Re: [PATCH 2/3] ptp: add a software clock based on clock_monotonic_raw

[torbenh <torbenh@gmx.de>]

On Fri, Mar 04, 2011 at 07:42:51AM +0100, Richard Cochran wrote:
> On Thu, Mar 03, 2011 at 06:26:13PM +0100, Torben Hohn wrote:
> > +config PTP_1588_CLOCK_SOFTWARE
> > +	tristate "software PTP clock"
> > +	depends on PTP_1588_CLOCK
> > +	help
> > +	  This driver adds a software PTP clock. 
> 
> As John said, this needs a longer description. You can look back at
> my V6 as an example.
> 
>   https://lkml.org/lkml/2010/9/23/220
> 
> > +static int ptp_ixp_adjfreq(struct ptp_clock_info *ptp, s32 ppb)
> 
> Once again, please rename.

i was planning to rename it.
i just posted my patches early on because i had found the permission
problem, and thought the software clock might be useful to play around
with the whole framework.
iE without the software clock, i wouldnt have found the actual problem,
because i dont have the right h/w.


> 
> Better yet, why not just use the V6 example instead?  If we are going
> to have a testing driver, then I would want the simulated ancillary
> code as well.

well... your example code messes with clock_realtime.
i was planning to emit real timed events, not just some simulation.
but i dont completely understand this part of the code yet.


> 
> > +
> > +static struct ptp_clock_info ptp_ixp_caps = {
> > +	.owner		= THIS_MODULE,
> > +	.name		= "software timer",
> > +	.max_adj	= 66666655,
> 
> You should calculate the actual limit of adjustment and use it here.

yeah. i posted it too early...
gonna fix it in the next iteration.

-- 
torben Hohn