + * For example, if nb_irqs = 8 and nr_vring = 64: + * 0 -> 0, 8, 16, 24, 32, 40, 48, 56; + * 1 -> 1, 9, 17, 25, 33, 41, 49, 57; + * ... + * 7 -> 7, 15, 23, 31, 39, 47, 55, 63; + */ + + for (i = irq - oct_hw->irqs[0]; i < oct_hw->nr_vring; i += oct_hw->nb_irqs) { + if (ioread8(oct_hw->vqs[i].cb_notify_addr)) { + /* Acknowledge the per ring notification to the device */ + iowrite8(0, oct_hw->vqs[i].cb_notify_addr); + + if (likely(oct_hw->vqs[i].cb.callback)) + oct_hw->vqs[i].cb.callback(oct_hw->vqs[i].cb.private); + break; } } @@ -63,44 +77,53 @@ static irqreturn_t octep_vdpa_intr_handler(int irq, void *data) static void octep_free_irqs(struct octep_hw *oct_hw) { struct pci_dev *pdev = oct_hw->pdev; + int irq; + + if (!oct_hw->irqs) + return; - if (oct_hw->irq != -1) { - devm_free_irq(&pdev->dev, oct_hw->irq, oct_hw); - oct_hw->irq = -1; + for (irq = 0; irq < oct_hw->nb_irqs; irq++) { + if (!oct_hw->irqs[irq]) + break; + + devm_free_irq(&pdev->dev, oct_hw->irqs[irq], oct_hw); } + pci_free_irq_vectors(pdev); + devm_kfree(&pdev->dev, oct_hw->irqs); + oct_hw->irqs = NULL; } static int octep_request_irqs(struct octep_hw *oct_hw) { struct pci_dev *pdev = oct_hw->pdev; - int ret, irq; + int ret, irq, idx; - /* Currently HW device provisions one IRQ per VF, hence - * allocate one IRQ for all virtqueues call interface. - */ - ret = pci_alloc_irq_vectors(pdev, 1, 1, PCI_IRQ_MSIX); + oct_hw->irqs = devm_kcalloc(&pdev->dev, oct_hw->nb_irqs, sizeof(int), GFP_KERNEL); + if (!oct_hw->irqs) + return -ENOMEM; + + ret = pci_alloc_irq_vectors(pdev, 1, oct_hw->nb_irqs, PCI_IRQ_MSIX); if (ret < 0) { dev_err(&pdev->dev, "Failed to alloc msix vector"); return ret; } - snprintf(oct_hw->vqs->msix_name, sizeof(oct_hw->vqs->msix_name), - OCTEP_VDPA_DRIVER_NAME "-vf-%d", pci_iov_vf_id(pdev)); - - irq = pci_irq_vector(pdev, 0); - ret = devm_request_irq(&pdev->dev, irq, octep_vdpa_intr_handler, 0, - oct_hw->vqs->msix_name, oct_hw); - if (ret) { - dev_err(&pdev->dev, "Failed to register interrupt handler\n"); - goto free_irq_vec; + for (idx = 0; idx < oct_hw->nb_irqs; idx++) { + irq = pci_irq_vector(pdev, idx); + ret = devm_request_irq(&pdev->dev, irq, octep_vdpa_intr_handler, 0, + dev_name(&pdev->dev), oct_hw); + if (ret) { + dev_err(&pdev->dev, "Failed to register interrupt handler\n"); + goto free_irqs; + } + oct_hw->irqs[idx] = irq; } - oct_hw->irq = irq; return 0; -free_irq_vec: - pci_free_irq_vectors(pdev); +free_irqs: + octep_free_irqs(oct_hw); return ret; } @@ -559,6 +582,7 @@ static void octep_vdpa_setup_task(struct work_struct *work) struct device *dev = &pdev->dev; struct octep_hw *oct_hw; unsigned long timeout; + u64 val; int ret; oct_hw = &mgmt_dev->oct_hw; @@ -590,6 +614,13 @@ static void octep_vdpa_setup_task(struct work_struct *work) if (ret) return; + val = readq(oct_hw->base[OCTEP_HW_MBOX_BAR] + OCTEP_VF_IN_CTRL(0)); + oct_hw->nb_irqs = OCTEP_VF_IN_CTRL_RPVF(val); + if (!oct_hw->nb_irqs || oct_hw->nb_irqs > OCTEP_MAX_CB_INTR) { + dev_err(dev, "Invalid number of interrupts %d\n", oct_hw->nb_irqs); + goto unmap_region; + } + ret = octep_hw_caps_read(oct_hw, pdev); if (ret < 0) goto unmap_region; @@ -768,12 +799,6 @@ static int octep_vdpa_pf_setup(struct octep_pf *octpf) return -EINVAL; } - if (OCTEP_EPF_RINFO_RPVF(val) != BIT_ULL(0)) { - val &= ~GENMASK_ULL(35, 32); - val |= BIT_ULL(32); - writeq(val, addr + OCTEP_EPF_RINFO(0)); - } - len = pci_resource_len(pdev, OCTEP_HW_CAPS_BAR); octpf->vf_stride = len / totalvfs; -- 2.25.1
