Commit Graph

96 Commits (9e07364ad389ac21dad1fc8ae7f6b1b7b39bd6bc)

Author SHA1 Message Date
leitner 421df2174c systems without fastopen support return EPIPE
leitner 1137831ddd make fastopen fall back to connect if len==0
leitner ed67042965 undo freebsd fix, does not compile
leitner 2cc0eec6cc #define _POSIX_SOURCE for FreeBSD
leitner 61db2acee6 no :: at the end if it's only :0
leitner a4a2aed3de fmt_ip6 was violating rfc5952 section 4.2.2
leitner 760598ec2f #include <foo.h> -> #include <libowfat/foo.h>
leitner a7f74699da add #include <string.h> for memset/strcpy
leitner 79cc92ec6f remove warning for crippled operating systems
leitner d26b8082d8 shut up new gcc 6 warnings
leitner 83712b7f23 attempt to support FreeBSD
leitner 189d8bad1d darwin compat
leitner f74f9b9b25 add socket_udp4b and socket_udp6b
leitner 7bdc1b29bb a few missing headers and minor fixes
fix corking behavior for the case where a buffer follows a buffer in
iob_send
leitner 695843c79e small buffer optimization
fix botched noipv6 case in buffer_tcp6
leitner d0e735e2c2 support noipv6 being set externally even when there is perfectly good
IPv6 support.
leitner 1856c8a2d7 damn it, fucked up the return value
leitner 6e022f98c9 ... but if errno==EINVAL still fall back to socket+fcntl (Robert Henney)
leitner 35f5a9692a if SOCK_NONBLOCK is defined, use it instead of socket+fcntl
leitner a691887e75 forgot to check in the actual code m(
leitner 88167b5ce1 add socket_fastopen, socket_fastopen_connect4, socket_fastopen_connect6 and socket_quickack
leitner dc4d204f4e also mention scan_ip6if(3)
leitner 8581b74a84 on FreeBSD, on a PF_INET6 socket, recvfrom and friends can actually
return a sockaddr with family PF_INET.  WTF?  Work around that.
leitner 53e2af7389 never close fd 0, 1 or 2 in buffer_close
leitner 4a04c40595 get rid of more -Wconversion warnings
leitner c099365ec7 mingw compile fix
leitner 82e6b4086b socket_remote4 does not need havescope.h
leitner 2cb6a12c92 #include <errno.h>
leitner b4469f1ed8 support SCTP in addition to TCP
leitner 78306e451e typo
leitner 3636c04609 use SO_REUSEPORT
leitner 67bfd53652 check for socket() returning -1 before trying to set the socket
non-blocking so we don't hide the "too many open files" error code
leitner c114890119 *cough*
leitner d4cdc92b4a remove warning
leitner 22f88f6550 add socket_deferaccept
leitner b0d5046428 fmt_ip6 compresses at best spot, not at first spot (Nikola Vladov)
leitner 0cda5dacc8 immer noch falsch
leitner 8420a7a724 fix man pages
leitner 3004b518ef switch to size_t and ssize_t
leitner 83c1186870 fix for NetBSD 3.0 (they broke their sys/types.h)
remove support for ip6.int reverse lookup in DNS routines (the domain is
no longer delegated)
leitner 144fb41918 remove warnings
leitner d361d81c64 make socket_(tc|ud)p[46] actually return non-blocking sockets as
documented (Richard Lyons)
leitner c1c50c7dbd further Windoze support (test/io5.c works, gatling still doesn't)
This is just to get gatling to work, I may remove it again after
    that.
leitner f141d3e24f socket_accept4 should cope with being passed NULL pointers
leitner 10658ebe60 only write if dest is non-NULL
leitner eeffc36387 remove even more warnings
64-bit cleanliness issue with auto-NULL-appending macros
leitner 6196f771cd remove a few gcc 4 warnings
work around freebsd 5.4 brokenness (if you don't have IPv6 in the
    kernel, socket(PF_INET6,SOCK_STREAM,0) returns EPROTONOSUPPORT
    instead of EPFNOSUPPORT, which basically says "yeah, I know IPv6,
    but TCP?  never heard of it")
leitner 9210337fa5 turns out there is one more error message that can signal missing IPv6
support in the kernel
leitner d6b7c29808 uh, the scope_id detection #defined the wrong constant. libowfat
always thought there was no scope_id.  Oops.
leitner 8873a442f2 man page update