Re: [Qemu-devel] [PATCH v4 00/54] plugins for TCG

[Peter Maydell]

On Fri, 6 Sep 2019 at 20:53, Alex Bennée <address@hidden> wrote:
> Markus Armbruster <address@hidden> writes:
>
> > Please advise why TCG plugins don't undermine the GPL.  Any proposal to
> > add a plugin interface needs to do that.
>
> I'm not sure what we can say about this apart from "ask your lawyer".

I suspect the underlying confusion here is "the term 'plugin'
has a very wide scope and you and Markus are probably not on
the same page about what this API/ABI is actually doing".

Specifically, this is not the 'plugin ABI that lets you write
device models out of tree' that is sometimes mooted, and the ABI
does not (unless it's changed since I last looked) expose arbitrary
bits of QEMU internals to the plugin or let plugins make calls to
random QEMU functions. It's a very limited scope ABI that allows
QEMU users to write code that can do useful introspection on what
running guest code is doing without having to modify QEMU itself
in non-sustainable non-upstreamable ways.
[If it doesn't have that kind of protection against misuse
then we should add it.]

More generally, your cover letter really needs to be much
more descriptive about what we're trying to do here, and
what the purpose, limitations, etc are here. If you say
"plugins" without giving any detail then you're going to
trigger a lot of (reasonable) reactions from people who
associate that word with a much more generalized "provide
arbitrary bits of QEMU functionality as 3rd-party blobs"
concept...

thanks
-- PMM