src/os/pidfd_linux.go

// Copyright 2023 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.

// Support for pidfd was added during the course of a few Linux releases:
//  v5.1: pidfd_send_signal syscall;
//  v5.2: CLONE_PIDFD flag for clone syscall;
//  v5.3: pidfd_open syscall, clone3 syscall;
//  v5.4: P_PIDFD idtype support for waitid syscall;
//  v5.6: pidfd_getfd syscall.
//
// N.B. Alternative Linux implementations may not follow this ordering. e.g.,
// QEMU user mode 7.2 added pidfd_open, but CLONE_PIDFD was not added until
// 8.0.

package os

import (
	"internal/syscall/unix"
	"runtime"
	"sync"
	"syscall"
	_ "unsafe" // for linkname
)

// ensurePidfd initializes the PidFD field in sysAttr if it is not already set.
// It returns the original or modified SysProcAttr struct and a flag indicating
// whether the PidFD should be duplicated before using.
func ensurePidfd(sysAttr *syscall.SysProcAttr) (*syscall.SysProcAttr, bool) {
	if !pidfdWorks() {
		return sysAttr, false
	}

	var pidfd int

	if sysAttr == nil {
		return &syscall.SysProcAttr{
			PidFD: &pidfd,
		}, false
	}
	if sysAttr.PidFD == nil {
		newSys := *sysAttr // copy
		newSys.PidFD = &pidfd
		return &newSys, false
	}

	return sysAttr, true
}

// getPidfd returns the value of sysAttr.PidFD (or its duplicate if needDup is
// set) and a flag indicating whether the value can be used.
func getPidfd(sysAttr *syscall.SysProcAttr, needDup bool) (uintptr, bool) {
	if !pidfdWorks() {
		return 0, false
	}

	h := *sysAttr.PidFD
	if needDup {
		dupH, e := unix.Fcntl(h, syscall.F_DUPFD_CLOEXEC, 0)
		if e != nil {
			return 0, false
		}
		h = dupH
	}
	return uintptr(h), true
}

// pidfdFind returns the process handle for pid.
func pidfdFind(pid int) (uintptr, error) {
	if !pidfdWorks() {
		return 0, syscall.ENOSYS
	}

	h, err := unix.PidFDOpen(pid, 0)
	if err != nil {
		return 0, convertESRCH(err)
	}
	return h, nil
}

// pidfdWait waits for the process to complete,
// and updates the process status to done.
func (p *Process) pidfdWait() (*ProcessState, error) {
	// When pidfd is used, there is no wait/kill race (described in CL 23967)
	// because the PID recycle issue doesn't exist (IOW, pidfd, unlike PID,
	// is guaranteed to refer to one particular process). Thus, there is no
	// need for the workaround (blockUntilWaitable + sigMu) from pidWait.
	//
	// We _do_ need to be careful about reuse of the pidfd FD number when
	// closing the pidfd. See handle for more details.
	handle, status := p.handleTransientAcquire()
	switch status {
	case statusDone:
		// Process already completed Wait, or was not found by
		// pidfdFind. Return ECHILD for consistency with what the wait
		// syscall would return.
		return nil, NewSyscallError("wait", syscall.ECHILD)
	case statusReleased:
		return nil, syscall.EINVAL
	}
	defer p.handleTransientRelease()

	var (
		info   unix.SiginfoChild
		rusage syscall.Rusage
	)
	err := ignoringEINTR(func() error {
		return unix.Waitid(unix.P_PIDFD, int(handle), &info, syscall.WEXITED, &rusage)
	})
	if err != nil {
		return nil, NewSyscallError("waitid", err)
	}

	// Update the Process status to statusDone.
	// This also releases a reference to the handle.
	p.doRelease(statusDone)

	return &ProcessState{
		pid:    int(info.Pid),
		status: info.WaitStatus(),
		rusage: &rusage,
	}, nil
}

// pidfdSendSignal sends a signal to the process.
func (p *Process) pidfdSendSignal(s syscall.Signal) error {
	handle, status := p.handleTransientAcquire()
	switch status {
	case statusDone:
		return ErrProcessDone
	case statusReleased:
		return errProcessReleased
	}
	defer p.handleTransientRelease()

	return convertESRCH(unix.PidFDSendSignal(handle, s))
}

// pidfdWorks returns whether we can use pidfd on this system.
func pidfdWorks() bool {
	return checkPidfdOnce() == nil
}

// checkPidfdOnce is used to only check whether pidfd works once.
var checkPidfdOnce = sync.OnceValue(checkPidfd)

// checkPidfd checks whether all required pidfd-related syscalls work. This
// consists of pidfd_open and pidfd_send_signal syscalls, waitid syscall with
// idtype of P_PIDFD, and clone(CLONE_PIDFD).
//
// Reasons for non-working pidfd syscalls include an older kernel and an
// execution environment in which the above system calls are restricted by
// seccomp or a similar technology.
func checkPidfd() error {
	// In Android version < 12, pidfd-related system calls are not allowed
	// by seccomp and trigger the SIGSYS signal. See issue #69065.
	if runtime.GOOS == "android" {
		ignoreSIGSYS()
		defer restoreSIGSYS()
	}

	// Get a pidfd of the current process (opening of "/proc/self" won't
	// work for waitid).
	fd, err := unix.PidFDOpen(syscall.Getpid(), 0)
	if err != nil {
		return NewSyscallError("pidfd_open", err)
	}
	defer syscall.Close(int(fd))

	// Check waitid(P_PIDFD) works.
	err = ignoringEINTR(func() error {
		var info unix.SiginfoChild
		// We don't actually care about the info, but passing a nil pointer
		// makes valgrind complain because 0x0 is unaddressable.
		return unix.Waitid(unix.P_PIDFD, int(fd), &info, syscall.WEXITED, nil)
	})
	// Expect ECHILD from waitid since we're not our own parent.
	if err != syscall.ECHILD {
		return NewSyscallError("pidfd_wait", err)
	}

	// Check pidfd_send_signal works (should be able to send 0 to itself).
	if err := unix.PidFDSendSignal(fd, 0); err != nil {
		return NewSyscallError("pidfd_send_signal", err)
	}

	// Verify that clone(CLONE_PIDFD) works.
	//
	// This shouldn't be necessary since pidfd_open was added in Linux 5.3,
	// after CLONE_PIDFD in Linux 5.2, but some alternative Linux
	// implementations may not adhere to this ordering.
	if err := checkClonePidfd(); err != nil {
		return err
	}

	return nil
}

// Provided by syscall.
//
//go:linkname checkClonePidfd
func checkClonePidfd() error

// Provided by runtime.
//
//go:linkname ignoreSIGSYS
func ignoreSIGSYS()

//go:linkname restoreSIGSYS
func restoreSIGSYS()