From: Thomas Gleixner <tglx@linutronix.de>
To: Martin Kaistra <martin.kaistra@linutronix.de>,
Florian Fainelli <f.fainelli@gmail.com>,
Andrew Lunn <andrew@lunn.ch>,
Vivien Didelot <vivien.didelot@gmail.com>
Cc: martin.kaistra@linutronix.de,
Richard Cochran <richardcochran@gmail.com>,
Vladimir Oltean <olteanv@gmail.com>,
"David S. Miller" <davem@davemloft.net>,
Jakub Kicinski <kuba@kernel.org>,
John Stultz <john.stultz@linaro.org>,
Stephen Boyd <sboyd@kernel.org>,
Russell King <linux@armlinux.org.uk>,
Marc Kleine-Budde <mkl@pengutronix.de>,
linux-kernel@vger.kernel.org, netdev@vger.kernel.org
Subject: Re: [PATCH v2 3/7] timecounter: allow for non-power of two overflow
Date: Wed, 24 Nov 2021 21:55:44 +0100 [thread overview]
Message-ID: <874k81l1db.ffs@tglx> (raw)
In-Reply-To: <20211109095013.27829-4-martin.kaistra@linutronix.de>
Martin,
On Tue, Nov 09 2021 at 10:50, Martin Kaistra wrote:
> * see CYCLECOUNTER_MASK() helper macro
> * @mult: cycle to nanosecond multiplier
> * @shift: cycle to nanosecond divisor (power of two)
> + * @overflow_point: non-power of two overflow point (optional),
> + * smaller than mask
> */
> struct cyclecounter {
> u64 (*read)(const struct cyclecounter *cc);
> u64 mask;
> u32 mult;
> u32 shift;
> + u64 overflow_point;
> };
>
> /**
> diff --git a/kernel/time/timecounter.c b/kernel/time/timecounter.c
> index e6285288d765..afd2910a9724 100644
> --- a/kernel/time/timecounter.c
> +++ b/kernel/time/timecounter.c
> @@ -39,6 +39,9 @@ static u64 timecounter_read_delta(struct timecounter *tc)
> /* calculate the delta since the last timecounter_read_delta(): */
> cycle_delta = (cycle_now - tc->cycle_last) & tc->cc->mask;
>
> + if (tc->cc->overflow_point && (cycle_now - tc->cycle_last) > tc->cc->mask)
> + cycle_delta -= tc->cc->mask - tc->cc->overflow_point;
TBH, this took me more than one twisted braincell to grok.
With support for clocks which do not wrap at power of 2 boundaries we
already lose the unconditional fast path no matter what. So what's the
point of having two conditions and doing this convoluted math here?
In timecounter_init():
tc->ovfl = cc->ovfl ? cc->ovfl : cc->mask + 1;
which makes it a common path in timecounter_read_delta():
cycle_delta = cycle_now - tc->cycle_last;
if ((s64)cycle_delta) < 0)
cycle_delta += tc->ovfl;
which produces way better binary code.
The conditional does not really matter for the timecounter use cases as
that calculation is noise compared to the actual cc->read() access.
Aside of that the same problem exists in timecounter_cyc2time()...
After that we probably should do a treewide sweep to get rid of cc->mask
to avoid confusion and subtle to understand errors when some code uses
cc->mask instead of cc->ovfl.
Thanks,
tglx
next prev parent reply other threads:[~2021-11-24 20:55 UTC|newest]
Thread overview: 32+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-11-09 9:50 [PATCH v2 0/7] Add PTP support for BCM53128 switch Martin Kaistra
2021-11-09 9:50 ` [PATCH v2 1/7] net: dsa: b53: Add BroadSync HD register definitions Martin Kaistra
2021-11-09 10:10 ` Kurt Kanzenbach
2021-11-09 18:04 ` Florian Fainelli
2021-11-10 8:19 ` Martin Kaistra
2021-11-09 9:50 ` [PATCH v2 2/7] net: dsa: b53: Move struct b53_device to include/linux/dsa/b53.h Martin Kaistra
2021-11-09 18:05 ` Florian Fainelli
2021-11-09 18:11 ` Florian Fainelli
2021-11-09 18:15 ` Vladimir Oltean
2021-11-09 18:20 ` Florian Fainelli
2021-11-09 18:49 ` Vladimir Oltean
2021-11-09 9:50 ` [PATCH v2 3/7] timecounter: allow for non-power of two overflow Martin Kaistra
2021-11-24 20:55 ` Thomas Gleixner [this message]
2021-11-09 9:50 ` [PATCH v2 4/7] net: dsa: b53: Add PHC clock support Martin Kaistra
2021-11-09 18:08 ` Florian Fainelli
2021-11-09 9:50 ` [PATCH v2 5/7] net: dsa: b53: Add logic for RX timestamping Martin Kaistra
2021-11-09 18:07 ` Florian Fainelli
2021-11-10 8:43 ` Martin Kaistra
2021-11-09 9:50 ` [PATCH v2 6/7] net: dsa: b53: Add logic for TX timestamping Martin Kaistra
2021-11-09 11:12 ` Vladimir Oltean
2021-11-10 7:14 ` Kurt Kanzenbach
2021-11-10 13:05 ` Vladimir Oltean
2021-11-10 13:30 ` Vladimir Oltean
2021-11-10 13:47 ` Kurt Kanzenbach
2021-11-10 14:00 ` Vladimir Oltean
2021-11-10 15:08 ` Richard Cochran
2021-11-10 15:23 ` Vladimir Oltean
2021-11-10 12:57 ` Vladimir Oltean
2021-11-09 9:50 ` [PATCH v2 7/7] net: dsa: b53: Expose PTP timestamping ioctls to userspace Martin Kaistra
2021-11-09 10:39 ` [PATCH v2 0/7] Add PTP support for BCM53128 switch Vladimir Oltean
2021-11-09 11:13 ` Martin Kaistra
2021-11-09 18:13 ` Florian Fainelli
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=874k81l1db.ffs@tglx \
--to=tglx@linutronix.de \
--cc=andrew@lunn.ch \
--cc=davem@davemloft.net \
--cc=f.fainelli@gmail.com \
--cc=john.stultz@linaro.org \
--cc=kuba@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux@armlinux.org.uk \
--cc=martin.kaistra@linutronix.de \
--cc=mkl@pengutronix.de \
--cc=netdev@vger.kernel.org \
--cc=olteanv@gmail.com \
--cc=richardcochran@gmail.com \
--cc=sboyd@kernel.org \
--cc=vivien.didelot@gmail.com \
/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