doc: update

README.md

@@ -7,16 +7,16 @@ This eBPF application implements an "Endpoint-Independent Mapping" and "Endpoint
 -   **eBPF**: IPv4 to IPv4 NAPT(Network Address Port Translation)
 -   **eBPF**: IPv6 to IPv6 NAPT
 -   **eBPF**: Endpoint-Independent(Full Cone) NAT for TCP, UDP and ICMP
--   **eBPF**: Partial port-range usage, allows reserving external ports for other usage
+-   **eBPF**: Partial external port range usage, allows reserving external ports for other usage
 -   **Frontend**: Automatic reconfiguration on interface address changes
--   **Frontend**: Automatic hairpin IP rule and route setup, see https://github.com/EHfive/einat-ebpf/issues/4
+-   **Frontend**: Automatic IP rule and route setup for hairpinning, see https://github.com/EHfive/einat-ebpf/issues/4
 
 ## Requirement
 
 -   Linux kernel >= 5.15 (compiled with BPF and BTF support) on target machine
 -   `libelf` from elfutils
 -   `zlib`
--   `clang` to compile BPF C code
+-   `clang` for bindgen and compiling BPF C code
 -   `cargo` and `rustfmt` for building
 
 It's also required the eBPF JIT implementation for target architecture in kernel has implemented support for BPF-to-BPF calls, which is not the case for MIPS and other architectures have less interests. This application is only tested to work on x86-64 or aarch64.
@@ -69,7 +69,7 @@ You would only need to specify external interface name in a minimal setup, and `
 sudo sysctl net.ipv4.ip_forward=1
 # With simplified CLI options,
 # this setup NAT for traffic forwarding to and from wan0 and setup hairpin
-# routing for traffic forwarding from lo and lan0 to wan0
+# routing for traffic forwarding from lo or lan0 to wan0
 sudo einat --ifname wan0 --hairpin-if lo lan0
 # With config file
 sudo einat --config /path/to/config.toml

config.sample.toml

@@ -35,7 +35,7 @@ nat66 = false
 # 0: disable, 1: error, 2: warn, 3: info, 4: debug, 5: trace
 # View logs with `cat /sys/kernel/debug/tracing/trace_pipe`
 bpf_log_level = 0
-# Enable external address(preferd source) lookup, recommend to enable.
+# Enable external address(preferd source) lookup, recommended to enable.
 # Only works on Linux kernel>=6.7, it's a no-op for kernel on lower version.
 bpf_fib_lookup_external = false
 # Set this to `false` for early disabling inbound ICMP binding initiation,
@@ -93,7 +93,7 @@ table_id = 4787
 address = "192.168.4.2"
 # The address would not be used as NAT external address if this is `true`.
 no_snat = false
-# Disable hairpin the address.
+# Disable hairpinning for the address.
 no_hairpin = false
 # Defaults to ranges in [defaults] if not specified.
 #tcp_ranges = ["10000-65535"]

docs/guide/cross.md

@@ -12,6 +12,12 @@ apt install libelf-dev zlib1g-dev gcc-aarch64-linux-gnu clang
 
 Install `rustup` to get Rust>=1.74, see https://www.rust-lang.org/tools/install. Also make sure `rustfmt` is installed as it's used by `libbpf-cargo`.
 
+Add required target to Rust toolchain:
+
+```
+rustup target add aarch64-unknown-linux-gnu
+```
+
 ### Target Dependencies
 
 ```

docs/guide/openwrt.md

@@ -3,7 +3,9 @@
 As said in [README](../../README.md), `einat` requires a kernel with BPF and BTF support enabled which is not the default.
 And `einat` requires running kernel for target architecture has support for BPF-to-BPF calls, which is not the case for MIPS and other older architectures.
 
-So if the architecture of your router is not x86-64 or aarch64 or other actively maintained architecture in kernel, your router would mostly not be able to run `einat`.
+So if the architecture of your router is not x86-64 or aarch64 or other actively maintained architecture in kernel, your router would mostly not be able to have `einat` working.
+
+The following is OpenWrt build configs required for `einat` to work.
 
 ### Build Configs