Safe Lock-free Primitives with iceoryx2's ByteAtomic

4 hours ago

Sequence locks detect concurrent modifications but still cause undefined behavior due to non-atomic memory copies.
iceoryx2's ByteAtomic provides byte-wise atomic read/write to prevent undefined behavior during memory copies.
ByteAtomic avoids copying uninitialized memory (like padding bytes) by using the AtomicCopy trait for field-wise copying.
The read operation returns MaybeTorn<T> to indicate potential torn reads, requiring additional synchronization checks.
A derive macro simplifies implementing AtomicCopy for structs, making the wrapper easy to use.
ByteAtomic enables correct and safe implementation of sequence locks and other lock-free primitives in safety-critical systems.

RocksDB uses an LSM-tree that creates new versions of keys on every write, providing half of MVCC; the other half requires snapshots and garbage collection aware of in-flight reads.
Each write assigns a monotonically increasing sequence number; reads capture the published sequence at start to see a consistent snapshot, skipping higher sequence numbers.
The memtable is a concurrent skip list ordered by user key ascending and sequence number descending; atomic writes use write batches to allocate sequence numbers atomically, ensuring no partial updates are visible.

Fil-C ensures memory safety via a capability system for pointers, dynamically tracking allowed objects and prohibiting out-of-bounds, use-after-free, and other unsafe accesses.
InvisiCaps are the latest capability model, replacing PLUT, SideCaps, and MonoCaps, offering 64-bit pointers on 64-bit systems and reduced performance overhead to about 4x.
InvisiCaps support meaningful union usage, thread safety, and flexible memory type reinterpretation without compromising safety.

Linux 6.13 introduces atomic block writes to prevent data tearing in databases.
Atomic writes require hardware support (NVMe/SCSI) and software stack updates (VFS, filesystem, block layer).
Databases use double writes for consistency, which atomic writes can replace to improve performance.

LMDB is a Btree-based embedded database management library with a simplified BerkeleyDB-like API.
It uses memory-mapped files for direct data access, eliminating the need for malloc or memcpy during fetches, resulting in high performance and simplicity.
The database is fully transactional with ACID semantics, thread-aware, and supports concurrent read/write access across multiple processes and threads.

Go 1.27 introduces type parameters on methods, allowing generic operations on types to be defined as methods instead of package-level functions.
Struct literal keys can now use promoted fields from embedded structs directly, simplifying initialization syntax.
Function type inference is generalized to work in conversions, composite literals, and other contexts, reducing the need for explicit type arguments.

RocksDB uses an LSM-tree that creates new versions of keys on every write, providing half of MVCC; the other half requires snapshots and garbage collection aware of in-flight reads.
Each write assigns a monotonically increasing sequence number; reads capture the published sequence at start to see a consistent snapshot, skipping higher sequence numbers.
The memtable is a concurrent skip list ordered by user key ascending and sequence number descending; atomic writes use write batches to allocate sequence numbers atomically, ensuring no partial updates are visible.

Fil-C ensures memory safety via a capability system for pointers, dynamically tracking allowed objects and prohibiting out-of-bounds, use-after-free, and other unsafe accesses.
InvisiCaps are the latest capability model, replacing PLUT, SideCaps, and MonoCaps, offering 64-bit pointers on 64-bit systems and reduced performance overhead to about 4x.
InvisiCaps support meaningful union usage, thread safety, and flexible memory type reinterpretation without compromising safety.

Linux 6.13 introduces atomic block writes to prevent data tearing in databases.
Atomic writes require hardware support (NVMe/SCSI) and software stack updates (VFS, filesystem, block layer).
Databases use double writes for consistency, which atomic writes can replace to improve performance.

LMDB is a Btree-based embedded database management library with a simplified BerkeleyDB-like API.
It uses memory-mapped files for direct data access, eliminating the need for malloc or memcpy during fetches, resulting in high performance and simplicity.
The database is fully transactional with ACID semantics, thread-aware, and supports concurrent read/write access across multiple processes and threads.

Go 1.27 introduces type parameters on methods, allowing generic operations on types to be defined as methods instead of package-level functions.
Struct literal keys can now use promoted fields from embedded structs directly, simplifying initialization syntax.
Function type inference is generalized to work in conversions, composite literals, and other contexts, reducing the need for explicit type arguments.

Hasty Briefsbeta