Le 24/09/2026 à 03:15, Yuyang Huang a écrit :
> A multi-part RTM_GETMULTICAST dump of dev->mc resumes by position, so
> entries added or removed between two dump rounds can be skipped or
> repeated. The IPv4 and IPv6 dumps report that with NLM_F_DUMP_INTR by
> stamping cb->seq from a per netns generation counter combined with
> dev_base_seq, see inet_base_seq().
>
> Add the equivalent for the device multicast lists: a per netns counter
> bumped whenever an entry is added to or removed from any dev->mc. The
> list helpers do not know which device a list belongs to, so give
> netdev_hw_addr_list an owner set for dev->mc only and bump the counter
> of dev_net(owner) where entries are created and freed. That covers the
> dev_mc_* helpers, both lists of a sync, the hardware sync helpers
> drivers call from their rx mode callbacks or their own workers and the
> reconciliation after an asynchronous rx mode update, while snapshot
> and other lists have no owner and are not tracked. It is atomic since
> the writers only hold the address lock of their own device.
>
> Used by the following patch for the AF_PACKET multicast dump.
>
> Signed-off-by: Yuyang Huang <[email protected]>
> ---
> include/linux/netdevice.h | 3 +++
> include/net/net_namespace.h | 1 +
> net/core/dev_addr_lists.c | 21 +++++++++++++++++++++
> 3 files changed, 25 insertions(+)
>
> diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
> index 5d16737167ee..be3804cdbb2d 100644
> --- a/include/linux/netdevice.h
> +++ b/include/linux/netdevice.h
> @@ -256,6 +256,9 @@ struct netdev_hw_addr_list {
>
> /* Auxiliary tree for faster lookup on addition and deletion */
> struct rb_root tree;
> +
> + /* Set for dev->mc, the owning device of a tracked list */
> + struct net_device *owner;
There are several lists owned by a device (dev->dev_addrs, dev->uc, dev->mc,
dev->rx_mode_addr_cache), the name 'owner' does not reflect that it is set only
when the list is dev->mc.
I wonder if the name should be changed or if __hw_addr_init() should be updated
to always set owner when the list is owned by a device.
> };
>
> #define netdev_hw_addr_list_count(l) ((l)->count)
> diff --git a/include/net/net_namespace.h b/include/net/net_namespace.h
> index 46b4c67e2966..d8c681ab5c74 100644
> --- a/include/net/net_namespace.h
> +++ b/include/net/net_namespace.h
> @@ -71,6 +71,7 @@ struct net {
> spinlock_t rules_mod_lock;
>
> unsigned int dev_base_seq; /* protected by rtnl_mutex */
> + atomic_t dev_mc_genid; /* bumped on dev->mc changes */
> u32 ifindex;
>
> spinlock_t nsid_lock;
> diff --git a/net/core/dev_addr_lists.c b/net/core/dev_addr_lists.c
> index 08528ca0a8b3..18acd874fdbd 100644
> --- a/net/core/dev_addr_lists.c
> +++ b/net/core/dev_addr_lists.c
> @@ -16,6 +16,19 @@
>
> #include "dev.h"
>
> +/**
> + * __hw_addr_changed - account a change of a tracked address list
> + * @list: the address list an entry was added to or removed from
> + *
> + * Bumps the netns generation counter RTM_GETMULTICAST dumps use to detect
> + * changes of dev->mc between dump rounds. Untracked lists have no owner.
> + */
> +static void __hw_addr_changed(struct netdev_hw_addr_list *list)
> +{
> + if (list->owner)
> + atomic_inc(&dev_net(list->owner)->dev_mc_genid);
> +}
> +
> /*
> * General list handling functions
> */
> @@ -126,6 +139,7 @@ static int __hw_addr_add_ex(struct netdev_hw_addr_list
> *list,
>
> list_add_tail_rcu(&ha->list, &list->list);
> list->count++;
> + __hw_addr_changed(list);
list->count and dev_mc_genid evolves together. Perhaps some helpers would be
less error-prone for future patches?
__hw_addr_count_inc()
__hw_addr_count_dec()
__hw_addr_count_reset()
?
>
> return 0;
> }
> @@ -162,6 +176,7 @@ static int __hw_addr_del_entry(struct netdev_hw_addr_list
> *list,
> list_del_rcu(&ha->list);
> kfree_rcu(ha, rcu_head);
> list->count--;
> + __hw_addr_changed(list);
> return 0;
> }
>
> @@ -487,6 +502,8 @@ void __hw_addr_flush(struct netdev_hw_addr_list *list)
> {
> struct netdev_hw_addr *ha, *tmp;
>
> + if (list->count)
> + __hw_addr_changed(list);
> list->tree = RB_ROOT;
> list_for_each_entry_safe(ha, tmp, &list->list, list) {
> list_del_rcu(&ha->list);
> @@ -501,6 +518,7 @@ void __hw_addr_init(struct netdev_hw_addr_list *list)
> INIT_LIST_HEAD(&list->list);
> list->count = 0;
> list->tree = RB_ROOT;
> + list->owner = NULL;
> }
> EXPORT_SYMBOL(__hw_addr_init);
>
> @@ -612,6 +630,7 @@ void __hw_addr_list_reconcile(struct netdev_hw_addr_list
> *real_list,
> __hw_addr_insert(real_list, ref_ha,
> addr_len);
> real_list->count++;
> + __hw_addr_changed(real_list);
> }
> continue;
> }
> @@ -623,6 +642,7 @@ void __hw_addr_list_reconcile(struct netdev_hw_addr_list
> *real_list,
> list_del_rcu(&real_ha->list);
> kfree_rcu(real_ha, rcu_head);
> real_list->count--;
> + __hw_addr_changed(real_list);
> }
> }
>
> @@ -1177,6 +1197,7 @@ EXPORT_SYMBOL(dev_mc_flush);
> void dev_mc_init(struct net_device *dev)
> {
> __hw_addr_init(&dev->mc);
> + dev->mc.owner = dev;
> }
> EXPORT_SYMBOL(dev_mc_init);
>
