> From: Mark H Weaver <m...@netris.org> > Thanks for nudging me to do the measurement. To be honest, the original > patch I posted slowed things down by 4.5%, which I found extremely > surprising. I fixed it by adding an internal 'static' version of > 'scm_fill_input'. So for those who doubt the cost of function calls > though the shared library PLT (which, within a shared library, is > *every* function call to a non-static function), let this be a lesson. > That's the only thing that changed here, and it made more than a 4% > difference, even though the procedure call in question was done only > once per 4 kilobyte buffer!
> I've attached the new patches. Okay to push now? I don't think I get a vote, but, I vote yes anyway. -Mike
