mirror of
https://github.com/AuxXxilium/linux_dsm_epyc7002.git
synced 2024-11-25 01:10:53 +07:00
1da177e4c3
Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
110 lines
3.1 KiB
C
110 lines
3.1 KiB
C
/*
|
|
* DECnet An implementation of the DECnet protocol suite for the LINUX
|
|
* operating system. DECnet is implemented using the BSD Socket
|
|
* interface as the means of communication with the user level.
|
|
*
|
|
* DECnet Socket Timer Functions
|
|
*
|
|
* Author: Steve Whitehouse <SteveW@ACM.org>
|
|
*
|
|
*
|
|
* Changes:
|
|
* Steve Whitehouse : Made keepalive timer part of the same
|
|
* timer idea.
|
|
* Steve Whitehouse : Added checks for sk->sock_readers
|
|
* David S. Miller : New socket locking
|
|
* Steve Whitehouse : Timer grabs socket ref.
|
|
*/
|
|
#include <linux/net.h>
|
|
#include <linux/socket.h>
|
|
#include <linux/skbuff.h>
|
|
#include <linux/netdevice.h>
|
|
#include <linux/timer.h>
|
|
#include <linux/spinlock.h>
|
|
#include <net/sock.h>
|
|
#include <asm/atomic.h>
|
|
#include <net/flow.h>
|
|
#include <net/dn.h>
|
|
|
|
/*
|
|
* Slow timer is for everything else (n * 500mS)
|
|
*/
|
|
|
|
#define SLOW_INTERVAL (HZ/2)
|
|
|
|
static void dn_slow_timer(unsigned long arg);
|
|
|
|
void dn_start_slow_timer(struct sock *sk)
|
|
{
|
|
sk->sk_timer.expires = jiffies + SLOW_INTERVAL;
|
|
sk->sk_timer.function = dn_slow_timer;
|
|
sk->sk_timer.data = (unsigned long)sk;
|
|
|
|
add_timer(&sk->sk_timer);
|
|
}
|
|
|
|
void dn_stop_slow_timer(struct sock *sk)
|
|
{
|
|
del_timer(&sk->sk_timer);
|
|
}
|
|
|
|
static void dn_slow_timer(unsigned long arg)
|
|
{
|
|
struct sock *sk = (struct sock *)arg;
|
|
struct dn_scp *scp = DN_SK(sk);
|
|
|
|
sock_hold(sk);
|
|
bh_lock_sock(sk);
|
|
|
|
if (sock_owned_by_user(sk)) {
|
|
sk->sk_timer.expires = jiffies + HZ / 10;
|
|
add_timer(&sk->sk_timer);
|
|
goto out;
|
|
}
|
|
|
|
/*
|
|
* The persist timer is the standard slow timer used for retransmits
|
|
* in both connection establishment and disconnection as well as
|
|
* in the RUN state. The different states are catered for by changing
|
|
* the function pointer in the socket. Setting the timer to a value
|
|
* of zero turns it off. We allow the persist_fxn to turn the
|
|
* timer off in a permant way by returning non-zero, so that
|
|
* timer based routines may remove sockets. This is why we have a
|
|
* sock_hold()/sock_put() around the timer to prevent the socket
|
|
* going away in the middle.
|
|
*/
|
|
if (scp->persist && scp->persist_fxn) {
|
|
if (scp->persist <= SLOW_INTERVAL) {
|
|
scp->persist = 0;
|
|
|
|
if (scp->persist_fxn(sk))
|
|
goto out;
|
|
} else {
|
|
scp->persist -= SLOW_INTERVAL;
|
|
}
|
|
}
|
|
|
|
/*
|
|
* Check for keepalive timeout. After the other timer 'cos if
|
|
* the previous timer caused a retransmit, we don't need to
|
|
* do this. scp->stamp is the last time that we sent a packet.
|
|
* The keepalive function sends a link service packet to the
|
|
* other end. If it remains unacknowledged, the standard
|
|
* socket timers will eventually shut the socket down. Each
|
|
* time we do this, scp->stamp will be updated, thus
|
|
* we won't try and send another until scp->keepalive has passed
|
|
* since the last successful transmission.
|
|
*/
|
|
if (scp->keepalive && scp->keepalive_fxn && (scp->state == DN_RUN)) {
|
|
if ((jiffies - scp->stamp) >= scp->keepalive)
|
|
scp->keepalive_fxn(sk);
|
|
}
|
|
|
|
sk->sk_timer.expires = jiffies + SLOW_INTERVAL;
|
|
|
|
add_timer(&sk->sk_timer);
|
|
out:
|
|
bh_unlock_sock(sk);
|
|
sock_put(sk);
|
|
}
|