Re: [PATCH/RFC] msleep() with hrtimers

[Nish Aravamudan]

On 7/16/07, Ray Lee <ray-lk@xxxxxxxxxxxxx> wrote:
> On 7/16/07, Roman Zippel <zippel@xxxxxxxxxxxxxx> wrote:
> > On Mon, 16 Jul 2007, Jonathan Corbet wrote:
> > > > One possible problem here is that setting up that timer can be
> > > > considerably more expensive, for a relative timer you have to read the
> > > > current time, which can be quite expensive (e.g. your machine now uses the
> > > > PIT timer, because TSC was deemed unstable).
> > >
> > > That's a possibility, I admit I haven't benchmarked it.  I will say that
> > > I don't think it will be enough to matter - msleep() is not a hot-path
> > > sort of function.  Once the system is up and running it almost never
> > > gets called at all - at least, on my setup.
> >
> > That's a bit my problem - we have to consider other setups as well.
> > Is it worth converting all msleep users behind their back or should we
> > just a provide a separate function for those who care?
>
> As a driver author (2.4 timeframe, embedded platform, see gitinc.com
> for the hardware description), I would rather msleep did what it says
> it's going to do. If the current one can wait 20 times longer than you
> ask for, then that's just broken.

Well, before these changes, the only guarantee msleep() could make,
just like the only guarantee schedule_timeout() could make, was that
it would not return early. The 1-jiffy sleep was always tough to deal
with, because of rounding and such. And it's simply exacerbated with
HZ=100. It's not technically 20 times longer in all cases, it's 2
jiffies longer, which depends on HZ, so varies from 2 msecs longer to
20 msecs longer.

Thanks,
Nish
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/