Re: silent semantic changes with reiser4

[Alex Zarochentsev]

On Thu, Aug 26, 2004 at 02:16:30PM +0200, Andrea Arcangeli wrote:
> On Thu, Aug 26, 2004 at 01:44:59PM +0200, Christophe Saout wrote:
> > I think you're not exactly understanding what a reiser4 "plugin" is.
> > 
> > Currently plugins are not modules. It just means that there's a defined
> > interface between the reiser4 core and the plugins. Just like you could
> > see filesystems as "VFS plugins".
> 
> but filesystems are exactly always modules (in precompiled kernels at
> least).
> 
> Anyways I don't see any visible EXPORT_SYMBOL in reiser4.

they would be there, I think.
> 
> > 
> > In fact, reiser4 plugins are
> > - users of the reiser4 core API
> > - some of them are implementing Linux VFS methods (thus being some sort
> > of glue code between the Reiser4 storage tree and the Linux VFS)
> 
> then it seems a bit misleading to call those things plugins.

There should be infracture to change plugin on fly, and choosing object's
plugins at object's creation time.  Currently only certain object's plugins
can be changed through <object>/metas/plugin interface. 

> 
> As you wrote those are "users of the reiser4 core API", with plugins I
> always meant dynamically loadable things, like plugins for web browers
> (hence kernel modules in kernel space). While here apparently you can't
> plugin anything at runtime, it's just code that uses the reiser4 core
> API that can be modified with a kernel patch as usual.

-- 
Alex.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/