From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mx1.redhat.com (mx1.redhat.com [209.132.183.28]) by dpdk.org (Postfix) with ESMTP id 68F71107A; Thu, 14 Dec 2017 15:21:46 +0100 (CET) Received: from smtp.corp.redhat.com (int-mx04.intmail.prod.int.phx2.redhat.com [10.5.11.14]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id BDE6CC058EAF; Thu, 14 Dec 2017 14:21:45 +0000 (UTC) Received: from colo-mx.corp.redhat.com (colo-mx02.intmail.prod.int.phx2.redhat.com [10.5.11.21]) by smtp.corp.redhat.com (Postfix) with ESMTPS id 9F81E6C94D; Thu, 14 Dec 2017 14:21:45 +0000 (UTC) Received: from zmail17.collab.prod.int.phx2.redhat.com (zmail17.collab.prod.int.phx2.redhat.com [10.5.83.19]) by colo-mx.corp.redhat.com (Postfix) with ESMTP id 2B8A64BB78; Thu, 14 Dec 2017 14:21:44 +0000 (UTC) Date: Thu, 14 Dec 2017 09:21:43 -0500 (EST) From: Victor Kaplansky To: Maxime Coquelin Cc: dev@dpdk.org, stable@dpdk.org, Jens Freimann , Yuanhan Liu , Tiwei Bie , Jianfeng Tan Message-ID: <1818102083.54728632.1513261303965.JavaMail.zimbra@redhat.com> In-Reply-To: <9742b405-433c-edf7-3cb5-084521a4a569@redhat.com> References: <20171214133531-mutt-send-email-victork@redhat.com> <9742b405-433c-edf7-3cb5-084521a4a569@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit X-Originating-IP: [10.35.206.49, 10.4.195.12] Thread-Topic: vhost_user: protect active rings from async ring changes Thread-Index: jdwZSG762zRuMXiv3wO+vGgtvj5j3A== X-Scanned-By: MIMEDefang 2.79 on 10.5.11.14 X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.5.16 (mx1.redhat.com [10.5.110.32]); Thu, 14 Dec 2017 14:21:45 +0000 (UTC) X-Mailman-Approved-At: Fri, 15 Dec 2017 11:33:10 +0100 Subject: Re: [dpdk-stable] [PATCH v3] vhost_user: protect active rings from async ring changes X-BeenThere: stable@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: patches for DPDK stable branches List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 14 Dec 2017 14:21:46 -0000 ----- Original Message ----- > From: "Maxime Coquelin" > To: "Victor Kaplansky" , dev@dpdk.org > Cc: stable@dpdk.org, "Jens Freimann" , "Yuanhan Liu" , "Tiwei Bie" > , "Jianfeng Tan" > Sent: Thursday, December 14, 2017 2:16:23 PM > Subject: Re: [PATCH v3] vhost_user: protect active rings from async ring changes > > Hi Victor, > > On 12/14/2017 12:35 PM, Victor Kaplansky wrote: > > When performing live migration or memory hot-plugging, > > the changes to the device and vrings made by message handler > > done independently from vring usage by PMD threads. > > > > This causes for example segfaults during live-migration > > with MQ enable, but in general virtually any request > > sent by qemu changing the state of device can cause > > problems. > > > > These patches fixes all above issues by adding a spinlock > > to every vring and requiring message handler to start operation > > only after ensuring that all PMD threads related to the device > > are out of critical section accessing the vring data. > > > > Each vring has its own lock in order to not create contention > > between PMD threads of different vrings and to prevent > > performance degradation by scaling queue pair number. > > > > See https://bugzilla.redhat.com/show_bug.cgi?id=1450680 > > > > Signed-off-by: Victor Kaplansky > > Tested-by: Maxime Coquelin > > Sorry, but I didn't tested this patch. I just benchmarked the fact to > add a lock in the hot paths. > > > --- > > > > v3: > > o Added locking to enqueue flow. > > o Enqueue path guarded as well as dequeue path. > > o Changed name of active_lock. > > o Added initialization of guarding spinlock. > > o Reworked functions skimming over all virt-queues. > > o Performance measurements done by Maxime Coquelin shows > > no degradation in bandwidth and throughput. > > o Spelling. > > o Taking lock only on set operations. > > o IOMMU messages are not guarded by access lock. > > > > v2: > > o Fixed checkpatch complains. > > o Added Signed-off-by. > > o Refined placement of guard to exclude IOMMU messages. > > o TODO: performance degradation measurement. > > > > > > lib/librte_vhost/vhost.h | 15 ++++++++++ > > lib/librte_vhost/vhost.c | 1 + > > lib/librte_vhost/vhost_user.c | 65 > > +++++++++++++++++++++++++++++++++++++++++++ > > lib/librte_vhost/virtio_net.c | 20 +++++++++++-- > > 4 files changed, 99 insertions(+), 2 deletions(-) > > > > diff --git a/lib/librte_vhost/vhost.h b/lib/librte_vhost/vhost.h > > index 1cc81c17..26e2c571 100644 > > --- a/lib/librte_vhost/vhost.h > > +++ b/lib/librte_vhost/vhost.h > > @@ -137,6 +137,8 @@ struct vhost_virtqueue { > > TAILQ_HEAD(, vhost_iotlb_entry) iotlb_list; > > int iotlb_cache_nr; > > TAILQ_HEAD(, vhost_iotlb_entry) iotlb_pending_list; > > + > > + rte_spinlock_t access_lock; > > On previous revision of the patch, Jianfeng mentioned taking the lock > has an impact when nothing to enqueue/dequeue (80 cycles vs. 50 cycles > IIRC). I wonder whether moving the lock closer to enabled field would > make it be in same cache line and so might improve performance. Good point, I'll move it closer to enable. > > + switch (msg.request.master) { > > + case VHOST_USER_SET_FEATURES: > > + case VHOST_USER_SET_PROTOCOL_FEATURES: > > + case VHOST_USER_SET_OWNER: > > + case VHOST_USER_RESET_OWNER: > > + case VHOST_USER_SET_MEM_TABLE: > > + case VHOST_USER_SET_LOG_BASE: > > + case VHOST_USER_SET_LOG_FD: > > + case VHOST_USER_SET_VRING_NUM: > > + case VHOST_USER_SET_VRING_ADDR: > > + case VHOST_USER_SET_VRING_BASE: > > + case VHOST_USER_SET_VRING_KICK: > > + case VHOST_USER_SET_VRING_CALL: > > + case VHOST_USER_SET_VRING_ERR: > > + case VHOST_USER_SET_VRING_ENABLE: > > For the VRING specific requests, I think we could only take the > corresponding VQ lock instead of blocking all queues. > > There might be some exceptions, like GET_VRING_BASE. > > Maybe only protecting critical sections would be better? I agree in general, but requests are so rare, that I prefer to have a robust solution instead of dealing with questionable optimizations prone to bugs. Anyway, we can add more fine tuning later on case by case basis, as we become aware of more performance bottlenecks. > > > > #include "iotlb.h" > > #include "vhost.h" > > @@ -326,8 +327,11 @@ virtio_dev_rx(struct virtio_net *dev, uint16_t > > queue_id, > > } > > > > vq = dev->virtqueue[queue_id]; > > + > > + vhost_user_access_lock(vq); > I suggested to use rte_spinlock_trylock() when reviewing previous > revision. It would have the advantage to not block other devices being > handled on the same CPU whlie handling the message. > > Any thoughts? If we want trylock, we have two options in this case. One is just return zero if trylock fails and give the caller to decide what to do: retry, issue an error, or yield the cpu to another thread, etc. I don't think any caller of rte_eth_tx_burst in current dpdk source doing this. Most frequently it is just considered as an error. The second option is to invoke sched_yield() yielding the cpu to other threads. I'm in favor of second option, but looking into current dpdk code base I don't see any instances of using sched_yield(). Maybe it is for a reason? > > > + > > if (unlikely(vq->enabled == 0)) > > - return 0; > > + goto out; > > > > if (dev->features & (1ULL << VIRTIO_F_IOMMU_PLATFORM)) > > vhost_user_iotlb_rd_lock(vq); > > @@ -416,6 +420,7 @@ virtio_dev_rx(struct virtio_net *dev, uint16_t > > queue_id, > > && (vq->callfd >= 0)) > > eventfd_write(vq->callfd, (eventfd_t)1); > > out: > > + vhost_user_access_unlock(vq); > > I think it should be moved after the iotlb lock unlock: > LOCK A > LOCK B > UNLOCK B > UNLOCK A Yep, good catch. > > > if (dev->features & (1ULL << VIRTIO_F_IOMMU_PLATFORM)) > > vhost_user_iotlb_rd_unlock(vq); > > > > @@ -651,8 +656,11 @@ virtio_dev_merge_rx(struct virtio_net *dev, uint16_t > > queue_id, > > } > > > > vq = dev->virtqueue[queue_id]; > > + > > + vhost_user_access_lock(vq); > > + > > if (unlikely(vq->enabled == 0)) > > - return 0; > > + goto out; > > > > if (dev->features & (1ULL << VIRTIO_F_IOMMU_PLATFORM)) > > vhost_user_iotlb_rd_lock(vq); > > @@ -712,6 +720,7 @@ virtio_dev_merge_rx(struct virtio_net *dev, uint16_t > > queue_id, > > } > > > > out: > > + vhost_user_access_unlock(vq); > > Ditto > > > if (dev->features & (1ULL << VIRTIO_F_IOMMU_PLATFORM)) > > vhost_user_iotlb_rd_unlock(vq); > > > > @@ -1180,9 +1189,12 @@ rte_vhost_dequeue_burst(int vid, uint16_t queue_id, > > } > > > > vq = dev->virtqueue[queue_id]; > > + > > if (unlikely(vq->enabled == 0)) > > return 0; > > > > + vhost_user_access_lock(vq); > > + > > To be consistent with other paths, I guess you should lock before > checking the vq is enabled. Oops. Missed this one, thanks. > > > vq->batch_copy_nb_elems = 0; > > > > if (dev->features & (1ULL << VIRTIO_F_IOMMU_PLATFORM)) > > @@ -1240,6 +1252,7 @@ rte_vhost_dequeue_burst(int vid, uint16_t queue_id, > > if (rarp_mbuf == NULL) { > > RTE_LOG(ERR, VHOST_DATA, > > "Failed to allocate memory for mbuf.\n"); > > + vhost_user_access_unlock(vq); > > return 0; > > } > > > > @@ -1356,6 +1369,8 @@ rte_vhost_dequeue_burst(int vid, uint16_t queue_id, > > if (dev->features & (1ULL << VIRTIO_F_IOMMU_PLATFORM)) > > vhost_user_iotlb_rd_unlock(vq); > > > > + vhost_user_access_unlock(vq); > > + > > if (unlikely(rarp_mbuf != NULL)) { > > /* > > * Inject it to the head of "pkts" array, so that switch's mac > > @@ -1366,5 +1381,6 @@ rte_vhost_dequeue_burst(int vid, uint16_t queue_id, > > i += 1; > > } > > > > + > > Remove trailing line. > > > return i; > > } > > > > Thanks, > Maxime >