From mboxrd@z Thu Jan 1 00:00:00 1970 From: Boris BREZILLON Subject: Re: [PATCH 4/5] rtc: at91sam9: retain slow clock and check its rate Date: Mon, 8 Sep 2014 21:22:18 +0200 Message-ID: <20140908212218.6a81af86@bbrezillon> References: <1409733934-14465-1-git-send-email-boris.brezillon@free-electrons.com> <1409733934-14465-5-git-send-email-boris.brezillon@free-electrons.com> <20140908173338.GD14623@piout.net> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20140908173338.GD14623@piout.net> Sender: linux-kernel-owner@vger.kernel.org To: Alexandre Belloni Cc: Nicolas Ferre , Jean-Christophe Plagniol-Villard , Andrew Victor , Alessandro Zummo , rtc-linux@googlegroups.com, Rob Herring , Pawel Moll , Mark Rutland , Ian Campbell , Kumar Gala , devicetree@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org List-Id: devicetree@vger.kernel.org On Mon, 8 Sep 2014 19:33:38 +0200 Alexandre Belloni wrote: > On 03/09/2014 at 10:45:33 +0200, Boris Brezillon wrote : > > The RTT block is using the slow clock and expect it to run at 32KHz. > > Now that we moved to the CCF it's better to retain the clk reference so > > that the CCF can't disable the slow clock considering it is unused. > > > > Signed-off-by: Boris BREZILLON > > --- > > drivers/rtc/rtc-at91sam9.c | 24 ++++++++++++++++++++++++ > > 1 file changed, 24 insertions(+) > > > > diff --git a/drivers/rtc/rtc-at91sam9.c b/drivers/rtc/rtc-at91sam9.c > > index 57014b7..5c5093b 100644 > > --- a/drivers/rtc/rtc-at91sam9.c > > +++ b/drivers/rtc/rtc-at91sam9.c > > @@ -21,6 +21,7 @@ > > #include > > #include > > #include > > +#include > > > > /* > > * This driver uses two configurable hardware resources that live in the > > @@ -74,6 +75,7 @@ struct sam9_rtc { > > u32 imr; > > void __iomem *gpbr; > > int irq; > > + struct clk *sclk; > > }; > > > > #define rtt_readl(rtc, field) \ > > @@ -373,6 +375,25 @@ static int at91_rtc_probe(struct platform_device *pdev) > > return ret; > > } > > > > + /* Retain slow clk if it is specified in the DT. > > + * Do not complain if slow clk is missing, but check its rate > > + * if it is available. > > + */ > > + rtc->sclk = devm_clk_get(&pdev->dev, NULL); > > + if (!IS_ERR(rtc->sclk)) { > > + if (clk_get_rate(rtc->sclk) != AT91_SLOW_CLOCK) { > > I would not bother doing that check but use the value for MR instead of > AT91_SLOW_CLOCK (see my previous mail). Unfortunately, we can't get rid of this macro without modifying the clk_lookup table in several arch/arm/mach-at91/.c files in order to handle non DT/CCF cases (which will remain until all non DT boards are moved to DT). But I agree that this should be removed as soon as possible (AFAIR, all SoCs have already been moved to the CCF). How about adding a TODO comment ? > > > + dev_err(&pdev->dev, > > + "Invalid slow clock rate (expecting %lu got %lu)", > > + (unsigned long)AT91_SLOW_CLOCK, > > + clk_get_rate(rtc->sclk)); > > + return -EINVAL; > > + } > > + > > + ret = clk_prepare_enable(rtc->sclk); > > + if (ret) > > + return ret; > > + } > > + > > /* NOTE: sam9260 rev A silicon has a ROM bug which resets the > > * RTT on at least some reboots. If you have that chip, you must > > * initialize the time from some external source like a GPS, wall > > @@ -397,6 +418,9 @@ static int at91_rtc_remove(struct platform_device *pdev) > > /* disable all interrupts */ > > rtt_writel(rtc, MR, mr & ~(AT91_RTT_ALMIEN | AT91_RTT_RTTINCIEN)); > > > > + if (!IS_ERR(rtc->sclk)) > > + clk_disable_unprepare(rtc->sclk); > > + > > return 0; > > } > > > > -- > > 1.9.1 > > > -- Boris Brezillon, Free Electrons Embedded Linux and Kernel engineering http://free-electrons.com