Re: [PATCH] [RFCv2] notify userspace about time changes

[Alexander Shishkin <virtuoso@slind.org>]

On Fri, Aug 20, 2010 at 08:33:46 -0700, Greg KH wrote:
> On Fri, Aug 20, 2010 at 11:37:23AM +0300, Kirill A. Shutemov wrote:
> > On Thu, Aug 19, 2010 at 08:31:27AM -0700, Greg KH wrote:
> > > On Thu, Aug 19, 2010 at 11:36:12AM +0300, Kirill A. Shutemov wrote:
> > > > On Wed, Aug 18, 2010 at 09:09:37PM -0700, Greg KH wrote:
> > > > > On Wed, Aug 18, 2010 at 04:53:03PM -0700, Andrew Morton wrote:
> > > > > > Is sysfs the right interface for this thing?  Bear in mind that
> > > > > > CONFIG_SYSFS does exist.  
> > > > > > 
> > > > > > > +		     fd = open("/sys/kernel/time_notify", O_WRONLY);
> > > > > > > +		     fdprintf(fd, "%d 1 0 1 1", efd);
> > > > > > 
> > > > > > why not
> > > > > > 
> > > > > > 		sys_time_notify(efd, 1, 0, 1, 1);
> > > > > 
> > > > > Yeah, that would be much better than a sysfs file, this is abusing the
> > > > > sysfs interface quite a lot.
> > > > 
> > > > Do you really think, that increasing number of syscalls is better then
> > > > fs-based interfaces?
> > > 
> > > As you are pretty much creating a new syscall here anyway, there is no
> > > problem with making it a real one, right?
> > 
> > I think Linux has too many syscalls. Significant part these interfaces
> > would be better to map to a filesystem[s].
> 
> What is the difference between a syscall and a filesystem interface?
> They are both things that we can not change in the future and need to be
> preserved and documented.
> 
> Don't be afraid of syscall's, they don't bite :)
> 
> > > That way you can properly
> > > handle the user/kernel documentation and persistance over time (i.e. you
> > > can't change it.)
> > 
> > On the other, hand properly designed fs-based interface requires less
> > modification of userspeace to use it. Acctually, you can use most of
> > fs-based intefaces directly from shell. No need in libc modifications and
> > utilities to use it from shell or other script language.
> > See cgroup, for example.
> > 
> > > So yes, a syscall would be better, especially as this does not exactly
> > > fit into the model of sysfs, right?
> > 
> > Yes, sysfs is not the best place for it, but...
> 
> You just answered your own question.  Please don't make it in sysfs,
> make it a syscall as it does not fit into sysfs.

Hmm, how about a syscallfs? :)

Regards,
--
Alex