On Thu, 2008-06-26 at 13:37 +0800, Li Zefan wrote:
> > In subsequent email echanges Li and Subrata wrote:
> > 
> >> Subrata Modak wrote:
> >>> Li,
> >>>
> >>> Will you be resending me a renewed patch that fixes those issues, so
> >>> that i can merge for the June 2008 LTP release ?
> >>>
> >> Will it be better to hold this until David Miller accepts my patch or
> >> Evgeniy Polyakov extends the connector core to allow querying of
> >> connector
> >> users ?
> > 
> >     I'm not sure exactly what you're referring to.
> > 
> 
> Here:
>       http://lkml.org/lkml/2008/6/23/465
> 
> I should have CCed you in that mail.
> 
> >     Are you concerned that there's no way to test if the interfaces exist?
> > If so, couldn't you consider using main() of pec_listener.c up to the
> > point above as a test to determine if the kernel feature(s) needed for
> > the LTP tests exist? Then you can determine whether or not to run the
> > tests by default based on the result. You might need to check for
> > specific errno values and you might also need to test to be sure netlink
> > sockets have been configured into the kernel too, but I think that may
> > be do-able. 
> > 
> 
> The problem is, even if process event connector is not supported, socket()
> and bind() will return successfully. Just I can recv() nothing, but I can't
> determine it's not supported by this.
> 
> I can check the existence of connector by looking into /proc/net/netlink,
> but this is not enough, one can set CONFIG_CONNECTOR but unse
> CONFIG_PROC_EVENTS.
> 
> >     Either way, I think putting the tests in as they are would be nice. You
> > can always update the tests to reflect your patch or extensions by
> > Evgeniy, but in the meantime your useful tests would be available to LTP
> > users (even if not run by default).
> > 
> 
> So how about put the tests into LTP but not run by default, and when my patch
> is accepted (but I'm not optimistic) or Evgeniy has done the extension, we
> can update the tests and make it run by default.

That would be really great, rather than holding it off from LTP
inclusion. I would merge your last updated patch with little
modifications from your side. I will make sure that it does not run by
default until your kernel patch is through, and in that case, you can
send me an updated patch.

Regards--
Subrata

> 
> >     That's just my 2 cents.
> > 
> 
> Thanks for your comments. :)
> 
> 


-------------------------------------------------------------------------
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://sourceforge.net/services/buy/index.php
_______________________________________________
Ltp-list mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ltp-list

Reply via email to