[vbox-dev] [PATCH RFC] Allow an extension pack to load an HGCM service

[Klaus Espenlaub]

Hi Jeff,

On 30.08.2013 22:12, Jeff Westfahl wrote:
> Klaus,
>
> What's the likelihood of something like this making it into a 4.2 update,

Right now I'm scratching my head why you designed this in such a 
backwards manner (at least that it feels to me right now, maybe it has 
just holes in its functionality which confuses me) - essentially I'm 
lacking an understanding how you think the new callback should be used 
to implement a HGCM module. A little pseudo code explaining where it is 
called (and where it gets its parameters from) should make this far 
easier to understand. Also don't forget to hint how such a loaded HGCM 
service would be used afterwards. I don't see any way for sending 
messages etc etc - please enlighten me :)

I "cleaned up" the modification to the HGCM loader code, it is now 
working the same as in many other places in VirtualBox. If the module 
name has a path then it's a plugin (usually from an extension pack), 
otherwise it's a core module. This simplifies the code, and should be no 
significant problem as the extpack "main module" can rather easily work 
out the absolute path name of the HGCM module:
...
     char szPath[RTPATH_MAX];
     rc = g_pHlp->pfnFindModule(g_pHlp, "MyHGCMModule", NULL, 
VBOXEXTPACKMODKIND_R3, szPath, sizeof(szPath), NULL);
...

> or maybe an early 4.3 update? There seems to be a lot of interest in the
> forums and on the mailing list for a way to add custom HGCM servers and
> clients, and the extension pack interface seems like a good way to support
> this use case.

The plan is to get this into the next VirtualBox 4.3 beta, which may not 
work out if the question marks in my mind don't get addressed.

No need to argue how useful this would be to some people, I'm currently 
more concerned if your contribution does enough to achieve its goal...

We'd hope to get more HGCM/... contributions once developers can test 
drive their new functionality in an extension pack. After all behind 
most HGCM services there is an implicit long term commitment, and thus a 
well validated initial design/implementation is a key item for a 
successful contribution in this area.

Thanks,
Klaus



>
> Jeff Westfahl
>
> Klaus Espenlaub <klaus.espenlaub at oracle.com> wrote on 08/06/2013 12:10:14
> PM:
>
>> From: Klaus Espenlaub <klaus.espenlaub at oracle.com>
>> To: vbox-dev at virtualbox.org,
>> Date: 08/06/2013 12:11 PM
>> Subject: Re: [vbox-dev] [PATCH RFC] Allow an extension pack to load an
> HGCM service
>>
>> Hi Jeff,
>>
>> thanks for this mail, was a bit faster than I could ask for such a
>> statement :)
>>
>> On 06.08.2013 18:03, Jeff Westfahl wrote:
>>> Author: Jeff Westfahl <jeff.westfahl at ni.com>
>>> License: Patch is under MIT License
>>
>> I'm still reviewing the change, which overall looks very sensible. We
>> want the product to be extensible, and this certainly wasn't easily
>> possible before in this particular area.
>>
>> Klaus