Andrew Tridgell
63fc397808
tsocket: not all systems have IPV6_V6ONLY
2010-03-26 16:50:48 +11:00
Andreas Schneider
1983256c70
tsocket: Improve the tsocket_address_bsd_sockaddr documentation.
...
Signed-off-by: Stefan Metzmacher <metze@samba.org>
2010-02-26 08:57:30 +01:00
Brad Hards
319d3bf9d9
Spelling fixes for tsocket API documentation.
...
Signed-off-by: Matthias Dieter Wallnöfer <mwallnoefer@yahoo.de>
2010-02-22 21:45:33 +01:00
Stefan Metzmacher
c42d9c4ec4
tsocket/bsd: fix comment in tdgram_bsd_recvfrom_handler()
...
metze
2010-02-18 09:38:00 +01:00
Jeremy Allison
936828de71
Fix commit d07cd37b99
...
Which was:
tsocket/bsd: fix bug #7115 FreeBSD includes the UDP header in FIONREAD
Metze, this has to have been wrong - you are throwing away the talloc_realloc
pointer returned. Also no error checking. Please review.
Thank goodness for gcc warnings :-).
Jeremy.
2010-02-17 09:24:34 -08:00
Stefan Metzmacher
d07cd37b99
tsocket/bsd: fix bug #7115 FreeBSD includes the UDP header in FIONREAD
...
metze
2010-02-17 14:46:39 +01:00
Stefan Metzmacher
1ffcb991a9
tsocket/bsd: set IPV6_V6ONLY on AF_INET6 sockets
...
Some system already have this as default. It's easier
to behave the same way on all systems and handle ipv6
and ipv4 sockets separate.
metze
2010-02-17 14:46:39 +01:00
Stefan Metzmacher
8a0949dfc8
tsocket/bsd: fix bug #7140 autodetect ipv4 and ipv6 based on the remote address if the local address is any
...
metze
2010-02-17 14:46:08 +01:00
Stefan Metzmacher
6637b2f4b0
tsocket/bsd: fix bug #7140 use calculated sa_socklen for bind() in tstream_bsd_connect_send()
...
This is needed because, we can't use sizeof(sockaddr_storage) for AF_UNIX
sockets. Also some platforms require exact values for AF_INET and AF_INET6.
metze
2010-02-17 14:45:34 +01:00
Stefan Metzmacher
135543b4c3
tsocket/bsd: fix do_bind logic for AF_INET
...
We want the explicit bind() when we don't use the any address.
metze
2010-02-17 14:13:57 +01:00
Kai Blin
88099bcc93
tsocket: Fix the description of tstream_writev_queue_send/recv
2010-01-28 11:16:24 +01:00
Kai Blin
9184f524f0
tsocket: Fix description for tstream_readv_pdu_queue_send/recv
2010-01-28 11:11:33 +01:00
Andreas Schneider
04f3f27f7e
tsocket: Added doxygen config file.
...
Signed-off-by: Stefan Metzmacher <metze@samba.org>
2010-01-11 11:23:04 +01:00
Andreas Schneider
574f77f89e
tsocket: Added complete doxygen documentation.
...
Signed-off-by: Stefan Metzmacher <metze@samba.org>
2010-01-11 11:22:59 +01:00
Stefan Metzmacher
17c804a675
tsocket: only copy the specific part of sockaddr*
...
This makes sure we fill the unspefic bytes with 0
and have zero termination for sockaddr_un.
metze
2009-12-24 17:38:29 +01:00
Stefan Metzmacher
4784c8c3be
tsocket: allow AF_UNIX sockaddrs smaller than sizeof(sockaddr_un)
...
This is needed as getpeername() truncates the length to 2 if the
peer socket has no name.
metze
2009-12-24 17:38:04 +01:00
Andreas Schneider
a5bdab8b32
tsocket: Fixed tsocket_guide.txt asciidoc syntax.
2009-12-15 23:34:23 +01:00
Andreas Schneider
e33ae22b87
tsocket: Fixed documentation for tsocket_address_bsd_sockaddr().
2009-12-15 23:34:22 +01:00
Andreas Schneider
ab04dd3be7
tsocket: Fixed typo in LGPL header.
2009-12-15 23:34:22 +01:00
Stefan Metzmacher
edbf2caa6f
tsocket: add tsocket_address_bsd_sockaddr() and tsocket_address_bsd_from_sockaddr()
...
metze
2009-12-15 23:34:22 +01:00
Stefan Metzmacher
97a32035be
tsocket: rewrite tsocket_guide.txt to reflect the current APIs
...
metze
2009-11-03 18:26:30 +01:00
Stefan Metzmacher
eb39f66940
tsocket_bsd: return -1 and set errno to ENAMETOOLONG if the unix path is too long
...
metze
2009-11-03 18:24:44 +01:00
Stefan Metzmacher
9b5d1e9e35
tsocket: remove prototype of non-existing tsocket_address_inet_set_broadcast()
...
metze
2009-11-03 18:14:23 +01:00
Stefan Metzmacher
1f12dc4409
tsocket: rename sa_len => sa_socklen, because sa_len is a macro on some platforms
...
metze
2009-07-16 08:21:08 +02:00
Stefan Metzmacher
857ed9ca5a
tsocket/bsd: also ask for TEVENT_FD_READ when we want to write into a stream socket
...
Otherwise we would not notice a broken connection.
metze
2009-06-29 16:03:57 +02:00
Stefan Metzmacher
b1d5e515b2
tsocket/bsd: more correctly check if the cached tevent_fd is still valid
...
I some cases the pointer value of tevent_context is the same again,
if we do something like:
ev1 = tevent_context_init();
...
fde = tevent_add_fd(ev1, fd, TEVENT_FD_READ...);
...
talloc_free(ev1);
...
ev2 = tevent_context_init();
if (ev1 == ev2) {
/* this can happen! */
}
if (tevent_fd_get_flags(fde) == 0) {
/* this is always true */
}
But the "talloc_free(ev1)" will set fde->event_ctx to NULL
and tevent_fd_get_flags() will always return 0.
metze
2009-06-29 16:03:57 +02:00
Volker Lendecke
c687eb15d9
Properly free a downlevel readv request. Metze, please check
2009-06-27 23:17:49 +02:00
Andrew Tridgell
c0965e0eec
It seems that IRIX doesn't have IOV_MAX
2009-06-09 14:54:15 +10:00
Stefan Metzmacher
edd9bd9b16
tsocket: allow empty vectors at the end for tstream_writev()/readv()
...
metze
2009-05-22 14:01:22 +02:00
Volker Lendecke
ac65ae9b1d
Set errno=ENOMEM if tevent_add_fd fails
...
tevent_add_fd does not properly set that. At least in epoll and select this is
the only error condition.
Metze, please check!
2009-05-19 23:51:22 +02:00
Volker Lendecke
f9d00fb58e
Simplify the logic of tsocket_bsd_pending
...
Remove two indentation levels by returning early on error.
Metze, please check!
2009-05-19 23:35:16 +02:00
Volker Lendecke
6ca4d727c4
Attempt to the build on NetBSD5 and a few others
2009-05-01 23:26:22 +02:00
Stefan Metzmacher
6ac61e6707
tsocket: add tstream_readv_pdu_queue_send/recv()
...
metze
2009-05-01 17:42:01 +02:00
Stefan Metzmacher
eb723c2061
tsocket: add tstream_writev_queue_send/recv()
...
metze
2009-05-01 17:42:00 +02:00
Stefan Metzmacher
a0830f4cb9
tsocket: add tstream_readv_pdu_send/recv()
...
metze
2009-05-01 17:41:56 +02:00
Stefan Metzmacher
ee6d796c19
tsocket: add tstream implementation for bsd sockets (inet and unix)
...
metze
2009-05-01 17:41:55 +02:00
Stefan Metzmacher
8a090c4b8b
tsocket: add tstream_context infrastructure similar to tdgram_context
...
metze
2009-05-01 17:41:53 +02:00
Stefan Metzmacher
182a4b30de
tsocket: rename tdgram_unix_dgram_socket() => tdgram_unix_socket()
...
metze
2009-05-01 17:41:38 +02:00
Stefan Metzmacher
b46599590f
tsocket: move tsocket_simple_int_recv() to tsocket.c
...
metze
2009-05-01 17:38:27 +02:00
Stefan Metzmacher
0db3b944bd
tsocket: move tsocket_sendto_queue_send/recv() to tsocket_helpers.c
...
metze
2009-05-01 17:38:26 +02:00
Stefan Metzmacher
b4c5387e9f
tsocket: tdgram move input checks to the common code as there're needed for all backends
...
We need to make sure that we free the sendto and recvfrom requests before
the tdgram_context
metze
2009-05-01 17:38:25 +02:00
Stefan Metzmacher
cc75ff1a37
tsocket: for unix domain sockets we need to use sizeof(struct sockaddr_un)
...
sizeof(struct sockaddr_storage) generates EINVAL.
metze
2009-05-01 17:38:24 +02:00
Stefan Metzmacher
66a2cd36c6
tsocket: return EINVAL when tdgram_sendto_send() is used with len == 0
...
metze
2009-05-01 17:38:23 +02:00
Stefan Metzmacher
997eb9eb43
tsocket: fix dependecies
...
metze
2009-05-01 17:38:22 +02:00
Stefan Metzmacher
9dbb535aa5
tsocket: remove tsocket_context related stuff
...
It will be replaced by tdgram_context and tstream_context.
metze
2009-05-01 17:38:21 +02:00
Stefan Metzmacher
56aae35a23
tsocket: fix the build without ipv6 support
...
metze
2009-04-16 09:31:14 +02:00
Stefan Metzmacher
522bdea1f3
tsocket: reduce size of dummy variable
...
metze
2009-04-06 16:37:18 +02:00
Stefan Metzmacher
9d9f945281
tsocket: use the correct 'ret' variable for tdgram_disconnect_send/recv()
...
metze
2009-04-06 15:49:08 +02:00
Stefan Metzmacher
3bbad34a02
tsocket: remove DGRAM support from tsocket_context
...
metze
2009-04-02 21:54:08 +02:00
Stefan Metzmacher
c59ee5a139
tsocket: optimize tdgram_bsd a lot
...
The desire is to do as less syscalls during the
tdgram_sendto_send/recv() and tdgram_recvfrom_send/recv()
operations.
1. we first try the sendto()/recvfrom() syscall and
only use a fd event if we got EAGAIN.
2. we cache the fd event and only change it's flags
if really needed.
For the highload case we do almost no epoll_ctl() and epoll_wait()/select()
syscalls anymore. This speeds up the LDAP-BENCH-CLDAP test
by more than 20%. (With a modified version of this test
which let the server skip any ldb calls and just return success
I'm getting about 8000 requests per second, while I'm getting
just about 6000 requests per second without optimization)
metze
2009-04-02 21:54:00 +02:00