michaelh/zephyr
1
0
Fork 0
Code Releases Activity

net/iface: Coalesce all DAD timers through one

Browse source 
This reduces the size of struct net_if_addr by 24 bytes by moving
the k_delayed_work attribute into net_if core code.
Then each net_if_addr can be added to the timer handler via a slist.

This does not make much gain if the system has only 1 unicast IPv6
address. It's a nice memory improvment once it has 2+ unicast IPv6
address. Note that having IPv4 enabled along with IPv6 will also see
memory improvements since both IPv6 and IPv4 use the same struct
net_if_addr.

Fixes #8728

Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
This commit is contained in:
Tomasz Bursztyka 2019-05-21 11:12:45 +02:00 • committed by Jukka Rissanen
commit 09e7262be8
2 changed files with 64 additions and 27 deletions
Download patch file Download diff file Expand all files Collapse all files

3
include/net/net_if.h
View file

@ -56,7 +56,8 @@ struct net_if_addr {
#if defined(CONFIG_NET_IPV6_DAD)
/** Duplicate address detection (DAD) timer */
struct k_delayed_work dad_timer;
sys_snode_t dad_node;
u32_t dad_start;
#endif
/** How the IP address was set */
enum net_addr_type addr_type;

76
subsys/net/ip/net_if.c
View file

@ -59,6 +59,12 @@ static struct k_delayed_work prefix_lifetime_timer;
/* Track currently active IPv6 prefix lifetime timers */
static sys_slist_t active_prefix_lifetime_timers;
#if defined(CONFIG_NET_IPV6_DAD)
/** Duplicate address detection (DAD) timer */
static struct k_delayed_work dad_timer;
static sys_slist_t active_dad_timers;
#endif
static struct {
struct net_if_ipv6 ipv6;
struct net_if *iface;
@ -570,32 +576,56 @@ static void join_mcast_nodes(struct net_if *iface, struct in6_addr *addr)
static void dad_timeout(struct k_work *work)
{
/* This means that the DAD succeed. */
struct net_if_addr *tmp, *ifaddr = CONTAINER_OF(work,
struct net_if_addr,
dad_timer);
struct net_if *iface = NULL;
u32_t current_time = k_uptime_get_32();
struct net_if_addr *ifaddr, *next;
ARG_UNUSED(work);
SYS_SLIST_FOR_EACH_CONTAINER_SAFE(&active_dad_timers,
ifaddr, next, dad_node) {
struct net_if_addr *tmp;
struct net_if *iface;
if ((s32_t)(ifaddr->dad_start +
DAD_TIMEOUT - current_time) > 0) {
break;
}
/* Removing the ifaddr from active_dad_timers list */
sys_slist_remove(&active_dad_timers, NULL, &ifaddr->dad_node);
NET_DBG("DAD succeeded for %s",
log_strdup(net_sprint_ipv6_addr(&ifaddr->address.in6_addr)));
log_strdup(net_sprint_ipv6_addr(
&ifaddr->address.in6_addr)));
ifaddr->addr_state = NET_ADDR_PREFERRED;
/* Because we do not know the interface at this point, we need to
* lookup for it.
/* Because we do not know the interface at this point,
* we need to lookup for it.
*/
tmp = net_if_ipv6_addr_lookup(&ifaddr->address.in6_addr, &iface);
iface = NULL;
tmp = net_if_ipv6_addr_lookup(&ifaddr->address.in6_addr,
&iface);
if (tmp == ifaddr) {
net_mgmt_event_notify_with_info(NET_EVENT_IPV6_DAD_SUCCEED,
iface,
&ifaddr->address.in6_addr,
net_mgmt_event_notify_with_info(
NET_EVENT_IPV6_DAD_SUCCEED,
iface, &ifaddr->address.in6_addr,
sizeof(struct in6_addr));
/* The address gets added to neighbor cache which is not needed
* in this case as the address is our own one.
/* The address gets added to neighbor cache which is not
* needed in this case as the address is our own one.
*/
net_ipv6_nbr_rm(iface, &ifaddr->address.in6_addr);
}
ifaddr = NULL;
}
if (ifaddr) {
k_delayed_work_submit(&dad_timer,
ifaddr->dad_start +
DAD_TIMEOUT - current_time);
}
}
static void net_if_ipv6_start_dad(struct net_if *iface,
@ -615,7 +645,12 @@ static void net_if_ipv6_start_dad(struct net_if *iface,
ifaddr->dad_count = 1U;
if (!net_ipv6_start_dad(iface, ifaddr)) {
k_delayed_work_submit(&ifaddr->dad_timer, DAD_TIMEOUT);
ifaddr->dad_start = k_uptime_get_32();
sys_slist_append(&active_dad_timers, &ifaddr->dad_node);
if (!k_delayed_work_remaining_get(&dad_timer)) {
k_delayed_work_submit(&dad_timer, DAD_TIMEOUT);
}
}
} else {
NET_DBG("Interface %p is down, starting DAD for %s later.",
@ -676,7 +711,7 @@ void net_if_ipv6_dad_failed(struct net_if *iface, const struct in6_addr *addr)
return;
}
k_delayed_work_cancel(&ifaddr->dad_timer);
sys_slist_find_and_remove(&active_dad_timers, &ifaddr->dad_node);
net_mgmt_event_notify_with_info(NET_EVENT_IPV6_DAD_FAILED, iface,
&ifaddr->address.in6_addr,
@ -685,9 +720,10 @@ void net_if_ipv6_dad_failed(struct net_if *iface, const struct in6_addr *addr)
net_if_ipv6_addr_rm(iface, addr);
}
static inline void iface_ipv6_dad_init(struct net_if_addr *ifaddr)
static inline void iface_ipv6_dad_init(void)
{
k_delayed_work_init(&ifaddr->dad_timer, dad_timeout);
k_delayed_work_init(&dad_timer, dad_timeout);
sys_slist_init(&active_dad_timers);
}
#else
@ -1027,8 +1063,6 @@ static inline void net_if_addr_init(struct net_if_addr *ifaddr,
ifaddr->addr_type = addr_type;
net_ipaddr_copy(&ifaddr->address.in6_addr, addr);
iface_ipv6_dad_init(ifaddr);
/* FIXME - set the mcast addr for this node */
if (vlifetime) {
@ -2172,6 +2206,8 @@ static void iface_ipv6_init(int if_count)
{
int i;
iface_ipv6_dad_init();
k_delayed_work_init(&address_lifetime_timer, address_lifetime_timeout);
k_delayed_work_init(&prefix_lifetime_timer, prefix_lifetime_timeout);