Actually, I wonder what the patch author intended to accomplish (and that his testing presumably demonstrated he did). The article mentions larger less scattered I/Os, but in a situation where the queue tends to be empty, I can't see where that would accomplish anything. When I look at this change, I just see a frontal assault on doing stuff in parallel with the I/O turnaround time.
Copyright © 2018, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds