
Currently, the BSD socket API is implemented in libc, translating the API calls to character driver operations underneath. This approach has several issues: - it is inefficient, as most character driver operations are specific to the socket type, thus requiring that each operation start by bruteforcing the socket protocol family and type of the given file descriptor using several system calls; - it requires that libc itself be changed every time system support for a new protocol is added; - various parts of the libc implementations violate the asynchronous signal safety POSIX requirements. In order to resolve all these issues at once, the plan is to turn the BSD socket calls into system calls, thus making the BSD socket API the "native" ABI, removing the complexity from libc and instead letting VFS deal with the socket calls. The overall change is going to break all networking functionality. In order to smoothen the transition, this patch introduces the fifteen new BSD socket system calls, and makes libc try these first before falling back on the old behavior. For now, the VFS implementations of the new calls fail such that libc will always use the fallback cases. Later on, when we introduce the actual implementation of the native BSD socket calls, all statically linked programs will automatically use the new ABI, thus limiting actual application breakage. In other words: by itself, this patch does nothing, except add a bit of transitional overhead that will disappear in the future. The largest part of the patch is concerned with adding full support for the new BSD socket system calls to trace(1) - this early addition has the advantage of making system call tracing output of several socket calls much more readable already. Both the system call interfaces and the trace(1) support have already been tested using code that will be committed later on. Change-Id: I3460812be50c78be662d857f9d3d6840f3ca917f
287 lines
6.5 KiB
C
287 lines
6.5 KiB
C
#include <sys/cdefs.h>
|
|
#include "namespace.h"
|
|
#include <lib.h>
|
|
|
|
#include <assert.h>
|
|
#include <errno.h>
|
|
#include <stdio.h>
|
|
#include <string.h>
|
|
#include <sys/ioctl.h>
|
|
#include <sys/socket.h>
|
|
#include <sys/types.h>
|
|
#include <sys/ucred.h>
|
|
#include <netinet/tcp.h>
|
|
|
|
#include <net/gen/in.h>
|
|
#include <net/gen/tcp.h>
|
|
#include <net/gen/tcp_io.h>
|
|
#include <net/gen/udp.h>
|
|
#include <net/gen/udp_io.h>
|
|
|
|
#include <minix/type.h>
|
|
|
|
#define DEBUG 0
|
|
|
|
static int _tcp_getsockopt(int sock, int level, int option_name,
|
|
void *__restrict option_value, socklen_t *__restrict option_len);
|
|
static int _udp_getsockopt(int sock, int level, int option_name,
|
|
void *__restrict option_value, socklen_t *__restrict option_len);
|
|
static int _uds_getsockopt(int sock, int level, int option_name,
|
|
void *__restrict option_value, socklen_t *__restrict option_len);
|
|
static void getsockopt_copy(void *return_value, size_t return_len,
|
|
void *__restrict option_value, socklen_t *__restrict option_len);
|
|
|
|
/*
|
|
* Get socket options.
|
|
*/
|
|
static int
|
|
__getsockopt(int fd, int level, int option_name,
|
|
void * __restrict option_value, socklen_t * __restrict option_len)
|
|
{
|
|
message m;
|
|
|
|
if (option_len == NULL) {
|
|
errno = EFAULT;
|
|
return -1;
|
|
}
|
|
|
|
memset(&m, 0, sizeof(m));
|
|
m.m_lc_vfs_sockopt.fd = fd;
|
|
m.m_lc_vfs_sockopt.level = level;
|
|
m.m_lc_vfs_sockopt.name = option_name;
|
|
m.m_lc_vfs_sockopt.buf = (vir_bytes)option_value;
|
|
m.m_lc_vfs_sockopt.len = *option_len;
|
|
|
|
if (_syscall(VFS_PROC_NR, VFS_GETSOCKOPT, &m) < 0)
|
|
return -1;
|
|
|
|
*option_len = m.m_vfs_lc_socklen.len;
|
|
return 0;
|
|
}
|
|
|
|
int getsockopt(int sock, int level, int option_name,
|
|
void *__restrict option_value, socklen_t *__restrict option_len)
|
|
{
|
|
int r;
|
|
nwio_tcpopt_t tcpopt;
|
|
nwio_udpopt_t udpopt;
|
|
struct sockaddr_un uds_addr;
|
|
|
|
r = __getsockopt(sock, level, option_name, option_value, option_len);
|
|
if (r != -1 || errno != ENOTSOCK)
|
|
return r;
|
|
|
|
r= ioctl(sock, NWIOGTCPOPT, &tcpopt);
|
|
if (r != -1 || errno != ENOTTY)
|
|
{
|
|
if (r == -1)
|
|
{
|
|
/* Bad file descriptor */
|
|
return -1;
|
|
}
|
|
return _tcp_getsockopt(sock, level, option_name,
|
|
option_value, option_len);
|
|
}
|
|
|
|
r= ioctl(sock, NWIOGUDPOPT, &udpopt);
|
|
if (r != -1 || errno != ENOTTY)
|
|
{
|
|
if (r == -1)
|
|
{
|
|
/* Bad file descriptor */
|
|
return -1;
|
|
}
|
|
return _udp_getsockopt(sock, level, option_name,
|
|
option_value, option_len);
|
|
}
|
|
|
|
r= ioctl(sock, NWIOGUDSADDR, &uds_addr);
|
|
if (r != -1 || errno != ENOTTY)
|
|
{
|
|
if (r == -1)
|
|
{
|
|
/* Bad file descriptor */
|
|
return -1;
|
|
}
|
|
return _uds_getsockopt(sock, level, option_name,
|
|
option_value, option_len);
|
|
}
|
|
|
|
errno = ENOTSOCK;
|
|
return -1;
|
|
}
|
|
|
|
static void getsockopt_copy(void *return_value, size_t return_len,
|
|
void *__restrict option_value, socklen_t *__restrict option_len)
|
|
{
|
|
/* copy as much data as possible */
|
|
if (*option_len < return_len)
|
|
memcpy(option_value, return_value, *option_len);
|
|
else
|
|
memcpy(option_value, return_value, return_len);
|
|
|
|
/* return length */
|
|
*option_len = return_len;
|
|
}
|
|
|
|
static int _tcp_getsockopt(int sock, int level, int option_name,
|
|
void *__restrict option_value, socklen_t *__restrict option_len)
|
|
{
|
|
int i, r, err;
|
|
|
|
if (level == SOL_SOCKET && option_name == SO_REUSEADDR)
|
|
{
|
|
i = 1; /* Binds to TIME_WAIT sockets never cause errors */
|
|
getsockopt_copy(&i, sizeof(i), option_value, option_len);
|
|
return 0;
|
|
}
|
|
if (level == SOL_SOCKET && option_name == SO_KEEPALIVE)
|
|
{
|
|
i = 1; /* Keepalive is always on */
|
|
getsockopt_copy(&i, sizeof(i), option_value, option_len);
|
|
return 0;
|
|
}
|
|
if (level == SOL_SOCKET && option_name == SO_ERROR)
|
|
{
|
|
r = ioctl(sock, NWIOTCPGERROR, &err);
|
|
if (r != 0)
|
|
return r;
|
|
|
|
getsockopt_copy(&err, sizeof(err), option_value, option_len);
|
|
return 0;
|
|
}
|
|
if (level == SOL_SOCKET && option_name == SO_RCVBUF)
|
|
{
|
|
i = 32 * 1024; /* Receive buffer in the current
|
|
* implementation
|
|
*/
|
|
getsockopt_copy(&i, sizeof(i), option_value, option_len);
|
|
return 0;
|
|
}
|
|
if (level == SOL_SOCKET && option_name == SO_SNDBUF)
|
|
{
|
|
i = 32 * 1024; /* Send buffer in the current implementation */
|
|
getsockopt_copy(&i, sizeof(i), option_value, option_len);
|
|
return 0;
|
|
}
|
|
if (level == SOL_SOCKET && option_name == SO_TYPE)
|
|
{
|
|
i = SOCK_STREAM; /* this is a TCP socket */
|
|
getsockopt_copy(&i, sizeof(i), option_value, option_len);
|
|
return 0;
|
|
}
|
|
if (level == IPPROTO_TCP && option_name == TCP_NODELAY)
|
|
{
|
|
i = 0; /* nodelay is always off */
|
|
getsockopt_copy(&i, sizeof(i), option_value, option_len);
|
|
return 0;
|
|
}
|
|
#if DEBUG
|
|
fprintf(stderr, "_tcp_getsocketopt: level %d, name %d\n",
|
|
level, option_name);
|
|
#endif
|
|
|
|
errno= ENOPROTOOPT;
|
|
return -1;
|
|
}
|
|
|
|
static int _udp_getsockopt(int sock, int level, int option_name,
|
|
void *__restrict option_value, socklen_t *__restrict option_len)
|
|
{
|
|
int i;
|
|
|
|
if (level == SOL_SOCKET && option_name == SO_TYPE)
|
|
{
|
|
i = SOCK_DGRAM; /* this is a UDP socket */
|
|
getsockopt_copy(&i, sizeof(i), option_value, option_len);
|
|
return 0;
|
|
}
|
|
#if DEBUG
|
|
fprintf(stderr, "_udp_getsocketopt: level %d, name %d\n",
|
|
level, option_name);
|
|
#endif
|
|
|
|
errno= ENOSYS;
|
|
return -1;
|
|
}
|
|
|
|
static int _uds_getsockopt(int sock, int level, int option_name,
|
|
void *__restrict option_value, socklen_t *__restrict option_len)
|
|
{
|
|
int i, r;
|
|
size_t size;
|
|
|
|
if (level == SOL_SOCKET && option_name == SO_RCVBUF)
|
|
{
|
|
r= ioctl(sock, NWIOGUDSRCVBUF, &size);
|
|
if (r == -1) {
|
|
return r;
|
|
}
|
|
|
|
getsockopt_copy(&size, sizeof(size), option_value, option_len);
|
|
return 0;
|
|
}
|
|
|
|
if (level == SOL_SOCKET && option_name == SO_SNDBUF)
|
|
{
|
|
r= ioctl(sock, NWIOGUDSSNDBUF, &size);
|
|
if (r == -1) {
|
|
return r;
|
|
}
|
|
|
|
getsockopt_copy(&size, sizeof(size), option_value, option_len);
|
|
return 0;
|
|
}
|
|
|
|
if (level == SOL_SOCKET && option_name == SO_TYPE)
|
|
{
|
|
r= ioctl(sock, NWIOGUDSSOTYPE, &i);
|
|
if (r == -1) {
|
|
return r;
|
|
}
|
|
|
|
getsockopt_copy(&i, sizeof(i), option_value, option_len);
|
|
return 0;
|
|
}
|
|
|
|
if (level == SOL_SOCKET && option_name == SO_PEERCRED)
|
|
{
|
|
struct uucred cred;
|
|
|
|
r= ioctl(sock, NWIOGUDSPEERCRED, &cred);
|
|
if (r == -1) {
|
|
return -1;
|
|
}
|
|
|
|
getsockopt_copy(&cred, sizeof(struct uucred), option_value,
|
|
option_len);
|
|
return 0;
|
|
}
|
|
|
|
|
|
if (level == SOL_SOCKET && option_name == SO_REUSEADDR)
|
|
{
|
|
i = 1; /* as long as nobody is listen()ing on the address,
|
|
* it can be reused without waiting for a
|
|
* timeout to expire.
|
|
*/
|
|
getsockopt_copy(&i, sizeof(i), option_value, option_len);
|
|
return 0;
|
|
}
|
|
|
|
if (level == SOL_SOCKET && option_name == SO_PASSCRED)
|
|
{
|
|
i = 1; /* option is always 'on' */
|
|
getsockopt_copy(&i, sizeof(i), option_value, option_len);
|
|
return 0;
|
|
}
|
|
|
|
#if DEBUG
|
|
fprintf(stderr, "_uds_getsocketopt: level %d, name %d\n",
|
|
level, option_name);
|
|
#endif
|
|
|
|
errno= ENOSYS;
|
|
return -1;
|
|
}
|