Hi,
Something is wrong with your installation. See below.
nobo...@web.de wrote:
Hi,
The FW42x parts have been supported (and the support actually used)
since last summer. That is why the timera1 is known during your
compilation (although with a typo upsetting things) - timera1 was added
to the headers specifically for the FW42x :-)
I tried it:
msp430-gcc -mmcu=msp430xW427 -Os -mendup-at=main -Wall -g -c main.c
Known MCU names:
msp1
msp2
msp430x110
msp430x112
msp430x122
msp430x122
msp430x1222
msp430x1122
msp430x1132
msp430x123
msp430x1232
msp430x133
msp430x135
msp430x147
msp430x148
msp430x149
msp430x412
msp430x413
msp430x311
msp430x312
msp430x313
msp430x314
msp430x315
msp430x323
msp430x325
msp430x336
msp430x337
msp430x1101
msp430x1111
msp430x1121
msp430x1331
msp430x1351
msp430x435
msp430x436
msp430x437
msp430x447
msp430x448
msp430x449
msp430x167
msp430x168
msp430x169
msp430x155
msp430x156
msp430x157
This is the output from a very old version of msp430-gcc. There are more
MCUs in versions less than one year old, and they are sorted into order
now. Which version of gcc are you using? Your script seems to indicate
you are using the right one, but they above output doesn't. The right
versions to use at the moment are GCC 3.2.3 and the mspgcc patches from
the 3.3 directory at the mspgcc site. These do *not* give the output
shown here. At least, they don't for everyone else :-)
I looked at the source file names, found some good looking files and with
-mmcu=msp430xW427 and #include <msp430xW42x.h> the MCU is known, so the above
error list is wrong.
I could start compiling but with the new MCU type i do get many errors:
main.c:69: `CACTL2' undeclared (first use in this function)
...
main.c:71: `CAON' undeclared (first use in this function)
isr.c:109: `CACTL1' undeclared (first use in this function)
...
modules.c:789: `CAF' undeclared (first use in this function)
...
make: *** [main.o] Error 1
What's wrong?
This is just a consequence of the compiler not knowing the device. It
has not selected the right headers, so it does not know these port
names. When the compiler is fixed, this problem will go away. I can use
those names OK with an FW427.
Is there no trick to reduce the size to the size the IAR compiler
produces?
- did you compile for the same target?
Yes, for F413.
- did you try -O2 too?
Yes, with -O2 it's bigger.
I'm using
CPU = msp430xW427
ASMOPT = -mmcu=${CPU} -std=gnu99 -ffunction-sections -fgcse -fssa
-fdelete-null-pointer-checks -fno-math-errno -mno-stack-init -Os
-mendup-at=main -Wall
COPT = -mmcu=${CPU} -std=gnu99 -Os -mendup-at=main -Wall -g
- does the code use floats?
Yes.
Does mspgcc produces more float code than the IAR compilers?
There are not many people using floats, so it hasn't received a lot of
attention. The integer code from mspgcc is generally smaller and faster
than IAR. However, IAR beats mspgcc very easily on floating point code.
Regards,
Steve
