From cfd08f09723c5408eb3025b945fff08a99343911 Mon Sep 17 00:00:00 2001 From: Dragos Tatulea Date: Wed, 13 May 2026 15:45:18 +0300 Subject: [PATCH] IB/IPoIB: ndo_set_rx_mode_async conversion The commit in the fixes tag added a warning for devices that are netdev ops locked that they should be converted to .ndo_set_rx_mode_async. IPoIB for mlx5 is such a driver which was missed during the conversion because the flow is more complex: - mlx5 part of IPoIB device was converted to ops-lock in commit [1]. - ipoib_intf_init() then overrides netdev_ops with ipoib_netdev_ops_{pf,vf}, which still wired ndo_set_rx_mode to the legacy sync path -- tripping the new warning on every probe. So now we have the following splat: netdevice: ib0 (uninitialized): ops-locked drivers should use ndo_set_rx_mode_async WARNING: net/core/dev.c:11366 at register_netdevice+0x83c/0x21d0 ... register_netdev+0x1f/0x40 ipoib_add_one+0x35c/0x880 [ib_ipoib] This patch implements .ndo_set_rx_mode_async but it simply schedules the multicast restart task like before. This is done to maintain the assumption that this task and others [2] must run on the same order workqueue to avoid racing with themselves. The race between ipoib_mcast_join_task() and ipoib_mcast_restart_task() would be the most obvious example. [1] 8f7b00307bf1, "net/mlx5e: Convert mlx5 netdevs to instance locking") [2] ipoib_mcast_join_task, ipoib_mcast_restart_task, ipoib_mcast_carrier_on_task, ipoib_reap_ah, ipoib_reap_neigh Fixes: 3cbd22938877 ("net: warn ops-locked drivers still using ndo_set_rx_mode") Signed-off-by: Dragos Tatulea Reviewed-by: Cosmin Ratiu Acked-by: Jason Gunthorpe Link: https://patch.msgid.link/20260513124519.3357165-1-dtatulea@nvidia.com Signed-off-by: Jakub Kicinski --- drivers/infiniband/ulp/ipoib/ipoib_main.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/drivers/infiniband/ulp/ipoib/ipoib_main.c b/drivers/infiniband/ulp/ipoib/ipoib_main.c index 402671567736..3e1e1e861739 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib_main.c +++ b/drivers/infiniband/ulp/ipoib/ipoib_main.c @@ -1297,7 +1297,9 @@ static int ipoib_hard_header(struct sk_buff *skb, return IPOIB_HARD_LEN; } -static void ipoib_set_mcast_list(struct net_device *dev) +static void ipoib_set_rx_mode_async(struct net_device *dev, + struct netdev_hw_addr_list *uc, + struct netdev_hw_addr_list *mc) { struct ipoib_dev_priv *priv = ipoib_priv(dev); @@ -2160,7 +2162,7 @@ static const struct net_device_ops ipoib_netdev_ops_pf = { .ndo_fix_features = ipoib_fix_features, .ndo_start_xmit = ipoib_start_xmit, .ndo_tx_timeout = ipoib_timeout, - .ndo_set_rx_mode = ipoib_set_mcast_list, + .ndo_set_rx_mode_async = ipoib_set_rx_mode_async, .ndo_get_iflink = ipoib_get_iflink, .ndo_set_vf_link_state = ipoib_set_vf_link_state, .ndo_get_vf_config = ipoib_get_vf_config, @@ -2183,7 +2185,7 @@ static const struct net_device_ops ipoib_netdev_ops_vf = { .ndo_fix_features = ipoib_fix_features, .ndo_start_xmit = ipoib_start_xmit, .ndo_tx_timeout = ipoib_timeout, - .ndo_set_rx_mode = ipoib_set_mcast_list, + .ndo_set_rx_mode_async = ipoib_set_rx_mode_async, .ndo_get_iflink = ipoib_get_iflink, .ndo_get_stats64 = ipoib_get_stats, .ndo_eth_ioctl = ipoib_ioctl,