generate

Documentation

Index

• Constants
• func BlockAccessToKernelFilesystems(privileged, pidModeIsHost bool, g *generate.Generator)
• func CompleteSpec(ctx context.Context, r *libpod.Runtime, s *specgen.SpecGenerator) ([]string, error)
• func CreateExitCommandArgs(storageConfig storage.StoreOptions, config *config.Config, ...) ([]string, error)
• func Device(d *configs.Device) spec.LinuxDevice
• func DevicesFromPath(g *generate.Generator, devicePath string) error
• func GetAvailableGids() (int64, error)
• func GetDefaultNamespaceMode(nsType string, cfg *config.Config, pod *libpod.Pod) (specgen.Namespace, error)
• func GetNamespaceOptions(ns []string) ([]libpod.PodCreateOption, error)
• func InitFSMounts(mounts []spec.Mount) error
• func IsValidDeviceMode(mode string) bool
• func MakeContainer(ctx context.Context, rt *libpod.Runtime, s *specgen.SpecGenerator) (*libpod.Container, error)
• func MakePod(p *specgen.PodSpecGenerator, rt *libpod.Runtime) (*libpod.Pod, error)
• func ParseDevice(device string) (string, string, string, error)
• func SpecGenToOCI(ctx context.Context, s *specgen.SpecGenerator, rt *libpod.Runtime, ...) (*spec.Spec, error)
• func SupercedeUserMounts(mounts []spec.Mount, configMount []spec.Mount) []spec.Mount

Constants

const (
	// TypeBind is the type for mounting host dir
	TypeBind = "bind"
	// TypeVolume is the type for named volumes
	TypeVolume = "volume"
	// TypeTmpfs is the type for mounting tmpfs
	TypeTmpfs = "tmpfs"
)

TODO unify this in one place - maybe libpod/define

Functions

func BlockAccessToKernelFilesystems

func BlockAccessToKernelFilesystems(privileged, pidModeIsHost bool, g *generate.Generator)

func CompleteSpec

func CompleteSpec(ctx context.Context, r *libpod.Runtime, s *specgen.SpecGenerator) ([]string, error)

Fill any missing parts of the spec generator (e.g. from the image). Returns a set of warnings or any fatal error that occurred.

func CreateExitCommandArgs

func CreateExitCommandArgs(storageConfig storage.StoreOptions, config *config.Config, syslog, rm, exec bool) ([]string, error)

func Device

func Device(d *configs.Device) spec.LinuxDevice

Device transforms a libcontainer configs.Device to a specs.LinuxDevice object.

func DevicesFromPath

func DevicesFromPath(g *generate.Generator, devicePath string) error

DevicesFromPath computes a list of devices

func GetAvailableGids

func GetAvailableGids() (int64, error)

func GetDefaultNamespaceMode

func GetDefaultNamespaceMode(nsType string, cfg *config.Config, pod *libpod.Pod) (specgen.Namespace, error)

Get the default namespace mode for any given namespace type.

func GetNamespaceOptions

func GetNamespaceOptions(ns []string) ([]libpod.PodCreateOption, error)

GetNamespaceOptions transforms a slice of kernel namespaces into a slice of pod create options. Currently, not all kernel namespaces are supported, and they will be returned in an error

func InitFSMounts

func InitFSMounts(mounts []spec.Mount) error

func IsValidDeviceMode

func IsValidDeviceMode(mode string) bool

IsValidDeviceMode checks if the mode for device is valid or not. IsValid mode is a composition of r (read), w (write), and m (mknod).

func MakeContainer

func MakeContainer(ctx context.Context, rt *libpod.Runtime, s *specgen.SpecGenerator) (*libpod.Container, error)

MakeContainer creates a container based on the SpecGenerator. Returns the created, container and any warnings resulting from creating the container, or an error.

func MakePod

func MakePod(p *specgen.PodSpecGenerator, rt *libpod.Runtime) (*libpod.Pod, error)

func ParseDevice

func ParseDevice(device string) (string, string, string, error)

ParseDevice parses device mapping string to a src, dest & permissions string

func SpecGenToOCI

func SpecGenToOCI(ctx context.Context, s *specgen.SpecGenerator, rt *libpod.Runtime, rtc *config.Config, newImage *image.Image, mounts []spec.Mount, pod *libpod.Pod, finalCmd []string) (*spec.Spec, error)

func SupercedeUserMounts

func SupercedeUserMounts(mounts []spec.Mount, configMount []spec.Mount) []spec.Mount

Supersede existing mounts in the spec with new, user-specified mounts. TODO: Should we unmount subtree mounts? E.g., if /tmp/ is mounted by one mount, and we already have /tmp/a and /tmp/b, should we remove the /tmp/a and /tmp/b mounts in favor of the more general /tmp?