[$] The ongoing quest for atomic buffered writes

There are many applications that need to be able to write multi-blockchunks of data to disk with the assurance that the operation will eithercomplete successfully or fail altogether - that the write will not bepartially completed (or "torn"), in other words. For years, kerneldevelopers have worked on providing atomic writes as a way of satisfyingthat need; see, for example, sessions from the Linux Storage, Filesystem,Memory Management, and BPF (LSFMM+BPF) Summit from 2023, 2024,and 2025 (twice). While atomic direct I/O is now supported by some filesystems, atomicbuffered I/O still is not. Fillingthat gap seems certain to be a 2026 LSFMM+BPF topic but, thanks to an earlydiscussion, the shape of a solution might already be coming into focus.