DPDK patches and discussions
 help / color / mirror / Atom feed
From: David Marchand <david.marchand@redhat.com>
To: Maxime Coquelin <maxime.coquelin@redhat.com>
Cc: dev@dpdk.org, chenbox@nvidia.com
Subject: Re: [PATCH 3/6] vhost: fix VDUSE reconnect device start failure
Date: Wed, 23 Oct 2024 18:06:54 +0200	[thread overview]
Message-ID: <CAJFAV8x5hU=V56BmAWOQ7hbMGnHdUStR9Yc2E5rjV77j9sDjTQ@mail.gmail.com> (raw)
In-Reply-To: <20241023151552.2863387-4-maxime.coquelin@redhat.com>

On Wed, Oct 23, 2024 at 5:16 PM Maxime Coquelin
<maxime.coquelin@redhat.com> wrote:
>
> This patch fixes a FD leak in the VDUSE device reconnect
> code fails to start the device.
>
> Also take the opportunity to refactor the related code into
> a dedicated function.
>
> Fixes: da79cc7fda76 ("vhost: add reconnection support to VDUSE")
>
> Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
> ---
>  lib/vhost/vduse.c | 63 +++++++++++++++++++++++++++++++----------------
>  1 file changed, 42 insertions(+), 21 deletions(-)
>
> diff --git a/lib/vhost/vduse.c b/lib/vhost/vduse.c
> index d1373d0549..c8c4761293 100644
> --- a/lib/vhost/vduse.c
> +++ b/lib/vhost/vduse.c
> @@ -476,6 +476,46 @@ vduse_reconnect_handler(int fd, void *arg, int *remove)
>         *remove = 1;
>  }
>
> +static int
> +vduse_reconnect_start_device(struct virtio_net *dev)
> +{
> +       int fd, ret;
> +
> +       /*
> +        * Make vduse_device_start() being executed in the same
> +        * context for both reconnection and fresh startup.
> +        */
> +       fd = eventfd(0, EFD_NONBLOCK | EFD_CLOEXEC);
> +       if (fd < 0) {
> +               VHOST_CONFIG_LOG(dev->ifname, ERR, "Failed to create reconnect efd: %s",
> +                               strerror(errno));
> +               ret = -1;
> +               goto out_err;
> +       }
> +
> +       ret = fdset_add(vduse.fdset, fd, vduse_reconnect_handler, NULL, dev);
> +       if (ret) {
> +               VHOST_CONFIG_LOG(dev->ifname, ERR, "Failed to add reconnect efd %d to vduse fdset",
> +                               fd);
> +               goto out_err_close;
> +       }
> +
> +       ret = eventfd_write(fd, (eventfd_t)1);
> +       if (ret < 0) {
> +               VHOST_CONFIG_LOG(dev->ifname, ERR, "Failed to write to reconnect eventfd");
> +               goto out_err_fdset;
> +       }
> +
> +       return 0;
> +
> +out_err_fdset:
> +       fdset_del(vduse.fdset, fd);
> +out_err_close:
> +       close(fd);
> +out_err:
> +       return ret;
> +}
> +
>  int
>  vduse_device_create(const char *path, bool compliant_ol_flags)
>  {
> @@ -741,28 +781,9 @@ vduse_device_create(const char *path, bool compliant_ol_flags)
>         }
>
>         if (reconnect && dev->status & VIRTIO_DEVICE_STATUS_DRIVER_OK)  {
> -               /*
> -                * Make vduse_device_start() being executed in the same
> -                * context for both reconnection and fresh startup.
> -                */
> -               reco_fd = eventfd(0, EFD_NONBLOCK | EFD_CLOEXEC);
> -               if (reco_fd < 0) {
> -                       VHOST_CONFIG_LOG(name, ERR, "Failed to create reco_fd: %s",
> -                                       strerror(errno));
> -                       ret = -1;
> -                       goto out_dev_destroy;
> -               }
> -
> -               ret = fdset_add(vduse.fdset, reco_fd, vduse_reconnect_handler, NULL, dev);
> +               ret = vduse_reconnect_start_device(dev);
>                 if (ret) {
> -                       VHOST_CONFIG_LOG(name, ERR, "Failed to add reconnect fd %d to vduse fdset",
> -                                       reco_fd);
> -                       goto out_dev_destroy;
> -               }
> -
> -               ret = eventfd_write(reco_fd, (eventfd_t)1);
> -               if (ret < 0) {
> -                       VHOST_CONFIG_LOG(name, ERR, "Failed to write to reconnect eventfd");
> +                       VHOST_CONFIG_LOG(name, ERR, "Failed to start device at reconnection");

We don't need a new log, there is already one for each error case in
vduse_reconnect_start_device().


-- 
David Marchand


  reply	other threads:[~2024-10-23 16:07 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-10-23 15:15 [PATCH 0/6] VDUSE reconnection fixes and cleanup Maxime Coquelin
2024-10-23 15:15 ` [PATCH 1/6] vhost: fix vduse device creation error handling Maxime Coquelin
2024-10-23 16:06   ` David Marchand
2024-10-23 18:55     ` Maxime Coquelin
2024-10-23 15:15 ` [PATCH 2/6] vhost: fix possible TOCTOU in VDUSE dev creation Maxime Coquelin
2024-10-23 16:06   ` David Marchand
2024-10-23 15:15 ` [PATCH 3/6] vhost: fix VDUSE reconnect device start failure Maxime Coquelin
2024-10-23 16:06   ` David Marchand [this message]
2024-10-23 18:57     ` Maxime Coquelin
2024-10-23 15:15 ` [PATCH 4/6] vhost: refactor VDUSE reconnection log mapping Maxime Coquelin
2024-10-23 16:07   ` David Marchand
2024-10-23 19:04     ` Maxime Coquelin
2024-10-23 15:15 ` [PATCH 5/6] vhost: fix and refactor VDUSE reconnect log check Maxime Coquelin
2024-10-23 16:08   ` David Marchand
2024-10-23 15:15 ` [PATCH 6/6] vhost: move VDUSE reconnection after device is created Maxime Coquelin

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='CAJFAV8x5hU=V56BmAWOQ7hbMGnHdUStR9Yc2E5rjV77j9sDjTQ@mail.gmail.com' \
    --to=david.marchand@redhat.com \
    --cc=chenbox@nvidia.com \
    --cc=dev@dpdk.org \
    --cc=maxime.coquelin@redhat.com \
    /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).