idle priority > 30 in top(1) and idprio > 31

[Anonymous]

Is the following a regression?

  $ uname -vm
  FreeBSD 9.0-HEAD-20101216-JPSNAP #0: Thu Dec 16 07:13:24 UTC 2010
  root at build-amd64-fbsd.allbsd.org:/usr/obj/usr/src/sys/GENERIC  amd64

  $ yes >/dev/null &
  $ idprio 31 -$(pgrep yes)
  $ top
    PID USERNAME  THR PRI NICE   SIZE    RES STATE    TIME   WCPU COMMAND
   2139 luser       1 117 i-23 10052K  1128K RUN      0:41 98.97% yes
  $ idprio 444 -$(pgrep yes)
  $ top
    PID USERNAME  THR PRI NICE   SIZE    RES STATE    TIME   WCPU COMMAND
   2139 luser       1 117 i-102 10052K  1128K RUN     0:42 97.56% yes

On 8.0-RELEASE it's different

  $ yes >/dev/null &
  $ idprio 31 -$(pgrep yes)
  $ top
    PID USERNAME  THR PRI NICE   SIZE    RES STATE    TIME   WCPU COMMAND
   2653 luser       1 171  i31  5808K   596K RUN      0:10 97.88% yes
  $ idprio 444 -$(pgrep yes)
  idprio: idprio: Invalid argument

- 8.1-RELEASE accepts idprio > 31, e.g. 444.
- 8.2-BETA1 shows negative values in top(1) like -CURRENT