[meta-xilinx] Linux/FreeRTOS AMP vrings

Nathan Rossi nathan at nathanrossi.com
Thu Sep 3 09:25:19 PDT 2015


On Thu, Sep 3, 2015 at 9:03 AM, Edward Wingate <edwingate8 at gmail.com> wrote:
> So I decided to go ahead and just try a kernel using linux-xlnx master
> tip, which is:
> Linux 4.0.0-xilinx-dev #1 SMP PREEMPT Wed Sep 2 15:30:11 PDT 2015
> armv7l GNU/Linux
>
> The vring buffer addresses are correct in this version; they are set
> correctly to 0x1e440xxx.  No more fudging needed on CPU1 side!
>
> Now I don't know the exact relationship between linux-xlnx kernel,
> meta-xilinx yocto layer, nor how new branches (e.g., fido) come into
> play here, so I want to ask how would you proceed?  I'm currently
> using meta-xilinx fido branch.
> 1) Find the commit(s) that fixed this in 4.0 and create a backported
> patch for linux-xlnx 3.14
> 2) Find the commit(s) that fixed this in 4.0 and create a backported
> patch for linux-xlnx-dev using kernel 3.19
> 3) Use linux-xlnx-dev with kernel 4.0, but perhaps point to a more
> stable 4.0 commit than tip (that still contains the fix)
>
> I'd like to go with option 3.  Are there any drawbacks to that?  Is

Option 3 is fine, the only draw back is that you would need to pick a
stable commit which might be problematic. I guess it all depends on
what sort of 'stable' target you need (prototyping/production/etc).

> possible for anyone to make a guesstimate for when a meta-xilinx
> branch with kernel 4.0 will be available?

No guesses here, but I would imagine a bit after it is tagged/released
as part of a Vivado version, possibly the next one (2015.3?) but that
is speculation :).

Regards,
Nathan

>
> Thanks everyone for your help and advice.



More information about the meta-xilinx mailing list