From: john stultz <johnstul@us.ibm.com>
To: Roman Zippel <zippel@linux-m68k.org>
Cc: lkml <linux-kernel@vger.kernel.org>,
George Anzinger <george@mvista.com>,
frank@tuxrocks.com, Anton Blanchard <anton@samba.org>,
benh@kernel.crashing.org, Nishanth Aravamudan <nacc@us.ibm.com>,
Ulrich Windl <ulrich.windl@rz.uni-regensburg.de>
Subject: Re: [RFC - 0/9] Generic timekeeping subsystem (v. B5)
Date: Mon, 22 Aug 2005 11:57:55 -0700 [thread overview]
Message-ID: <1124737075.22195.114.camel@cog.beaverton.ibm.com> (raw)
In-Reply-To: <Pine.LNX.4.61.0508202204240.3728@scrub.home>
On Mon, 2005-08-22 at 01:19 +0200, Roman Zippel wrote:
> Hi,
>
> On Fri, 19 Aug 2005, john stultz wrote:
>
> > timekeeping_perioidic_hook():
> >
> > /* get ntp adjusted interval length*/
> > interval_length = get_timesource_interval(ppm)
>
> Here starts the problem, this requires more expensive math than necessary,
> as every time you first have to scale the values.
Hmmm. I feel like we're mixing signals. See below for more on this.
> Let's take a standard PIT timer as an example. With HZ=100 we program it
> with 11932, for simplicity let's assume this corresponds to 10^7ns and
> scale this by 2^8. This means the timer multiplier is initially 214549,
> this updates the system time by 214549*11932 and the reference time by
> 10^7*2^8 every tick. We can now just ignore the error or as soon as it
> exceeds 11932/2 we increase/decrease the mutiplier. The error calculation
> is rather simple, usually just adds and shifts, only if the error exceeds
> 2*11932 it gets a little more complicated, but even here the possible
> divide is avoidable.
I feel like we're talking about different problems. Which reference
clock (other then the system clock) are you wanting to increment at the
tick time? Do you mean the ntp time_offset value? A little bit of psudo
code might go a long way in helping me understand your solution.
Also I'm not sure how this is connects to the continuous timesource
situation where we do not assume timer ticks are not lost or late.
> The gettimeofday would then basically be "xtime + (cycle_offset * mult +
> error_offset) / 2^8". Depending on the update frequency and the required
> precision it's even possible to keep this within 32bit. The ntp part stays
> pretty much the same and the time source can add anything it wants on top
> of that. The basic math is also pretty much the same so we can generate
> most of the code depending on various parameters.
Again, I must not be understanding what you're suggesting. Above where
you called get_timesource_interval(ppm) too expensive, what you're
suggesting here is almost exactly what get_timesource_interval(ppm)
would do. In my timeofday patches, its called cyc2ns() and gettimeofday
looks like:
xtime + cyc2ns(timesource, ntp_adjustment, cycle_delta)
Where cyc2ns does:
(cycle_delta * (timesource->mult + ntp_adjustment))>>timesource->shift
The reason why we calculate the interval_length in the continuous
timesource case is because we are not assuming anything about the
frequency that the timekeeping_periodic_hook() is called.
Again, I'm really wanting to address your concerns, but I still do not
really understand the specific objections.
thanks
-john
next prev parent reply other threads:[~2005-08-22 18:59 UTC|newest]
Thread overview: 71+ messages / expand[flat|nested] mbox.gz Atom feed top
2005-08-11 1:21 [RFC - 0/13] NTP cleanup work " john stultz
2005-08-11 1:23 ` [RFC][PATCH - 1/13] NTP cleanup: Move NTP code into ntp.c john stultz
2005-08-11 1:25 ` [RFC][PATCH - 2/13] NTP cleanup: Move arches to new ntp interfaces john stultz
2005-08-11 1:26 ` [RFC][PATCH - 3/13] NTP cleanup: Remove unused NTP PPS code john stultz
2005-08-11 1:27 ` [RFC][PATCH - 4/13] NTP cleanup: Breakup ntp_adjtimex() john stultz
2005-08-11 1:28 ` [RFC][PATCH - 5/13] NTP cleanup: Break out leapsecond processing john stultz
2005-08-11 1:28 ` [RFC][PATCH - 6/13] NTP cleanup: Clean up ntp_adjtimex() arguement checking john stultz
2005-08-11 1:31 ` [RFC][PATCH - 7/13] NTP cleanup: Cleanup signed shifting logic john stultz
2005-08-11 1:31 ` [RFC][PATCH - 8/13] NTP cleanup: Integrate second_overflow() logic john stultz
2005-08-11 1:33 ` [RFC][PATCH - 9/13] NTP cleanup: Improve NTP variable names john stultz
2005-08-11 1:33 ` [RFC][PATCH - 10/13] NTP cleanup: Use ntp_lock instead of xtime_lock john stultz
2005-08-11 1:35 ` [RFC][PATCH - 11/13] NTP cleanup: Introduce PPM adjustment variables john stultz
2005-08-11 1:36 ` [RFC][PATCH - 12/13] NTP cleanup: cleanup ntp_advance() adjtime code john stultz
2005-08-11 1:38 ` [RFC][PATCH - 13/13] NTP cleanup: drop time_phase and time_adj add copyright john stultz
2005-08-16 2:08 ` [RFC][PATCH - 4/13] NTP cleanup: Breakup ntp_adjtimex() john stultz
2005-08-11 2:13 ` [RFC - 0/9] Generic timekeeping subsystem (v. B5) john stultz
2005-08-11 2:14 ` [PATCH 1/9] Timesource management code john stultz
2005-08-11 2:16 ` [PATCH 2/9] Generic timekeeping core subsystem john stultz
2005-08-11 2:18 ` [PATCH 3/9] Generic timekeeping i386 arch specific changes, part 1 john stultz
2005-08-11 2:19 ` [PATCH 4/9] generic timekeeping i386 arch specific changes, part 2 john stultz
2005-08-11 2:20 ` [PATCH 5/9] generic timekeeping i386 arch specific changes, part 3 john stultz
2005-08-11 2:21 ` [PATCH 6/9] generic timekeeping i386 arch specific changes, part 4 john stultz
2005-08-11 2:23 ` [PATCH 7/9] generic timekeeping i386 arch specific changes, part 5 john stultz
2005-08-11 2:24 ` [PATCH 8/9] generic timekeeping i386 arch specific changes, part 6 john stultz
2005-08-11 2:25 ` [PATCH 9/9] generic timekeeping i386 specific timesources john stultz
2005-08-11 2:32 ` [RFC - 0/9] Generic timekeeping subsystem (v. B5) Lee Revell
2005-08-11 2:39 ` john stultz
2005-08-11 2:44 ` Lee Revell
2005-08-11 6:17 ` Ulrich Windl
2005-08-11 2:37 ` [RFC] Cumulative NTP cleanujp and generic timekeeping patch (v B5) john stultz
2005-08-15 22:14 ` [RFC - 0/9] Generic timekeeping subsystem (v. B5) Roman Zippel
2005-08-16 0:10 ` john stultz
2005-08-16 18:25 ` Christoph Lameter
2005-08-16 23:48 ` john stultz
2005-08-17 0:14 ` Christoph Lameter
2005-08-17 0:17 ` john stultz
2005-08-17 0:21 ` Christoph Lameter
2005-08-17 6:08 ` Ulrich Windl
2005-08-17 14:07 ` Christoph Lameter
2005-08-17 0:28 ` Roman Zippel
2005-08-17 1:17 ` john stultz
2005-08-17 7:40 ` Ulrich Windl
2005-08-19 0:27 ` Roman Zippel
2005-08-20 2:32 ` john stultz
2005-08-21 23:19 ` Roman Zippel
2005-08-22 18:57 ` john stultz [this message]
2005-08-23 11:30 ` Roman Zippel
2005-08-23 18:52 ` john stultz
2005-08-23 20:51 ` john stultz
2005-08-23 21:34 ` Roman Zippel
2005-08-23 23:14 ` john stultz
2005-08-23 23:54 ` Roman Zippel
2005-08-24 0:29 ` George Anzinger
2005-08-24 20:36 ` john stultz
2005-08-24 23:46 ` George Anzinger
2005-08-25 0:42 ` john stultz
2005-08-25 1:44 ` George Anzinger
2005-08-25 2:13 ` john stultz
2005-08-24 6:34 ` Ulrich Windl
2005-08-24 9:47 ` Roman Zippel
2005-08-24 18:00 ` john stultz
2005-08-24 18:48 ` Roman Zippel
2005-08-24 19:15 ` john stultz
2005-08-24 19:49 ` Roman Zippel
2005-08-24 22:40 ` john stultz
2005-08-25 0:45 ` Roman Zippel
2005-08-25 18:08 ` john stultz
2005-08-17 19:03 ` George Anzinger
2005-08-15 22:12 ` [RFC - 0/13] NTP cleanup work " Roman Zippel
2005-08-15 22:46 ` john stultz
2005-08-17 0:10 ` Roman Zippel
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1124737075.22195.114.camel@cog.beaverton.ibm.com \
--to=johnstul@us.ibm.com \
--cc=anton@samba.org \
--cc=benh@kernel.crashing.org \
--cc=frank@tuxrocks.com \
--cc=george@mvista.com \
--cc=linux-kernel@vger.kernel.org \
--cc=nacc@us.ibm.com \
--cc=ulrich.windl@rz.uni-regensburg.de \
--cc=zippel@linux-m68k.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®