From mboxrd@z Thu Jan  1 00:00:00 1970
Return-path: <linux-watchdog-owner@vger.kernel.org>
Received: from mx0b-0016f401.pphosted.com ([67.231.156.173]:18170 "EHLO
	mx0b-0016f401.pphosted.com" rhost-flags-OK-OK-OK-OK)
	by vger.kernel.org with ESMTP id S1757162AbcG0Dje (ORCPT
	<rfc822;linux-watchdog@vger.kernel.org>);
	Tue, 26 Jul 2016 23:39:34 -0400
Date: Wed, 27 Jul 2016 11:33:42 +0800
From: Jisheng Zhang <jszhang@marvell.com>
To: Timur Tabi <timur@codeaurora.org>
CC: Will Deacon <will.deacon@arm.com>, Fu Wei <fu.wei@linaro.org>,
	"Linaro ACPI Mailman List" <linaro-acpi@lists.linaro.org>,
	Catalin Marinas <catalin.marinas@arm.com>,
	<rruigrok@codeaurora.org>, Wim Van Sebroeck <wim@iguana.be>,
	<wei@redhat.com>, Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>,
	Al Stone <al.stone@linaro.org>,
	Daniel Lezcano <daniel.lezcano@linaro.org>,
	ACPI Devel Maling List <linux-acpi@vger.kernel.org>,
	Guenter Roeck <linux@roeck-us.net>,
	Len Brown <lenb@kernel.org>, <harba@codeaurora.org>,
	<linux-watchdog@vger.kernel.org>, Arnd Bergmann <arnd@arndb.de>,
	"Marc Zyngier" <marc.zyngier@arm.com>,
	Jon Masters <jcm@redhat.com>,
	"Christopher Covington" <cov@codeaurora.org>,
	Thomas Gleixner <tglx@linutronix.de>,
	<linux-arm-kernel@lists.infradead.org>,
	G Gregory <graeme.gregory@linaro.org>,
	"Rafael J. Wysocki" <rjw@rjwysocki.net>,
	Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	Leo Duran <leo.duran@amd.com>,
	Hanjun Guo <hanjun.guo@linaro.org>,
	"Suravee Suthikulpanit" <Suravee.Suthikulpanit@amd.com>,
	Sudeep Holla <sudeep.holla@arm.com>
Subject: Re: [PATCH v9 4/9] clocksource/drivers/arm_arch_timer: use readq to
 get 64-bit CNTVCT
Message-ID: <20160727113342.2a839c1a@xhacker>
In-Reply-To: <57976FA5.2070802@codeaurora.org>
References: <1469460427-8643-1-git-send-email-fu.wei@linaro.org>
	<1469460427-8643-5-git-send-email-fu.wei@linaro.org>
	<20160725153118.GD19209@arm.com>
	<CADyBb7ubKceQiB7+u7sSA=1+9_VMUrGjiabf_FkN2-n4rq7Kgg@mail.gmail.com>
	<20160725163144.GE19209@arm.com>	<57976FA5.2070802@codeaurora.org>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
Sender: linux-watchdog-owner@vger.kernel.org
List-Id: linux-watchdog@vger.kernel.org

+1

On Tue, 26 Jul 2016 09:11:49 -0500 Timur Tabi  wrote:

> Will Deacon wrote:
> > The kernel really needs to support both of those platforms :/
> >
> > For the memory-mapped counter registers, the architecture says:
> >
> >    `If the implementation supports 64-bit atomic accesses, then the
> >     CNTV_CVAL register must be accessible as an atomic 64-bit value.'
> >
> > which is borderline tautological. If we take the generous reading that
> > this means AArch64 CPUs can use readq (and I'm not completely
> > comfortable with that assertion, particularly as you say that it breaks
> > the model), then you still need to use readq_relaxed here to avoid a
> > DSB. Furthermore, what are you going to do for AArch32? readq doesn't
> > exist over there, and if you use the generic implementation then it's
> > not atomic. In which case, we end up with the current code, as well as a
> > readq_relaxed guarded by a questionable #ifdef that is known to break a
> > supported platform for an unknown performance improvement. Hardly a big
> > win.  
> 
> I know Fu dropped this patch, and I don't want to kick a dead horse, but 
> I was wondering if it would be okay to do this:
> 
> static u64 arch_counter_get_cntvct_mem(void)
> {
> #ifdef readq_relaxed
> 	return readq_relaxed(arch_counter_base + CNTVCT_LO);
> #else
> 	u32 vct_lo, vct_hi, tmp_hi;
> 
> 	do {
> 		vct_hi = readl_relaxed(arch_counter_base + CNTVCT_HI);
> 		vct_lo = readl_relaxed(arch_counter_base + CNTVCT_LO);
> 		tmp_hi = readl_relaxed(arch_counter_base + CNTVCT_HI);
> 	} while (vct_hi != tmp_hi);
> 
> 	return ((u64) vct_hi << 32) | vct_lo;
> #endif
> }
> 
> readq and readq_relaxed are defined in arch/arm64/include/asm/io.h.  Why 
> would the function exist if AArch64 CPUs can't use it?

+1

I measured the performance on berlin arm64 platforms:

compared with original version, using readq_relaxed could reduce
time of arch_counter_get_cntvct_mem() by about 42%!

Thanks,
Jisheng