Hi,
Andreas Persson, while investigating for kern/14246, found out
that if he avoids the microtime()  in kern/sys_pipe.c:vfs_timestamp()
and use a stub instead, the performance of the benchmark using 2KB
pipe buffer raises like 3 times on his PIII. Thus, it seems microtime()
incurrs significant overhead to the pipe code.

Upon further investigation, it shows that microtime(9) on i386 uses
port i/o, which is very slow on modern processors. Is there a way
to improve this? This could help system overall, since microtime(9)
is used quite often.

Jaromir
-- 
Jaromir Dolecek <jdolecek@NetBSD.org> http://www.NetBSD.org/Ports/i386/ps2.html
-=  Those who would give up liberty for a little temporary safety deserve  =-
-=  neither liberty nor safety, and will lose both.  -- Benjamin Franklin  =-