On Sun, Feb 25, 2024 at 08:58:28PM -0500, Kent Overstreet wrote: > I think everything works; we need the end result to be consistent with > some total ordering of all the writes, IOW, thread B's write (if fully > within thread A's write) should be fully overwritten or not at all, and > that clearly is the case. But there may be situations involving more > than two threads where things get weirder. Not sure if there's anything interesting or deep here, but this situations where sometimes we've got a global lock and sometimes we've got little locks do seem to come up - it reminded me of Suren's work on the mmap sem recently, and I think I've seen it in other places. Feels like it should at least have a name...
