cmsis-svd 0.1

Install It

You can install the latest stable version from pypi:

pip install -U cmsis-svd

To install the latest development version by doing. If this fails, you may need to update the version of pip you are using.

pip install -U -e 'git+https://github.com/posborne/cmsis-svd.git#egg=cmsis-svd&subdirectory=python'

Example

There’s a lot of information you can glean from the SVDs for various platforms. Let’s say, for instance, that I wanted to see the names and base address of each peripheral on the Freescale K20 (D7). Since the K20 SVD is packaged with the library, I can do the following:

from cmsis_svd.parser import SVDParser

parser = SVDParser.for_packaged_svd('Freescale', 'MK20D7.xml')
for peripheral in parser.get_device().peripherals:
    print("%s @ 0x%08x" % (peripheral.name, peripheral.base_address))

This generates the following output:

FTFL_FlashConfig @ 0x00000400
AIPS0 @ 0x40000000
AIPS1 @ 0x40080000
AXBS @ 0x40004000
DMA @ 0x40008000
FB @ 0x4000c000
FMC @ 0x4001f000
FTFL @ 0x40020000
DMAMUX @ 0x40021000
CAN0 @ 0x40024000
SPI0 @ 0x4002c000
SPI1 @ 0x4002d000
...

Example 2: Convert to JSON

The data structures representing the SVD data have the ability to convert themselves to a dictionary suitable for serialization as JSON. This works recursively. To generate JSON data and pretty print it you can do something like the following:

from cmsis_svd.parser import SVDParser

parser = SVDParser.for_packaged_svd('Freescale', 'MK20D7.xml')
svd_dict = parser.get_device().to_dict()
print(json.dumps(svd_dict, sort_key=True,
                 indent=4, separators=(',', ': ')))

Development

Once you have the code checked out, you can run the following from this directory to install dependencies:

virtualenv env
source env/bin/activate
pip install -r dev-requirements.txt

Then, to run the tests:

nosetests .

There are quite a few SVD files, so the tests take a bit. If you have some extra CPUs to throw at the problem, you can do the following:

nosetes --process=8 .

Where 8 can be replaced with as many processes as you see fit. Generally, 2x the number of processors in your machine is a good starting place.

Contributing

Please open issues and submit pull requests on Github.