summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2019-06-20Still need to walk the options tree for user options.Nicholas Marriott
2019-06-20prefix_updateall() is only used internally, make it a static function.Claudio Jeker
2019-06-20Change nexthop_update to run the list walk over all prefixes to runClaudio Jeker
asynchronously and therefor other tasks can make progress at the same time. Additionally prefixes belonging to a RIB which does not run the the decision process are no longer linked into the nexthop list. This replaces the early return in prefix_updateall() and reduces the time spent in nexthop_update(). OK benno@
2019-06-20tweaks with help from jmc@Otto Moerbeek
2019-06-20Add a per-pane option set. Pane options inherit from window options (soNicholas Marriott
there should be no change to existing behaviour) and are set and shown with set-option -p and show-options -p. Change remain-on-exit and window-style/window-active-style to be pane options (some others will be changed later). This makes select-pane -P and -g unnecessary so no longer document them (they still work) and no longer document set-window-option and show-window-options in favour of set-option -w and show-options -w.
2019-06-20Tell a bit about automatic mode; ok jmc@Otto Moerbeek
2019-06-20Add a helper function to work out option table from name.Nicholas Marriott
2019-06-20Do a quick DNS probe to decide to stay in the forground and attemptOtto Moerbeek
an (auto) settime or give up. 15s timeout is still in effect. ok florian@
2019-06-20Add a -A flag to show-options to show parent options as well.Nicholas Marriott
2019-06-20Expand command formats in %if and move the config file loading later (toNicholas Marriott
when the first client has identified) so all the client formats are available, fixes problems reported by Thomas Sattler.
2019-06-20Work around locking issues in logwakeup(). Instead of actually waking upVisa Hankala
waiters, just set a flag in logwakeup(). The flag is later noted through periodic polling. This lets the wakeup code run with sufficient locking. logwakeup() is a very tricky place to take locks because the function can be called in many different contexts. By not requiring locks in the routine helps to keep printf(9) as usable as possible. OK mpi@
2019-06-20auth_approval() returns 0 on failure, so don't test for <= 0Theo de Raadt
(error has no effect. just striving for an accurate following of contract) ok millert
2019-06-19from tim:Jason McIntyre
- for reput, it is remote-path which is optional, not local-path - sync help from deraadt: - prefer -R and undocument -r (but add a comment for future editors) from schwarze: - prefer -p and undocument -P (as above. the comment was schwarze's too) more: - add the -f flag to reput and reget - sort help (i can;t remember who suggested this originally) djm and deraadt were ok with earlier versions of this; tim and schwarze ok
2019-06-19Allow string greater/less than than operators to work with test aka [.Todd C. Miller
Previously they were only recognized in [[ ... ]] expressions. This changes sh/ksh to be consistent with test(1) as well as shells like bash and dash. OK jca@ jmc@
2019-06-19the pledge STATLIE code is no longer needed, as discussed with beck.Theo de Raadt
it actually isn't reached...
2019-06-19The main process does not need to toggle the sockets from blocking toClaudio Jeker
nonblocking just for poll(). OK deraadt@
2019-06-19Close proc fd used to talk to proc_parser when spaning the rsync processClaudio Jeker
else the fd leaks into the rsync process. OK deraadt@
2019-06-19use $OpenBSD$ headersTheo de Raadt
2019-06-19repair a cluster of "< 0" error checks for libc / syscalls, whichTheo de Raadt
should be "== -1"
2019-06-19Fix regression introduced in rev. 1.59 where the wrong shell isTodd C. Miller
checked in -m mode for non-root. The intent is that non-root shouldn't be able to run their own shell as the target user if the target user has a shell that is not listed in /etc/shells. This makes the code match the manual once again. OK schwarze@
2019-06-19syncTheo de Raadt
2019-06-19move tals to /etc, where they can be upgraded by a "sysupgrade" ifTheo de Raadt
such a circumstance ever occurs. ok job
2019-06-19indentTheo de Raadt
2019-06-19use asprintf, rather than an intermediate bufferTheo de Raadt
ok claudio
2019-06-19Add another test file, this way CERT_IP_RANGE is also tested.Claudio Jeker
2019-06-19Adjust a bit since string representation of IPv6 addrs changed.Claudio Jeker
2019-06-19Instead of the strange ASN.1 encoding of IP addresses with a size in bytesClaudio Jeker
and a count of how many bits of the last byte are unused switch to a much more normal prefixlen in bits encoding for internal use. This makes the code easier to understand. Also switch to inet_ntop() for converting a struct ip_addr into a string. OK job@
2019-06-19rpki-client(1) -> rpki-client(8)job
2019-06-19use REGRESS_EXPECTED_FAILURESMarc Espie
okay bluhm@
2019-06-19use new REGRESS_EXPECT_FAILURESMarc Espie
okay bluhm@
2019-06-19Adjust file glob so that CVS is not picked up anymore.Claudio Jeker
2019-06-19more precisely describe what happens to the environment without keepenv;Ingo Schwarze
OK tedu@
2019-06-17Import Kristaps Dzonsons' RPKI validator into the treejob
rpki-client(1) is an implementation of the Resource Public Key Infrastructure (RPKI), specified by RFC 6480. The client is responsible for downloading, validating and converting Route Origin Authorisations (ROAs) into Validated ROA Payloads (VRPs). The client's output (VRPs) can be used by bgpd(8) to perform BGP Origin Validation (RFC 6811). The current rpki-client(1) version depends on the CMS functions in OpenSSL, this of course needs to be addressed urgently. Thanks to NetNod, IIS.SE, SUNET & 6connect for supporting this effort! OK deraadt@
2019-06-19mention that doas(1) resets the umask(2);Ingo Schwarze
OK tedu@
2019-06-19Improve terminology usedjob
2019-06-19With the community rewrite the output code needs to also check theClaudio Jeker
communities member of struct prefix to decide if prefixes can be put together into the same update. This bug results in prefixes being sent out with the wrong comminities.
2019-06-19Add checks to copyin(9), copyout(9), copyinstr(9) and copyoutstr(9)Patrick Wildt
to make sure that we don't even attempt to access user-provided addresses that look like kernel space. ok kettenis@
2019-06-19Add rpki-clientClaudio Jeker
2019-06-19This currently needs the openssl-1.0.2 package to run. So skip test ifClaudio Jeker
the package is not installed.
2019-06-19indentation adjustments, in particular near warn statementsTheo de Raadt
ok claudio
2019-06-17Import Kristaps Dzonsons' RPKI validator into the treejob
rpki-client(1) is an implementation of the Resource Public Key Infrastructure (RPKI), specified by RFC 6480. The client is responsible for downloading, validating and converting Route Origin Authorisations (ROAs) into Validated ROA Payloads (VRPs). The client's output (VRPs) can be used by bgpd(8) to perform BGP Origin Validation (RFC 6811). The current rpki-client(1) version depends on the CMS functions in OpenSSL, this of course needs to be addressed urgently. Thanks to NetNod, IIS.SE, SUNET & 6connect for supporting this effort! OK deraadt@
2019-06-19swap comparisonsTheo de Raadt
2019-06-19Move variables into struct in openssl(1) genrsaKinichiro Inoguchi
- Move local variables in genrsa_main() to struct genrsa_config - Leave long lines more than 80, still ok bcook@
2019-06-18Stop probing a Tx rate if transmission failure or too many retries occur.Stefan Sperling
Fixes stalling TCP connections on wifi in some situations. testing jmatthew@, jmc@, Bruno Flueckiger, Micah Muer, Matthias Schmidt
2019-06-18When the parser detects that it needs more tokens it returnsFlorian Obser
JSMN_ERROR_NOMEM. We then need to allocate more tokens and call the parser with its current state again. It will continue where it left of. For this to work we also need to pass in the old tokens from the previous run and not just more space. Found the hard way by Renaud Allard. OK millert
2019-06-18Ensure that timeout p_sleep_to is not left running when finishing sleep.Visa Hankala
This is necessary when invoking sleep_finish_timeout() without the kernel lock. If not cancelled properly, an already running endtsleep() might cause a spurious wakeup on the thread if the thread re-enters a sleep queue very quickly before the handler completes. The flag P_TIMEOUT should stay cleared across the timeout cancellation. Add an assertion for that. OK mpi@
2019-06-18do not need to allocate memory for execvp argumentsTheo de Raadt
2019-06-18Use correct terminilogy in rpki-client manpagejob
2019-06-18Use the test-files from rpki-client and some files from the rpki cache toClaudio Jeker
implement a basic regress test. Needs more work but should be a start.
2019-06-18Remove tests here and move them to regress. At least then they can useClaudio Jeker
the existing framework. OK job@
generated by cgit v1.2.3 (git 2.25.1) at 2025-01-03 15:18:45 +0000