On Sat, Jan 28, 2023 at 6:06 PM Alan Bateman <alan.bate...@oracle.com> wrote:
> 8k is the threshold at which RAF spills into a temporary malloc'ed > buffer for the I/O operation. At some point the RAF implementation will > be replaced (we need to do this for a Project Loom), but in the mean > time, it probably would require testing on a wide set of platforms and > conditions to see the impact. > Makes sense! Would it be worth going through a PR to add a comment to document this somewhat obscure code? https://github.com/openjdk/jdk/compare/master...eirbjo:jdk:raf-buf-threshold Thanks, Eirik.