Remove internal type references from documentation

Update docs to match cleaned-up public API surface. Remove references
to internal types (BackendMode, TopicRole, TopicConfig, RtKernelInfo,
RtCpuInfo, LogEntry, SIMD_COPY_THRESHOLD, TensorPoolStats fields) and
internal helper functions (pin_thread_to_core, detect_isolated_cpus,
get_rt_recommended_cpus, prefault_stack). Fix import paths for
RtScheduler and RtApplyResult which moved out of the prelude.

Author: neos-builder

content/docs/advanced/rt-config.mdx

@@ -19,9 +19,11 @@ RtConfig is designed for **development on any OS** with **production deployment
 | **macOS** | Graceful degradation (no RT) | Development & testing |
 | **Windows** | Graceful degradation (no RT) | Development & testing |
 
-On non-Linux platforms, `RtConfig::apply()` returns `RtApplyResult::Degraded` and your code continues running normally - just without kernel-level RT guarantees. This allows you to develop and test on any OS.
+On non-Linux platforms, `RtConfig::apply()` returns a degraded result and your code continues running normally - just without kernel-level RT guarantees. This allows you to develop and test on any OS.
 
 ```rust
+use horus::core::rt_config::RtApplyResult;
+
 let config = RtConfig::hard_realtime(Some(&[2, 3]));
 match config.apply() {
     Ok(RtApplyResult::FullSuccess) => println!("Full RT enabled (Linux)"),
@@ -48,7 +50,7 @@ match config.apply() {
 ## Quick Start
 
 ```rust
-use horus::prelude::*;  // Provides RtConfig, RtScheduler
+use horus::prelude::*;
 
 // Apply hard real-time configuration
 let config = RtConfig::hard_realtime(Some(&[2, 3]));  // Pin to cores 2, 3
@@ -104,6 +106,7 @@ For fine-grained control:
 
 ```rust
 use horus::prelude::*;
+use horus::core::rt_config::RtScheduler;
 
 let config = RtConfig::new()
     .memory_locked(true)           // mlockall()
@@ -130,14 +133,14 @@ config.apply()?;
 
 ### Scheduling Classes
 
-```rust
-pub enum RtScheduler {
-    Normal,      // Default Linux scheduler (SCHED_OTHER)
-    Fifo,        // SCHED_FIFO - First-in-first-out RT
-    RoundRobin,  // SCHED_RR - Round-robin RT
-    Deadline,    // SCHED_DEADLINE - For periodic tasks (not yet implemented)
-}
-```
+Available via `use horus::core::rt_config::RtScheduler`:
+
+| Variant | Linux Scheduler | Description |
+|---------|----------------|-------------|
+| `Normal` | `SCHED_OTHER` | Default Linux scheduler |
+| `Fifo` | `SCHED_FIFO` | First-in-first-out RT |
+| `RoundRobin` | `SCHED_RR` | Round-robin RT |
+| `Deadline` | `SCHED_DEADLINE` | Periodic tasks (not yet implemented) |
 
 | Class | Preemption | Use Case |
 |-------|------------|----------|
@@ -257,6 +260,8 @@ With full RT configuration on PREEMPT_RT kernel:
 `RtConfig` degrades gracefully when features aren't available:
 
 ```rust
+use horus::core::rt_config::RtApplyResult;
+
 let config = RtConfig::hard_realtime(Some(&[2, 3]));
 match config.apply() {
     Ok(RtApplyResult::FullSuccess) => {
@@ -307,53 +312,6 @@ let cpus = RtConfig::get_current_affinity()?;
 println!("CPU affinity: {:?}", cpus);
 ```
 
-### Detect System RT Capabilities
-
-```rust
-// Get detailed kernel RT information
-let info = RtKernelInfo::detect();
-println!("PREEMPT_RT: {}", info.preempt_rt);
-println!("Kernel: {}", info.kernel_version);
-println!("Max RT priority: {}", info.max_rt_priority);
-println!("mlockall permitted: {}", info.mlockall_permitted);
-println!("CPU count: {}", info.cpu_count);
-```
-
-### Find Best CPUs for RT
-
-```rust
-// Get recommended CPUs for RT tasks (prefers isolated + nohz_full cores)
-let rt_cpus = get_rt_recommended_cpus(2);
-println!("Recommended RT CPUs: {:?}", rt_cpus);
-
-// Or get full CPU info
-let cpu_info = RtCpuInfo::detect();
-println!("{}", cpu_info.summary());
-```
-
-## Helper Functions
-
-HORUS provides standalone helper functions for common RT operations:
-
-```rust
-use horus::prelude::*;
-
-// Pin current thread to a single CPU core
-pin_thread_to_core(2)?;
-
-// Detect kernel-isolated CPUs (from isolcpus boot parameter)
-let isolated = detect_isolated_cpus();
-
-// Detect tickless CPUs (from nohz_full boot parameter)
-let nohz = detect_nohz_full_cpus();
-
-// Get recommended CPUs for RT (prefers isolated+nohz_full, then isolated, etc.)
-let best_cpus = get_rt_recommended_cpus(2);
-
-// Prefault stack memory to avoid page faults
-prefault_stack(512 * 1024);  // 512KB
-```
-
 ## Troubleshooting
 
 ### "Operation not permitted" on mlockall
@@ -400,6 +358,7 @@ cat /sys/devices/system/cpu/isolated
 
 ```rust
 use horus::prelude::*;
+use horus::core::rt_config::{RtApplyResult, RtScheduler};
 
 fn main() -> Result<()> {
     // Configure system for hard real-time

content/docs/concepts/communication-configuration.mdx

@@ -93,17 +93,16 @@ println!("Camera endpoint: {}", endpoint);
 // For direct: "sensor@192.168.1.5:9000"
 ```
 
-### Creating Topics from Config
+### Creating Topics
 
 ```rust
 use horus::prelude::*;
 
-// Create a topic with TopicConfig
-let topic_config = TopicConfig::new("sensors").with_capacity(16);
-let topic: Topic<SensorData> = Topic::from_config(topic_config)?;
-
-// Or create directly with Topic::new using the hub name
+// Create a topic using the hub name
 let topic: Topic<SensorData> = Topic::new("sensors")?;
+
+// Or with custom capacity
+let topic: Topic<SensorData> = Topic::with_capacity("sensors", 32, None)?;
 ```
 
 ## EndpointConfig Fields

content/docs/concepts/communication-transport.mdx

@@ -117,15 +117,6 @@ let camera: Topic<Image> = Topic::new("camera")?;
 
 ## Topic Configuration
 
-### From Config Struct
-
-```rust
-use horus::prelude::*;
-
-let config = TopicConfig::new("sensors").with_capacity(16);
-let topic: Topic<SensorData> = Topic::from_config(config)?;
-```
-
 ### Custom Capacity
 
 ```rust
@@ -172,7 +163,7 @@ mDNS allows nodes to discover each other without hardcoding IP addresses. This i
 
 ## Endpoint Configuration (Planned)
 
-> **Note:** The endpoint configuration system (`HorusConfig`, `EndpointConfig`, transport types) is **planned but not yet implemented**. See [Communication Configuration](/concepts/communication-configuration) for the intended design. Currently, all topics use local shared memory via `Topic::new("name")` or `Topic::from_config(TopicConfig)`.
+> **Note:** The endpoint configuration system (`HorusConfig`, `EndpointConfig`, transport types) is **planned but not yet implemented**. See [Communication Configuration](/concepts/communication-configuration) for the intended design. Currently, all topics use local shared memory via `Topic::new("name")`.
 
 ## Summary
 
@@ -183,7 +174,7 @@ mDNS allows nodes to discover each other without hardcoding IP addresses. This i
 | **Throughput** | 2.5M+ msgs/sec (SPSC), 1.5M+ msgs/sec (MPMC) |
 | **Backend Selection** | Automatic based on usage pattern |
 | **Platforms** | Linux (`/dev/shm`), macOS (`shm_open`), Windows (page file) |
-| **Configuration** | `Topic::new("name")` or `Topic::from_config(TopicConfig)` |
+| **Configuration** | `Topic::new("name")` or `Topic::with_capacity("name", N, None)` |
 
 ## Next Steps
 

content/docs/concepts/core-concepts-topic.mdx

@@ -306,19 +306,6 @@ Creates a topic with a custom ring buffer capacity. By default, HORUS auto-sizes
 let topic: Topic<f32> = Topic::with_capacity("velocity", 256, None)?;
 ```
 
-### from_config()
-
-```rust
-pub fn from_config(config: TopicConfig) -> Result<Self>
-```
-
-Creates a topic from a `TopicConfig` struct. Primarily used by Python bindings and config-file-based topic creation.
-
-```rust
-let config = TopicConfig::new("velocity").with_capacity(128);
-let topic: Topic<f32> = Topic::from_config(config)?;
-```
-
 ## Type-Safe Topic Descriptors
 
 The `topics!` macro defines compile-time topic descriptors that prevent topic name typos and type mismatches across your codebase.

content/docs/concepts/zero-overhead-ipc.mdx

@@ -123,12 +123,12 @@ The `with_logging()` method requires `T: LogSummary`. Use `#[derive(LogSummary)]
 2. Records `Instant::now()` before the IPC operation
 3. Performs the IPC send
 4. Calculates elapsed IPC time
-5. Publishes a `LogEntry` (timestamp, node name, tick number, topic, summary, IPC latency)
+5. Records a log entry with timestamp, node name, tick number, topic, summary, and IPC latency
 
 **What `with_logging()` adds to `recv()`:**
 1. Records `Instant::now()` before the IPC operation
 2. Performs the IPC receive
-3. If a message was received, calls `msg.log_summary()` and publishes a `LogEntry`
+3. If a message was received, calls `msg.log_summary()` and records a log entry
 
 **Latency:** +200-500ns overhead per operation
 

content/docs/performance/performance.mdx

@@ -179,7 +179,7 @@ RUSTFLAGS="-C target-feature=+avx2" cargo build --release
 
 SIMD acceleration automatically applies to:
 
-- **Topic send/recv**: Message copying for messages >= 4KB (`SIMD_COPY_THRESHOLD`)
+- **Topic send/recv**: Message copying for messages >= 4KB
 - **Shared memory operations**: Large buffer transfers using non-temporal streaming stores
 
 For messages smaller than 4KB, standard `ptr::copy_nonoverlapping` is used (setup overhead exceeds SIMD benefit for small copies).
@@ -194,7 +194,7 @@ Benchmarks on Intel Core i9-13900K (AVX2 enabled):
 | 304B (IMU) | ~280ns | ~160ns | 1.75x |
 | 1.5KB (LaserScan) | ~800ns | ~400ns | 2x |
 
-> **Note**: SIMD acceleration applies to messages >= 4KB (`SIMD_COPY_THRESHOLD`). Smaller messages use standard `ptr::copy_nonoverlapping`. The latency improvements above are from the overall Topic pipeline, including cache-line alignment and lock-free algorithms.
+> **Note**: SIMD acceleration applies to messages >= 4KB. Smaller messages use standard `ptr::copy_nonoverlapping`. The latency improvements above are from the overall Topic pipeline, including cache-line alignment and lock-free algorithms.
 
 ### Fallback Behavior
 
@@ -561,7 +561,7 @@ HORUS automatically tracks IPC timing for each topic operation. The `horus monit
 Tick: 12μs | IPC: 296ns
 ```
 
-Each `LogEntry` includes `tick_us` (node tick time in microseconds) and `ipc_ns` (IPC write time in nanoseconds).
+Each log entry includes `tick_us` (node tick time in microseconds) and `ipc_ns` (IPC write time in nanoseconds).
 
 ### Manual Profiling