ktracepoint

A Rust crate for implementing tracepoints in kernel. This crate provides a flexible and efficient way to add tracing capabilities to your kernel, similar to Linux kernel's tracepoint mechanism.

Features

• Define and manage kernel tracepoints with custom event data

• Hierarchical organization of tracepoints through subsystems

• Thread-safe implementation using mutexes

• Configurable tracepoint enable/disable functionality

• Customizable trace record formatting

• Support for tracing pipe for collecting trace records

• No-std compatible for kernel space usage

Usage

Basic Example

use spin::Mutex;
use ktracepoint::{define_event_trace, KernelTraceOps};
// Define kernel operations
pub static TRACE_RAW_PIPE: Mutex<ktracepoint::TracePipeRaw> =
    Mutex::new(ktracepoint::TracePipeRaw::new(1024));

pub struct Kops;

impl KernelTraceOps for Kops {
    ... // Implement required methods
}

// Define tracepoint
define_event_trace!(
    TEST,
    TP_lock(Mutex<()>),
    TP_kops(Kops),
    TP_system(tracepoint_test),
    TP_PROTO(a: u32, b: u32),
    TP_STRUCT__entry{
        a: u32,
        b: u32
    },
    TP_fast_assign{
        a:a,
        b:b
    },
    TP_ident(__entry),
    TP_printk(format_args!("Hello from tracepoint! a={}, b={}", __entry.a, __entry.b))
);
// Use the tracepoint in kernel code
 pub fn test_trace(a: u32, b: u32) {
    // call the tracepoint
    trace_TEST(a, &x);
    println!("Tracepoint TEST called with a={}, b={}", a, b);
}

See example in examples/usage.rs for a complete example.

Managing Tracepoints

// Initialize the tracing system in kernel module
let manager = global_init_events::<Mutex<()>>().unwrap();

// Enable/disable tracepoints
let subsystem = manager.get_subsystem("my_subsystem").unwrap();
let tracepoint_info = subsystem.get_event("my_event").unwrap();
tracepoint_info.enable_file().write(true); // Enable
tracepoint_info.enable_file().write(false); // Disable


// other operations
let tracepoint_map = manager.tracepoint_map();
// Iterate over all tracepoints
for (name, tracepoint) in tracepoint_map.iter() {
    println!("Tracepoint: {}, Enabled: {}", name, tracepoint.is_enabled());
}

Examples

See examples/usage.rs for a complete example of defining and using tracepoints.

cd examples
cargo run --example usage

Architecture

The crate provides several key components:

1. TracingEventsManager: Manages subsystems and their tracepoints
2. EventsSubsystem: Groups related tracepoints
3. EventInfo: Contains tracepoint metadata and control
4. TracePointEnableFile: Controls tracepoint enable/disable state
5. KernelTraceOps: Trait for implementing kernel-level operations

Safety

This crate is designed for kernel space usage and:

• Uses #![no_std] for kernel compatibility

• Provides safe abstractions for kernel tracepoint management

Example

• See DragonOS tracepoint for more details.
• See Hermit tracepoint for more details.