OpenBSD src changes summary for 2015-12-15
==========================================

etc/skel/dot.login                      etc/unbound.conf
games/atc                               lib/libcrypto
regress/bin                             sbin/mountd
sys/arch/amd64/amd64                    sys/arch/amd64/include
sys/dev/isa                             sys/net
sys/sys                                 usr.bin/fmt
usr.bin/mandoc                          usr.bin/tmux
usr.sbin/relayd                         usr.sbin/smtpd
usr.sbin/unbound                        usr.sbin/vmd

== etc =============================================================== 01/08 ==

  http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/etc

skel/dot.login

  ~ skel/dot.login                        

  > automatic HOSTALIASES setup is a bad idea (deraadt@)

unbound.conf

  ~ unbound.conf                          

  > add commented-out unbound.conf entries for dns64 (sitting in my tree and
  > ok'd
  > some time ago by phessler and IIRC also mikeb), and for qname-minimisation
  > (sthen@)

== games ============================================================= 02/08 ==

  http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/games

atc

  ~ update.c                              

  > Replace 'arc4random() % range' by 'arc4random_uniform(range)'.
  > Do this by introducing atcrandom_uniform(range) which returns
  > deterministic randomness or good randomness depending on whether
  > a seed was specified with the -r option.
  > Diff by Matthew Martin, reviewed by deraadt@ and me.
  > ok deraadt@ (tb@)

  ~ include.h                             

  > Cleanup: pwd.h is no longer needed. (tb@)

== lib =============================================================== 03/08 ==

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

libcrypto

  ~ cert.pem                              

  > Remove "C=US, O=VeriSign, Inc., OU=Class 3 Public Primary Certification
  > Authority" (serial 3c:91:31:cb:1f:f6:d0:1b:0e:9a:b8:d0:44:bf:12:be) root
  > certificate from cert.pem. ok rpe@
  > Symantec/VeriSign say "Browsers/root store operators are encouraged to
  > remove/untrust this root from their root stores" and "hasn't been used to
  > generate new certificates in several years, and will now be repurposed to
  > provide transition support for some of our enterprise customers' legacy,
  > non-public applications" (https://www.symantec.com/page.jsp?id=roots,
  > http://www.scmagazine.com/google-will-remove-trust-of-symantecs-pca3-g1-cer
  > tificate/article/459688/).
  > Also see
  > https://knowledge.symantec.com/support/ssl-certificates-support/index?page=
  > content&id=ALERT1941
  > https://googleonlinesecurity.blogspot.co.uk/2015/12/proactive-measures-in-d
  > igital.html (sthen@)

== regress =========================================================== 04/08 ==

  http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/regress

bin

  ~ ps/Makefile                           + ps/command.sh
  + ps/shortsleep.c                       

  > regression suite for the "command" keyword (schwarze@)

  ~ ps/Makefile                           + ps/cwd.sh
  + ps/user.sh                            

  > tests for the following keywords: cwd login group rgroup ruser user
  > (schwarze@)

  ~ ps/command.sh                         

  > test vis(3)ing (schwarze@)

  ~ ps/command.sh                         

  > test encoding of a few invalid UTF-8 sequences (schwarze@)

  ~ ps/command.sh                         

  > test some valid UTF-8, but in the C locale (schwarze@)

== sbin ============================================================== 05/08 ==

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

mountd

  ~ mountd.c                              

  > When (re)reading /etc/exports, handle the case where the mount point of an
  > export is not also the mount point of a local filesystem. In this case, the
  > local filesystem would not be removed from the "unexport list", causing the
  > export to be "unexported" again.
  > This fix then also allows us to actually delete an export that has been
  > removed
  > from /etc/exports, by passing the MNT_DELEXPORT export flag to mount(2).
  > Makes sense to millert@. (tim@)

  ~ mountd.8                              

  > Document /var/db/mountdtab (tim@)

  ~ exports.5                             

  > Include RFC number in reference. While here, improve formatting of another
  > reference. (tim@)

== sys =============================================================== 06/08 ==

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

arch/amd64/amd64

  ~ vmm.c                                 

  > support reset vcpu by triple fault (kernel part, userland fix will come
  > later).
  > discussed with deraadt@ and reyk@ at length. (mlarkin@)

  ~ vmm.c                                 

  > remove some not interesting debug printfs (mlarkin@)

arch/amd64/include

  ~ vmmvar.h                              

  > support reset vcpu by triple fault (kernel part, userland fix will come
  > later).
  > discussed with deraadt@ and reyk@ at length. (mlarkin@)

dev/isa

  ~ asmc.c                                

  > make sure the read fan description is always nul terminated
  > hint from kettenis (jung@)

  ~ asmc.c                                

  > add more temperature keys found in newer macbook airs (6,1 and 7,2)
  > encountered via key dump diff, run by Bryan Vyhmeister (jung@)

  ~ asmc.c                                

  > add 2 more temperature keys found in macbook pro (10,2)
  > found via key dump diff, run by Bryan Vyhmeister
  > also remove 3 temperature keys which likely do not return useful values
  > verified with help from macbook pro output of tb (4,1 and 5,5),
  > kettenis (12,1), and Cristoph R. Murauer (9,1) via misc@ (jung@)

  ~ asmc.c                                

  > ignore unlikely temperature values like -127 degC (jung@)

net

  ~ rtable.c                              

  > Do not panic when trying to delete an non-existing route with ART.
  > Reported by bluhm@, ok jmatthew@ (mpi@)

sys

  ~ videoio.h                             + v4l2-common.h
  + v4l2-controls.h                       

  > Sync V4L2 header files with recent Linux kernel.
  > This introduces the two header files v4l2-common.h and v4l2-controls.h
  > which were split off from videodev2.h (which corresponds to our videoio.h).
  > There will be a second commit just addressing whitespace and formatting.
  > The
  > current version has only real changes whereas the second commit will bring
  > videoio.h in line with videodev2.h formatting so that potential syncs in
  > the
  > future are easier to implement.
  > Bulk build by ajacoutot@; ok robert@, ajacoutot@ (feinerer@)

  - v4l2-common.h                         - v4l2-controls.h
  ~ videoio.h                             

  > Revert the previous commit; the licencing needs to be checked very
  > carefully. (feinerer@)

== usr.bin =========================================================== 07/08 ==

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

fmt

  ~ fmt.1                                 ~ fmt.c

  > UTF-8 support; does not yet handle the -c option.
  > No longer expand tabs up front in get_line(), their width depends on the
  > width of characters earlier on the line.  Always NUL-terminate the input
  > buffer for easier and safer handling.  Get rid of the hand-rolled output
  > buffer, just let stdio do its work.
  > OK tedu@ (schwarze@)

  ~ fmt.1                                 

  > missing space in previous; (jmc@)

mandoc

  ~ main.c                                ~ mandocdb.c

  > pledge(2) style:
  > Make sure to always use the idiom 'if (pledge("'
  > such that it can easily be searched for.
  > No functional change.
  > Requested by deraadt@ some time ago. (schwarze@)

tmux

  ~ cmd-find.c                            ~ cmd-select-pane.c
  ~ screen-redraw.c                       ~ server.c
  ~ tmux.h                                ~ window.c

  > Make the marked pane a cmd_find_state. (nicm@)

  ~ cmd-find.c                            

  > Don't copy marked pane when can just point to it. (nicm@)

  ~ cmd-find.c                            

  > We changed somewhat recently to us the pty when tmux was run inside
  > itself to work out the current pane. This is confusing in many cases
  > (particularly notable is that "tmux neww\; splitw" would not split the
  > new window), and the few advantages do not make up for the confusion.
  > So drop this behaviour and return to using the current window and pane;
  > keep the pty check but only use it to limit the list of possible current
  > sessions. (nicm@)

  ~ cmd-list-keys.c                       

  > Allow list-keys and list-commands to be run without a running server.
  > (nicm@)

  ~ cmd-attach-session.c                  ~ hooks.c
  ~ server-client.c                       ~ tmux.h

  > Some hooks API changes to fire a hook while waiting another cmdq and
  > infrastructure that will be needed soon. (nicm@)

  ~ cmd-find.c                            

  > Copy state directly rather than dereferencing wl (which could be NULL).
  > (nicm@)

== usr.sbin ========================================================== 08/08 ==

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

relayd

  ~ relay_http.c                          

  > PATCH is like PUT, data is expected.
  > From mxb at alumni chalmers se (reyk@)

smtpd

  ~ makemap.c                             

  > Sync the DB file once when done with fsync(), not on each write with
  > O_SYNC.
  > The DB file being written is a temp file, so O_EXLOCK is unnecesary.
  > ok sunil@ gilles@ (guenther@)

  ~ smtpd.conf.5                          

  > less macro and lines
  > ok millert (jung@)

unbound

  ~ Makefile.in                           ~ README
  ~ acx_nlnetlabs.m4                      ~ config.h.in
  ~ configure                             ~ configure.ac
  ~ daemon/unbound.c                      ~ daemon/worker.c
  ~ doc/Changelog                         ~ doc/README
  ~ doc/example.conf.in                   ~ doc/libunbound.3.in
  ~ doc/unbound-anchor.8.in               ~ doc/unbound-checkconf.8.in
  ~ doc/unbound-control.8.in              ~ doc/unbound-host.1.in
  ~ doc/unbound.8.in                      ~ doc/unbound.conf.5.in
  ~ iterator/iter_scrub.c                 ~ iterator/iterator.c
  ~ iterator/iterator.h                   ~ libunbound/libunbound.c
  ~ services/cache/rrset.c                ~ smallapp/unbound-anchor.c
  ~ smallapp/unbound-checkconf.c          ~ smallapp/unbound-control.c
  ~ util/config_file.c                    ~ util/config_file.h
  ~ util/configlexer.lex                  ~ util/configparser.y
  ~ util/iana_ports.inc                   ~ util/random.c
  ~ validator/autotrust.c                 ~ validator/val_nsec.c
  ~ validator/val_nsec3.c                 ~ validator/val_secalgo.c
  ~ validator/val_secalgo.h               ~ validator/val_sigcrypt.c

  > update to Unbound 1.5.7, looks fine to deraadt@.
  > Notable addition: implements optional support for qname minimisation
  > (a privacy improvement; see draft-ietf-dnsop-qname-minimisation-07)
  > (sthen@)

vmd

  ~ vmm.c                                 

  > if a vcpu run thread in vmd exits unexpectedly, teardown the VM's kernel
  > components as well.
  > ok deraadt@ (mlarkin@)

===============================================================================
_______________________________________________
odc mailing list
[email protected]
http://www.squish.net/mailman/listinfo/odc

Reply via email to