CMSIS Version 5 will focus on improvements and further industry adoption. The license will be changed to the permissive Apache 2.0 license, to enable contributions from interested third parties.
Support for the new ARMv8-M architecture will be added as well as improvements for ARM Cortex-A/Cortex-M based hybrid devices (with a clear focus on Cortex-M interaction).
The CMSIS-RTOS API and RTX reference implementation with get several enhancements:
CMSIS-Pack will get additions for generic example projects, project templates, and multiple download portals. It will also adopt the Flash loader technology from IAR Systems.
As announced on embedded world, the development repository of CMSIS Version 5 is now available on GitHub: https://github.com/ARM-software/CMSIS_5
ARM invites all interested parties to contribute and/or to provide feedback for the CMSIS project using GitHub.
Embedded
Sorry for the question but I don't know where to ask this.
I'm having problems fetchin CMSIS_5 using repo. I have the following in my default.xml file corresponding to CMSIS
<remote name="arm" fetch="https://github.com/ARM-software"/>
<project name="CMSIS_5" remote="arm" revision="refs/tags/5.7.0" path="cmsis"/>
But I'm getting this error:
git-lfs filter-process --skip: 1: git-lfs: not foundfatal: the remote end hung up unexpectedlyerror.GitError: Cannot checkout CMSIS_5: Cannot initialize work tree for CMSIS_5error: Cannot checkout CMSIS_5
I hope you can help me figure it out
Every so often I come back to this comment and get such a kick out of it.
Because we all see the told you so coming, it's just a matter of time....
> "CMSIS++ seems a long way to go ... but I hope that CMSIS++ or something equivalent (near Posix) will arrive in the next months/years"
hrusch, CMSIS++ is indeed a long way to go, but not that long. right now the reference synchronisation objects (native C++) are fully functional (i.e. passed the Keil CMSIS RTOS validator), and the reference scheduler (also native C++) is "almost" ready (for convenience now it is partly running inside a Darwin process, with Cortex-M support to follow soon).
but if are really interested in CMSIS++, the only way to make it happen sooner is to contribute to the project.
ilg: I will. Working with XMC4800 right now and DAVE (which is an integrated IDE on eclipse) from Infineon. I am tired of old style development in C on uCs and think that is is time to really switch to C++ in the coming projects (Cortex M4 are powerfull enough and 256kByte+ ram is sufficent for a lot of stuff. CMSIS++ seems a long way to go and I may choose an intermediate way right now (project concerns), but I hope that CMSIS++ or something equivalent (near Posix) will arrive in the next months/years.
> I would love to see if CMSIS is going the way Liviu started to go...
hrusch, in CMSIS++ custom allocators are already planned and will be available. anyway, your suggestions are welcomed.