const (
	CgroupNamePrefix = "name="
)

const (
	CgroupProcesses = "cgroup.procs"
)

var (
	// ErrDevicesUnsupported is an error returned when a cgroup manager
	// is not configured to set device rules.
	ErrDevicesUnsupported = errors.New("cgroup manager is not configured to set device rules")

	// ErrRootless is returned by [Manager.Apply] when there is an error
	// creating cgroup directory, and cgroup.Rootless is set. In general,
	// this error is to be ignored.
	ErrRootless = errors.New("cgroup manager can not access cgroup (rootless container)")

	// DevicesSetV1 and DevicesSetV2 are functions to set devices for
	// cgroup v1 and v2, respectively. Unless
	// [github.com/opencontainers/cgroups/devices]
	// package is imported, it is set to nil, so cgroup managers can't
	// manage devices.
	DevicesSetV1 func(path string, r *Resources) error
	DevicesSetV2 func(path string, r *Resources) error
)

var (
	ErrV1NoUnified = errors.New("invalid configuration: cannot use unified on cgroup v1")
)

var (
	// TestMode is set to true by unit tests that need "fake" cgroupfs.
	TestMode bool
)

type BlkioStatEntry struct {
	Major uint64 `json:"major,omitempty"`
	Minor uint64 `json:"minor,omitempty"`
	Op    string `json:"op,omitempty"`
	Value uint64 `json:"value,omitempty"`
}

type BlkioStats struct {
	// number of bytes transferred to and from the block device
	IoServiceBytesRecursive []BlkioStatEntry `json:"io_service_bytes_recursive,omitempty"`
	IoServicedRecursive     []BlkioStatEntry `json:"io_serviced_recursive,omitempty"`
	IoQueuedRecursive       []BlkioStatEntry `json:"io_queue_recursive,omitempty"`
	IoServiceTimeRecursive  []BlkioStatEntry `json:"io_service_time_recursive,omitempty"`
	IoWaitTimeRecursive     []BlkioStatEntry `json:"io_wait_time_recursive,omitempty"`
	IoMergedRecursive       []BlkioStatEntry `json:"io_merged_recursive,omitempty"`
	IoTimeRecursive         []BlkioStatEntry `json:"io_time_recursive,omitempty"`
	SectorsRecursive        []BlkioStatEntry `json:"sectors_recursive,omitempty"`
	PSI                     *PSIStats        `json:"psi,omitempty"`
}

type BlockIODevice struct {
	// Major is the device's major number
	Major int64 `json:"major"`
	// Minor is the device's minor number
	Minor int64 `json:"minor"`
}

type CPUSetStats struct {
	// List of the physical numbers of the CPUs on which processes
	// in that cpuset are allowed to execute
	CPUs []uint16 `json:"cpus,omitempty"`
	// cpu_exclusive flag
	CPUExclusive uint64 `json:"cpu_exclusive"`
	// List of memory nodes on which processes in that cpuset
	// are allowed to allocate memory
	Mems []uint16 `json:"mems,omitempty"`
	// mem_hardwall flag
	MemHardwall uint64 `json:"mem_hardwall"`
	// mem_exclusive flag
	MemExclusive uint64 `json:"mem_exclusive"`
	// memory_migrate flag
	MemoryMigrate uint64 `json:"memory_migrate"`
	// memory_spread page flag
	MemorySpreadPage uint64 `json:"memory_spread_page"`
	// memory_spread slab flag
	MemorySpreadSlab uint64 `json:"memory_spread_slab"`
	// memory_pressure
	MemoryPressure uint64 `json:"memory_pressure"`
	// sched_load balance flag
	SchedLoadBalance uint64 `json:"sched_load_balance"`
	// sched_relax_domain_level
	SchedRelaxDomainLevel int64 `json:"sched_relax_domain_level"`
}

type Cgroup struct {
	// Name specifies the name of the cgroup
	Name string `json:"name,omitempty"`

	// Parent specifies the name of parent of cgroup or slice
	Parent string `json:"parent,omitempty"`

	// Path specifies the path to cgroups that are created and/or joined by the container.
	// The path is assumed to be relative to the host system cgroup mountpoint.
	Path string `json:"path"`

	// ScopePrefix describes prefix for the scope name
	ScopePrefix string `json:"scope_prefix"`

	// Resources contains various cgroups settings to apply
	*Resources

	// Systemd tells if systemd should be used to manage cgroups.
	Systemd bool

	// SystemdProps are any additional properties for systemd,
	// derived from org.systemd.property.xxx annotations.
	// Ignored unless systemd is used for managing cgroups.
	SystemdProps []systemdDbus.Property `json:"-"`

	// Rootless tells if rootless cgroups should be used.
	Rootless bool

	// The host UID that should own the cgroup, or nil to accept
	// the default ownership.  This should only be set when the
	// cgroupfs is to be mounted read/write.
	// Not all cgroup manager implementations support changing
	// the ownership.
	OwnerUID *int `json:"owner_uid,omitempty"`
}

type CpuStats struct {
	CpuUsage       CpuUsage       `json:"cpu_usage,omitempty"`
	ThrottlingData ThrottlingData `json:"throttling_data,omitempty"`
	PSI            *PSIStats      `json:"psi,omitempty"`
}

type CpuUsage struct {
	// Total CPU time consumed.
	// Units: nanoseconds.
	TotalUsage uint64 `json:"total_usage,omitempty"`
	// Total CPU time consumed per core.
	// Units: nanoseconds.
	PercpuUsage []uint64 `json:"percpu_usage,omitempty"`
	// CPU time consumed per core in kernel mode
	// Units: nanoseconds.
	PercpuUsageInKernelmode []uint64 `json:"percpu_usage_in_kernelmode"`
	// CPU time consumed per core in user mode
	// Units: nanoseconds.
	PercpuUsageInUsermode []uint64 `json:"percpu_usage_in_usermode"`
	// Time spent by tasks of the cgroup in kernel mode.
	// Units: nanoseconds.
	UsageInKernelmode uint64 `json:"usage_in_kernelmode"`
	// Time spent by tasks of the cgroup in user mode.
	// Units: nanoseconds.
	UsageInUsermode uint64 `json:"usage_in_usermode"`
}

type FreezerState string

const (
	Undefined FreezerState = ""
	Frozen    FreezerState = "FROZEN"
	Thawed    FreezerState = "THAWED"
)

type HugepageLimit struct {
	// which type of hugepage to limit.
	Pagesize string `json:"page_size"`

	// usage limit for hugepage.
	Limit uint64 `json:"limit"`
}

type HugetlbStats struct {
	// current res_counter usage for hugetlb
	Usage uint64 `json:"usage,omitempty"`
	// maximum usage ever recorded.
	MaxUsage uint64 `json:"max_usage,omitempty"`
	// number of times hugetlb usage allocation failure.
	Failcnt uint64 `json:"failcnt"`
}

type IfPrioMap struct {
	Interface string `json:"interface"`
	Priority  int64  `json:"priority"`
}

type LinuxRdma struct {
	// Maximum number of HCA handles that can be opened. Default is "no limit".
	HcaHandles *uint32 `json:"hca_handles,omitempty"`
	// Maximum number of HCA objects that can be created. Default is "no limit".
	HcaObjects *uint32 `json:"hca_objects,omitempty"`
}

type Manager interface {
	// Apply creates a cgroup, if not yet created, and adds a process
	// with the specified pid into that cgroup.  A special value of -1
	// can be used to merely create a cgroup.
	Apply(pid int) error

	// GetPids returns the PIDs of all processes inside the cgroup.
	GetPids() ([]int, error)

	// GetAllPids returns the PIDs of all processes inside the cgroup
	// any all its sub-cgroups.
	GetAllPids() ([]int, error)

	// GetStats returns cgroups statistics.
	GetStats() (*Stats, error)

	// Freeze sets the freezer cgroup to the specified state.
	Freeze(state FreezerState) error

	// Destroy removes cgroup.
	Destroy() error

	// Path returns a cgroup path to the specified controller/subsystem.
	// For cgroupv2, the argument is unused and can be empty.
	Path(string) string

	// Set sets cgroup resources parameters/limits. If the argument is nil,
	// the resources specified during Manager creation (or the previous call
	// to Set) are used.
	Set(r *Resources) error

	// GetPaths returns cgroup path(s) to save in a state file in order to
	// restore later.
	//
	// For cgroup v1, a key is cgroup subsystem name, and the value is the
	// path to the cgroup for this subsystem.
	//
	// For cgroup v2 unified hierarchy, a key is "", and the value is the
	// unified path.
	GetPaths() map[string]string

	// GetCgroups returns the cgroup data as configured.
	GetCgroups() (*Cgroup, error)

	// GetFreezerState retrieves the current FreezerState of the cgroup.
	GetFreezerState() (FreezerState, error)

	// Exists returns whether the cgroup path exists or not.
	Exists() bool

	// OOMKillCount reports OOM kill count for the cgroup.
	OOMKillCount() (uint64, error)
}

type MemoryData struct {
	Usage    uint64 `json:"usage,omitempty"`
	MaxUsage uint64 `json:"max_usage,omitempty"`
	Failcnt  uint64 `json:"failcnt"`
	Limit    uint64 `json:"limit"`
}

type MemoryStats struct {
	// memory used for cache
	Cache uint64 `json:"cache,omitempty"`
	// usage of memory
	Usage MemoryData `json:"usage,omitempty"`
	// usage of memory + swap
	SwapUsage MemoryData `json:"swap_usage,omitempty"`
	// usage of swap only
	SwapOnlyUsage MemoryData `json:"swap_only_usage,omitempty"`
	// usage of kernel memory
	KernelUsage MemoryData `json:"kernel_usage,omitempty"`
	// usage of kernel TCP memory
	KernelTCPUsage MemoryData `json:"kernel_tcp_usage,omitempty"`
	// usage of memory pages by NUMA node
	// see chapter 5.6 of memory controller documentation
	PageUsageByNUMA PageUsageByNUMA `json:"page_usage_by_numa,omitempty"`
	// if true, memory usage is accounted for throughout a hierarchy of cgroups.
	UseHierarchy bool `json:"use_hierarchy"`

	Stats map[string]uint64 `json:"stats,omitempty"`
	PSI   *PSIStats         `json:"psi,omitempty"`
}

type MiscStats struct {
	// current resource usage for a key in misc
	Usage uint64 `json:"usage,omitempty"`
	// number of times the resource usage was about to go over the max boundary
	Events uint64 `json:"events,omitempty"`
}

type Mount struct {
	Mountpoint string
	Root       string
	Subsystems []string
}

type NotFoundError struct {
	Subsystem string
}

type PSIData struct {
	Avg10  float64 `json:"avg10"`
	Avg60  float64 `json:"avg60"`
	Avg300 float64 `json:"avg300"`
	Total  uint64  `json:"total"`
}

type PSIStats struct {
	Some PSIData `json:"some,omitempty"`
	Full PSIData `json:"full,omitempty"`
}

type PageStats struct {
	Total uint64           `json:"total,omitempty"`
	Nodes map[uint8]uint64 `json:"nodes,omitempty"`
}

type PageUsageByNUMA struct {
	// Embedding is used as types can't be recursive.
	PageUsageByNUMAInner
	Hierarchical PageUsageByNUMAInner `json:"hierarchical,omitempty"`
}

type PageUsageByNUMAInner struct {
	Total       PageStats `json:"total,omitempty"`
	File        PageStats `json:"file,omitempty"`
	Anon        PageStats `json:"anon,omitempty"`
	Unevictable PageStats `json:"unevictable,omitempty"`
}

type PidsStats struct {
	// number of pids in the cgroup
	Current uint64 `json:"current,omitempty"`
	// active pids hard limit
	Limit uint64 `json:"limit,omitempty"`
}

type RdmaEntry struct {
	Device     string `json:"device,omitempty"`
	HcaHandles uint32 `json:"hca_handles,omitempty"`
	HcaObjects uint32 `json:"hca_objects,omitempty"`
}

type RdmaStats struct {
	RdmaLimit   []RdmaEntry `json:"rdma_limit,omitempty"`
	RdmaCurrent []RdmaEntry `json:"rdma_current,omitempty"`
}

type Resources struct {
	// Devices is the set of access rules for devices in the container.
	Devices []*devices.Rule `json:"devices"`

	// Memory limit (in bytes)
	Memory int64 `json:"memory"`

	// Memory reservation or soft_limit (in bytes)
	MemoryReservation int64 `json:"memory_reservation"`

	// Total memory usage (memory + swap); set `-1` to enable unlimited swap
	MemorySwap int64 `json:"memory_swap"`

	// CPU shares (relative weight vs. other containers)
	CpuShares uint64 `json:"cpu_shares"`

	// CPU hardcap limit (in usecs). Allowed cpu time in a given period.
	CpuQuota int64 `json:"cpu_quota"`

	// CPU hardcap burst limit (in usecs). Allowed accumulated cpu time additionally for burst in a given period.
	CpuBurst *uint64 `json:"cpu_burst"` //nolint:revive

	// CPU period to be used for hardcapping (in usecs). 0 to use system default.
	CpuPeriod uint64 `json:"cpu_period"`

	// How many time CPU will use in realtime scheduling (in usecs).
	CpuRtRuntime int64 `json:"cpu_rt_quota"`

	// CPU period to be used for realtime scheduling (in usecs).
	CpuRtPeriod uint64 `json:"cpu_rt_period"`

	// CPU to use
	CpusetCpus string `json:"cpuset_cpus"`

	// MEM to use
	CpusetMems string `json:"cpuset_mems"`

	// cgroup SCHED_IDLE
	CPUIdle *int64 `json:"cpu_idle,omitempty"`

	// Process limit; set <= `0' to disable limit.
	PidsLimit int64 `json:"pids_limit"`

	// Specifies per cgroup weight, range is from 10 to 1000.
	BlkioWeight uint16 `json:"blkio_weight"`

	// Specifies tasks' weight in the given cgroup while competing with the cgroup's child cgroups, range is from 10 to 1000, cfq scheduler only
	BlkioLeafWeight uint16 `json:"blkio_leaf_weight"`

	// Weight per cgroup per device, can override BlkioWeight.
	BlkioWeightDevice []*WeightDevice `json:"blkio_weight_device"`

	// IO read rate limit per cgroup per device, bytes per second.
	BlkioThrottleReadBpsDevice []*ThrottleDevice `json:"blkio_throttle_read_bps_device"`

	// IO write rate limit per cgroup per device, bytes per second.
	BlkioThrottleWriteBpsDevice []*ThrottleDevice `json:"blkio_throttle_write_bps_device"`

	// IO read rate limit per cgroup per device, IO per second.
	BlkioThrottleReadIOPSDevice []*ThrottleDevice `json:"blkio_throttle_read_iops_device"`

	// IO write rate limit per cgroup per device, IO per second.
	BlkioThrottleWriteIOPSDevice []*ThrottleDevice `json:"blkio_throttle_write_iops_device"`

	// set the freeze value for the process
	Freezer FreezerState `json:"freezer"`

	// Hugetlb limit (in bytes)
	HugetlbLimit []*HugepageLimit `json:"hugetlb_limit"`

	// Whether to disable OOM Killer
	OomKillDisable bool `json:"oom_kill_disable"`

	// Tuning swappiness behaviour per cgroup
	MemorySwappiness *uint64 `json:"memory_swappiness"`

	// Set priority of network traffic for container
	NetPrioIfpriomap []*IfPrioMap `json:"net_prio_ifpriomap"`

	// Set class identifier for container's network packets
	NetClsClassid uint32 `json:"net_cls_classid_u"`

	// Rdma resource restriction configuration
	Rdma map[string]LinuxRdma `json:"rdma"`

	// CpuWeight sets a proportional bandwidth limit.
	CpuWeight uint64 `json:"cpu_weight"`

	// Unified is cgroupv2-only key-value map.
	Unified map[string]string `json:"unified"`

	// SkipDevices allows to skip configuring device permissions.
	// Used by e.g. kubelet while creating a parent cgroup (kubepods)
	// common for many containers, and by runc update.
	//
	// NOTE it is impossible to start a container which has this flag set.
	SkipDevices bool `json:"-"`

	// SkipFreezeOnSet is a flag for cgroup manager to skip the cgroup
	// freeze when setting resources. Only applicable to systemd legacy
	// (i.e. cgroup v1) manager (which uses freeze by default to avoid
	// spurious permission errors caused by systemd inability to update
	// device rules in a non-disruptive manner).
	//
	// If not set, a few methods (such as looking into cgroup's
	// devices.list and querying the systemd unit properties) are used
	// during Set() to figure out whether the freeze is required. Those
	// methods may be relatively slow, thus this flag.
	SkipFreezeOnSet bool `json:"-"`

	// MemoryCheckBeforeUpdate is a flag for cgroup v2 managers to check
	// if the new memory limits (Memory and MemorySwap) being set are lower
	// than the current memory usage, and reject if so.
	MemoryCheckBeforeUpdate bool `json:"memory_check_before_update"`
}

type Stats struct {
	CpuStats    CpuStats    `json:"cpu_stats,omitempty"`
	CPUSetStats CPUSetStats `json:"cpuset_stats,omitempty"`
	MemoryStats MemoryStats `json:"memory_stats,omitempty"`
	PidsStats   PidsStats   `json:"pids_stats,omitempty"`
	BlkioStats  BlkioStats  `json:"blkio_stats,omitempty"`
	// the map is in the format "size of hugepage: stats of the hugepage"
	HugetlbStats map[string]HugetlbStats `json:"hugetlb_stats,omitempty"`
	RdmaStats    RdmaStats               `json:"rdma_stats,omitempty"`
	// the map is in the format "misc resource name: stats of the key"
	MiscStats map[string]MiscStats `json:"misc_stats,omitempty"`
}

type ThrottleDevice struct {
	BlockIODevice
	// Rate is the IO rate limit per cgroup per device
	Rate uint64 `json:"rate"`
}

type ThrottlingData struct {
	// Number of periods with throttling active
	Periods uint64 `json:"periods,omitempty"`
	// Number of periods when the container hit its throttling limit.
	ThrottledPeriods uint64 `json:"throttled_periods,omitempty"`
	// Aggregate time the container was throttled for in nanoseconds.
	ThrottledTime uint64 `json:"throttled_time,omitempty"`
}

type WeightDevice struct {
	BlockIODevice
	// Weight is the bandwidth rate for the device, range is from 10 to 1000
	Weight uint16 `json:"weight"`
	// LeafWeight is the bandwidth rate for the device while competing with the cgroup's child cgroups, range is from 10 to 1000, cfq scheduler only
	LeafWeight uint16 `json:"leafWeight"`
}