Subject: Re: lock-free data structures
To: Jonathan Stone <jonathan@Pescadero.dsg.stanford.edu>
From: None <zvrba@globalnet.hr>
List: tech-kern
Date: 01/05/2006 15:48:06
-----BEGIN PGP SIGNED MESSAGE-----
Hash: RIPEMD160

On Wed, Jan 04, 2006 at 01:57:57PM -0800, Jonathan Stone wrote:
> 
> Huh? To quote ras(9):
> 
> DESCRIPTION
>      Restartable atomic sequences are user code sequences which are guaranteed
>                                       ^^^^^^^^^^^^^^^^^^^
>      to execute without preemption.  
> 
> Emphasis added, not in original.
> 
NHF, but the word "user" is overloaded in this sentence. I took "user code" to
mean "code produced by the user" (i.e. the programmer). I mean, I was surprised
to see something like ras(9) implemented, so I took this sentence to mean: the
user (= programmer) can create his own restartable sequences. (and the rest of
the manpage documents how...)

IMHO, putting "user-level" instead of "user" would be much more explicit.

Anyway, thanks for clarifying the issue.

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)

iD8DBQFDvTGmFtofFpCIfhMRAxHPAJ4vjIh/sV0g286XwFjDGwASnsC+LwCgiZOg
CCFliwBqGlQPi/71/R5tCno=
=alv8
-----END PGP SIGNATURE-----