matterbridge/vendor/golang.org/x/sys/unix/syscall_solaris.go

1134 lines
32 KiB
Go
Raw Normal View History

2017-03-25 20:45:10 +01:00
// Copyright 2009 The Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
// Solaris system calls.
// This file is compiled as ordinary Go code,
// but it is also input to mksyscall,
// which parses the //sys lines and generates system call stubs.
// Note that sometimes we use a lowercase //sys name and wrap
// it in our own nicer implementation, either here or in
// syscall_solaris.go or syscall_unix.go.
package unix
import (
"fmt"
"os"
2020-11-22 15:55:57 +01:00
"runtime"
"sync"
2017-03-25 20:45:10 +01:00
"syscall"
"unsafe"
)
// Implemented in runtime/syscall_solaris.go.
type syscallFunc uintptr
func rawSysvicall6(trap, nargs, a1, a2, a3, a4, a5, a6 uintptr) (r1, r2 uintptr, err syscall.Errno)
func sysvicall6(trap, nargs, a1, a2, a3, a4, a5, a6 uintptr) (r1, r2 uintptr, err syscall.Errno)
// SockaddrDatalink implements the Sockaddr interface for AF_LINK type sockets.
2017-03-25 20:45:10 +01:00
type SockaddrDatalink struct {
Family uint16
Index uint16
Type uint8
Nlen uint8
Alen uint8
Slen uint8
Data [244]int8
raw RawSockaddrDatalink
}
func direntIno(buf []byte) (uint64, bool) {
return readInt(buf, unsafe.Offsetof(Dirent{}.Ino), unsafe.Sizeof(Dirent{}.Ino))
}
func direntReclen(buf []byte) (uint64, bool) {
return readInt(buf, unsafe.Offsetof(Dirent{}.Reclen), unsafe.Sizeof(Dirent{}.Reclen))
}
func direntNamlen(buf []byte) (uint64, bool) {
reclen, ok := direntReclen(buf)
if !ok {
return 0, false
}
return reclen - uint64(unsafe.Offsetof(Dirent{}.Name)), true
}
2017-03-25 20:45:10 +01:00
//sysnb pipe(p *[2]_C_int) (n int, err error)
func Pipe(p []int) (err error) {
if len(p) != 2 {
return EINVAL
}
var pp [2]_C_int
n, err := pipe(&pp)
if n != 0 {
return err
}
if err == nil {
p[0] = int(pp[0])
p[1] = int(pp[1])
}
2017-03-25 20:45:10 +01:00
return nil
}
2021-03-20 22:40:23 +01:00
//sysnb pipe2(p *[2]_C_int, flags int) (err error)
func Pipe2(p []int, flags int) error {
if len(p) != 2 {
return EINVAL
}
var pp [2]_C_int
err := pipe2(&pp, flags)
if err == nil {
p[0] = int(pp[0])
p[1] = int(pp[1])
}
2021-03-20 22:40:23 +01:00
return err
}
2017-03-25 20:45:10 +01:00
func (sa *SockaddrInet4) sockaddr() (unsafe.Pointer, _Socklen, error) {
if sa.Port < 0 || sa.Port > 0xFFFF {
return nil, 0, EINVAL
}
sa.raw.Family = AF_INET
p := (*[2]byte)(unsafe.Pointer(&sa.raw.Port))
p[0] = byte(sa.Port >> 8)
p[1] = byte(sa.Port)
sa.raw.Addr = sa.Addr
2017-03-25 20:45:10 +01:00
return unsafe.Pointer(&sa.raw), SizeofSockaddrInet4, nil
}
func (sa *SockaddrInet6) sockaddr() (unsafe.Pointer, _Socklen, error) {
if sa.Port < 0 || sa.Port > 0xFFFF {
return nil, 0, EINVAL
}
sa.raw.Family = AF_INET6
p := (*[2]byte)(unsafe.Pointer(&sa.raw.Port))
p[0] = byte(sa.Port >> 8)
p[1] = byte(sa.Port)
sa.raw.Scope_id = sa.ZoneId
sa.raw.Addr = sa.Addr
2017-03-25 20:45:10 +01:00
return unsafe.Pointer(&sa.raw), SizeofSockaddrInet6, nil
}
func (sa *SockaddrUnix) sockaddr() (unsafe.Pointer, _Socklen, error) {
name := sa.Name
n := len(name)
if n >= len(sa.raw.Path) {
return nil, 0, EINVAL
}
sa.raw.Family = AF_UNIX
for i := 0; i < n; i++ {
sa.raw.Path[i] = int8(name[i])
}
// length is family (uint16), name, NUL.
sl := _Socklen(2)
if n > 0 {
sl += _Socklen(n) + 1
}
if sa.raw.Path[0] == '@' {
sa.raw.Path[0] = 0
// Don't count trailing NUL for abstract address.
sl--
}
return unsafe.Pointer(&sa.raw), sl, nil
}
//sys getsockname(fd int, rsa *RawSockaddrAny, addrlen *_Socklen) (err error) = libsocket.getsockname
func Getsockname(fd int) (sa Sockaddr, err error) {
var rsa RawSockaddrAny
var len _Socklen = SizeofSockaddrAny
if err = getsockname(fd, &rsa, &len); err != nil {
return
}
return anyToSockaddr(fd, &rsa)
}
// GetsockoptString returns the string value of the socket option opt for the
// socket associated with fd at the given socket level.
func GetsockoptString(fd, level, opt int) (string, error) {
buf := make([]byte, 256)
vallen := _Socklen(len(buf))
err := getsockopt(fd, level, opt, unsafe.Pointer(&buf[0]), &vallen)
if err != nil {
return "", err
}
return string(buf[:vallen-1]), nil
2017-03-25 20:45:10 +01:00
}
const ImplementsGetwd = true
//sys Getcwd(buf []byte) (n int, err error)
func Getwd() (wd string, err error) {
var buf [PathMax]byte
// Getcwd will return an error if it failed for any reason.
_, err = Getcwd(buf[0:])
if err != nil {
return "", err
}
n := clen(buf[:])
if n < 1 {
return "", EINVAL
}
return string(buf[:n]), nil
}
/*
* Wrapped
*/
//sysnb getgroups(ngid int, gid *_Gid_t) (n int, err error)
//sysnb setgroups(ngid int, gid *_Gid_t) (err error)
func Getgroups() (gids []int, err error) {
n, err := getgroups(0, nil)
2017-12-03 01:24:05 +01:00
// Check for error and sanity check group count. Newer versions of
2017-03-25 20:45:10 +01:00
// Solaris allow up to 1024 (NGROUPS_MAX).
if n < 0 || n > 1024 {
if err != nil {
return nil, err
}
return nil, EINVAL
} else if n == 0 {
return nil, nil
}
a := make([]_Gid_t, n)
n, err = getgroups(n, &a[0])
if n == -1 {
return nil, err
}
gids = make([]int, n)
for i, v := range a[0:n] {
gids[i] = int(v)
}
return
}
func Setgroups(gids []int) (err error) {
if len(gids) == 0 {
return setgroups(0, nil)
}
a := make([]_Gid_t, len(gids))
for i, v := range gids {
a[i] = _Gid_t(v)
}
return setgroups(len(a), &a[0])
}
// ReadDirent reads directory entries from fd and writes them into buf.
2017-03-25 20:45:10 +01:00
func ReadDirent(fd int, buf []byte) (n int, err error) {
// Final argument is (basep *uintptr) and the syscall doesn't take nil.
// TODO(rsc): Can we use a single global basep for all calls?
return Getdents(fd, buf, new(uintptr))
}
// Wait status is 7 bits at bottom, either 0 (exited),
// 0x7F (stopped), or a signal number that caused an exit.
// The 0x80 bit is whether there was a core dump.
// An extra number (exit code, signal causing a stop)
// is in the high bits.
type WaitStatus uint32
const (
mask = 0x7F
core = 0x80
shift = 8
exited = 0
stopped = 0x7F
)
func (w WaitStatus) Exited() bool { return w&mask == exited }
func (w WaitStatus) ExitStatus() int {
if w&mask != exited {
return -1
}
return int(w >> shift)
}
func (w WaitStatus) Signaled() bool { return w&mask != stopped && w&mask != 0 }
func (w WaitStatus) Signal() syscall.Signal {
sig := syscall.Signal(w & mask)
if sig == stopped || sig == 0 {
return -1
}
return sig
}
func (w WaitStatus) CoreDump() bool { return w.Signaled() && w&core != 0 }
func (w WaitStatus) Stopped() bool { return w&mask == stopped && syscall.Signal(w>>shift) != SIGSTOP }
func (w WaitStatus) Continued() bool { return w&mask == stopped && syscall.Signal(w>>shift) == SIGSTOP }
func (w WaitStatus) StopSignal() syscall.Signal {
if !w.Stopped() {
return -1
}
return syscall.Signal(w>>shift) & 0xFF
}
func (w WaitStatus) TrapCause() int { return -1 }
//sys wait4(pid int32, statusp *_C_int, options int, rusage *Rusage) (wpid int32, err error)
func Wait4(pid int, wstatus *WaitStatus, options int, rusage *Rusage) (int, error) {
var status _C_int
rpid, err := wait4(int32(pid), &status, options, rusage)
wpid := int(rpid)
if wpid == -1 {
return wpid, err
}
if wstatus != nil {
*wstatus = WaitStatus(status)
}
return wpid, nil
}
//sys gethostname(buf []byte) (n int, err error)
func Gethostname() (name string, err error) {
var buf [MaxHostNameLen]byte
n, err := gethostname(buf[:])
if n != 0 {
return "", err
}
n = clen(buf[:])
if n < 1 {
return "", EFAULT
}
return string(buf[:n]), nil
}
//sys utimes(path string, times *[2]Timeval) (err error)
func Utimes(path string, tv []Timeval) (err error) {
if tv == nil {
return utimes(path, nil)
}
if len(tv) != 2 {
return EINVAL
}
return utimes(path, (*[2]Timeval)(unsafe.Pointer(&tv[0])))
}
//sys utimensat(fd int, path string, times *[2]Timespec, flag int) (err error)
func UtimesNano(path string, ts []Timespec) error {
if ts == nil {
return utimensat(AT_FDCWD, path, nil, 0)
}
if len(ts) != 2 {
return EINVAL
}
return utimensat(AT_FDCWD, path, (*[2]Timespec)(unsafe.Pointer(&ts[0])), 0)
}
func UtimesNanoAt(dirfd int, path string, ts []Timespec, flags int) error {
if ts == nil {
return utimensat(dirfd, path, nil, flags)
}
if len(ts) != 2 {
return EINVAL
}
return utimensat(dirfd, path, (*[2]Timespec)(unsafe.Pointer(&ts[0])), flags)
}
//sys fcntl(fd int, cmd int, arg int) (val int, err error)
// FcntlInt performs a fcntl syscall on fd with the provided command and argument.
func FcntlInt(fd uintptr, cmd, arg int) (int, error) {
valptr, _, errno := sysvicall6(uintptr(unsafe.Pointer(&procfcntl)), 3, uintptr(fd), uintptr(cmd), uintptr(arg), 0, 0, 0)
var err error
if errno != 0 {
err = errno
}
return int(valptr), err
}
2017-03-25 20:45:10 +01:00
// FcntlFlock performs a fcntl syscall for the F_GETLK, F_SETLK or F_SETLKW command.
func FcntlFlock(fd uintptr, cmd int, lk *Flock_t) error {
_, _, e1 := sysvicall6(uintptr(unsafe.Pointer(&procfcntl)), 3, uintptr(fd), uintptr(cmd), uintptr(unsafe.Pointer(lk)), 0, 0, 0)
if e1 != 0 {
return e1
}
return nil
}
//sys futimesat(fildes int, path *byte, times *[2]Timeval) (err error)
func Futimesat(dirfd int, path string, tv []Timeval) error {
pathp, err := BytePtrFromString(path)
if err != nil {
return err
}
if tv == nil {
return futimesat(dirfd, pathp, nil)
}
if len(tv) != 2 {
return EINVAL
}
return futimesat(dirfd, pathp, (*[2]Timeval)(unsafe.Pointer(&tv[0])))
}
// Solaris doesn't have an futimes function because it allows NULL to be
2017-12-03 01:24:05 +01:00
// specified as the path for futimesat. However, Go doesn't like
2017-03-25 20:45:10 +01:00
// NULL-style string interfaces, so this simple wrapper is provided.
func Futimes(fd int, tv []Timeval) error {
if tv == nil {
return futimesat(fd, nil, nil)
}
if len(tv) != 2 {
return EINVAL
}
return futimesat(fd, nil, (*[2]Timeval)(unsafe.Pointer(&tv[0])))
}
func anyToSockaddr(fd int, rsa *RawSockaddrAny) (Sockaddr, error) {
2017-03-25 20:45:10 +01:00
switch rsa.Addr.Family {
case AF_UNIX:
pp := (*RawSockaddrUnix)(unsafe.Pointer(rsa))
sa := new(SockaddrUnix)
// Assume path ends at NUL.
// This is not technically the Solaris semantics for
// abstract Unix domain sockets -- they are supposed
// to be uninterpreted fixed-size binary blobs -- but
// everyone uses this convention.
n := 0
for n < len(pp.Path) && pp.Path[n] != 0 {
n++
}
2020-01-09 21:02:56 +01:00
bytes := (*[len(pp.Path)]byte)(unsafe.Pointer(&pp.Path[0]))[0:n]
2017-03-25 20:45:10 +01:00
sa.Name = string(bytes)
return sa, nil
case AF_INET:
pp := (*RawSockaddrInet4)(unsafe.Pointer(rsa))
sa := new(SockaddrInet4)
p := (*[2]byte)(unsafe.Pointer(&pp.Port))
sa.Port = int(p[0])<<8 + int(p[1])
sa.Addr = pp.Addr
2017-03-25 20:45:10 +01:00
return sa, nil
case AF_INET6:
pp := (*RawSockaddrInet6)(unsafe.Pointer(rsa))
sa := new(SockaddrInet6)
p := (*[2]byte)(unsafe.Pointer(&pp.Port))
sa.Port = int(p[0])<<8 + int(p[1])
sa.ZoneId = pp.Scope_id
sa.Addr = pp.Addr
2017-03-25 20:45:10 +01:00
return sa, nil
}
return nil, EAFNOSUPPORT
}
//sys accept(s int, rsa *RawSockaddrAny, addrlen *_Socklen) (fd int, err error) = libsocket.accept
func Accept(fd int) (nfd int, sa Sockaddr, err error) {
var rsa RawSockaddrAny
var len _Socklen = SizeofSockaddrAny
nfd, err = accept(fd, &rsa, &len)
if nfd == -1 {
return
}
sa, err = anyToSockaddr(fd, &rsa)
2017-03-25 20:45:10 +01:00
if err != nil {
Close(nfd)
nfd = 0
}
return
}
2017-12-03 01:24:05 +01:00
//sys recvmsg(s int, msg *Msghdr, flags int) (n int, err error) = libsocket.__xnet_recvmsg
2017-03-25 20:45:10 +01:00
func recvmsgRaw(fd int, iov []Iovec, oob []byte, flags int, rsa *RawSockaddrAny) (n, oobn int, recvflags int, err error) {
2017-03-25 20:45:10 +01:00
var msg Msghdr
2022-04-25 23:50:10 +02:00
msg.Name = (*byte)(unsafe.Pointer(rsa))
2017-03-25 20:45:10 +01:00
msg.Namelen = uint32(SizeofSockaddrAny)
var dummy byte
2017-03-25 20:45:10 +01:00
if len(oob) > 0 {
// receive at least one normal byte
if emptyIovecs(iov) {
var iova [1]Iovec
iova[0].Base = &dummy
iova[0].SetLen(1)
iov = iova[:]
2017-03-25 20:45:10 +01:00
}
2017-12-03 01:24:05 +01:00
msg.Accrightslen = int32(len(oob))
2017-03-25 20:45:10 +01:00
}
if len(iov) > 0 {
msg.Iov = &iov[0]
msg.SetIovlen(len(iov))
}
2017-03-25 20:45:10 +01:00
if n, err = recvmsg(fd, &msg, flags); n == -1 {
return
}
oobn = int(msg.Accrightslen)
return
}
2017-12-03 01:24:05 +01:00
//sys sendmsg(s int, msg *Msghdr, flags int) (n int, err error) = libsocket.__xnet_sendmsg
2017-03-25 20:45:10 +01:00
func sendmsgN(fd int, iov []Iovec, oob []byte, ptr unsafe.Pointer, salen _Socklen, flags int) (n int, err error) {
2017-03-25 20:45:10 +01:00
var msg Msghdr
msg.Name = (*byte)(unsafe.Pointer(ptr))
msg.Namelen = uint32(salen)
var dummy byte
var empty bool
2017-03-25 20:45:10 +01:00
if len(oob) > 0 {
// send at least one normal byte
empty = emptyIovecs(iov)
if empty {
var iova [1]Iovec
iova[0].Base = &dummy
iova[0].SetLen(1)
iov = iova[:]
2017-03-25 20:45:10 +01:00
}
2017-12-03 01:24:05 +01:00
msg.Accrightslen = int32(len(oob))
2017-03-25 20:45:10 +01:00
}
if len(iov) > 0 {
msg.Iov = &iov[0]
msg.SetIovlen(len(iov))
}
2017-03-25 20:45:10 +01:00
if n, err = sendmsg(fd, &msg, flags); err != nil {
return 0, err
}
if len(oob) > 0 && empty {
2017-03-25 20:45:10 +01:00
n = 0
}
return n, nil
}
//sys acct(path *byte) (err error)
func Acct(path string) (err error) {
if len(path) == 0 {
// Assume caller wants to disable accounting.
return acct(nil)
}
pathp, err := BytePtrFromString(path)
if err != nil {
return err
}
return acct(pathp)
}
2017-12-03 01:24:05 +01:00
//sys __makedev(version int, major uint, minor uint) (val uint64)
func Mkdev(major, minor uint32) uint64 {
return __makedev(NEWDEV, uint(major), uint(minor))
}
//sys __major(version int, dev uint64) (val uint)
func Major(dev uint64) uint32 {
return uint32(__major(NEWDEV, dev))
}
//sys __minor(version int, dev uint64) (val uint)
func Minor(dev uint64) uint32 {
return uint32(__minor(NEWDEV, dev))
}
2017-03-25 20:45:10 +01:00
/*
* Expose the ioctl function
*/
2021-05-30 00:25:30 +02:00
//sys ioctlRet(fd int, req uint, arg uintptr) (ret int, err error) = libc.ioctl
func ioctl(fd int, req uint, arg uintptr) (err error) {
_, err = ioctlRet(fd, req, arg)
return err
}
2017-03-25 20:45:10 +01:00
2020-11-22 15:55:57 +01:00
func IoctlSetTermio(fd int, req uint, value *Termio) error {
err := ioctl(fd, req, uintptr(unsafe.Pointer(value)))
runtime.KeepAlive(value)
return err
2017-03-25 20:45:10 +01:00
}
2017-12-03 01:24:05 +01:00
func IoctlGetTermio(fd int, req uint) (*Termio, error) {
2017-03-25 20:45:10 +01:00
var value Termio
err := ioctl(fd, req, uintptr(unsafe.Pointer(&value)))
return &value, err
}
2021-03-20 22:40:23 +01:00
//sys poll(fds *PollFd, nfds int, timeout int) (n int, err error)
2017-12-03 01:24:05 +01:00
func Poll(fds []PollFd, timeout int) (n int, err error) {
if len(fds) == 0 {
return poll(nil, 0, timeout)
}
return poll(&fds[0], len(fds), timeout)
}
func Sendfile(outfd int, infd int, offset *int64, count int) (written int, err error) {
if raceenabled {
raceReleaseMerge(unsafe.Pointer(&ioSync))
}
return sendfile(outfd, infd, offset, count)
}
2017-03-25 20:45:10 +01:00
/*
* Exposed directly
*/
//sys Access(path string, mode uint32) (err error)
//sys Adjtime(delta *Timeval, olddelta *Timeval) (err error)
//sys Chdir(path string) (err error)
//sys Chmod(path string, mode uint32) (err error)
//sys Chown(path string, uid int, gid int) (err error)
//sys Chroot(path string) (err error)
2023-01-28 22:57:53 +01:00
//sys ClockGettime(clockid int32, time *Timespec) (err error)
2017-03-25 20:45:10 +01:00
//sys Close(fd int) (err error)
//sys Creat(path string, mode uint32) (fd int, err error)
//sys Dup(fd int) (nfd int, err error)
//sys Dup2(oldfd int, newfd int) (err error)
//sys Exit(code int)
//sys Faccessat(dirfd int, path string, mode uint32, flags int) (err error)
2017-03-25 20:45:10 +01:00
//sys Fchdir(fd int) (err error)
//sys Fchmod(fd int, mode uint32) (err error)
//sys Fchmodat(dirfd int, path string, mode uint32, flags int) (err error)
//sys Fchown(fd int, uid int, gid int) (err error)
//sys Fchownat(dirfd int, path string, uid int, gid int, flags int) (err error)
//sys Fdatasync(fd int) (err error)
//sys Flock(fd int, how int) (err error)
2017-03-25 20:45:10 +01:00
//sys Fpathconf(fd int, name int) (val int, err error)
//sys Fstat(fd int, stat *Stat_t) (err error)
//sys Fstatat(fd int, path string, stat *Stat_t, flags int) (err error)
2017-12-03 01:24:05 +01:00
//sys Fstatvfs(fd int, vfsstat *Statvfs_t) (err error)
2017-03-25 20:45:10 +01:00
//sys Getdents(fd int, buf []byte, basep *uintptr) (n int, err error)
//sysnb Getgid() (gid int)
//sysnb Getpid() (pid int)
//sysnb Getpgid(pid int) (pgid int, err error)
//sysnb Getpgrp() (pgid int, err error)
//sys Geteuid() (euid int)
//sys Getegid() (egid int)
//sys Getppid() (ppid int)
//sys Getpriority(which int, who int) (n int, err error)
//sysnb Getrlimit(which int, lim *Rlimit) (err error)
//sysnb Getrusage(who int, rusage *Rusage) (err error)
//sysnb Getsid(pid int) (sid int, err error)
2017-03-25 20:45:10 +01:00
//sysnb Gettimeofday(tv *Timeval) (err error)
//sysnb Getuid() (uid int)
//sys Kill(pid int, signum syscall.Signal) (err error)
//sys Lchown(path string, uid int, gid int) (err error)
//sys Link(path string, link string) (err error)
2017-12-03 01:24:05 +01:00
//sys Listen(s int, backlog int) (err error) = libsocket.__xnet_llisten
2017-03-25 20:45:10 +01:00
//sys Lstat(path string, stat *Stat_t) (err error)
//sys Madvise(b []byte, advice int) (err error)
//sys Mkdir(path string, mode uint32) (err error)
//sys Mkdirat(dirfd int, path string, mode uint32) (err error)
//sys Mkfifo(path string, mode uint32) (err error)
//sys Mkfifoat(dirfd int, path string, mode uint32) (err error)
//sys Mknod(path string, mode uint32, dev int) (err error)
//sys Mknodat(dirfd int, path string, mode uint32, dev int) (err error)
//sys Mlock(b []byte) (err error)
//sys Mlockall(flags int) (err error)
//sys Mprotect(b []byte, prot int) (err error)
2017-12-03 01:24:05 +01:00
//sys Msync(b []byte, flags int) (err error)
2017-03-25 20:45:10 +01:00
//sys Munlock(b []byte) (err error)
//sys Munlockall() (err error)
//sys Nanosleep(time *Timespec, leftover *Timespec) (err error)
//sys Open(path string, mode int, perm uint32) (fd int, err error)
//sys Openat(dirfd int, path string, flags int, mode uint32) (fd int, err error)
//sys Pathconf(path string, name int) (val int, err error)
//sys Pause() (err error)
2022-04-25 23:50:10 +02:00
//sys pread(fd int, p []byte, offset int64) (n int, err error)
//sys pwrite(fd int, p []byte, offset int64) (n int, err error)
2017-03-25 20:45:10 +01:00
//sys read(fd int, p []byte) (n int, err error)
//sys Readlink(path string, buf []byte) (n int, err error)
//sys Rename(from string, to string) (err error)
//sys Renameat(olddirfd int, oldpath string, newdirfd int, newpath string) (err error)
//sys Rmdir(path string) (err error)
//sys Seek(fd int, offset int64, whence int) (newoffset int64, err error) = lseek
2020-01-09 21:02:56 +01:00
//sys Select(nfd int, r *FdSet, w *FdSet, e *FdSet, timeout *Timeval) (n int, err error)
2017-03-25 20:45:10 +01:00
//sysnb Setegid(egid int) (err error)
//sysnb Seteuid(euid int) (err error)
//sysnb Setgid(gid int) (err error)
//sys Sethostname(p []byte) (err error)
//sysnb Setpgid(pid int, pgid int) (err error)
//sys Setpriority(which int, who int, prio int) (err error)
//sysnb Setregid(rgid int, egid int) (err error)
//sysnb Setreuid(ruid int, euid int) (err error)
//sysnb Setrlimit(which int, lim *Rlimit) (err error)
//sysnb Setsid() (pid int, err error)
//sysnb Setuid(uid int) (err error)
//sys Shutdown(s int, how int) (err error) = libsocket.shutdown
//sys Stat(path string, stat *Stat_t) (err error)
2017-12-03 01:24:05 +01:00
//sys Statvfs(path string, vfsstat *Statvfs_t) (err error)
2017-03-25 20:45:10 +01:00
//sys Symlink(path string, link string) (err error)
//sys Sync() (err error)
2021-03-20 22:40:23 +01:00
//sys Sysconf(which int) (n int64, err error)
2017-03-25 20:45:10 +01:00
//sysnb Times(tms *Tms) (ticks uintptr, err error)
//sys Truncate(path string, length int64) (err error)
//sys Fsync(fd int) (err error)
//sys Ftruncate(fd int, length int64) (err error)
//sys Umask(mask int) (oldmask int)
//sysnb Uname(buf *Utsname) (err error)
//sys Unmount(target string, flags int) (err error) = libc.umount
//sys Unlink(path string) (err error)
//sys Unlinkat(dirfd int, path string, flags int) (err error)
//sys Ustat(dev int, ubuf *Ustat_t) (err error)
//sys Utime(path string, buf *Utimbuf) (err error)
2017-12-03 01:24:05 +01:00
//sys bind(s int, addr unsafe.Pointer, addrlen _Socklen) (err error) = libsocket.__xnet_bind
//sys connect(s int, addr unsafe.Pointer, addrlen _Socklen) (err error) = libsocket.__xnet_connect
2017-03-25 20:45:10 +01:00
//sys mmap(addr uintptr, length uintptr, prot int, flag int, fd int, pos int64) (ret uintptr, err error)
//sys munmap(addr uintptr, length uintptr) (err error)
//sys sendfile(outfd int, infd int, offset *int64, count int) (written int, err error) = libsendfile.sendfile
2017-12-03 01:24:05 +01:00
//sys sendto(s int, buf []byte, flags int, to unsafe.Pointer, addrlen _Socklen) (err error) = libsocket.__xnet_sendto
//sys socket(domain int, typ int, proto int) (fd int, err error) = libsocket.__xnet_socket
//sysnb socketpair(domain int, typ int, proto int, fd *[2]int32) (err error) = libsocket.__xnet_socketpair
2017-03-25 20:45:10 +01:00
//sys write(fd int, p []byte) (n int, err error)
2017-12-03 01:24:05 +01:00
//sys getsockopt(s int, level int, name int, val unsafe.Pointer, vallen *_Socklen) (err error) = libsocket.__xnet_getsockopt
2017-03-25 20:45:10 +01:00
//sysnb getpeername(fd int, rsa *RawSockaddrAny, addrlen *_Socklen) (err error) = libsocket.getpeername
//sys setsockopt(s int, level int, name int, val unsafe.Pointer, vallen uintptr) (err error) = libsocket.setsockopt
//sys recvfrom(fd int, p []byte, flags int, from *RawSockaddrAny, fromlen *_Socklen) (n int, err error) = libsocket.recvfrom
func readlen(fd int, buf *byte, nbuf int) (n int, err error) {
r0, _, e1 := sysvicall6(uintptr(unsafe.Pointer(&procread)), 3, uintptr(fd), uintptr(unsafe.Pointer(buf)), uintptr(nbuf), 0, 0, 0)
n = int(r0)
if e1 != 0 {
err = e1
}
return
}
func writelen(fd int, buf *byte, nbuf int) (n int, err error) {
r0, _, e1 := sysvicall6(uintptr(unsafe.Pointer(&procwrite)), 3, uintptr(fd), uintptr(unsafe.Pointer(buf)), uintptr(nbuf), 0, 0, 0)
n = int(r0)
if e1 != 0 {
err = e1
}
return
}
var mapper = &mmapper{
active: make(map[*byte][]byte),
mmap: mmap,
munmap: munmap,
}
func Mmap(fd int, offset int64, length int, prot int, flags int) (data []byte, err error) {
return mapper.Mmap(fd, offset, length, prot, flags)
}
func Munmap(b []byte) (err error) {
return mapper.Munmap(b)
}
// Event Ports
type fileObjCookie struct {
fobj *fileObj
cookie interface{}
}
// EventPort provides a safe abstraction on top of Solaris/illumos Event Ports.
type EventPort struct {
port int
mu sync.Mutex
2022-04-25 23:50:10 +02:00
fds map[uintptr]*fileObjCookie
paths map[string]*fileObjCookie
2022-04-25 23:50:10 +02:00
// The user cookie presents an interesting challenge from a memory management perspective.
// There are two paths by which we can discover that it is no longer in use:
// 1. The user calls port_dissociate before any events fire
// 2. An event fires and we return it to the user
// The tricky situation is if the event has fired in the kernel but
// the user hasn't requested/received it yet.
// If the user wants to port_dissociate before the event has been processed,
// we should handle things gracefully. To do so, we need to keep an extra
// reference to the cookie around until the event is processed
// thus the otherwise seemingly extraneous "cookies" map
2022-11-27 00:42:16 +01:00
// The key of this map is a pointer to the corresponding fCookie
cookies map[*fileObjCookie]struct{}
}
// PortEvent is an abstraction of the port_event C struct.
// Compare Source against PORT_SOURCE_FILE or PORT_SOURCE_FD
// to see if Path or Fd was the event source. The other will be
// uninitialized.
type PortEvent struct {
Cookie interface{}
Events int32
Fd uintptr
Path string
Source uint16
fobj *fileObj
}
// NewEventPort creates a new EventPort including the
// underlying call to port_create(3c).
func NewEventPort() (*EventPort, error) {
port, err := port_create()
if err != nil {
return nil, err
}
e := &EventPort{
2022-04-25 23:50:10 +02:00
port: port,
fds: make(map[uintptr]*fileObjCookie),
paths: make(map[string]*fileObjCookie),
2022-11-27 00:42:16 +01:00
cookies: make(map[*fileObjCookie]struct{}),
}
return e, nil
}
//sys port_create() (n int, err error)
//sys port_associate(port int, source int, object uintptr, events int, user *byte) (n int, err error)
//sys port_dissociate(port int, source int, object uintptr) (n int, err error)
//sys port_get(port int, pe *portEvent, timeout *Timespec) (n int, err error)
//sys port_getn(port int, pe *portEvent, max uint32, nget *uint32, timeout *Timespec) (n int, err error)
// Close closes the event port.
func (e *EventPort) Close() error {
e.mu.Lock()
defer e.mu.Unlock()
2022-04-25 23:50:10 +02:00
err := Close(e.port)
if err != nil {
return err
}
e.fds = nil
e.paths = nil
2022-11-27 00:42:16 +01:00
e.cookies = nil
2022-04-25 23:50:10 +02:00
return nil
}
// PathIsWatched checks to see if path is associated with this EventPort.
func (e *EventPort) PathIsWatched(path string) bool {
e.mu.Lock()
defer e.mu.Unlock()
_, found := e.paths[path]
return found
}
// FdIsWatched checks to see if fd is associated with this EventPort.
func (e *EventPort) FdIsWatched(fd uintptr) bool {
e.mu.Lock()
defer e.mu.Unlock()
_, found := e.fds[fd]
return found
}
// AssociatePath wraps port_associate(3c) for a filesystem path including
// creating the necessary file_obj from the provided stat information.
func (e *EventPort) AssociatePath(path string, stat os.FileInfo, events int, cookie interface{}) error {
e.mu.Lock()
defer e.mu.Unlock()
if _, found := e.paths[path]; found {
return fmt.Errorf("%v is already associated with this Event Port", path)
}
2022-11-27 00:42:16 +01:00
fCookie, err := createFileObjCookie(path, stat, cookie)
if err != nil {
return err
}
2022-11-27 00:42:16 +01:00
_, err = port_associate(e.port, PORT_SOURCE_FILE, uintptr(unsafe.Pointer(fCookie.fobj)), events, (*byte)(unsafe.Pointer(fCookie)))
if err != nil {
return err
}
e.paths[path] = fCookie
2022-11-27 00:42:16 +01:00
e.cookies[fCookie] = struct{}{}
return nil
}
// DissociatePath wraps port_dissociate(3c) for a filesystem path.
func (e *EventPort) DissociatePath(path string) error {
e.mu.Lock()
defer e.mu.Unlock()
f, ok := e.paths[path]
if !ok {
return fmt.Errorf("%v is not associated with this Event Port", path)
}
_, err := port_dissociate(e.port, PORT_SOURCE_FILE, uintptr(unsafe.Pointer(f.fobj)))
2022-04-25 23:50:10 +02:00
// If the path is no longer associated with this event port (ENOENT)
// we should delete it from our map. We can still return ENOENT to the caller.
// But we need to save the cookie
if err != nil && err != ENOENT {
return err
}
2022-04-25 23:50:10 +02:00
if err == nil {
// dissociate was successful, safe to delete the cookie
fCookie := e.paths[path]
2022-11-27 00:42:16 +01:00
delete(e.cookies, fCookie)
2022-04-25 23:50:10 +02:00
}
delete(e.paths, path)
2022-04-25 23:50:10 +02:00
return err
}
// AssociateFd wraps calls to port_associate(3c) on file descriptors.
func (e *EventPort) AssociateFd(fd uintptr, events int, cookie interface{}) error {
e.mu.Lock()
defer e.mu.Unlock()
if _, found := e.fds[fd]; found {
return fmt.Errorf("%v is already associated with this Event Port", fd)
}
2022-11-27 00:42:16 +01:00
fCookie, err := createFileObjCookie("", nil, cookie)
if err != nil {
return err
}
_, err = port_associate(e.port, PORT_SOURCE_FD, fd, events, (*byte)(unsafe.Pointer(fCookie)))
if err != nil {
return err
}
2022-04-25 23:50:10 +02:00
e.fds[fd] = fCookie
2022-11-27 00:42:16 +01:00
e.cookies[fCookie] = struct{}{}
return nil
}
// DissociateFd wraps calls to port_dissociate(3c) on file descriptors.
func (e *EventPort) DissociateFd(fd uintptr) error {
e.mu.Lock()
defer e.mu.Unlock()
_, ok := e.fds[fd]
if !ok {
return fmt.Errorf("%v is not associated with this Event Port", fd)
}
_, err := port_dissociate(e.port, PORT_SOURCE_FD, fd)
2022-04-25 23:50:10 +02:00
if err != nil && err != ENOENT {
return err
}
2022-04-25 23:50:10 +02:00
if err == nil {
// dissociate was successful, safe to delete the cookie
fCookie := e.fds[fd]
2022-11-27 00:42:16 +01:00
delete(e.cookies, fCookie)
2022-04-25 23:50:10 +02:00
}
delete(e.fds, fd)
2022-04-25 23:50:10 +02:00
return err
}
2022-11-27 00:42:16 +01:00
func createFileObjCookie(name string, stat os.FileInfo, cookie interface{}) (*fileObjCookie, error) {
fCookie := new(fileObjCookie)
fCookie.cookie = cookie
if name != "" && stat != nil {
fCookie.fobj = new(fileObj)
bs, err := ByteSliceFromString(name)
if err != nil {
return nil, err
}
fCookie.fobj.Name = (*int8)(unsafe.Pointer(&bs[0]))
s := stat.Sys().(*syscall.Stat_t)
fCookie.fobj.Atim.Sec = s.Atim.Sec
fCookie.fobj.Atim.Nsec = s.Atim.Nsec
fCookie.fobj.Mtim.Sec = s.Mtim.Sec
fCookie.fobj.Mtim.Nsec = s.Mtim.Nsec
fCookie.fobj.Ctim.Sec = s.Ctim.Sec
fCookie.fobj.Ctim.Nsec = s.Ctim.Nsec
}
return fCookie, nil
}
// GetOne wraps port_get(3c) and returns a single PortEvent.
func (e *EventPort) GetOne(t *Timespec) (*PortEvent, error) {
pe := new(portEvent)
_, err := port_get(e.port, pe, t)
if err != nil {
return nil, err
}
p := new(PortEvent)
e.mu.Lock()
defer e.mu.Unlock()
2022-11-27 00:42:16 +01:00
err = e.peIntToExt(pe, p)
if err != nil {
return nil, err
}
2022-04-25 23:50:10 +02:00
return p, nil
}
// peIntToExt converts a cgo portEvent struct into the friendlier PortEvent
// NOTE: Always call this function while holding the e.mu mutex
2022-11-27 00:42:16 +01:00
func (e *EventPort) peIntToExt(peInt *portEvent, peExt *PortEvent) error {
if e.cookies == nil {
return fmt.Errorf("this EventPort is already closed")
}
2022-04-25 23:50:10 +02:00
peExt.Events = peInt.Events
peExt.Source = peInt.Source
2022-11-27 00:42:16 +01:00
fCookie := (*fileObjCookie)(unsafe.Pointer(peInt.User))
_, found := e.cookies[fCookie]
if !found {
panic("unexpected event port address; may be due to kernel bug; see https://go.dev/issue/54254")
}
peExt.Cookie = fCookie.cookie
delete(e.cookies, fCookie)
2022-04-25 23:50:10 +02:00
switch peInt.Source {
case PORT_SOURCE_FD:
2022-04-25 23:50:10 +02:00
peExt.Fd = uintptr(peInt.Object)
// Only remove the fds entry if it exists and this cookie matches
if fobj, ok := e.fds[peExt.Fd]; ok {
2022-11-27 00:42:16 +01:00
if fobj == fCookie {
2022-04-25 23:50:10 +02:00
delete(e.fds, peExt.Fd)
}
}
case PORT_SOURCE_FILE:
2022-11-27 00:42:16 +01:00
peExt.fobj = fCookie.fobj
2022-04-25 23:50:10 +02:00
peExt.Path = BytePtrToString((*byte)(unsafe.Pointer(peExt.fobj.Name)))
// Only remove the paths entry if it exists and this cookie matches
if fobj, ok := e.paths[peExt.Path]; ok {
2022-11-27 00:42:16 +01:00
if fobj == fCookie {
2022-04-25 23:50:10 +02:00
delete(e.paths, peExt.Path)
}
}
}
2022-11-27 00:42:16 +01:00
return nil
}
// Pending wraps port_getn(3c) and returns how many events are pending.
func (e *EventPort) Pending() (int, error) {
var n uint32 = 0
_, err := port_getn(e.port, nil, 0, &n, nil)
return int(n), err
}
// Get wraps port_getn(3c) and fills a slice of PortEvent.
// It will block until either min events have been received
// or the timeout has been exceeded. It will return how many
// events were actually received along with any error information.
func (e *EventPort) Get(s []PortEvent, min int, timeout *Timespec) (int, error) {
if min == 0 {
return 0, fmt.Errorf("need to request at least one event or use Pending() instead")
}
if len(s) < min {
return 0, fmt.Errorf("len(s) (%d) is less than min events requested (%d)", len(s), min)
}
got := uint32(min)
max := uint32(len(s))
var err error
2022-11-27 00:42:16 +01:00
ps := make([]portEvent, max)
_, err = port_getn(e.port, &ps[0], max, &got, timeout)
// got will be trustworthy with ETIME, but not any other error.
if err != nil && err != ETIME {
return 0, err
}
e.mu.Lock()
defer e.mu.Unlock()
2022-11-27 00:42:16 +01:00
valid := 0
for i := 0; i < int(got); i++ {
2022-11-27 00:42:16 +01:00
err2 := e.peIntToExt(&ps[i], &s[i])
if err2 != nil {
if valid == 0 && err == nil {
// If err2 is the only error and there are no valid events
// to return, return it to the caller.
err = err2
}
break
}
valid = i + 1
}
return valid, err
}
//sys putmsg(fd int, clptr *strbuf, dataptr *strbuf, flags int) (err error)
func Putmsg(fd int, cl []byte, data []byte, flags int) (err error) {
var clp, datap *strbuf
if len(cl) > 0 {
clp = &strbuf{
Len: int32(len(cl)),
Buf: (*int8)(unsafe.Pointer(&cl[0])),
}
}
2022-11-27 00:42:16 +01:00
if len(data) > 0 {
datap = &strbuf{
Len: int32(len(data)),
Buf: (*int8)(unsafe.Pointer(&data[0])),
}
}
return putmsg(fd, clp, datap, flags)
}
//sys getmsg(fd int, clptr *strbuf, dataptr *strbuf, flags *int) (err error)
func Getmsg(fd int, cl []byte, data []byte) (retCl []byte, retData []byte, flags int, err error) {
var clp, datap *strbuf
if len(cl) > 0 {
clp = &strbuf{
Maxlen: int32(len(cl)),
Buf: (*int8)(unsafe.Pointer(&cl[0])),
}
}
if len(data) > 0 {
datap = &strbuf{
Maxlen: int32(len(data)),
Buf: (*int8)(unsafe.Pointer(&data[0])),
}
}
if err = getmsg(fd, clp, datap, &flags); err != nil {
return nil, nil, 0, err
}
if len(cl) > 0 {
retCl = cl[:clp.Len]
}
if len(data) > 0 {
retData = data[:datap.Len]
}
return retCl, retData, flags, nil
}
func IoctlSetIntRetInt(fd int, req uint, arg int) (int, error) {
return ioctlRet(fd, req, uintptr(arg))
}
func IoctlSetString(fd int, req uint, val string) error {
bs := make([]byte, len(val)+1)
copy(bs[:len(bs)-1], val)
err := ioctl(fd, req, uintptr(unsafe.Pointer(&bs[0])))
runtime.KeepAlive(&bs[0])
return err
}
// Lifreq Helpers
func (l *Lifreq) SetName(name string) error {
if len(name) >= len(l.Name) {
return fmt.Errorf("name cannot be more than %d characters", len(l.Name)-1)
}
for i := range name {
l.Name[i] = int8(name[i])
}
return nil
}
func (l *Lifreq) SetLifruInt(d int) {
*(*int)(unsafe.Pointer(&l.Lifru[0])) = d
}
func (l *Lifreq) GetLifruInt() int {
return *(*int)(unsafe.Pointer(&l.Lifru[0]))
}
func (l *Lifreq) SetLifruUint(d uint) {
*(*uint)(unsafe.Pointer(&l.Lifru[0])) = d
}
func (l *Lifreq) GetLifruUint() uint {
return *(*uint)(unsafe.Pointer(&l.Lifru[0]))
}
func IoctlLifreq(fd int, req uint, l *Lifreq) error {
return ioctl(fd, req, uintptr(unsafe.Pointer(l)))
}
// Strioctl Helpers
func (s *Strioctl) SetInt(i int) {
s.Len = int32(unsafe.Sizeof(i))
s.Dp = (*int8)(unsafe.Pointer(&i))
}
func IoctlSetStrioctlRetInt(fd int, req uint, s *Strioctl) (int, error) {
return ioctlRet(fd, req, uintptr(unsafe.Pointer(s)))
}