DPDK patches and discussions
 help / color / mirror / Atom feed
From: "Zhang, Qi Z" <qi.z.zhang@intel.com>
To: "Deng, KaiwenX" <kaiwenx.deng@intel.com>, "dev@dpdk.org" <dev@dpdk.org>
Cc: "stable@dpdk.org" <stable@dpdk.org>,
	"Yang, Qiming" <qiming.yang@intel.com>,
	"Zhou, YidingX" <yidingx.zhou@intel.com>,
	"Wu, Jingjing" <jingjing.wu@intel.com>,
	"Xing, Beilei" <beilei.xing@intel.com>
Subject: RE: [PATCH v2] net/iavf: fix iavf query stats in intr thread
Date: Wed, 15 Mar 2023 13:40:54 +0000	[thread overview]
Message-ID: <DM4PR11MB599445CEFB754FF95FAD7A59D7BF9@DM4PR11MB5994.namprd11.prod.outlook.com> (raw)
In-Reply-To: <20230307025533.1950861-1-kaiwenx.deng@intel.com>



> -----Original Message-----
> From: Deng, KaiwenX <kaiwenx.deng@intel.com>
> Sent: Tuesday, March 7, 2023 10:56 AM
> To: dev@dpdk.org
> Cc: stable@dpdk.org; Yang, Qiming <qiming.yang@intel.com>; Zhou, YidingX
> <yidingx.zhou@intel.com>; Deng, KaiwenX <kaiwenx.deng@intel.com>; Wu,
> Jingjing <jingjing.wu@intel.com>; Xing, Beilei <beilei.xing@intel.com>; Zhang,
> Qi Z <qi.z.zhang@intel.com>
> Subject: [PATCH v2] net/iavf: fix iavf query stats in intr thread
> 
> When iavf send query-stats command in eal-intr-thread through virtual
> channel, there will be no response received from iavf_dev_virtchnl_handler
> for this command during block and wait.

This is because the improper implementation in iavf_execute_vf_cmd
I assume without interrupt (in eal_intr_thread), we should able to query stats through virtchnl by pulling mode.
Check out how "VIRTCHNL_OP_VERSION" was implemented.


> Because iavf_dev_virtchnl_handler is also registered in eal-intr-thread.
> 
> When vf device is bonded as BONDING_MODE_TLB mode, the slave device
> update callback will registered in alarm and called by eal-intr-thread, it
> would also raise the above issue.
> 
> This commit adds local stats return to iavf_dev_stats_get immediately when
> it is called by eal-intr-thread. And updates local stats in iavf-virtchnl-thread.
> 
> Fixes: cb5c1b91f76f ("net/iavf: add thread for event callbacks")
> Fixes: 22b123a36d07 ("net/avf: initialize PMD")
> Cc: stable@dpdk.org
> 
> Signed-off-by: Kaiwen Deng <kaiwenx.deng@intel.com>
> ---
> Changes since v1:
> - Add lock to avoid race condition.
> ---
> ---
>  drivers/net/iavf/iavf.h        |  10 ++-
>  drivers/net/iavf/iavf_ethdev.c |  37 ++++++++++--
> drivers/net/iavf/iavf_vchnl.c  | 107 ++++++++++++++++++++++++---------
>  3 files changed, 118 insertions(+), 36 deletions(-)
> 
> diff --git a/drivers/net/iavf/iavf.h b/drivers/net/iavf/iavf.h index
> 1edebab8dc..641dfa2e3b 100644
> --- a/drivers/net/iavf/iavf.h
> +++ b/drivers/net/iavf/iavf.h
> @@ -128,6 +128,8 @@ struct iavf_vsi {
>  	uint16_t base_vector;
>  	uint16_t msix_intr;      /* The MSIX interrupt binds to VSI */
>  	struct iavf_eth_xstats eth_stats_offset;
> +	struct virtchnl_eth_stats eth_stats;
> +	rte_spinlock_t stats_lock;
>  };
> 
>  struct rte_flow;
> @@ -325,6 +327,8 @@ struct iavf_adapter {
>  	struct iavf_devargs devargs;
>  };
> 
> +typedef void (*virtchnl_callback)(struct rte_eth_dev *dev, void *args);
> +
>  /* IAVF_DEV_PRIVATE_TO */
>  #define IAVF_DEV_PRIVATE_TO_ADAPTER(adapter) \
>  	((struct iavf_adapter *)adapter)
> @@ -424,8 +428,10 @@ _atomic_set_async_response_cmd(struct iavf_info
> *vf, enum virtchnl_ops ops)  }  int iavf_check_api_version(struct iavf_adapter
> *adapter);  int iavf_get_vf_resource(struct iavf_adapter *adapter); -void
> iavf_dev_event_handler_fini(void);
> -int iavf_dev_event_handler_init(void);
> +void iavf_dev_virtchnl_handler_fini(void);
> +void iavf_dev_virtchnl_callback_post(struct rte_eth_dev *dev,
> +			 virtchnl_callback cb, void *args);
> +int iavf_dev_virtchnl_handler_init(void);
>  void iavf_handle_virtchnl_msg(struct rte_eth_dev *dev);  int
> iavf_enable_vlan_strip(struct iavf_adapter *adapter);  int
> iavf_disable_vlan_strip(struct iavf_adapter *adapter); diff --git
> a/drivers/net/iavf/iavf_ethdev.c b/drivers/net/iavf/iavf_ethdev.c index
> 3196210f2c..772859d157 100644
> --- a/drivers/net/iavf/iavf_ethdev.c
> +++ b/drivers/net/iavf/iavf_ethdev.c
> @@ -1729,6 +1729,22 @@ iavf_update_stats(struct iavf_vsi *vsi, struct
> virtchnl_eth_stats *nes)
>  	iavf_stat_update_32(&oes->tx_discards, &nes->tx_discards);  }
> 
> +static void iavf_dev_stats_get_callback(struct rte_eth_dev *dev, void
> +*args) {
> +	struct virtchnl_eth_stats *eth_stats = (struct virtchnl_eth_stats
> *)args;
> +	struct virtchnl_eth_stats *pstats = NULL;
> +	struct iavf_adapter *adapter =
> +		IAVF_DEV_PRIVATE_TO_ADAPTER(dev->data->dev_private);
> +	struct iavf_info *vf = IAVF_DEV_PRIVATE_TO_VF(dev->data-
> >dev_private);
> +	struct iavf_vsi *vsi = &vf->vsi;
> +	int ret = iavf_query_stats(adapter, &pstats);
> +	if (ret == 0) {
> +		rte_spinlock_lock(&vsi->stats_lock);
> +		rte_memcpy(eth_stats, pstats, sizeof(struct
> virtchnl_eth_stats));
> +		rte_spinlock_unlock(&vsi->stats_lock);
> +	}
> +}
> +
>  static int
>  iavf_dev_stats_get(struct rte_eth_dev *dev, struct rte_eth_stats *stats)
> { @@ -1738,9 +1754,17 @@ iavf_dev_stats_get(struct rte_eth_dev *dev,
> struct rte_eth_stats *stats)
>  	struct iavf_vsi *vsi = &vf->vsi;
>  	struct virtchnl_eth_stats *pstats = NULL;
>  	int ret;
> -
> -	ret = iavf_query_stats(adapter, &pstats);
> +	int is_intr_thread = rte_thread_is_intr();
> +	if (is_intr_thread) {
> +		pstats = &vsi->eth_stats;
> +		iavf_dev_virtchnl_callback_post(dev,
> iavf_dev_stats_get_callback, (void *)pstats);
> +		ret = 0;
> +	} else {
> +		ret = iavf_query_stats(adapter, &pstats);
> +	}
>  	if (ret == 0) {
> +		if (is_intr_thread)
> +			rte_spinlock_lock(&vsi->stats_lock);
>  		uint8_t crc_stats_len = (dev->data-
> >dev_conf.rxmode.offloads &
>  					 RTE_ETH_RX_OFFLOAD_KEEP_CRC) ?
> 0 :
>  					 RTE_ETHER_CRC_LEN;
> @@ -1754,6 +1778,8 @@ iavf_dev_stats_get(struct rte_eth_dev *dev, struct
> rte_eth_stats *stats)
>  		stats->ibytes = pstats->rx_bytes;
>  		stats->ibytes -= stats->ipackets * crc_stats_len;
>  		stats->obytes = pstats->tx_bytes;
> +		if (is_intr_thread)
> +			rte_spinlock_unlock(&vsi->stats_lock);
>  	} else {
>  		PMD_DRV_LOG(ERR, "Get statistics failed");
>  	}
> @@ -2571,10 +2597,13 @@ iavf_dev_init(struct rte_eth_dev *eth_dev)
>  	struct iavf_hw *hw = IAVF_DEV_PRIVATE_TO_HW(adapter);
>  	struct iavf_info *vf = IAVF_DEV_PRIVATE_TO_VF(adapter);
>  	struct rte_pci_device *pci_dev = RTE_ETH_DEV_TO_PCI(eth_dev);
> +	struct iavf_vsi *vsi = &vf->vsi;
>  	int ret = 0;
> 
>  	PMD_INIT_FUNC_TRACE();
> 
> +	rte_spinlock_init(&vsi->stats_lock);
> +
>  	/* assign ops func pointer */
>  	eth_dev->dev_ops = &iavf_eth_dev_ops;
>  	eth_dev->rx_queue_count = iavf_dev_rxq_count; @@ -2634,7
> +2663,7 @@ iavf_dev_init(struct rte_eth_dev *eth_dev)
>  	rte_ether_addr_copy((struct rte_ether_addr *)hw->mac.addr,
>  			&eth_dev->data->mac_addrs[0]);
> 
> -	if (iavf_dev_event_handler_init())
> +	if (iavf_dev_virtchnl_handler_init())
>  		goto init_vf_err;
> 
>  	if (vf->vf_res->vf_cap_flags & VIRTCHNL_VF_OFFLOAD_WB_ON_ITR)
> { @@ -2791,7 +2820,7 @@ iavf_dev_uninit(struct rte_eth_dev *dev)
> 
>  	iavf_dev_close(dev);
> 
> -	iavf_dev_event_handler_fini();
> +	iavf_dev_virtchnl_handler_fini();
> 
>  	return 0;
>  }
> diff --git a/drivers/net/iavf/iavf_vchnl.c b/drivers/net/iavf/iavf_vchnl.c index
> f92daf97f2..4136c97c45 100644
> --- a/drivers/net/iavf/iavf_vchnl.c
> +++ b/drivers/net/iavf/iavf_vchnl.c
> @@ -31,24 +31,36 @@
> 
>  #define MAX_EVENT_PENDING 16
> 
> -struct iavf_event_element {
> -	TAILQ_ENTRY(iavf_event_element) next;
> +struct iavf_virtchnl_element {
> +	TAILQ_ENTRY(iavf_virtchnl_element) next;
>  	struct rte_eth_dev *dev;
> -	enum rte_eth_event_type event;
> -	void *param;
> -	size_t param_alloc_size;
> -	uint8_t param_alloc_data[0];
> +	enum iavf_virchnl_handle_type {
> +		EVENT_TYPE = 0,
> +		CALL_TYPE
> +	} handle_type;
> +	union {
> +		struct event_param {
> +			enum rte_eth_event_type event;
> +			void *param;
> +			size_t param_alloc_size;
> +			uint8_t param_alloc_data[0];
> +		} ep;
> +		struct call_param {
> +			virtchnl_callback cb;
> +			void *args;
> +		} cp;
> +	};
>  };
> 
> -struct iavf_event_handler {
> +struct iavf_virtchnl_handler {
>  	uint32_t ndev;
>  	pthread_t tid;
>  	int fd[2];
>  	pthread_mutex_t lock;
> -	TAILQ_HEAD(event_list, iavf_event_element) pending;
> +	TAILQ_HEAD(event_list, iavf_virtchnl_element) pending;
>  };
> 
> -static struct iavf_event_handler event_handler = {
> +static struct iavf_virtchnl_handler event_handler = {
>  	.fd = {-1, -1},
>  };
> 
> @@ -60,10 +72,10 @@ static struct iavf_event_handler event_handler =
> {  #endif
> 
>  static void *
> -iavf_dev_event_handle(void *param __rte_unused)
> +iavf_dev_virtchnl_handle(void *param __rte_unused)
>  {
> -	struct iavf_event_handler *handler = &event_handler;
> -	TAILQ_HEAD(event_list, iavf_event_element) pending;
> +	struct iavf_virtchnl_handler *handler = &event_handler;
> +	TAILQ_HEAD(event_list, iavf_virtchnl_element) pending;
> 
>  	while (true) {
>  		char unused[MAX_EVENT_PENDING];
> @@ -76,10 +88,22 @@ iavf_dev_event_handle(void *param __rte_unused)
>  		TAILQ_CONCAT(&pending, &handler->pending, next);
>  		pthread_mutex_unlock(&handler->lock);
> 
> -		struct iavf_event_element *pos, *save_next;
> +		struct iavf_virtchnl_element *pos, *save_next;
>  		TAILQ_FOREACH_SAFE(pos, &pending, next, save_next) {
>  			TAILQ_REMOVE(&pending, pos, next);
> -			rte_eth_dev_callback_process(pos->dev, pos->event,
> pos->param);
> +
> +			switch (pos->handle_type) {
> +			case EVENT_TYPE:
> +				rte_eth_dev_callback_process(pos->dev,
> +					pos->ep.event, pos->ep.param);
> +				break;
> +			case CALL_TYPE:
> +				pos->cp.cb(pos->dev, pos->cp.args);
> +				break;
> +			default:
> +				break;
> +			}
> +
>  			rte_free(pos);
>  		}
>  	}
> @@ -92,19 +116,20 @@ iavf_dev_event_post(struct rte_eth_dev *dev,
>  		enum rte_eth_event_type event,
>  		void *param, size_t param_alloc_size)  {
> -	struct iavf_event_handler *handler = &event_handler;
> +	struct iavf_virtchnl_handler *handler = &event_handler;
>  	char notify_byte;
> -	struct iavf_event_element *elem = rte_malloc(NULL, sizeof(*elem) +
> param_alloc_size, 0);
> +	struct iavf_virtchnl_element *elem = rte_malloc(NULL, sizeof(*elem)
> +
> +param_alloc_size, 0);
>  	if (!elem)
>  		return;
> -
> +	elem->handle_type = EVENT_TYPE;
> +	struct event_param *ep = &elem->ep;
>  	elem->dev = dev;
> -	elem->event = event;
> -	elem->param = param;
> -	elem->param_alloc_size = param_alloc_size;
> +	ep->event = event;
> +	ep->param = param;
> +	ep->param_alloc_size = param_alloc_size;
>  	if (param && param_alloc_size) {
> -		rte_memcpy(elem->param_alloc_data, param,
> param_alloc_size);
> -		elem->param = elem->param_alloc_data;
> +		rte_memcpy(ep->param_alloc_data, param,
> param_alloc_size);
> +		ep->param = ep->param_alloc_data;
>  	}
> 
>  	pthread_mutex_lock(&handler->lock);
> @@ -115,10 +140,32 @@ iavf_dev_event_post(struct rte_eth_dev *dev,
>  	RTE_SET_USED(nw);
>  }
> 
> +void
> +iavf_dev_virtchnl_callback_post(struct rte_eth_dev *dev,
> +virtchnl_callback cb, void *args) {
> +	struct iavf_virtchnl_handler *handler = &event_handler;
> +	char notify_byte;
> +	struct iavf_virtchnl_element *elem = rte_malloc(NULL, sizeof(*elem),
> 0);
> +	if (!elem)
> +		return;
> +	elem->dev = dev;
> +	elem->handle_type = CALL_TYPE;
> +	struct call_param *cp = &elem->cp;
> +	cp->cb = cb;
> +	cp->args = args;
> +
> +	pthread_mutex_lock(&handler->lock);
> +	TAILQ_INSERT_TAIL(&handler->pending, elem, next);
> +	pthread_mutex_unlock(&handler->lock);
> +
> +	ssize_t nw = write(handler->fd[1], &notify_byte, 1);
> +	RTE_SET_USED(nw);
> +}
> +
>  int
> -iavf_dev_event_handler_init(void)
> +iavf_dev_virtchnl_handler_init(void)
>  {
> -	struct iavf_event_handler *handler = &event_handler;
> +	struct iavf_virtchnl_handler *handler = &event_handler;
> 
>  	if (__atomic_add_fetch(&handler->ndev, 1, __ATOMIC_RELAXED) != 1)
>  		return 0;
> @@ -135,8 +182,8 @@ iavf_dev_event_handler_init(void)
>  	TAILQ_INIT(&handler->pending);
>  	pthread_mutex_init(&handler->lock, NULL);
> 
> -	if (rte_ctrl_thread_create(&handler->tid, "iavf-event-thread",
> -				NULL, iavf_dev_event_handle, NULL)) {
> +	if (rte_ctrl_thread_create(&handler->tid, "iavf-virtchnl-thread",
> +				NULL, iavf_dev_virtchnl_handle, NULL)) {
>  		__atomic_sub_fetch(&handler->ndev, 1,
> __ATOMIC_RELAXED);
>  		return -1;
>  	}
> @@ -145,9 +192,9 @@ iavf_dev_event_handler_init(void)  }
> 
>  void
> -iavf_dev_event_handler_fini(void)
> +iavf_dev_virtchnl_handler_fini(void)
>  {
> -	struct iavf_event_handler *handler = &event_handler;
> +	struct iavf_virtchnl_handler *handler = &event_handler;
> 
>  	if (__atomic_sub_fetch(&handler->ndev, 1, __ATOMIC_RELAXED) != 0)
>  		return;
> @@ -162,7 +209,7 @@ iavf_dev_event_handler_fini(void)
>  	pthread_join(handler->tid, NULL);
>  	pthread_mutex_destroy(&handler->lock);
> 
> -	struct iavf_event_element *pos, *save_next;
> +	struct iavf_virtchnl_element *pos, *save_next;
>  	TAILQ_FOREACH_SAFE(pos, &handler->pending, next, save_next) {
>  		TAILQ_REMOVE(&handler->pending, pos, next);
>  		rte_free(pos);
> @@ -1408,7 +1455,7 @@ iavf_query_stats(struct iavf_adapter *adapter,
>  	struct iavf_cmd_info args;
>  	int err;
> 
> -	if (adapter->closed)
> +	if (!adapter || adapter->closed)
>  		return -EIO;
> 
>  	memset(&q_stats, 0, sizeof(q_stats));
> --
> 2.34.1


  reply	other threads:[~2023-03-15 13:41 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-02-22  4:40 [PATCH] " Kaiwen Deng
2023-02-27  0:56 ` Zhang, Qi Z
2023-02-27  6:02   ` Deng, KaiwenX
2023-03-07  3:27   ` Deng, KaiwenX
2023-03-07  2:55 ` [PATCH v2] " Kaiwen Deng
2023-03-15 13:40   ` Zhang, Qi Z [this message]
2023-03-22  7:26   ` [PATCH v3] " Kaiwen Deng
2023-03-23 15:39     ` Ferruh Yigit
2023-03-27  5:31       ` Deng, KaiwenX
2023-03-27 12:31         ` Ferruh Yigit
2023-03-27 12:37           ` Ferruh Yigit
2023-03-29  7:53             ` Deng, KaiwenX
2023-05-05  2:31             ` Deng, KaiwenX
2023-05-23  1:45               ` Deng, KaiwenX
2023-05-23 10:26                 ` Ferruh Yigit
2023-03-29  6:41           ` Deng, KaiwenX
2023-06-06  5:41     ` Jiale, SongX
2023-06-07  2:03     ` [PATCH v4] " Kaiwen Deng
2023-06-07  4:01       ` Zhang, Qi Z

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=DM4PR11MB599445CEFB754FF95FAD7A59D7BF9@DM4PR11MB5994.namprd11.prod.outlook.com \
    --to=qi.z.zhang@intel.com \
    --cc=beilei.xing@intel.com \
    --cc=dev@dpdk.org \
    --cc=jingjing.wu@intel.com \
    --cc=kaiwenx.deng@intel.com \
    --cc=qiming.yang@intel.com \
    --cc=stable@dpdk.org \
    --cc=yidingx.zhou@intel.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).