Switch to dep for dependency management

vendor/golang.org/x/sys/unix/syscall_linux.go

@@ -36,6 +36,59 @@ func Creat(path string, mode uint32) (fd int, err error) {
 	return Open(path, O_CREAT|O_WRONLY|O_TRUNC, mode)
 }
 
+//sys	fchmodat(dirfd int, path string, mode uint32) (err error)
+
+func Fchmodat(dirfd int, path string, mode uint32, flags int) (err error) {
+	// Linux fchmodat doesn't support the flags parameter. Mimick glibc's behavior
+	// and check the flags. Otherwise the mode would be applied to the symlink
+	// destination which is not what the user expects.
+	if flags&^AT_SYMLINK_NOFOLLOW != 0 {
+		return EINVAL
+	} else if flags&AT_SYMLINK_NOFOLLOW != 0 {
+		return EOPNOTSUPP
+	}
+	return fchmodat(dirfd, path, mode)
+}
+
+//sys	ioctl(fd int, req uint, arg uintptr) (err error)
+
+// ioctl itself should not be exposed directly, but additional get/set
+// functions for specific types are permissible.
+
+// IoctlSetInt performs an ioctl operation which sets an integer value
+// on fd, using the specified request number.
+func IoctlSetInt(fd int, req uint, value int) error {
+	return ioctl(fd, req, uintptr(value))
+}
+
+func ioctlSetWinsize(fd int, req uint, value *Winsize) error {
+	return ioctl(fd, req, uintptr(unsafe.Pointer(value)))
+}
+
+func ioctlSetTermios(fd int, req uint, value *Termios) error {
+	return ioctl(fd, req, uintptr(unsafe.Pointer(value)))
+}
+
+// IoctlGetInt performs an ioctl operation which gets an integer value
+// from fd, using the specified request number.
+func IoctlGetInt(fd int, req uint) (int, error) {
+	var value int
+	err := ioctl(fd, req, uintptr(unsafe.Pointer(&value)))
+	return value, err
+}
+
+func IoctlGetWinsize(fd int, req uint) (*Winsize, error) {
+	var value Winsize
+	err := ioctl(fd, req, uintptr(unsafe.Pointer(&value)))
+	return &value, err
+}
+
+func IoctlGetTermios(fd int, req uint) (*Termios, error) {
+	var value Termios
+	err := ioctl(fd, req, uintptr(unsafe.Pointer(&value)))
+	return &value, err
+}
+
 //sys	Linkat(olddirfd int, oldpath string, newdirfd int, newpath string, flags int) (err error)
 
 func Link(oldpath string, newpath string) (err error) {
@@ -95,8 +148,6 @@ func Unlink(path string) error {
 
 //sys	Unlinkat(dirfd int, path string, flags int) (err error)
 
-//sys	utimes(path string, times *[2]Timeval) (err error)
-
 func Utimes(path string, tv []Timeval) error {
 	if tv == nil {
 		err := utimensat(AT_FDCWD, path, nil, 0)
@@ -154,20 +205,14 @@ func UtimesNanoAt(dirfd int, path string, ts []Timespec, flags int) error {
 	return utimensat(dirfd, path, (*[2]Timespec)(unsafe.Pointer(&ts[0])), flags)
 }
 
-//sys	futimesat(dirfd 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)
+		return futimesat(dirfd, path, nil)
 	}
 	if len(tv) != 2 {
 		return EINVAL
 	}
-	return futimesat(dirfd, pathp, (*[2]Timeval)(unsafe.Pointer(&tv[0])))
+	return futimesat(dirfd, path, (*[2]Timeval)(unsafe.Pointer(&tv[0])))
 }
 
 func Futimes(fd int, tv []Timeval) (err error) {
@@ -202,7 +247,7 @@ func Getgroups() (gids []int, err error) {
 		return nil, nil
 	}
 
-	// Sanity check group count.  Max is 1<<16 on Linux.
+	// Sanity check group count. Max is 1<<16 on Linux.
 	if n < 0 || n > 1<<20 {
 		return nil, EINVAL
 	}
@@ -237,8 +282,8 @@ type WaitStatus uint32
 // 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.  At least that's the idea.
-// There are various irregularities.  For example, the
+// is in the high bits. At least that's the idea.
+// There are various irregularities. For example, the
 // "continued" status is 0xFFFF, distinguishing itself
 // from stopped via the core dump bit.
 
@@ -299,10 +344,14 @@ func Wait4(pid int, wstatus *WaitStatus, options int, rusage *Rusage) (wpid int,
 	return
 }
 
-func Mkfifo(path string, mode uint32) (err error) {
+func Mkfifo(path string, mode uint32) error {
 	return Mknod(path, mode|S_IFIFO, 0)
 }
 
+func Mkfifoat(dirfd int, path string, mode uint32) error {
+	return Mknodat(dirfd, path, mode|S_IFIFO, 0)
+}
+
 func (sa *SockaddrInet4) sockaddr() (unsafe.Pointer, _Socklen, error) {
 	if sa.Port < 0 || sa.Port > 0xFFFF {
 		return nil, 0, EINVAL
@@ -356,6 +405,7 @@ func (sa *SockaddrUnix) sockaddr() (unsafe.Pointer, _Socklen, error) {
 	return unsafe.Pointer(&sa.raw), sl, nil
 }
 
+// SockaddrLinklayer implements the Sockaddr interface for AF_PACKET type sockets.
 type SockaddrLinklayer struct {
 	Protocol uint16
 	Ifindex  int
@@ -382,6 +432,7 @@ func (sa *SockaddrLinklayer) sockaddr() (unsafe.Pointer, _Socklen, error) {
 	return unsafe.Pointer(&sa.raw), SizeofSockaddrLinklayer, nil
 }
 
+// SockaddrNetlink implements the Sockaddr interface for AF_NETLINK type sockets.
 type SockaddrNetlink struct {
 	Family uint16
 	Pad    uint16
@@ -398,6 +449,8 @@ func (sa *SockaddrNetlink) sockaddr() (unsafe.Pointer, _Socklen, error) {
 	return unsafe.Pointer(&sa.raw), SizeofSockaddrNetlink, nil
 }
 
+// SockaddrHCI implements the Sockaddr interface for AF_BLUETOOTH type sockets
+// using the HCI protocol.
 type SockaddrHCI struct {
 	Dev     uint16
 	Channel uint16
@@ -411,6 +464,72 @@ func (sa *SockaddrHCI) sockaddr() (unsafe.Pointer, _Socklen, error) {
 	return unsafe.Pointer(&sa.raw), SizeofSockaddrHCI, nil
 }
 
+// SockaddrL2 implements the Sockaddr interface for AF_BLUETOOTH type sockets
+// using the L2CAP protocol.
+type SockaddrL2 struct {
+	PSM      uint16
+	CID      uint16
+	Addr     [6]uint8
+	AddrType uint8
+	raw      RawSockaddrL2
+}
+
+func (sa *SockaddrL2) sockaddr() (unsafe.Pointer, _Socklen, error) {
+	sa.raw.Family = AF_BLUETOOTH
+	psm := (*[2]byte)(unsafe.Pointer(&sa.raw.Psm))
+	psm[0] = byte(sa.PSM)
+	psm[1] = byte(sa.PSM >> 8)
+	for i := 0; i < len(sa.Addr); i++ {
+		sa.raw.Bdaddr[i] = sa.Addr[len(sa.Addr)-1-i]
+	}
+	cid := (*[2]byte)(unsafe.Pointer(&sa.raw.Cid))
+	cid[0] = byte(sa.CID)
+	cid[1] = byte(sa.CID >> 8)
+	sa.raw.Bdaddr_type = sa.AddrType
+	return unsafe.Pointer(&sa.raw), SizeofSockaddrL2, nil
+}
+
+// SockaddrRFCOMM implements the Sockaddr interface for AF_BLUETOOTH type sockets
+// using the RFCOMM protocol.
+//
+// Server example:
+//
+//      fd, _ := Socket(AF_BLUETOOTH, SOCK_STREAM, BTPROTO_RFCOMM)
+//      _ = unix.Bind(fd, &unix.SockaddrRFCOMM{
+//      	Channel: 1,
+//      	Addr:    [6]uint8{0, 0, 0, 0, 0, 0}, // BDADDR_ANY or 00:00:00:00:00:00
+//      })
+//      _ = Listen(fd, 1)
+//      nfd, sa, _ := Accept(fd)
+//      fmt.Printf("conn addr=%v fd=%d", sa.(*unix.SockaddrRFCOMM).Addr, nfd)
+//      Read(nfd, buf)
+//
+// Client example:
+//
+//      fd, _ := Socket(AF_BLUETOOTH, SOCK_STREAM, BTPROTO_RFCOMM)
+//      _ = Connect(fd, &SockaddrRFCOMM{
+//      	Channel: 1,
+//      	Addr:    [6]byte{0x11, 0x22, 0x33, 0xaa, 0xbb, 0xcc}, // CC:BB:AA:33:22:11
+//      })
+//      Write(fd, []byte(`hello`))
+type SockaddrRFCOMM struct {
+	// Addr represents a bluetooth address, byte ordering is little-endian.
+	Addr [6]uint8
+
+	// Channel is a designated bluetooth channel, only 1-30 are available for use.
+	// Since Linux 2.6.7 and further zero value is the first available channel.
+	Channel uint8
+
+	raw RawSockaddrRFCOMM
+}
+
+func (sa *SockaddrRFCOMM) sockaddr() (unsafe.Pointer, _Socklen, error) {
+	sa.raw.Family = AF_BLUETOOTH
+	sa.raw.Channel = sa.Channel
+	sa.raw.Bdaddr = sa.Addr
+	return unsafe.Pointer(&sa.raw), SizeofSockaddrRFCOMM, nil
+}
+
 // SockaddrCAN implements the Sockaddr interface for AF_CAN type sockets.
 // The RxID and TxID fields are used for transport protocol addressing in
 // (CAN_TP16, CAN_TP20, CAN_MCNET, and CAN_ISOTP), they can be left with
@@ -452,7 +571,128 @@ func (sa *SockaddrCAN) sockaddr() (unsafe.Pointer, _Socklen, error) {
 	return unsafe.Pointer(&sa.raw), SizeofSockaddrCAN, nil
 }
 
-func anyToSockaddr(rsa *RawSockaddrAny) (Sockaddr, error) {
+// SockaddrALG implements the Sockaddr interface for AF_ALG type sockets.
+// SockaddrALG enables userspace access to the Linux kernel's cryptography
+// subsystem. The Type and Name fields specify which type of hash or cipher
+// should be used with a given socket.
+//
+// To create a file descriptor that provides access to a hash or cipher, both
+// Bind and Accept must be used. Once the setup process is complete, input
+// data can be written to the socket, processed by the kernel, and then read
+// back as hash output or ciphertext.
+//
+// Here is an example of using an AF_ALG socket with SHA1 hashing.
+// The initial socket setup process is as follows:
+//
+//      // Open a socket to perform SHA1 hashing.
+//      fd, _ := unix.Socket(unix.AF_ALG, unix.SOCK_SEQPACKET, 0)
+//      addr := &unix.SockaddrALG{Type: "hash", Name: "sha1"}
+//      unix.Bind(fd, addr)
+//      // Note: unix.Accept does not work at this time; must invoke accept()
+//      // manually using unix.Syscall.
+//      hashfd, _, _ := unix.Syscall(unix.SYS_ACCEPT, uintptr(fd), 0, 0)
+//
+// Once a file descriptor has been returned from Accept, it may be used to
+// perform SHA1 hashing. The descriptor is not safe for concurrent use, but
+// may be re-used repeatedly with subsequent Write and Read operations.
+//
+// When hashing a small byte slice or string, a single Write and Read may
+// be used:
+//
+//      // Assume hashfd is already configured using the setup process.
+//      hash := os.NewFile(hashfd, "sha1")
+//      // Hash an input string and read the results. Each Write discards
+//      // previous hash state. Read always reads the current state.
+//      b := make([]byte, 20)
+//      for i := 0; i < 2; i++ {
+//          io.WriteString(hash, "Hello, world.")
+//          hash.Read(b)
+//          fmt.Println(hex.EncodeToString(b))
+//      }
+//      // Output:
+//      // 2ae01472317d1935a84797ec1983ae243fc6aa28
+//      // 2ae01472317d1935a84797ec1983ae243fc6aa28
+//
+// For hashing larger byte slices, or byte streams such as those read from
+// a file or socket, use Sendto with MSG_MORE to instruct the kernel to update
+// the hash digest instead of creating a new one for a given chunk and finalizing it.
+//
+//      // Assume hashfd and addr are already configured using the setup process.
+//      hash := os.NewFile(hashfd, "sha1")
+//      // Hash the contents of a file.
+//      f, _ := os.Open("/tmp/linux-4.10-rc7.tar.xz")
+//      b := make([]byte, 4096)
+//      for {
+//          n, err := f.Read(b)
+//          if err == io.EOF {
+//              break
+//          }
+//          unix.Sendto(hashfd, b[:n], unix.MSG_MORE, addr)
+//      }
+//      hash.Read(b)
+//      fmt.Println(hex.EncodeToString(b))
+//      // Output: 85cdcad0c06eef66f805ecce353bec9accbeecc5
+//
+// For more information, see: http://www.chronox.de/crypto-API/crypto/userspace-if.html.
+type SockaddrALG struct {
+	Type    string
+	Name    string
+	Feature uint32
+	Mask    uint32
+	raw     RawSockaddrALG
+}
+
+func (sa *SockaddrALG) sockaddr() (unsafe.Pointer, _Socklen, error) {
+	// Leave room for NUL byte terminator.
+	if len(sa.Type) > 13 {
+		return nil, 0, EINVAL
+	}
+	if len(sa.Name) > 63 {
+		return nil, 0, EINVAL
+	}
+
+	sa.raw.Family = AF_ALG
+	sa.raw.Feat = sa.Feature
+	sa.raw.Mask = sa.Mask
+
+	typ, err := ByteSliceFromString(sa.Type)
+	if err != nil {
+		return nil, 0, err
+	}
+	name, err := ByteSliceFromString(sa.Name)
+	if err != nil {
+		return nil, 0, err
+	}
+
+	copy(sa.raw.Type[:], typ)
+	copy(sa.raw.Name[:], name)
+
+	return unsafe.Pointer(&sa.raw), SizeofSockaddrALG, nil
+}
+
+// SockaddrVM implements the Sockaddr interface for AF_VSOCK type sockets.
+// SockaddrVM provides access to Linux VM sockets: a mechanism that enables
+// bidirectional communication between a hypervisor and its guest virtual
+// machines.
+type SockaddrVM struct {
+	// CID and Port specify a context ID and port address for a VM socket.
+	// Guests have a unique CID, and hosts may have a well-known CID of:
+	//  - VMADDR_CID_HYPERVISOR: refers to the hypervisor process.
+	//  - VMADDR_CID_HOST: refers to other processes on the host.
+	CID  uint32
+	Port uint32
+	raw  RawSockaddrVM
+}
+
+func (sa *SockaddrVM) sockaddr() (unsafe.Pointer, _Socklen, error) {
+	sa.raw.Family = AF_VSOCK
+	sa.raw.Port = sa.Port
+	sa.raw.Cid = sa.CID
+
+	return unsafe.Pointer(&sa.raw), SizeofSockaddrVM, nil
+}
+
+func anyToSockaddr(fd int, rsa *RawSockaddrAny) (Sockaddr, error) {
 	switch rsa.Addr.Family {
 	case AF_NETLINK:
 		pp := (*RawSockaddrNetlink)(unsafe.Pointer(rsa))
@@ -521,6 +761,38 @@ func anyToSockaddr(rsa *RawSockaddrAny) (Sockaddr, error) {
 			sa.Addr[i] = pp.Addr[i]
 		}
 		return sa, nil
+
+	case AF_VSOCK:
+		pp := (*RawSockaddrVM)(unsafe.Pointer(rsa))
+		sa := &SockaddrVM{
+			CID:  pp.Cid,
+			Port: pp.Port,
+		}
+		return sa, nil
+	case AF_BLUETOOTH:
+		proto, err := GetsockoptInt(fd, SOL_SOCKET, SO_PROTOCOL)
+		if err != nil {
+			return nil, err
+		}
+		// only BTPROTO_L2CAP and BTPROTO_RFCOMM can accept connections
+		switch proto {
+		case BTPROTO_L2CAP:
+			pp := (*RawSockaddrL2)(unsafe.Pointer(rsa))
+			sa := &SockaddrL2{
+				PSM:      pp.Psm,
+				CID:      pp.Cid,
+				Addr:     pp.Bdaddr,
+				AddrType: pp.Bdaddr_type,
+			}
+			return sa, nil
+		case BTPROTO_RFCOMM:
+			pp := (*RawSockaddrRFCOMM)(unsafe.Pointer(rsa))
+			sa := &SockaddrRFCOMM{
+				Channel: pp.Channel,
+				Addr:    pp.Bdaddr,
+			}
+			return sa, nil
+		}
 	}
 	return nil, EAFNOSUPPORT
 }
@@ -532,7 +804,7 @@ func Accept(fd int) (nfd int, sa Sockaddr, err error) {
 	if err != nil {
 		return
 	}
-	sa, err = anyToSockaddr(&rsa)
+	sa, err = anyToSockaddr(fd, &rsa)
 	if err != nil {
 		Close(nfd)
 		nfd = 0
@@ -550,7 +822,7 @@ func Accept4(fd int, flags int) (nfd int, sa Sockaddr, err error) {
 	if len > SizeofSockaddrAny {
 		panic("RawSockaddrAny too small")
 	}
-	sa, err = anyToSockaddr(&rsa)
+	sa, err = anyToSockaddr(fd, &rsa)
 	if err != nil {
 		Close(nfd)
 		nfd = 0
@@ -564,20 +836,7 @@ func Getsockname(fd int) (sa Sockaddr, err error) {
 	if err = getsockname(fd, &rsa, &len); err != nil {
 		return
 	}
-	return anyToSockaddr(&rsa)
-}
-
-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
+	return anyToSockaddr(fd, &rsa)
 }
 
 func GetsockoptIPMreqn(fd, level, opt int) (*IPMreqn, error) {
@@ -587,27 +846,6 @@ func GetsockoptIPMreqn(fd, level, opt int) (*IPMreqn, error) {
 	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 GetsockoptUcred(fd, level, opt int) (*Ucred, error) {
 	var value Ucred
 	vallen := _Socklen(SizeofUcred)
@@ -615,10 +853,142 @@ func GetsockoptUcred(fd, level, opt int) (*Ucred, error) {
 	return &value, err
 }
 
+func GetsockoptTCPInfo(fd, level, opt int) (*TCPInfo, error) {
+	var value TCPInfo
+	vallen := _Socklen(SizeofTCPInfo)
+	err := getsockopt(fd, level, opt, unsafe.Pointer(&value), &vallen)
+	return &value, err
+}
+
+// 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 {
+		if err == ERANGE {
+			buf = make([]byte, vallen)
+			err = getsockopt(fd, level, opt, unsafe.Pointer(&buf[0]), &vallen)
+		}
+		if err != nil {
+			return "", err
+		}
+	}
+	return string(buf[:vallen-1]), nil
+}
+
 func SetsockoptIPMreqn(fd, level, opt int, mreq *IPMreqn) (err error) {
 	return setsockopt(fd, level, opt, unsafe.Pointer(mreq), unsafe.Sizeof(*mreq))
 }
 
+// Keyctl Commands (http://man7.org/linux/man-pages/man2/keyctl.2.html)
+
+// KeyctlInt calls keyctl commands in which each argument is an int.
+// These commands are KEYCTL_REVOKE, KEYCTL_CHOWN, KEYCTL_CLEAR, KEYCTL_LINK,
+// KEYCTL_UNLINK, KEYCTL_NEGATE, KEYCTL_SET_REQKEY_KEYRING, KEYCTL_SET_TIMEOUT,
+// KEYCTL_ASSUME_AUTHORITY, KEYCTL_SESSION_TO_PARENT, KEYCTL_REJECT,
+// KEYCTL_INVALIDATE, and KEYCTL_GET_PERSISTENT.
+//sys	KeyctlInt(cmd int, arg2 int, arg3 int, arg4 int, arg5 int) (ret int, err error) = SYS_KEYCTL
+
+// KeyctlBuffer calls keyctl commands in which the third and fourth
+// arguments are a buffer and its length, respectively.
+// These commands are KEYCTL_UPDATE, KEYCTL_READ, and KEYCTL_INSTANTIATE.
+//sys	KeyctlBuffer(cmd int, arg2 int, buf []byte, arg5 int) (ret int, err error) = SYS_KEYCTL
+
+// KeyctlString calls keyctl commands which return a string.
+// These commands are KEYCTL_DESCRIBE and KEYCTL_GET_SECURITY.
+func KeyctlString(cmd int, id int) (string, error) {
+	// We must loop as the string data may change in between the syscalls.
+	// We could allocate a large buffer here to reduce the chance that the
+	// syscall needs to be called twice; however, this is unnecessary as
+	// the performance loss is negligible.
+	var buffer []byte
+	for {
+		// Try to fill the buffer with data
+		length, err := KeyctlBuffer(cmd, id, buffer, 0)
+		if err != nil {
+			return "", err
+		}
+
+		// Check if the data was written
+		if length <= len(buffer) {
+			// Exclude the null terminator
+			return string(buffer[:length-1]), nil
+		}
+
+		// Make a bigger buffer if needed
+		buffer = make([]byte, length)
+	}
+}
+
+// Keyctl commands with special signatures.
+
+// KeyctlGetKeyringID implements the KEYCTL_GET_KEYRING_ID command.
+// See the full documentation at:
+// http://man7.org/linux/man-pages/man3/keyctl_get_keyring_ID.3.html
+func KeyctlGetKeyringID(id int, create bool) (ringid int, err error) {
+	createInt := 0
+	if create {
+		createInt = 1
+	}
+	return KeyctlInt(KEYCTL_GET_KEYRING_ID, id, createInt, 0, 0)
+}
+
+// KeyctlSetperm implements the KEYCTL_SETPERM command. The perm value is the
+// key handle permission mask as described in the "keyctl setperm" section of
+// http://man7.org/linux/man-pages/man1/keyctl.1.html.
+// See the full documentation at:
+// http://man7.org/linux/man-pages/man3/keyctl_setperm.3.html
+func KeyctlSetperm(id int, perm uint32) error {
+	_, err := KeyctlInt(KEYCTL_SETPERM, id, int(perm), 0, 0)
+	return err
+}
+
+//sys	keyctlJoin(cmd int, arg2 string) (ret int, err error) = SYS_KEYCTL
+
+// KeyctlJoinSessionKeyring implements the KEYCTL_JOIN_SESSION_KEYRING command.
+// See the full documentation at:
+// http://man7.org/linux/man-pages/man3/keyctl_join_session_keyring.3.html
+func KeyctlJoinSessionKeyring(name string) (ringid int, err error) {
+	return keyctlJoin(KEYCTL_JOIN_SESSION_KEYRING, name)
+}
+
+//sys	keyctlSearch(cmd int, arg2 int, arg3 string, arg4 string, arg5 int) (ret int, err error) = SYS_KEYCTL
+
+// KeyctlSearch implements the KEYCTL_SEARCH command.
+// See the full documentation at:
+// http://man7.org/linux/man-pages/man3/keyctl_search.3.html
+func KeyctlSearch(ringid int, keyType, description string, destRingid int) (id int, err error) {
+	return keyctlSearch(KEYCTL_SEARCH, ringid, keyType, description, destRingid)
+}
+
+//sys	keyctlIOV(cmd int, arg2 int, payload []Iovec, arg5 int) (err error) = SYS_KEYCTL
+
+// KeyctlInstantiateIOV implements the KEYCTL_INSTANTIATE_IOV command. This
+// command is similar to KEYCTL_INSTANTIATE, except that the payload is a slice
+// of Iovec (each of which represents a buffer) instead of a single buffer.
+// See the full documentation at:
+// http://man7.org/linux/man-pages/man3/keyctl_instantiate_iov.3.html
+func KeyctlInstantiateIOV(id int, payload []Iovec, ringid int) error {
+	return keyctlIOV(KEYCTL_INSTANTIATE_IOV, id, payload, ringid)
+}
+
+//sys	keyctlDH(cmd int, arg2 *KeyctlDHParams, buf []byte) (ret int, err error) = SYS_KEYCTL
+
+// KeyctlDHCompute implements the KEYCTL_DH_COMPUTE command. This command
+// computes a Diffie-Hellman shared secret based on the provide params. The
+// secret is written to the provided buffer and the returned size is the number
+// of bytes written (returning an error if there is insufficient space in the
+// buffer). If a nil buffer is passed in, this function returns the minimum
+// buffer length needed to store the appropriate data. Note that this differs
+// from KEYCTL_READ's behavior which always returns the requested payload size.
+// See the full documentation at:
+// http://man7.org/linux/man-pages/man3/keyctl_dh_compute.3.html
+func KeyctlDHCompute(params *KeyctlDHParams, buffer []byte) (size int, err error) {
+	return keyctlDH(KEYCTL_DH_COMPUTE, params, buffer)
+}
+
 func Recvmsg(fd int, p, oob []byte, flags int) (n, oobn int, recvflags int, from Sockaddr, err error) {
 	var msg Msghdr
 	var rsa RawSockaddrAny
@@ -626,17 +996,24 @@ func Recvmsg(fd int, p, oob []byte, flags int) (n, oobn int, recvflags int, from
 	msg.Namelen = uint32(SizeofSockaddrAny)
 	var iov Iovec
 	if len(p) > 0 {
-		iov.Base = (*byte)(unsafe.Pointer(&p[0]))
+		iov.Base = &p[0]
 		iov.SetLen(len(p))
 	}
 	var dummy byte
 	if len(oob) > 0 {
-		// receive at least one normal byte
 		if len(p) == 0 {
-			iov.Base = &dummy
-			iov.SetLen(1)
+			var sockType int
+			sockType, err = GetsockoptInt(fd, SOL_SOCKET, SO_TYPE)
+			if err != nil {
+				return
+			}
+			// receive at least one normal byte
+			if sockType != SOCK_DGRAM {
+				iov.Base = &dummy
+				iov.SetLen(1)
+			}
 		}
-		msg.Control = (*byte)(unsafe.Pointer(&oob[0]))
+		msg.Control = &oob[0]
 		msg.SetControllen(len(oob))
 	}
 	msg.Iov = &iov
@@ -648,7 +1025,7 @@ func Recvmsg(fd int, p, oob []byte, flags int) (n, oobn int, recvflags int, from
 	recvflags = int(msg.Flags)
 	// source address is only specified if the socket is unconnected
 	if rsa.Addr.Family != AF_UNSPEC {
-		from, err = anyToSockaddr(&rsa)
+		from, err = anyToSockaddr(fd, &rsa)
 	}
 	return
 }
@@ -669,21 +1046,28 @@ func SendmsgN(fd int, p, oob []byte, to Sockaddr, flags int) (n int, err error)
 		}
 	}
 	var msg Msghdr
-	msg.Name = (*byte)(unsafe.Pointer(ptr))
+	msg.Name = (*byte)(ptr)
 	msg.Namelen = uint32(salen)
 	var iov Iovec
 	if len(p) > 0 {
-		iov.Base = (*byte)(unsafe.Pointer(&p[0]))
+		iov.Base = &p[0]
 		iov.SetLen(len(p))
 	}
 	var dummy byte
 	if len(oob) > 0 {
-		// send at least one normal byte
 		if len(p) == 0 {
-			iov.Base = &dummy
-			iov.SetLen(1)
+			var sockType int
+			sockType, err = GetsockoptInt(fd, SOL_SOCKET, SO_TYPE)
+			if err != nil {
+				return 0, err
+			}
+			// send at least one normal byte
+			if sockType != SOCK_DGRAM {
+				iov.Base = &dummy
+				iov.SetLen(1)
+			}
 		}
-		msg.Control = (*byte)(unsafe.Pointer(&oob[0]))
+		msg.Control = &oob[0]
 		msg.SetControllen(len(oob))
 	}
 	msg.Iov = &iov
@@ -713,7 +1097,7 @@ func ptracePeek(req int, pid int, addr uintptr, out []byte) (count int, err erro
 
 	var buf [sizeofPtr]byte
 
-	// Leading edge.  PEEKTEXT/PEEKDATA don't require aligned
+	// Leading edge. PEEKTEXT/PEEKDATA don't require aligned
 	// access (PEEKUSER warns that it might), but if we don't
 	// align our reads, we might straddle an unmapped page
 	// boundary and not get the bytes leading up to the page
@@ -752,6 +1136,10 @@ func PtracePeekData(pid int, addr uintptr, out []byte) (count int, err error) {
 	return ptracePeek(PTRACE_PEEKDATA, pid, addr, out)
 }
 
+func PtracePeekUser(pid int, addr uintptr, out []byte) (count int, err error) {
+	return ptracePeek(PTRACE_PEEKUSR, pid, addr, out)
+}
+
 func ptracePoke(pokeReq int, peekReq int, pid int, addr uintptr, data []byte) (count int, err error) {
 	// As for ptracePeek, we need to align our accesses to deal
 	// with the possibility of straddling an invalid page.
@@ -811,6 +1199,10 @@ func PtracePokeData(pid int, addr uintptr, data []byte) (count int, err error) {
 	return ptracePoke(PTRACE_POKEDATA, PTRACE_PEEKDATA, pid, addr, data)
 }
 
+func PtracePokeUser(pid int, addr uintptr, data []byte) (count int, err error) {
+	return ptracePoke(PTRACE_POKEUSR, PTRACE_PEEKUSR, pid, addr, data)
+}
+
 func PtraceGetRegs(pid int, regsout *PtraceRegs) (err error) {
 	return ptrace(PTRACE_GETREGS, pid, 0, uintptr(unsafe.Pointer(regsout)))
 }
@@ -850,40 +1242,10 @@ func Reboot(cmd int) (err error) {
 	return reboot(LINUX_REBOOT_MAGIC1, LINUX_REBOOT_MAGIC2, cmd, "")
 }
 
-func clen(n []byte) int {
-	for i := 0; i < len(n); i++ {
-		if n[i] == 0 {
-			return i
-		}
-	}
-	return len(n)
-}
-
 func ReadDirent(fd int, buf []byte) (n int, err error) {
 	return Getdents(fd, buf)
 }
 
-func ParseDirent(buf []byte, max int, names []string) (consumed int, count int, newnames []string) {
-	origlen := len(buf)
-	count = 0
-	for max != 0 && len(buf) > 0 {
-		dirent := (*Dirent)(unsafe.Pointer(&buf[0]))
-		buf = buf[dirent.Reclen:]
-		if dirent.Ino == 0 { // File absent in directory.
-			continue
-		}
-		bytes := (*[10000]byte)(unsafe.Pointer(&dirent.Name[0]))
-		var name = string(bytes[0:clen(bytes[:])])
-		if name == "." || name == ".." { // Useless names
-			continue
-		}
-		max--
-		count++
-		names = append(names, name)
-	}
-	return origlen - len(buf), count, names
-}
-
 //sys	mount(source string, target string, fstype string, flags uintptr, data *byte) (err error)
 
 func Mount(source string, target string, fstype string, flags uintptr, data string) (err error) {
@@ -907,26 +1269,30 @@ func Mount(source string, target string, fstype string, flags uintptr, data stri
  * Direct access
  */
 //sys	Acct(path string) (err error)
+//sys	AddKey(keyType string, description string, payload []byte, ringid int) (id int, err error)
 //sys	Adjtimex(buf *Timex) (state int, err error)
 //sys	Chdir(path string) (err error)
 //sys	Chroot(path string) (err error)
 //sys	ClockGettime(clockid int32, time *Timespec) (err error)
 //sys	Close(fd int) (err error)
+//sys	CopyFileRange(rfd int, roff *int64, wfd int, woff *int64, len int, flags int) (n int, err error)
 //sys	Dup(oldfd int) (fd int, err error)
 //sys	Dup3(oldfd int, newfd int, flags int) (err error)
-//sysnb	EpollCreate(size int) (fd int, err error)
 //sysnb	EpollCreate1(flag int) (fd int, err error)
 //sysnb	EpollCtl(epfd int, op int, fd int, event *EpollEvent) (err error)
+//sys	Eventfd(initval uint, flags int) (fd int, err error) = SYS_EVENTFD2
 //sys	Exit(code int) = SYS_EXIT_GROUP
-//sys	Faccessat(dirfd int, path string, mode uint32, flags int) (err error)
 //sys	Fallocate(fd int, mode uint32, off int64, len int64) (err error)
 //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	Fchownat(dirfd int, path string, uid int, gid int, flags int) (err error)
 //sys	fcntl(fd int, cmd int, arg int) (val int, err error)
 //sys	Fdatasync(fd int) (err error)
+//sys	Fgetxattr(fd int, attr string, dest []byte) (sz int, err error)
+//sys	Flistxattr(fd int, dest []byte) (sz int, err error)
 //sys	Flock(fd int, how int) (err error)
+//sys	Fremovexattr(fd int, attr string) (err error)
+//sys	Fsetxattr(fd int, attr string, dest []byte, flags int) (err error)
 //sys	Fsync(fd int) (err error)
 //sys	Getdents(fd int, buf []byte) (n int, err error) = SYS_GETDENTS64
 //sysnb	Getpgid(pid int) (pgid int, err error)
@@ -939,6 +1305,7 @@ func Getpgrp() (pid int) {
 //sysnb	Getpid() (pid int)
 //sysnb	Getppid() (ppid int)
 //sys	Getpriority(which int, who int) (prio int, err error)
+//sys	Getrandom(buf []byte, flags int) (n int, err error)
 //sysnb	Getrusage(who int, rusage *Rusage) (err error)
 //sysnb	Getsid(pid int) (sid int, err error)
 //sysnb	Gettid() (tid int)
@@ -948,16 +1315,24 @@ func Getpgrp() (pid int) {
 //sysnb	InotifyRmWatch(fd int, watchdesc uint32) (success int, err error)
 //sysnb	Kill(pid int, sig syscall.Signal) (err error)
 //sys	Klogctl(typ int, buf []byte) (n int, err error) = SYS_SYSLOG
+//sys	Lgetxattr(path string, attr string, dest []byte) (sz int, err error)
 //sys	Listxattr(path string, dest []byte) (sz int, err error)
+//sys	Llistxattr(path string, dest []byte) (sz int, err error)
+//sys	Lremovexattr(path string, attr string) (err error)
+//sys	Lsetxattr(path string, attr string, data []byte, flags int) (err error)
 //sys	Mkdirat(dirfd int, path string, mode uint32) (err error)
 //sys	Mknodat(dirfd int, path string, mode uint32, dev int) (err error)
 //sys	Nanosleep(time *Timespec, leftover *Timespec) (err error)
+//sys	PerfEventOpen(attr *PerfEventAttr, pid int, cpu int, groupFd int, flags int) (fd int, err error)
 //sys	PivotRoot(newroot string, putold string) (err error) = SYS_PIVOT_ROOT
 //sysnb prlimit(pid int, resource int, newlimit *Rlimit, old *Rlimit) (err error) = SYS_PRLIMIT64
 //sys   Prctl(option int, arg2 uintptr, arg3 uintptr, arg4 uintptr, arg5 uintptr) (err error)
+//sys	Pselect(nfd int, r *FdSet, w *FdSet, e *FdSet, timeout *Timespec, sigmask *Sigset_t) (n int, err error) = SYS_PSELECT6
 //sys	read(fd int, p []byte) (n int, err error)
 //sys	Removexattr(path string, attr string) (err error)
 //sys	Renameat(olddirfd int, oldpath string, newdirfd int, newpath string) (err error)
+//sys	Renameat2(olddirfd int, oldpath string, newdirfd int, newpath string, flags uint) (err error)
+//sys	RequestKey(keyType string, description string, callback string, destRingid int) (id int, err error)
 //sys	Setdomainname(p []byte) (err error)
 //sys	Sethostname(p []byte) (err error)
 //sysnb	Setpgid(pid int, pgid int) (err error)
@@ -980,7 +1355,9 @@ func Setgid(uid int) (err error) {
 
 //sys	Setpriority(which int, who int, prio int) (err error)
 //sys	Setxattr(path string, attr string, data []byte, flags int) (err error)
+//sys	Statx(dirfd int, path string, flags int, mask int, stat *Statx_t) (err error)
 //sys	Sync()
+//sys	Syncfs(fd int) (err error)
 //sysnb	Sysinfo(info *Sysinfo_t) (err error)
 //sys	Tee(rfd int, wfd int, len int, flags int) (n int64, err error)
 //sysnb	Tgkill(tgid int, tid int, sig syscall.Signal) (err error)
@@ -989,7 +1366,6 @@ func Setgid(uid int) (err error) {
 //sysnb	Uname(buf *Utsname) (err error)
 //sys	Unmount(target string, flags int) (err error) = SYS_UMOUNT2
 //sys	Unshare(flags int) (err error)
-//sys	Ustat(dev int, ubuf *Ustat_t) (err error)
 //sys	write(fd int, p []byte) (n int, err error)
 //sys	exitThread(code int) (err error) = SYS_EXIT
 //sys	readlen(fd int, p *byte, np int) (n int, err error) = SYS_READ
@@ -1015,14 +1391,104 @@ func Munmap(b []byte) (err error) {
 //sys	Madvise(b []byte, advice int) (err error)
 //sys	Mprotect(b []byte, prot int) (err error)
 //sys	Mlock(b []byte) (err error)
-//sys	Munlock(b []byte) (err error)
 //sys	Mlockall(flags int) (err error)
+//sys	Msync(b []byte, flags int) (err error)
+//sys	Munlock(b []byte) (err error)
 //sys	Munlockall() (err error)
 
+// Vmsplice splices user pages from a slice of Iovecs into a pipe specified by fd,
+// using the specified flags.
+func Vmsplice(fd int, iovs []Iovec, flags int) (int, error) {
+	n, _, errno := Syscall6(
+		SYS_VMSPLICE,
+		uintptr(fd),
+		uintptr(unsafe.Pointer(&iovs[0])),
+		uintptr(len(iovs)),
+		uintptr(flags),
+		0,
+		0,
+	)
+	if errno != 0 {
+		return 0, syscall.Errno(errno)
+	}
+
+	return int(n), nil
+}
+
+//sys	faccessat(dirfd int, path string, mode uint32) (err error)
+
+func Faccessat(dirfd int, path string, mode uint32, flags int) (err error) {
+	if flags & ^(AT_SYMLINK_NOFOLLOW|AT_EACCESS) != 0 {
+		return EINVAL
+	}
+
+	// The Linux kernel faccessat system call does not take any flags.
+	// The glibc faccessat implements the flags itself; see
+	// https://sourceware.org/git/?p=glibc.git;a=blob;f=sysdeps/unix/sysv/linux/faccessat.c;hb=HEAD
+	// Because people naturally expect syscall.Faccessat to act
+	// like C faccessat, we do the same.
+
+	if flags == 0 {
+		return faccessat(dirfd, path, mode)
+	}
+
+	var st Stat_t
+	if err := Fstatat(dirfd, path, &st, flags&AT_SYMLINK_NOFOLLOW); err != nil {
+		return err
+	}
+
+	mode &= 7
+	if mode == 0 {
+		return nil
+	}
+
+	var uid int
+	if flags&AT_EACCESS != 0 {
+		uid = Geteuid()
+	} else {
+		uid = Getuid()
+	}
+
+	if uid == 0 {
+		if mode&1 == 0 {
+			// Root can read and write any file.
+			return nil
+		}
+		if st.Mode&0111 != 0 {
+			// Root can execute any file that anybody can execute.
+			return nil
+		}
+		return EACCES
+	}
+
+	var fmode uint32
+	if uint32(uid) == st.Uid {
+		fmode = (st.Mode >> 6) & 7
+	} else {
+		var gid int
+		if flags&AT_EACCESS != 0 {
+			gid = Getegid()
+		} else {
+			gid = Getgid()
+		}
+
+		if uint32(gid) == st.Gid {
+			fmode = (st.Mode >> 3) & 7
+		} else {
+			fmode = st.Mode & 7
+		}
+	}
+
+	if fmode&mode == mode {
+		return nil
+	}
+
+	return EACCES
+}
+
 /*
  * Unimplemented
  */
-// AddKey
 // AfsSyscall
 // Alarm
 // ArchPrctl
@@ -1038,13 +1504,8 @@ func Munmap(b []byte) (err error) {
 // EpollCtlOld
 // EpollPwait
 // EpollWaitOld
-// Eventfd
 // Execve
-// Fgetxattr
-// Flistxattr
 // Fork
-// Fremovexattr
-// Fsetxattr
 // Futex
 // GetKernelSyms
 // GetMempolicy
@@ -1057,23 +1518,16 @@ func Munmap(b []byte) (err error) {
 // IoGetevents
 // IoSetup
 // IoSubmit
-// Ioctl
 // IoprioGet
 // IoprioSet
 // KexecLoad
-// Keyctl
-// Lgetxattr
-// Llistxattr
 // LookupDcookie
-// Lremovexattr
-// Lsetxattr
 // Mbind
 // MigratePages
 // Mincore
 // ModifyLdt
 // Mount
 // MovePages
-// Mprotect
 // MqGetsetattr
 // MqNotify
 // MqOpen
@@ -1085,8 +1539,6 @@ func Munmap(b []byte) (err error) {
 // Msgget
 // Msgrcv
 // Msgsnd
-// Msync
-// Newfstatat
 // Nfsservctl
 // Personality
 // Pselect6
@@ -1097,7 +1549,6 @@ func Munmap(b []byte) (err error) {
 // Readahead
 // Readv
 // RemapFilePages
-// RequestKey
 // RestartSyscall
 // RtSigaction
 // RtSigpending
@@ -1108,11 +1559,9 @@ func Munmap(b []byte) (err error) {
 // RtSigtimedwait
 // SchedGetPriorityMax
 // SchedGetPriorityMin
-// SchedGetaffinity
 // SchedGetparam
 // SchedGetscheduler
 // SchedRrGetInterval
-// SchedSetaffinity
 // SchedSetparam
 // SchedYield
 // Security
@@ -1146,7 +1595,6 @@ func Munmap(b []byte) (err error) {
 // Utimensat
 // Vfork
 // Vhangup
-// Vmsplice
 // Vserver
 // Waitid
 // _Sysctl