Skip to content

Commit b85d8d4

Browse files
poelzipoelzi
committed
new docs
renamed README to NEWS and started a real README
1 parent bec4d09 commit b85d8d4

File tree

2 files changed

+153
-0
lines changed

2 files changed

+153
-0
lines changed

‎NEWS

+129
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,129 @@
1+
2010-06-20 - OKAY TO FLASH THE WATCH WITH THIS PROJECT
2+
It turns out there was never a problem with the build.
3+
I discovered later that I had a Windows 7 problem that
4+
prevented the BM Firmware Update Tool and also TI's
5+
BSL_Scripter tool from accessing the the USB FET.
6+
7+
I've since rebooted my machine and have repeated flashed
8+
the watch with mspgcc4 project with not problems.
9+
10+
Paul
11+
12+
13+
============ FIXED 2010-06-20 ========================
14+
2010-06-18 - DO NOT FLASH YOUR WATCH WITH THIS BUILD!!
15+
Since flashing the watch with this build a couple of days ago
16+
it has been working fine, until I tried to flash it again with
17+
the USB FET and the BM Firmware Flash Tool , the same tool that
18+
worked before I flashed the watch with this build.
19+
20+
Although I was not able to flash the device with the USB FET
21+
I was able to flash it with the RFBSL. But even after reflashing
22+
it with the factory TI firmware I still could not flash it with
23+
the USB FET.
24+
25+
However, using the IAR Workbench debug process I was able to flash
26+
it with the USB FET, and after flashing it this way with the original
27+
TI firmware, the USB FET capability is back. I am able to flash it
28+
once again with the BM Firmware Flash Tool via the USB FET.
29+
30+
So a little research is in order to determine how to incorporate the
31+
USB BSL into this project.
32+
33+
34+
35+
36+
2010-06-16 (2:30 AM CDT) - IT WORKS!!!
37+
I just completed building the project. I then flashed my watch,
38+
and all the functions I have tested work. Hour and Minutes,
39+
Seconds, Month and Day, Day of the Week, and YES! even sync.
40+
41+
42+
2010-06-15 - Project now compiles and links without errors. This
43+
was accomplished by compiling and linking in separate steps.
44+
Thanks go to Daniel Poelzleitherner for this contribution.
45+
Daniel also contributed Python scripts that convert the build's elf
46+
file to a TI txt file. This function is incorporated into the
47+
makefile and is done automatically at the end of the build.
48+
49+
50+
ITEMS NEEDING TESTING: (as of 2010-06-14
51+
52+
1) four intrinsic functions had to be written.
53+
As of 2010-06-13 they are untested and couple of
54+
them probably don't work yet.
55+
56+
2) bsp_msp430_defs.h - A compiler specific stanza for
57+
MSPgcc was added near line #135. This stanza was copied
58+
from the previous stanza specific to IAR and modified. The
59+
only modifications were to the __bsp_ENABLE_INTERRUPTS__()
60+
and __bsp_DISABLE_INTERRUPTS__() macros. The other macros
61+
may or may not work with MSPgcc. They do compile without errors or
62+
warnings. The one macro I am not sure about, haven't looked
63+
into what it means, is __bsp_QUOTED_PRAGMA__(x)
64+
65+
3) FIXED: (see note on 2010-06-15)
66+
With MSPgcc4 this project compiles cleanly (see minor warning
67+
in "Before you build the first time:") but there are linking errors
68+
"warning: internal error: unsupported relocation error"
69+
70+
71+
BEFORE YOU BUILD THE FIRST TIME:
72+
73+
1) Many of the TI source code files #include <intrinsics.h> MSPgcc does
74+
not come with an intrinsics.h file. Therefore the intrinsics.h included
75+
in this project will need to be placed in a system directory; e.g., in
76+
MSPGCC4\lib\gcc\msp430\4.4.3\include mspgcc4/msp430/include
77+
78+
2) I haven't figured out how to have an empty directory in git so that
79+
clones will be complete. So, before the first build, it is necessary to
80+
create the build directory specified in the makefile.
81+
82+
3) I am using Windows 7, and have experiemented with three different 'make'
83+
utilities - CygWin, MinGW, and GNU make. The Unix based CygWin 'make'
84+
had no problems with a blank in a directory name as long as the blank was
85+
escaped, but one or both of the DOS versions - MinGW and/or GNU make - did
86+
have a problem. To avoid this problem I have added an underscore to the
87+
directory "simpliciti/Applications/application/End Device/". It is now
88+
"simpliciti/Applications/application/End_Device/". And the reference in the
89+
makefile has been adjusted accordingly. Other than this one thing the code
90+
should still compile with either IAR, CCS, or MSPgcc4.
91+
92+
4) In general the source files have not been changed in such a way that the
93+
code will not still compile with either IAR or CCS compilers (except for
94+
the aforementioned adding of an underscore). However, I have not included
95+
the precompiled library files that are need to copile with the code-size
96+
limited versions of IAR and CCS.
97+
98+
5) MSPgcc4's signal.h issues an apparently harmless, but annoying warning,
99+
at line 39. This warning can fill a page of the output from make, making it
100+
difficult to identify the real errors. I commented this line out in my copy
101+
of signal.h. You may want to consider doing the same.
102+
103+
104+
HOW THE PORT WAS DONE:
105+
106+
1) All places where I modified the original code hopefully) are marked with my
107+
initials PFS or pfs. The only actual code changes were to the two area
108+
mentioned above in "Items needing testing:" and the #5 below.
109+
110+
2) bluerobin code had to be elimiated because mspgcc cannot link to the provided
111+
bluerobin library and the source code is not available.
112+
113+
3) All compiler specific areas that accommodated IAR and CCS were enhanced to
114+
accommodate MSPgcc4
115+
116+
4) The Simpliciti *.dat files were converted to *.h files
117+
118+
5) Since I never need to look at my watch to determine the year, but I often
119+
look to my watch to determine the day of the week (I'm retired now), in date.c
120+
I changed the date display to display the day of the week (as a number from 0 to 6).
121+
The original code is still there, just commented out, just one line. BTW,
122+
while doing this I noticed that TI did not use the RTC for the time and date
123+
functions which might account for some of the inaccurate timekeeping that has
124+
been reported.
125+
126+
127+
Paul F. Sehorne
128+
2010-06-14
129+

‎README

+24
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,24 @@
1+
== Requirements ==
2+
3+
msp430-gcc4 http://mspgcc4.sourceforge.net/
4+
make
5+
python http://python.org
6+
7+
8+
== HOWTO ==
9+
10+
To configure your image, run:
11+
12+
make config
13+
14+
which will generate a config.h file that contains the settings for your build.
15+
16+
To compile the image run:
17+
18+
make
19+
20+
21+
It is HIGHLY suggested to make a clean build befor you flash the image with:
22+
23+
make clean main
24+

0 commit comments

Comments
 (0)