[dpdk-dev] [PATCH] net/ice: fix link status recovery

[Yang, Qiming]

Yes, I know this bug, will fix it in the next version.

-----Original Message-----
From: Stillwell Jr, Paul M 
Sent: Monday, November 4, 2019 23:03
To: Yang, Qiming <qiming.yang at intel.com>; dev at dpdk.org
Cc: Lu, Wenzhuo <wenzhuo.lu at intel.com>; Yang, Qiming <qiming.yang at intel.com>; stable at dpdk.org
Subject: RE: [dpdk-dev] [PATCH] net/ice: fix link status recovery

> -----Original Message-----
> From: dev <dev-bounces at dpdk.org> On Behalf Of Qiming Yang
> Sent: Sunday, November 3, 2019 9:40 PM
> To: dev at dpdk.org
> Cc: Lu, Wenzhuo <wenzhuo.lu at intel.com>; Yang, Qiming 
> <qiming.yang at intel.com>; stable at dpdk.org
> Subject: [dpdk-dev] [PATCH] net/ice: fix link status recovery
> 
> This patch fix a kernel driver link status issue by recovering link 
> status when device stop.
> 
> Fixes: e6161345d8a9 ("net/ice: support link status change")
> Cc: stable at dpdk.org
> 
> Signed-off-by: Qiming Yang <qiming.yang at intel.com>
> ---
>  drivers/net/ice/ice_ethdev.c | 26 +++++++++++++++++++++++++- 
> drivers/net/ice/ice_ethdev.h |  1 +
>  2 files changed, 26 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/net/ice/ice_ethdev.c 
> b/drivers/net/ice/ice_ethdev.c index d746758..c9972cf 100644
> --- a/drivers/net/ice/ice_ethdev.c
> +++ b/drivers/net/ice/ice_ethdev.c
> @@ -2277,7 +2277,10 @@ ice_dev_stop(struct rte_eth_dev *dev)
>  /* Clear all queues and release mbufs */  ice_clear_queues(dev);
> 
> -ice_dev_set_link_down(dev);
> +if (pf->init_link_status)
> +ice_dev_set_link_up(dev);
> +else
> +ice_dev_set_link_down(dev);
> 
>  /* Clean datapath event and queue/vec mapping */  
> rte_intr_efd_disable(intr_handle);
> @@ -2648,6 +2651,25 @@ ice_rxq_intr_setup(struct rte_eth_dev *dev)  
> return 0;  }
> 
> +static void ice_get_init_link_status(struct rte_eth_dev *dev) { 
> +struct ice_hw *hw = ICE_DEV_PRIVATE_TO_HW(dev->data-
> >dev_private);
> +struct ice_pf *pf = ICE_DEV_PRIVATE_TO_PF(dev->data-
> >dev_private);
> +bool enable_lse = dev->data->dev_conf.intr_conf.lsc ? true : false; 
> +struct ice_link_status link_status; int ret;
> +
> +ret = ice_aq_get_link_info(hw->port_info, enable_lse,
> +   &link_status, NULL);
> +if (ret != ICE_SUCCESS) {
> +PMD_DRV_LOG(ERR, "Failed to get link info");
> +pf->init_link_status = false;

You should probably just return here instead of continuing on. If the FW doesn't clear out the link_status variable then there could be garbage in there that would match the test below for link_info & ICE_AQ_LINK_UP, thus setting init_link_status to true after you just set it to false.

> +}
> +
> +if (link_status.link_info & ICE_AQ_LINK_UP)
> +pf->init_link_status = true;
> +}
> +
>  static int
>  ice_dev_start(struct rte_eth_dev *dev)  { @@ -2717,6 +2739,8 @@ 
> ice_dev_start(struct rte_eth_dev *dev)  if (ret != ICE_SUCCESS)  
> PMD_DRV_LOG(WARNING, "Fail to set phy mask");
> 
> +ice_get_init_link_status(dev);
> +
>  ice_dev_set_link_up(dev);
> 
>  /* Call get_link_info aq commond to enable/disable LSE */ diff --git 
> a/drivers/net/ice/ice_ethdev.h b/drivers/net/ice/ice_ethdev.h index
> de67e59..27db4e7 100644
> --- a/drivers/net/ice/ice_ethdev.h
> +++ b/drivers/net/ice/ice_ethdev.h
> @@ -369,6 +369,7 @@ struct ice_pf {
>  struct ice_parser_list rss_parser_list;  struct ice_parser_list 
> perm_parser_list;  struct ice_parser_list dist_parser_list;
> +bool init_link_status;
>  };
> 
>  #define ICE_MAX_QUEUE_NUM  2048
> --
> 2.9.5