patches for DPDK stable branches
 help / color / mirror / Atom feed
From: Luc Pelletier <lucp.at.work@gmail.com>
To: Olivier Matz <olivier.matz@6wind.com>
Cc: dev@dpdk.org, jianfeng.tan@intel.com, david.marchand@redhat.com,
	 stable@dpdk.org
Subject: Re: [dpdk-stable] [PATCH v2] eal: fix race in ctrl thread creation
Date: Thu, 25 Mar 2021 10:42:38 -0400
Message-ID: <CAFeRdtA_5MXnCgxtXZUy6tBm3ZVQKu47Nu+6fWytSuM=9q_3Lg@mail.gmail.com> (raw)
In-Reply-To: <20210325112731.16324-1-olivier.matz@6wind.com>

Hi Olivier,

> Thank you for reporting this problem and submitting the patch.
> I think the issue can be fixed without any loop, like in this
> patch. What do you think?

I think getting rid of the loop is an excellent idea. Good thinking.
Your version is much cleaner.

> +       __atomic_store_n(&params->barrier_refcnt, 2, __ATOMIC_RELEASE);

I don't mean to nitpick but I don't think you need to use
__atomic_store_n to initialize the refcnt. Either way is fine of
course :)

Thanks.

Le jeu. 25 mars 2021 à 07:27, Olivier Matz <olivier.matz@6wind.com> a écrit :
>
> As reported by Luc, there is a race where the barrier is destroyed by
> one thread, while the other thread did not yet leave
> pthread_barrier_wait.
>
> This patch fixes the race condition by adding an atomic counter to
> ensure that the barrier is destroyed only it is not used by any thread.
>
> Fixes: 3a0d465d4c53 ("eal: fix use-after-free on control thread creation")
> Cc: jianfeng.tan@intel.com
> Cc: stable@dpdk.org
>
> Reported-by: Luc Pelletier <lucp.at.work@gmail.com>
> Signed-off-by: David Marchand <david.marchand@redhat.com>
> Signed-off-by: Olivier Matz <olivier.matz@6wind.com>
> ---
>
> Hi Luc,
>
> Thank you for reporting this problem and submitting the patch.
> I think the issue can be fixed without any loop, like in this
> patch. What do you think?
>
> Regards,
> Olivier
>
>
>  lib/librte_eal/common/eal_common_thread.c | 38 +++++++++++++----------
>  1 file changed, 21 insertions(+), 17 deletions(-)
>
> diff --git a/lib/librte_eal/common/eal_common_thread.c b/lib/librte_eal/common/eal_common_thread.c
> index 73a055902a..891f825e87 100644
> --- a/lib/librte_eal/common/eal_common_thread.c
> +++ b/lib/librte_eal/common/eal_common_thread.c
> @@ -170,11 +170,11 @@ struct rte_thread_ctrl_params {
>         void *(*start_routine)(void *);
>         void *arg;
>         pthread_barrier_t configured;
> +       unsigned int barrier_refcnt;
>  };
>
>  static void *ctrl_thread_init(void *arg)
>  {
> -       int ret;
>         struct internal_config *internal_conf =
>                 eal_get_internal_configuration();
>         rte_cpuset_t *cpuset = &internal_conf->ctrl_cpuset;
> @@ -184,8 +184,9 @@ static void *ctrl_thread_init(void *arg)
>
>         __rte_thread_init(rte_lcore_id(), cpuset);
>
> -       ret = pthread_barrier_wait(&params->configured);
> -       if (ret == PTHREAD_BARRIER_SERIAL_THREAD) {
> +       pthread_barrier_wait(&params->configured);
> +       if (__atomic_sub_fetch(&params->barrier_refcnt, 1,
> +                               __ATOMIC_ACQ_REL) == 0) {
>                 pthread_barrier_destroy(&params->configured);
>                 free(params);
>         }
> @@ -210,15 +211,17 @@ rte_ctrl_thread_create(pthread_t *thread, const char *name,
>
>         params->start_routine = start_routine;
>         params->arg = arg;
> -
> -       pthread_barrier_init(&params->configured, NULL, 2);
> -
> -       ret = pthread_create(thread, attr, ctrl_thread_init, (void *)params);
> +       __atomic_store_n(&params->barrier_refcnt, 2, __ATOMIC_RELEASE);
> +       ret = pthread_barrier_init(&params->configured, NULL, 2);
>         if (ret != 0) {
>                 free(params);
>                 return -ret;
>         }
>
> +       ret = pthread_create(thread, attr, ctrl_thread_init, (void *)params);
> +       if (ret != 0)
> +               goto fail;
> +
>         if (name != NULL) {
>                 ret = rte_thread_setname(*thread, name);
>                 if (ret < 0)
> @@ -227,25 +230,26 @@ rte_ctrl_thread_create(pthread_t *thread, const char *name,
>         }
>
>         ret = pthread_setaffinity_np(*thread, sizeof(*cpuset), cpuset);
> -       if (ret)
> -               goto fail;
> +       if (ret != 0)
> +               goto fail_cancel;
>
> -       ret = pthread_barrier_wait(&params->configured);
> -       if (ret == PTHREAD_BARRIER_SERIAL_THREAD) {
> +       pthread_barrier_wait(&params->configured);
> +       if (__atomic_sub_fetch(&params->barrier_refcnt, 1,
> +                               __ATOMIC_ACQ_REL) == 0) {
>                 pthread_barrier_destroy(&params->configured);
>                 free(params);
>         }
>
>         return 0;
>
> -fail:
> -       if (PTHREAD_BARRIER_SERIAL_THREAD ==
> -           pthread_barrier_wait(&params->configured)) {
> -               pthread_barrier_destroy(&params->configured);
> -               free(params);
> -       }
> +fail_cancel:
>         pthread_cancel(*thread);
>         pthread_join(*thread, NULL);
> +
> +fail:
> +       pthread_barrier_destroy(&params->configured);
> +       free(params);
> +
>         return -ret;
>  }
>
> --
> 2.29.2
>

  reply	other threads:[~2021-03-25 14:42 UTC|newest]

Thread overview: 24+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-03-24 13:04 [dpdk-stable] [PATCH] eal: fix possible UB on creation of ctrl thread Luc Pelletier
2021-03-25 11:27 ` [dpdk-stable] [PATCH v2] eal: fix race in ctrl thread creation Olivier Matz
2021-03-25 14:42   ` Luc Pelletier [this message]
2021-04-02  4:34   ` [dpdk-stable] [dpdk-dev] " Honnappa Nagarahalli
2021-04-06 16:15 ` [dpdk-stable] [PATCH v3] " Luc Pelletier
2021-04-06 21:10   ` Honnappa Nagarahalli
2021-04-07 12:35     ` [dpdk-stable] [PATCH v4] " Luc Pelletier
2021-04-07 12:53       ` [dpdk-stable] [PATCH v5] " Luc Pelletier
2021-04-07 13:22         ` Luc Pelletier
2021-04-07 13:31         ` Olivier Matz
2021-04-07 14:42           ` [dpdk-stable] [PATCH v6] " Luc Pelletier
2021-04-07 14:57             ` Olivier Matz
2021-04-07 15:29               ` [dpdk-stable] [PATCH v7] " Luc Pelletier
2021-04-07 17:15                 ` Honnappa Nagarahalli
2021-04-07 15:15           ` [dpdk-stable] [PATCH v5] " Honnappa Nagarahalli
2021-04-07 20:16             ` [dpdk-stable] [PATCH 1/2] " Luc Pelletier
2021-04-08 14:17               ` Olivier Matz
2021-04-08 17:06               ` Honnappa Nagarahalli
2021-04-07 20:16             ` [dpdk-stable] [PATCH 2/2] eal: fix hang in ctrl thread creation error logic Luc Pelletier
2021-04-08 14:20               ` Olivier Matz
2021-04-08 18:01                 ` Luc Pelletier
2021-04-09  8:13                   ` [dpdk-stable] [dpdk-dev] " David Marchand
2021-04-08 17:07               ` [dpdk-stable] " Honnappa Nagarahalli
2021-04-09 14:34               ` David Marchand

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='CAFeRdtA_5MXnCgxtXZUy6tBm3ZVQKu47Nu+6fWytSuM=9q_3Lg@mail.gmail.com' \
    --to=lucp.at.work@gmail.com \
    --cc=david.marchand@redhat.com \
    --cc=dev@dpdk.org \
    --cc=jianfeng.tan@intel.com \
    --cc=olivier.matz@6wind.com \
    --cc=stable@dpdk.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

patches for DPDK stable branches

This inbox may be cloned and mirrored by anyone:

	git clone --mirror https://inbox.dpdk.org/stable/0 stable/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 stable stable/ https://inbox.dpdk.org/stable \
		stable@dpdk.org
	public-inbox-index stable

Example config snippet for mirrors.
Newsgroup available over NNTP:
	nntp://inbox.dpdk.org/inbox.dpdk.stable


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git