chore: release v2.2.1

Cargo.toml

@@ -132,7 +132,7 @@ members = [
 # different files that never got the section. Fixed here, in both, and
 # verified by extracting the packaged `.crate` before publishing rather than
 # checking the live page afterwards.
-version = "2.2.0"
+version = "2.2.1"
 edition = "2024"
 # MSRV. Declared for the first time in 1.0.1 because this release genuinely
 # needs it: `asm!` label blocks (`asm_goto`, stable 1.87) carry the free path's

README.md

@@ -39,10 +39,14 @@ does not offer.
   measured on a XIAO ESP32-S3 at 240 MHz, both allocators built from one source.
   It costs more RAM to get that (64 KiB vs 8 KiB); both numbers are below.
 
-> **Status: `2.2.0`.** The API is frozen and changes follow semver. 2.0.1
-> through 2.0.5 were patch releases; 2.1.0 and 2.2.0 are minors because they
-> ADD public items and move none — `cargo-semver-checks` calls both
-> API-compatible. **2.2.0 also carries the largest embedded speed fix in this
+> **Status: `2.2.1`.** The API is frozen and changes follow semver. 2.0.1
+> through 2.0.5 and 2.2.1 are patch releases; 2.1.0 and 2.2.0 are minors because
+> they ADD public items and move none — `cargo-semver-checks` calls every one
+> API-compatible. **2.2.1 is a speed release** (it also adds `malloc_aligned_pow2` and `prim::getenv`, moving nothing): four rounds of exact-instruction
+> work (the cross-thread free 32 % cheaper, a 2 MiB allocate-and-free 57 %),
+> and a `GlobalAlloc` fast path worth up to 27 % whole-program on Rust
+> workloads — see [Performance](#performance-deterministic-instruction-counts).
+> **2.2.0 also carries the largest embedded speed fix in this
 > series:** under `--cfg ra_small_profile` every page was being extended one
 > block at a time, so `malloc_generic` ran on 100 % of allocations; on a XIAO
 > ESP32-S3 fixing it is **13–16 % faster on every binned workload**.

crates/rusty_alloc/CHANGELOG.md

@@ -7,6 +7,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [2.2.1](https://github.com/Remade-With-Rust/rusty_alloc/compare/rusty_alloc-v2.2.0...rusty_alloc-v2.2.1) - 2026-09-25
+
 ### Fixed
 
 - **Growing a huge block with `realloc` copied its whole reservation.** A

crates/rusty_alloc_api/CHANGELOG.md

@@ -7,6 +7,23 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [2.2.1](https://github.com/Remade-With-Rust/rusty_alloc/compare/rusty_alloc-api-v2.2.0...rusty_alloc-api-v2.2.1) - 2026-09-25
+
+### Performance
+
+- **`GlobalAlloc` methods are `#[inline]`**, as the `mimalloc` crate's are, so
+  rustc's `__rust_alloc` / `__rust_dealloc` shims carry the fast path instead
+  of jumping to it, and `dealloc` is the free fast path with its null test
+  folded away. Measured whole-program against 2.2.0 on two deterministic Rust
+  workloads (`bench/rust-globalalloc.sh`): **-21.1 %** and **-5.8 %**.
+- **Layouts aligned up to two words (16 bytes on 64-bit) come from the natural
+  size classes**, which are already aligned that far, instead of the aligned
+  path; every hashbrown table is such a layout. `tests/natural_align.rs` pins
+  the alignment for every size class through alloc, alloc_zeroed and realloc.
+- **`realloc` keeps a block in place at any alignment when it fits** instead
+  of always allocating, copying and freeing above 8 bytes of alignment:
+  **-27.3 %** on an over-aligned buffer workload.
+
 ## [2.2.0](https://github.com/Remade-With-Rust/rusty_alloc/compare/rusty_alloc-api-v2.1.0...rusty_alloc-api-v2.2.0) - 2026-09-10
 
 ### Other

crates/rusty_alloc_api/Cargo.toml

@@ -14,7 +14,7 @@ readme = "README.md"
 [dependencies]
 # A published crate needs a VERSION alongside the path: the path is what the
 # workspace builds against, the version is what crates.io resolves.
-rusty_alloc = { path = "../rusty_alloc", version = "2.2.0", default-features = false }
+rusty_alloc = { path = "../rusty_alloc", version = "2.2.1", default-features = false }
 
 [features]
 # Mirrors the core's (P3, docs/plans/small-metal.md). Default-on, so every