adamdunkels
26e3b6fdc0
Bugfix by Simon Duquennoy: don't transmit packet until previous packet has been sent by the RDC layer.
2011-01-13 19:06:22 +00:00
adamdunkels
a0960154ba
Cleaned up parts of the code; added comments
2011-01-09 21:07:01 +00:00
dak664
322d835fee
Fix avr-gcc compiler warnings
2010-12-18 22:12:53 +00:00
adamdunkels
4a89cc9db3
Rework of the new CSMA in-order code that fixes a number of smalllish bugs in the in-order code.
2010-12-16 22:44:02 +00:00
adamdunkels
79f2c25dae
Rework of the phase-lock optimization mechanism. There were several bugs in the existing code, such as the phase-lock would lose synchronization when a packet was lost.
2010-12-16 22:43:07 +00:00
adamdunkels
952dfc9384
Rework of the CSMA layer: instead of using one timer per packet, the module now uses one central timer and a queue of outgoing packets. This helps to keep packets in order, which is extremely useful for TCP transfers, and reduces the amount of congestion since consecutive packets now are spaced in time if there is a collision. Based on the observations and ideas of Simon Duquennoy.
2010-12-14 07:57:14 +00:00
adamdunkels
af21afa750
Make sure that the cycle timing is exact
2010-12-10 13:36:06 +00:00
adamdunkels
c19a967973
Bugfix: with the recent bugfixes, it was no longer possible to turn duty cycling back on after it had been turned off
2010-12-07 23:12:54 +00:00
adamdunkels
4c9793f428
Let ContikiMAC remember the state that the radio should have when duty cycling is turned off. This fixes a bug in which the radio was turned on after sending, even when NETSTACK_RDC.off(0) was enabled.
2010-12-06 09:08:22 +00:00
dak664
e54d7d6f9c
Fix compiler warnings
2010-12-02 15:55:17 +00:00
adamdunkels
6d556d5c13
Fixed bug with link-layer ACKs. Refactored the ACK reception code.
2010-11-25 08:44:34 +00:00
nifi
3bd78893e5
Renamed nullrdc to nullrdc-noframer and nullrdc-framer to nullrdc.
...
nullrdc (previously nullrdc-framer) is now a RDC implementation that
does not do any radio duty cycling but supports link layer addresses
and frame format via framer (for example 802.15.4). nullrdc can now be
used in the same way as contikimac, xmac, etc.
nullrdc-noframer (previously nullrdc) is a null RDC implementation
that does not do any radio duty cycling, does not add any header of its
own to the packets, and does not support link layer addresses.
2010-11-23 18:11:00 +00:00
nifi
05d78522d3
Added MAC-layer duplicate message suppression when using 802.15.4 auto ack
2010-11-12 12:53:16 +00:00
adamdunkels
b7df18a96a
Turned off streaming by default to save code space; made guard time a little more conservative
2010-11-02 11:00:08 +00:00
adamdunkels
8fa7bf6c31
Always restore the queuebuf to the packetbuf before calling the upper layers. Reduce the waiting time between successive MAC layer retransmissions.
2010-10-24 21:07:00 +00:00
adamdunkels
73f5535754
Made the guard time slightly more conservative - found an occasion where this was needed
2010-10-24 21:06:17 +00:00
adamdunkels
9cb830ed37
Timing bugfix in the waiting period between channel samples to check if the coast is clear before sending a packet.
2010-10-20 22:25:24 +00:00
adamdunkels
3ffcff0656
Added mechanisms for explicit de-synchronization of synchronized neighbors. Added link-layer ACKs. Fixed bugs with MAC-layer interaction. Now seems to work somewhat (didn't work at all before, in some cases), but there are still problems with packet loss.
2010-10-20 15:23:43 +00:00
nifi
e2b4956a7f
Bug fix: release the packet and notify the sender if RDC returned an error
2010-10-14 19:08:39 +00:00
joxe
4f3fc1c404
fixed xmac to use new channel check define name
2010-10-03 22:46:53 +00:00
adamdunkels
ceebd7020a
Bugfix: the return value sometimes was wrong when a collision was detected.
2010-10-03 20:39:24 +00:00
adamdunkels
ae88ed04f1
Moved the definition of the channel check rate used by the radio duty
...
cycling layer. This definition was previously held in net/mac/mac.h,
with the name MAC_CONF_CHANNEL_CHECK_RATE. But since the rate is used
by the radio duty cycling layer, it makes more sense to change its
name to reflect this. Also, the configuration option should be tied to
the netstack configuration instead.
So the new configuration option is called
NETSTACK_CONF_RDC_CHANNEL_CHECK_RATE.
2010-10-03 20:37:32 +00:00
adamdunkels
eb58968459
Added a timer to each phase structure that keeps track of for how long
...
a particular receiver has not ACKed packets. After some time
(currently 16 seconds), the sender will begin sending full strobe
periods again.
2010-09-13 13:39:05 +00:00
dak664
12479c32bc
Fix compiler warnings
2010-08-01 21:18:07 +00:00
nifi
21828ad782
Changed the retransmission time to be based on the configured MAC channel check rate when the RDC has no channel check interval.
2010-06-16 10:08:30 +00:00
adamdunkels
7647cbd8aa
Use the list_item_next() API
2010-06-15 19:22:25 +00:00
adamdunkels
3fa8ffda1a
Moved the modules packetbuf, queuebuf, and packetqueue from net/rime to net/, since they are not Rime-specific
2010-06-14 19:19:16 +00:00
adamdunkels
951fd78f6c
ctimer moved from net/rime to sys/
2010-06-14 18:58:45 +00:00
adamdunkels
8f8393aadc
Moved ctimer.[ch] from their old and illogical location in core/net/rime to a more appropriate place in core/sys
2010-06-14 07:34:36 +00:00
adamdunkels
8d1680fed5
Filter duplicate packets
2010-06-14 06:52:41 +00:00
nifi
ac9b862cd0
Added configurable support for 802.15.4 autoack
2010-05-26 14:12:33 +00:00
nvt-se
b8a3a4109e
Adjusted the code so as to remove compilation warnings.
2010-04-30 22:18:12 +00:00
adamdunkels
0859d4b95f
Avoid turning off the radio in the middle of a packet reception
2010-04-30 07:32:39 +00:00
adamdunkels
e606c674c2
Rewrote LPP so that it uses 802.15.4 headers. A number of bugfixes with respect to alignment. Avoid turning off the radio in the middle of packet reception.
2010-04-30 07:31:44 +00:00
adamdunkels
3c735fcea4
Added a packetbuf flag that indicates if the link layer acks should be used or not
2010-04-30 07:25:51 +00:00
fros4943
12b6f35ce7
PRINTF()ed debugging output
2010-04-27 12:56:35 +00:00
nifi
85d6f6c514
removed unused variable
2010-04-26 22:41:50 +00:00
nifi
b26f45a35b
code cleanup
2010-04-26 17:55:11 +00:00
nifi
d761b6e89a
Added MAC header to handle padding
2010-04-26 17:46:21 +00:00
nifi
14b1bb9f57
Fixed to only use leds when DEBUG is set
2010-04-20 11:41:16 +00:00
adamdunkels
3670fac618
Bugfix: timing fixed
2010-04-08 18:23:24 +00:00
adamdunkels
91ff8574ad
Improved handling of duplicate packets
2010-04-08 09:32:56 +00:00
nvt-se
ac4ba199bc
Turn off DEBUG mode by default.
2010-04-06 21:37:54 +00:00
adamdunkels
497111bcc9
Added an optional rate limit for broadcast transmissions
2010-04-06 11:57:43 +00:00
adamdunkels
e06a30fdb0
Added MAC-layer duplicate message suppression, which is very helpful when running with duty cycling turned off
2010-04-05 19:28:07 +00:00
adamdunkels
c68ee6e577
Bugfix: memory from the wrong memb was freed.
2010-04-04 21:02:09 +00:00
adamdunkels
a4c328785a
Small timing tweaks
2010-04-04 21:01:24 +00:00
adamdunkels
62039a530d
Removed non-existant cooja-debug.h include
2010-04-04 12:31:47 +00:00
adamdunkels
7d6ff0289e
Fiddled more with the timing: the phase seems stable now at least. Still bugs left though...
2010-04-04 12:28:29 +00:00
adamdunkels
be117afe04
After a lot of tweaking and twiddling with the timing of the ContikiMAC code and the mechanisms in the CC2420 driver, things seem to work somewhat stable now - knock on wood!
2010-04-04 07:49:30 +00:00