Module perf 

Re-exports

use libc::pid_t; 🔒

use std::os::raw::c_int; 🔒

use std::os::raw::c_ulong; 🔒

use anyhow::Context as _; 🔒

use anyhow::Result; 🔒

use libbpf_rs::MapCore; 🔒

use libbpf_rs::MapFlags; 🔒

Modules

bindings

ioctls

Structs

PerfEventSpec

Perf event specification: either hex (0xN) or a symbolic name (e.g. cache-misses). event_id is the opaque id written to BPF rodata (it must match between install and read); type_/config drive perf_event_open.

Constants

PERF_MAP_STRIDE

Must match lib/pmu.bpf.c SCX_PMU_STRIDE for the perf_events map key layout.

Functions

parse_hardware_event 🔒

parse_hw_cache_event 🔒

parse_perf_event

Parse a perf event value: hex (0xN) or a symbolic name (e.g. cache-misses, LLC-load-misses, page-faults). Intended for use as a clap value_parser.

parse_software_event 🔒

perf_event_open^⚠

The perf_event_open system call.

setup_perf_events

Open a perf event on @cpu and register its fd in the given BPF PMU map (the scheduler’s scx_pmu_map) at the slot for (@cpu, @counter_idx). counter_idx 0 is the migration event, 1 the sticky event, matching the PMU library install order.