Update of Std/Split Control interface definitions to be in line with TEP115. Also added the EALREADY type to TinyError.h so that providers of the SplitControl interface can return this value
Updated TestDissemination and TestNetwork to start the dissemination service manually. I've tried TestDissemination, but haven't tried TestNetwork. Let me know if you experience any problems.
Added a StdControl interface, and a top-level DisseminationC component. Added a set() command to the DisseminationValue interface that changes the locally stored value w/o disseminating it.
Patch from John Regehr. Make queue catch it when the underlying radio
stack issues a spurious sendDone. Currently, if the queue is empty this
will cause it to access invalid memory (past the end of an array).
The real bug is the spurious sendDone, which he observed on the mica2.
However, until this bug is fixed, we need to catch it.
Supporting test suite for the newest CC2420 stack. Includes a low power version of RadioCountToLeds, TxThroughput analysis, TestAck, Unicast and Broadcast deliveries, RssiToSerial channel analyzer, and the PacketLink test app.
Update of the default CC2420 to the low power listening stack. The "acknowledgement" LPL version is the default, while the continuous-modulation "no-acknowledgement" LPL layer is still considered experimental. RSSI can be directly read from the radio through CC2420ConrolC.
include support in storage volume xml files
note that there's now a PLATFORMDIR define that points to the actual
platform directory and that can be overridden in .target files
Include I2C in sleep calculation, and base SPI off of SPE, not SPIE;
it may be enabled but not with interrupts. It would be a weird use case,
but it is possible.
> during event void AdcResource.granted[uint8_t client]()
> If your sref is not REFERENCE_VREFplus_AVss ||
> REFERENCE_VREFplus_VREFnegterm
> then you will never call Refvolt_1/2_5V.start()
> Now while releasing the resource, imho, we should do the same check
> else we end up with a post switchOff() loop
> as call RefVolt_1_5V.stop() will fail.
Fix: changed the "owner" variable to "syncOwner", which is now only modified in sync (task) context and have the async path "parallel" and unaffecting the "syncOwner" variable (hard to explain, have a look at the diff) - there's a test app in tinyos-2.x/apps/tests/msp430/Adc12.
switch to metric MHz for SMCLK, adjust constants, include SmclkManager
reduce baud rate for radio -- it can not handle more than 23kBit/s with a
50kHz frequency shift
idgay [Thu, 29 Mar 2007 21:07:25 +0000 (21:07 +0000)]
a nicer version, w/ cleaner HPL access
in particular, HPL for timer0 now has less magic in it (this means that
users of it have to be more aware of the asynchronous timer access rules,
though)
vlahan [Tue, 13 Mar 2007 17:57:46 +0000 (17:57 +0000)]
Quick and dirty solution for the uninitialized timers problem (after the removal of the periodic calibration components) leading to high Null app consumption. The initialization needs to be restructured to handle this properly...
andreaskoepke [Sat, 10 Mar 2007 22:14:21 +0000 (22:14 +0000)]
removed bug that could stop the MAC from operating. Lesson re-learned:
a splitphase call may trigger the done event before the call returns -- hence:
make sure that all state that the done event uses is initialized befor the call.