On Wed, Dec 20, 2023 at 7:21 AM Juraj Marcin <juraj@xxxxxxxxxxxxxxx> wrote: > > Hi Paul > > thank you for your valuable feedback. I will incorporate your > suggestions to my code. I will then report back with benchmark results > with included optimizations. I suspect, kstrdup() in > filename_compute_type() might also be part of the reason, the patched > kernel is few tenths of microsecond slower, so I will also look into > that. Thank you. I did look at the dup'd string for a bit, but didn't see a relatively quick way to do away with the additional string copy; if you can come up with something that isn't too ugly I think that would be great. -- paul-moore.com
