
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
295 lines
5.6 KiB
C
295 lines
5.6 KiB
C
#include <sys/cdefs.h>
|
|
#include "namespace.h"
|
|
#include <lib.h>
|
|
|
|
#include <string.h>
|
|
#include <assert.h>
|
|
#include <errno.h>
|
|
#include <stdio.h>
|
|
#include <sys/ioctl.h>
|
|
#include <sys/socket.h>
|
|
#include <sys/types.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>
|
|
|
|
#define DEBUG 0
|
|
|
|
static int _tcp_setsockopt(int sock, int level, int option_name,
|
|
const void *option_value, socklen_t option_len);
|
|
|
|
static int _udp_setsockopt(int sock, int level, int option_name,
|
|
const void *option_value, socklen_t option_len);
|
|
|
|
static int _uds_setsockopt(int sock, int level, int option_name,
|
|
const void *option_value, socklen_t option_len);
|
|
|
|
/*
|
|
* Set socket options.
|
|
*/
|
|
static int
|
|
__setsockopt(int fd, int level, int option_name, const void * option_value,
|
|
socklen_t option_len)
|
|
{
|
|
message m;
|
|
|
|
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;
|
|
|
|
return _syscall(VFS_PROC_NR, VFS_SETSOCKOPT, &m);
|
|
}
|
|
|
|
int setsockopt(int sock, int level, int option_name,
|
|
const void *option_value, socklen_t option_len)
|
|
{
|
|
int r;
|
|
nwio_tcpopt_t tcpopt;
|
|
nwio_udpopt_t udpopt;
|
|
struct sockaddr_un uds_addr;
|
|
|
|
r = __setsockopt(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_setsockopt(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_setsockopt(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_setsockopt(sock, level, option_name,
|
|
option_value, option_len);
|
|
}
|
|
|
|
errno = ENOTSOCK;
|
|
return -1;
|
|
}
|
|
|
|
static int _tcp_setsockopt(int sock, int level, int option_name,
|
|
const void *option_value, socklen_t option_len)
|
|
{
|
|
int i;
|
|
|
|
if (level == SOL_SOCKET && option_name == SO_REUSEADDR)
|
|
{
|
|
if (option_len != sizeof(i))
|
|
{
|
|
errno= EINVAL;
|
|
return -1;
|
|
}
|
|
i= *(const int *)option_value;
|
|
if (!i)
|
|
{
|
|
/* At the moment there is no way to turn off
|
|
* reusing addresses.
|
|
*/
|
|
errno= ENOSYS;
|
|
return -1;
|
|
}
|
|
return 0;
|
|
}
|
|
if (level == SOL_SOCKET && option_name == SO_KEEPALIVE)
|
|
{
|
|
if (option_len != sizeof(i))
|
|
{
|
|
errno= EINVAL;
|
|
return -1;
|
|
}
|
|
i= *(const int *)option_value;
|
|
if (!i)
|
|
{
|
|
/* At the moment there is no way to turn off
|
|
* keepalives.
|
|
*/
|
|
errno= ENOSYS;
|
|
return -1;
|
|
}
|
|
return 0;
|
|
}
|
|
if (level == SOL_SOCKET && option_name == SO_RCVBUF)
|
|
{
|
|
if (option_len != sizeof(i))
|
|
{
|
|
errno= EINVAL;
|
|
return -1;
|
|
}
|
|
i= *(const int *)option_value;
|
|
if (i > 32*1024)
|
|
{
|
|
/* The receive buffer is limited to 32K at the moment.
|
|
*/
|
|
errno= ENOSYS;
|
|
return -1;
|
|
}
|
|
/* There is no way to reduce the receive buffer, do we have to
|
|
* let this call fail for smaller buffers?
|
|
*/
|
|
return 0;
|
|
}
|
|
if (level == SOL_SOCKET && option_name == SO_SNDBUF)
|
|
{
|
|
if (option_len != sizeof(i))
|
|
{
|
|
errno= EINVAL;
|
|
return -1;
|
|
}
|
|
i= *(const int *)option_value;
|
|
if (i > 32*1024)
|
|
{
|
|
/* The send buffer is limited to 32K at the moment.
|
|
*/
|
|
errno= ENOSYS;
|
|
return -1;
|
|
}
|
|
/* There is no way to reduce the send buffer, do we have to
|
|
* let this call fail for smaller buffers?
|
|
*/
|
|
return 0;
|
|
}
|
|
if (level == IPPROTO_TCP && option_name == TCP_NODELAY)
|
|
{
|
|
if (option_len != sizeof(i))
|
|
{
|
|
errno= EINVAL;
|
|
return -1;
|
|
}
|
|
i= *(const int *)option_value;
|
|
if (i)
|
|
{
|
|
/* At the moment there is no way to turn on
|
|
* nodelay.
|
|
*/
|
|
errno= ENOSYS;
|
|
return -1;
|
|
}
|
|
return 0;
|
|
}
|
|
#if DEBUG
|
|
fprintf(stderr, "_tcp_setsocketopt: level %d, name %d\n",
|
|
level, option_name);
|
|
#endif
|
|
|
|
errno= ENOSYS;
|
|
return -1;
|
|
}
|
|
|
|
static int _udp_setsockopt(int sock, int level, int option_name,
|
|
const void *option_value, socklen_t option_len)
|
|
{
|
|
#if DEBUG
|
|
fprintf(stderr, "_udp_setsocketopt: level %d, name %d\n",
|
|
level, option_name);
|
|
#endif
|
|
|
|
errno= ENOSYS;
|
|
return -1;
|
|
}
|
|
|
|
|
|
static int _uds_setsockopt(int sock, int level, int option_name,
|
|
const void *option_value, socklen_t option_len)
|
|
{
|
|
int i;
|
|
size_t size;
|
|
|
|
if (level == SOL_SOCKET && option_name == SO_RCVBUF)
|
|
{
|
|
if (option_len != sizeof(size))
|
|
{
|
|
errno= EINVAL;
|
|
return -1;
|
|
}
|
|
size= *(const size_t *)option_value;
|
|
return ioctl(sock, NWIOSUDSRCVBUF, &size);
|
|
}
|
|
|
|
if (level == SOL_SOCKET && option_name == SO_SNDBUF)
|
|
{
|
|
if (option_len != sizeof(size))
|
|
{
|
|
errno= EINVAL;
|
|
return -1;
|
|
}
|
|
size= *(const size_t *)option_value;
|
|
return ioctl(sock, NWIOSUDSSNDBUF, &size);
|
|
}
|
|
|
|
if (level == SOL_SOCKET && option_name == SO_REUSEADDR)
|
|
{
|
|
if (option_len != sizeof(i))
|
|
{
|
|
errno= EINVAL;
|
|
return -1;
|
|
}
|
|
i= *(const int *)option_value;
|
|
if (!i)
|
|
{
|
|
/* At the moment there is no way to turn off
|
|
* reusing addresses.
|
|
*/
|
|
errno= ENOSYS;
|
|
return -1;
|
|
}
|
|
return 0;
|
|
}
|
|
|
|
if (level == SOL_SOCKET && option_name == SO_PASSCRED)
|
|
{
|
|
if (option_len != sizeof(i))
|
|
{
|
|
errno= EINVAL;
|
|
return -1;
|
|
}
|
|
i= *(const int *)option_value;
|
|
if (!i)
|
|
{
|
|
/* credentials can always be received. */
|
|
errno= ENOSYS;
|
|
return -1;
|
|
}
|
|
return 0;
|
|
}
|
|
|
|
#if DEBUG
|
|
fprintf(stderr, "_uds_setsocketopt: level %d, name %d\n",
|
|
level, option_name);
|
|
#endif
|
|
|
|
errno= ENOSYS;
|
|
return -1;
|
|
}
|