UNIXソケットバッファサイズ
sk->sk_sndbufに送信バッファサイズ、sk->sk_rcvbufに受信バッファサイズが設定され、unixソケットではDGRAMだとsk_sndbufサイズ以下のデータしか送信できません。STREAMではsk_sndbufサイズ以上のデータは、sk_sndbufサイズ毎に繰り返し送信されます。(ただし送信したデータがアクセスされないとウエイトします。)なお、UNIXソケットでは、受信バッファサイズにかかる実装はされていません。
/proc/sys/net/coreコールバック。
/proc/sys/net/coreコールバック。
static struct ctl_table net_core_table[] = {
#ifdef CONFIG_NET
{
.procname = "wmem_max",
.data = &sysctl_wmem_max,
.maxlen = sizeof(int),
.mode = 0644,
.proc_handler = proc_dointvec
},
{
.procname = "rmem_max",
.data = &sysctl_rmem_max,
.maxlen = sizeof(int),
.mode = 0644,
.proc_handler = proc_dointvec
},
{
.procname = "wmem_default",
.data = &sysctl_wmem_default,
.maxlen = sizeof(int),
.mode = 0644,
.proc_handler = proc_dointvec
},
{
.procname = "rmem_default",
.data = &sysctl_rmem_default,
.maxlen = sizeof(int),
.mode = 0644,
.proc_handler = proc_dointvec
},
:
};
初期送信受信バッファサイズはsysctl_wmem_default/sysctl_rmem_defaultで、サイズは設定値の2倍とし、送信範囲:SOCK_MIN_SNDBUF〜sysctl_wmem_max/受信範囲:SOCK_MIN_RCVBUF〜sysctl_rmem_maxtonとなります。
#define _SK_MEM_PACKETS 256
#define _SK_MEM_OVERHEAD SKB_TRUESIZE(256)
#define SK_WMEM_MAX (_SK_MEM_OVERHEAD * _SK_MEM_PACKETS)
#define SK_RMEM_MAX (_SK_MEM_OVERHEAD * _SK_MEM_PACKETS)
#define SOCK_MIN_SNDBUF 2048
#define SOCK_MIN_RCVBUF (2048 + sizeof(struct sk_buff))
__u32 sysctl_wmem_max __read_mostly = SK_WMEM_MAX;
__u32 sysctl_rmem_max __read_mostly = SK_RMEM_MAX;
__u32 sysctl_wmem_default __read_mostly = SK_WMEM_MAX;
__u32 sysctl_rmem_default __read_mostly = SK_RMEM_MAX;
void __init sk_init(void)
{
if (totalram_pages <= 4096) {
sysctl_wmem_max = 32767;
sysctl_rmem_max = 32767;
sysctl_wmem_default = 32767;
sysctl_rmem_default = 32767;
} else if (totalram_pages >= 131072) {
sysctl_wmem_max = 131071;
sysctl_rmem_max = 131071;
}
}
void sock_init_data(struct socket *sock, struct sock *sk)
{
skb_queue_head_init(&sk->sk_receive_queue);
skb_queue_head_init(&sk->sk_write_queue);
skb_queue_head_init(&sk->sk_error_queue);
#ifdef CONFIG_NET_DMA
skb_queue_head_init(&sk->sk_async_wait_queue);
#endif
sk->sk_send_head = NULL;
init_timer(&sk->sk_timer);
sk->sk_allocation = GFP_KERNEL;
sk->sk_rcvbuf = sysctl_rmem_default;
sk->sk_sndbuf = sysctl_wmem_default;
sk->sk_state = TCP_CLOSE;
sk_set_socket(sk, sock);
sock_set_flag(sk, SOCK_ZAPPED);
if (sock) {
sk->sk_type = sock->type;
sk->sk_wq = sock->wq;
sock->sk = sk;
} else
sk->sk_wq = NULL;
spin_lock_init(&sk->sk_dst_lock);
rwlock_init(&sk->sk_callback_lock);
lockdep_set_class_and_name(&sk->sk_callback_lock,
af_callback_keys + sk->sk_family,
af_family_clock_key_strings[sk->sk_family]);
sk->sk_state_change = sock_def_wakeup;
sk->sk_data_ready = sock_def_readable;
sk->sk_write_space = sock_def_write_space;
sk->sk_error_report = sock_def_error_report;
sk->sk_destruct = sock_def_destruct;
sk->sk_sndmsg_page = NULL;
sk->sk_sndmsg_off = 0;
sk->sk_peer_pid = NULL;
sk->sk_peer_cred = NULL;
sk->sk_write_pending = 0;
sk->sk_rcvlowat = 1;
sk->sk_rcvtimeo = MAX_SCHEDULE_TIMEOUT;
sk->sk_sndtimeo = MAX_SCHEDULE_TIMEOUT;
sk->sk_stamp = ktime_set(-1L, 0);
smp_wmb();
atomic_set(&sk->sk_refcnt, 1);
atomic_set(&sk->sk_drops, 0);
}
level=SOL_SOCKETなら、ジェネリックなsocketの属性値とするsock_setsockopt()で、そうでないならソケットコールバック.setsockoptによる実装となります。UNIXソケットは.setsockoptは実装されておらずエラーとなります。
#define SOL_SOCKET 1
SYSCALL_DEFINE5(setsockopt, int, fd, int, level, int, optname,
char __user *, optval, int, optlen)
{
int err, fput_needed;
struct socket *sock;
if (optlen < 0)
return -EINVAL;
sock = sockfd_lookup_light(fd, &err, &fput_needed);
if (sock != NULL) {
err = security_socket_setsockopt(sock, level, optname);
if (err)
goto out_put;
if (level == SOL_SOCKET)
err =
sock_setsockopt(sock, level, optname, optval,
optlen);
else
err =
sock->ops->setsockopt(sock, level, optname, optval,
optlen);
out_put:
fput_light(sock->file, fput_needed);
}
return err;
}
int sock_setsockopt(struct socket *sock, int level, int optname,
char __user *optval, unsigned int optlen)
{
struct sock *sk = sock->sk;
int val;
int valbool;
struct linger ling;
int ret = 0;
if (optname == SO_BINDTODEVICE)
return sock_bindtodevice(sk, optval, optlen);
if (optlen < sizeof(int))
return -EINVAL;
if (get_user(val, (int __user *)optval))
return -EFAULT;
valbool = val ? 1 : 0;
lock_sock(sk);
switch (optname) {
case SO_DEBUG:
if (val && !capable(CAP_NET_ADMIN))
ret = -EACCES;
else
sock_valbool_flag(sk, SOCK_DBG, valbool);
break;
case SO_REUSEADDR:
sk->sk_reuse = valbool;
break;
case SO_TYPE:
case SO_PROTOCOL:
case SO_DOMAIN:
case SO_ERROR:
ret = -ENOPROTOOPT;
break;
case SO_DONTROUTE:
sock_valbool_flag(sk, SOCK_LOCALROUTE, valbool);
break;
case SO_BROADCAST:
sock_valbool_flag(sk, SOCK_BROADCAST, valbool);
break;
case SO_SNDBUF:
if (val > sysctl_wmem_max)
val = sysctl_wmem_max;
set_sndbuf:
sk->sk_userlocks |= SOCK_SNDBUF_LOCK;
if ((val * 2) < SOCK_MIN_SNDBUF)
sk->sk_sndbuf = SOCK_MIN_SNDBUF;
else
sk->sk_sndbuf = val * 2;
sk->sk_write_space(sk);
break;
case SO_SNDBUFFORCE:
if (!capable(CAP_NET_ADMIN)) {
ret = -EPERM;
break;
}
goto set_sndbuf;
case SO_RCVBUF:
if (val > sysctl_rmem_max)
val = sysctl_rmem_max;
set_rcvbuf:
sk->sk_userlocks |= SOCK_RCVBUF_LOCK;
if ((val * 2) < SOCK_MIN_RCVBUF)
sk->sk_rcvbuf = SOCK_MIN_RCVBUF;
else
sk->sk_rcvbuf = val * 2;
break;
case SO_RCVBUFFORCE:
if (!capable(CAP_NET_ADMIN)) {
ret = -EPERM;
break;
}
goto set_rcvbuf;
case SO_KEEPALIVE:
#ifdef CONFIG_INET
if (sk->sk_protocol == IPPROTO_TCP)
tcp_set_keepalive(sk, valbool);
#endif
sock_valbool_flag(sk, SOCK_KEEPOPEN, valbool);
break;
case SO_OOBINLINE:
sock_valbool_flag(sk, SOCK_URGINLINE, valbool);
break;
case SO_NO_CHECK:
sk->sk_no_check = valbool;
break;
case SO_PRIORITY:
if ((val >= 0 && val <= 6) || capable(CAP_NET_ADMIN))
sk->sk_priority = val;
else
ret = -EPERM;
break;
case SO_LINGER:
if (optlen < sizeof(ling)) {
ret = -EINVAL; /* 1003.1g */
break;
}
if (copy_from_user(&ling, optval, sizeof(ling))) {
ret = -EFAULT;
break;
}
if (!ling.l_onoff)
sock_reset_flag(sk, SOCK_LINGER);
else {
#if (BITS_PER_LONG == 32)
if ((unsigned int)ling.l_linger >= MAX_SCHEDULE_TIMEOUT/HZ)
sk->sk_lingertime = MAX_SCHEDULE_TIMEOUT;
else
#endif
sk->sk_lingertime = (unsigned int)ling.l_linger * HZ;
sock_set_flag(sk, SOCK_LINGER);
}
break;
case SO_BSDCOMPAT:
sock_warn_obsolete_bsdism("setsockopt");
break;
case SO_PASSCRED:
if (valbool)
set_bit(SOCK_PASSCRED, &sock->flags);
else
clear_bit(SOCK_PASSCRED, &sock->flags);
break;
case SO_TIMESTAMP:
case SO_TIMESTAMPNS:
if (valbool) {
if (optname == SO_TIMESTAMP)
sock_reset_flag(sk, SOCK_RCVTSTAMPNS);
else
sock_set_flag(sk, SOCK_RCVTSTAMPNS);
sock_set_flag(sk, SOCK_RCVTSTAMP);
sock_enable_timestamp(sk, SOCK_TIMESTAMP);
} else {
sock_reset_flag(sk, SOCK_RCVTSTAMP);
sock_reset_flag(sk, SOCK_RCVTSTAMPNS);
}
break;
case SO_TIMESTAMPING:
if (val & ~SOF_TIMESTAMPING_MASK) {
ret = -EINVAL;
break;
}
sock_valbool_flag(sk, SOCK_TIMESTAMPING_TX_HARDWARE,
val & SOF_TIMESTAMPING_TX_HARDWARE);
sock_valbool_flag(sk, SOCK_TIMESTAMPING_TX_SOFTWARE,
val & SOF_TIMESTAMPING_TX_SOFTWARE);
sock_valbool_flag(sk, SOCK_TIMESTAMPING_RX_HARDWARE,
val & SOF_TIMESTAMPING_RX_HARDWARE);
if (val & SOF_TIMESTAMPING_RX_SOFTWARE)
sock_enable_timestamp(sk,
SOCK_TIMESTAMPING_RX_SOFTWARE);
else
sock_disable_timestamp(sk,
(1UL << SOCK_TIMESTAMPING_RX_SOFTWARE));
sock_valbool_flag(sk, SOCK_TIMESTAMPING_SOFTWARE,
val & SOF_TIMESTAMPING_SOFTWARE);
sock_valbool_flag(sk, SOCK_TIMESTAMPING_SYS_HARDWARE,
val & SOF_TIMESTAMPING_SYS_HARDWARE);
sock_valbool_flag(sk, SOCK_TIMESTAMPING_RAW_HARDWARE,
val & SOF_TIMESTAMPING_RAW_HARDWARE);
break;
case SO_RCVLOWAT:
if (val < 0)
val = INT_MAX;
sk->sk_rcvlowat = val ? : 1;
break;
case SO_RCVTIMEO:
ret = sock_set_timeout(&sk->sk_rcvtimeo, optval, optlen);
break;
case SO_SNDTIMEO:
ret = sock_set_timeout(&sk->sk_sndtimeo, optval, optlen);
break;
case SO_ATTACH_FILTER:
ret = -EINVAL;
if (optlen == sizeof(struct sock_fprog)) {
struct sock_fprog fprog;
ret = -EFAULT;
if (copy_from_user(&fprog, optval, sizeof(fprog)))
break;
ret = sk_attach_filter(&fprog, sk);
}
break;
case SO_DETACH_FILTER:
ret = sk_detach_filter(sk);
break;
case SO_PASSSEC:
if (valbool)
set_bit(SOCK_PASSSEC, &sock->flags);
else
clear_bit(SOCK_PASSSEC, &sock->flags);
break;
case SO_MARK:
if (!capable(CAP_NET_ADMIN))
ret = -EPERM;
else
sk->sk_mark = val;
break;
case SO_RXQ_OVFL:
sock_valbool_flag(sk, SOCK_RXQ_OVFL, valbool);
break;
case SO_WIFI_STATUS:
sock_valbool_flag(sk, SOCK_WIFI_STATUS, valbool);
break;
default:
ret = -ENOPROTOOPT;
break;
}
release_sock(sk);
return ret;
}
検証サンプル
[root@localhost c]# cat socketsize.c
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/types.h>
#include <sys/socket.h>
void set_buffsize(int sockfd, int optname, int buff_size);
void main() {
int s_socket, d_socket;
int val[2];
int len = sizeof(sizeof(int));
s_socket = socket(AF_UNIX, SOCK_STREAM, 0);
d_socket = socket(AF_UNIX, SOCK_DGRAM, 0);
len = sizeof(sizeof(int));
getsockopt(d_socket, SOL_SOCKET, SO_RCVBUF, &val[0], &len);
len = sizeof(sizeof(int));
getsockopt(d_socket, SOL_SOCKET, SO_SNDBUF, &val[1], &len);
printf("DGRAM INIT\n");
printf(" rcvbuf:%6d\n",val[0]);
printf(" sndbuf:%6d\n",val[1]);
len = sizeof(sizeof(int));
getsockopt(s_socket, SOL_SOCKET, SO_RCVBUF, &val[0], &len);
len = sizeof(sizeof(int));
getsockopt(s_socket, SOL_SOCKET, SO_SNDBUF, &val[1], &len);
printf("STREAM INIT\n");
printf(" rcvbuf:%6d\n",val[0]);
printf(" sndbuf:%6d\n",val[1]);
printf("DGRAM UPDATE\n");
set_buffsize(d_socket, SO_RCVBUF, 0);
set_buffsize(d_socket, SO_SNDBUF, 0);
set_buffsize(d_socket, SO_RCVBUF, 3000);
set_buffsize(d_socket, SO_SNDBUF, 3000);
set_buffsize(d_socket, SO_RCVBUF, 163840);
set_buffsize(d_socket, SO_SNDBUF, 163840)
set_buffsize(d_socket, SO_RCVBUF, 200000);
set_buffsize(d_socket, SO_SNDBUF, 200000);
printf("STREAM UPDATE\n");
set_buffsize(s_socket, SO_RCVBUF, 0);
set_buffsize(s_socket, SO_SNDBUF, 0);
set_buffsize(s_socket, SO_RCVBUF, 3000);
set_buffsize(s_socket, SO_SNDBUF, 3000);
set_buffsize(s_socket, SO_RCVBUF, 163840);
set_buffsize(s_socket, SO_SNDBUF, 163840);
set_buffsize(s_socket, SO_RCVBUF, 200000);
set_buffsize(s_socket, SO_SNDBUF, 200000);
close(s_socket);
close(d_socket);
}
void set_buffsize(int socket, int optname, int buf_size)
{
int val;
int len = sizeof(sizeof(int));
char *name;
val = buf_size;
setsockopt(socket, SOL_SOCKET, optname, (char *)&val, len);
getsockopt(socket, SOL_SOCKET, optname, &val, &len);
name = (optname == SO_RCVBUF)? "rcvbuf": "sndbuf";
printf(" %s:%6d by(%6d)\n",name, val, buf_size);
}
検証結果[root@localhost c]# cat /proc/sys/net/core/wmem_max 163840 [root@localhost c]# cat /proc/sys/net/core/rmem_max 163840 [root@localhost c]# cat /proc/sys/net/core/wmem_default 163840 [root@localhost c]# cat /proc/sys/net/core/rmem_default 163840 [root@localhost c]# ./socketsize.o DGRAM INIT rcvbuf:163840 sndbuf:163840 STREAM INIT rcvbuf:163840 sndbuf:163840 DGRAM UPDATE rcvbuf: 2240 by( 0) sndbuf: 2048 by( 0) rcvbuf: 6000 by( 3000) sndbuf: 6000 by( 3000) rcvbuf:327680 by(163840) sndbuf:327680 by(163840) rcvbuf:327680 by(200000) sndbuf:327680 by(200000) STREAM UPDATE rcvbuf: 2240 by( 0) sndbuf: 2048 by( 0) rcvbuf: 6000 by( 3000) sndbuf: 6000 by( 3000) rcvbuf:327680 by(163840) sndbuf:327680 by(163840) rcvbuf:327680 by(200000) sndbuf:327680 by(200000)
