Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754183AbYAQCgS (ORCPT ); Wed, 16 Jan 2008 21:36:18 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1750941AbYAQCgD (ORCPT ); Wed, 16 Jan 2008 21:36:03 -0500 Received: from tomts36-srv.bellnexxia.net ([209.226.175.93]:51675 "EHLO tomts36-srv.bellnexxia.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750926AbYAQCgA (ORCPT ); Wed, 16 Jan 2008 21:36:00 -0500 X-IronPort-Anti-Spam-Filtered: true X-IronPort-Anti-Spam-Result: Aq4HAOBNjkdMROHU/2dsb2JhbACBWKxe Date: Wed, 16 Jan 2008 21:35:57 -0500 From: Mathieu Desnoyers To: john stultz Cc: Steven Rostedt , LKML , Ingo Molnar , Linus Torvalds , Andrew Morton , Peter Zijlstra , Christoph Hellwig , Gregory Haskins , Arnaldo Carvalho de Melo , Thomas Gleixner , Tim Bird , Sam Ravnborg , "Frank Ch. Eigler" , Steven Rostedt , Paul Mackerras , Daniel Walker Subject: Re: [RFC PATCH 16/22 -v2] add get_monotonic_cycles Message-ID: <20080117023557.GC2322@Krystal> References: <20080115220824.GB22242@Krystal> <20080116031730.GA2164@Krystal> <20080116145604.GB31329@Krystal> <1f1b08da0801161436k4a7ac1e3kd83590951e7bebb9@mail.gmail.com> <1200523867.6127.5.camel@localhost.localdomain> <20080116233927.GB23895@Krystal> <1200536409.6127.47.camel@localhost.localdomain> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Content-Disposition: inline In-Reply-To: <1200536409.6127.47.camel@localhost.localdomain> X-Editor: vi X-Info: http://krystal.dyndns.org:8080 X-Operating-System: Linux/2.6.21.3-grsec (i686) X-Uptime: 21:23:29 up 74 days, 7:28, 5 users, load average: 0.20, 0.42, 0.52 User-Agent: Mutt/1.5.16 (2007-06-11) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 6758 Lines: 196 * john stultz (johnstul@us.ibm.com) wrote: > On Wed, 2008-01-16 at 18:39 -0500, Mathieu Desnoyers wrote: > > I would disable preemption in clocksource_get_basecycles. We would not > > want to be scheduled out while we hold a pointer to the old array > > element. > > > > > + int num = cs->base_num; > > > > Since you deal with base_num in a shared manner (not per cpu), you will > > need a smp_read_barrier_depend() here after the cs->base_num read. > > > > You should think about reading the cs->base_num first, and _after_ that > > read the real clocksource. Here, the clocksource value is passed as > > parameter. It means that the read clocksource may have been read in the > > previous RCU window. > > Here's an updated version of the patch w/ the suggested memory barrier > changes and favored (1-x) inversion change. ;) Let me know if you see > any other holes, or have any other suggestions or ideas. > > Still un-tested (my test box will free up soon, I promise!), but builds. > > Signed-off-by: John Stultz > > Index: monotonic-cleanup/include/linux/clocksource.h > =================================================================== > --- monotonic-cleanup.orig/include/linux/clocksource.h 2008-01-16 12:22:04.000000000 -0800 > +++ monotonic-cleanup/include/linux/clocksource.h 2008-01-16 18:12:53.000000000 -0800 > @@ -87,9 +87,17 @@ > * more than one cache line. > */ > struct { > - cycle_t cycle_last, cycle_accumulated, cycle_raw; > - } ____cacheline_aligned_in_smp; > + cycle_t cycle_last, cycle_accumulated; > > + /* base structure provides lock-free read > + * access to a virtualized 64bit counter > + * Uses RCU-like update. > + */ > + struct { > + cycle_t cycle_base_last, cycle_base; > + } base[2]; > + int base_num; > + } ____cacheline_aligned_in_smp; > u64 xtime_nsec; > s64 error; > > @@ -175,19 +183,29 @@ > } > > /** > - * clocksource_get_cycles: - Access the clocksource's accumulated cycle value > + * clocksource_get_basecycles: - get the clocksource's accumulated cycle value > * @cs: pointer to clocksource being read > * @now: current cycle value > * > * Uses the clocksource to return the current cycle_t value. > * NOTE!!!: This is different from clocksource_read, because it > - * returns the accumulated cycle value! Must hold xtime lock! > + * returns a 64bit wide accumulated value. > */ > static inline cycle_t > -clocksource_get_cycles(struct clocksource *cs, cycle_t now) > +clocksource_get_basecycles(struct clocksource *cs) > { > - cycle_t offset = (now - cs->cycle_last) & cs->mask; > - offset += cs->cycle_accumulated; > + int num; > + cycle_t now, offset; > + > + preempt_disable(); > + num = cs->base_num; > + smp_read_barrier_depends(); > + now = clocksource_read(cs); > + offset = (now - cs->base[num].cycle_base_last); > + offset &= cs->mask; > + offset += cs->base[num].cycle_base; > + preempt_enable(); > + > return offset; > } > > @@ -197,14 +215,26 @@ > * @now: current cycle value > * > * Used to avoids clocksource hardware overflow by periodically > - * accumulating the current cycle delta. Must hold xtime write lock! > + * accumulating the current cycle delta. Uses RCU-like update, but > + * ***still requires the xtime_lock is held for writing!*** > */ > static inline void clocksource_accumulate(struct clocksource *cs, cycle_t now) > { > - cycle_t offset = (now - cs->cycle_last) & cs->mask; > + /* First update the monotonic base portion. > + * The dual array update method allows for lock-free reading. > + */ > + int num = 1 - cs->base_num; (nitpick) right here, you could probably express 1-num with cs->base_num, since we are the only ones supposed to touch it. > + cycle_t offset = (now - cs->base[1-num].cycle_base_last); > + offset &= cs->mask; here too. > + cs->base[num].cycle_base = cs->base[1-num].cycle_base + offset; > + cs->base[num].cycle_base_last = now; > + wmb(); As I just emailed : smp_smb() *should* be enough. I don't see which architecture could reorder writes wrt local interrupts ? (please tell me if I am grossly mistaken) Mathieu > + cs->base_num = num; > + > + /* Now update the cycle_accumulated portion */ > + offset = (now - cs->cycle_last) & cs->mask; > cs->cycle_last = now; > cs->cycle_accumulated += offset; > - cs->cycle_raw += offset; > } > > /** > Index: monotonic-cleanup/kernel/time/timekeeping.c > =================================================================== > --- monotonic-cleanup.orig/kernel/time/timekeeping.c 2008-01-16 12:21:46.000000000 -0800 > +++ monotonic-cleanup/kernel/time/timekeeping.c 2008-01-16 17:51:50.000000000 -0800 > @@ -71,10 +71,12 @@ > */ > static inline s64 __get_nsec_offset(void) > { > - cycle_t cycle_delta; > + cycle_t now, cycle_delta; > s64 ns_offset; > > - cycle_delta = clocksource_get_cycles(clock, clocksource_read(clock)); > + now = clocksource_read(clock); > + cycle_delta = (now - clock->cycle_last) & clock->mask; > + cycle_delta += clock->cycle_accumulated; > ns_offset = cyc2ns(clock, cycle_delta); > > return ns_offset; > @@ -105,35 +107,7 @@ > > cycle_t notrace get_monotonic_cycles(void) > { > - cycle_t cycle_now, cycle_delta, cycle_raw, cycle_last; > - > - do { > - /* > - * cycle_raw and cycle_last can change on > - * another CPU and we need the delta calculation > - * of cycle_now and cycle_last happen atomic, as well > - * as the adding to cycle_raw. We don't need to grab > - * any locks, we just keep trying until get all the > - * calculations together in one state. > - * > - * In fact, we __cant__ grab any locks. This > - * function is called from the latency_tracer which can > - * be called anywhere. To grab any locks (including > - * seq_locks) we risk putting ourselves into a deadlock. > - */ > - cycle_raw = clock->cycle_raw; > - cycle_last = clock->cycle_last; > - > - /* read clocksource: */ > - cycle_now = clocksource_read(clock); > - > - /* calculate the delta since the last update_wall_time: */ > - cycle_delta = (cycle_now - cycle_last) & clock->mask; > - > - } while (cycle_raw != clock->cycle_raw || > - cycle_last != clock->cycle_last); > - > - return cycle_raw + cycle_delta; > + return clocksource_get_basecycles(clock); > } > > unsigned long notrace cycles_to_usecs(cycle_t cycles) > > -- Mathieu Desnoyers Computer Engineering Ph.D. Student, Ecole Polytechnique de Montreal OpenPGP key fingerprint: 8CD5 52C3 8E3C 4140 715F BA06 3F25 A8FE 3BAE 9A68 -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/