From: "Sören Brinkmann" <soren.brinkmann@xilinx.com>
To: "Uwe Kleine-König" <u.kleine-koenig@pengutronix.de>
Cc: Mike Turquette <mturquette@linaro.org>,
"Rafael J. Wysocki" <rjw@rjwysocki.net>,
Viresh Kumar <viresh.kumar@linaro.org>,
Russell King <linux@arm.linux.org.uk>,
Michal Simek <michal.simek@xilinx.com>,
<linux-pm@vger.kernel.org>, <linux-kernel@vger.kernel.org>,
<cpufreq@vger.kernel.org>, <linux-arm-kernel@lists.infradead.org>
Subject: Re: [RFC PATCH 2/5] clk: Introduce 'clk_round_rate_nearest()'
Date: Mon, 19 May 2014 10:29:05 -0700 [thread overview]
Message-ID: <df742b46-c630-49ea-b234-4d746bb66ec1@BL2FFO11FD040.protection.gbl> (raw)
In-Reply-To: <db8633e5-01a9-4d3a-80e8-74e23a456d52@BN1BFFO11FD051.protection.gbl>
Hi Uwe,
On Mon, 2014-05-19 at 09:41AM -0700, Sören Brinkmann wrote:
> On Mon, 2014-05-19 at 06:19PM +0200, Uwe Kleine-König wrote:
> > Hi Sören,
> >
> > On Sun, May 18, 2014 at 05:51:05PM -0700, Sören Brinkmann wrote:
> > > ------------------8<-----------------8<---------------------8<-------------8<---
> > > From: Soren Brinkmann <soren.brinkmann@xilinx.com>
> > > Date: Tue, 2 Apr 2013 10:08:13 -0700
> > > Subject: [PATCH] clk: Introduce 'clk_round_rate_nearest()'
> > >
> > > Introduce a new API function to round a rate to the closest possible
> > > rate the HW clock can generate.
> > > In contrast to 'clk_round_rate()' which works similar, but always returns
> > > a frequency <= its input rate.
> > >
> > > Cc: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
> > > Signed-off-by: Soren Brinkmann <soren.brinkmann@xilinx.com>
> > > ---
> > > drivers/clk/clk.c | 43 +++++++++++++++++++++++++++++++++++++++++--
> > > include/linux/clk.h | 14 ++++++++++++--
> > > 2 files changed, 53 insertions(+), 4 deletions(-)
> > >
> > > diff --git a/drivers/clk/clk.c b/drivers/clk/clk.c
> > > index dff0373f53c1..faf24d0569df 100644
> > > --- a/drivers/clk/clk.c
> > > +++ b/drivers/clk/clk.c
> > > @@ -1011,8 +1011,9 @@ unsigned long __clk_round_rate(struct clk *clk, unsigned long rate)
> > > * @rate: the rate which is to be rounded
> > > *
> > > * Takes in a rate as input and rounds it to a rate that the clk can actually
> > > - * use which is then returned. If clk doesn't support round_rate operation
> > > - * then the parent rate is returned.
> > > + * use and does not exceed the requested frequency, which is then returned.
> > > + * If clk doesn't support round_rate operation then the parent rate
> > > + * is returned.
> > > */
> > > long clk_round_rate(struct clk *clk, unsigned long rate)
> > > {
> > > @@ -1027,6 +1028,44 @@ long clk_round_rate(struct clk *clk, unsigned long rate)
> > > EXPORT_SYMBOL_GPL(clk_round_rate);
> > >
> > > /**
> > > + * clk_round_rate_nearest - round the given rate for a clk
> > > + * @clk: the clk for which we are rounding a rate
> > > + * @rate: the rate which is to be rounded
> > > + *
> > > + * Takes in a rate as input and rounds it to the closest rate that the clk
> > > + * can actually use which is then returned. If clk doesn't support
> > > + * round_rate operation then the parent rate is returned.
> > > + */
> > > +long clk_round_rate_nearest(struct clk *clk, unsigned long rate)
> > Why does this function doesn't return an unsigned long when it never
> > returns a negative value? Ditto for clk_round_rate?
>
> I matched the definition of clk_round_rate(). But you're probably right,
> it may be the right thing to change clk_round_rate to return an
> unsigned, but with that being exposed API it would be a risky change.
>
> >
> > > +{
> > > + unsigned long lower, upper, cur, lower_last, upper_last;
> > > +
> > > + lower = clk_round_rate(clk, rate);
> > > + if (lower >= rate)
> > > + return lower;
> > Is the >-case worth a warning?
>
> No, it's correct behavior. If you request a rate that is way lower than what the
> clock can generate, returning something larger is perfectly valid, IMHO.
> Which reveals one problem in this whole discussion. The API does not
> require clk_round_rate() to round down. It is actually an implementation
> choice that had been made for clk-divider.
>
> >
> > > +
> > > + upper = clk_round_rate(clk, rate + rate - lower);
> > This was parenthesized in my original patch on purpose. If rate is big
> >
> > rate + rate - lower
> >
> > might overflow when
> >
> > rate + (rate - lower)
> >
> > doesn't. Thinking again, there is no real problem, because this is
> > unsigned arithmetic. To be save we still need to check if rate + (rate -
> > lower) overflows.
>
> Good point. I'll add the parentheses.
>
> >
> > > + if (upper == lower)
> > if (upper <= rate) is the better check here. (= would be a bug.)
>
> I don't understand. Passing rate + x to round rate can never return
> something below 'lower'. Only something in the range [lower,lower+x].
> So, if upper == lower we found our closest frequency and return it.
> Otherwise we have to iterate over [lower+1,upper]. Or what did I miss?
>
> >
> > > + return upper;
> > > +
> > > + lower = rate + 1;
> > ok, so your loop invariant is that the best freq is in [lower; upper].
>
> right.
>
> >
> > > + do {
> > > + upper_last = upper;
> > > + lower_last = lower;
> > > +
> > > + cur = clk_round_rate(clk, lower + ((upper - lower) >> 1));
> > > + if (cur < lower)
> > > + lower += (upper - lower) >> 1;
> > You already know that lower + ((upper - lower) >> 1) is too small, so
> > you can better do
> >
> > lower += ((upper - lower) >> 1) + 1;
>
> right. I'll add the '+1'
>
> >
> > > + else
> > > + upper = cur;
> > > +
> > > + } while (lower_last != lower && upper_last != upper);
> > > +
> > > + return upper;
> > > +}
> > > +EXPORT_SYMBOL_GPL(clk_round_rate_nearest);
> > I think the function still has potential for optimisation, what about:
>
> At first glance, I don't see many differences except for the comments
> you made above. I'll have a closer look though.
>
> >
> > unsigned long clk_round_rate_nearest(struct clk *clk, unsigned long rate)
> > {
> > unsigned long lower, upper, rounded;
> >
> > rounded = clk_round_rate(clk, rate);
> >
> > if (rounded >= rate)
> > return rounded;
> >
> > /*
> > * rounded is the best approximation for rate that is not
> > * bigger than rate. If there is a better one, it must be in the
> > * interval (rate; rate + (rate - rounded)).
> > * Note that the upper limit isn't better than rate itself, so
> > * that one doesn't need to be considered.
> > */
> >
> > upper = rate + (rate - rounded) - 1;
> > if (upper < rate)
> > upper = ULONG_MAX;
>
> Aren't we done here? Your search for an upper boundary resulted in
> 'lower'. Hence there is no valid frequency closer to 'rate' than 'lower'. Why do
> you extend to ULONG_MAX?
With the improvements suggested by you I have this now:
long clk_round_rate_nearest(struct clk *clk, unsigned long rate)
{
unsigned long lower, upper;
lower = clk_round_rate(clk, rate);
if (lower >= rate)
return lower;
upper = clk_round_rate(clk, rate + (rate - lower) - 1);
if (upper == lower)
return upper;
lower = rate + 1;
while (lower < upper) {
unsigned long rounded, mid;
mid = lower + ((upper - lower) >> 1);
rounded = clk_round_rate(clk, mid);
if (rounded < lower)
lower = mid + 1;
else
upper = rounded;
}
return upper;
}
Sören
next prev parent reply other threads:[~2014-05-19 17:31 UTC|newest]
Thread overview: 41+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-05-14 22:30 [RFC PATCH 0/5] Frequency resolution in CCF vs. cpufreq Soren Brinkmann
2014-05-14 22:30 ` [RFC PATCH 1/5] cpufreq: stats: Allow small rounding errors Soren Brinkmann
2014-05-14 22:30 ` [RFC PATCH 2/5] clk: Introduce 'clk_round_rate_nearest()' Soren Brinkmann
2014-05-15 7:38 ` Uwe Kleine-König
2014-05-15 14:10 ` Sören Brinkmann
2014-05-19 0:51 ` Sören Brinkmann
2014-05-19 16:19 ` Uwe Kleine-König
2014-05-19 16:41 ` Sören Brinkmann
2014-05-19 17:29 ` Sören Brinkmann [this message]
2014-05-20 7:51 ` Uwe Kleine-König
2014-05-20 7:33 ` Uwe Kleine-König
2014-05-20 16:01 ` Sören Brinkmann
2014-05-20 17:48 ` Stephen Boyd
2014-05-20 21:48 ` Sören Brinkmann
2014-05-21 7:34 ` Uwe Kleine-König
2014-05-21 15:58 ` Sören Brinkmann
2014-05-21 18:23 ` Uwe Kleine-König
2014-05-21 20:19 ` Sören Brinkmann
2014-05-21 20:33 ` Mike Turquette
2014-05-22 18:03 ` Sören Brinkmann
2014-05-22 18:20 ` Uwe Kleine-König
2014-05-22 20:32 ` Sören Brinkmann
2014-05-22 21:03 ` Mike Turquette
2014-05-22 23:44 ` Sören Brinkmann
[not found] ` <20140523013732.9521.70820@quantum>
2014-05-23 16:14 ` Sören Brinkmann
2014-05-26 6:29 ` Viresh Kumar
2014-05-26 11:22 ` Rafael J. Wysocki
2014-05-26 11:07 ` Viresh Kumar
2014-05-26 11:47 ` Rafael J. Wysocki
2014-05-26 21:52 ` Sören Brinkmann
2014-05-28 2:05 ` Mike Turquette
2014-05-28 16:17 ` Rafael J. Wysocki
2014-06-07 0:44 ` Sören Brinkmann
2014-05-14 22:30 ` [RFC PATCH 3/5] cpufreq: cpu0: Use clk_round_rate_nearest() Soren Brinkmann
2014-05-14 22:30 ` [RFC PATCH 4/5] ARM: zynq: dt: Use properly rounded frequencies in OPPs Soren Brinkmann
2014-05-14 22:30 ` [RFC PATCH 5/5] net: macb: Use clk_round_rate_nearest() API Soren Brinkmann
2014-05-15 6:12 ` [RFC PATCH 0/5] Frequency resolution in CCF vs. cpufreq Viresh Kumar
2014-05-15 14:05 ` Sören Brinkmann
2014-05-15 7:47 ` Uwe Kleine-König
2014-05-15 12:14 ` Nishanth Menon
2014-05-15 14:00 ` Sören Brinkmann
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=df742b46-c630-49ea-b234-4d746bb66ec1@BL2FFO11FD040.protection.gbl \
--to=soren.brinkmann@xilinx.com \
--cc=cpufreq@vger.kernel.org \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-pm@vger.kernel.org \
--cc=linux@arm.linux.org.uk \
--cc=michal.simek@xilinx.com \
--cc=mturquette@linaro.org \
--cc=rjw@rjwysocki.net \
--cc=u.kleine-koenig@pengutronix.de \
--cc=viresh.kumar@linaro.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;
as well as URLs for NNTP newsgroup(s).