From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mga07.intel.com (mga07.intel.com [134.134.136.100]) by dpdk.org (Postfix) with ESMTP id 46F791B161 for ; Tue, 9 Jan 2018 08:34:21 +0100 (CET) X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False Received: from orsmga001.jf.intel.com ([10.7.209.18]) by orsmga105.jf.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 08 Jan 2018 23:34:20 -0800 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.46,334,1511856000"; d="scan'208";a="22542676" Received: from fmsmsx107.amr.corp.intel.com ([10.18.124.205]) by orsmga001.jf.intel.com with ESMTP; 08 Jan 2018 23:34:19 -0800 Received: from shsmsx104.ccr.corp.intel.com (10.239.4.70) by fmsmsx107.amr.corp.intel.com (10.18.124.205) with Microsoft SMTP Server (TLS) id 14.3.319.2; Mon, 8 Jan 2018 23:34:19 -0800 Received: from shsmsx102.ccr.corp.intel.com ([169.254.2.189]) by SHSMSX104.ccr.corp.intel.com ([169.254.5.152]) with mapi id 14.03.0319.002; Tue, 9 Jan 2018 15:34:17 +0800 From: "Yao, Lei A" To: "Chen, Junjie J" , "yliu@fridaylinux.org" , "maxime.coquelin@redhat.com" , "Wang, Xiao W" , "Bie, Tiwei" CC: "dev@dpdk.org" Thread-Topic: [PATCH v7] vhost: support virtqueue interrupt/notification suppression Thread-Index: AQHTiPk5Qp4C/v0my0uUSqmeHsqVTqNrJTmQ Date: Tue, 9 Jan 2018 07:34:17 +0000 Message-ID: <2DBBFF226F7CF64BAFCA79B681719D953A3220A3@shsmsx102.ccr.corp.intel.com> References: <1514310190-140916-1-git-send-email-junjie.j.chen@intel.com> <1515495828-190409-1-git-send-email-junjie.j.chen@intel.com> In-Reply-To: <1515495828-190409-1-git-send-email-junjie.j.chen@intel.com> Accept-Language: en-US Content-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: x-titus-metadata-40: eyJDYXRlZ29yeUxhYmVscyI6IiIsIk1ldGFkYXRhIjp7Im5zIjoiaHR0cDpcL1wvd3d3LnRpdHVzLmNvbVwvbnNcL0ludGVsMyIsImlkIjoiMWE2YjYwM2YtNzEwYS00OGM5LWE5ZDUtNGY2MDVmMjE3YTc2IiwicHJvcHMiOlt7Im4iOiJDVFBDbGFzc2lmaWNhdGlvbiIsInZhbHMiOlt7InZhbHVlIjoiQ1RQX05UIn1dfV19LCJTdWJqZWN0TGFiZWxzIjpbXSwiVE1DVmVyc2lvbiI6IjE2LjUuOS4zIiwiVHJ1c3RlZExhYmVsSGFzaCI6Ik0yVGhFOVI2Tk1KWlhueEpXTE1ZYjA5OHBtTU5aZm9QcU5wYURyT0ZjOHc9In0= x-ctpclassification: CTP_NT dlp-product: dlpe-windows dlp-version: 11.0.0.116 dlp-reaction: no-action x-originating-ip: [10.239.127.40] Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 Subject: Re: [dpdk-dev] [PATCH v7] vhost: support virtqueue interrupt/notification suppression 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: , X-List-Received-Date: Tue, 09 Jan 2018 07:34:21 -0000 > -----Original Message----- > From: Chen, Junjie J > Sent: Tuesday, January 9, 2018 7:04 PM > To: yliu@fridaylinux.org; maxime.coquelin@redhat.com; Wang, Xiao W > ; Bie, Tiwei ; Yao, Lei A > > Cc: dev@dpdk.org; Chen, Junjie J > Subject: [PATCH v7] vhost: support virtqueue interrupt/notification > suppression >=20 > The driver can suppress interrupt when VIRTIO_F_EVENT_IDX feature bit is > negotiated. The driver set vring flags to 0, and MAY use used_event in > available ring to advise device interrupt util reach an index specified > by used_event. The device ignore the lower bit of vring flags, and send > an interrupt when index reach used_event. >=20 > The device can suppress notification in a manner analogous to the ways > driver suppress interrupt. The device manipulates flags or avail_event in > the used ring in the same way the driver manipulates flags or used_event = in > available ring. >=20 > Signed-off-by: Junjie Chen Tested-by: Lei Yao VM2VM Iperf test has been executed with virtio-net driver after apply this = patch. No performance drop.=20 CPU: Intel(R) Xeon(R) CPU E5-2699 v4 @ 2.20GHz Host OS: Ubuntu 16.04 Guest OS: Ubuntu 16.04 Kernel: 4.4.0 > --- > v7: > Add vhost_need_event definition and update code for next virtio. >=20 > v6: > Use volatile qualifier to access avail event idx. >=20 > v5: > Remove updating avail event index in backend. >=20 > v2-v4: > Use definition of VIRTIO_F_EVENT_IDX from kernel. >=20 > lib/librte_vhost/vhost.c | 2 +- > lib/librte_vhost/vhost.h | 44 > ++++++++++++++++++++++++++++++++++++++----- > lib/librte_vhost/virtio_net.c | 6 +++--- > 3 files changed, 43 insertions(+), 9 deletions(-) >=20 > diff --git a/lib/librte_vhost/vhost.c b/lib/librte_vhost/vhost.c > index 6f7ef7f..400caa0 100644 > --- a/lib/librte_vhost/vhost.c > +++ b/lib/librte_vhost/vhost.c > @@ -538,7 +538,7 @@ rte_vhost_vring_call(int vid, uint16_t vring_idx) > if (!vq) > return -1; >=20 > - vhost_vring_call(vq); > + vhost_vring_call(dev, vq); > return 0; > } >=20 > diff --git a/lib/librte_vhost/vhost.h b/lib/librte_vhost/vhost.h > index 1d9366e..ec79991 100644 > --- a/lib/librte_vhost/vhost.h > +++ b/lib/librte_vhost/vhost.h > @@ -103,6 +103,8 @@ struct vhost_virtqueue { >=20 > uint16_t last_avail_idx; > uint16_t last_used_idx; > + /* Last used index we notify to front end. */ > + uint16_t signalled_used; > #define VIRTIO_INVALID_EVENTFD (-1) > #define VIRTIO_UNINITIALIZED_EVENTFD (-2) >=20 > @@ -214,6 +216,7 @@ struct vhost_msg { > (1ULL << VIRTIO_NET_F_GUEST_TSO6) | \ > (1ULL << VIRTIO_NET_F_GUEST_UFO) | \ > (1ULL << VIRTIO_RING_F_INDIRECT_DESC) | > \ > + (1ULL << VIRTIO_RING_F_EVENT_IDX) | \ > (1ULL << VIRTIO_NET_F_MTU) | \ > (1ULL << VIRTIO_F_IOMMU_PLATFORM)) >=20 > @@ -399,16 +402,47 @@ vhost_iova_to_vva(struct virtio_net *dev, struct > vhost_virtqueue *vq, > return __vhost_iova_to_vva(dev, vq, iova, size, perm); > } >=20 > +#define vhost_used_event(vr) \ > + (*(volatile uint16_t*)&(vr)->avail->ring[(vr)->size]) > + > +/* > + * The following is used with VIRTIO_RING_F_EVENT_IDX. > + * Assuming a given event_idx value from the other size, if we have > + * just incremented index from old to new_idx, should we trigger an > + * event? > + */ > +static __rte_always_inline int > +vhost_need_event(uint16_t event_idx, uint16_t new_idx, uint16_t old) > +{ > + return (uint16_t)(new_idx - event_idx - 1) < (uint16_t)(new_idx - > old); > +} > + > static __rte_always_inline void > -vhost_vring_call(struct vhost_virtqueue *vq) > +vhost_vring_call(struct virtio_net *dev, struct vhost_virtqueue *vq) > { > /* Flush used->idx update before we read avail->flags. */ > rte_mb(); >=20 > - /* Kick the guest if necessary. */ > - if (!(vq->avail->flags & VRING_AVAIL_F_NO_INTERRUPT) > - && (vq->callfd >=3D 0)) > - eventfd_write(vq->callfd, (eventfd_t)1); > + /* Don't kick guest if we don't reach index specified by guest. */ > + if (dev->features & (1ULL << VIRTIO_RING_F_EVENT_IDX)) { > + uint16_t old =3D vq->signalled_used; > + uint16_t new =3D vq->last_used_idx; > + > + LOG_DEBUG(VHOST_DATA, "%s: used_event_idx=3D%d, > old=3D%d, new=3D%d\n", > + __func__, > + vhost_used_event(vq), > + old, new); > + if (vhost_need_event(vhost_used_event(vq), new, old) > + && (vq->callfd >=3D 0)) { > + vq->signalled_used =3D vq->last_used_idx; > + eventfd_write(vq->callfd, (eventfd_t) 1); > + } > + } else { > + /* Kick the guest if necessary. */ > + if (!(vq->avail->flags & VRING_AVAIL_F_NO_INTERRUPT) > + && (vq->callfd >=3D 0)) > + eventfd_write(vq->callfd, (eventfd_t)1); > + } > } >=20 > #endif /* _VHOST_NET_CDEV_H_ */ > diff --git a/lib/librte_vhost/virtio_net.c b/lib/librte_vhost/virtio_net.= c > index eb496e1..d0f9221 100644 > --- a/lib/librte_vhost/virtio_net.c > +++ b/lib/librte_vhost/virtio_net.c > @@ -413,7 +413,7 @@ virtio_dev_rx(struct virtio_net *dev, uint16_t > queue_id, > offsetof(struct vring_used, idx), > sizeof(vq->used->idx)); >=20 > - vhost_vring_call(vq); > + vhost_vring_call(dev, vq); > out: > if (dev->features & (1ULL << VIRTIO_F_IOMMU_PLATFORM)) > vhost_user_iotlb_rd_unlock(vq); > @@ -700,7 +700,7 @@ virtio_dev_merge_rx(struct virtio_net *dev, uint16_t > queue_id, >=20 > if (likely(vq->shadow_used_idx)) { > flush_shadow_used_ring(dev, vq); > - vhost_vring_call(vq); > + vhost_vring_call(dev, vq); > } >=20 > out: > @@ -1104,7 +1104,7 @@ update_used_idx(struct virtio_net *dev, struct > vhost_virtqueue *vq, > vq->used->idx +=3D count; > vhost_log_used_vring(dev, vq, offsetof(struct vring_used, idx), > sizeof(vq->used->idx)); > - vhost_vring_call(vq); > + vhost_vring_call(dev, vq); > } >=20 > static __rte_always_inline struct zcopy_mbuf * > -- > 2.0.1