From patchwork Tue Feb 22 11:54:27 2022 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: "Zhu, Lingshan" X-Patchwork-Id: 12754948 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 9BDA1C433EF for ; Tue, 22 Feb 2022 12:02:13 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S231934AbiBVMCf (ORCPT ); Tue, 22 Feb 2022 07:02:35 -0500 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:48290 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S231892AbiBVMCb (ORCPT ); Tue, 22 Feb 2022 07:02:31 -0500 Received: from mga05.intel.com (mga05.intel.com [192.55.52.43]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 0D72B24F0C for ; Tue, 22 Feb 2022 04:02:06 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=intel.com; i=@intel.com; q=dns/txt; s=Intel; t=1645531325; x=1677067325; h=from:to:cc:subject:date:message-id:in-reply-to: references:mime-version:content-transfer-encoding; bh=oMTWR9yxVq54Coe0uIi0r/mi4lE0WlrS1FdggqWU+gQ=; b=PmnlAVp4ZIPdV8EgitRdjNWNN2hrlN+SMQUvOfZec5H2JejN/TnJSLqA jdnn9Tn6xoz8GysgkhPglM84NFId54K7GSL5G8YTgTj3zlfSzYHZ1vWji F1cQIUF7T0XGWzJuFCzmKyhKLw0CdNsoGq+QkdKsRlE62lyQr2/pQKJbO IzyEk3vkolpknqSl8wCvtbP7oVpbgyxHFPTX0zkK/aU//BDpykuS6yyVj RJyCawSC8HbkGncqpOXVcR7vTL14FlKyILJmocGPQiTLB3p6rGoNqYM2s 5NRqERyqvdbqcvBZDKcxrcoq1IPx4w7+C6Hs2Ir0Vnz+sp8/uXfVNCn7N w==; X-IronPort-AV: E=McAfee;i="6200,9189,10265"; a="338117859" X-IronPort-AV: E=Sophos;i="5.88,387,1635231600"; d="scan'208";a="338117859" Received: from orsmga003.jf.intel.com ([10.7.209.27]) by fmsmga105.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 22 Feb 2022 04:02:05 -0800 X-IronPort-AV: E=Sophos;i="5.88,387,1635231600"; d="scan'208";a="490772593" Received: from unknown (HELO cra01infra01.deacluster.intel.com) ([10.240.193.73]) by orsmga003-auth.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 22 Feb 2022 04:02:03 -0800 From: Zhu Lingshan To: mst@redhat.com, jasowang@redhat.com Cc: netdev@vger.kernel.org, virtualization@lists.linux-foundation.org, Zhu Lingshan Subject: [PATCH V6 4/5] vDPA/ifcvf: implement shared IRQ feature Date: Tue, 22 Feb 2022 19:54:27 +0800 Message-Id: <20220222115428.998334-5-lingshan.zhu@intel.com> X-Mailer: git-send-email 2.27.0 In-Reply-To: <20220222115428.998334-1-lingshan.zhu@intel.com> References: <20220222115428.998334-1-lingshan.zhu@intel.com> MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: netdev@vger.kernel.org On some platforms/devices, there may not be enough MSI vectors allocated for the virtqueues and config changes. In such a case, the interrupt sources(virtqueues, config changes) must share an IRQ/vector, to avoid initialization failures, keep the device functional. This commit handles three cases: (1) number of the allocated vectors == the number of virtqueues + 1 (config changes), every virtqueue and the config interrupt has a separated vector/IRQ, the best and the most likely case. (2) number of the allocated vectors is less than the best case, but greater than 1. In this case, all virtqueues share a vector/IRQ, the config interrupt has a separated vector/IRQ (3) only one vector is allocated, in this case, the virtqueues and the config interrupt share a vector/IRQ. The worst and most unlikely case. Otherwise, it needs to fail. This commit introduces some helper functions: ifcvf_set_vq_vector() and ifcvf_set_config_vector() sets virtqueue vector and config vector in the device config space, so that the device can send interrupt DMA. Signed-off-by: Zhu Lingshan --- drivers/vdpa/ifcvf/ifcvf_base.c | 48 +++--- drivers/vdpa/ifcvf/ifcvf_base.h | 15 +- drivers/vdpa/ifcvf/ifcvf_main.c | 294 ++++++++++++++++++++++++++++---- 3 files changed, 300 insertions(+), 57 deletions(-) diff --git a/drivers/vdpa/ifcvf/ifcvf_base.c b/drivers/vdpa/ifcvf/ifcvf_base.c index b9fdc5258611..8aba3ab4a2f3 100644 --- a/drivers/vdpa/ifcvf/ifcvf_base.c +++ b/drivers/vdpa/ifcvf/ifcvf_base.c @@ -15,6 +15,26 @@ struct ifcvf_adapter *vf_to_adapter(struct ifcvf_hw *hw) return container_of(hw, struct ifcvf_adapter, vf); } +u16 ifcvf_set_vq_vector(struct ifcvf_hw *hw, u16 qid, int vector) +{ + struct virtio_pci_common_cfg __iomem *cfg = hw->common_cfg; + + vp_iowrite16(qid, &cfg->queue_select); + vp_iowrite16(vector, &cfg->queue_msix_vector); + + return vp_ioread16(&cfg->queue_msix_vector); +} + +u16 ifcvf_set_config_vector(struct ifcvf_hw *hw, int vector) +{ + struct virtio_pci_common_cfg __iomem *cfg = hw->common_cfg; + + cfg = hw->common_cfg; + vp_iowrite16(vector, &cfg->msix_config); + + return vp_ioread16(&cfg->msix_config); +} + static void __iomem *get_cap_addr(struct ifcvf_hw *hw, struct virtio_pci_cap *cap) { @@ -131,6 +151,7 @@ int ifcvf_init_hw(struct ifcvf_hw *hw, struct pci_dev *pdev) notify_off * hw->notify_off_multiplier; hw->vring[i].notify_pa = hw->notify_base_pa + notify_off * hw->notify_off_multiplier; + hw->vring[i].irq = -EINVAL; } hw->lm_cfg = hw->base[IFCVF_LM_BAR]; @@ -140,6 +161,9 @@ int ifcvf_init_hw(struct ifcvf_hw *hw, struct pci_dev *pdev) hw->common_cfg, hw->notify_base, hw->isr, hw->dev_cfg, hw->notify_off_multiplier); + hw->vqs_reused_irq = -EINVAL; + hw->config_irq = -EINVAL; + return 0; } @@ -321,13 +345,6 @@ static int ifcvf_hw_enable(struct ifcvf_hw *hw) ifcvf = vf_to_adapter(hw); cfg = hw->common_cfg; - vp_iowrite16(IFCVF_MSI_CONFIG_OFF, &cfg->msix_config); - - if (vp_ioread16(&cfg->msix_config) == VIRTIO_MSI_NO_VECTOR) { - IFCVF_ERR(ifcvf->pdev, "No msix vector for device config\n"); - return -EINVAL; - } - for (i = 0; i < hw->nr_vring; i++) { if (!hw->vring[i].ready) break; @@ -340,15 +357,6 @@ static int ifcvf_hw_enable(struct ifcvf_hw *hw) vp_iowrite64_twopart(hw->vring[i].used, &cfg->queue_used_lo, &cfg->queue_used_hi); vp_iowrite16(hw->vring[i].size, &cfg->queue_size); - vp_iowrite16(i + IFCVF_MSI_QUEUE_OFF, &cfg->queue_msix_vector); - - if (vp_ioread16(&cfg->queue_msix_vector) == - VIRTIO_MSI_NO_VECTOR) { - IFCVF_ERR(ifcvf->pdev, - "No msix vector for queue %u\n", i); - return -EINVAL; - } - ifcvf_set_vq_state(hw, i, hw->vring[i].last_avail_idx); vp_iowrite16(1, &cfg->queue_enable); } @@ -362,14 +370,10 @@ static void ifcvf_hw_disable(struct ifcvf_hw *hw) u32 i; cfg = hw->common_cfg; - vp_iowrite16(VIRTIO_MSI_NO_VECTOR, &cfg->msix_config); - + ifcvf_set_config_vector(hw, VIRTIO_MSI_NO_VECTOR); for (i = 0; i < hw->nr_vring; i++) { - vp_iowrite16(i, &cfg->queue_select); - vp_iowrite16(VIRTIO_MSI_NO_VECTOR, &cfg->queue_msix_vector); + ifcvf_set_vq_vector(hw, i, VIRTIO_MSI_NO_VECTOR); } - - vp_ioread16(&cfg->queue_msix_vector); } int ifcvf_start_hw(struct ifcvf_hw *hw) diff --git a/drivers/vdpa/ifcvf/ifcvf_base.h b/drivers/vdpa/ifcvf/ifcvf_base.h index 25c591a3eae2..dcd31accfce5 100644 --- a/drivers/vdpa/ifcvf/ifcvf_base.h +++ b/drivers/vdpa/ifcvf/ifcvf_base.h @@ -28,8 +28,6 @@ #define IFCVF_QUEUE_ALIGNMENT PAGE_SIZE #define IFCVF_QUEUE_MAX 32768 -#define IFCVF_MSI_CONFIG_OFF 0 -#define IFCVF_MSI_QUEUE_OFF 1 #define IFCVF_PCI_MAX_RESOURCE 6 #define IFCVF_LM_CFG_SIZE 0x40 @@ -43,6 +41,13 @@ #define ifcvf_private_to_vf(adapter) \ (&((struct ifcvf_adapter *)adapter)->vf) +/* all vqs and config interrupt has its own vector */ +#define MSIX_VECTOR_PER_VQ_AND_CONFIG 1 +/* all vqs share a vector, and config interrupt has a separate vector */ +#define MSIX_VECTOR_SHARED_VQ_AND_CONFIG 2 +/* all vqs and config interrupt share a vector */ +#define MSIX_VECTOR_DEV_SHARED 3 + struct vring_info { u64 desc; u64 avail; @@ -77,9 +82,11 @@ struct ifcvf_hw { void __iomem * const *base; char config_msix_name[256]; struct vdpa_callback config_cb; - unsigned int config_irq; + int config_irq; + int vqs_reused_irq; /* virtio-net or virtio-blk device config size */ u32 config_size; + u8 msix_vector_status; }; struct ifcvf_adapter { @@ -124,4 +131,6 @@ int ifcvf_set_vq_state(struct ifcvf_hw *hw, u16 qid, u16 num); struct ifcvf_adapter *vf_to_adapter(struct ifcvf_hw *hw); int ifcvf_probed_virtio_net(struct ifcvf_hw *hw); u32 ifcvf_get_config_size(struct ifcvf_hw *hw); +u16 ifcvf_set_vq_vector(struct ifcvf_hw *hw, u16 qid, int vector); +u16 ifcvf_set_config_vector(struct ifcvf_hw *hw, int vector); #endif /* _IFCVF_H_ */ diff --git a/drivers/vdpa/ifcvf/ifcvf_main.c b/drivers/vdpa/ifcvf/ifcvf_main.c index 964f7ac142ba..3b48e717e89f 100644 --- a/drivers/vdpa/ifcvf/ifcvf_main.c +++ b/drivers/vdpa/ifcvf/ifcvf_main.c @@ -27,7 +27,7 @@ static irqreturn_t ifcvf_config_changed(int irq, void *arg) return IRQ_HANDLED; } -static irqreturn_t ifcvf_intr_handler(int irq, void *arg) +static irqreturn_t ifcvf_vq_intr_handler(int irq, void *arg) { struct vring_info *vring = arg; @@ -37,24 +37,98 @@ static irqreturn_t ifcvf_intr_handler(int irq, void *arg) return IRQ_HANDLED; } +static irqreturn_t ifcvf_vqs_reused_intr_handler(int irq, void *arg) +{ + struct ifcvf_hw *vf = arg; + struct vring_info *vring; + int i; + + for (i = 0; i < vf->nr_vring; i++) { + vring = &vf->vring[i]; + if (vring->cb.callback) + vf->vring->cb.callback(vring->cb.private); + } + + return IRQ_HANDLED; +} + +static irqreturn_t ifcvf_dev_intr_handler(int irq, void *arg) +{ + struct ifcvf_hw *vf = arg; + u8 isr; + + isr = vp_ioread8(vf->isr); + if (isr & VIRTIO_PCI_ISR_CONFIG) + ifcvf_config_changed(irq, arg); + + return ifcvf_vqs_reused_intr_handler(irq, arg); +} + static void ifcvf_free_irq_vectors(void *data) { pci_free_irq_vectors(data); } -static void ifcvf_free_irq(struct ifcvf_adapter *adapter, int queues) +static void ifcvf_free_per_vq_irq(struct ifcvf_adapter *adapter) { struct pci_dev *pdev = adapter->pdev; struct ifcvf_hw *vf = &adapter->vf; int i; + for (i = 0; i < vf->nr_vring; i++) { + if (vf->vring[i].irq != -EINVAL) { + devm_free_irq(&pdev->dev, vf->vring[i].irq, &vf->vring[i]); + vf->vring[i].irq = -EINVAL; + } + } +} - for (i = 0; i < queues; i++) { - devm_free_irq(&pdev->dev, vf->vring[i].irq, &vf->vring[i]); - vf->vring[i].irq = -EINVAL; +static void ifcvf_free_vqs_reused_irq(struct ifcvf_adapter *adapter) +{ + struct pci_dev *pdev = adapter->pdev; + struct ifcvf_hw *vf = &adapter->vf; + + if (vf->vqs_reused_irq != -EINVAL) { + devm_free_irq(&pdev->dev, vf->vqs_reused_irq, vf); + vf->vqs_reused_irq = -EINVAL; } - devm_free_irq(&pdev->dev, vf->config_irq, vf); +} + +static void ifcvf_free_vq_irq(struct ifcvf_adapter *adapter) +{ + struct ifcvf_hw *vf = &adapter->vf; + + if (vf->msix_vector_status == MSIX_VECTOR_PER_VQ_AND_CONFIG) + ifcvf_free_per_vq_irq(adapter); + else + ifcvf_free_vqs_reused_irq(adapter); +} + +static void ifcvf_free_config_irq(struct ifcvf_adapter *adapter) +{ + struct pci_dev *pdev = adapter->pdev; + struct ifcvf_hw *vf = &adapter->vf; + + if (vf->config_irq == -EINVAL) + return; + + /* If the irq is shared by all vqs and the config interrupt, + * it is already freed in ifcvf_free_vq_irq, so here only + * need to free config irq when msix_vector_status != MSIX_VECTOR_DEV_SHARED + */ + if (vf->msix_vector_status != MSIX_VECTOR_DEV_SHARED) { + devm_free_irq(&pdev->dev, vf->config_irq, vf); + vf->config_irq = -EINVAL; + } +} + +static void ifcvf_free_irq(struct ifcvf_adapter *adapter) +{ + struct pci_dev *pdev = adapter->pdev; + + ifcvf_free_vq_irq(adapter); + ifcvf_free_config_irq(adapter); ifcvf_free_irq_vectors(pdev); } @@ -86,48 +160,201 @@ static int ifcvf_alloc_vectors(struct ifcvf_adapter *adapter) return ret; } -static int ifcvf_request_irq(struct ifcvf_adapter *adapter) +static int ifcvf_request_per_vq_irq(struct ifcvf_adapter *adapter) { struct pci_dev *pdev = adapter->pdev; struct ifcvf_hw *vf = &adapter->vf; - int vector, nvectors, i, ret, irq; + int i, vector, ret, irq; - nvectors = ifcvf_alloc_vectors(adapter); - if (nvectors <= 0) - return -EFAULT; + vf->vqs_reused_irq = -EINVAL; + for (i = 0; i < vf->nr_vring; i++) { + snprintf(vf->vring[i].msix_name, 256, "ifcvf[%s]-%d\n", pci_name(pdev), i); + vector = i; + irq = pci_irq_vector(pdev, vector); + ret = devm_request_irq(&pdev->dev, irq, + ifcvf_vq_intr_handler, 0, + vf->vring[i].msix_name, + &vf->vring[i]); + if (ret) { + IFCVF_ERR(pdev, "Failed to request irq for vq %d\n", i); + goto err; + } + + vf->vring[i].irq = irq; + ret = ifcvf_set_vq_vector(vf, i, vector); + if (ret == VIRTIO_MSI_NO_VECTOR) { + IFCVF_ERR(pdev, "No msix vector for vq %u\n", i); + goto err; + } + } + + return 0; +err: + ifcvf_free_irq(adapter); + + return -EFAULT; +} + +static int ifcvf_request_vqs_reused_irq(struct ifcvf_adapter *adapter) +{ + struct pci_dev *pdev = adapter->pdev; + struct ifcvf_hw *vf = &adapter->vf; + int i, vector, ret, irq; + + vector = 0; + snprintf(vf->vring[0].msix_name, 256, "ifcvf[%s]-vqs-reused-irq\n", pci_name(pdev)); + irq = pci_irq_vector(pdev, vector); + ret = devm_request_irq(&pdev->dev, irq, + ifcvf_vqs_reused_intr_handler, 0, + vf->vring[0].msix_name, vf); + if (ret) { + IFCVF_ERR(pdev, "Failed to request reused irq for the device\n"); + goto err; + } + + vf->vqs_reused_irq = irq; + for (i = 0; i < vf->nr_vring; i++) { + vf->vring[i].irq = -EINVAL; + ret = ifcvf_set_vq_vector(vf, i, vector); + if (ret == VIRTIO_MSI_NO_VECTOR) { + IFCVF_ERR(pdev, "No msix vector for vq %u\n", i); + goto err; + } + } + + return 0; +err: + ifcvf_free_irq(adapter); + + return -EFAULT; +} + +static int ifcvf_request_dev_irq(struct ifcvf_adapter *adapter) +{ + struct pci_dev *pdev = adapter->pdev; + struct ifcvf_hw *vf = &adapter->vf; + int i, vector, ret, irq; + + vector = 0; + snprintf(vf->vring[0].msix_name, 256, "ifcvf[%s]-dev-irq\n", pci_name(pdev)); + irq = pci_irq_vector(pdev, vector); + ret = devm_request_irq(&pdev->dev, irq, + ifcvf_dev_intr_handler, 0, + vf->vring[0].msix_name, vf); + if (ret) { + IFCVF_ERR(pdev, "Failed to request irq for the device\n"); + goto err; + } + + vf->vqs_reused_irq = irq; + for (i = 0; i < vf->nr_vring; i++) { + vf->vring[i].irq = -EINVAL; + ret = ifcvf_set_vq_vector(vf, i, vector); + if (ret == VIRTIO_MSI_NO_VECTOR) { + IFCVF_ERR(pdev, "No msix vector for vq %u\n", i); + goto err; + } + } + + vf->config_irq = irq; + ret = ifcvf_set_config_vector(vf, vector); + if (ret == VIRTIO_MSI_NO_VECTOR) { + IFCVF_ERR(pdev, "No msix vector for device config\n"); + goto err; + } + + return 0; +err: + ifcvf_free_irq(adapter); + + return -EFAULT; + +} + +static int ifcvf_request_vq_irq(struct ifcvf_adapter *adapter) +{ + struct ifcvf_hw *vf = &adapter->vf; + int ret; + + if (vf->msix_vector_status == MSIX_VECTOR_PER_VQ_AND_CONFIG) + ret = ifcvf_request_per_vq_irq(adapter); + else + ret = ifcvf_request_vqs_reused_irq(adapter); + + return ret; +} + +static int ifcvf_request_config_irq(struct ifcvf_adapter *adapter) +{ + struct pci_dev *pdev = adapter->pdev; + struct ifcvf_hw *vf = &adapter->vf; + int config_vector, ret; + + if (vf->msix_vector_status == MSIX_VECTOR_DEV_SHARED) + return 0; + + if (vf->msix_vector_status == MSIX_VECTOR_PER_VQ_AND_CONFIG) + /* vector 0 ~ vf->nr_vring for vqs, num vf->nr_vring vector for config interrupt */ + config_vector = vf->nr_vring; + + if (vf->msix_vector_status == MSIX_VECTOR_SHARED_VQ_AND_CONFIG) + /* vector 0 for vqs and 1 for config interrupt */ + config_vector = 1; snprintf(vf->config_msix_name, 256, "ifcvf[%s]-config\n", pci_name(pdev)); - vector = 0; - vf->config_irq = pci_irq_vector(pdev, vector); + vf->config_irq = pci_irq_vector(pdev, config_vector); ret = devm_request_irq(&pdev->dev, vf->config_irq, ifcvf_config_changed, 0, vf->config_msix_name, vf); if (ret) { IFCVF_ERR(pdev, "Failed to request config irq\n"); - return ret; + goto err; } - for (i = 0; i < vf->nr_vring; i++) { - snprintf(vf->vring[i].msix_name, 256, "ifcvf[%s]-%d\n", - pci_name(pdev), i); - vector = i + IFCVF_MSI_QUEUE_OFF; - irq = pci_irq_vector(pdev, vector); - ret = devm_request_irq(&pdev->dev, irq, - ifcvf_intr_handler, 0, - vf->vring[i].msix_name, - &vf->vring[i]); - if (ret) { - IFCVF_ERR(pdev, - "Failed to request irq for vq %d\n", i); - ifcvf_free_irq(adapter, i); + ret = ifcvf_set_config_vector(vf, config_vector); + if (ret == VIRTIO_MSI_NO_VECTOR) { + IFCVF_ERR(pdev, "No msix vector for device config\n"); + goto err; + } - return ret; - } + return 0; +err: + ifcvf_free_irq(adapter); - vf->vring[i].irq = irq; + return -EFAULT; +} + +static int ifcvf_request_irq(struct ifcvf_adapter *adapter) +{ + struct ifcvf_hw *vf = &adapter->vf; + int nvectors, ret, max_intr; + + nvectors = ifcvf_alloc_vectors(adapter); + if (nvectors <= 0) + return -EFAULT; + + vf->msix_vector_status = MSIX_VECTOR_PER_VQ_AND_CONFIG; + max_intr = vf->nr_vring + 1; + if (nvectors < max_intr) + vf->msix_vector_status = MSIX_VECTOR_SHARED_VQ_AND_CONFIG; + + if (nvectors == 1) { + vf->msix_vector_status = MSIX_VECTOR_DEV_SHARED; + ret = ifcvf_request_dev_irq(adapter); + + return ret; } + ret = ifcvf_request_vq_irq(adapter); + if (ret) + return ret; + + ret = ifcvf_request_config_irq(adapter); + + if (ret) + return ret; + return 0; } @@ -284,7 +511,7 @@ static int ifcvf_vdpa_reset(struct vdpa_device *vdpa_dev) if (status_old & VIRTIO_CONFIG_S_DRIVER_OK) { ifcvf_stop_datapath(adapter); - ifcvf_free_irq(adapter, vf->nr_vring); + ifcvf_free_irq(adapter); } ifcvf_reset_vring(adapter); @@ -431,7 +658,10 @@ static int ifcvf_vdpa_get_vq_irq(struct vdpa_device *vdpa_dev, { struct ifcvf_hw *vf = vdpa_to_vf(vdpa_dev); - return vf->vring[qid].irq; + if (vf->vqs_reused_irq < 0) + return vf->vring[qid].irq; + else + return -EINVAL; } static struct vdpa_notification_area ifcvf_get_vq_notification(struct vdpa_device *vdpa_dev,