On Tue, Nov 21, 2023 at 12:12:15AM -0500, Theodore Ts'o wrote: > Yeah, agreed, that would be a nice optimization. However, in the > unfortunate case where (a) it's non-ASCII, and (b) the input string is > non-normalized and/or differs in case, we end up scanning some portion > of the two strings twice; once doing the strcmp, and once doing the > Unicode slow path. The first pass is cheap and the second one will be running entirely from the cache, so I doubt that scanning them twice will be a loss...