DPDK patches and discussions
 help / color / mirror / Atom feed
From: Phil Yang <Phil.Yang@arm.com>
To: Honnappa Nagarahalli <Honnappa.Nagarahalli@arm.com>,
	"dev@dpdk.org" <dev@dpdk.org>,
	Honnappa Nagarahalli <Honnappa.Nagarahalli@arm.com>,
	"thomas@monjalon.net" <thomas@monjalon.net>,
	"arybchenko@solarflare.com" <arybchenko@solarflare.com>,
	"ferruh.yigit@intel.com" <ferruh.yigit@intel.com>
Cc: "abhinandan.gujjar@intel.com" <abhinandan.gujjar@intel.com>,
	nd <nd@arm.com>,
	"bruce.richardson@intel.com" <bruce.richardson@intel.com>,
	"john.mcnamara@intel.com" <john.mcnamara@intel.com>,
	"reshma.pattan@intel.com" <reshma.pattan@intel.com>,
	"stable@dpdk.org" <stable@dpdk.org>, nd <nd@arm.com>
Subject: Re: [dpdk-dev] [PATCH 2/2] lib/ethdev: fix memory ordering for call back functions
Date: Fri, 9 Oct 2020 02:34:14 +0000	[thread overview]
Message-ID: <DB7PR08MB3865B0185DF6C0C82ECF27F3E9080@DB7PR08MB3865.eurprd08.prod.outlook.com> (raw)
In-Reply-To: <20201002000711.41511-2-honnappa.nagarahalli@arm.com>

> -----Original Message-----
> From: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
> Sent: Friday, October 2, 2020 8:07 AM
> To: dev@dpdk.org; Honnappa Nagarahalli
> <Honnappa.Nagarahalli@arm.com>; Phil Yang <Phil.Yang@arm.com>;
> thomas@monjalon.net; arybchenko@solarflare.com; ferruh.yigit@intel.com
> Cc: abhinandan.gujjar@intel.com; nd <nd@arm.com>;
> bruce.richardson@intel.com; john.mcnamara@intel.com;
> reshma.pattan@intel.com; stable@dpdk.org
> Subject: [PATCH 2/2] lib/ethdev: fix memory ordering for call back functions
> 
> Call back functions are registered on the control plane. They
> are accessed from the data plane. Hence, correct memory orderings
> should be used to avoid race conditions.
> 
> Fixes: 4dc294158cac ("ethdev: support optional Rx and Tx callbacks")
> Fixes: c8231c63ddcb ("ethdev: insert Rx callback as head of list")
> Cc: bruce.richardson@intel.com
> Cc: john.mcnamara@intel.com
> Cc: reshma.pattan@intel.com
> Cc: stable@dpdk.org
> 
> Signed-off-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
> Reviewed-by: Ola Liljedahl <ola.liljedahl@arm.com>


Reviewed-by: Phil Yang <phil.yang@arm.com>


> ---
>  lib/librte_ethdev/rte_ethdev.c | 28 +++++++++++++++++++++------
>  lib/librte_ethdev/rte_ethdev.h | 35 ++++++++++++++++++++++++++-------
> -
>  2 files changed, 49 insertions(+), 14 deletions(-)
> 
> diff --git a/lib/librte_ethdev/rte_ethdev.c b/lib/librte_ethdev/rte_ethdev.c
> index 59a41c07f..d89fcdc77 100644
> --- a/lib/librte_ethdev/rte_ethdev.c
> +++ b/lib/librte_ethdev/rte_ethdev.c
> @@ -4486,12 +4486,20 @@ rte_eth_add_rx_callback(uint16_t port_id,
> uint16_t queue_id,
>  		rte_eth_devices[port_id].post_rx_burst_cbs[queue_id];
> 
>  	if (!tail) {
> -		rte_eth_devices[port_id].post_rx_burst_cbs[queue_id] = cb;
> +		/* Stores to cb->fn and cb->param should complete before
> +		 * cb is visible to data plane.
> +		 */
> +		__atomic_store_n(
> +
> 	&rte_eth_devices[port_id].post_rx_burst_cbs[queue_id],
> +			cb, __ATOMIC_RELEASE);
> 
>  	} else {
>  		while (tail->next)
>  			tail = tail->next;
> -		tail->next = cb;
> +		/* Stores to cb->fn and cb->param should complete before
> +		 * cb is visible to data plane.
> +		 */
> +		__atomic_store_n(&tail->next, cb, __ATOMIC_RELEASE);
>  	}
>  	rte_spinlock_unlock(&rte_eth_rx_cb_lock);
> 
> @@ -4576,12 +4584,20 @@ rte_eth_add_tx_callback(uint16_t port_id,
> uint16_t queue_id,
>  		rte_eth_devices[port_id].pre_tx_burst_cbs[queue_id];
> 
>  	if (!tail) {
> -		rte_eth_devices[port_id].pre_tx_burst_cbs[queue_id] = cb;
> +		/* Stores to cb->fn and cb->param should complete before
> +		 * cb is visible to data plane.
> +		 */
> +		__atomic_store_n(
> +
> 	&rte_eth_devices[port_id].pre_tx_burst_cbs[queue_id],
> +			cb, __ATOMIC_RELEASE);
> 
>  	} else {
>  		while (tail->next)
>  			tail = tail->next;
> -		tail->next = cb;
> +		/* Stores to cb->fn and cb->param should complete before
> +		 * cb is visible to data plane.
> +		 */
> +		__atomic_store_n(&tail->next, cb, __ATOMIC_RELEASE);
>  	}
>  	rte_spinlock_unlock(&rte_eth_tx_cb_lock);
> 
> @@ -4612,7 +4628,7 @@ rte_eth_remove_rx_callback(uint16_t port_id,
> uint16_t queue_id,
>  		cb = *prev_cb;
>  		if (cb == user_cb) {
>  			/* Remove the user cb from the callback list. */
> -			*prev_cb = cb->next;
> +			__atomic_store_n(prev_cb, cb->next,
> __ATOMIC_RELAXED);
>  			ret = 0;
>  			break;
>  		}
> @@ -4646,7 +4662,7 @@ rte_eth_remove_tx_callback(uint16_t port_id,
> uint16_t queue_id,
>  		cb = *prev_cb;
>  		if (cb == user_cb) {
>  			/* Remove the user cb from the callback list. */
> -			*prev_cb = cb->next;
> +			__atomic_store_n(prev_cb, cb->next,
> __ATOMIC_RELAXED);
>  			ret = 0;
>  			break;
>  		}
> diff --git a/lib/librte_ethdev/rte_ethdev.h b/lib/librte_ethdev/rte_ethdev.h
> index 70295d7ab..d810e3e38 100644
> --- a/lib/librte_ethdev/rte_ethdev.h
> +++ b/lib/librte_ethdev/rte_ethdev.h
> @@ -3734,7 +3734,8 @@ struct rte_eth_rxtx_callback;
>   *   The callback function
>   * @param user_param
>   *   A generic pointer parameter which will be passed to each invocation of
> the
> - *   callback function on this port and queue.
> + *   callback function on this port and queue. Inter-thread synchronization
> + *   of any user data changes is the responsibility of the user.
>   *
>   * @return
>   *   NULL on error.
> @@ -3763,7 +3764,8 @@ rte_eth_add_rx_callback(uint16_t port_id, uint16_t
> queue_id,
>   *   The callback function
>   * @param user_param
>   *   A generic pointer parameter which will be passed to each invocation of
> the
> - *   callback function on this port and queue.
> + *   callback function on this port and queue. Inter-thread synchronization
> + *   of any user data changes is the responsibility of the user.
>   *
>   * @return
>   *   NULL on error.
> @@ -3791,7 +3793,8 @@ rte_eth_add_first_rx_callback(uint16_t port_id,
> uint16_t queue_id,
>   *   The callback function
>   * @param user_param
>   *   A generic pointer parameter which will be passed to each invocation of
> the
> - *   callback function on this port and queue.
> + *   callback function on this port and queue. Inter-thread synchronization
> + *   of any user data changes is the responsibility of the user.
>   *
>   * @return
>   *   NULL on error.
> @@ -3816,7 +3819,9 @@ rte_eth_add_tx_callback(uint16_t port_id, uint16_t
> queue_id,
>   *   on that queue.
>   *
>   * - After a short delay - where the delay is sufficient to allow any
> - *   in-flight callbacks to complete.
> + *   in-flight callbacks to complete. Alternately, the RCU mechanism can be
> + *   used to detect when data plane threads have ceased referencing the
> + *   callback memory.
>   *
>   * @param port_id
>   *   The port identifier of the Ethernet device.
> @@ -3849,7 +3854,9 @@ int rte_eth_remove_rx_callback(uint16_t port_id,
> uint16_t queue_id,
>   *   on that queue.
>   *
>   * - After a short delay - where the delay is sufficient to allow any
> - *   in-flight callbacks to complete.
> + *   in-flight callbacks to complete. Alternately, the RCU mechanism can be
> + *   used to detect when data plane threads have ceased referencing the
> + *   callback memory.
>   *
>   * @param port_id
>   *   The port identifier of the Ethernet device.
> @@ -4510,10 +4517,16 @@ rte_eth_rx_burst(uint16_t port_id, uint16_t
> queue_id,
>  				     rx_pkts, nb_pkts);
> 
>  #ifdef RTE_ETHDEV_RXTX_CALLBACKS
> -	if (unlikely(dev->post_rx_burst_cbs[queue_id] != NULL)) {
> -		struct rte_eth_rxtx_callback *cb =
> -				dev->post_rx_burst_cbs[queue_id];
> +	/* __ATOMIC_RELEASE memory order was used when the
> +	 * call back was inserted into the list.
> +	 * Since there is a clear dependency between loading
> +	 * cb and cb->fn/cb->next, __ATOMIC_ACQUIRE memory order is
> +	 * not required.
> +	 */
> +	struct rte_eth_rxtx_callback *cb =
> +			dev->post_rx_burst_cbs[queue_id];
> 
> +	if (unlikely(cb != NULL)) {
>  		do {
>  			nb_rx = cb->fn.rx(port_id, queue_id, rx_pkts, nb_rx,
>  						nb_pkts, cb->param);
> @@ -4775,6 +4788,12 @@ rte_eth_tx_burst(uint16_t port_id, uint16_t
> queue_id,
>  #endif
> 
>  #ifdef RTE_ETHDEV_RXTX_CALLBACKS
> +	/* __ATOMIC_RELEASE memory order was used when the
> +	 * call back was inserted into the list.
> +	 * Since there is a clear dependency between loading
> +	 * cb and cb->fn/cb->next, __ATOMIC_ACQUIRE memory order is
> +	 * not required.
> +	 */
>  	struct rte_eth_rxtx_callback *cb = dev-
> >pre_tx_burst_cbs[queue_id];
> 
>  	if (unlikely(cb != NULL)) {
> --
> 2.17.1


  reply	other threads:[~2020-10-09  2:34 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-10-02  0:07 [dpdk-dev] [PATCH 1/2] lib/ethdev: replace full barrier with relaxed barrier Honnappa Nagarahalli
2020-10-02  0:07 ` [dpdk-dev] [PATCH 2/2] lib/ethdev: fix memory ordering for call back functions Honnappa Nagarahalli
2020-10-09  2:34   ` Phil Yang [this message]
2020-10-13 10:32   ` Ferruh Yigit
2020-10-13 15:25     ` Honnappa Nagarahalli
2020-10-13 14:25   ` Ananyev, Konstantin
2020-10-13 10:30 ` [dpdk-dev] [PATCH 1/2] lib/ethdev: replace full barrier with relaxed barrier Ferruh Yigit
2020-10-13 14:24 ` Ananyev, Konstantin
2020-10-13 16:25 ` [dpdk-dev] [PATCH v2 " Honnappa Nagarahalli
2020-10-13 16:25   ` [dpdk-dev] [PATCH v2 2/2] lib/ethdev: fix memory ordering for call back functions Honnappa Nagarahalli
2020-10-14 16:36   ` [dpdk-dev] [PATCH v2 1/2] lib/ethdev: replace full barrier with relaxed barrier Ferruh Yigit

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=DB7PR08MB3865B0185DF6C0C82ECF27F3E9080@DB7PR08MB3865.eurprd08.prod.outlook.com \
    --to=phil.yang@arm.com \
    --cc=Honnappa.Nagarahalli@arm.com \
    --cc=abhinandan.gujjar@intel.com \
    --cc=arybchenko@solarflare.com \
    --cc=bruce.richardson@intel.com \
    --cc=dev@dpdk.org \
    --cc=ferruh.yigit@intel.com \
    --cc=john.mcnamara@intel.com \
    --cc=nd@arm.com \
    --cc=reshma.pattan@intel.com \
    --cc=stable@dpdk.org \
    --cc=thomas@monjalon.net \
    /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).