From: Maxime Coquelin <maxime.coquelin@redhat.com>
To: Gaoxiang Liu <gaoxiangliu0@163.com>, chenbo.xia@intel.com
Cc: dev@dpdk.org, liugaoxiang@huawei.com
Subject: Re: [dpdk-dev] [PATCH v7] vhost: fix crash on port deletion
Date: Tue, 14 Sep 2021 13:29:33 +0200 [thread overview]
Message-ID: <cbaf5718-2524-22c9-f0ca-97cfa8a822fc@redhat.com> (raw)
In-Reply-To: <20210902154553.249-1-gaoxiangliu0@163.com>
On 9/2/21 5:45 PM, Gaoxiang Liu wrote:
> The rte_vhost_driver_unregister() and vhost_user_read_cb()
> can be called at the same time by 2 threads.
> when memory of vsocket is freed in rte_vhost_driver_unregister(),
> the invalid memory of vsocket is accessed in vhost_user_read_cb().
> It's a bug of both mode for vhost as server or client.
>
> E.g., vhostuser port is created as server.
> Thread1 calls rte_vhost_driver_unregister().
> Before the listen fd is deleted from poll waiting fds,
> "vhost-events" thread then calls vhost_user_server_new_connection(),
> then a new conn fd is added in fdset when trying to reconnect.
> "vhost-events" thread then calls vhost_user_read_cb() and
> accesses invalid memory of socket while thread1 frees the memory of
> vsocket.
>
> E.g., vhostuser port is created as client.
> Thread1 calls rte_vhost_driver_unregister().
> Before vsocket of reconn is deleted from reconn list,
> "vhost_reconn" thread then calls vhost_user_add_connection()
> then a new conn fd is added in fdset when trying to reconnect.
> "vhost-events" thread then calls vhost_user_read_cb() and
> accesses invalid memory of socket while thread1 frees the memory of
> vsocket.
>
> The fix is to move the "fdset_try_del" in front of free memory of conn,
> then avoid the race condition.
>
> The core trace is:
> Program terminated with signal 11, Segmentation fault.
>
> Fixes: 52d874dc6705 ("vhost: fix crash on closing in client mode")
>
> Signed-off-by: Gaoxiang Liu <liugaoxiang@huawei.com>
> ---
>
> v2:
> * Fix coding style issues.
>
> v3:
> * Add detailed log.
>
> v4:
> * Add the reason, when vhostuser port is created as server.
>
> v5:
> * Add detailed log when vhostuser port is created as client
>
> v6:
> * Add 'path' check before deleting listen fd
> * Fix spelling issues
>
> v7:
> * Fix coding style issues.
> ---
> lib/vhost/socket.c | 107 ++++++++++++++++++++++-----------------------
> 1 file changed, 53 insertions(+), 54 deletions(-)
>
Applied to dpdk-next-virtio/main.
Thanks,
Maxime
prev parent reply other threads:[~2021-09-14 11:29 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-08-07 8:25 [dpdk-dev] [PATCH] vhost: fix coredump " Gaoxiang Liu
2021-08-07 23:12 ` [dpdk-dev] [PATCH v2] " Gaoxiang Liu
2021-08-13 14:02 ` [dpdk-dev] [PATCH] vhost: fix crash on port deletion The rte_vhost_driver_unregister() and vhost_user_read_cb() can be called at the same time by 2 threads. Eg thread1 calls rte_vhost_driver_unregister() and frees memory of "conn". Because socket fd has not been deleted from poll waiting fds, "vhost-events" thread calls fdset_event_dispatch, then calls vhost_user_read_cb(), and accesses invalid memory of "conn" Gaoxiang Liu
2021-08-13 14:22 ` [dpdk-dev] [PATCH] vhost: fix crash on port deletion Gaoxiang Liu
2021-08-16 6:44 ` Xia, Chenbo
2021-08-20 15:53 ` Gaoxiang Liu
2021-08-18 16:08 ` [dpdk-dev] [PATCH v4] " Gaoxiang Liu
2021-08-20 15:46 ` [dpdk-dev] [PATCH v5] " Gaoxiang Liu
2021-08-26 8:37 ` Xia, Chenbo
2021-08-27 14:19 ` [dpdk-dev] [PATCH v6] " Gaoxiang Liu
2021-08-31 5:37 ` Xia, Chenbo
2021-09-02 15:38 ` Gaoxiang Liu
2021-09-06 3:18 ` Xia, Chenbo
2021-09-06 3:32 ` Xia, Chenbo
2021-09-06 3:54 ` Gaoxiang Liu
2021-09-02 15:45 ` [dpdk-dev] [PATCH v7] " Gaoxiang Liu
2021-09-06 3:24 ` Xia, Chenbo
2021-09-06 5:19 ` Xia, Chenbo
2021-09-14 11:29 ` 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=cbaf5718-2524-22c9-f0ca-97cfa8a822fc@redhat.com \
--to=maxime.coquelin@redhat.com \
--cc=chenbo.xia@intel.com \
--cc=dev@dpdk.org \
--cc=gaoxiangliu0@163.com \
--cc=liugaoxiang@huawei.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).