From mboxrd@z Thu Jan 1 00:00:00 1970 From: Tetsuya Mukawa Subject: Re: [PATCH 1/3] vhost: Fix wrong handling of virtqueue array index Date: Tue, 27 Oct 2015 16:28:58 +0900 Message-ID: <562F27BA.5060501@igel.co.jp> References: <1445507150-1481-2-git-send-email-mukawa@igel.co.jp> <1445926375-18986-1-git-send-email-mukawa@igel.co.jp> <1445926375-18986-2-git-send-email-mukawa@igel.co.jp> <20151027064720.GD3115@yliu-dev.sh.intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset=windows-1252 Content-Transfer-Encoding: 7bit Cc: dev@dpdk.org, ann.zhuangyanying@huawei.com To: Yuanhan Liu Return-path: Received: from mail-pa0-f53.google.com (mail-pa0-f53.google.com [209.85.220.53]) by dpdk.org (Postfix) with ESMTP id 2260A6A80 for ; Tue, 27 Oct 2015 08:29:02 +0100 (CET) Received: by pasz6 with SMTP id z6so214018103pas.2 for ; Tue, 27 Oct 2015 00:29:01 -0700 (PDT) In-Reply-To: <20151027064720.GD3115@yliu-dev.sh.intel.com> List-Id: patches and discussions about DPDK List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org Sender: "dev" Hi Yuanhan, I appreciate your checking. I haven't noticed SET_BACKEND is only supported by vhost-cuse. :-( I will follow your comments, then submit again. Thanks, Tetsuya On 2015/10/27 15:47, Yuanhan Liu wrote: > On Tue, Oct 27, 2015 at 03:12:53PM +0900, Tetsuya Mukawa wrote: >> The patch fixes wrong handling of virtqueue array index. >> >> GET_VRING_BASE: >> The vhost backend will receive the message per virtqueue. >> Also we should call a destroy callback when both RXQ and TXQ receives >> the message. >> >> SET_BACKEND: >> Because vhost library supports multiple queue, the index may be over 2. >> Also a vhost frontend(QEMU) may send such a index. > Note that only vhost-user supports MQ. vhost-cuse does not. > >> Signed-off-by: Tetsuya Mukawa >> --- >> lib/librte_vhost/vhost_user/virtio-net-user.c | 22 +++++++++++----------- >> lib/librte_vhost/virtio-net.c | 5 +++-- >> 2 files changed, 14 insertions(+), 13 deletions(-) >> >> diff --git a/lib/librte_vhost/vhost_user/virtio-net-user.c b/lib/librte_vhost/vhost_user/virtio-net-user.c >> index a998ad8..3e8dfea 100644 >> --- a/lib/librte_vhost/vhost_user/virtio-net-user.c >> +++ b/lib/librte_vhost/vhost_user/virtio-net-user.c >> @@ -283,12 +283,10 @@ user_get_vring_base(struct vhost_device_ctx ctx, >> struct vhost_vring_state *state) >> { >> struct virtio_net *dev = get_device(ctx); >> + uint16_t base_idx = state->index / VIRTIO_QNUM; > So, fixing what my 1st reply said, for Nth queue pair, state->index > is "N * 2 + is_tx". So, the base should be "state->index / 2 * 2". > >> >> if (dev == NULL) >> return -1; >> - /* We have to stop the queue (virtio) if it is running. */ >> - if (dev->flags & VIRTIO_DEV_RUNNING) >> - notify_ops->destroy_device(dev); >> >> /* Here we are safe to get the last used index */ >> ops->get_vring_base(ctx, state->index, state); >> @@ -300,15 +298,17 @@ user_get_vring_base(struct vhost_device_ctx ctx, >> * sent and only sent in vhost_vring_stop. >> * TODO: cleanup the vring, it isn't usable since here. >> */ >> - if (dev->virtqueue[state->index + VIRTIO_RXQ]->kickfd >= 0) { >> - close(dev->virtqueue[state->index + VIRTIO_RXQ]->kickfd); >> - dev->virtqueue[state->index + VIRTIO_RXQ]->kickfd = -1; >> - } >> - if (dev->virtqueue[state->index + VIRTIO_TXQ]->kickfd >= 0) { >> - close(dev->virtqueue[state->index + VIRTIO_TXQ]->kickfd); >> - dev->virtqueue[state->index + VIRTIO_TXQ]->kickfd = -1; >> + if (dev->virtqueue[state->index]->kickfd >= 0) { >> + close(dev->virtqueue[state->index]->kickfd); >> + dev->virtqueue[state->index]->kickfd = -1; >> } >> >> + /* We have to stop the queue (virtio) if it is running. */ >> + if ((dev->flags & VIRTIO_DEV_RUNNING) && >> + (dev->virtqueue[base_idx + VIRTIO_RXQ]->kickfd == -1) && >> + (dev->virtqueue[base_idx + VIRTIO_TXQ]->kickfd == -1)) >> + notify_ops->destroy_device(dev); > This is a proper fix then. (You just need fix base_idx). > >> return 0; >> } >> >> @@ -321,7 +321,7 @@ user_set_vring_enable(struct vhost_device_ctx ctx, >> struct vhost_vring_state *state) >> { >> struct virtio_net *dev = get_device(ctx); >> - uint16_t base_idx = state->index; >> + uint16_t base_idx = state->index / VIRTIO_QNUM; > user_set_vring_enable is sent per queue pair (I'm sure this time), so > base_idx equals to state->index. No need fix here. > >> int enable = (int)state->num; >> >> RTE_LOG(INFO, VHOST_CONFIG, >> diff --git a/lib/librte_vhost/virtio-net.c b/lib/librte_vhost/virtio-net.c >> index 97213c5..ee2e84d 100644 >> --- a/lib/librte_vhost/virtio-net.c >> +++ b/lib/librte_vhost/virtio-net.c >> @@ -778,6 +778,7 @@ static int >> set_backend(struct vhost_device_ctx ctx, struct vhost_vring_file *file) >> { >> struct virtio_net *dev; >> + uint32_t base_idx = file->index / VIRTIO_QNUM; > As stated, vhost-cuse doesn't not support MQ. > > --yliu >> >> dev = get_device(ctx); >> if (dev == NULL) >> @@ -791,8 +792,8 @@ set_backend(struct vhost_device_ctx ctx, struct vhost_vring_file *file) >> * we add the device. >> */ >> if (!(dev->flags & VIRTIO_DEV_RUNNING)) { >> - if (((int)dev->virtqueue[VIRTIO_TXQ]->backend != VIRTIO_DEV_STOPPED) && >> - ((int)dev->virtqueue[VIRTIO_RXQ]->backend != VIRTIO_DEV_STOPPED)) { >> + if (((int)dev->virtqueue[base_idx + VIRTIO_RXQ]->backend != VIRTIO_DEV_STOPPED) && >> + ((int)dev->virtqueue[base_idx + VIRTIO_TXQ]->backend != VIRTIO_DEV_STOPPED)) { >> return notify_ops->new_device(dev); >> } >> /* Otherwise we remove it. */ >> -- >> 2.1.4