]> oss.titaniummirror.com Git - tinyos-2.x.git/log
tinyos-2.x.git
9 years agostm25p: fix seek error master
R. Steve McKown [Fri, 2 Jan 2015 21:06:50 +0000 (14:06 -0700)]
stm25p: fix seek error

Ensure correct computation of m_log_info[ id ].read_addr when client code
requests seek() to a cookie whose value is below the current virtual
address range by a distance, in address value, greater than 255 sectors.

The sector component of a virtual address is 16 bits, given
STM25P_SECTOR_SIZE_LOG2 is 16 and sizeof(storage_cookie_t) is 32.
However, in ClientResource.granted[]() in the S_SEEK case, the client's
read_addr value can have its upper bits silently discarded because
sector value are handled in uint8_t variables.

The symptom of upper bit discards is that client code receives the
contents of a full log many times until the client read_addr value
finally increments to the point that allows satisfaction of the end of
log condition,

    read_addr >= m_log_info[ client ].write_addr

Example case exhibiting this behavior:

read_addr = 0x59f0000
write_addr = 5a6dbf8
LogRead.seek(0x2a71400)

The resulting address handling in ClientResource.granted[]() in the
S_SEEK request case, around Stm25pLogP.nc:229:

readSector = 0xa7; // from incoming cookie value, should be 0x2a7
writeSector = 0xa7; // should be 0x5a7
if (writeSector - readSector > numSectors): FALSE, should be TRUE
  Had the if expression above been TRUE, the read address would
  still be computed incorrectly due to 8-bit match in the if's
  'then' clause.
m_log_state[ id  ].read_addr = 0x2a7000; // should be 0x59f000

9 years agomsp430 usci: fix DMA TX byte overwrite
R. Steve McKown [Mon, 29 Dec 2014 23:02:27 +0000 (16:02 -0700)]
msp430 usci: fix DMA TX byte overwrite

The DMA transmit operation for USCI was a little naive, in that it was
not accounting for the possibility that UCxxTXBUF already has a byte to
be transmitted.  Wait for this byte to be transmitted before starting
the DMA operation.

Also note in the code that the DMA transfer operations, trigered by
UCxxTXIFG, are edge sensitive.

9 years agomsp430 usci: TX operations via DMA
R. Steve McKown [Tue, 23 Dec 2014 19:50:45 +0000 (12:50 -0700)]
msp430 usci: TX operations via DMA

USCI UART A0 and A1 support DMA transmit via Msp430UartDmaA0C and
Msp430UartDmaA1C, respectively.  Since both components are implemented
using DMA channel 0, they cannot be both utilized in the same program.

A future parameterized DMA interface could be used to allow
serialization of DMA operations between UARTs using a single DMA
channel.

9 years agomsp430 usci: add setIfgTx(), setIfgRx()
R. Steve McKown [Tue, 23 Dec 2014 19:42:27 +0000 (12:42 -0700)]
msp430 usci: add setIfgTx(), setIfgRx()

With upcoming support for DMA transfers with the USCI peripheral,
support for setting the IFG bits will be required.

9 years agomsp430: correct DMA interrupt vector name
R. Steve McKown [Tue, 23 Dec 2014 19:41:47 +0000 (12:41 -0700)]
msp430: correct DMA interrupt vector name

9 years agoInvoke.nc: new interface
R. Steve McKown [Fri, 31 Oct 2014 22:16:53 +0000 (16:16 -0600)]
Invoke.nc: new interface

9 years agoMass platform updates for GCC 4.6.3
R. Steve McKown [Tue, 19 Aug 2014 15:37:49 +0000 (09:37 -0600)]
Mass platform updates for GCC 4.6.3

* -mmcu option now uses different strings to reference specific msp430
  parts.  For example, where the string msp430x1611 was valid before,
  the new equivalent string is msp430f1611.
* The option -mdisable-hwmul has been changed to -mmpy=none

9 years agoMass platform updates for GCC 4.6.3
R. Steve McKown [Tue, 19 Aug 2014 15:31:09 +0000 (09:31 -0600)]
Mass platform updates for GCC 4.6.3

11 years agoFix TMI copyright attributions
R. Steve McKown [Wed, 3 Apr 2013 14:50:54 +0000 (08:50 -0600)]
Fix TMI copyright attributions

TMI copyright attributions, originally derived from another attribution
in TinyOS source code, failed to replace all instances of the entity
name in the original attribution.  This problem was found because that
original entity name contains characters that breaks nesdoc (aka make
<platform> docs).

11 years agoAdd GetSetNow interface
R. Steve McKown [Tue, 2 Apr 2013 21:56:42 +0000 (15:56 -0600)]
Add GetSetNow interface

11 years agomsp430: 115200 baud entries for 2 and 4 MiHz SMCLK
R. Steve McKown [Tue, 26 Mar 2013 14:41:28 +0000 (08:41 -0600)]
msp430: 115200 baud entries for 2 and 4 MiHz SMCLK

1 MiHz = 1024^2 Hz

11 years agoUpdate tos-bsl for win32 python
R. Steve McKown [Fri, 15 Mar 2013 15:04:38 +0000 (09:04 -0600)]
Update tos-bsl for win32 python

tos-bsl uses the cp210xrt module, whose source is provided herein.  This
is a simple extension for Python that allows access to the
CP210x_WriteLatch function in the CP210xRuntime.dll.  The compiled
Python extension module is included, as well as the runtime dll and its
associated .lib and .h files.  The module was compiled using Visual
Studio C++ Express 2008, which is apparently the same version of the
MSVC used to compile ActivePython 2.7.1 from ActiveState, on which the
module was tested.

This uses the latest CP210xRuntime.dll and requires a 6.x version of the
SiLabs VCP to be installed.  Tested with VCP_Windows v6.6.1.

The code changes should not change behavior on non-win32 platforms.

11 years agotos-storage-stm25p: fix family support
R. Steve McKown [Thu, 10 Jan 2013 16:56:33 +0000 (09:56 -0700)]
tos-storage-stm25p: fix family support

The original commit, 125997899ab48cee5128a2dbf89776a4f446935d contained
an error where no family file would cause program abort.

11 years agoSearch for Makelocal in TOSMAKE_PATH
R. Steve McKown [Sun, 9 Dec 2012 17:35:34 +0000 (10:35 -0700)]
Search for Makelocal in TOSMAKE_PATH

User code may provide its own Makelocal via TOSMAKE_PATH.
This only works correct if TOS_MAKE_PATH is added to the end of
TOSMAKE_PATH so local paths take precedence.

11 years agodebug: add cfprintf macro
R. Steve McKown [Sun, 28 Oct 2012 20:01:40 +0000 (14:01 -0600)]
debug: add cfprintf macro

Where cprintf implicitly uses a Uprint interface, cfprintf is provided
the interface name.

11 years agotmimsp: clean up open drain for sht1x DATA
R. Steve McKown [Sat, 27 Oct 2012 03:55:30 +0000 (21:55 -0600)]
tmimsp: clean up open drain for sht1x DATA

Rename Msp430OpenDrainC to Msp430ODPUC and fix it up.  This is an
open-drain component that also provides an internal pull-up so an
external one is not required.  It only works on MSP430 pins offering
internal pull up/down feature.

The OpenDrainC component provides an open-drain emulation, does not
provide an internal pull-up, and should work for any microcontroller
having a GeneralIO implementation for its GPIO pins.

11 years agosht11: correct measurement timeout values
R. Steve McKown [Sat, 27 Oct 2012 03:28:55 +0000 (21:28 -0600)]
sht11: correct measurement timeout values

11 years agotmimsp: support 4 KB sectors on AT25DF041A
R. Steve McKown [Tue, 23 Oct 2012 23:45:35 +0000 (17:45 -0600)]
tmimsp: support 4 KB sectors on AT25DF041A

The tmimsp platform family by default uses an Atmel AT25DF041A serial
FLASH part which supports 4 KB erase blocks.  Since the part's
communications protocol is a super-set of that supported by the STM25P
family, a private Stm25PSpiP that issues 0x20 4 KB erase block instead
of the default 0xd8 64 KB erase block command, and a companion Stm25p.h
that properly defines the sector size (aka erase block size) allows the
part to be treated as an array of 4 KB sectors.  This is a huge win for
most application uses of FLASH.

11 years agoAdd debug library; use with tmimsp platforms
R. Steve McKown [Mon, 15 Oct 2012 00:59:23 +0000 (18:59 -0600)]
Add debug library; use with tmimsp platforms

11 years agoAdd uprintf library; use it in tmimsp platforms
R. Steve McKown [Mon, 15 Oct 2012 00:55:11 +0000 (18:55 -0600)]
Add uprintf library; use it in tmimsp platforms

11 years agotmimsp family: flash is now working
R. Steve McKown [Thu, 4 Oct 2012 23:22:18 +0000 (17:22 -0600)]
tmimsp family: flash is now working

The AT25DF family, on power up or reset (not reset of the board, but the
AT25DF part itself) has by default write protection active for its
sectors.  This code change in MotePlatformC for both parts in the family
globally unprotects all sectors.  At this point, the AT25DF family are
functionally equivalent to the ST/Numonyx/whomever M25P family, so the
latter's driver works for the former.

11 years agotmimsp family: configure flash chip support
R. Steve McKown [Thu, 4 Oct 2012 16:24:39 +0000 (10:24 -0600)]
tmimsp family: configure flash chip support

11 years agoTeach tos-storage-stm25p about platform families
R. Steve McKown [Thu, 4 Oct 2012 16:23:55 +0000 (10:23 -0600)]
Teach tos-storage-stm25p about platform families

11 years agotmimsp platform family better uart abstraction
R. Steve McKown [Thu, 4 Oct 2012 15:27:12 +0000 (09:27 -0600)]
tmimsp platform family better uart abstraction

11 years agotmicore and tmirws platforms -> tmimsp family
R. Steve McKown [Thu, 4 Oct 2012 15:16:13 +0000 (09:16 -0600)]
tmicore and tmirws platforms -> tmimsp family

11 years agoTemp fix for ADC_VECTOR undefined for msp430f2x1x
R. Steve McKown [Tue, 2 Oct 2012 02:04:06 +0000 (20:04 -0600)]
Temp fix for ADC_VECTOR undefined for msp430f2x1x

11 years agoAdded a temporary workaround that allows to compile msp430 ADC code with mspgcc versi...
jan.hauer [Wed, 9 Mar 2011 15:26:12 +0000 (15:26 +0000)]
Added a temporary workaround that allows to compile msp430 ADC code with mspgcc version >= 4. The workaround is re-defining the ADC bitfield structs, see also https://www.millennium.berkeley.edu/pipermail/tinyos-devel/2011-March/004804.html

git-svn-id: http://tinyos-main.googlecode.com/svn/trunk@5503 285f258d-8064-b571-fbc6-9656dc998e90

11 years agoRevert "GCC 4.4.3 seems to have proper bitfield support."
R. Steve McKown [Tue, 2 Oct 2012 01:44:07 +0000 (19:44 -0600)]
Revert "GCC 4.4.3 seems to have proper bitfield support."

This reverts commit 6c5c2220161bf2df3244c860a009e6438f6945e8.

12 years agoRemove hwmul-disable; it works in newer GCC release/2.1.1-4.5
R. Steve McKown [Thu, 3 Nov 2011 23:55:24 +0000 (17:55 -0600)]
Remove hwmul-disable; it works in newer GCC

12 years agoUpdates for GCC 4.5.3
R. Steve McKown [Thu, 3 Nov 2011 23:45:36 +0000 (17:45 -0600)]
Updates for GCC 4.5.3

* Include correct GCC headers depending upon version
* GCC does not stop the watchdog, so do it in PlatformP.nc
* Stopping watchdog in MotePlatformC no longer needed (olimexp16,
  tmicore, tmirws)
* tmicore writes TxIV in MoteClockP so needs MSP430REG_NORACE()

12 years agoSupport %F substitution for @includes
R. Steve McKown [Wed, 21 Mar 2012 15:27:48 +0000 (09:27 -0600)]
Support %F substitution for @includes

.platform and .family files can use %F when adding to includes.  %F is
replaced with the family dir at build time.

12 years agoAllow custom platform families
R. Steve McKown [Wed, 21 Mar 2012 14:50:47 +0000 (08:50 -0600)]
Allow custom platform families

ncc can find a custom platform if its PLATFORMDIR is included into
PFLAGS as an include: -I$(PLATFORMDIR) or equivalent.  Teach ncc to
identify when such a platform is a family member and in such cases
process the corresponding .family file.

13 years agoAdd types for ports 7 and 8, allowing TOSH_ASSIGN_PIN() and friends.
R. Steve McKown [Mon, 29 Nov 2010 23:48:20 +0000 (16:48 -0700)]
Add types for ports 7 and 8, allowing TOSH_ASSIGN_PIN() and friends.

13 years agoUSCI I2C support.
R. Steve McKown [Wed, 27 Oct 2010 23:54:17 +0000 (17:54 -0600)]
USCI I2C support.

Current limitations and features:

* Master mode only.
* No multi-master bus configurations.
* Will automatically reset a hung bus, if for example the master reset in the
  middle of a transaction and a slave currently is driving SDA low.
* Repeated starts are supported and can be very time efficient.
* An I2C transactions may be processed with multiple I2CPacket.read() or
  I2CPacket.write() commands, as appropriate.  The first must include I2C_START
  and the last must include I2C_STOP.
* Conditions exist during read where the hardware may clock out an extra byte
  from the slave before stop.  This byte is not passed to the user.

13 years agoi2cNak() -> i2cNack()
R. Steve McKown [Fri, 29 Oct 2010 18:03:01 +0000 (12:03 -0600)]
i2cNak() -> i2cNack()

Make this change because the MSP430 documentation calls the bit UCNACK...

13 years agoFix a typo in a comment.
R. Steve McKown [Wed, 27 Oct 2010 23:53:52 +0000 (17:53 -0600)]
Fix a typo in a comment.

13 years agoFix up io pin restoration after Uart peripheral is released. release/2.1.1-4.4
R. Steve McKown [Mon, 27 Sep 2010 18:21:51 +0000 (12:21 -0600)]
Fix up io pin restoration after Uart peripheral is released.

13 years agoImplement proper request states in SharedArbiterP.
R. Steve McKown [Fri, 24 Sep 2010 22:33:50 +0000 (16:33 -0600)]
Implement proper request states in SharedArbiterP.

13 years agoRace condition clearing bit vectors that are already clear in Init.init.
R. Steve McKown [Fri, 24 Sep 2010 22:32:11 +0000 (16:32 -0600)]
Race condition clearing bit vectors that are already clear in Init.init.

There's no point to clear the bit vectors in SharedArbiterP.Init.init,
because the compiler guarantees they are already clear, see BitVectorC.nc.
However, manually clearing them can create a situation where someone requesting
a shared resource has their requesting bit erased, depending upon the order of
calls to Init.init providers during boot.

13 years agoFix client immediateRequest() if S_IDLE and ResourceDefaultOwner releases.
R. Steve McKown [Tue, 21 Sep 2010 17:04:24 +0000 (11:04 -0600)]
Fix client immediateRequest() if S_IDLE and ResourceDefaultOwner releases.

13 years agoAdd SharedSplitControlC, for N clients needing shared access to something.
R. Steve McKown [Sun, 19 Sep 2010 21:53:50 +0000 (15:53 -0600)]
Add SharedSplitControlC, for N clients needing shared access to something.

13 years agoSharedArbiterP doesn't need to know the number of clients.
R. Steve McKown [Sun, 19 Sep 2010 15:20:16 +0000 (09:20 -0600)]
SharedArbiterP doesn't need to know the number of clients.

That's being handled indrectly by the BitVectors.

13 years agoNew Write and WriteNow interfaces.
R. Steve McKown [Fri, 17 Sep 2010 01:35:25 +0000 (19:35 -0600)]
New Write and WriteNow interfaces.

13 years agoConform SharedArbiterP better to TEP108.
R. Steve McKown [Wed, 15 Sep 2010 18:44:11 +0000 (12:44 -0600)]
Conform SharedArbiterP better to TEP108.

13 years agoAdd SharedArbiterC.
R. Steve McKown [Wed, 15 Sep 2010 18:44:11 +0000 (12:44 -0600)]
Add SharedArbiterC.

13 years agoAdd an async version of ReadRef.
R. Steve McKown [Wed, 15 Sep 2010 01:09:29 +0000 (19:09 -0600)]
Add an async version of ReadRef.

13 years agoAdd 19200 baud definition when using 1MHz (10^6 Hz) SMCLK.
R. Steve McKown [Tue, 14 Sep 2010 23:46:56 +0000 (17:46 -0600)]
Add 19200 baud definition when using 1MHz (10^6 Hz) SMCLK.

And change baud parameter names using 1E6MHZ to use 1E6HZ.  The clock is not
running at 1 GHz...

13 years agoGCC 4.4.3 seems to have proper bitfield support.
R. Steve McKown [Fri, 3 Sep 2010 21:43:21 +0000 (15:43 -0600)]
GCC 4.4.3 seems to have proper bitfield support.

13 years agoVolatile keyword tricks GCC into not optimizing out busy wait loop.
R. Steve McKown [Wed, 1 Sep 2010 02:36:44 +0000 (20:36 -0600)]
Volatile keyword tricks GCC into not optimizing out busy wait loop.

13 years agoPlatformUartC provides Resource; PlatformSerialC provides StdControl.
R. Steve McKown [Fri, 30 Jul 2010 22:34:41 +0000 (16:34 -0600)]
PlatformUartC provides Resource; PlatformSerialC provides StdControl.

PlatformSerialC is then derived from PlatformUartC.  Only for TMI created
platforms for now.

13 years agoTOS expects an enabled UART to have rx interrupt on by default.
R. Steve McKown [Mon, 2 Aug 2010 20:26:48 +0000 (14:26 -0600)]
TOS expects an enabled UART to have rx interrupt on by default.

In other words, one will receive UartStream.received() without first having to
call UartStream.enableReceiveInterrupt.

13 years agoEnsure Uart state variables are reset on resource config and unconfig.
R. Steve McKown [Mon, 2 Aug 2010 20:26:29 +0000 (14:26 -0600)]
Ensure Uart state variables are reset on resource config and unconfig.

13 years agoInterrupts.tx() signal is only relevant when UartStream.send() pending.
R. Steve McKown [Mon, 2 Aug 2010 20:25:37 +0000 (14:25 -0600)]
Interrupts.tx() signal is only relevant when UartStream.send() pending.

13 years agoAdd TINYOS_MAKE_PATH to TOSMAKE_PATH at end not beginning.
R. Steve McKown [Fri, 23 Jul 2010 14:52:51 +0000 (08:52 -0600)]
Add TINYOS_MAKE_PATH to TOSMAKE_PATH at end not beginning.

2.1.0 added it to the end.  2.1.1 added it to the beginning.  Revert the change
done for 2.1.1.

13 years agoUpdate tos-bsl. cp210x driver ioctls now based on SIOCDEVPRIVATE (0x89f0). release/2.1.1-4.3
R. Steve McKown [Wed, 26 May 2010 19:30:38 +0000 (13:30 -0600)]
Update tos-bsl.  cp210x driver ioctls now based on SIOCDEVPRIVATE (0x89f0).

13 years agoMerge TinyOS 2.1.1 into master.
R. Steve McKown [Fri, 7 May 2010 20:42:29 +0000 (14:42 -0600)]
Merge TinyOS 2.1.1 into master.

How this merge was done:
* Rebase patchset/2.1.0-4.2 onto tinyos/2.1.1 as topic/patchup211
* cp topic/patchup211 tree to tmp dir
* Checkout master, rm all files, cp in from tmp dir, git add .

Future work on TinyOS 2.1.0 w/TMI enhancements via release/2.1.0.

13 years agoProper MSP_BSL_FLAGS for telosa and telosb based platforms using TMI modular tos...
R. Steve McKown [Thu, 6 May 2010 19:30:48 +0000 (13:30 -0600)]
Proper MSP_BSL_FLAGS for telosa and telosb based platforms using TMI modular tos-bsl.

13 years agoUse HOME instead of tilde. Consistent use of quotes around TOSCFG. release/2.1.0 release/2.1.0-4.2
R. Steve McKown [Sat, 24 Apr 2010 01:32:09 +0000 (19:32 -0600)]
Use HOME instead of tilde.  Consistent use of quotes around TOSCFG.

13 years agoThe previous TOSROOT was being writting to TOSCFG.
R. Steve McKown [Sat, 24 Apr 2010 01:31:22 +0000 (19:31 -0600)]
The previous TOSROOT was being writting to TOSCFG.

13 years agoDon't hard-code for ncc in mig and ncg scripts. release/2.1.0-4.1
R. Steve McKown [Fri, 23 Apr 2010 22:39:53 +0000 (16:39 -0600)]
Don't hard-code for ncc in mig and ncg scripts.

The rationalization for this change is that nesc shouldn't really know where
tinyos-tools is installed, since they are separate packages.  The pragmatic
reason for this change is that it allows using ncc from tinyos-tools before
installation to allow a single debian tinyos source package to build both
tinyos-tools and tinyos-source-VERSION packages.

13 years agoGet rid of TOS and use TOSDIR like the other Makefiles in support/sdk/java.
R. Steve McKown [Fri, 23 Apr 2010 21:52:33 +0000 (15:52 -0600)]
Get rid of TOS and use TOSDIR like the other Makefiles in support/sdk/java.

13 years agoFix trailing spaces in cp210xrtmodule.c
R. Steve McKown [Fri, 23 Apr 2010 18:30:21 +0000 (12:30 -0600)]
Fix trailing spaces in cp210xrtmodule.c

13 years agoFix tmicore's PlatformSerialC to provide the conventional set of interfaces.
R. Steve McKown [Fri, 23 Apr 2010 18:24:41 +0000 (12:24 -0600)]
Fix tmicore's PlatformSerialC to provide the conventional set of interfaces.

13 years agoImproved tinyos.sh
R. Steve McKown [Fri, 23 Apr 2010 16:25:39 +0000 (10:25 -0600)]
Improved tinyos.sh

* Automatically installs tinyos.sh into user shell rc files when sourced.
* Adds an uninstall option if a user no longer wants to do TinyOS dev.
* Smarter handling of CLASSPATH.
* tinyos.jar is added to CLASSPATH instead of support/sdk/java.  The latter is
  a TOS 1.x idiom.
* Better error and help messages.

14 years agoFix copyrights now that Sporian has granted TMI the right to create derivative release/2.1.0-4
R. Steve McKown [Thu, 22 Apr 2010 22:39:08 +0000 (16:39 -0600)]
Fix copyrights now that Sporian has granted TMI the right to create derivative
works releasable under tos-bsl's BSD license.

14 years agoSubstitue for newest getTime32() by Miklos.
R. Steve McKown [Sat, 13 Feb 2010 01:19:10 +0000 (18:19 -0700)]
Substitue for newest getTime32() by Miklos.

14 years agoUpstream commits to tos/chips/cc2420, Aug 7 2008 thru Jul 16 2009, except one.
R. Steve McKown [Fri, 13 Nov 2009 23:47:49 +0000 (16:47 -0700)]
Upstream commits to tos/chips/cc2420, Aug 7 2008 thru Jul 16 2009, except one.

commit df004818be0944a5459b2d23dd4b49c55b716448
    kusy <kusy> Thu Aug 7 00:06:53 2008 +0000
    Previously reported timestamping bug was only fixed partially: *timesync...

NOT COMMITTED f8120ec47efebf578a6edf99882d009ea6dedcd1
    kusy <kusy> Wed Aug 13 07:37:05 2008 +0000
    Fixing T32khz to TMilli conversion - coeficient 32 needs to be...

commit 42b450d2e6d5eda17245d38f4d2bceb6b60e5400
    rincon <rincon> Mon Aug 18 22:04:15 2008 +0000
    Added a length check to make sure we don't try to send a packet that is...

commit 62d558a6db9cf823c7926fc74ed5610a0dcbad23
    rincon <rincon> Fri Sep 5 20:39:00 2008 +0000
    added more descriptive comments about the header

commit 72c8a4f1a2b4f331aeedc6bfd5262253ce8d6d04
    janhauer <janhauer> Wed Nov 26 10:13:31 2008 +0000
    Added an interface for accessing the RXFIFO in a non-split phase...

commit 10c02f9ece9b5ee20aa93c877d70fa6689ef7a7d
    razvanm <razvanm> Mon Dec 1 19:56:53 2008 +0000
    Remove a redundant definition.

commit 979965da13886883fe7399d60ff4d025f53bfd4c
    razvanm <razvanm> Mon Dec 1 23:51:38 2008 +0000
    Add a comment about the polarity of FIFOP.

commit 98d4eae39755d369b001971ec9c5f1dd741bc82d
    andreaskoepke <andreaskoepke> Wed Dec 17 17:42:22 2008 +0000
    avoid integer overflow

commit 31bef23ef3377514af595000d2fa2cce07bd4531
    kusy <kusy> Fri Feb 6 06:38:49 2009 +0000
    committing new timestamping code

commit 9be3c9111199378d00acd0d946a0c1772947165a
    kusy <kusy> Mon Mar 2 07:02:32 2009 +0000
    timestamping patch: improved test to fall-through on SFD

commit cd6c99ffa038524d2532860acb79e3ba52540ad7
    kusy <kusy> Thu Jul 16 06:41:41 2009 +0000
    Using AMSenderC infrastructure to fix CTP+FTSP bug

commit fbde8e8b590651b32583d9250bb4921d5b6ab1a3
    kusy <kusy> Thu Jul 16 06:46:07 2009 +0000
    thomas's FTSP patch

14 years agoPatch incorporate changes on CVS head through 8/23/09 for tos/lib/ftsp only.
R. Steve McKown [Tue, 25 Aug 2009 03:13:21 +0000 (21:13 -0600)]
Patch incorporate changes on CVS head through 8/23/09 for tos/lib/ftsp only.

14 years agoFix startup in SoftwareInit.init overwriting changes to StateC's.
R. Steve McKown [Mon, 19 Apr 2010 22:13:15 +0000 (16:13 -0600)]
Fix startup in SoftwareInit.init overwriting changes to StateC's.

If a component wires its Init.init() via SoftwareInit and ncc calls
StateImplP.Init.init() after the component's Init.init(), any change to its
StateC, such as call State.forceState(), is undone.  The easy fix is to remove
the extra initiation of state[] in StateImplP.Init.init(), since state[] is a
static and is guaranteed to have members with value zero (S_IDLE).

14 years agoUSCI fixes.
R. Steve McKown [Thu, 8 Apr 2010 01:16:46 +0000 (19:16 -0600)]
USCI fixes.

* Watch out for sendData() clobbering TXBUF.
* May solve a loss of UART TX chars.
* USCI UartStream.send() gains nothing by immediately sending the 1st byte.
* m_sobuf cannot be null in Interrupts.tx, so remove the check.

14 years agoLet UartByte.send() interleave with any in-progress UartStream.send().
R. Steve McKown [Thu, 8 Apr 2010 13:46:32 +0000 (07:46 -0600)]
Let UartByte.send() interleave with any in-progress UartStream.send().

14 years agoSome comments about the UART ISRs.
R. Steve McKown [Thu, 8 Apr 2010 00:46:48 +0000 (18:46 -0600)]
Some comments about the UART ISRs.

14 years agoRemove redundant atomic blocks.
R. Steve McKown [Thu, 8 Apr 2010 00:17:22 +0000 (18:17 -0600)]
Remove redundant atomic blocks.

14 years agoFix USCI hanging caused by clearing the TXIFG bit.
R. Steve McKown [Thu, 8 Apr 2010 00:12:00 +0000 (18:12 -0600)]
Fix USCI hanging caused by clearing the TXIFG bit.

Also removed the commands that allow set/clear of TXIFG, as there is never a
valid reason to do so.

14 years agoImprove interrupt handling for USCI.
R. Steve McKown [Thu, 8 Apr 2010 00:02:56 +0000 (18:02 -0600)]
Improve interrupt handling for USCI.

Because some interrupts share an interrupt vector, we use the relevant interrupt
enable bit to decide if any user is interested in the interrupt.  If not, that
interrupt is skipped so that another interrupt in the if/then/elseif chain can
be taken.  We also defensively try to prevent I2C interrupts from being a
problem if not used, although no I2C code using these interrupts has yet been
written.

14 years agoFix last UartSend.send() lost when immediately followed by UartStream.send().
R. Steve McKown [Wed, 7 Apr 2010 19:54:02 +0000 (13:54 -0600)]
Fix last UartSend.send() lost when immediately followed by UartStream.send().

14 years agoFix up a duplicate error message in tos-storage-stm25p.
R. Steve McKown [Tue, 13 Apr 2010 19:20:44 +0000 (13:20 -0600)]
Fix up a duplicate error message in tos-storage-stm25p.

14 years agoMake RS-232 signalling docs in tos-bsl easier to understand.
R. Steve McKown [Mon, 12 Apr 2010 23:04:37 +0000 (17:04 -0600)]
Make RS-232 signalling docs in tos-bsl easier to understand.

14 years agoTeach tos-storage-stm25p to derive NUM_SECTORS and SECTOR_SIZE.
R. Steve McKown [Thu, 8 Apr 2010 19:50:41 +0000 (13:50 -0600)]
Teach tos-storage-stm25p to derive NUM_SECTORS and SECTOR_SIZE.

The stm25p driver is suitable for different sizes of flash parts, such as the
commonly used 8Mb M25P80 and the less commonly used 1Mb M25P10A.  The driver
defines the flash configuration in Stm25p.h.  By having tos-storage-stm25p walk
the @includes for the platform, as found in the .platform file, each platform
can ensure that a Stm25p.h with correct content for its flash part is found
for use by the driver and this script.

14 years agoOlimex P169 platform.
R. Steve McKown [Thu, 4 Mar 2010 20:37:26 +0000 (13:37 -0700)]
Olimex P169 platform.

14 years agoRemove unnecessary wiring for tmicore and derivative platforms.
R. Steve McKown [Thu, 4 Mar 2010 23:36:58 +0000 (16:36 -0700)]
Remove unnecessary wiring for tmicore and derivative platforms.

14 years agoUse platform variable name to build make dependency.
R. Steve McKown [Thu, 4 Mar 2010 23:16:48 +0000 (16:16 -0700)]
Use platform variable name to build make dependency.

Minimize the patch to upstream official, so don't make the same modification to
target files for platforms we didn't create.

14 years agoUSCI using uint8_t for len instead of required uint16_t.
R. Steve McKown [Thu, 25 Feb 2010 21:00:16 +0000 (14:00 -0700)]
USCI using uint8_t for len instead of required uint16_t.

14 years agoFix typo in msp430 BusyWait32khzC.nc.
R. Steve McKown [Wed, 3 Feb 2010 01:12:52 +0000 (18:12 -0700)]
Fix typo in msp430 BusyWait32khzC.nc.

14 years agoMake USCI interrupt comments more clear and accurate.
R. Steve McKown [Fri, 15 Jan 2010 15:42:58 +0000 (08:42 -0700)]
Make USCI interrupt comments more clear and accurate.

14 years agoCollapse two sections that are exactly the same.
R. Steve McKown [Fri, 15 Jan 2010 15:32:26 +0000 (08:32 -0700)]
Collapse two sections that are exactly the same.

14 years agoSeparate filters in tools/release/tinyos.files for use by debian packaging too. release/2.1.0-3
R. Steve McKown [Mon, 7 Dec 2009 17:29:57 +0000 (10:29 -0700)]
Separate filters in tools/release/tinyos.files for use by debian packaging too.

14 years agoAdd tinyos.sh, a utility for changing the active tree (TOSROOT).
R. Steve McKown [Thu, 3 Dec 2009 20:45:13 +0000 (13:45 -0700)]
Add tinyos.sh, a utility for changing the active tree (TOSROOT).

14 years agoUpdate tools/release/tinyos.files so it also works with debian/git builds.
R. Steve McKown [Mon, 7 Dec 2009 15:50:36 +0000 (08:50 -0700)]
Update tools/release/tinyos.files so it also works with debian/git builds.

14 years agoRemove unnecessary execute permission on text files.
R. Steve McKown [Thu, 3 Dec 2009 19:14:02 +0000 (12:14 -0700)]
Remove unnecessary execute permission on text files.

14 years agoFormally indicate support/sdk/c/sf/bootstrap is a shell script.
R. Steve McKown [Thu, 3 Dec 2009 19:13:31 +0000 (12:13 -0700)]
Formally indicate support/sdk/c/sf/bootstrap is a shell script.

14 years agoWe don't want README.mkd in the tree right now, though.
R. Steve McKown [Mon, 7 Dec 2009 15:39:42 +0000 (08:39 -0700)]
We don't want README.mkd in the tree right now, though.

14 years agoA possible README.mkd for when this repo is made public.
R. Steve McKown [Thu, 3 Dec 2009 02:55:59 +0000 (19:55 -0700)]
A possible README.mkd for when this repo is made public.

This file would work with github, for example.

14 years agoRemove unneeded references to oww dirs in tmirws .platform release/2.1.0-2
R. Steve McKown [Thu, 3 Dec 2009 00:39:34 +0000 (17:39 -0700)]
Remove unneeded references to oww dirs in tmirws .platform

14 years agoRemove sensors code from tmirws platform.
R. Steve McKown [Wed, 2 Dec 2009 23:37:00 +0000 (16:37 -0700)]
Remove sensors code from tmirws platform.

The sensors code is not stable enough for inclusion into core tinyos, nor is
there yet a clean separation between platform code and application code.  The
sensors code removed here was imported with its history into the RWS
application repository.

14 years agoMove NoMsp430GpioC out of tos/chips/msp430 and into tos/platforms/tmicore
R. Steve McKown [Wed, 2 Dec 2009 22:35:51 +0000 (15:35 -0700)]
Move NoMsp430GpioC out of tos/chips/msp430 and into tos/platforms/tmicore

While this component is very generic for msp430, there is no compelling case
yet to argue for its inclusion into core tinyos.

14 years agoUse AverageAngleC in AeroVaneReadC.
smckown [Mon, 9 Nov 2009 19:31:01 +0000 (19:31 +0000)]
Use AverageAngleC in AeroVaneReadC.

14 years agoAddress possible dead zone discontinuities when reading in WindVaneReadP.
smckown [Mon, 9 Nov 2009 19:31:00 +0000 (19:31 +0000)]
Address possible dead zone discontinuities when reading in WindVaneReadP.

14 years agoTeach aerovector_t that it can be null, via a special value in its dir member.
smckown [Mon, 9 Nov 2009 19:30:59 +0000 (19:30 +0000)]
Teach aerovector_t that it can be null, via a special value in its dir member.

14 years agoRedo Average interface.
smckown [Mon, 9 Nov 2009 19:30:58 +0000 (19:30 +0000)]
Redo Average interface.

There is no point in having the averagers pre-know their number of samples.
The user should know this information.  Therefore we have a more flexible
solution and avoid a signal to return the averate to the user when that user is
very likely to want to call a command instead.