From mboxrd@z Thu Jan 1 00:00:00 1970 From: Uwe =?iso-8859-1?Q?Kleine-K=F6nig?= Subject: Re: [RFC PATCH 2/5] clk: Introduce 'clk_round_rate_nearest()' Date: Thu, 22 May 2014 20:20:40 +0200 Message-ID: <20140522182040.GB20155@pengutronix.de> References: <20140520073358.GJ16662@pengutronix.de> <4bb5f44a-60bb-4e34-8f88-f91b8419be8d@BL2FFO11FD050.protection.gbl> <537B957B.5010001@codeaurora.org> <668683e3-856e-4f30-9b11-8f3e91e12d1d@BL2FFO11FD038.protection.gbl> <20140521073457.GD31687@pengutronix.de> <20140521182308.GN31687@pengutronix.de> <20140521203300.9521.67546@quantum> Mime-Version: 1.0 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: Content-Disposition: inline In-Reply-To: Sender: cpufreq-owner@vger.kernel.org List-ID: Content-Type: text/plain; charset="iso-8859-1" To: =?iso-8859-1?Q?S=F6ren?= Brinkmann Cc: Mike Turquette , Russell King , Stephen Boyd , linux-pm@vger.kernel.org, Viresh Kumar , "Rafael J. Wysocki" , Michal Simek , cpufreq@vger.kernel.org, linux-kernel@vger.kernel.org, linux-arm-kernel@lists.infradead.org Hello S=F6ren, On Thu, May 22, 2014 at 11:03:00AM -0700, S=F6ren Brinkmann wrote: > On Wed, 2014-05-21 at 01:33PM -0700, Mike Turquette wrote: > > Quoting Uwe Kleine-K=F6nig (2014-05-21 11:23:08) > > > Hello S=F6ren, > > >=20 > > > On Wed, May 21, 2014 at 08:58:10AM -0700, S=F6ren Brinkmann wrote= : > > > > On Wed, 2014-05-21 at 09:34AM +0200, Uwe Kleine-K=F6nig wrote: > > > > > On Tue, May 20, 2014 at 02:48:20PM -0700, S=F6ren Brinkmann w= rote: > > > > > > On Tue, 2014-05-20 at 10:48AM -0700, Stephen Boyd wrote: > > > > > > > On 05/20/14 09:01, S=F6ren Brinkmann wrote: > > > > > > > > > > > > > > > >>>>> +{ > > > > > > > >>>>> + unsigned long lower, upper, cur, lower_last, uppe= r_last; > > > > > > > >>>>> + > > > > > > > >>>>> + lower =3D clk_round_rate(clk, rate); > > > > > > > >>>>> + if (lower >=3D 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 per= fectly valid, IMHO. > > > > > > > >>> Which reveals one problem in this whole discussion. T= he API does not > > > > > > > >>> require clk_round_rate() to round down. It is actuall= y an implementation > > > > > > > >>> choice that had been made for clk-divider. > > > > > > > >> I'm sure it's more than an implementation choice for c= lk-divider. But I > > > > > > > >> don't find any respective documentation (but I didn't = try hard). > > > > > > > > A similar discussion - without final conclusion: > > > > > > > > https://lkml.org/lkml/2010/7/14/260 > > > > > > > > > > > > > > > > > > > > > > >=20 > > > > > > > Please call this new API something like clk_find_nearest_= rate() or > > > > > > > something. clk_round_rate() is supposed to return the rat= e that will be > > > > > > > set if you call clk_set_rate() with the same arguments. I= t's up to the > > > > > > > implementation to decide if that means rounding the rate = up or down or > > > > > > > to the nearest value. > > > > > >=20 > > > > > > Sounds good to me. Are there any cases of clocks that round= up? I think > > > > > > that case would not be handled correctly. But I also don't = see a use > > > > > > case for such an implementation. > > > > > I don't really care which semantic (i.e. round up, round down= or round > > > > > closest) is picked, but I'd vote that all should pick up the = same. I > > > > > think the least surprising definition is to choose rounding d= own and add > > > > > the function that is under discussion here to get a nearest m= atch. > > > > >=20 > > > > > So I suggest: > > > > >=20 > > > > > - if round_rate is given a rate that is smaller than the > > > > > smallest available rate, return 0 > > > > > - add WARN_ONCE to round_rate and set_rate if they return= with a > > > > > rate bigger than requested > > > >=20 > > > > Why do you think 0 is always valid? I think for a clock that ca= n > > > > generate 40, 70, 120, clk_round_rate(20) should return 40. > > > I didn't say it's a valid value. It just makes the it possible to= check > > > for clk_round_rate(clk, rate) <=3D rate. > > >=20 > > > I grepped a bit around and found da850_round_armrate which implem= ents a > > > round_rate callback returning the best match. > > > omap1_clk_round_rate_ckctl_arm can return a value < 0. > > > s3c2412_roundrate_usbsrc can return values that are bigger than > > > requested. (I wonder if that is a bug though.) > > >=20 > > > > > - change the return values to unsigned long > > > >=20 > > > > Yep, I agree, this should happen. > > > And we're using 0 as error value? e.g. for the case where > > > omap1_clk_round_rate_ckctl_arm returns -EIO now? > >=20 > > No. clk_round_rate returns long for a reason, which is that we can > > provide an error code to the caller. From include/linux/clk.h: > >=20 > > /** > > * clk_round_rate - adjust a rate to the exact rate a clock can pro= vide > > * @clk: clock source > > * @rate: desired clock rate in Hz > > * > > * Returns rounded clock rate in Hz, or negative errno. > > */ > >=20 > > This has the unfortunate side effect that the max value we can retu= rn > > safely is 2147483647 (~2GHz). So another issue here is converting c= lock > > rates to 64-bit values. >=20 > So, let's assume > - a clock does either of these > - round down > - round nearest > - round up (is there any such case? I don't see a use-case for thi= s) > - or return an error >=20 > I think my latest try handles such cases, with the limitation of > for a clock that rounds up, the up-rounded value is found instead of = the > nearest. >=20 >=20 > static long clk_find_nearest_rate(struct clk *clk, unsigned long rate= ) > { > long ret; > unsigned long lower, upper; >=20 > clk_prepare_lock(); >=20 > lower =3D __clk_round_rate(clk, rate); this is CCF specific while I don't see a need for it. (But yes, a lock-less clk_find_nearest_rate is of course racy.) > if (lower >=3D rate || (long)lower < 0) { If you made lower and upper a signed long, you could drop the casting here. BTW, why does __clk_round_rate return an unsigned long?? There seem to be several more type mismatches in that area. Maybe we should add a waring if rate is > LONG_MAX? (And ISTR that the C standard doesn't specify what the result of (long)lower is given that lower is of type unsigned long and holding a value > LONG_MAX.) > ret =3D lower; > goto unlock; > } >=20 > upper =3D rate + (rate - lower) - 1; > if (upper > LONG_MAX) > upper =3D LONG_MAX; >=20 > upper =3D __clk_round_rate(clk, upper); > if (upper <=3D lower || (long)upper < 0) { > ret =3D lower; > goto unlock; > } >=20 > lower =3D rate + 1; > while (lower < upper) { > unsigned long rounded, mid; >=20 > mid =3D lower + ((upper - lower) >> 1); > rounded =3D __clk_round_rate(clk, mid); > if (rounded < lower) > lower =3D mid + 1; > else > upper =3D rounded; > } >=20 > ret =3D upper; >=20 > unlock: > clk_prepare_unlock(); >=20 > return ret; > } --=20 Pengutronix e.K. | Uwe Kleine-K=F6nig = | Industrial Linux Solutions | http://www.pengutronix.de/= |