[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20120627153023.GF17154@phenom.dumpdata.com>
Date: Wed, 27 Jun 2012 11:30:23 -0400
From: Konrad Rzeszutek Wilk <konrad.wilk@...cle.com>
To: Rafael Aquini <aquini@...hat.com>
Cc: Konrad Rzeszutek Wilk <konrad@...nok.org>, linux-mm@...ck.org,
linux-kernel@...r.kernel.org,
virtualization@...ts.linux-foundation.org,
Rusty Russell <rusty@...tcorp.com.au>,
"Michael S. Tsirkin" <mst@...hat.com>,
Rik van Riel <riel@...hat.com>
Subject: Re: [PATCH 1/4] mm: introduce compaction and migration for virtio
ballooned pages
On Wed, Jun 27, 2012 at 12:17:17PM -0300, Rafael Aquini wrote:
> On Tue, Jun 26, 2012 at 07:57:55PM -0400, Konrad Rzeszutek Wilk wrote:
> > > +#if defined(CONFIG_VIRTIO_BALLOON) || defined(CONFIG_VIRTIO_BALLOON_MODULE)
.. snip..
> > > +struct address_space *balloon_mapping;
> > > +EXPORT_SYMBOL(balloon_mapping);
> >
> > Why don't you call this kvm_balloon_mapping - and when other balloon
> > drivers use it, then change it to something more generic. Also at that
> > future point the other balloon drivers might do it a bit differently so
> > it might be that will be reworked completly.
>
> Ok, I see your point. However I really think it's better to keep the naming as
> generic as possible today and, in the future, those who need to change it a bit can
> do it with no pain at all. I believe this way we potentially prevent unnecessary code
> duplication, as it will just be a matter of adjusting those preprocessor checking to
> include other balloon driver to the scheme, or get rid of all of them (in case all
> balloon drivers assume the very same technique for their page mobility primitives).
Either way, if a driver is going to use this, they would need to adjust the
preprocessor checking (as you pointed out) to include: #ifdef CONFIG_HYPERVISORX_BALLOON
in this file. At which point they might as well rename the exported symbol to be more
generic - and do whatever else they need to do (add extra stuff maybe?).
>
> As I can be utterly wrong on this, lets see if other folks raise the same
> concerns about this naming scheme I'm using here. If it ends up being a general
> concern that it would be better not being generic at this point, I'll happily
> switch my approach to whatever comes up to be the most feasible way of doing it.
My point here is that its more of name-space pollution. I've gotten flak on doing
this with drivers - which had very generic sounding names, and it made more sense
to rename them with a proper prefix. You are adding pieces of code for the
benefit of one driver.
But that (getting flak on the namespace) might be because the mailing list where I
had posted had more aggressive reviewers and this one is composed of more mellow folks
who are OK with this. Andrew is the final man - and I am not sure what he
prefers.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists