[odc] Daily src changes for 2010-10-09

[ODC]

OpenBSD src changes summary for 2010-10-09
==========================================

lib/libpcap                             sys/dev/ic
sys/dev/pci                             sys/dev/usb
usr.bin/tmux                            usr.sbin/openssl
usr.sbin/smtpd                          usr.sbin/tcpdump

== lib =============================================================== 01/04 ==

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

libpcap

  ~ gencode.c                             

  > Fix typo in pflog header size computation. Fixes filtering with the new
  > pflog headers. reported by jmc@, tested by jmc@ and sthen@, ok sthen@
  > (canacar@)

== sys =============================================================== 02/04 ==

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

dev/ic

  ~ trm.c                                 

  > Make trm(4) work again. Avoid use-after-free of xs fields, don't
  > discard devices because 2nd INQUIRY (a.k.a. scsi_inquire_vpd())
  > fails, avoid NULL dereferences, reset bus after a timeout since I
  > can't see how else to account for an unsupported instruction (e.g.
  > 0x9e, a.k.a. READ_CAPACITY_16).
  > All found while preparing for iopoolification. (krw@)

dev/pci

  ~ cmpci.c                               

  > only print out of sync messages when compiled for debugging (jakemsr@)

dev/usb

  ~ uvideo.c                              ~ uvideo.h

  > control values can be 1 or 2 bytes, signed or unsigned.  fixes hue
  > control. (jakemsr@)

  ~ uvideo.c                              ~ uvideo.h

  > * define structure that represents USB descriptor with variable sized
  > member the way uaudio does, make the variable sized member as large
  > as it can be and don't define the following members.
  > * reenable control support for devices where bControlSize != 2 in
  > the processing unit descriptor.
  > tested on alpha, amd64, i386, sparc64 and zaurus, as well as the machine
  > where previous support for variable bControlSize caused a hang. (jakemsr@)

== usr.bin =========================================================== 03/04 ==

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

tmux

  ~ server-fn.c                           

  > Set cause when failing due to linking a window to itself, from Martin
  > Pieuchot. (nicm@)

== usr.sbin ========================================================== 04/04 ==

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

openssl

  ~ openssl.1                             

  > document "openssl genpkey"; (jmc@)

  ~ openssl.1                             

  > document "openssl pkey"; (jmc@)

smtpd

  + bounce.c                              

  > missing from previous commit (gilles@)

  + queue_shared.c                        + runner.c

  > these need to be re-added too (gilles@)

  - queue_backend.c                       - queue_backend.h

  > no longer compiled, i will reintroduce them later, lots of stuff needs to
  > be done before we can write queue backends anyway ... (gilles@)

  ~ aliases.c                             ~ client.c
  ~ control.c                             ~ enqueue.c
  ~ expand.c                              ~ lka.c
  ~ log.c                                 ~ map.c
  ~ mda.c                                 ~ mfa.c
  ~ mta.c                                 ~ parse.y
  ~ parser.c                              ~ parser.h
  ~ queue.c                               ~ smtp.c
  ~ smtp_session.c                        ~ smtpctl.8
  ~ smtpctl.c                             ~ smtpd.c
  ~ smtpd.h                               ~ ssl.c
  ~ util.c                                ~ smtpctl/Makefile
  ~ smtpd/Makefile                        

  > backout the "new" queue code commited 4 months ago. it has many good ideas,
  > is way more optimized than what we had earlier and there's definitely stuff
  > we want to keep, however it is early optimization that doesn't account for
  > many features and makes them hard (if not impossible) to write without
  > ugly workarounds that ruin the purpose of the optimizations.
  > the backout goes to 30 May's right before the commit and catches up on all
  > the non-queue related commits that happened since then.
  > i'll work on reintroducing the ideas from this queue when the basic
  > features we expect from a MTA are implemented.
  > suggested on tech@ about a week ago, no objections, several "please make
  > smtpd move forward" mails from hackers and tech readers. (gilles@)

tcpdump

  ~ print-pflog.c                         

  > Do not use BPF_WORDALIGN when computing pflog header length as it aligns
  > to 8-byte boundary on 64-bit architectures. Instead explicitly round up
  > to a 4-byte boundary. Reported and tested by sthen@ (canacar@)

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