[odc] Daily src changes for 2009-03-02

[ODC]

OpenBSD src changes summary for 2009-03-02
==========================================

libexec/ld.so                           share/man
sys/kern                                sys/lib/libsa
sys/net                                 sys/sys

== libexec =========================================================== 01/03 ==

  http://www.openbsd.org/cgi-bin/cvsweb/src/libexec

ld.so

  ~ ldd/ldd.1                             ~ ldd/ldd.c

  > ldd(1) accepts more than one program as argument; while here,
  > synchronize synopsis and usage. (sobrado@)

== share ============================================================= 02/03 ==

  http://www.openbsd.org/cgi-bin/cvsweb/src/share

man

  ~ man4/man4.mvme68k/lrc.4               ~ man4/man4.mvme68k/ofobio.4
  ~ man4/man4.mvme68k/pcctwo.4            

  > tweak previous; (jmc@)

  ~ man9/Makefile                         ~ man9/mbuf.9

  > Remove MEXTMALLOC and MCHTYPE from mbuf.9, as they have not existed
  > for some time.
  > ok jmc@ (blambert@)

== sys =============================================================== 03/03 ==

  http://www.openbsd.org/cgi-bin/cvsweb/src/sys

kern

  ~ uipc_mbuf.c                           

  > the packet length passed to m_clget is a u_int, print it with %u not %d in
  > the panic string. (dlg@)

lib/libsa

  ~ tftp.c                                ~ tftp.h

  > Send valid ERROR packets to prematurely terminate a transfer.
  > According to RFC1350 there should be always be a terminating NUL.
  > Andre Gillibert on bugs@ pointed out that the misformed packets
  > caused Gentoo Linux's tftpd (atftpd) to crash trying to transfer
  > pxeboot to an OpenBSD machine.
  > This is slightly different diff than the one proposed by Andre, but
  > he confirms it also fixes the problem.
  > "looks like a better diff to me" deraadt at . (krw@)

net

  ~ radix_mpath.c                         

  > Make sure rt_gateway is not NULL in rt_mpath_matchgate(). It is possible to
  > hit this case with a root node that comes with no real rtentry attached to
  > it.
  > Problem found by Mischa Diehm, OK henning@ (claudio@)

sys

  ~ tree.h                                

  > Fixup empty macro.	Original comment from harti at freebsd.org:
  > Make the default RB_AUGMENT() produce a 'do {} while (0)' instead
  > of nothing. This prevents the compiler from complaining about empty
  > if statements when compiled with higher WARN levels.
  > ok krw (mikeb@)

===============================================================================