| Age | Commit message (Expand) | Author |
|---|
| 2015-10-29 | put timeout nullification in correct spot | Ted Unangst |
| 2015-10-29 | be a little more precise about checking filters | Ted Unangst |
| 2015-10-29 | refold a few lines | Ted Unangst |
| 2015-10-29 | collect some cool stats and print them out with SIGUSR1 | Ted Unangst |
| 2015-10-28 | impose some limit on the cache size as well. | Ted Unangst |
| 2015-10-28 | if accept() fails due to fd exhaustion, stop accepting for one second. | Ted Unangst |
| 2015-10-28 | bug in the request tree code: requests were being added to the tree before | Ted Unangst |
| 2015-10-28 | twiddle with rlimit to give us a few more files. use rlimit as a hint as | Ted Unangst |
| 2015-10-28 | add pledge to main process, though it needs a few more options | Ted Unangst |
| 2015-10-28 | use an rb tree for the cache too. modelled after the reqtree. | Ted Unangst |
| 2015-10-28 | use an rb tree for finding requests. from Dimitris Papastamos | Ted Unangst |
| 2015-10-26 | one second amnesty for timeouts so we don't spin with short timeouts | Ted Unangst |
| 2015-10-26 | set listening sockets to nonblocking, then handle as many incoming requests | Ted Unangst |
| 2015-10-17 | don't need fcntl for non blocking socket, just ask for it upfront | Ted Unangst |
| 2015-10-16 | save some file descriptors. instead of a pipe, use kevent to watch parent | Ted Unangst |
| 2015-10-16 | naddy would like the child to exit when the parent dies. | Ted Unangst |
| 2015-10-16 | life is simpler if all requests go in the fifo, and then just remove them | Ted Unangst |
| 2015-10-16 | deraadt tells me i'm supposed to check if connect() actually worked. | Ted Unangst |
| 2015-10-16 | two phase handling for tcp so that slow connects don't stall the process | Ted Unangst |
| 2015-10-16 | save request length in cache. naddy noticed we weren't getting any hits. | Ted Unangst |
| 2015-10-16 | fine tune the logging some more | Ted Unangst |
| 2015-10-16 | simplify logging functions. once a daemon, always a daemon | Ted Unangst |
| 2015-10-16 | safety check that we're dealing with the filter we expect | Ted Unangst |
| 2015-10-16 | most things should be static | Ted Unangst |
| 2015-10-16 | exit(1) is better for the impossible condition | Ted Unangst |
| 2015-10-15 | it is perhaps better style to not call close() on -1, even if harmless | Ted Unangst |
| 2015-10-15 | make sure req is zeroed in tcp case | Ted Unangst |
| 2015-10-15 | better memory handling of the request/cache chain | Ted Unangst |
| 2015-10-15 | do not insert entry into cache until it's fully formed | Ted Unangst |
| 2015-10-15 | doh, not all requests are the same size. check len first. | Ted Unangst |
| 2015-10-15 | assert is the wrong tool | Ted Unangst |
| 2015-10-15 | the inet sockets don't work well with inet6 addrs. pick family from addr. | Ted Unangst |
| 2015-10-15 | introduce logerr, since most logging is followed by exit | Ted Unangst |
| 2015-10-15 | now with _rebound user, we can try a little harder at privdrop | Ted Unangst |
| 2015-10-15 | make the HUP interlocking in the parent work better. | Ted Unangst |
| 2015-10-15 | trivial KNF | Theo de Raadt |
| 2015-10-15 | just a space in usage, from deraadt | Ted Unangst |
| 2015-10-15 | add a hint about the config file, until it changes | Ted Unangst |
| 2015-10-15 | child can be pledged down a bit to just sockets and io | Ted Unangst |
| 2015-10-15 | import rebound, a lightweight dns proxy, for further polishing | Ted Unangst |
