Struct Opts 

pub(crate) struct Opts {

    pub(crate) verbose: u8,
    pub(crate) autopilot: bool,
    pub(crate) autopower: bool,
    pub(crate) performance: bool,
    pub(crate) powersave: bool,
    pub(crate) balanced: bool,
    pub(crate) slice_max_us: u64,
    pub(crate) slice_min_us: u64,
    pub(crate) mig_delta_pct: u8,
    pub(crate) pinned_slice_us: Option<u64>,
    pub(crate) preempt_shift: u8,
    pub(crate) cpu_pref_order: String,
    pub(crate) no_use_em: bool,
    pub(crate) no_futex_boost: bool,
    pub(crate) no_preemption: bool,
    pub(crate) no_wake_sync: bool,
    pub(crate) no_slice_boost: bool,
    pub(crate) per_cpu_dsq: bool,
    pub(crate) enable_cpu_bw: bool,
    pub(crate) no_core_compaction: bool,
    pub(crate) no_freq_scaling: bool,
    pub(crate) stats: Option<f64>,
    pub(crate) monitor: Option<f64>,
    pub(crate) monitor_sched_samples: Option<u64>,
    pub(crate) log_level: String,
    pub(crate) version: bool,
    pub(crate) run_id: Option<u64>,
    pub(crate) help_stats: bool,
    pub libbpf: LibbpfOpts,
    pub(crate) topology: Option<TopologyArgs>,
}

scx_lavd: Latency-criticality Aware Virtual Deadline (LAVD) scheduler

The rust part is minimal. It processes command line options and logs out scheduling statistics. The BPF part makes all the scheduling decisions. See the more detailed overview of the LAVD design at main.bpf.c.

Fields

verbose: u8

Depricated, noop, use RUST_LOG or –log-level instead.

autopilot: bool

Automatically decide the scheduler’s power mode (performance vs. powersave vs. balanced), CPU preference order, etc, based on system load. The options affecting the power mode and the use of core compaction (–autopower, –performance, –powersave, –balanced, –no-core-compaction) cannot be used with this option. When no option is specified, this is a default mode.

autopower: bool

Automatically decide the scheduler’s power mode (performance vs. powersave vs. balanced) based on the system’s active power profile. The scheduler’s power mode decides the CPU preference order and the use of core compaction, so the options affecting these (–autopilot, –performance, –powersave, –balanced, –no-core-compaction) cannot be used with this option.

performance: bool

Run the scheduler in performance mode to get maximum performance. This option cannot be used with other conflicting options (–autopilot, –autopower, –balanced, –powersave, –no-core-compaction) affecting the use of core compaction.

powersave: bool

Run the scheduler in powersave mode to minimize powr consumption. This option cannot be used with other conflicting options (–autopilot, –autopower, –performance, –balanced, –no-core-compaction) affecting the use of core compaction.

balanced: bool

Run the scheduler in balanced mode aiming for sweetspot between power and performance. This option cannot be used with other conflicting options (–autopilot, –autopower, –performance, –powersave, –no-core-compaction) affecting the use of core compaction.

slice_max_us: u64

Maximum scheduling slice duration in microseconds.

slice_min_us: u64

Minimum scheduling slice duration in microseconds.

mig_delta_pct: u8

Migration delta threshold percentage (0-100). When set to a non-zero value, uses average utilization for threshold calculation instead of current utilization, and the threshold is calculated as: avg_load * (mig-delta-pct / 100). Additionally, disables force task stealing in the consume path, relying only on the is_stealer/is_stealee thresholds for more predictable load balancing. Default is 0 (disabled, uses dynamic threshold based on load with both probabilistic and force task stealing enabled). This is an experimental feature.

pinned_slice_us: Option<u64>

Slice duration in microseconds to use for all tasks when pinned tasks are running on a CPU. Must be between slice-min-us and slice-max-us. When this option is enabled, pinned tasks are always enqueued to per-CPU DSQs and the dispatch logic compares vtimes across all DSQs to select the lowest vtime task. This helps improve responsiveness for pinned tasks. By default, this option is on with a default value of 5000 (5 msec). To turn off the option, explicitly set the value to 0.

preempt_shift: u8

Limit the ratio of preemption to the roughly top P% of latency-critical tasks. When N is given as an argument, P is 0.5^N * 100. The default value is 6, which limits the preemption for the top 1.56% of latency-critical tasks.

cpu_pref_order: String

List of CPUs in preferred order (e.g., “0-3,7,6,5,4”). The scheduler uses the CPU preference mode only when the core compaction is enabled (i.e., balanced or powersave mode is specified as an option or chosen in the autopilot or autopower mode). When “–cpu-pref-order” is given, it implies “–no-use-em”.

no_use_em: bool

Do not use the energy model in making CPU preference order decisions.

no_futex_boost: bool

Do not boost futex holders.

no_preemption: bool

Disable preemption.

no_wake_sync: bool

Disable an optimization for synchronous wake-up.

no_slice_boost: bool

Disable dynamic slice boost for long-running tasks.

per_cpu_dsq: bool

Enables DSQs per CPU, this enables task queuing and dispatching from CPU specific DSQs. This generally increases L1/L2 cache locality for tasks and lowers lock contention compared to shared DSQs, but at the cost of higher load balancing complexity. This is a highly experimental feature.

enable_cpu_bw: bool

Enable CPU bandwidth control using cpu.max in cgroup v2. This is a highly experimental feature.

no_core_compaction: bool

Disable core compaction so the scheduler uses all the online CPUs. The core compaction attempts to minimize the number of actively used CPUs for unaffinitized tasks, respecting the CPU preference order. Normally, the core compaction is enabled by the power mode (i.e., balanced or powersave mode is specified as an option or chosen in the autopilot or autopower mode). This option cannot be used with the other options that control the core compaction (–autopilot, –autopower, –performance, –balanced, –powersave).

no_freq_scaling: bool

Disable controlling the CPU frequency.

stats: Option<f64>

Enable stats monitoring with the specified interval.

monitor: Option<f64>

Run in stats monitoring mode with the specified interval. Scheduler is not launched.

monitor_sched_samples: Option<u64>

Run in monitoring mode. Show the specified number of scheduling samples every second.

log_level: String

Specify the logging level. Accepts rust’s envfilter syntax for modular logging: https://docs.rs/tracing-subscriber/latest/tracing_subscriber/filter/struct.EnvFilter.html#example-syntax. Examples: [“info”, “warn,tokio=info”]

version: bool

Print scheduler version and exit.

run_id: Option<u64>

Optional run ID for tracking scheduler instances.

help_stats: bool

Show descriptions for statistics.

libbpf: LibbpfOpts

topology: Option<TopologyArgs>

Topology configuration options

Implementations

impl Opts

pub(crate) fn can_autopilot(&self) -> bool

pub(crate) fn can_autopower(&self) -> bool

pub(crate) fn can_performance(&self) -> bool

pub(crate) fn can_balanced(&self) -> bool

pub(crate) fn can_powersave(&self) -> bool

pub(crate) fn proc(&mut self) -> Option<&mut Self>

pub(crate) fn preempt_shift_range(s: &str) -> Result<u8, String>

pub(crate) fn mig_delta_pct_range(s: &str) -> Result<u8, String>

Trait Implementations

impl Args for Opts

fn group_id() -> Option<Id>

Report the [ArgGroup::id][crate::ArgGroup::id] for this set of arguments

fn augment_args<'b>(__clap_app: Command) -> Command

Append to [Command] so it can instantiate Self via [FromArgMatches::from_arg_matches_mut]

fn augment_args_for_update<'b>(__clap_app: Command) -> Command

Append to [Command] so it can instantiate self via [FromArgMatches::update_from_arg_matches_mut]

impl CommandFactory for Opts

fn command<'b>() -> Command

Build a [Command] that can instantiate Self.

fn command_for_update<'b>() -> Command

Build a [Command] that can update self.

impl Debug for Opts

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl FromArgMatches for Opts

fn from_arg_matches(__clap_arg_matches: &ArgMatches) -> Result<Self, Error>

Instantiate Self from [ArgMatches], parsing the arguments as needed.

fn from_arg_matches_mut( __clap_arg_matches: &mut ArgMatches, ) -> Result<Self, Error>

Instantiate Self from [ArgMatches], parsing the arguments as needed.

fn update_from_arg_matches( &mut self, __clap_arg_matches: &ArgMatches, ) -> Result<(), Error>

Assign values from ArgMatches to self.

fn update_from_arg_matches_mut( &mut self, __clap_arg_matches: &mut ArgMatches, ) -> Result<(), Error>

Assign values from ArgMatches to self.

impl Parser for Opts

fn parse() -> Self

Parse from std::env::args_os(), [exit][Error::exit] on error.

fn try_parse() -> Result<Self, Error>

Parse from std::env::args_os(), return Err on error.

fn parse_from<I, T>(itr: I) -> Self

where I: IntoIterator<Item = T>, T: Into<OsString> + Clone,

Parse from iterator, [exit][Error::exit] on error.

fn try_parse_from<I, T>(itr: I) -> Result<Self, Error>

where I: IntoIterator<Item = T>, T: Into<OsString> + Clone,

Parse from iterator, return Err on error.

fn update_from<I, T>(&mut self, itr: I)

where I: IntoIterator<Item = T>, T: Into<OsString> + Clone,

Update from iterator, [exit][Error::exit] on error.

fn try_update_from<I, T>(&mut self, itr: I) -> Result<(), Error>

where I: IntoIterator<Item = T>, T: Into<OsString> + Clone,

Update from iterator, return Err on error.