Since commit ac8c58f5b535 ("igb: fix deadlock caused by taking RTNL in RPM resume path") __igb_resume() doesn't take RTNL for runtime resume, but __igb_open() still calls netif_set_real_num_{tx,rx}_queues(), which require it. An asynchronous runtime resume (e.g. after the link comes back) then triggers "RTNL: assertion failed". Only set the real queue counts in __igb_open() when RTNL is held (open and system sleep resume), and set them from a work item taking RTNL on runtime resume. Until it runs, after a fallback to fewer queues on resume (MSI-X to MSI), the stack can use queues the device doesn't have: drop those packets in igb_xmit_frame() instead of mapping several stack queues, each with its own xmit lock, onto the same ring. The stack can't be kept off those queues instead, as __igb_open() and netif_device_attach() start all of them and changing the count needs RTNL. Tested with the 82576 emulated by QEMU. Fixes: ac8c58f5b535 ("igb: fix deadlock caused by taking RTNL in RPM resume path") Cc: stable@vger.kernel.org Closes: https://bugzilla.ipfire.org/show_bug.cgi?id=13562 Signed-off-by: Antonio Quartulli --- v3: - set the real queue counts when RTNL is held (open and system sleep resume), and from a work item on runtime resume, instead of not at all on resume (Janis, Sashiko) - until then, drop packets for queues the device doesn't have - stable: netif_set_real_num_queues() needs 5.15+ and dev_core_stats_tx_dropped_inc() 5.18+; older backports can use the separate tx/rx calls and netdev->stats.tx_dropped v2: - keep setting the queue counts in __igb_open() after igb_request_irq(), only skipping them on resume (Sashiko) - drop the reference to igc Note: igc has the same issue: it sets the real queue counts only in igc_open(), before requesting the IRQs, and never on resume. A fix will follow separately once it has been tested on igc hardware (Joe offered to test). drivers/net/ethernet/intel/igb/igb.h | 1 + drivers/net/ethernet/intel/igb/igb_main.c | 73 ++++++++++++++++------- 2 files changed, 54 insertions(+), 20 deletions(-) diff --git a/drivers/net/ethernet/intel/igb/igb.h b/drivers/net/ethernet/intel/igb/igb.h index 8c9b020..79d7fcc 100644 --- a/drivers/net/ethernet/intel/igb/igb.h +++ b/drivers/net/ethernet/intel/igb/igb.h @@ -586,6 +586,7 @@ struct igb_adapter { struct work_struct reset_task; struct work_struct watchdog_task; + struct work_struct rtnl_resume_task; bool fc_autoneg; u8 tx_timeout_factor; struct timer_list blink_timer; diff --git a/drivers/net/ethernet/intel/igb/igb_main.c b/drivers/net/ethernet/intel/igb/igb_main.c index d4a897a..2f51edc 100644 --- a/drivers/net/ethernet/intel/igb/igb_main.c +++ b/drivers/net/ethernet/intel/igb/igb_main.c @@ -110,6 +110,7 @@ static void igb_init_queue_configuration(struct igb_adapter *adapter); static int igb_sw_init(struct igb_adapter *); int igb_open(struct net_device *); int igb_close(struct net_device *); +static void igb_rtnl_resume_task(struct work_struct *); static void igb_configure(struct igb_adapter *); static void igb_configure_tx(struct igb_adapter *); static void igb_configure_rx(struct igb_adapter *); @@ -3426,6 +3427,7 @@ static int igb_probe(struct pci_dev *pdev, const struct pci_device_id *ent) INIT_WORK(&adapter->reset_task, igb_reset_task); INIT_WORK(&adapter->watchdog_task, igb_watchdog_task); + INIT_WORK(&adapter->rtnl_resume_task, igb_rtnl_resume_task); /* Initialize link properties that are user-changeable */ adapter->fc_autoneg = true; @@ -3889,6 +3891,7 @@ static void igb_remove(struct pci_dev *pdev) cancel_work_sync(&adapter->reset_task); cancel_work_sync(&adapter->watchdog_task); + cancel_work_sync(&adapter->rtnl_resume_task); #ifdef CONFIG_IGB_DCA if (adapter->flags & IGB_FLAG_DCA_ENABLED) { @@ -4123,10 +4126,31 @@ static int igb_sw_init(struct igb_adapter *adapter) return 0; } +/* Some settings need RTNL, which isn't held when resuming from runtime + * suspend: apply them from here instead, if the device is still up. + */ +static void igb_rtnl_resume_task(struct work_struct *work) +{ + struct igb_adapter *adapter = container_of(work, struct igb_adapter, + rtnl_resume_task); + + rtnl_lock(); + if (!test_bit(__IGB_DOWN, &adapter->state) && + netif_running(adapter->netdev)) { + if (netif_set_real_num_queues(adapter->netdev, + adapter->num_tx_queues, + adapter->num_rx_queues)) + netdev_err(adapter->netdev, + "error setting real queue count\n"); + } + rtnl_unlock(); +} + /** * __igb_open - Called when a network interface is made active * @netdev: network interface device structure * @resuming: indicates whether we are in a resume call + * @rtnl_held: indicates whether RTNL is held * * Returns 0 on success, negative value on failure * @@ -4136,7 +4160,8 @@ static int igb_sw_init(struct igb_adapter *adapter) * handler is registered with the OS, the watchdog timer is started, * and the stack is notified that the interface is ready. **/ -static int __igb_open(struct net_device *netdev, bool resuming) +static int __igb_open(struct net_device *netdev, bool resuming, + bool rtnl_held) { struct igb_adapter *adapter = netdev_priv(netdev); struct pci_dev *pdev = adapter->pdev; @@ -4179,16 +4204,15 @@ static int __igb_open(struct net_device *netdev, bool resuming) if (err) goto err_req_irq; - /* Notify the stack of the actual queue counts. */ - err = netif_set_real_num_tx_queues(adapter->netdev, - adapter->num_tx_queues); - if (err) - goto err_set_queues; - - err = netif_set_real_num_rx_queues(adapter->netdev, - adapter->num_rx_queues); - if (err) - goto err_set_queues; + /* Notify the stack of the actual queue counts. This needs RTNL: on + * runtime resume igb_rtnl_resume_task() does it. + */ + if (rtnl_held) { + err = netif_set_real_num_queues(netdev, adapter->num_tx_queues, + adapter->num_rx_queues); + if (err) + goto err_set_queues; + } /* From here on the code is the same as igb_up() */ clear_bit(__IGB_DOWN, &adapter->state); @@ -4242,7 +4266,7 @@ err_setup_tx: int igb_open(struct net_device *netdev) { - return __igb_open(netdev, false); + return __igb_open(netdev, false, true); } /** @@ -6617,12 +6641,7 @@ cleanup_tx_tstamp: static inline struct igb_ring *igb_tx_queue_mapping(struct igb_adapter *adapter, struct sk_buff *skb) { - unsigned int r_idx = skb->queue_mapping; - - if (r_idx >= adapter->num_tx_queues) - r_idx = r_idx % adapter->num_tx_queues; - - return adapter->tx_ring[r_idx]; + return adapter->tx_ring[skb->queue_mapping]; } static netdev_tx_t igb_xmit_frame(struct sk_buff *skb, @@ -6636,6 +6655,17 @@ static netdev_tx_t igb_xmit_frame(struct sk_buff *skb, if (skb_put_padto(skb, 17)) return NETDEV_TX_OK; + /* After a fallback to fewer queues on runtime resume, the stack can + * use more queues than the device has until igb_rtnl_resume_task() + * updates it: drop those packets instead of sharing a ring between + * queues, which aren't serialized with each other. + */ + if (unlikely(skb->queue_mapping >= adapter->num_tx_queues)) { + dev_kfree_skb_any(skb); + dev_core_stats_tx_dropped_inc(netdev); + return NETDEV_TX_OK; + } + return igb_xmit_frame_ring(skb, igb_tx_queue_mapping(adapter, skb)); } @@ -9639,8 +9669,11 @@ static int __igb_resume(struct device *dev, bool rpm) if (!rpm) rtnl_lock(); - if (!err && netif_running(netdev)) - err = __igb_open(netdev, true); + if (!err && netif_running(netdev)) { + err = __igb_open(netdev, true, !rpm); + if (!err && rpm) + schedule_work(&adapter->rtnl_resume_task); + } if (!err) netif_device_attach(netdev); -- 2.56.1