Case 42: Type Alignment Changed (standalone alignas)¶

Field	Value
Verdict	🔴 BREAKING
Category	Breaking
Platforms	Linux, macOS
Flags	ABI break, API break
Detected `ChangeKind`s	—
Source files	browse on GitHub

Category: Type Layout / DWARF | Verdict: BREAKING

What this case is about¶

v1 defines CacheBlock with aligned(8). v2 increases alignment to aligned(64) for cache-line optimization. The fields and their types are identical — only the alignment attribute changes.

This is a clean, isolated alignment change (unlike case41 which bundles alignment with type removal and enum changes).

What breaks at binary level¶

sizeof may change: Compilers pad structs to a multiple of their alignment. aligned(64) makes sizeof(CacheBlock) = 64 (padded to alignment boundary). With aligned(8) it's also 64 here, but the ABI contract about where the struct can live in memory changes.
Stack allocation misaligned: Old binaries allocate CacheBlock with 8-byte alignment. The v2 library may use SIMD instructions (e.g., vmovdqa) that require 64-byte alignment → SIGBUS / SIGSEGV.
Array stride changes: If sizeof changes, &blocks[i] computes wrong offsets.
malloc alignment: malloc typically returns 16-byte aligned memory. 64-byte aligned structs need aligned_alloc(64, sizeof(CacheBlock)).

What abicheck detects¶

TYPE_ALIGNMENT_CHANGED: Detected via DWARF DW_AT_alignment or inferred from DW_AT_byte_size changes caused by alignment padding.

Overall verdict: BREAKING

How to reproduce¶

gcc -shared -fPIC -g bad.c  -include bad.h  -o libbad.so
gcc -shared -fPIC -g good.c -include good.h -o libgood.so

python3 -m abicheck.cli dump libbad.so  -o /tmp/v1.json
python3 -m abicheck.cli dump libgood.so -o /tmp/v2.json
python3 -m abicheck.cli compare /tmp/v1.json /tmp/v2.json
# → BREAKING: TYPE_ALIGNMENT_CHANGED

How to fix¶

Use opaque types so callers never allocate or embed the struct directly:

/* header */
typedef struct CacheBlock CacheBlock;
CacheBlock* block_alloc(void);  /* library controls alignment */

/* implementation */
struct CacheBlock __attribute__((aligned(64))) { ... };
CacheBlock* block_alloc(void) {
    return aligned_alloc(64, sizeof(CacheBlock));
}

Real-world examples¶

DPDK packet buffers require cache-line alignment (64 bytes).
Intel TBB / oneTBB uses alignas(64) for scalable allocator metadata.
Changing alignment in a public struct after release broke ABI in several multimedia libraries (FFmpeg, GStreamer).