summary refs log tree commit diff
path: root/bounce.c (follow)
Commit message (Collapse)AuthorAge
* Set client sockets non-blockingJune McEnroe2020-11-16
| | | | | | | | | | | | | | | | | | | | | Except for during writes. This prevents pounce getting blocked on a client sending only a partial TLS record, for example. Writes still need to block because pounce doesn't have a way to resume them. (And it would do so by having a buffer, but sockets already have a send buffer, so what would be the point of that?) I don't think it should be a problem since outside of stateSync, writes only happen when poll returns POLLOUT. I feel like ideally SO_SNDLOWAT would be set to guarantee a full IRC message can always be written on POLLOUT, but since it's actually TLS records being sent, it's not obvious what the size would be. I'm also making an assumption here that tls_read returning TLS_WANT_POLLOUT is unlikely to happen, since I don't actually set pollfd.events based on that. I'm not sure how wanting to resume a tls_read after a POLLOUT could be cleanly handled. I'm just going to hope that if it does happen, the regular poll loop will eventually sort it out...
* Swap localAccept parameter orderJune McEnroe2020-11-14
|
* Only send shutdown QUIT and ERROR to registered clientsJune McEnroe2020-11-14
|
* Make struct Client publicJune McEnroe2020-11-14
|
* Wait for POLLIN to do client tls_handshakeJune McEnroe2020-11-13
| | | | | | | | | | | Otherwise a client could cause pounce to hang (since the sockets are left blocking) by opening a connection without handshaking! Oops, that's pretty bad. Since the sockets are still blocking, a hang can still be caused by a client sending a partial handshake then waiting. More fixes to follow. pounce is slightly protected from this when used with calico, as it applies a timeout to waiting for the ClientHello.
* Report paths in unveil errorsJune McEnroe2020-11-10
|
* Handle signals before the main loopJune McEnroe2020-10-11
| | | | This is a long-standing issue I ignored.
* Fix possibliy uninitialized error 2.0June McEnroe2020-09-09
| | | | It won't be, but gcc thinks it might.
* Rearrange bounce.c, move non-main mains below mainJune McEnroe2020-08-30
|
* Sandbox pounce with unveil(2)June McEnroe2020-08-30
|
* Refactor certificate loading and load all certs from config pathsJune McEnroe2020-08-28
|
* Sandbox pounce with pledge(2)June McEnroe2020-08-27
| | | | | unveil(2) is a bit complicated to apply to this, I'll have to think about it more.
* Add support for OpenBSDJune McEnroe2020-08-27
|
* Remove deprecated option namesJune McEnroe2020-08-25
| | | | The next release will be 2.0 so these can be removed now.
* Use dataOpen for save fileJune McEnroe2020-08-24
|
* Use configOpen to load localCAJune McEnroe2020-08-24
|
* Replace “RAND_bytes” by “getentropy”Issam E. Maghni2020-08-23
| | | | | | This removes the dependency on libcrypto. Signed-off-by: Issam E. Maghni <issam.e.maghni@mailbox.org>
* Implement stub of palaverapp.com capabilityJune McEnroe2020-08-11
| | | | | | This needs to be documented! But the documentation won't make any sense until there's something that can implement the actual functionality of the capability.
* Refactor clientCA and clientSTS as clientCapsJune McEnroe2020-08-10
|
* Say "OpenSSL" in additional permission noticesJune McEnroe2020-08-06
| | | | LibreSSL is "a modified version of that library".
* Fix signing certificates with -A and -gJune McEnroe2020-08-01
| | | | | Always generate a certificate request and pipe it to be signed, either by the CA or by itself.
* Remove compat.hJune McEnroe2020-08-01
|
* Use snprintf instead of strlcatJune McEnroe2020-08-01
|
* Use snprintf instead of strlcpyJune McEnroe2020-07-31
|
* Use RAND_bytes instead of arc4random_bufJune McEnroe2020-07-31
| | | | | This adds an actual dependency on libcrypto, but removes a dependency on BSD (or LibreSSL libcrypto specifically).
* Add -R blind-req optionJune McEnroe2020-07-06
| | | | | Allows requesting userhost-in-names on freenode, which is available but hidden.
* Add additional permission for linking with LibreSSLJune McEnroe2020-06-08
| | | | https://www.gnu.org/licenses/gpl-faq.en.html#GPLIncompatibleLibs
* Add server send queueing with time intervalJune McEnroe2020-05-11
| | | | | | | | | This addresses pounce getting killed with "Excess flood" when it sends NAMES commands for too many channels when a client connects. These commands, as well as automatic AWAY commands, are by default throttled to 5 per second. Tested on freenode with 36 channels and 200ms interval.
* Pass -1 as backlog to listen(2)June McEnroe2020-05-11
| | | | | | There seems to be no guidance on how an application should set this parameter. However, every system I've looked at will limit the value to some default maximum, usually 128.
* Check signals after handling ready socketsJune McEnroe2020-05-11
| | | | | | | In the case where a signal arrives while handling a ready socket, it should be handled as soon as possible, rather than waiting for poll to return again. Signals will still be handled immediately if poll returns -1 due to EINTR.
* Generate short option string from long optionsJune McEnroe2020-04-02
|
* Do not crash on error from acceptJune McEnroe2020-04-02
|
* Replace .mk files with configure scriptJune McEnroe2020-03-01
| | | | Copied and expanded from catgirl.
* Advertise STS policyJune McEnroe2020-02-27
| | | | | Duration is set to INT_MAX since pounce will never accept cleartext connections.
* Implement source address selectionmultiplexd2020-02-18
| | | | | | This commit introduces a '-S' command line option and a "bind" configuration file option for selecting the source address when making outbound TCP connections (similar to the corresponding option in catgirl(1)).
* Set certificate expiry to 10 yearsJune McEnroe2020-01-16
| | | | | I'm pretty sure any kind of "renewing" of these is going to suck, so just set it long enough that the world will probably be ash by then.
* Allow signing by CA in -gJune McEnroe2020-01-12
|
* Rename all local-related options to local-June McEnroe2020-01-12
| | | | | | | This disambiguates client-ca and client-pass from client-cert and client-key, which apply to opposite sides of the program. The old option names will continue to work.
* Use the CAP_PREAD rightJune McEnroe2020-01-12
| | | | Required for the rewind call when loading the CA.
* Add option to set local client CAJune McEnroe2020-01-12
| | | | | | | | | This is a little bit messy. Allows setting either -A or -W or both. Implements SASL EXTERNAL for clients that expect that when connecting with a client certificate. Need to test that reloading still works inside capsicum, since I suspect that rewind call may be blocked.
* Rename -A and -Q to -y and -qJune McEnroe2020-01-07
| | | | | | | | | | | The other upper-case options are related to the listening side of things, not the server side, so this is more consistent. This is incompatible, but will fail loudly, and I expect these options are more likely set in a configuration file, if they are set at all. I also want to free up -A for setting a client CA, but assuming your away message is not also an existing file path, that will continue to fail loudly.
* Revert "Exit when server gets POLLHUP"June McEnroe2019-12-22
| | | | | | | Turns out the more likely thing is that the fd will just continue to be POLLIN and produce zero-length reads. This reverts commit 5707b15920a1ce57f01db0d592487d833218be9d.
* Exit when server gets POLLHUPJune McEnroe2019-12-20
| | | | | | This should maybe gracefully inform clients of what happened, but for now this is much better than the infinite poll loop that happened previously.
* Reference openssl(1) by absolute pathedef2019-11-27
| | | | | This still allows using openssl(1) from PATH, but defaults to using ${LIBRESSL_PREFIX}/bin/openssl.
* Declare globals as extern in headers, and define in source fileMichael Forney2019-11-21
| | | | | | | | | | | | | | | Otherwise, each source file that includes the header gets its own definition, and according to the C standard (C99 6.9p5): > If an identifier declared with external linkage is used in an > expression (other than as part of the operand of a sizeof operator > whose result is an integer constant), somewhere in the entire > program there shall be exactly one external definition for the > identifier Most compilers use the .bss section for zero data, but if it uses .data instead, or if -Wl,--warn-common is used, this will cause a linking error.
* Request NAMES on sync by defaultJune McEnroe2019-11-15
| | | | This inverts the meaning of -N!
* Rename listen to localJune McEnroe2019-11-10
|
* Check that password is hashedJune McEnroe2019-11-08
|
* Only change AWAY status for registered clientsJune McEnroe2019-11-08
| | | | | | Turns out I did eventually fix this, because I may want to implement "passive clients" for logging or notification stuff, which wouldn't affect AWAY status either.
* Include path in readlinkat errorJune McEnroe2019-11-07
|