Age | Commit message (Collapse) | Author | |
---|---|---|---|
2017-01-25 | Hopefully the last of the struct rib rototilling. Peer just points to a | Claudio Jeker | |
struct rib and not rib_desc since the full descriptor is almost never needed. This should now allow the update code to be changed. | |||
2017-01-25 | Switch rde_generate_update and rde_send_kroute to accept a struct rib instead | Claudio Jeker | |
of the id. For this we move the rtableid into struct rib. Also move the update code in rib.c up to where the kroute code is. Makes more senses like that. | |||
2017-01-25 | Flag the Loc-RIB with F_RIB_LOCAL so we can remove one ugly hack somewhere else | Claudio Jeker | |
2017-01-24 | Save some space in struct rib_entry so it is back to 64bytes (on 64bit archs). | Claudio Jeker | |
Doing this by folding the lock flag into a pointer and providing an accessor function for the rib pointer. This is an acceptable middle path for this important structure. OK benno@ on an earlier version | |||
2017-01-24 | do not need sys/proc.h | Theo de Raadt | |
2017-01-24 | Tell traceroute6 to choose source address based on correct routing | Florian Obser | |
table. From dhill, tweeks by me. OK benno | |||
2017-01-24 | Make returns consistent and remove parentheses per style(9). | Joel Sing | |
Transformed with coccinelle. Requested by and ok tb@ | |||
2017-01-24 | sort SEE ALSO; | Jason McIntyre | |
2017-01-24 | nuke extra brackets. discussed with jsing | Theo Buehler | |
2017-01-24 | Complete jsing's coccinelle cleanup... by hand | Theo de Raadt | |
ok jsing | |||
2017-01-24 | correct usage format; ok beck claudio benno | Theo de Raadt | |
2017-01-24 | errx() does not need extra newline | Theo de Raadt | |
2017-01-24 | add ocspcheck to see also | Bob Beck | |
2017-01-24 | fix mode on open() and ftruncate(), noticed by | Bob Beck | |
bcook@ | |||
2017-01-24 | hook ocspcheck into the build | Bob Beck | |
2017-01-24 | Say no to two line error messages on failure | Bob Beck | |
2017-01-24 | Replace comparisons between a constant or enum and an expression, with | Joel Sing | |
a comparison between the expression and the constant or enum. This significantly improves readability. Transformed with coccinelle. Requested by deraadt@ | |||
2017-01-24 | s/returns/exits/ | Bob Beck | |
2017-01-24 | Break run-on sentence into two. | Bob Beck | |
2017-01-24 | sort previous; | Jason McIntyre | |
2017-01-24 | string terminator is called a NUL | Theo de Raadt | |
2017-01-24 | Actually load the cafile when providede, and error message cleanup | Bob Beck | |
2017-01-24 | move the opening of /dev/pf from the parent process to the pfe process | Sebastian Benoit | |
where it is used. Currently pf is opened on every reload, that will no longer be possible in the future with pledged programms that do ioctls. This prepares relayd for that change. ok deraadt@, meinetwegen reyk@ | |||
2017-01-24 | use warn, I have errno here. noticed by theo | Bob Beck | |
2017-01-24 | Yes the "if (const == val" idiom provides some safety, but it grates on | Theo de Raadt | |
us too much. ok beck jsing | |||
2017-01-24 | knf | Bob Beck | |
2017-01-24 | revert accidental commit of theo diff | Bob Beck | |
2017-01-24 | Just don't bother with OpenSSL error strings, they are mostly | Bob Beck | |
irrelevant and look gross here anyway.. we don't need them | |||
2017-01-24 | SVM: vmd(8) support for SVM | Mike Larkin | |
2017-01-24 | Add a workaround for the fact that 'install -d' without explicit mode | Theo Buehler | |
defaults to 755, which is incompatible with WOBJUMASK. Make sure that obj/util has permissions :wobj 770, as all other directories in obj/. Issue also found by jmc, actual reason tracked down by ajacoutot. ok ajacoutot | |||
2017-01-24 | various cleanup; | Jason McIntyre | |
2017-01-24 | slight cleanups | Theo de Raadt | |
2017-01-24 | s/exit/exist/ typo | Bob Beck | |
2017-01-24 | mark start -m as optional; reorder start synopsis to match vmctl usage | Richard Procter | |
ok mlarkin@ | |||
2017-01-24 | New ocspcheck utility to validate a certificate against its ocsp responder | Bob Beck | |
and save the reply for stapling ok deraadt@ jsing@ | |||
2017-01-24 | string terminator is called a NUL | Theo de Raadt | |
2017-01-24 | vmctl start -k is optional; also emphasise it specifies a path on the host | Richard Procter | |
ok mlarkin@ | |||
2017-01-24 | sync log.c from relayd et al to ospfd. | Sebastian Benoit | |
there is still a little difference regarding handling of the verbosity value that will be handled later. ok claudio@ | |||
2017-01-24 | sync log.c from relayd et al to bgpd. | Sebastian Benoit | |
there is still a little difference regarding handling of the verbosity value that will be handled later. ok claudio@ florian@ | |||
2017-01-24 | remove unused junk definitions | Theo de Raadt | |
2017-01-23 | Mission accomplished, bgpctl is now fatal free. | Claudio Jeker | |
2017-01-23 | Kill the timer.c reacharound it is not needed. | Claudio Jeker | |
2017-01-23 | cleanup log* functions that break with bgpd reacharound | Sebastian Benoit | |
ok claudio@ | |||
2017-01-23 | Make util.c fatal() free by allowing undefined behaviour in prefix_compare. | Claudio Jeker | |
If you pass in crap then you will not get gold back. | |||
2017-01-23 | Rename rib pointer in struct prefix to re since it points to a rib_entry. | Claudio Jeker | |
While there also remove a comment that is since a few years at least. OK gcc | |||
2017-01-23 | Introduce a struct rib sitting between struct rib_desc and struct rib_tree. | Claudio Jeker | |
This way the tree becomes a bit better decoupled. | |||
2017-01-23 | Revert the struct rib_tree rename. I need a struct in between because of | Claudio Jeker | |
how struct rib_entry is used. | |||
2017-01-23 | Build on arm64. | Mark Kettenis | |
2017-01-23 | More rototilling, make rib_new and rib_find return a point to struct rib_desc | Claudio Jeker | |
2017-01-23 | Now rename struct rib_tree to struct rib. Again OK gcc | Claudio Jeker | |