summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2021-04-25Remove old mman headerDirk Engling
2021-04-25Turn mmaps into mallocDirk Engling
2021-04-25Retry successful writes immediatelyDirk Engling
2021-04-24Split huge iovecs over multiple io_batchesDirk Engling
2021-04-22gzip iovecs always end on the boundary and don't need to be fixedDirk Engling
2021-04-21Add a handler for the EAGAIN reason code to iob_sendDirk Engling
2021-04-21Rework fullscrape worker, unifying non-gzip and gzip code was a bad ideaDirk Engling
2021-04-20accesslist checker should not operate on an empty listDirk Engling
2021-04-20srandomdev is not available on linuxes by defaultDirk Engling
2021-04-20Use arc4random whereever we need strong entropyDirk Engling
2021-04-20Use dev random by defaultDirk Engling
2021-04-20Turn random() to nrand48 in inner loop where it is not cryptographically relevant to avoid futex storm. Thanks and credits to Bart SmienkDirk Engling
2021-04-20White spaceDirk Engling
2021-04-19unmap file if a memory allocation error occursDirk Engling
2021-04-19Make accesslist reload thread safe. The last commit actually would make a free possible while another thread was bsearching that memoryDirk Engling
2021-04-19Try accessing the access lists without locks by making the replacement process as atomic as possible.Dirk Engling
2021-04-19De-bottleneck mutex access codeDirk Engling
2018-05-26Remove debug code for self pipesDirk Engling
2018-05-26Tell libowfat about EWOULDBLOCK conditionsDirk Engling
2018-01-29register both ends of self pipe to libowfat in order to use themDirk Engling
2017-04-28break belongs inside the guardDirk Engling
2017-04-28incorporate a more verbose ascii dump, provided by Tom <tom@foscore.com>Dirk Engling
2017-04-28Get rid of some warnings regarding header guardsDirk Engling
2016-10-02Avoid double free when iob_addbuf_free fails, thanks to Sami FarinDirk Engling
2015-12-11Forgot to include one error code. Thanks to Lars SeipelDirk Engling
2015-11-26dont count scrapes as full scrapes, fall through failDirk Engling
2015-11-25Accept was counted twiceDirk Engling
2015-08-30Add missing enum value. Thanks to zhangdongmaoDirk Engling
2015-06-08Increase number of locks to accomodate for more cores on contemporary CPUsDirk Engling
2015-06-05Avoid reacting to our own error packets, should the happen to end up on our interface with a spoofed source IP. Thanks to zorun.Dirk Engling
2015-04-27Fix typo in stat reporting, use udp and not tcp counter. Thanks to zebulonDirk Engling
2014-10-07Protect static data by mutexerdgeist
2014-10-07Only kick off udp handlers, when everything is initialisederdgeist
2014-10-07Initialise all values of aes keyerdgeist
2013-08-04Fix immediate overwrite in the X-forwarded-for passing case. Thanks to Alexander Luetjen.erdgeist
2013-06-07Fix spelling error ;)erdgeist
2012-06-08Dont make srandomdev the default, it exists only in BSDserdgeist
2012-06-03Make woodpecker spotting more usable by adjusting count every 2 minuteserdgeist
2012-06-03Calculate the old hash for the ip address only, when the current one mismatcheserdgeist
2012-06-03Dont ruin the srandomdev initialisation in trackerlogic's initerdgeist
2012-05-31Forgot variable declarationerdgeist
2012-05-29reduce zlib warnings to when there really is something failingerdgeist
2012-05-29new flag to enforce gzip even if client did not request itdenis
2012-05-29Add rijndael to version reporterdgeist
2012-05-28Remove debug infoerdgeist
2012-05-28Report connection id missmatch counterdgeist
2012-05-28udp now generates a cryptographically secure token for connecting clients. This is later verified.erdgeist
2012-05-28Init prng before udp code needs it.erdgeist
2012-05-28Add connection id missmatch codeerdgeist
2012-05-28Adding rijndael code for udp connection id calculationerdgeist