lkml.org 
[lkml]   [1999]   [Nov]   [26]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: Request for comments (kdev_t and friends...)
Oliver Xymoron writes:
> On Fri, 26 Nov 1999, Richard Gooch wrote:
>
> > Oliver Xymoron writes:
> > > In fact, we should add at least a void *private so that devices can
> > > be freed of managing their instance data with static arrays indexed
> > > off minor number.
> >
> > With devfs this is really clean. A driver can allocate an instance
> > structure at probe time, and pass the pointer to devfs_register().
> > When the device node is opened, file->private_data is initialised with
> > that pointer. So the driver open() method already has a handle to the
> > device instance structure, without any lookups being required.
>
> That part of the scheme is good, but I'm still not convinced the
> lookup is a big deal. To open a file, you still end up doing a
> lookup that matches name to device (ie a dentry hash
> lookup). There's no reason opening a device couldn't use a similarly
> fast (or even identical!) method without requiring that the /dev dir
> be managed inside the kernel. In fact, I think I earlier suggested
> it's possible to cache the lookup in the inode.

I'm not saying the extra lookup is going to make all the difference in
the world. However:

- one less lookup is always going to be faster

- devfs is cleaner (which is what matters more to me)

- why have an extra lookup when there is a simpler, cleaner method?

Regards,

Richard....
Permanent: rgooch@atnf.csiro.au
Current: rgooch@ras.ucalgary.ca

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

\
 
 \ /
  Last update: 2005-03-22 13:55    [from the cache]
©2003-2011 Jasper Spaans