Re: [PATCH] alternative API for raw devices

["H. Peter Anvin" <hpa@zytor.com>]

Followup to:  <Pine.GSO.4.21.0205011555450.12640-100000@weyl.math.psu.edu>
By author:    Alexander Viro <viro@math.psu.edu>
In newsgroup: linux.dev.kernel
> 
> 	* it's _not_ a character device - stat() will give you S_IFREG.
> 	  To check that <foo> is a new-style raw device call statfs(2) and
> 	  compare .f_type with rawfs magic (0x726177).  It doesn't conflict
> 	  with existing check for raw devices (stat(), check that it's
> 	  a character device and compare major with RAW_MAJOR), so existing
> 	  software can be taught to check for raw devices in
> 	  backwards-compatible way.
> 

I really don't know if it's a good idea for this to be S_IFREG, which
software has a reasonable expecation to behave like a normal file,
which a raw device *definitely* doesn't.

I would really like things like this as well as a lot of the
zero-length magic files in /proc to be S_ISCHR with i_rdev ==
MK_DEV(0,0) (the latter to let user space know that this isn't a
standard device node.)

	-hpa
-- 
<hpa@transmeta.com> at work, <hpa@zytor.com> in private!
"Unix gives you enough rope to shoot yourself in the foot."
http://www.zytor.com/~hpa/puzzle.txt	<amsp@zytor.com>