* Re: unlimited read buffer support on configfs
[not found] ` <1160615452.6389.91.camel@linuxchandra>
@ 2006-10-12 6:54 ` Joel Becker
0 siblings, 0 replies; only message in thread
From: Joel Becker @ 2006-10-12 6:54 UTC (permalink / raw)
To: Chandra Seetharaman; +Cc: Paul Menage, Greg KH, linux-kernel
On Wed, Oct 11, 2006 at 06:10:52PM -0700, Chandra Seetharaman wrote:
> So, what you are saying is that we should not be using configfs, even
> though it fits nicely except the fact that we are not fitting the "one
> file == one attribute" model ?
>
> In other words, write our own file system instead of expanding the
> existing infrastructure (just to have one additional feature) ?
No, I'm not saying that you shouldn't use configfs. Greg is
more adamant than I, actually, on the "file == attribute" model.
Here's the thing. For most users, there is no reason they can't
use configfs for _config_ and something like netlink for bulk data
movement. configfs isn't a kitchen sink, and it never should be.
Now, I know that your group/pids list fits really nicely as a
concept in the configfs tree. You certainly can't be calling a usermode
helper for each fork() and exit(). So this is why we're still having a
discussion and working on it.
> I think we should be talking these in lkml as it is more on the
> philosophical discusiion than a technical discussion.
Fair enough, Cc'd!
Joel
--
"The question of whether computers can think is just like the question
of whether submarines can swim."
- Edsger W. Dijkstra
Joel Becker
Principal Software Developer
Oracle
E-mail: joel.becker@oracle.com
Phone: (650) 506-8127
^ permalink raw reply [flat|nested] only message in thread