Date: Wed, 2 Nov 2016 01:19:25 +0100
From: Daniel Lezcano <daniel.lezcano@linaro.org>
To: Vineet Gupta <Vineet.Gupta1@synopsys.com>
Cc: tglx@linutronix.de, linux-kernel@vger.kernel.org,
        linux-snps-arc@lists.infradead.org, Noam Camus <noamca@mellanox.com>,
        Alexey.Brodkin@synopsys.com
Subject: Re: [PATCH 9/9] clocksource: import ARC timer driver
Message-ID: <20161102001925.GH1506@mai>
References: <35bde193-8492-83e0-fb03-8385d8afd007@synopsys.com>
 <1477954096-770-1-git-send-email-vgupta@synopsys.com>
 <1477954096-770-10-git-send-email-vgupta@synopsys.com>
 <20161101204257.GG1506@mai>
 <188a2912-3369-19a3-86af-cbb154ff7e44@synopsys.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <188a2912-3369-19a3-86af-cbb154ff7e44@synopsys.com>
User-Agent: Mutt/1.5.24 (2015-08-30)
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 2117
Lines: 52

On Tue, Nov 01, 2016 at 01:57:05PM -0700, Vineet Gupta wrote:
> Hi Daniel,
> 
> On 11/01/2016 01:42 PM, Daniel Lezcano wrote:
> > Please stay consistent with the rest of the Kconfig.
> > 
> > config ARC_TIMER_RTC
> > 	bool "64-bit cycle counter in HS38 cores" if COMPILE_TEST
> > 	select CLKSRC_OF
> > 	help
> > 	  This counter provides 64-bit resolution vs. the 32-bit TIMER1.
> > 	  It is implemented inside the core thus can't be used in SMP systems.
> > 
> > config ARC_TIMER_GFRC
> > 	bool "64-bit cycle counter in ARConnect block in HS38x cores" if COMPILE_TEST
> > 	select CLKSRC_OF
> > 	help
> > 	  This counter can be used as clocksource in SMP HS38 SoCs.
> > 	  It sits outside the core thus can be used in SMP systems
> > 
> 
> Yes I did so already :-) Although I also added a default y if ARC to both, but as
> you say that is better done in ARC Kconfig.
> 
> > Then in the ARC's Kconfig you select ARC_TIMER_RTC or ARC_TIMER_GFRC depending
> > it is SMP or not.
> > 
> > One question:
> > 
> > Why ARC_TIMER_RTC can't be used in a SMP system ? Doesn't have each core its
> > own clocksource ? It seems you are assuming a clocksource can be used on SMP
> > only if the clocksource is unique and shared across the cores.
> 
> Thats what I thought so far. Thing is, the individual core's counters could get
> out of sync, simply because non masters cores were halted to begin with and came
> up at different points in real time. so a gtod might return different value
> depending on what core it landed on. Does clocksource also does ticks broadcasts
> and such to keep things in sync ?

Sounds like it is similar than the TSC. Do you agree to have a try by setting
the CONFIG_HAVE_UNSTABLE_SCHED_CLOCK option ?

If you can use those per cpu clocksource, performances on your system may
improve with the sched_clock().
 
> Because of the git mv you, diff didn't include bulk of driver code which would
> make for bulk of review anyways. So perhaps in v2 I don't do the git mv. OK ?

That means I will review and comment existing code. It is not a problem for me
if you agree to do the changes.

  -- Daniel