genode/repos/ports/include/noux_session/sysio.h

496 lines
14 KiB
C
Raw Normal View History

2011-12-22 16:19:25 +01:00
/*
* \brief Facility for passing system-call arguments
* \author Norman Feske
* \date 2011-02-15
*
* The 'Sysio' data structure is shared between the noux environment
* and the child. It is used to pass system-call arguments that would
* traditionally be transferred via 'copy_from_user' and 'copy_to_user'.
*/
/*
* Copyright (C) 2011-2017 Genode Labs GmbH
2011-12-22 16:19:25 +01:00
*
* This file is part of the Genode OS framework, which is distributed
* under the terms of the GNU Affero General Public License version 3.
2011-12-22 16:19:25 +01:00
*/
#ifndef _INCLUDE__NOUX_SESSION__SYSIO_H_
#define _INCLUDE__NOUX_SESSION__SYSIO_H_
/* Genode includes */
#include <os/ring_buffer.h>
2011-12-22 16:19:25 +01:00
#include <util/misc_math.h>
#include <vfs/file_system.h>
2011-12-22 16:19:25 +01:00
#define SYSIO_DECL(syscall_name, args, results) \
struct args syscall_name##_in; \
struct results syscall_name##_out;
namespace Noux {
using namespace Genode;
struct Sysio;
}
2011-12-22 16:19:25 +01:00
struct Noux::Sysio
{
/* signal numbers must match with libc signal numbers */
enum Signal {
2018-02-07 17:49:57 +01:00
SIG_INT = 2,
SIG_CHLD = 20,
SIG_WINCH = 28
};
2011-12-22 16:19:25 +01:00
enum { SIGNAL_QUEUE_SIZE = 32 };
Ring_buffer<enum Signal, SIGNAL_QUEUE_SIZE,
Ring_buffer_unsynchronized> pending_signals;
2011-12-22 16:19:25 +01:00
enum { MAX_PATH_LEN = 512 };
typedef char Path[MAX_PATH_LEN];
2011-12-22 16:19:25 +01:00
enum { CHUNK_SIZE = 64*1024 };
typedef char Chunk[CHUNK_SIZE];
2011-12-22 16:19:25 +01:00
enum { ARGS_MAX_LEN = 16*1024 };
typedef char Args[ARGS_MAX_LEN];
2011-12-22 16:19:25 +01:00
enum { ENV_MAX_LEN = 6*1024 };
typedef char Env[ENV_MAX_LEN];
2011-12-22 16:19:25 +01:00
typedef __SIZE_TYPE__ size_t;
typedef long int ssize_t;
/**
* Flags of 'mode' argument of open syscall
*/
enum {
OPEN_MODE_RDONLY = 0,
OPEN_MODE_WRONLY = 1,
OPEN_MODE_RDWR = 2,
OPEN_MODE_ACCMODE = 3,
OPEN_MODE_CREATE = 0x0800, /* libc O_EXCL */
};
2011-12-22 16:19:25 +01:00
typedef Vfs::Directory_service::Stat Stat;
/**
* Argument structure used for ioctl syscall
*/
struct Ioctl_in
{
typedef Vfs::File_io_service::Ioctl_opcode Opcode;
2011-12-22 16:19:25 +01:00
typedef Vfs::File_io_service::Ioctl_value Value;
2011-12-22 16:19:25 +01:00
Opcode request;
int argp;
};
2011-12-22 16:19:25 +01:00
/**
* Structure carrying the result values of 'ioctl' syscalls
*/
typedef Vfs::File_io_service::Ioctl_out Ioctl_out;
enum Lseek_whence { LSEEK_SET, LSEEK_CUR, LSEEK_END };
2011-12-22 16:19:25 +01:00
enum { DIRENT_MAX_NAME_LEN = Vfs::Directory_service::Dirent::Name::MAX_LEN };
typedef Vfs::Directory_service::Dirent_type Dirent_type;
2011-12-22 16:19:25 +01:00
/*
* Must be POD (in contrast to the VFS type) because it's used in a union
*/
struct Dirent
{
unsigned long fileno;
Dirent_type type;
char name[DIRENT_MAX_NAME_LEN];
Dirent & operator= (Vfs::Directory_service::Dirent const &dirent)
2011-12-22 16:19:25 +01:00
{
fileno = dirent.fileno;
type = dirent.type;
memcpy(name, dirent.name.buf, DIRENT_MAX_NAME_LEN);
2011-12-22 16:19:25 +01:00
return *this;
}
};
2011-12-22 16:19:25 +01:00
enum Fcntl_cmd {
FCNTL_CMD_GET_FILE_STATUS_FLAGS,
FCNTL_CMD_SET_FILE_STATUS_FLAGS,
FCNTL_CMD_SET_FD_FLAGS,
FCNTL_CMD_GET_FD_FLAGS
};
2011-12-22 16:19:25 +01:00
enum {
FCNTL_FILE_STATUS_FLAG_NONBLOCK = 4
};
/**
* Input and output argument type of select syscall
*/
struct Select_fds
{
/**
* Maximum number of file descriptors supported
*/
enum { MAX_FDS = 32U };
2011-12-22 16:19:25 +01:00
/**
* Number of file descriptors to watch for read operations (rd),
* write operations (wr), or exceptions (ex).
*/
size_t num_rd,
num_wr,
num_ex;
2011-12-22 16:19:25 +01:00
/**
* Array containing the file descriptors, starting with those
* referring to rd, followed by wr, and finally ex
*/
int array[MAX_FDS];
2011-12-22 16:19:25 +01:00
/**
* Return total number of file descriptors contained in the array
*/
size_t total_fds() const {
return min(num_rd + num_wr + num_ex, (size_t)MAX_FDS); }
2011-12-22 16:19:25 +01:00
/**
* Check for maximum population of fds array
*/
bool max_fds_exceeded() const
2011-12-22 16:19:25 +01:00
{
/*
* Note that even though the corner case of num_rd + num_wr +
* num_ex == MAX_FDS is technically valid, this condition hints
* at a possible attempt to over popupate the array (see the
* implementation of 'select' in the Noux libc plugin). Hence,
* we regard this case as an error, too.
2011-12-22 16:19:25 +01:00
*/
return total_fds() >= MAX_FDS;
}
2011-12-22 16:19:25 +01:00
/**
* Return true of fd set index should be watched for reading
*/
bool watch_for_rd(unsigned i) const { return i < num_rd; }
/**
* Return true if fd set index should be watched for writing
*/
bool watch_for_wr(unsigned i) const {
return (i >= num_rd) && (i < num_rd + num_wr); }
/**
* Return true if fd set index should be watched for exceptions
*/
bool watch_for_ex(unsigned i) const {
return (i >= num_rd + num_wr) && (i < total_fds()); }
};
struct Select_timeout
{
long sec, usec;
/**
* Set timeout to infinity
*/
void set_infinite() { sec = -1; usec = -1; }
/**
* Return true if the timeout is infinite
*/
bool infinite() const { return (sec == -1) && (usec == -1); }
/**
* Return true if the timeout is zero
*/
bool zero() const { return (sec == 0) && (usec == 0); }
};
/**
* Socket related structures
*/
enum { MAX_HOSTNAME_LEN = 255 };
typedef char Hostname[MAX_HOSTNAME_LEN];
enum { MAX_SERVNAME_LEN = 255 };
typedef char Servname[MAX_SERVNAME_LEN];
enum { MAX_ADDRINFO_RESULTS = 4 };
struct in_addr
{
unsigned int s_addr;
};
struct sockaddr
{
unsigned char sa_len;
unsigned char sa_family;
char sa_data[14];
};
struct sockaddr_in {
unsigned char sin_len;
unsigned char sin_family;
unsigned short sin_port;
struct in_addr sin_addr;
char sin_zero[8];
};
typedef unsigned socklen_t;
struct addrinfo {
int ai_flags;
int ai_family;
int ai_socktype;
int ai_protocol;
socklen_t ai_addrlen;
struct sockaddr *ai_addr;
char *ai_canonname;
struct addrinfo *ai_next;
};
struct Addrinfo {
struct addrinfo addrinfo;
struct sockaddr ai_addr;
char ai_canonname[255];
};
/**
* user info defintions
*/
enum { USERINFO_GET_ALL = 0, USERINFO_GET_UID, USERINFO_GET_GID };
enum { MAX_USERNAME_LEN = 32 };
typedef char User[MAX_USERNAME_LEN];
enum { MAX_SHELL_LEN = 16 };
typedef char Shell[MAX_SHELL_LEN];
enum { MAX_HOME_LEN = 128 };
typedef char Home[MAX_HOME_LEN];
typedef unsigned int Uid;
/**
* time/clock definitions
*/
enum Clock_Id { CLOCK_ID_SECOND };
enum Fcntl_error { FCNTL_ERR_CMD_INVALID = Vfs::Directory_service::NUM_GENERAL_ERRORS };
VFS: nonblocking interface The VFS library can be used in single-threaded or multi-threaded environments and depending on that, signals are handled by the same thread which uses the VFS library or possibly by a different thread. If a VFS plugin needs to block to wait for a signal, there is currently no way which works reliably in both environments. For this reason, this commit makes the interface of the VFS library nonblocking, similar to the File_system session interface. The most important changes are: - Directories are created and opened with the 'opendir()' function and the directory entries are read with the recently introduced 'queue_read()' and 'complete_read()' functions. - Symbolic links are created and opened with the 'openlink()' function and the link target is read with the 'queue_read()' and 'complete_read()' functions and written with the 'write()' function. - The 'write()' function does not wait for signals anymore. This can have the effect that data written by a VFS library user has not been processed by a file system server yet when the library user asks for the size of the file or closes it (both done with RPC functions at the file system server). For this reason, a user of the VFS library should request synchronization before calling 'stat()' or 'close()'. To make sure that a file system server has processed all write request packets which a client submitted before the synchronization request, synchronization is now requested at the file system server with a synchronization packet instead of an RPC function. Because of this change, the synchronization interface of the VFS library is now split into 'queue_sync()' and 'complete_sync()' functions. Fixes #2399
2017-08-15 20:51:53 +02:00
enum Mkdir_error { MKDIR_ERR_EXISTS, MKDIR_ERR_NO_ENTRY,
MKDIR_ERR_NO_SPACE, MKDIR_ERR_NO_PERM,
MKDIR_ERR_NAME_TOO_LONG };
enum Readlink_error { READLINK_ERR_NO_ENTRY, READLINK_ERR_NO_PERM };
enum Symlink_error { SYMLINK_ERR_EXISTS, SYMLINK_ERR_NO_ENTRY,
SYMLINK_ERR_NO_SPACE, SYMLINK_ERR_NO_PERM,
SYMLINK_ERR_NAME_TOO_LONG };
enum Execve_error { EXECVE_ERR_NO_ENTRY = Vfs::Directory_service::NUM_GENERAL_ERRORS,
EXECVE_ERR_NO_MEMORY,
EXECVE_ERR_NO_EXEC,
EXECVE_ERR_ACCESS};
enum Fork_error { FORK_NOMEM = Vfs::Directory_service::NUM_GENERAL_ERRORS };
enum Select_error { SELECT_ERR_INTERRUPT };
/**
* Socket related errors
*/
enum Accept_error { ACCEPT_ERR_AGAIN, ACCEPT_ERR_WOULD_BLOCK,
ACCEPT_ERR_INVALID, ACCEPT_ERR_NO_MEMORY,
ACCEPT_ERR_NOT_SUPPORTED };
enum Bind_error { BIND_ERR_ACCESS, BIND_ERR_ADDR_IN_USE,
BIND_ERR_INVALID, BIND_ERR_NO_MEMORY };
enum Connect_error { CONNECT_ERR_ACCESS, CONNECT_ERR_AGAIN,
CONNECT_ERR_ALREADY, CONNECT_ERR_CONN_REFUSED,
CONNECT_ERR_NO_PERM, CONNECT_ERR_ADDR_IN_USE,
CONNECT_ERR_IN_PROGRESS, CONNECT_ERR_IS_CONNECTED,
CONNECT_ERR_RESET, CONNECT_ERR_ABORTED,
CONNECT_ERR_NO_ROUTE };
enum Listen_error { LISTEN_ERR_ADDR_IN_USE, LISTEN_ERR_NOT_SUPPORTED };
enum Recv_error { RECV_ERR_AGAIN, RECV_ERR_WOULD_BLOCK,
RECV_ERR_CONN_REFUSED, RECV_ERR_INVALID,
RECV_ERR_NOT_CONNECTED, RECV_ERR_NO_MEMORY };
enum Send_error { SEND_ERR_AGAIN, SEND_ERR_WOULD_BLOCK,
SEND_ERR_CONNECTION_RESET, SEND_ERR_INVALID,
SEND_ERR_IS_CONNECTED, SEND_ERR_NO_MEMORY };
enum Shutdown_error { SHUTDOWN_ERR_NOT_CONNECTED };
enum Socket_error { SOCKET_ERR_ACCESS, SOCKET_ERR_NO_AF_SUPPORT,
SOCKET_ERR_INVALID, SOCKET_ERR_NO_MEMORY };
enum Clock_error { CLOCK_ERR_INVALID, CLOCK_ERR_FAULT, CLOCK_ERR_NO_PERM };
enum Utimes_error { UTIMES_ERR_ACCESS, UTIMES_ERR_FAUL, UTIMES_ERR_EIO,
UTIMES_ERR_NAME_TOO_LONG, UTIMES_ERR_NO_ENTRY,
UTIMES_ERR_NOT_DIRECTORY, UTIMES_ERR_NO_PERM,
UTIMES_ERR_READ_ONLY };
enum Wait4_error { WAIT4_ERR_INTERRUPT };
enum Kill_error { KILL_ERR_SRCH };
union {
Vfs::Directory_service::General_error general;
Vfs::Directory_service::Stat_result stat;
Vfs::File_io_service::Ftruncate_result ftruncate;
Vfs::Directory_service::Open_result open;
Vfs::Directory_service::Unlink_result unlink;
Vfs::Directory_service::Rename_result rename;
Vfs::File_io_service::Read_result read;
Vfs::File_io_service::Write_result write;
Vfs::File_io_service::Ioctl_result ioctl;
Fcntl_error fcntl;
VFS: nonblocking interface The VFS library can be used in single-threaded or multi-threaded environments and depending on that, signals are handled by the same thread which uses the VFS library or possibly by a different thread. If a VFS plugin needs to block to wait for a signal, there is currently no way which works reliably in both environments. For this reason, this commit makes the interface of the VFS library nonblocking, similar to the File_system session interface. The most important changes are: - Directories are created and opened with the 'opendir()' function and the directory entries are read with the recently introduced 'queue_read()' and 'complete_read()' functions. - Symbolic links are created and opened with the 'openlink()' function and the link target is read with the 'queue_read()' and 'complete_read()' functions and written with the 'write()' function. - The 'write()' function does not wait for signals anymore. This can have the effect that data written by a VFS library user has not been processed by a file system server yet when the library user asks for the size of the file or closes it (both done with RPC functions at the file system server). For this reason, a user of the VFS library should request synchronization before calling 'stat()' or 'close()'. To make sure that a file system server has processed all write request packets which a client submitted before the synchronization request, synchronization is now requested at the file system server with a synchronization packet instead of an RPC function. Because of this change, the synchronization interface of the VFS library is now split into 'queue_sync()' and 'complete_sync()' functions. Fixes #2399
2017-08-15 20:51:53 +02:00
Mkdir_error mkdir;
Readlink_error readlink;
Symlink_error symlink;
Execve_error execve;
Select_error select;
Accept_error accept;
Bind_error bind;
Connect_error connect;
Listen_error listen;
Recv_error recv;
Send_error send;
Shutdown_error shutdown;
Socket_error socket;
Clock_error clock;
Utimes_error utimes;
Wait4_error wait4;
Kill_error kill;
Fork_error fork;
} error;
2011-12-22 16:19:25 +01:00
union {
2011-12-22 16:19:25 +01:00
SYSIO_DECL(write, { int fd; size_t count; Chunk chunk; },
{ size_t count; });
2011-12-22 16:19:25 +01:00
SYSIO_DECL(stat, { Path path; }, { Stat st; });
2011-12-22 16:19:25 +01:00
SYSIO_DECL(symlink, { Path oldpath; Path newpath; }, { });
SYSIO_DECL(fstat, { int fd; }, { Stat st; });
2011-12-22 16:19:25 +01:00
SYSIO_DECL(ftruncate, { int fd; off_t length; }, { });
2011-12-22 16:19:25 +01:00
SYSIO_DECL(fcntl, { int fd; long long_arg; Fcntl_cmd cmd; },
{ int result; });
2011-12-22 16:19:25 +01:00
SYSIO_DECL(open, { Path path; int mode; }, { int fd; });
2011-12-22 16:19:25 +01:00
SYSIO_DECL(close, { int fd; }, { });
2011-12-22 16:19:25 +01:00
SYSIO_DECL(ioctl, : Ioctl_in { int fd; }, : Ioctl_out { });
SYSIO_DECL(lseek, { int fd; off_t offset; Lseek_whence whence; },
{ off_t offset; });
2011-12-22 16:19:25 +01:00
SYSIO_DECL(dirent, { int fd; }, { Dirent entry; });
2011-12-22 16:19:25 +01:00
SYSIO_DECL(read, { int fd; size_t count; },
{ Chunk chunk; size_t count; });
2011-12-22 16:19:25 +01:00
SYSIO_DECL(readlink, { Path path; size_t bufsiz; },
{ Chunk chunk; size_t count; });
SYSIO_DECL(execve, { Path filename; Args args; Env env; }, { });
2012-02-15 19:51:13 +01:00
SYSIO_DECL(select, { Select_fds fds; Select_timeout timeout; },
{ Select_fds fds; });
2012-02-15 19:51:13 +01:00
SYSIO_DECL(fork, { addr_t ip; addr_t sp;
addr_t parent_cap_addr; },
{ int pid; });
SYSIO_DECL(getpid, { }, { int pid; });
SYSIO_DECL(wait4, { int pid; bool nohang; },
{ int pid; int status; });
SYSIO_DECL(pipe, { }, { int fd[2]; });
SYSIO_DECL(dup2, { int fd; int to_fd; }, { int fd; });
SYSIO_DECL(unlink, { Path path; }, { });
SYSIO_DECL(rename, { Path from_path; Path to_path; }, { });
SYSIO_DECL(mkdir, { Path path; int mode; }, { });
SYSIO_DECL(socket, { int domain; int type; int protocol; },
{ int fd; });
/* XXX for now abuse Chunk for passing optval */
SYSIO_DECL(getsockopt, { int fd; int level; int optname; Chunk optval;
socklen_t optlen; }, { int result; });
SYSIO_DECL(setsockopt, { int fd; int level;
int optname; Chunk optval;
socklen_t optlen; }, { });
SYSIO_DECL(accept, { int fd; struct sockaddr addr; socklen_t addrlen; },
{ int fd; });
SYSIO_DECL(bind, { int fd; struct sockaddr addr;
socklen_t addrlen; }, { int result; });
SYSIO_DECL(getpeername, { int fd; struct sockaddr addr;
socklen_t addrlen; }, { });
SYSIO_DECL(listen, { int fd; int type; int backlog; },
{ int result; });
SYSIO_DECL(send, { int fd; Chunk buf; size_t len; int flags; },
{ ssize_t len; });
SYSIO_DECL(sendto, { int fd; Chunk buf; size_t len; int flags;
struct sockaddr dest_addr; socklen_t addrlen; },
{ ssize_t len; });
SYSIO_DECL(recv, { int fd; Chunk buf; size_t len; int flags; },
{ size_t len; });
SYSIO_DECL(recvfrom, { int fd; Chunk buf; size_t len; int flags;
struct sockaddr src_addr; socklen_t addrlen; },
{ size_t len; });
SYSIO_DECL(shutdown, { int fd; int how; }, { });
SYSIO_DECL(connect, { int fd; struct sockaddr addr; socklen_t addrlen; },
{ int result; });
SYSIO_DECL(userinfo, { int request; Uid uid; },
{ User name; Uid uid; Uid gid; Shell shell;
Home home; });
SYSIO_DECL(gettimeofday, { }, { unsigned long sec; unsigned int usec; });
SYSIO_DECL(clock_gettime, { Clock_Id clock_id; },
{ unsigned long sec; unsigned long nsec; });
SYSIO_DECL(utimes, { Path path; unsigned long sec; unsigned long usec; },
{ });
SYSIO_DECL(sync, { }, { });
SYSIO_DECL(kill, { int pid; Signal sig; }, { });
SYSIO_DECL(getdtablesize, { }, { int n; });
2011-12-22 16:19:25 +01:00
};
};
#undef SYSIO_DECL
#endif /* _INCLUDE__NOUX_SESSION__SYSIO_H_ */