From patchwork Fri Sep 27 18:37:57 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Adrian Moreno X-Patchwork-Id: 60057 X-Patchwork-Delegate: maxime.coquelin@redhat.com Return-Path: X-Original-To: patchwork@dpdk.org Delivered-To: patchwork@dpdk.org Received: from [92.243.14.124] (localhost [127.0.0.1]) by dpdk.org (Postfix) with ESMTP id 2F51D1BEF7; Fri, 27 Sep 2019 20:38:23 +0200 (CEST) Received: from mx1.redhat.com (mx1.redhat.com [209.132.183.28]) by dpdk.org (Postfix) with ESMTP id CE9101BEDA; Fri, 27 Sep 2019 20:38:19 +0200 (CEST) Received: from smtp.corp.redhat.com (int-mx06.intmail.prod.int.phx2.redhat.com [10.5.11.16]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id 24CA0E8CDA; Fri, 27 Sep 2019 18:38:19 +0000 (UTC) Received: from amorenoz.users.ipa.redhat.com (ovpn-116-96.ams2.redhat.com [10.36.116.96]) by smtp.corp.redhat.com (Postfix) with ESMTP id 9983A5C21A; Fri, 27 Sep 2019 18:38:17 +0000 (UTC) From: Adrian Moreno To: dev@dpdk.org Cc: tiwei.bie@intel.com, zhihong.wang@intel.com, maxime.coquelin@redhat.com, Adrian Moreno , stable@dpdk.org Date: Fri, 27 Sep 2019 20:37:57 +0200 Message-Id: <20190927183759.18816-2-amorenoz@redhat.com> In-Reply-To: <20190927183759.18816-1-amorenoz@redhat.com> References: <20190927183759.18816-1-amorenoz@redhat.com> MIME-Version: 1.0 X-Scanned-By: MIMEDefang 2.79 on 10.5.11.16 X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.5.16 (mx1.redhat.com [10.5.110.38]); Fri, 27 Sep 2019 18:38:19 +0000 (UTC) Subject: [dpdk-dev] [PATCH v3 1/3] vhost: translate incoming log address to gpa X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org Sender: "dev" When IOMMU is enabled the incoming log address is in IOVA space. In that case, look in IOTLB table and translate the resulting HVA to GPA. If IOMMU is not enabled, the incoming log address is already a GPA so no transformation is needed. Fixes: 69c90e98f483 ("vhost: enable IOMMU support") Cc: maxime.coquelin@redhat.com Cc: stable@dpdk.org Signed-off-by: Adrian Moreno --- lib/librte_vhost/vhost.c | 1 + lib/librte_vhost/vhost.h | 20 ++++++++++++++++ lib/librte_vhost/vhost_user.c | 44 ++++++++++++++++++++++++++++++++++- 3 files changed, 64 insertions(+), 1 deletion(-) diff --git a/lib/librte_vhost/vhost.c b/lib/librte_vhost/vhost.c index 981837b5d..e57dda22f 100644 --- a/lib/librte_vhost/vhost.c +++ b/lib/librte_vhost/vhost.c @@ -383,6 +383,7 @@ vring_invalidate(struct virtio_net *dev, struct vhost_virtqueue *vq) vq->desc = NULL; vq->avail = NULL; vq->used = NULL; + vq->log_guest_addr = 0; if (dev->features & (1ULL << VIRTIO_F_IOMMU_PLATFORM)) vhost_user_iotlb_wr_unlock(vq); diff --git a/lib/librte_vhost/vhost.h b/lib/librte_vhost/vhost.h index 884befa85..9d3883385 100644 --- a/lib/librte_vhost/vhost.h +++ b/lib/librte_vhost/vhost.h @@ -447,6 +447,26 @@ gpa_to_hpa(struct virtio_net *dev, uint64_t gpa, uint64_t size) return 0; } +static __rte_always_inline uint64_t +hva_to_gpa(struct virtio_net *dev, uint64_t vva, uint64_t len) +{ + struct rte_vhost_mem_region *r; + uint32_t i; + + if (unlikely(!dev || !dev->mem)) + return 0; + + for (i = 0; i < dev->mem->nregions; i++) { + r = &dev->mem->regions[i]; + + if (vva >= r->host_user_addr && + vva + len < r->host_user_addr + r->size) { + return r->guest_phys_addr + vva - r->host_user_addr; + } + } + return 0; +} + static __rte_always_inline struct virtio_net * get_device(int vid) { diff --git a/lib/librte_vhost/vhost_user.c b/lib/librte_vhost/vhost_user.c index 0b72648a5..fec448ea2 100644 --- a/lib/librte_vhost/vhost_user.c +++ b/lib/librte_vhost/vhost_user.c @@ -570,6 +570,40 @@ ring_addr_to_vva(struct virtio_net *dev, struct vhost_virtqueue *vq, return qva_to_vva(dev, ra, size); } +/* + * Converts vring log address to GPA + * If IOMMU is enabled, the log address is IOVA + * If IOMMU not enabled, the log address is already GPA + */ +static uint64_t +translate_log_addr(struct virtio_net *dev, struct vhost_virtqueue *vq, + uint64_t log_addr) +{ + if (dev->features & (1ULL << VIRTIO_F_IOMMU_PLATFORM)) { + const uint64_t exp_size = sizeof(struct vring_used) + + sizeof(struct vring_used_elem) * vq->size; + uint64_t hva, gpa; + uint64_t size = exp_size; + + hva = vhost_iova_to_vva(dev, vq, log_addr, + &size, VHOST_ACCESS_RW); + if (size != exp_size) + return 0; + + gpa = hva_to_gpa(dev, hva, exp_size); + if (!gpa) { + RTE_LOG(ERR, VHOST_CONFIG, + "VQ: Failed to find GPA mapping for log_addr." + "log_addr: 0x%0lx hva: 0x%0lx\n", + log_addr, hva); + return 0; + } + return gpa; + + } else + return log_addr; +} + static struct virtio_net * translate_ring_addresses(struct virtio_net *dev, int vq_index) { @@ -676,7 +710,15 @@ translate_ring_addresses(struct virtio_net *dev, int vq_index) vq->last_avail_idx = vq->used->idx; } - vq->log_guest_addr = addr->log_guest_addr; + vq->log_guest_addr = + translate_log_addr(dev, vq, addr->log_guest_addr); + if (vq->log_guest_addr == 0) { + RTE_LOG(DEBUG, VHOST_CONFIG, + "(%d) failed to map log_guest_addr .\n", + dev->vid); + return dev; + } + VHOST_LOG_DEBUG(VHOST_CONFIG, "(%d) mapped address desc: %p\n", dev->vid, vq->desc); From patchwork Fri Sep 27 18:37:58 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Adrian Moreno X-Patchwork-Id: 60058 X-Patchwork-Delegate: maxime.coquelin@redhat.com Return-Path: X-Original-To: patchwork@dpdk.org Delivered-To: patchwork@dpdk.org Received: from [92.243.14.124] (localhost [127.0.0.1]) by dpdk.org (Postfix) with ESMTP id 164451BEEF; Fri, 27 Sep 2019 20:38:26 +0200 (CEST) Received: from mx1.redhat.com (mx1.redhat.com [209.132.183.28]) by dpdk.org (Postfix) with ESMTP id 89E691BEDA; Fri, 27 Sep 2019 20:38:21 +0200 (CEST) Received: from smtp.corp.redhat.com (int-mx06.intmail.prod.int.phx2.redhat.com [10.5.11.16]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id 02BC0308219E; Fri, 27 Sep 2019 18:38:21 +0000 (UTC) Received: from amorenoz.users.ipa.redhat.com (ovpn-116-96.ams2.redhat.com [10.36.116.96]) by smtp.corp.redhat.com (Postfix) with ESMTP id 7A2A75C21A; Fri, 27 Sep 2019 18:38:19 +0000 (UTC) From: Adrian Moreno To: dev@dpdk.org Cc: tiwei.bie@intel.com, zhihong.wang@intel.com, maxime.coquelin@redhat.com, Adrian Moreno , stable@dpdk.org Date: Fri, 27 Sep 2019 20:37:58 +0200 Message-Id: <20190927183759.18816-3-amorenoz@redhat.com> In-Reply-To: <20190927183759.18816-1-amorenoz@redhat.com> References: <20190927183759.18816-1-amorenoz@redhat.com> MIME-Version: 1.0 X-Scanned-By: MIMEDefang 2.79 on 10.5.11.16 X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.5.16 (mx1.redhat.com [10.5.110.47]); Fri, 27 Sep 2019 18:38:21 +0000 (UTC) Subject: [dpdk-dev] [PATCH v3 2/3] vhost: convert buffer addresses to GPA for logging X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org Sender: "dev" Note that the API facing rte_vhost_log_write is not modified. So, make explicit that it expects the address in GPA space. Fixes: 69c90e98f483 ("vhost: enable IOMMU support") Cc: maxime.coquelin@redhat.com Cc: stable@dpdk.org Signed-off-by: Adrian Moreno --- lib/librte_vhost/rte_vhost.h | 2 +- lib/librte_vhost/vdpa.c | 4 +++- lib/librte_vhost/vhost.c | 21 +++++++++++++++++++++ lib/librte_vhost/vhost.h | 16 ++++++++++++++++ lib/librte_vhost/virtio_net.c | 12 +++++++----- 5 files changed, 48 insertions(+), 7 deletions(-) diff --git a/lib/librte_vhost/rte_vhost.h b/lib/librte_vhost/rte_vhost.h index 7fb172912..16c7e4243 100644 --- a/lib/librte_vhost/rte_vhost.h +++ b/lib/librte_vhost/rte_vhost.h @@ -265,7 +265,7 @@ rte_vhost_va_from_guest_pa(struct rte_vhost_memory *mem, * @param vid * vhost device ID * @param addr - * the starting address for write + * the starting address for write (in guest physical address space) * @param len * the length to write */ diff --git a/lib/librte_vhost/vdpa.c b/lib/librte_vhost/vdpa.c index 8e45ce9f8..11fd5aede 100644 --- a/lib/librte_vhost/vdpa.c +++ b/lib/librte_vhost/vdpa.c @@ -201,7 +201,9 @@ rte_vdpa_relay_vring_used(int vid, uint16_t qid, void *vring_m) goto fail; desc = desc_ring[desc_id]; if (desc.flags & VRING_DESC_F_WRITE) - vhost_log_write(dev, desc.addr, desc.len); + vhost_log_cache_write_iova(dev, vq, + desc.addr, + desc.len); desc_id = desc.next; } while (desc.flags & VRING_DESC_F_NEXT); diff --git a/lib/librte_vhost/vhost.c b/lib/librte_vhost/vhost.c index e57dda22f..78457a77c 100644 --- a/lib/librte_vhost/vhost.c +++ b/lib/librte_vhost/vhost.c @@ -200,6 +200,27 @@ __vhost_log_cache_write(struct virtio_net *dev, struct vhost_virtqueue *vq, } } +void +__vhost_log_cache_write_iova(struct virtio_net *dev, struct vhost_virtqueue *vq, + uint64_t iova, uint64_t len) +{ + uint64_t hva, gpa, map_len; + map_len = len; + + hva = __vhost_iova_to_vva(dev, vq, iova, &map_len, VHOST_ACCESS_RW); + if (map_len != len) { + RTE_LOG(ERR, VHOST_CONFIG, + "Failed to write log for IOVA 0x%0lx." + "No IOTLB entry found\n", + iova); + return; + } + + gpa = hva_to_gpa(dev, hva, len); + if (gpa) + __vhost_log_cache_write(dev, vq, gpa, len); +} + void * vhost_alloc_copy_ind_table(struct virtio_net *dev, struct vhost_virtqueue *vq, uint64_t desc_addr, uint64_t desc_len) diff --git a/lib/librte_vhost/vhost.h b/lib/librte_vhost/vhost.h index 9d3883385..9fa70bc3c 100644 --- a/lib/librte_vhost/vhost.h +++ b/lib/librte_vhost/vhost.h @@ -353,6 +353,9 @@ desc_is_avail(struct vring_packed_desc *desc, bool wrap_counter) void __vhost_log_cache_write(struct virtio_net *dev, struct vhost_virtqueue *vq, uint64_t addr, uint64_t len); +void __vhost_log_cache_write_iova(struct virtio_net *dev, + struct vhost_virtqueue *vq, + uint64_t iova, uint64_t len); void __vhost_log_cache_sync(struct virtio_net *dev, struct vhost_virtqueue *vq); void __vhost_log_write(struct virtio_net *dev, uint64_t addr, uint64_t len); @@ -393,6 +396,19 @@ vhost_log_used_vring(struct virtio_net *dev, struct vhost_virtqueue *vq, vhost_log_write(dev, vq->log_guest_addr + offset, len); } +static __rte_always_inline void +vhost_log_cache_write_iova(struct virtio_net *dev, struct vhost_virtqueue *vq, + uint64_t iova, uint64_t len) +{ + if (likely(!(dev->features & (1ULL << VHOST_F_LOG_ALL)))) + return; + + if (dev->features & (1ULL << VIRTIO_F_IOMMU_PLATFORM)) + __vhost_log_cache_write_iova(dev, vq, iova, len); + else + __vhost_log_cache_write(dev, vq, iova, len); +} + /* Macros for printing using RTE_LOG */ #define RTE_LOGTYPE_VHOST_CONFIG RTE_LOGTYPE_USER1 #define RTE_LOGTYPE_VHOST_DATA RTE_LOGTYPE_USER1 diff --git a/lib/librte_vhost/virtio_net.c b/lib/librte_vhost/virtio_net.c index 5b85b832d..22961cafe 100644 --- a/lib/librte_vhost/virtio_net.c +++ b/lib/librte_vhost/virtio_net.c @@ -178,7 +178,8 @@ do_data_copy_enqueue(struct virtio_net *dev, struct vhost_virtqueue *vq) for (i = 0; i < count; i++) { rte_memcpy(elem[i].dst, elem[i].src, elem[i].len); - vhost_log_cache_write(dev, vq, elem[i].log_addr, elem[i].len); + vhost_log_cache_write_iova(dev, vq, elem[i].log_addr, + elem[i].len); PRINT_PACKET(dev, (uintptr_t)elem[i].dst, elem[i].len, 0); } @@ -633,7 +634,7 @@ copy_vnet_hdr_to_desc(struct virtio_net *dev, struct vhost_virtqueue *vq, PRINT_PACKET(dev, (uintptr_t)dst, (uint32_t)len, 0); - vhost_log_cache_write(dev, vq, + vhost_log_cache_write_iova(dev, vq, iova, len); remain -= len; @@ -733,7 +734,7 @@ copy_mbuf_to_desc(struct virtio_net *dev, struct vhost_virtqueue *vq, } else { PRINT_PACKET(dev, (uintptr_t)hdr_addr, dev->vhost_hlen, 0); - vhost_log_cache_write(dev, vq, + vhost_log_cache_write_iova(dev, vq, buf_vec[0].buf_iova, dev->vhost_hlen); } @@ -748,8 +749,9 @@ copy_mbuf_to_desc(struct virtio_net *dev, struct vhost_virtqueue *vq, rte_memcpy((void *)((uintptr_t)(buf_addr + buf_offset)), rte_pktmbuf_mtod_offset(m, void *, mbuf_offset), cpy_len); - vhost_log_cache_write(dev, vq, buf_iova + buf_offset, - cpy_len); + vhost_log_cache_write_iova(dev, vq, + buf_iova + buf_offset, + cpy_len); PRINT_PACKET(dev, (uintptr_t)(buf_addr + buf_offset), cpy_len, 0); } else { From patchwork Fri Sep 27 18:37:59 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Adrian Moreno X-Patchwork-Id: 60059 X-Patchwork-Delegate: maxime.coquelin@redhat.com Return-Path: X-Original-To: patchwork@dpdk.org Delivered-To: patchwork@dpdk.org Received: from [92.243.14.124] (localhost [127.0.0.1]) by dpdk.org (Postfix) with ESMTP id 7EEF41BF13; Fri, 27 Sep 2019 20:38:28 +0200 (CEST) Received: from mx1.redhat.com (mx1.redhat.com [209.132.183.28]) by dpdk.org (Postfix) with ESMTP id 693041BEFD; Fri, 27 Sep 2019 20:38:23 +0200 (CEST) Received: from smtp.corp.redhat.com (int-mx06.intmail.prod.int.phx2.redhat.com [10.5.11.16]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id D77C3210E; Fri, 27 Sep 2019 18:38:22 +0000 (UTC) Received: from amorenoz.users.ipa.redhat.com (ovpn-116-96.ams2.redhat.com [10.36.116.96]) by smtp.corp.redhat.com (Postfix) with ESMTP id 58D645C21A; Fri, 27 Sep 2019 18:38:21 +0000 (UTC) From: Adrian Moreno To: dev@dpdk.org Cc: tiwei.bie@intel.com, zhihong.wang@intel.com, maxime.coquelin@redhat.com, Adrian Moreno , stable@dpdk.org Date: Fri, 27 Sep 2019 20:37:59 +0200 Message-Id: <20190927183759.18816-4-amorenoz@redhat.com> In-Reply-To: <20190927183759.18816-1-amorenoz@redhat.com> References: <20190927183759.18816-1-amorenoz@redhat.com> MIME-Version: 1.0 X-Scanned-By: MIMEDefang 2.79 on 10.5.11.16 X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.6.2 (mx1.redhat.com [10.5.110.71]); Fri, 27 Sep 2019 18:38:22 +0000 (UTC) Subject: [dpdk-dev] [PATCH v3 3/3] vhost: prevent zero copy mode if iommu is on X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org Sender: "dev" The simltaneous use of dequeue_zero_copy and IOMMU is problematic. Not only because IOVA_VA mode is not supported but also because the potential invalidation of guest pages while the buffers are in use, is not handled. Prevent these two features to be enabled simultaneosly. Fixes: 69c90e98f483 ("vhost: enable IOMMU support") Cc: maxime.coquelin@redhat.com Cc: stable@dpdk.org Signed-off-by: Adrian Moreno Reviewed-by: Tiwei Bie --- lib/librte_vhost/socket.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/lib/librte_vhost/socket.c b/lib/librte_vhost/socket.c index 274988c4d..534a7ffe3 100644 --- a/lib/librte_vhost/socket.c +++ b/lib/librte_vhost/socket.c @@ -871,6 +871,14 @@ rte_vhost_driver_register(const char *path, uint64_t flags) } vsocket->dequeue_zero_copy = flags & RTE_VHOST_USER_DEQUEUE_ZERO_COPY; + if (vsocket->dequeue_zero_copy && + (flags & RTE_VHOST_USER_IOMMU_SUPPORT)) { + RTE_LOG(ERR, VHOST_CONFIG, + "error: enabling dequeue zero copy and IOMMU features " + "simultaneously is not supported\n"); + goto out_mutex; + } + /* * Set the supported features correctly for the builtin vhost-user * net driver.