[Intel-wired-lan] [PATCH net v1] iavf: Fix race between iavf_close and iavf_reset_task

Tony Nguyen anthony.l.nguyen at intel.com
Thu Aug 18 00:00:12 UTC 2022



On 8/11/2022 3:01 AM, Mateusz Palczewski wrote:
> From: Michal Jaron <michalx.jaron at intel.com>
> 
> During stress tests with adding VF to namespace and changing vf's
> trust there was a race between iavf_reset_task and iavf_close.
> Sometimes when IAVF_FLAG_AQ_DISABLE_QUEUES from iavf_close was sent
> to PF after reset and before IAVF_AQ_GET_CONFIG was sent then PF
> returns error IAVF_NOT_SUPPORTED to disable queues request and
> following requests. There is need to get_config before other
> aq_required will be send but iavf_close clears all flags, if
> get_config was not sent before iavf_close, then it will not be send
> at all.
> 
> In case when IAVF_FLAG_AQ_GET_OFFLOAD_VLAN_V2_CAPS was sent before
> IAVF_FLAG_AQ_DISABLE_QUEUES then there was rtnl_lock deadlock
> between iavf_close and iavf_adminq_task until iavf_close timeouts
> and disable queues was sent after iavf_close ends.
> 
> There was also a problem with sending delete/add filters.
> Sometimes when filters was not yet added to PF and in
> iavf_close all filters was set to remove there might be a try
> to remove nonexistent filters on PF.
> 
> Add aq_required_tmp to save aq_required flags and send them after
> disable_queues will be handled. Clear flags given to iavf_down
> different than IAVF_FLAG_AQ_GET_CONFIG as this flag is necessary
> to sent other aq_required. Remove some flags that we don't
> want to send as we are in iavf_close and we want to disable
> interface. Remove filters which was not yet sent and send del
> filters flags only when there are filters to remove.
> 
> Fixes: 53d0b3ae2590 ("i40evf: refactor shutdown code")
> Signed-off-by: Michal Jaron <michalx.jaron at intel.com>
> Signed-off-by: Mateusz Palczewski <mateusz.palczewski at intel.com>
> ---
>   drivers/net/ethernet/intel/iavf/iavf_main.c | 177 ++++++++++++++++----
>   1 file changed, 141 insertions(+), 36 deletions(-)

These changes seem a bit big for net. It would probably be better suited 
for them to go to next.

> diff --git a/drivers/net/ethernet/intel/iavf/iavf_main.c b/drivers/net/ethernet/intel/iavf/iavf_main.c
> index 35dcb1ce12b6..9cfcab5e68e5 100644
> --- a/drivers/net/ethernet/intel/iavf/iavf_main.c
> +++ b/drivers/net/ethernet/intel/iavf/iavf_main.c
> @@ -1270,66 +1270,138 @@ static void iavf_up_complete(struct iavf_adapter *adapter)
>   }
>   
>   /**
> - * iavf_down - Shutdown the connection processing
> + * iavf_clear_mac_vlan_filters - Remove mac and vlan filters not sent to PF
> + * yet and mark other to be removed.
>    * @adapter: board private structure
> - *
> - * Expects to be called while holding the __IAVF_IN_CRITICAL_TASK bit lock.
>    **/
> -void iavf_down(struct iavf_adapter *adapter)
> +static void iavf_clear_mac_vlan_filters(struct iavf_adapter *adapter)
>   {
> -	struct net_device *netdev = adapter->netdev;
> -	struct iavf_vlan_filter *vlf;
> -	struct iavf_cloud_filter *cf;
> -	struct iavf_fdir_fltr *fdir;
> -	struct iavf_mac_filter *f;
> -	struct iavf_adv_rss *rss;
> -
> -	if (adapter->state <= __IAVF_DOWN_PENDING)
> -		return;
> -
> -	netif_carrier_off(netdev);
> -	netif_tx_disable(netdev);
> -	adapter->link_up = false;
> -	iavf_napi_disable_all(adapter);
> -	iavf_irq_disable(adapter);
> +	struct iavf_vlan_filter *vlf, *vlftmp;
> +	struct iavf_mac_filter *f, *ftmp;
>   
>   	spin_lock_bh(&adapter->mac_vlan_list_lock);
> -
>   	/* clear the sync flag on all filters */
>   	__dev_uc_unsync(adapter->netdev, NULL);
>   	__dev_mc_unsync(adapter->netdev, NULL);
>   
>   	/* remove all MAC filters */
> -	list_for_each_entry(f, &adapter->mac_filter_list, list) {
> -		f->remove = true;
> +	list_for_each_entry_safe(f, ftmp, &adapter->mac_filter_list,
> +				 list) {
> +		if (f->add) {
> +			list_del(&f->list);
> +			kfree(f);
> +		} else {
> +			f->remove = true;
> +		}
>   	}
>   
>   	/* remove all VLAN filters */
> -	list_for_each_entry(vlf, &adapter->vlan_filter_list, list) {
> -		vlf->remove = true;
> +	list_for_each_entry_safe(vlf, vlftmp, &adapter->vlan_filter_list,
> +				 list) {
> +		if (vlf->add) {
> +			list_del(&vlf->list);
> +			kfree(vlf);
> +		} else {
> +			vlf->remove = true;
> +		}
>   	}
> -
>   	spin_unlock_bh(&adapter->mac_vlan_list_lock);
> +}
> +
> +/**
> + * iavf_clear_cloud_filters - Remove cloud filters not sent to PF yet and
> + * mark other to be removed.
> + * @adapter: board private structure
> + **/
> +static void iavf_clear_cloud_filters(struct iavf_adapter *adapter)
> +{
> +	struct iavf_cloud_filter *cf, *cftmp;
>   
>   	/* remove all cloud filters */
>   	spin_lock_bh(&adapter->cloud_filter_list_lock);
> -	list_for_each_entry(cf, &adapter->cloud_filter_list, list) {
> -		cf->del = true;
> +	list_for_each_entry_safe(cf, cftmp, &adapter->cloud_filter_list,
> +				 list) {
> +		if (cf->add) {
> +			list_del(&cf->list);
> +			kfree(cf);
> +			adapter->num_cloud_filters--;
> +		} else {
> +			cf->del = true;
> +		}
>   	}
>   	spin_unlock_bh(&adapter->cloud_filter_list_lock);
> +}
> +
> +/**
> + * iavf_clear_fdir_filters - Remove fdir filters not sent to PF yet and mark
> + * other to be removed.
> + * @adapter: board private structure
> + **/
> +static void iavf_clear_fdir_filters(struct iavf_adapter *adapter)
> +{
> +	struct iavf_fdir_fltr *fdir, *fdirtmp;
>   
>   	/* remove all Flow Director filters */
>   	spin_lock_bh(&adapter->fdir_fltr_lock);
> -	list_for_each_entry(fdir, &adapter->fdir_list_head, list) {
> -		fdir->state = IAVF_FDIR_FLTR_DEL_REQUEST;
> +	list_for_each_entry_safe(fdir, fdirtmp, &adapter->fdir_list_head,
> +				 list) {
> +		if (fdir->state == IAVF_FDIR_FLTR_ADD_REQUEST) {
> +			list_del(&fdir->list);
> +			kfree(fdir);
> +			adapter->fdir_active_fltr--;
> +		} else {
> +			fdir->state = IAVF_FDIR_FLTR_DEL_REQUEST;
> +		}
>   	}
>   	spin_unlock_bh(&adapter->fdir_fltr_lock);
> +}
> +
> +/**
> + * iavf_clear_adv_rss_conf - Remove adv rss conf not sent to PF yet and mark
> + * other to be removed.
> + * @adapter: board private structure
> + **/
> +static void iavf_clear_adv_rss_conf(struct iavf_adapter *adapter)
> +{
> +	struct iavf_adv_rss *rss, *rsstmp;
>   
>   	/* remove all advance RSS configuration */
>   	spin_lock_bh(&adapter->adv_rss_lock);
> -	list_for_each_entry(rss, &adapter->adv_rss_list_head, list)
> -		rss->state = IAVF_ADV_RSS_DEL_REQUEST;
> +	list_for_each_entry_safe(rss, rsstmp, &adapter->adv_rss_list_head,
> +				 list) {
> +		if (rss->state == IAVF_ADV_RSS_ADD_REQUEST) {
> +			list_del(&rss->list);
> +			kfree(rss);
> +		} else {
> +			rss->state = IAVF_ADV_RSS_DEL_REQUEST;
> +		}
> +	}
>   	spin_unlock_bh(&adapter->adv_rss_lock);
> +}
> +
> +/**
> + * iavf_down - Shutdown the connection processing
> + * @adapter: board private structure
> + *
> + * Expects to be called while holding the __IAVF_IN_CRITICAL_TASK bit lock.
> + **/
> +void iavf_down(struct iavf_adapter *adapter)
> +{
> +	struct net_device *netdev = adapter->netdev;
> +
> +	if (adapter->state <= __IAVF_DOWN_PENDING)
> +		return;
> +
> +	netif_carrier_off(netdev);
> +	netif_tx_disable(netdev);
> +	adapter->link_up = false;
> +	iavf_napi_disable_all(adapter);
> +	iavf_irq_disable(adapter);
> +
> +	iavf_clear_mac_vlan_filters(adapter);
> +	iavf_clear_cloud_filters(adapter);
> +	iavf_clear_fdir_filters(adapter);
> +	iavf_clear_adv_rss_conf(adapter);
>   
>   	if (!(adapter->flags & IAVF_FLAG_PF_COMMS_FAILED)) {
>   		/* cancel any current operation */
> @@ -1338,11 +1410,16 @@ void iavf_down(struct iavf_adapter *adapter)
>   		 * here for this to complete. The watchdog is still running
>   		 * and it will take care of this.
>   		 */
> -		adapter->aq_required = IAVF_FLAG_AQ_DEL_MAC_FILTER;
> -		adapter->aq_required |= IAVF_FLAG_AQ_DEL_VLAN_FILTER;
> -		adapter->aq_required |= IAVF_FLAG_AQ_DEL_CLOUD_FILTER;
> -		adapter->aq_required |= IAVF_FLAG_AQ_DEL_FDIR_FILTER;
> -		adapter->aq_required |= IAVF_FLAG_AQ_DEL_ADV_RSS_CFG;
> +		if (!list_empty(&adapter->mac_filter_list))
> +			adapter->aq_required |= IAVF_FLAG_AQ_DEL_MAC_FILTER;
> +		if (!list_empty(&adapter->vlan_filter_list))
> +			adapter->aq_required |= IAVF_FLAG_AQ_DEL_VLAN_FILTER;
> +		if (!list_empty(&adapter->cloud_filter_list))
> +			adapter->aq_required |= IAVF_FLAG_AQ_DEL_CLOUD_FILTER;
> +		if (!list_empty(&adapter->fdir_list_head))
> +			adapter->aq_required |= IAVF_FLAG_AQ_DEL_FDIR_FILTER;
> +		if (!list_empty(&adapter->adv_rss_list_head))
> +			adapter->aq_required |= IAVF_FLAG_AQ_DEL_ADV_RSS_CFG;
>   		adapter->aq_required |= IAVF_FLAG_AQ_DISABLE_QUEUES;
>   	}
>   
> @@ -4164,6 +4241,7 @@ static int iavf_open(struct net_device *netdev)
>   static int iavf_close(struct net_device *netdev)
>   {
>   	struct iavf_adapter *adapter = netdev_priv(netdev);
> +	u64 aq_to_restore;
>   	int status;
>   
>   	mutex_lock(&adapter->crit_lock);
> @@ -4176,6 +4254,29 @@ static int iavf_close(struct net_device *netdev)
>   	set_bit(__IAVF_VSI_DOWN, adapter->vsi.state);
>   	if (CLIENT_ENABLED(adapter))
>   		adapter->flags |= IAVF_FLAG_CLIENT_NEEDS_CLOSE;
> +	/* We cannot send IAVF_FLAG_AQ_GET_OFFLOAD_VLAN_V2_CAPS before
> +	 * IAVF_FLAG_AQ_DISABLE_QUEUES because in such case there is rtnl
> +	 * deadlock with adminq_task() until iavf_close timeouts. We must send
> +	 * IAVF_FLAG_AQ_GET_CONFIG before IAVF_FLAG_AQ_DISABLE_QUEUES to make
> +	 * disable queues possible for vf. Give only necessary flags to
> +	 * iavf_down and save other to set them right before iavf_close()
> +	 * returns, when IAVF_FLAG_AQ_DISABLE_QUEUES will be already sent and
> +	 * iavf will be in DOWN state.
> +	 */
> +	aq_to_restore = adapter->aq_required;
> +	adapter->aq_required &= IAVF_FLAG_AQ_GET_CONFIG;
> +
> +	/* Remove flags which we do not want to send after close or we want to
> +	 * send before disable queues.
> +	 */
> +	aq_to_restore &= ~(IAVF_FLAG_AQ_GET_CONFIG		|
> +			   IAVF_FLAG_AQ_ENABLE_QUEUES		|
> +			   IAVF_FLAG_AQ_CONFIGURE_QUEUES	|
> +			   IAVF_FLAG_AQ_ADD_VLAN_FILTER		|
> +			   IAVF_FLAG_AQ_ADD_MAC_FILTER		|
> +			   IAVF_FLAG_AQ_ADD_CLOUD_FILTER	|
> +			   IAVF_FLAG_AQ_ADD_FDIR_FILTER		|
> +			   IAVF_FLAG_AQ_ADD_ADV_RSS_CFG);
>   
>   	iavf_down(adapter);
>   	iavf_change_state(adapter, __IAVF_DOWN_PENDING);
> @@ -4199,6 +4300,10 @@ static int iavf_close(struct net_device *netdev)
>   				    msecs_to_jiffies(500));
>   	if (!status)
>   		netdev_warn(netdev, "Device resources not yet released\n");
> +
> +	mutex_lock(&adapter->crit_lock);
> +	adapter->aq_required |= aq_to_restore;
> +	mutex_unlock(&adapter->crit_lock);
>   	return 0;
>   }
>   


More information about the Intel-wired-lan mailing list