Merge pull request #20 from doublegate/copilot/sub-pr-18

Migrate from bincode to rkyv v0.8.14 for memory-mapped serialization

Author: doublegate

Cargo.lock

@@ -43,7 +43,7 @@ rusqlite = { version = "0.31", features = ["bundled"] }
 
 # Memory-mapped I/O (Sprint 6.6)
 memmap2 = "0.9"
-bincode = "1.3"
+rkyv = "0.8.14"
 
 # CSV export (promote to workspace)
 csv = "1.3"

Cargo.toml

@@ -30,6 +30,7 @@ rand = { workspace = true }
 sysinfo = { workspace = true }
 flate2 = "1.0"
 dirs = "5.0"
+rkyv = { workspace = true }
 
 [dev-dependencies]
 tokio = { workspace = true }

crates/prtip-core/Cargo.toml

@@ -136,4 +136,4 @@ pub use resource_monitor::{
 pub use retry::{retry_with_backoff, RetryConfig};
 pub use service_db::{ServiceMatch, ServiceProbe, ServiceProbeDb};
 pub use templates::{ScanTemplate, TemplateManager};
-pub use types::{PortRange, PortState, Protocol, ScanResult, ScanTarget, ScanType, TimingTemplate};
+pub use types::{PortRange, PortState, Protocol, ScanResult, ScanResultRkyv, ScanTarget, ScanType, TimingTemplate};

crates/prtip-core/src/lib.rs

@@ -3,6 +3,7 @@
 use crate::error::{Error, Result};
 use chrono::{DateTime, Utc};
 use ipnetwork::IpNetwork;
+use rkyv::{Archive, Deserialize as RkyvDeserialize, Serialize as RkyvSerialize};
 use serde::{Deserialize, Serialize};
 use std::fmt;
 use std::net::IpAddr;
@@ -315,7 +316,8 @@ impl Iterator for PortRangeIterator {
 }
 
 /// State of a scanned port
-#[derive(Debug, Clone, Copy, PartialEq, Eq, PartialOrd, Ord, Serialize, Deserialize)]
+#[derive(Debug, Clone, Copy, PartialEq, Eq, PartialOrd, Ord, Serialize, Deserialize, Archive, RkyvSerialize, RkyvDeserialize)]
+#[rkyv(derive(Debug))]
 pub enum PortState {
     /// Port is open and accepting connections
     Open,
@@ -475,6 +477,53 @@ impl fmt::Display for TimingTemplate {
     }
 }
 
+/// Serializable representation of ScanResult for rkyv
+#[derive(Debug, Clone, Archive, RkyvSerialize, RkyvDeserialize)]
+#[rkyv(derive(Debug))]
+pub struct ScanResultRkyv {
+    target_ip: IpAddr,
+    port: u16,
+    state: PortState,
+    response_time_nanos: u128,
+    timestamp_nanos: i64,
+    banner: Option<String>,
+    service: Option<String>,
+    version: Option<String>,
+    raw_response: Option<Vec<u8>>,
+}
+
+impl From<&ScanResult> for ScanResultRkyv {
+    fn from(result: &ScanResult) -> Self {
+        Self {
+            target_ip: result.target_ip,
+            port: result.port,
+            state: result.state,
+            response_time_nanos: result.response_time.as_nanos(),
+            timestamp_nanos: result.timestamp.timestamp_nanos_opt().unwrap_or(0),
+            banner: result.banner.clone(),
+            service: result.service.clone(),
+            version: result.version.clone(),
+            raw_response: result.raw_response.clone(),
+        }
+    }
+}
+
+impl From<ScanResultRkyv> for ScanResult {
+    fn from(rkyv: ScanResultRkyv) -> Self {
+        Self {
+            target_ip: rkyv.target_ip,
+            port: rkyv.port,
+            state: rkyv.state,
+            response_time: Duration::from_nanos(rkyv.response_time_nanos as u64),
+            timestamp: DateTime::from_timestamp_nanos(rkyv.timestamp_nanos),
+            banner: rkyv.banner,
+            service: rkyv.service,
+            version: rkyv.version,
+            raw_response: rkyv.raw_response,
+        }
+    }
+}
+
 /// Result of scanning a single port
 #[derive(Debug, Clone, Serialize, Deserialize)]
 pub struct ScanResult {

crates/prtip-core/src/types.rs

@@ -87,7 +87,7 @@ pcap-file = "2.0"
 
 # Memory-mapped I/O
 memmap2 = { workspace = true }
-bincode = { workspace = true }
+rkyv = { workspace = true, features = ["alloc"] }
 
 [dev-dependencies]
 tokio = { workspace = true }

crates/prtip-scanner/Cargo.toml

@@ -1,13 +1,14 @@
 //! Memory-mapped result reader for zero-copy access to scan results
 
 use memmap2::Mmap;
-use prtip_core::ScanResult;
+use prtip_core::{ScanResult, ScanResultRkyv};
 use std::fs::File;
 use std::io;
 use std::path::Path;
 
 const HEADER_SIZE: usize = 64;
 const ENTRY_SIZE: usize = 512;
+const LENGTH_PREFIX_SIZE: usize = 8; // 8 bytes for length to maintain alignment
 
 /// Memory-mapped result reader
 pub struct MmapResultReader {
@@ -75,10 +76,24 @@ impl MmapResultReader {
         }
 
         let offset = HEADER_SIZE + (index * self.entry_size);
-        let entry_bytes = &self.mmap[offset..offset + self.entry_size];
+        
+        // Read length prefix (8 bytes)
+        let len_bytes: [u8; 8] = self.mmap[offset..offset + LENGTH_PREFIX_SIZE].try_into().ok()?;
+        let len = u64::from_le_bytes(len_bytes) as usize;
+        
+        if len == 0 || len + LENGTH_PREFIX_SIZE > self.entry_size {
+            return None;
+        }
 
-        // Deserialize the entry (bincode handles trailing zeros)
-        bincode::deserialize(entry_bytes).ok()
+        // Copy data to an aligned buffer (rkyv requires alignment)
+        let entry_bytes = &self.mmap[offset + LENGTH_PREFIX_SIZE..offset + LENGTH_PREFIX_SIZE + len];
+        let aligned_data: Vec<u8> = entry_bytes.to_vec();
+
+        // Deserialize using rkyv
+        match rkyv::from_bytes::<ScanResultRkyv, rkyv::rancor::Error>(&aligned_data) {
+            Ok(rkyv_result) => Some(rkyv_result.into()),
+            Err(_) => None,
+        }
     }
 
     /// Create an iterator over all entries