Pull up following revision(s) (requested by nat in ticket #485):
sys/dev/auconv.c: 1.32-1.35
sys/dev/auconv.h: 1.20-1.21
sys/dev/mulaw.c: 1.33
sys/dev/mulaw.h: 1.25
Improve LINEAR_LINEARN macro - inparticular BE/LE conversions.
Addresses part of PR kern/52586 analyzed and reported by isaki@.
--
Allow maxused to be an odd number of bytes.  This makes single byte
conversions work for the test described in the PR.
Addresses another part of PR kern/52586.
--
Remove 24 bit in a 32 bit container filters.
At present this filter did not work well and has caused problems with
the audio mixer disabled for hdaudio(4).
These filters may return when they are fixed.
--
Replace LINEARN_TO_LINEARN macro with individual optimized functions for
8, 16, 24 and 32 bits linear to linear conversions.
Addresses PR kern/52586.
--
mulaw/alaw_to_linearN macros replaced with individual faster filters.
NFC.

(snj)