Hi Nathan,
On 04/19/2016 02:48 AM, Nathan Rossi wrote:
For a task this simple, I really don't see any benefits in using HSI,
since it is not as if there is any platform processing it is simply
extracting the pre-generated files. In fact I think there is a
potential for more problems with using the HSI tools (version
incompatibilities are the biggest issue) as opposed to simply
extracting the HDF file, at least for this case.
I agree it is relatively simple, but it is a bad methodology according
to me. The contents of HDF file are tool dependent, they are meant to
work with HSI or XSCT etc. Writing a code to extract contents without
the required tools can cause potential problems in future, for example
if there were changes to HDF contents etc (ex: formats).
The other issue is if this patch is not accepted, the board support will be
missing.
With the changes discussed previously, I am happy to apply it. I doubt
there will be a better solution to this problem any time in the
relatively near future.
The only relatively near future solution will be meta-xilinx-tools.
I have always been interested in having HSI available for use within
Yocto/OE/meta-xilinx. The biggest problem has always been that HSI is
tied to the Xilinx tools which is a really hefty dependency (and the
implied issues with licensing, EULA, support, platform compatibility,
version-ing, etc. that come with it) for what is a relatively simple
task (comparatively to generated bitstreams). Maybe this is something
Xilinx has a solution for? standalone distribution or open-source HSI?
definitely would make it considerably easier for project like Yocto/OE
and others to allow for tool automation.
I think as we evolve providing more features to our customers, there
will be licensing issues (EULA etc). Unfortunately, this will become
unavoidable.
Thanks
Manju
--
_______________________________________________
meta-xilinx mailing list
[email protected]
https://lists.yoctoproject.org/listinfo/meta-xilinx
