From: Maxime Coquelin <maxime.coquelin@redhat.com>
To: Keiichi Watanabe <keiichiw@chromium.org>, dev@dpdk.org
Cc: dgreid@chromium.org, chirantan@chromium.org,
Chenbo Xia <chenbo.xia@intel.com>
Subject: Re: [dpdk-dev] [PATCH] examples/vhost_blk: check protocol feature before get inflight info
Date: Fri, 19 Mar 2021 10:16:23 +0100 [thread overview]
Message-ID: <63a57aa5-e409-7ca4-b0b6-2eb79dd7d55a@redhat.com> (raw)
In-Reply-To: <20210202091039.3353569-1-keiichiw@chromium.org>
Hi Keiichi,
On 2/2/21 10:10 AM, Keiichi Watanabe wrote:
> Avoid calling rte_vhost_get_vhost_ring_inflight() and
> rte_vhost_get_vring_base_from_inflight() when
> VHOST_USER_PROTOCOL_F_INFLIGHT_SHMFD is not set.
>
> Signed-off-by: Keiichi Watanabe <keiichiw@chromium.org>
> ---
> examples/vhost_blk/vhost_blk.c | 22 +++++++++++++++++-----
> lib/librte_vhost/rte_vhost.h | 13 +++++++++++++
> lib/librte_vhost/vhost.c | 14 ++++++++++++++
> 3 files changed, 44 insertions(+), 5 deletions(-)
>
> diff --git a/examples/vhost_blk/vhost_blk.c b/examples/vhost_blk/vhost_blk.c
> index 7ea60863d..9f3e157a0 100644
> --- a/examples/vhost_blk/vhost_blk.c
> +++ b/examples/vhost_blk/vhost_blk.c
> @@ -603,10 +603,10 @@ new_device(int vid)
> struct vhost_blk_ctrlr *ctrlr;
> struct vhost_blk_queue *vq;
> char path[PATH_MAX];
> - uint64_t features;
> + uint64_t features, protocol_features;
> pthread_t tid;
> int i, ret;
> - bool packed_ring;
> + bool packed_ring, inflight_shmfd;
>
> ret = rte_vhost_get_ifname(vid, path, PATH_MAX);
> if (ret) {
> @@ -631,6 +631,15 @@ new_device(int vid)
> }
> packed_ring = !!(features & (1ULL << VIRTIO_F_RING_PACKED));
>
> + ret = rte_vhost_get_negotiated_protocol_features(
> + vid, &protocol_features);
> + if (ret) {
> + fprintf(stderr,
> + "Failed to get the negotiated protocol features\n");
> + return -1;
> + }
> + inflight_shmfd = !!(features & (1ULL << VHOST_USER_PROTOCOL_F_INFLIGHT_SHMFD));
> +
> /* Disable Notifications and init last idx */
> for (i = 0; i < NUM_OF_BLK_QUEUES; i++) {
> vq = &ctrlr->queues[i];
> @@ -641,10 +650,13 @@ new_device(int vid)
> assert(rte_vhost_get_vring_base(ctrlr->vid, i,
> &vq->last_avail_idx,
> &vq->last_used_idx) == 0);
> - assert(rte_vhost_get_vhost_ring_inflight(ctrlr->vid, i,
> - &vq->inflight_ring) == 0);
>
> - if (packed_ring) {
> + if (inflight_shmfd)
> + assert(rte_vhost_get_vhost_ring_inflight(
> + ctrlr->vid, i,
> + &vq->inflight_ring) == 0);
> +
> + if (packed_ring && inflight_shmfd) {
> /* for the reconnection */
> assert(rte_vhost_get_vring_base_from_inflight(
> ctrlr->vid, i,
> diff --git a/lib/librte_vhost/rte_vhost.h b/lib/librte_vhost/rte_vhost.h
> index 010f16086..2744374af 100644
> --- a/lib/librte_vhost/rte_vhost.h
> +++ b/lib/librte_vhost/rte_vhost.h
> @@ -567,6 +567,19 @@ rte_vhost_driver_get_queue_num(const char *path, uint32_t *queue_num);
> */
> int rte_vhost_get_negotiated_features(int vid, uint64_t *features);
>
> +/**
> + * Get the protocol feature bits after negotiation
> + *
> + * @param vid
> + * Vhost device ID
> + * @param protocol_features
> + * A pointer to store the queried protocol feature bits
> + * @return
> + * 0 on success, -1 on failure
> + */
> +int rte_vhost_get_negotiated_protocol_features(int vid,
> + uint64_t *protocol_features);
> +
> /* Register callbacks. */
> int rte_vhost_driver_callback_register(const char *path,
> struct vhost_device_ops const * const ops);
> diff --git a/lib/librte_vhost/vhost.c b/lib/librte_vhost/vhost.c
> index efb136edd..61189ce93 100644
> --- a/lib/librte_vhost/vhost.c
> +++ b/lib/librte_vhost/vhost.c
> @@ -868,6 +868,20 @@ rte_vhost_get_negotiated_features(int vid, uint64_t *features)
> return 0;
> }
>
> +int
> +rte_vhost_get_negotiated_protocol_features(int vid,
> + uint64_t *protocol_features)
> +{
> + struct virtio_net *dev;
> +
> + dev = get_device(vid);
> + if (dev == NULL || protocol_features == NULL)
> + return -1;
> +
> + *protocol_features = dev->protocol_features;
> + return 0;
> +}
> +
> int
> rte_vhost_get_mem_table(int vid, struct rte_vhost_memory **mem)
> {
>
The Vhost lib part should be in a dedicated patch and new API should be
added in its version.map as experimental.
Thanks,
Maxime
prev parent reply other threads:[~2021-03-19 9:16 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-02-02 9:10 Keiichi Watanabe
2021-03-19 9:16 ` Maxime Coquelin [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=63a57aa5-e409-7ca4-b0b6-2eb79dd7d55a@redhat.com \
--to=maxime.coquelin@redhat.com \
--cc=chenbo.xia@intel.com \
--cc=chirantan@chromium.org \
--cc=dev@dpdk.org \
--cc=dgreid@chromium.org \
--cc=keiichiw@chromium.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).