summaryrefslogtreecommitdiff
path: root/usr.sbin
AgeCommit message (Collapse)Author
2015-02-12Allow constraints URL without leading path (eg. "https://www.openbsd.org").Reyk Floeter
Fixes segfault on configuration load time, as reported by Donovan Watteau.
2015-02-12add missing .Pa macros; patch from Theo Buehler <theo at math dot ethz dot ch>Ingo Schwarze
2015-02-12Rename escape_uri() to url_encode() because it is the opposite ofReyk Floeter
url_decode(). No functional change.
2015-02-12Allow TLS protocols to be specified via a "tls protocols" configurationJoel Sing
option. ok reyk@
2015-02-12Change TLS_PROTOCOLS_DEFAULT to be TLSv1.2 only. Add a TLS_PROTOCOLS_ALLJoel Sing
that includes all currently supported protocols (TLSv1.0, TLSv1.1 and TLSv1.2). Change all users of libtls to use TLS_PROTOCOLS_ALL so that they maintain existing behaviour. Discussed with tedu@ and reyk@.
2015-02-12Use ntpd's deferred DNS resolving for constraints as well. ThisReyk Floeter
allows to get constraint addresses even if network/DNS is not available at startup (or system boot). thumbs up & OK henning@
2015-02-12ber_printf_elements should return NULL if any of its parts fail.Martin Pelikan
Leave the error handling up to its callers. ok reyk
2015-02-11show if we are using a transport security mechanism other than "none"Peter Hessler
OK claudio@
2015-02-11More http status codes.Florian Obser
OK benno@, reyk@
2015-02-11Replace base 10 strtol with strtonum.Florian Obser
OK krw@, tedu@, deraadt@
2015-02-11Use sizeof(u_short) in the first check since there are RT messages thatClaudio Jeker
are less then sizeof(*rtm) bytes long (e.g. interface announcements). Found the hard way by phessler@
2015-02-11Use sizeof(u_short) in the first check since there are RT messages thatClaudio Jeker
are less then sizeof(*rtm) bytes long (e.g. interface announcements). Found the hard way by phessler@
2015-02-11initialize a variable in case "goto done" makes us compare itMartin Pelikan
found by clang, ok henning
2015-02-11Remove dead code (IMSG_HOST_DNS has been moved from the parent toReyk Floeter
ntp_dns some years ago). OK henning@
2015-02-11YPPROC_XFR will trigger a crash, even if using it doesn't make sense.Martin Pelikan
ok deraadt miod
2015-02-11use reallocarray; ok dougTheo de Raadt
2015-02-11Merge in a commit from upstream..Brad Smith
- Fix scrubber with harden-glue turned off to reject NS (and other not-address) records. ok sthen@
2015-02-10Be less chatty on constraint errors.Reyk Floeter
OK deraadt@
2012-03-26Import Unbound 1.4.16 to work on in-tree (not yet linked to the build).Stuart Henderson
These are the direct sources from NLnet Labs upstream, minus these: compat contrib libunbound/python pythonmod testcode testdata winrc ok deraadt@ jakob@
2015-02-10Keep track of the last time we scanned the leases to find expiredKenneth R Westerback
ones that needed to be booted out of the pf table process. This avoids removing the same addess over and over. Problem report and original diff from Bertrand Provost.
2015-02-10pounce on a reallocarray opportunity before teduTheo de Raadt
2015-02-10Merge in a commit from upstream..Brad Smith
- Fix validation failure in case upstream forwarder (ISC BIND) does not have the same trust anchors and decides to insert unsigned NS record in authority section. ok sthen@
2015-02-10Merge in a commit from upstream..Brad Smith
- Fix tcp waiting list for zone transfers where the bind and connect calls fail. ok sthen@
2015-02-10tweak previous;Jason McIntyre
2015-02-10Oops, I accidently reverted the two previous commits in syslog.conf.5.Alexander Bluhm
Bring back revision 1.29.
2015-02-10Make error check consistent in all recvfrom(2) callbacks.Alexander Bluhm
OK henning@
2015-02-10Fix typo in previous commit.Alexander Bluhm
2012-03-26Import Unbound 1.4.16 to work on in-tree (not yet linked to the build).Stuart Henderson
These are the direct sources from NLnet Labs upstream, minus these: compat contrib libunbound/python pythonmod testcode testdata winrc ok deraadt@ jakob@
2015-02-10Document that RFC 5905 (ie. NTP protocol) has a limit which will beTheo de Raadt
hit before Y2038.
2015-02-10spacingReyk Floeter
2015-02-10After successfully getting a constraint from an HTTPS server, there isReyk Floeter
no need to request it ever again. The only exception is the escalation of failed constraint checks that might lead into re-requesting the constraint time from all servers. Adjust the states accordingly. OK henning@
2015-02-10Don't show the subseconds when displaying the constraint offset.Reyk Floeter
OK henning@ deraadt@
2015-02-10Fix -d output of smaller than 64 bit values on big-endian systems.Miod Vallat
ok mpi@ tedu@
2015-02-10Remove a lie that "snmpd does not fully work yet". It works like a charm.Reyk Floeter
2015-02-10Specifying the port is non-optional for (non-TLS) syslog over TCP, ok bluhm@Stuart Henderson
2015-02-10redo the tolower/string conversion, but retain itsabbr which is strangelyTed Unangst
not quite the same as strncasecmp
2015-02-10More SOCK_NONBLOCK adjustments similar to the other daemons.Claudio Jeker
2015-02-10Forgot the EAGAIN or EINTR check on read.Claudio Jeker
2015-02-10Sync kroute code with bgpd/ospfd code regarding EAGAIN and short readsClaudio Jeker
2015-02-10Sync kroute code with what we do in ospfd/bgpd (EAGAIN handling).Claudio Jeker
2015-02-10Encode directory listings.Florian Obser
Problem pointed out by remco AT d-compu.dyndns.org some time ago. Input / OK reyk@
2015-02-10Move the constraints in a new section and add a preamble to explainReyk Floeter
the functionality. Requested by henning@ OK beck@ deraadt@
2015-02-10Add support for "constraints": when configured, ntpd(8) will query theReyk Floeter
time from HTTPS servers, by parsing the Date: header, and use the median constraint time as a boundary to verify NTP responses. This adds some level of authentication and protection against MITM attacks while preserving the accuracy of the NTP protocol; without relying on authentication options for NTP that are basically unavailable at present. This is an initial implementation and the semantics will be improved once it is in the tree. Discussed with deraadt@ and henning@ OK henning@
2015-02-10be more verbose when logging privsep errors.Brent Cook
ok phessler@ deraadt@
2015-02-10use correct formatters for s/size_t data types.Brent Cook
ok deraadt@
2015-02-10revert rev 1.12 (tolower conversion). didn't get it quite right.Ted Unangst
2015-02-10move scheck.c into zic.cTed Unangst
2015-02-10Same session_socket_blockmode() changes as done to ospfd. Also do the sameClaudio Jeker
kroute change (make socket non-blocking and add trigger for partial reads).
2015-02-10Convert ospfd over to SOCK_CLOEXEC | SOCK_NONBLOCK and make the routeClaudio Jeker
socket non-blocking. Introduce the same trigger for partial rt msgs.
2015-02-10Make also the special sockets SOCK_NONBLOCK. For the routing socket addClaudio Jeker
a trigger for the case that not a full message has been read. Should not be possible but lets see if this triggers somewhen. With and OK henning@