On Mon, 8 Sep 2008, Warner Losh wrote:

 Commit the no-brainer parts of my space saving kernel experiments.  We
 don't inline the locking primitives, and only grab those parts of mii
 that we really need.  Other space optimizations are too agressive for
 the generic file (removing all of usb, and loading it as kernel
 modules).

It's been a few years since we selected defaults on lock inlining (and other optional parameters), and it might well be time to re-evaluate our defaults. Cache footprint on code continues to play a significant role in performance, the trade-offs for function invocation have changed quite a bit over the years. A more thorough performance analysis of lock inlining is probably overdue for all platforms, not just small embedded ones.

Robert N M Watson
Computer Laboratory
University of Cambridge
_______________________________________________
[email protected] mailing list
http://lists.freebsd.org/mailman/listinfo/cvs-all
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Reply via email to