olsrd.git
11 years agoAutomated merge with http://gredler.at/hg/olsrd
Bernd Petrovitsch [Sat, 18 Oct 2008 20:53:53 +0000 (22:53 +0200)]
Automated merge with http://gredler.at/hg/olsrd

11 years agocleanup strange looking MID code
Hannes Gredler [Fri, 17 Oct 2008 16:42:42 +0000 (18:42 +0200)]
cleanup strange looking MID code

11 years agoComplete the update the txtinfo plugin to not use the pollrate callback function
Bernd Petrovitsch [Fri, 17 Oct 2008 01:06:23 +0000 (03:06 +0200)]
Complete the update the txtinfo plugin to not use the pollrate callback function

Nothing more to say.

11 years agoConvert the numbers to a bitmask
Bernd Petrovitsch [Fri, 17 Oct 2008 01:06:15 +0000 (03:06 +0200)]
Convert the numbers to a bitmask

Convert the constants which specify the outpout from
numbers to a bitmask to simplify the selection.

11 years agoCleanup and bugfix the txtinfo plugin
Bernd Petrovitsch [Fri, 17 Oct 2008 01:06:05 +0000 (03:06 +0200)]
Cleanup and bugfix the txtinfo plugin

Contents:
- fix file descriptor leaks if initialization find's an error
- remove ipc_open and ipc_socket_up away. We use -1 in the respective
  file descriptors for "invalid".
- killed unused variables and return types
- killed trivially duplicated code
- localized variable defintions
- added a few "const"
- removed superflous #include's
- Some minor coding style issues.

11 years agoimported patch extend-scheduler
Bernd Petrovitsch [Fri, 17 Oct 2008 00:38:31 +0000 (02:38 +0200)]
imported patch extend-scheduler

11 years agoMisc build fixes
Bernd Petrovitsch [Fri, 17 Oct 2008 00:36:38 +0000 (02:36 +0200)]
Misc build fixes

Namely:
- added a target "show-ignored-warnings"
- added a few warnings (from gcc-4.3.2)
- fixed typos

11 years agoimported patch scheduler-fixes
Bernd Petrovitsch [Thu, 16 Oct 2008 17:46:41 +0000 (19:46 +0200)]
imported patch scheduler-fixes

11 years agoAutomated merge with http://gredler.at/hg/olsrd
Bernd Petrovitsch [Thu, 9 Oct 2008 20:34:20 +0000 (22:34 +0200)]
Automated merge with http://gredler.at/hg/olsrd

11 years agoCleanup net_olsr.[ch]
Bernd Petrovitsch [Mon, 6 Oct 2008 23:35:08 +0000 (01:35 +0200)]
Cleanup net_olsr.[ch]

While we are at it.

Coding style cleanups:
- create a "typedef" for the ifchg functions
- removed lots of superflous type casts cluttering the source
- use "tmp" as temporary variable name instead of e.g. "tmp_ptf_list"

Code cleanup:
- Use "ip4equal(&ifp->int_addr.sin_addr, &addr->v4)" instead of
  "((struct sockaddr_in *)&ifp->int_addr)->sin_addr.s_addr == addr->v4.s_addr"
  Ditto for IPv6.
- rewrite all occurences of sockaddr4_to_string() with ip4_to_string()
- replaced sockaddr4_to_string() with a function which prints all of the data
  (and not only the IP-address).
- added a sockaddr_to_string() and sockaddr6_to_string() to print the whole data
- move "init_net();" earlier as it doesn't depend on anything else.
- use a "union" for the struct sockaddr/sockaddr_in/sockaddr_in6 handling in
  net_output and rewrite the address copy operations killing several variables.
  And we have now one "olsr_sendto()" call.
- made one-line functions "static INLINE"
- #if 0 ... #endif out unused funtions: net_reserve_bufspace(), net_outbuffer_push_reserved()

BTW the whole chk_if_changed() function in src/unix/ifnet.c needs a real cleanup.

11 years agoReplace the nanosleep() with select()
Bernd Petrovitsch [Mon, 6 Oct 2008 23:35:08 +0000 (01:35 +0200)]
Replace the nanosleep() with select()

And now to the more gory stuff. We replace the nanosleep() in
the central loop with a select() (which doesn't has any file
descriptors there yet).

We basically select() and recalculate the timeout until the
time is over. And we make sure that we select() at least once -
even if the timeout is "0" (or < "0").
This is to give the funtions chance to do something even in
high-load situations.
The olsr_scheduler_sleep() function is thus superflous.

11 years agoSimplify scheduler.c
Bernd Petrovitsch [Mon, 6 Oct 2008 23:35:08 +0000 (01:35 +0200)]
Simplify scheduler.c

Make scheduler.c easier to read and reduce the code. And kill a
superflous memset() just before initializing the array.

11 years agoCleanup the signal handlers
Bernd Petrovitsch [Mon, 6 Oct 2008 23:35:08 +0000 (01:35 +0200)]
Cleanup the signal handlers

Not really belonging to that theme but I stumbled over it and it annoyed me
since ages:
The signal handlers where quite large functions very probably using system-calls
which one better do not use in a signal handler. Read Stevens'  "Programming in
the Unix environment" for more details into signal handlers.
The worse for Win32: Apperentyl Win32 needs two extra variables to let one
thread break out of the select() loop and continue afterwards with the shutdown
in the signal handler.

The simple and clean solution is quite easy: We add a state variable which is
changed in the signal handlers and do the whole action afterwards.

And the usual coding style cleanup.

This has been tested on Linux and Win32.

11 years agoMerge socket_parser.[ch] into scheduler.[ch]
Bernd Petrovitsch [Mon, 6 Oct 2008 23:35:07 +0000 (01:35 +0200)]
Merge socket_parser.[ch] into scheduler.[ch]

There is no real and clear border between the two above. And the files
are not that large. So we simply merge them and loose one global
function (which is now "static").

11 years agoCleanup scheduler.[ch]
Bernd Petrovitsch [Mon, 6 Oct 2008 23:35:06 +0000 (01:35 +0200)]
Cleanup scheduler.[ch]

This is done by
- making local-only used variables and funtions "static"
- move variable definitions to more inner blocks
- simplify the code by reducing the number of parallel incremented variables. gcc
  is smart enough.

11 years agoMake the socket parser more general
Bernd Petrovitsch [Mon, 6 Oct 2008 23:35:06 +0000 (01:35 +0200)]
Make the socket parser more general

The general goal is to get to a flexible central event loop to allow
decoupling of long time actions (e.g. a slow browser over a slow
connection) from the rest.

We extend the central select() loop to get an additional select() for
file descriptors where the data will be handled immediately.
We do this via an additional callback function (as to not disturb the
current ones).
The caller must also specify, if he wants to be called on a possible
read andor write on the "immediate" and/or the "pollrate" callback
function.
While we're at it, we add a "void *" which is passed back to the caller.
So the caller can actually register more than one instance on different
file descriptors and differentiate the instances via that pointer. E.g.
the httpinfo plugin is now able to really accept more than one client
connection at the same time.

11 years agoCleanup parser.[ch]
Bernd Petrovitsch [Mon, 6 Oct 2008 23:35:05 +0000 (01:35 +0200)]
Cleanup parser.[ch]

This is done by
- de-obfuscating the code

11 years agoCleanup ipc_frontend
Bernd Petrovitsch [Mon, 6 Oct 2008 23:35:05 +0000 (01:35 +0200)]
Cleanup ipc_frontend

This is done by
- making ipc_check_allowed_ip(), ipc_accept() and frontend_msgparser() "static"
  functions as they are only uised locally.
- ditto for "struct ipcmsg" and "struct ipc_net_msg"
- replacing unused return types with "void"
- killing superfluous #include's
- coding style cleanup, adding "local Variables" sections
- convert a pointer - malloc()ed at begin of a function and free()d at the end
  into a normal variable on the stack.
  This saves code to malloc(), chack that and free() it.
- Killing the boolean "ipc_active" variable as we can use the value "-1" in
  "icp_conn" for the same.

11 years agoCleanup socket_parser.[ch]
Bernd Petrovitsch [Mon, 6 Oct 2008 23:35:05 +0000 (01:35 +0200)]
Cleanup socket_parser.[ch]

This is done by
- coding style cleanup, adding "local Variables" sections
- removed the unused "arp_type" field
- creating a "typedef" for the socket handler functions
- calculating the "highest fd+1" for select() with the fdsets. It
  is more like almost allother similar apps and we avoid the ugly
  recalculation on socket removal and reduce the code.
- Retry the select() immediately on EINTR and EAGAIN (which
  shouldn't happen in the first place).

11 years agoAutomated merge with http://gredler.at/hg/olsrd
Bernd Petrovitsch [Mon, 6 Oct 2008 21:58:05 +0000 (23:58 +0200)]
Automated merge with http://gredler.at/hg/olsrd

11 years agoBugfix and work on the httpinfo plugin
Bernd Petrovitsch [Mon, 6 Oct 2008 21:54:20 +0000 (23:54 +0200)]
Bugfix and work on the httpinfo plugin

This changesets includes:
- The make ADMIN_INTERFACE variable must be == 1 to be "activated" as specified
  in the docs.
  And the C preprocessor #define is set to 0 or 1. So we can use #if (and not
  #ifdef) to check it. This saves against typos in the variables names.
- Without the admin_interface, the admin_interface.{c,h} files can be
  completely ignored. So we do that. We keep #ifdef's inside of the 2 files
  though.
- trivial cleanups:
  - reduced a few lines and using snprintf()s format strings with the
    snprintf(). Otherwise I regurarly wondered why %% was in the constant
    strings;-)
  - remove superflous #include's, added needed ones
  - save few bytes in the CSS. It is unreadable anyways.
- Removed code for the dead "lq_wsize" option.
- a few layout improvements.

Last but not least and the by far important part as it is a real bugfix:
- Removed the "for (idx = 0; idx < HASHSIZE; idx++) {" loop around the
  "OLSR_FOR_ALL_TC_ENTRIES(tc)" loop.
  Getting a > 11MB (IIRC) HTML file (if one tries it on the FunkFeuer net)
  didn't make FireFox happy. And reading the same 2 lines lots of times on
  a 2 node "net" is not better.
  I think it was forgotten/overseen with the MID-set refactoring.

Actually testet on Linux and Win32.

11 years ago* * *
Bernd Petrovitsch [Mon, 6 Oct 2008 21:45:41 +0000 (23:45 +0200)]
* * *

11 years agoadd make target for arprefresh
Eric Malkowski [Mon, 6 Oct 2008 12:40:29 +0000 (14:40 +0200)]
add make target for arprefresh

11 years ago* grrrmmbl, fixed unwanted changes
Bernd Petrovitsch [Sat, 4 Oct 2008 22:31:32 +0000 (00:31 +0200)]
* grrrmmbl, fixed unwanted changes

11 years ago* added missing __attribute__((unsed))
Bernd Petrovitsch [Sat, 4 Oct 2008 22:28:07 +0000 (00:28 +0200)]
* added missing __attribute__((unsed))
* IMHO olsr_print_hna_set() should have looked like this.

11 years agoAutomated merge with http://gredler.at/hg/olsrd
Bernd Petrovitsch [Sat, 4 Oct 2008 19:07:57 +0000 (21:07 +0200)]
Automated merge with http://gredler.at/hg/olsrd

11 years ago[mq]: use-c99-bool
Bernd Petrovitsch [Sat, 4 Oct 2008 18:57:26 +0000 (20:57 +0200)]
[mq]: use-c99-bool

11 years agoimported patch avoid-duplicate-define
Bernd Petrovitsch [Sat, 4 Oct 2008 18:17:11 +0000 (20:17 +0200)]
imported patch avoid-duplicate-define

11 years ago* make it compile without -DUSE_FPM
Bernd Petrovitsch [Sat, 4 Oct 2008 17:27:24 +0000 (19:27 +0200)]
* make it compile without -DUSE_FPM
* untested
Or we want to get rid of the old code?

11 years ago removed superflous -ifndef
Bernd Petrovitsch [Sat, 4 Oct 2008 16:59:04 +0000 (18:59 +0200)]
 removed superflous -ifndef

11 years ago- fix building on x86_64: "long" is 64bit there and according to the
Bernd Petrovitsch [Sat, 4 Oct 2008 16:55:50 +0000 (18:55 +0200)]
- fix building on x86_64: "long" is 64bit there and according to the
  comment, we want 32bit only. So we use "int"

11 years agoAutomated merge with http://gredler.at/hg/olsrd
Bernd Petrovitsch [Thu, 25 Sep 2008 22:20:17 +0000 (00:20 +0200)]
Automated merge with http://gredler.at/hg/olsrd

11 years agorework the HNA input parser
Hannes Gredler [Mon, 22 Sep 2008 21:26:05 +0000 (23:26 +0200)]
rework the HNA input parser

11 years agoremove max_jitter references since its not required after conversion to the new timer API
Hannes Gredler [Mon, 22 Sep 2008 16:30:17 +0000 (18:30 +0200)]
remove max_jitter references since its not required after conversion to the new timer API

11 years agobump version to pre-0.5.7
Hannes Gredler [Mon, 22 Sep 2008 14:03:42 +0000 (16:03 +0200)]
bump version to pre-0.5.7

11 years agore-indent interface code
Hannes Gredler [Mon, 22 Sep 2008 14:02:24 +0000 (16:02 +0200)]
re-indent interface code

11 years agorework the buffer holddown timer to use the new timer API
Hannes Gredler [Mon, 22 Sep 2008 13:23:19 +0000 (15:23 +0200)]
rework the buffer holddown timer to use the new timer API

11 years ago* fixed broken push
Bernd Petrovitsch [Thu, 18 Sep 2008 22:30:25 +0000 (00:30 +0200)]
* fixed broken push

11 years agoAutomated merge with http://gredler.at/hg/olsrd
Bernd Petrovitsch [Thu, 18 Sep 2008 22:24:25 +0000 (00:24 +0200)]
Automated merge with http://gredler.at/hg/olsrd

11 years ago* fix win build
Bernd Petrovitsch [Thu, 18 Sep 2008 22:24:19 +0000 (00:24 +0200)]
* fix win build

11 years agofix windows build
Bernd Petrovitsch [Thu, 18 Sep 2008 22:17:12 +0000 (00:17 +0200)]
fix windows build

11 years agoAutomated merge with http://gredler.at/hg/olsrd
Bernd Petrovitsch [Thu, 18 Sep 2008 22:12:28 +0000 (00:12 +0200)]
Automated merge with http://gredler.at/hg/olsrd

11 years ago[mq]: fix-win-build
Bernd Petrovitsch [Thu, 18 Sep 2008 22:12:23 +0000 (00:12 +0200)]
[mq]: fix-win-build

11 years agoAutomated merge with http://gredler.at/hg/olsrd
Bernd Petrovitsch [Mon, 15 Sep 2008 21:59:01 +0000 (23:59 +0200)]
Automated merge with http://gredler.at/hg/olsrd

11 years agobackout debug settings from last commit
Hannes Gredler [Mon, 15 Sep 2008 14:12:27 +0000 (16:12 +0200)]
backout debug settings from last commit

11 years agoprint correct HNA net when DEBUG is on
Hannes Gredler [Mon, 15 Sep 2008 14:00:03 +0000 (16:00 +0200)]
print correct HNA net when DEBUG is on

11 years agoHNA refactoring phase #1
Hannes Gredler [Mon, 15 Sep 2008 13:56:38 +0000 (15:56 +0200)]
HNA refactoring phase #1

11 years agodo not reset the mid_timer since its needed for a proper clean up
Hannes Gredler [Fri, 12 Sep 2008 12:46:43 +0000 (14:46 +0200)]
do not reset the mid_timer since its needed for a proper clean up

11 years agoAutomated merge with http://gredler.at/hg/olsrd
Bernd Petrovitsch [Thu, 11 Sep 2008 18:43:08 +0000 (20:43 +0200)]
Automated merge with http://gredler.at/hg/olsrd

11 years agoinit MID subtree in the lsdb
Hannes Gredler [Wed, 10 Sep 2008 19:30:20 +0000 (21:30 +0200)]
init MID subtree in the lsdb

11 years agoset the IPv6 hop limit (aka ttl) for outgoing packets to 1.
Joerg Pommnitz [Wed, 10 Sep 2008 12:16:13 +0000 (14:16 +0200)]
set the IPv6 hop limit (aka ttl) for outgoing packets to 1.
Currently olsrd only works with global multicast addresses.
If your olsr box routes multicast traffic it will happily
(and correctly) forward the OLSR messages, which is undesired.

11 years agoDo not automatically add a filter entry for ourselves
Hannes Gredler [Tue, 9 Sep 2008 10:43:39 +0000 (12:43 +0200)]
Do not automatically add a filter entry for ourselves

11 years agoconvert the inbound msg filter to an AVL tree
Hannes Gredler [Tue, 9 Sep 2008 10:28:41 +0000 (12:28 +0200)]
convert the inbound msg filter to an AVL tree

11 years ago* warning fixes
Bernd Petrovitsch [Mon, 8 Sep 2008 20:43:14 +0000 (22:43 +0200)]
* warning fixes

11 years ago[mq]: larger-inline
Bernd Petrovitsch [Sun, 7 Sep 2008 22:50:29 +0000 (00:50 +0200)]
[mq]: larger-inline

11 years agoAutomated merge with http://gredler.at/hg/olsrd
Bernd Petrovitsch [Sun, 7 Sep 2008 22:47:08 +0000 (00:47 +0200)]
Automated merge with http://gredler.at/hg/olsrd

11 years ago* x86_64 warning fix
Bernd Petrovitsch [Sun, 7 Sep 2008 22:43:10 +0000 (00:43 +0200)]
* x86_64 warning fix

11 years agoAutomated merge with http://gredler.at/hg/olsrd
Bernd Petrovitsch [Sun, 7 Sep 2008 22:42:11 +0000 (00:42 +0200)]
Automated merge with http://gredler.at/hg/olsrd

11 years ago* merge with tip
Bernd Petrovitsch [Sun, 7 Sep 2008 22:41:39 +0000 (00:41 +0200)]
* merge with tip

11 years ago* ooops, make it compile again
Bernd Petrovitsch [Sun, 7 Sep 2008 22:36:15 +0000 (00:36 +0200)]
* ooops, make it compile again

11 years agoresolve FreeBSD build break
Hannes Gredler [Sun, 7 Sep 2008 22:28:57 +0000 (00:28 +0200)]
resolve FreeBSD build break

11 years agoAutomated merge with http://gredler.at/hg/olsrd
Bernd Petrovitsch [Sun, 7 Sep 2008 21:28:02 +0000 (23:28 +0200)]
Automated merge with http://gredler.at/hg/olsrd

11 years ago* fxied compiler warning
Bernd Petrovitsch [Sun, 7 Sep 2008 21:27:03 +0000 (23:27 +0200)]
* fxied compiler warning

11 years ago* activate -Wshadow
Bernd Petrovitsch [Sun, 7 Sep 2008 21:26:43 +0000 (23:26 +0200)]
* activate -Wshadow

11 years agoAutomated merge with http://gredler.at/hg/olsrd
Bernd Petrovitsch [Sun, 7 Sep 2008 21:26:15 +0000 (23:26 +0200)]
Automated merge with http://gredler.at/hg/olsrd

11 years agoAutomated merge with http://gredler.at/hg/olsrd
Bernd Petrovitsch [Sun, 7 Sep 2008 21:25:59 +0000 (23:25 +0200)]
Automated merge with http://gredler.at/hg/olsrd

11 years ago* added (again) a macro to annotate debugging-only variables. And use it.
Bernd Petrovitsch [Sun, 7 Sep 2008 21:23:38 +0000 (23:23 +0200)]
* added (again) a macro to annotate debugging-only variables. And use it.

11 years ago* compile fixes for *BSD
Bernd Petrovitsch [Sun, 7 Sep 2008 20:30:50 +0000 (22:30 +0200)]
* compile fixes for *BSD
* removed duplicated #include

11 years agomerged tip
bernd@moria.at.home [Sun, 7 Sep 2008 20:22:55 +0000 (22:22 +0200)]
merged tip

11 years agoAutomated merge with http://gredler.at/hg/olsrd
Bernd Petrovitsch [Sun, 7 Sep 2008 20:22:14 +0000 (22:22 +0200)]
Automated merge with http://gredler.at/hg/olsrd

11 years ago merge bmf fix
Bernd Petrovitsch [Sun, 7 Sep 2008 20:20:27 +0000 (22:20 +0200)]
 merge bmf fix

11 years ago* made bmf plugin compile. I don't know if that is useful or even correct.
Bernd Petrovitsch [Sun, 7 Sep 2008 20:19:32 +0000 (22:19 +0200)]
* made bmf plugin compile. I don't know if that is useful or even correct.

11 years ago* fixed Win32 compile errors and warnings
Bernd Petrovitsch [Sun, 7 Sep 2008 20:18:37 +0000 (22:18 +0200)]
* fixed Win32 compile errors and warnings

11 years ago* build fix
Bernd Petrovitsch [Sun, 7 Sep 2008 20:13:28 +0000 (22:13 +0200)]
* build fix

11 years agoAutomated merge with http://gredler.at/hg/olsrd
Bernd Petrovitsch [Sun, 7 Sep 2008 20:07:35 +0000 (22:07 +0200)]
Automated merge with http://gredler.at/hg/olsrd

11 years agomerged with tip
Bernd Petrovitsch [Sun, 7 Sep 2008 20:07:21 +0000 (22:07 +0200)]
merged with tip

11 years ago fixed windows warnings and build errors
bernd@moria.at.home [Sun, 7 Sep 2008 19:51:18 +0000 (21:51 +0200)]
 fixed windows warnings and build errors

11 years ago* fixed the first compile error on Win32
Bernd Petrovitsch [Sun, 7 Sep 2008 16:35:37 +0000 (18:35 +0200)]
* fixed the first compile error on Win32

11 years ago* added support of EXTRA_WARNINGS similar to EXTRA_CFLAGS and EXTRA_CPPFLAGS
Bernd Petrovitsch [Sun, 7 Sep 2008 16:22:25 +0000 (18:22 +0200)]
* added support of EXTRA_WARNINGS similar to EXTRA_CFLAGS and EXTRA_CPPFLAGS

11 years agomake the duplicate set code use our lightweight memory manager
Hannes Gredler [Sun, 7 Sep 2008 15:25:49 +0000 (17:25 +0200)]
make the duplicate set code use our lightweight memory manager

11 years agorefactor MID message parsing
Hannes Gredler [Sun, 7 Sep 2008 15:02:39 +0000 (17:02 +0200)]
refactor MID message parsing
avoid memory churn when parsing MID messages

11 years ago* kill unused variables
Bernd Petrovitsch [Fri, 5 Sep 2008 16:04:17 +0000 (18:04 +0200)]
* kill unused variables

11 years ago* grrmml, and we need another #include of course
Bernd Petrovitsch [Fri, 5 Sep 2008 15:30:12 +0000 (17:30 +0200)]
* grrmml, and we need another #include of course

11 years agoMID refactoring phase #1
Hannes Gredler [Fri, 5 Sep 2008 15:21:51 +0000 (17:21 +0200)]
MID refactoring phase #1
clean up the code towards a unified linkstate database where all the information
is organized in a nodal oriented database. we use the TC tree as a hook to hold various
information. A MID entry is inserted both in a global tree for alias lookup
and in a per-tc subtree for housekeeping like timeouts etc.
replace MID traversals with the appropriate macros in the plugins.

11 years agoMajor cleanup with the goal to free src/defs.h from all sorts of stuff
Bernd Petrovitsch [Fri, 5 Sep 2008 11:38:54 +0000 (13:38 +0200)]
Major cleanup with the goal to free src/defs.h from all sorts of stuff
not belonging into a global .h file but somewhere else. src/defs.h now contains
only global (and application/OLSR-independent items) and serves as the .h file file
for main.c and builddata.c (or do we want to split out quite small main.h and builddata.h?).

More in detail:
- strscpy() and  strscat() are full blown functions (and not just "one-liners").
  So they are IMHO too large that inlining makes sense (even if debugging is disabled and
  src and dest != NULL. Even than a loop and a funtion call is left).
  So we move them into the new common/string.[ch] files
- The declaration of the "now_times" variable is moved into src/scheduler.h since
  it's definition is in src/scheduler.c.
  The timestamp-related #defines are also moved into src/scheduler.h with it.
- MIN_PACKET_SIZE() is only used in parser.c. IMHO it should go into that .c file.
  Or at least - as the first step - into parser.h.
- The declaration of the "olsr_cnf" variable clearly belongs into src/olsr_cfg.h
- Almost all #include's are removed as no longer needed.
  And the missing #include's are added to that .c and .h files where necessary.
  The bsd/ and win32/ subdirectories also got #include's added as they use errno etc.
  Sorry, if I missed something.
Now we can use src/defs.h for more global "generic" stuff and e.g. #include <src/defs.h>
in e.g. common/*.[ch] - in only for the '#define INLINE'

11 years ago* added missing #include
Bernd Petrovitsch [Fri, 5 Sep 2008 08:57:27 +0000 (10:57 +0200)]
* added missing #include

11 years agotypo fixes
Bernd Petrovitsch [Sun, 31 Aug 2008 21:08:44 +0000 (23:08 +0200)]
typo fixes

11 years ago- moved the declarations of ipc_init() and shutdown_ipc() from src/defs.h
Bernd Petrovitsch [Sun, 31 Aug 2008 21:02:42 +0000 (23:02 +0200)]
- moved the declarations of ipc_init() and shutdown_ipc() from src/defs.h
  to src/ipc_frontend.h where they better belong
- removed the declaration of ipc_output() as there is no function anywhere

11 years agoThe integer plugin parameter handler should be used for dyn gw's
Eric Malkowski [Thu, 4 Sep 2008 05:06:20 +0000 (07:06 +0200)]
The integer plugin parameter handler should be used for dyn gw's
"interval" as it's more correct since it's the number of whole seconds
between pings, not a floating point number (the nanosleep used to
implement it wants a struct timespec and the number of seconds is
integer and nanoseconds is set to 0L in the code).

11 years ago- moved the declarations of ipc_init() and shutdown_ipc() from src/defs.h
Bernd Petrovitsch [Sun, 31 Aug 2008 21:02:42 +0000 (23:02 +0200)]
- moved the declarations of ipc_init() and shutdown_ipc() from src/defs.h
  to src/ipc_frontend.h where they better belong
- removed the declaration of ipc_output() as there is no function anywhere

11 years agoadded "-funit-at-a-time" to the gcc options. Newer gccs (e.g. 4.3.0 on Fedora 9)
Bernd Petrovitsch [Sun, 31 Aug 2008 20:59:22 +0000 (22:59 +0200)]
added "-funit-at-a-time" to the gcc options. Newer gccs (e.g. 4.3.0 on Fedora 9)
need it for "-finline-functions-called-once"

11 years agotypo fixes
Bernd Petrovitsch [Sun, 31 Aug 2008 21:08:44 +0000 (23:08 +0200)]
typo fixes

11 years agodraft bugfix to the times(2) overrun situation
Bernd Petrovitsch [Tue, 2 Sep 2008 23:03:14 +0000 (01:03 +0200)]
draft bugfix to the times(2) overrun situation

11 years ago- the "cookies" array is used only here. So make it "static".
Bernd Petrovitsch [Tue, 2 Sep 2008 22:57:37 +0000 (00:57 +0200)]
- the "cookies" array is used only here. So make it "static".
- the "cookies" array is a global variable and thus initialized to
  0 anyways (Google it or read only e.g.
  http://gcc.gnu.org/ml/gcc-help/2007-07/msg00097.html).
  Make it explciti so and remove the now superfluous initializetion code.

11 years agoremove unnecessary cleanup call for empty TC messages
Hannes Gredler [Tue, 2 Sep 2008 14:18:14 +0000 (16:18 +0200)]
remove unnecessary cleanup call for empty TC messages

11 years agodo not build empty TC messages, kill the empty_tc_timer
Hannes Gredler [Tue, 2 Sep 2008 14:15:13 +0000 (16:15 +0200)]
do not build empty TC messages, kill the empty_tc_timer

11 years agore-indent duplicate_set code
Hannes Gredler [Tue, 2 Sep 2008 11:58:47 +0000 (13:58 +0200)]
re-indent duplicate_set code

11 years agoRe-enabled and fixed max_jitter resulting in sending out larger OLSR packets
Sven-Ola Tuecke [Mon, 1 Sep 2008 20:01:42 +0000 (22:01 +0200)]
Re-enabled and fixed max_jitter resulting in sending out larger OLSR packets

11 years agoRevert changes to etx_ff plugin
Sven-Ola Tuecke [Mon, 1 Sep 2008 19:58:37 +0000 (21:58 +0200)]
Revert changes to etx_ff plugin

11 years agoadded "-funit-at-a-time" to the gcc options. Newer gccs (e.g. 4.3.0 on Fedora 9)
Bernd Petrovitsch [Sun, 31 Aug 2008 20:59:22 +0000 (22:59 +0200)]
added "-funit-at-a-time" to the gcc options. Newer gccs (e.g. 4.3.0 on Fedora 9)
need it for "-finline-functions-called-once"