vendor: update buildkit to v0.15.0-rc1

Signed-off-by: Tonis Tiigi <tonistiigi@gmail.com>
This commit is contained in:
Tonis Tiigi
2024-07-03 11:39:35 -07:00
parent 74374ea418
commit 50aa895477
82 changed files with 2379 additions and 1183 deletions

View File

@ -1,42 +1,44 @@
// Copyright 2015 Tim Heckman. All rights reserved.
// Copyright 2018-2024 The Gofrs. All rights reserved.
// Use of this source code is governed by the BSD 3-Clause
// license that can be found in the LICENSE file.
// +build !aix,!windows
//go:build darwin || dragonfly || freebsd || illumos || linux || netbsd || openbsd
package flock
import (
"errors"
"os"
"syscall"
"golang.org/x/sys/unix"
)
// Lock is a blocking call to try and take an exclusive file lock. It will wait
// until it is able to obtain the exclusive file lock. It's recommended that
// TryLock() be used over this function. This function may block the ability to
// query the current Locked() or RLocked() status due to a RW-mutex lock.
// Lock is a blocking call to try and take an exclusive file lock.
// It will wait until it is able to obtain the exclusive file lock.
// It's recommended that TryLock() be used over this function.
// This function may block the ability to query the current Locked() or RLocked() status due to a RW-mutex lock.
//
// If we are already exclusive-locked, this function short-circuits and returns
// immediately assuming it can take the mutex lock.
// If we are already exclusive-locked,
// this function short-circuits and returns immediately assuming it can take the mutex lock.
//
// If the *Flock has a shared lock (RLock), this may transparently replace the
// shared lock with an exclusive lock on some UNIX-like operating systems. Be
// careful when using exclusive locks in conjunction with shared locks
// (RLock()), because calling Unlock() may accidentally release the exclusive
// lock that was once a shared lock.
// If the *Flock has a shared lock (RLock),
// this may transparently replace the shared lock with an exclusive lock on some UNIX-like operating systems.
// Be careful when using exclusive locks in conjunction with shared locks (RLock()),
// because calling Unlock() may accidentally release the exclusive lock that was once a shared lock.
func (f *Flock) Lock() error {
return f.lock(&f.l, syscall.LOCK_EX)
return f.lock(&f.l, unix.LOCK_EX)
}
// RLock is a blocking call to try and take a shared file lock. It will wait
// until it is able to obtain the shared file lock. It's recommended that
// TryRLock() be used over this function. This function may block the ability to
// query the current Locked() or RLocked() status due to a RW-mutex lock.
// RLock is a blocking call to try and take a shared file lock.
// It will wait until it is able to obtain the shared file lock.
// It's recommended that TryRLock() be used over this function.
// This function may block the ability to query the current Locked() or RLocked() status due to a RW-mutex lock.
//
// If we are already shared-locked, this function short-circuits and returns
// immediately assuming it can take the mutex lock.
// If we are already shared-locked,
// this function short-circuits and returns immediately assuming it can take the mutex lock.
func (f *Flock) RLock() error {
return f.lock(&f.r, syscall.LOCK_SH)
return f.lock(&f.r, unix.LOCK_SH)
}
func (f *Flock) lock(locked *bool, flag int) error {
@ -51,10 +53,12 @@ func (f *Flock) lock(locked *bool, flag int) error {
if err := f.setFh(); err != nil {
return err
}
defer f.ensureFhState()
}
if err := syscall.Flock(int(f.fh.Fd()), flag); err != nil {
err := unix.Flock(int(f.fh.Fd()), flag)
if err != nil {
shouldRetry, reopenErr := f.reopenFDOnError(err)
if reopenErr != nil {
return reopenErr
@ -64,71 +68,74 @@ func (f *Flock) lock(locked *bool, flag int) error {
return err
}
if err = syscall.Flock(int(f.fh.Fd()), flag); err != nil {
err = unix.Flock(int(f.fh.Fd()), flag)
if err != nil {
return err
}
}
*locked = true
return nil
}
// Unlock is a function to unlock the file. This file takes a RW-mutex lock, so
// while it is running the Locked() and RLocked() functions will be blocked.
// Unlock is a function to unlock the file.
// This file takes a RW-mutex lock,
// so while it is running the Locked() and RLocked() functions will be blocked.
//
// This function short-circuits if we are unlocked already. If not, it calls
// syscall.LOCK_UN on the file and closes the file descriptor. It does not
// remove the file from disk. It's up to your application to do.
// This function short-circuits if we are unlocked already.
// If not, it calls unix.LOCK_UN on the file and closes the file descriptor.
// It does not remove the file from disk. It's up to your application to do.
//
// Please note, if your shared lock became an exclusive lock this may
// unintentionally drop the exclusive lock if called by the consumer that
// believes they have a shared lock. Please see Lock() for more details.
// Please note,
// if your shared lock became an exclusive lock,
// this may unintentionally drop the exclusive lock if called by the consumer that believes they have a shared lock.
// Please see Lock() for more details.
func (f *Flock) Unlock() error {
f.m.Lock()
defer f.m.Unlock()
// if we aren't locked or if the lockfile instance is nil
// just return a nil error because we are unlocked
// If we aren't locked or if the lockfile instance is nil
// just return a nil error because we are unlocked.
if (!f.l && !f.r) || f.fh == nil {
return nil
}
// mark the file as unlocked
if err := syscall.Flock(int(f.fh.Fd()), syscall.LOCK_UN); err != nil {
// Mark the file as unlocked.
err := unix.Flock(int(f.fh.Fd()), unix.LOCK_UN)
if err != nil {
return err
}
f.fh.Close()
f.l = false
f.r = false
f.fh = nil
f.reset()
return nil
}
// TryLock is the preferred function for taking an exclusive file lock. This
// function takes an RW-mutex lock before it tries to lock the file, so there is
// the possibility that this function may block for a short time if another
// goroutine is trying to take any action.
// TryLock is the preferred function for taking an exclusive file lock.
// This function takes an RW-mutex lock before it tries to lock the file,
// so there is the possibility that this function may block for a short time
// if another goroutine is trying to take any action.
//
// The actual file lock is non-blocking. If we are unable to get the exclusive
// file lock, the function will return false instead of waiting for the lock. If
// we get the lock, we also set the *Flock instance as being exclusive-locked.
// The actual file lock is non-blocking.
// If we are unable to get the exclusive file lock,
// the function will return false instead of waiting for the lock.
// If we get the lock, we also set the *Flock instance as being exclusive-locked.
func (f *Flock) TryLock() (bool, error) {
return f.try(&f.l, syscall.LOCK_EX)
return f.try(&f.l, unix.LOCK_EX)
}
// TryRLock is the preferred function for taking a shared file lock. This
// function takes an RW-mutex lock before it tries to lock the file, so there is
// the possibility that this function may block for a short time if another
// goroutine is trying to take any action.
// TryRLock is the preferred function for taking a shared file lock.
// This function takes an RW-mutex lock before it tries to lock the file,
// so there is the possibility that this function may block for a short time
// if another goroutine is trying to take any action.
//
// The actual file lock is non-blocking. If we are unable to get the shared file
// lock, the function will return false instead of waiting for the lock. If we
// get the lock, we also set the *Flock instance as being share-locked.
// The actual file lock is non-blocking.
// If we are unable to get the shared file lock,
// the function will return false instead of waiting for the lock.
// If we get the lock, we also set the *Flock instance as being share-locked.
func (f *Flock) TryRLock() (bool, error) {
return f.try(&f.r, syscall.LOCK_SH)
return f.try(&f.r, unix.LOCK_SH)
}
func (f *Flock) try(locked *bool, flag int) (bool, error) {
@ -143,22 +150,25 @@ func (f *Flock) try(locked *bool, flag int) (bool, error) {
if err := f.setFh(); err != nil {
return false, err
}
defer f.ensureFhState()
}
var retried bool
retry:
err := syscall.Flock(int(f.fh.Fd()), flag|syscall.LOCK_NB)
err := unix.Flock(int(f.fh.Fd()), flag|unix.LOCK_NB)
switch err {
case syscall.EWOULDBLOCK:
switch {
case errors.Is(err, unix.EWOULDBLOCK):
return false, nil
case nil:
case err == nil:
*locked = true
return true, nil
}
if !retried {
if shouldRetry, reopenErr := f.reopenFDOnError(err); reopenErr != nil {
shouldRetry, reopenErr := f.reopenFDOnError(err)
if reopenErr != nil {
return false, reopenErr
} else if shouldRetry {
retried = true
@ -169,29 +179,34 @@ retry:
return false, err
}
// reopenFDOnError determines whether we should reopen the file handle
// in readwrite mode and try again. This comes from util-linux/sys-utils/flock.c:
// Since Linux 3.4 (commit 55725513)
// Probably NFSv4 where flock() is emulated by fcntl().
// reopenFDOnError determines whether we should reopen the file handle in readwrite mode and try again.
// This comes from `util-linux/sys-utils/flock.c`:
// > Since Linux 3.4 (commit 55725513)
// > Probably NFSv4 where flock() is emulated by fcntl().
func (f *Flock) reopenFDOnError(err error) (bool, error) {
if err != syscall.EIO && err != syscall.EBADF {
if !errors.Is(err, unix.EIO) && !errors.Is(err, unix.EBADF) {
return false, nil
}
if st, err := f.fh.Stat(); err == nil {
// if the file is able to be read and written
if st.Mode()&0600 == 0600 {
f.fh.Close()
f.fh = nil
// reopen in read-write mode and set the filehandle
fh, err := os.OpenFile(f.path, os.O_CREATE|os.O_RDWR, os.FileMode(0600))
if err != nil {
return false, err
}
f.fh = fh
return true, nil
}
st, err := f.fh.Stat()
if err != nil {
return false, nil
}
return false, nil
if st.Mode()&f.perm != f.perm {
return false, nil
}
_ = f.fh.Close()
f.fh = nil
// reopen in read-write mode and set the file handle
fh, err := os.OpenFile(f.path, f.flag, f.perm)
if err != nil {
return false, err
}
f.fh = fh
return true, nil
}