On Fri, Jul 24, 2026 at 05:56:03AM -0400, Michael S. Tsirkin wrote: > As any optimization, I expect actual perf measurements > showing an improvement to be included please. > > So it is still O(N^2) right? It just skips calling it for > some sections.
You are right. The change only skipped non-overlapping sections and did not really address the complexity, and I do not have measurements showing a meaningful win on the only call site (full-range sync in vhost_log_put). I will drop this patch. Thanks, Xiong Weimin
