mirror of
https://github.com/superseriousbusiness/gotosocial.git
synced 2024-12-29 10:36:31 +00:00
49009fbd8f
This fixes some linkname shenanigans previous versions of the library were using. It's now safe to upgrade to Go 1.23 and beyond once they become available.
616 lines
17 KiB
Go
616 lines
17 KiB
Go
// 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.
|
|
|
|
//go:build aix || darwin || dragonfly || freebsd || linux || netbsd || openbsd || solaris
|
|
|
|
package unix
|
|
|
|
import (
|
|
"bytes"
|
|
"sort"
|
|
"sync"
|
|
"syscall"
|
|
"unsafe"
|
|
)
|
|
|
|
var (
|
|
Stdin = 0
|
|
Stdout = 1
|
|
Stderr = 2
|
|
)
|
|
|
|
// Do the interface allocations only once for common
|
|
// Errno values.
|
|
var (
|
|
errEAGAIN error = syscall.EAGAIN
|
|
errEINVAL error = syscall.EINVAL
|
|
errENOENT error = syscall.ENOENT
|
|
)
|
|
|
|
var (
|
|
signalNameMapOnce sync.Once
|
|
signalNameMap map[string]syscall.Signal
|
|
)
|
|
|
|
// errnoErr returns common boxed Errno values, to prevent
|
|
// allocations at runtime.
|
|
func errnoErr(e syscall.Errno) error {
|
|
switch e {
|
|
case 0:
|
|
return nil
|
|
case EAGAIN:
|
|
return errEAGAIN
|
|
case EINVAL:
|
|
return errEINVAL
|
|
case ENOENT:
|
|
return errENOENT
|
|
}
|
|
return e
|
|
}
|
|
|
|
// ErrnoName returns the error name for error number e.
|
|
func ErrnoName(e syscall.Errno) string {
|
|
i := sort.Search(len(errorList), func(i int) bool {
|
|
return errorList[i].num >= e
|
|
})
|
|
if i < len(errorList) && errorList[i].num == e {
|
|
return errorList[i].name
|
|
}
|
|
return ""
|
|
}
|
|
|
|
// SignalName returns the signal name for signal number s.
|
|
func SignalName(s syscall.Signal) string {
|
|
i := sort.Search(len(signalList), func(i int) bool {
|
|
return signalList[i].num >= s
|
|
})
|
|
if i < len(signalList) && signalList[i].num == s {
|
|
return signalList[i].name
|
|
}
|
|
return ""
|
|
}
|
|
|
|
// SignalNum returns the syscall.Signal for signal named s,
|
|
// or 0 if a signal with such name is not found.
|
|
// The signal name should start with "SIG".
|
|
func SignalNum(s string) syscall.Signal {
|
|
signalNameMapOnce.Do(func() {
|
|
signalNameMap = make(map[string]syscall.Signal, len(signalList))
|
|
for _, signal := range signalList {
|
|
signalNameMap[signal.name] = signal.num
|
|
}
|
|
})
|
|
return signalNameMap[s]
|
|
}
|
|
|
|
// clen returns the index of the first NULL byte in n or len(n) if n contains no NULL byte.
|
|
func clen(n []byte) int {
|
|
i := bytes.IndexByte(n, 0)
|
|
if i == -1 {
|
|
i = len(n)
|
|
}
|
|
return i
|
|
}
|
|
|
|
// Mmap manager, for use by operating system-specific implementations.
|
|
|
|
type mmapper struct {
|
|
sync.Mutex
|
|
active map[*byte][]byte // active mappings; key is last byte in mapping
|
|
mmap func(addr, length uintptr, prot, flags, fd int, offset int64) (uintptr, error)
|
|
munmap func(addr uintptr, length uintptr) error
|
|
}
|
|
|
|
func (m *mmapper) Mmap(fd int, offset int64, length int, prot int, flags int) (data []byte, err error) {
|
|
if length <= 0 {
|
|
return nil, EINVAL
|
|
}
|
|
|
|
// Map the requested memory.
|
|
addr, errno := m.mmap(0, uintptr(length), prot, flags, fd, offset)
|
|
if errno != nil {
|
|
return nil, errno
|
|
}
|
|
|
|
// Use unsafe to convert addr into a []byte.
|
|
b := unsafe.Slice((*byte)(unsafe.Pointer(addr)), length)
|
|
|
|
// Register mapping in m and return it.
|
|
p := &b[cap(b)-1]
|
|
m.Lock()
|
|
defer m.Unlock()
|
|
m.active[p] = b
|
|
return b, nil
|
|
}
|
|
|
|
func (m *mmapper) Munmap(data []byte) (err error) {
|
|
if len(data) == 0 || len(data) != cap(data) {
|
|
return EINVAL
|
|
}
|
|
|
|
// Find the base of the mapping.
|
|
p := &data[cap(data)-1]
|
|
m.Lock()
|
|
defer m.Unlock()
|
|
b := m.active[p]
|
|
if b == nil || &b[0] != &data[0] {
|
|
return EINVAL
|
|
}
|
|
|
|
// Unmap the memory and update m.
|
|
if errno := m.munmap(uintptr(unsafe.Pointer(&b[0])), uintptr(len(b))); errno != nil {
|
|
return errno
|
|
}
|
|
delete(m.active, p)
|
|
return nil
|
|
}
|
|
|
|
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)
|
|
}
|
|
|
|
func MmapPtr(fd int, offset int64, addr unsafe.Pointer, length uintptr, prot int, flags int) (ret unsafe.Pointer, err error) {
|
|
xaddr, err := mapper.mmap(uintptr(addr), length, prot, flags, fd, offset)
|
|
return unsafe.Pointer(xaddr), err
|
|
}
|
|
|
|
func MunmapPtr(addr unsafe.Pointer, length uintptr) (err error) {
|
|
return mapper.munmap(uintptr(addr), length)
|
|
}
|
|
|
|
func Read(fd int, p []byte) (n int, err error) {
|
|
n, err = read(fd, p)
|
|
if raceenabled {
|
|
if n > 0 {
|
|
raceWriteRange(unsafe.Pointer(&p[0]), n)
|
|
}
|
|
if err == nil {
|
|
raceAcquire(unsafe.Pointer(&ioSync))
|
|
}
|
|
}
|
|
return
|
|
}
|
|
|
|
func Write(fd int, p []byte) (n int, err error) {
|
|
if raceenabled {
|
|
raceReleaseMerge(unsafe.Pointer(&ioSync))
|
|
}
|
|
n, err = write(fd, p)
|
|
if raceenabled && n > 0 {
|
|
raceReadRange(unsafe.Pointer(&p[0]), n)
|
|
}
|
|
return
|
|
}
|
|
|
|
func Pread(fd int, p []byte, offset int64) (n int, err error) {
|
|
n, err = pread(fd, p, offset)
|
|
if raceenabled {
|
|
if n > 0 {
|
|
raceWriteRange(unsafe.Pointer(&p[0]), n)
|
|
}
|
|
if err == nil {
|
|
raceAcquire(unsafe.Pointer(&ioSync))
|
|
}
|
|
}
|
|
return
|
|
}
|
|
|
|
func Pwrite(fd int, p []byte, offset int64) (n int, err error) {
|
|
if raceenabled {
|
|
raceReleaseMerge(unsafe.Pointer(&ioSync))
|
|
}
|
|
n, err = pwrite(fd, p, offset)
|
|
if raceenabled && n > 0 {
|
|
raceReadRange(unsafe.Pointer(&p[0]), n)
|
|
}
|
|
return
|
|
}
|
|
|
|
// For testing: clients can set this flag to force
|
|
// creation of IPv6 sockets to return EAFNOSUPPORT.
|
|
var SocketDisableIPv6 bool
|
|
|
|
// Sockaddr represents a socket address.
|
|
type Sockaddr interface {
|
|
sockaddr() (ptr unsafe.Pointer, len _Socklen, err error) // lowercase; only we can define Sockaddrs
|
|
}
|
|
|
|
// SockaddrInet4 implements the Sockaddr interface for AF_INET type sockets.
|
|
type SockaddrInet4 struct {
|
|
Port int
|
|
Addr [4]byte
|
|
raw RawSockaddrInet4
|
|
}
|
|
|
|
// SockaddrInet6 implements the Sockaddr interface for AF_INET6 type sockets.
|
|
type SockaddrInet6 struct {
|
|
Port int
|
|
ZoneId uint32
|
|
Addr [16]byte
|
|
raw RawSockaddrInet6
|
|
}
|
|
|
|
// SockaddrUnix implements the Sockaddr interface for AF_UNIX type sockets.
|
|
type SockaddrUnix struct {
|
|
Name string
|
|
raw RawSockaddrUnix
|
|
}
|
|
|
|
func Bind(fd int, sa Sockaddr) (err error) {
|
|
ptr, n, err := sa.sockaddr()
|
|
if err != nil {
|
|
return err
|
|
}
|
|
return bind(fd, ptr, n)
|
|
}
|
|
|
|
func Connect(fd int, sa Sockaddr) (err error) {
|
|
ptr, n, err := sa.sockaddr()
|
|
if err != nil {
|
|
return err
|
|
}
|
|
return connect(fd, ptr, n)
|
|
}
|
|
|
|
func Getpeername(fd int) (sa Sockaddr, err error) {
|
|
var rsa RawSockaddrAny
|
|
var len _Socklen = SizeofSockaddrAny
|
|
if err = getpeername(fd, &rsa, &len); err != nil {
|
|
return
|
|
}
|
|
return anyToSockaddr(fd, &rsa)
|
|
}
|
|
|
|
func GetsockoptByte(fd, level, opt int) (value byte, err error) {
|
|
var n byte
|
|
vallen := _Socklen(1)
|
|
err = getsockopt(fd, level, opt, unsafe.Pointer(&n), &vallen)
|
|
return n, err
|
|
}
|
|
|
|
func GetsockoptInt(fd, level, opt int) (value int, err error) {
|
|
var n int32
|
|
vallen := _Socklen(4)
|
|
err = getsockopt(fd, level, opt, unsafe.Pointer(&n), &vallen)
|
|
return int(n), err
|
|
}
|
|
|
|
func GetsockoptInet4Addr(fd, level, opt int) (value [4]byte, err error) {
|
|
vallen := _Socklen(4)
|
|
err = getsockopt(fd, level, opt, unsafe.Pointer(&value[0]), &vallen)
|
|
return value, err
|
|
}
|
|
|
|
func GetsockoptIPMreq(fd, level, opt int) (*IPMreq, error) {
|
|
var value IPMreq
|
|
vallen := _Socklen(SizeofIPMreq)
|
|
err := getsockopt(fd, level, opt, unsafe.Pointer(&value), &vallen)
|
|
return &value, err
|
|
}
|
|
|
|
func GetsockoptIPv6Mreq(fd, level, opt int) (*IPv6Mreq, error) {
|
|
var value IPv6Mreq
|
|
vallen := _Socklen(SizeofIPv6Mreq)
|
|
err := getsockopt(fd, level, opt, unsafe.Pointer(&value), &vallen)
|
|
return &value, err
|
|
}
|
|
|
|
func GetsockoptIPv6MTUInfo(fd, level, opt int) (*IPv6MTUInfo, error) {
|
|
var value IPv6MTUInfo
|
|
vallen := _Socklen(SizeofIPv6MTUInfo)
|
|
err := getsockopt(fd, level, opt, unsafe.Pointer(&value), &vallen)
|
|
return &value, err
|
|
}
|
|
|
|
func GetsockoptICMPv6Filter(fd, level, opt int) (*ICMPv6Filter, error) {
|
|
var value ICMPv6Filter
|
|
vallen := _Socklen(SizeofICMPv6Filter)
|
|
err := getsockopt(fd, level, opt, unsafe.Pointer(&value), &vallen)
|
|
return &value, err
|
|
}
|
|
|
|
func GetsockoptLinger(fd, level, opt int) (*Linger, error) {
|
|
var linger Linger
|
|
vallen := _Socklen(SizeofLinger)
|
|
err := getsockopt(fd, level, opt, unsafe.Pointer(&linger), &vallen)
|
|
return &linger, err
|
|
}
|
|
|
|
func GetsockoptTimeval(fd, level, opt int) (*Timeval, error) {
|
|
var tv Timeval
|
|
vallen := _Socklen(unsafe.Sizeof(tv))
|
|
err := getsockopt(fd, level, opt, unsafe.Pointer(&tv), &vallen)
|
|
return &tv, err
|
|
}
|
|
|
|
func GetsockoptUint64(fd, level, opt int) (value uint64, err error) {
|
|
var n uint64
|
|
vallen := _Socklen(8)
|
|
err = getsockopt(fd, level, opt, unsafe.Pointer(&n), &vallen)
|
|
return n, err
|
|
}
|
|
|
|
func Recvfrom(fd int, p []byte, flags int) (n int, from Sockaddr, err error) {
|
|
var rsa RawSockaddrAny
|
|
var len _Socklen = SizeofSockaddrAny
|
|
if n, err = recvfrom(fd, p, flags, &rsa, &len); err != nil {
|
|
return
|
|
}
|
|
if rsa.Addr.Family != AF_UNSPEC {
|
|
from, err = anyToSockaddr(fd, &rsa)
|
|
}
|
|
return
|
|
}
|
|
|
|
// Recvmsg receives a message from a socket using the recvmsg system call. The
|
|
// received non-control data will be written to p, and any "out of band"
|
|
// control data will be written to oob. The flags are passed to recvmsg.
|
|
//
|
|
// The results are:
|
|
// - n is the number of non-control data bytes read into p
|
|
// - oobn is the number of control data bytes read into oob; this may be interpreted using [ParseSocketControlMessage]
|
|
// - recvflags is flags returned by recvmsg
|
|
// - from is the address of the sender
|
|
//
|
|
// If the underlying socket type is not SOCK_DGRAM, a received message
|
|
// containing oob data and a single '\0' of non-control data is treated as if
|
|
// the message contained only control data, i.e. n will be zero on return.
|
|
func Recvmsg(fd int, p, oob []byte, flags int) (n, oobn int, recvflags int, from Sockaddr, err error) {
|
|
var iov [1]Iovec
|
|
if len(p) > 0 {
|
|
iov[0].Base = &p[0]
|
|
iov[0].SetLen(len(p))
|
|
}
|
|
var rsa RawSockaddrAny
|
|
n, oobn, recvflags, err = recvmsgRaw(fd, iov[:], oob, flags, &rsa)
|
|
// source address is only specified if the socket is unconnected
|
|
if rsa.Addr.Family != AF_UNSPEC {
|
|
from, err = anyToSockaddr(fd, &rsa)
|
|
}
|
|
return
|
|
}
|
|
|
|
// RecvmsgBuffers receives a message from a socket using the recvmsg system
|
|
// call. This function is equivalent to Recvmsg, but non-control data read is
|
|
// scattered into the buffers slices.
|
|
func RecvmsgBuffers(fd int, buffers [][]byte, oob []byte, flags int) (n, oobn int, recvflags int, from Sockaddr, err error) {
|
|
iov := make([]Iovec, len(buffers))
|
|
for i := range buffers {
|
|
if len(buffers[i]) > 0 {
|
|
iov[i].Base = &buffers[i][0]
|
|
iov[i].SetLen(len(buffers[i]))
|
|
} else {
|
|
iov[i].Base = (*byte)(unsafe.Pointer(&_zero))
|
|
}
|
|
}
|
|
var rsa RawSockaddrAny
|
|
n, oobn, recvflags, err = recvmsgRaw(fd, iov, oob, flags, &rsa)
|
|
if err == nil && rsa.Addr.Family != AF_UNSPEC {
|
|
from, err = anyToSockaddr(fd, &rsa)
|
|
}
|
|
return
|
|
}
|
|
|
|
// Sendmsg sends a message on a socket to an address using the sendmsg system
|
|
// call. This function is equivalent to SendmsgN, but does not return the
|
|
// number of bytes actually sent.
|
|
func Sendmsg(fd int, p, oob []byte, to Sockaddr, flags int) (err error) {
|
|
_, err = SendmsgN(fd, p, oob, to, flags)
|
|
return
|
|
}
|
|
|
|
// SendmsgN sends a message on a socket to an address using the sendmsg system
|
|
// call. p contains the non-control data to send, and oob contains the "out of
|
|
// band" control data. The flags are passed to sendmsg. The number of
|
|
// non-control bytes actually written to the socket is returned.
|
|
//
|
|
// Some socket types do not support sending control data without accompanying
|
|
// non-control data. If p is empty, and oob contains control data, and the
|
|
// underlying socket type is not SOCK_DGRAM, p will be treated as containing a
|
|
// single '\0' and the return value will indicate zero bytes sent.
|
|
//
|
|
// The Go function Recvmsg, if called with an empty p and a non-empty oob,
|
|
// will read and ignore this additional '\0'. If the message is received by
|
|
// code that does not use Recvmsg, or that does not use Go at all, that code
|
|
// will need to be written to expect and ignore the additional '\0'.
|
|
//
|
|
// If you need to send non-empty oob with p actually empty, and if the
|
|
// underlying socket type supports it, you can do so via a raw system call as
|
|
// follows:
|
|
//
|
|
// msg := &unix.Msghdr{
|
|
// Control: &oob[0],
|
|
// }
|
|
// msg.SetControllen(len(oob))
|
|
// n, _, errno := unix.Syscall(unix.SYS_SENDMSG, uintptr(fd), uintptr(unsafe.Pointer(msg)), flags)
|
|
func SendmsgN(fd int, p, oob []byte, to Sockaddr, flags int) (n int, err error) {
|
|
var iov [1]Iovec
|
|
if len(p) > 0 {
|
|
iov[0].Base = &p[0]
|
|
iov[0].SetLen(len(p))
|
|
}
|
|
var ptr unsafe.Pointer
|
|
var salen _Socklen
|
|
if to != nil {
|
|
ptr, salen, err = to.sockaddr()
|
|
if err != nil {
|
|
return 0, err
|
|
}
|
|
}
|
|
return sendmsgN(fd, iov[:], oob, ptr, salen, flags)
|
|
}
|
|
|
|
// SendmsgBuffers sends a message on a socket to an address using the sendmsg
|
|
// system call. This function is equivalent to SendmsgN, but the non-control
|
|
// data is gathered from buffers.
|
|
func SendmsgBuffers(fd int, buffers [][]byte, oob []byte, to Sockaddr, flags int) (n int, err error) {
|
|
iov := make([]Iovec, len(buffers))
|
|
for i := range buffers {
|
|
if len(buffers[i]) > 0 {
|
|
iov[i].Base = &buffers[i][0]
|
|
iov[i].SetLen(len(buffers[i]))
|
|
} else {
|
|
iov[i].Base = (*byte)(unsafe.Pointer(&_zero))
|
|
}
|
|
}
|
|
var ptr unsafe.Pointer
|
|
var salen _Socklen
|
|
if to != nil {
|
|
ptr, salen, err = to.sockaddr()
|
|
if err != nil {
|
|
return 0, err
|
|
}
|
|
}
|
|
return sendmsgN(fd, iov, oob, ptr, salen, flags)
|
|
}
|
|
|
|
func Send(s int, buf []byte, flags int) (err error) {
|
|
return sendto(s, buf, flags, nil, 0)
|
|
}
|
|
|
|
func Sendto(fd int, p []byte, flags int, to Sockaddr) (err error) {
|
|
var ptr unsafe.Pointer
|
|
var salen _Socklen
|
|
if to != nil {
|
|
ptr, salen, err = to.sockaddr()
|
|
if err != nil {
|
|
return err
|
|
}
|
|
}
|
|
return sendto(fd, p, flags, ptr, salen)
|
|
}
|
|
|
|
func SetsockoptByte(fd, level, opt int, value byte) (err error) {
|
|
return setsockopt(fd, level, opt, unsafe.Pointer(&value), 1)
|
|
}
|
|
|
|
func SetsockoptInt(fd, level, opt int, value int) (err error) {
|
|
var n = int32(value)
|
|
return setsockopt(fd, level, opt, unsafe.Pointer(&n), 4)
|
|
}
|
|
|
|
func SetsockoptInet4Addr(fd, level, opt int, value [4]byte) (err error) {
|
|
return setsockopt(fd, level, opt, unsafe.Pointer(&value[0]), 4)
|
|
}
|
|
|
|
func SetsockoptIPMreq(fd, level, opt int, mreq *IPMreq) (err error) {
|
|
return setsockopt(fd, level, opt, unsafe.Pointer(mreq), SizeofIPMreq)
|
|
}
|
|
|
|
func SetsockoptIPv6Mreq(fd, level, opt int, mreq *IPv6Mreq) (err error) {
|
|
return setsockopt(fd, level, opt, unsafe.Pointer(mreq), SizeofIPv6Mreq)
|
|
}
|
|
|
|
func SetsockoptICMPv6Filter(fd, level, opt int, filter *ICMPv6Filter) error {
|
|
return setsockopt(fd, level, opt, unsafe.Pointer(filter), SizeofICMPv6Filter)
|
|
}
|
|
|
|
func SetsockoptLinger(fd, level, opt int, l *Linger) (err error) {
|
|
return setsockopt(fd, level, opt, unsafe.Pointer(l), SizeofLinger)
|
|
}
|
|
|
|
func SetsockoptString(fd, level, opt int, s string) (err error) {
|
|
var p unsafe.Pointer
|
|
if len(s) > 0 {
|
|
p = unsafe.Pointer(&[]byte(s)[0])
|
|
}
|
|
return setsockopt(fd, level, opt, p, uintptr(len(s)))
|
|
}
|
|
|
|
func SetsockoptTimeval(fd, level, opt int, tv *Timeval) (err error) {
|
|
return setsockopt(fd, level, opt, unsafe.Pointer(tv), unsafe.Sizeof(*tv))
|
|
}
|
|
|
|
func SetsockoptUint64(fd, level, opt int, value uint64) (err error) {
|
|
return setsockopt(fd, level, opt, unsafe.Pointer(&value), 8)
|
|
}
|
|
|
|
func Socket(domain, typ, proto int) (fd int, err error) {
|
|
if domain == AF_INET6 && SocketDisableIPv6 {
|
|
return -1, EAFNOSUPPORT
|
|
}
|
|
fd, err = socket(domain, typ, proto)
|
|
return
|
|
}
|
|
|
|
func Socketpair(domain, typ, proto int) (fd [2]int, err error) {
|
|
var fdx [2]int32
|
|
err = socketpair(domain, typ, proto, &fdx)
|
|
if err == nil {
|
|
fd[0] = int(fdx[0])
|
|
fd[1] = int(fdx[1])
|
|
}
|
|
return
|
|
}
|
|
|
|
var ioSync int64
|
|
|
|
func CloseOnExec(fd int) { fcntl(fd, F_SETFD, FD_CLOEXEC) }
|
|
|
|
func SetNonblock(fd int, nonblocking bool) (err error) {
|
|
flag, err := fcntl(fd, F_GETFL, 0)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
if (flag&O_NONBLOCK != 0) == nonblocking {
|
|
return nil
|
|
}
|
|
if nonblocking {
|
|
flag |= O_NONBLOCK
|
|
} else {
|
|
flag &= ^O_NONBLOCK
|
|
}
|
|
_, err = fcntl(fd, F_SETFL, flag)
|
|
return err
|
|
}
|
|
|
|
// Exec calls execve(2), which replaces the calling executable in the process
|
|
// tree. argv0 should be the full path to an executable ("/bin/ls") and the
|
|
// executable name should also be the first argument in argv (["ls", "-l"]).
|
|
// envv are the environment variables that should be passed to the new
|
|
// process (["USER=go", "PWD=/tmp"]).
|
|
func Exec(argv0 string, argv []string, envv []string) error {
|
|
return syscall.Exec(argv0, argv, envv)
|
|
}
|
|
|
|
// Lutimes sets the access and modification times tv on path. If path refers to
|
|
// a symlink, it is not dereferenced and the timestamps are set on the symlink.
|
|
// If tv is nil, the access and modification times are set to the current time.
|
|
// Otherwise tv must contain exactly 2 elements, with access time as the first
|
|
// element and modification time as the second element.
|
|
func Lutimes(path string, tv []Timeval) error {
|
|
if tv == nil {
|
|
return UtimesNanoAt(AT_FDCWD, path, nil, AT_SYMLINK_NOFOLLOW)
|
|
}
|
|
if len(tv) != 2 {
|
|
return EINVAL
|
|
}
|
|
ts := []Timespec{
|
|
NsecToTimespec(TimevalToNsec(tv[0])),
|
|
NsecToTimespec(TimevalToNsec(tv[1])),
|
|
}
|
|
return UtimesNanoAt(AT_FDCWD, path, ts, AT_SYMLINK_NOFOLLOW)
|
|
}
|
|
|
|
// emptyIovecs reports whether there are no bytes in the slice of Iovec.
|
|
func emptyIovecs(iov []Iovec) bool {
|
|
for i := range iov {
|
|
if iov[i].Len > 0 {
|
|
return false
|
|
}
|
|
}
|
|
return true
|
|
}
|
|
|
|
// Setrlimit sets a resource limit.
|
|
func Setrlimit(resource int, rlim *Rlimit) error {
|
|
// Just call the syscall version, because as of Go 1.21
|
|
// it will affect starting a new process.
|
|
return syscall.Setrlimit(resource, (*syscall.Rlimit)(rlim))
|
|
}
|