diff mbox

[PATCHv2,2/2] clocksource: dw_apb_timer_of: Fix read_sched_clock

Message ID 20130821190729.GA9657@amd.pavel.ucw.cz (mailing list archive)
State New, archived
Headers show

Commit Message

Pavel Machek Aug. 21, 2013, 7:07 p.m. UTC
On Wed 2013-08-21 09:04:07, Linus Walleij wrote:
> On Thu, Aug 15, 2013 at 1:29 AM,  <dinguyen@altera.com> wrote:
> 
> > -       sched_io_base = iobase + 0x04;
> > +       sched_io_base = iobase;
> >         sched_rate = rate;
> >  }
> >
> >  static u32 read_sched_clock(void)
> >  {
> > -       return __raw_readl(sched_io_base);
> > +       return ~__raw_readl(sched_io_base + 0x4);
> 
> So what about #define what 0x04 is?
> 
> #define MY_FOO_REGISTER_OFFSET 0x04
> 
> raw_readl(sched_io_base + MY_FOO_REGISTER_OFFSET);

That define is already there, #define APBTMR_N_CURRENT_VALUE 0x04, but
it is in .c file, not in header.

Actually I believe I had patch that did that, and it was even applied,
but it clashed with same other patches, and was dropped. (It is below,
for reference).

Currently, the code _does not work_, and we have three platforms using
it. It would be good to solve this now, and polish it later.

								Pavel

From pavel@denx.de Tue May  7 22:11:26 2013
Date: Tue, 7 May 2013 22:11:26 +0200
From: Pavel Machek <pavel@denx.de>
To: John Stultz <john.stultz@linaro.org>
Cc: Arnd Bergmann <arnd@arndb.de>, tglx@linutronix.de,
	Jamie Iles <jamie@jamieiles.com>, dinguyen@altera.com, wd@denx.de,
	linux-arm-kernel@lists.infradead.org, olof@lixom.net,
	kernel list <linux-kernel@vger.kernel.org>
Subject: Re: dw_apb_timer_of.c: remove parts that were picoxcell-specific
Message-ID: <20130507201126.GA8169@amd.pavel.ucw.cz>
References: <20130426121433.GA16249@amd.pavel.ucw.cz>
 <201305061545.22587.arnd@arndb.de>
 <20130506155304.GA6645@amd.pavel.ucw.cz>
 <201305062324.17080.arnd@arndb.de>
 <20130507135752.GA3500@amd.pavel.ucw.cz>
 <51892E5E.8090909@linaro.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <51892E5E.8090909@linaro.org>
User-Agent: Mutt/1.5.20 (2009-06-14)
Status: RO
Content-Length: 4987
Lines: 170

Hi!

> >>>So I guess it should go through the timekeeping tree...?
> >>I would prefer that, but we can also take it through arm-soc, it doesn't
> >>really matter.
> >I see no response from timekeeping people... so if you could take it,
> >that would be great.
> 
> I don't think the original patch ever made it to my inbox.

For some reason, I cced johnstul@us.ibm.com. (I guess
get_maintainer.pl? But it seems to behave now.) Anyway, patch is
below...

It would be great if you could apply it... Thanks,
							Pavel

----

It seems we made a mistake when creating dw_apb_timer_of.c:
picoxcell sched_clock had parts that were not related to
dw_apb_timer, yet we moved them to dw_apb_timer_of, and tried to
use them on socfpga.

This results in system where user/system time is not measured
properly, as demonstrated by
    
    time dd if=/dev/urandom of=/dev/zero bs=100000 count=100
    
So this patch switches sched_clock to hardware that exists on both
platforms, and adds missing of_node_put() in dw_apb_timer_init().
    
Signed-off-by: Pavel Machek <pavel@denx.de>
Acked-by: Jamie Iles <jamie@jamieiles.com>

Comments

Dinh Nguyen Aug. 21, 2013, 8:02 p.m. UTC | #1
On Wed, 2013-08-21 at 21:07 +0200, ZY - pavel wrote:
> On Wed 2013-08-21 09:04:07, Linus Walleij wrote:
> > On Thu, Aug 15, 2013 at 1:29 AM,  <dinguyen@altera.com> wrote:
> > 
> > > -       sched_io_base = iobase + 0x04;
> > > +       sched_io_base = iobase;
> > >         sched_rate = rate;
> > >  }
> > >
> > >  static u32 read_sched_clock(void)
> > >  {
> > > -       return __raw_readl(sched_io_base);
> > > +       return ~__raw_readl(sched_io_base + 0x4);
> > 
> > So what about #define what 0x04 is?
> > 
> > #define MY_FOO_REGISTER_OFFSET 0x04
> > 
> > raw_readl(sched_io_base + MY_FOO_REGISTER_OFFSET);
> 
> That define is already there, #define APBTMR_N_CURRENT_VALUE 0x04, but
> it is in .c file, not in header.

I'll send a V2 that will address this comment and Stephen Warren's.

Dinh
> 
> Actually I believe I had patch that did that, and it was even applied,
> but it clashed with same other patches, and was dropped. (It is below,
> for reference).
> 
> Currently, the code _does not work_, and we have three platforms using
> it. It would be good to solve this now, and polish it later.
> 
> 								Pavel
> 
> From pavel@denx.de Tue May  7 22:11:26 2013
> Date: Tue, 7 May 2013 22:11:26 +0200
> From: Pavel Machek <pavel@denx.de>
> To: John Stultz <john.stultz@linaro.org>
> Cc: Arnd Bergmann <arnd@arndb.de>, tglx@linutronix.de,
> 	Jamie Iles <jamie@jamieiles.com>, dinguyen@altera.com, wd@denx.de,
> 	linux-arm-kernel@lists.infradead.org, olof@lixom.net,
> 	kernel list <linux-kernel@vger.kernel.org>
> Subject: Re: dw_apb_timer_of.c: remove parts that were picoxcell-specific
> Message-ID: <20130507201126.GA8169@amd.pavel.ucw.cz>
> References: <20130426121433.GA16249@amd.pavel.ucw.cz>
>  <201305061545.22587.arnd@arndb.de>
>  <20130506155304.GA6645@amd.pavel.ucw.cz>
>  <201305062324.17080.arnd@arndb.de>
>  <20130507135752.GA3500@amd.pavel.ucw.cz>
>  <51892E5E.8090909@linaro.org>
> MIME-Version: 1.0
> Content-Type: text/plain; charset=us-ascii
> Content-Disposition: inline
> In-Reply-To: <51892E5E.8090909@linaro.org>
> User-Agent: Mutt/1.5.20 (2009-06-14)
> Status: RO
> Content-Length: 4987
> Lines: 170
> 
> Hi!
> 
> > >>>So I guess it should go through the timekeeping tree...?
> > >>I would prefer that, but we can also take it through arm-soc, it doesn't
> > >>really matter.
> > >I see no response from timekeeping people... so if you could take it,
> > >that would be great.
> > 
> > I don't think the original patch ever made it to my inbox.
> 
> For some reason, I cced johnstul@us.ibm.com. (I guess
> get_maintainer.pl? But it seems to behave now.) Anyway, patch is
> below...
> 
> It would be great if you could apply it... Thanks,
> 							Pavel
> 
> ----
> 
> It seems we made a mistake when creating dw_apb_timer_of.c:
> picoxcell sched_clock had parts that were not related to
> dw_apb_timer, yet we moved them to dw_apb_timer_of, and tried to
> use them on socfpga.
> 
> This results in system where user/system time is not measured
> properly, as demonstrated by
>     
>     time dd if=/dev/urandom of=/dev/zero bs=100000 count=100
>     
> So this patch switches sched_clock to hardware that exists on both
> platforms, and adds missing of_node_put() in dw_apb_timer_init().
>     
> Signed-off-by: Pavel Machek <pavel@denx.de>
> Acked-by: Jamie Iles <jamie@jamieiles.com>
> 
> diff --git a/arch/arm/mach-picoxcell/common.h b/arch/arm/mach-picoxcell/common.h
> index 481b42a..237fb3b 100644
> --- a/arch/arm/mach-picoxcell/common.h
> +++ b/arch/arm/mach-picoxcell/common.h
> @@ -12,6 +12,4 @@
>  
>  #include <asm/mach/time.h>
>  
> -extern void dw_apb_timer_init(void);
> -
>  #endif /* __PICOXCELL_COMMON_H__ */
> diff --git a/drivers/clocksource/dw_apb_timer.c b/drivers/clocksource/dw_apb_timer.c
> index 8c2a35f..460ac99 100644
> --- a/drivers/clocksource/dw_apb_timer.c
> +++ b/drivers/clocksource/dw_apb_timer.c
> @@ -21,12 +21,6 @@
>  #define APBT_MIN_PERIOD			4
>  #define APBT_MIN_DELTA_USEC		200
>  
> -#define APBTMR_N_LOAD_COUNT		0x00
> -#define APBTMR_N_CURRENT_VALUE		0x04
> -#define APBTMR_N_CONTROL		0x08
> -#define APBTMR_N_EOI			0x0c
> -#define APBTMR_N_INT_STATUS		0x10
> -
>  #define APBTMRS_INT_STATUS		0xa0
>  #define APBTMRS_EOI			0xa4
>  #define APBTMRS_RAW_INT_STATUS		0xa8
> diff --git a/drivers/clocksource/dw_apb_timer_of.c b/drivers/clocksource/dw_apb_timer_of.c
> index ab09ed3..0fa3104 100644
> --- a/drivers/clocksource/dw_apb_timer_of.c
> +++ b/drivers/clocksource/dw_apb_timer_of.c
> @@ -57,6 +57,15 @@ static void add_clockevent(struct device_node *event_timer)
>  	dw_apb_clockevent_register(ced);
>  }
>  
> +static void __iomem *sched_io_base;
> +
> +/* This is actually same as __apbt_read_clocksource(), but with
> +   different interface */
> +static u32 read_sched_clock_sptimer(void)
> +{
> +	return ~__raw_readl(sched_io_base + APBTMR_N_CURRENT_VALUE);
> +}
> +
>  static void add_clocksource(struct device_node *source_timer)
>  {
>  	void __iomem *iobase;
> @@ -71,41 +80,27 @@ static void add_clocksource(struct device_node *source_timer)
>  
>  	dw_apb_clocksource_start(cs);
>  	dw_apb_clocksource_register(cs);
> -}
>  
> -static void __iomem *sched_io_base;
> -
> -static u32 read_sched_clock(void)
> -{
> -	return __raw_readl(sched_io_base);
> +	sched_io_base = iobase;
> +	setup_sched_clock(read_sched_clock_sptimer, 32, rate);
>  }
>  
> -static const struct of_device_id sptimer_ids[] __initconst = {
> -	{ .compatible = "picochip,pc3x2-rtc" },
> +static const struct of_device_id osctimer_ids[] __initconst = {
> +	{ .compatible = "picochip,pc3x2-timer" },
> +	{ .compatible = "snps,dw-apb-timer-osc" },
>  	{ .compatible = "snps,dw-apb-timer-sp" },
> -	{ /* Sentinel */ },
> +	{  /* Sentinel */ },
>  };
>  
> -static void init_sched_clock(void)
> -{
> -	struct device_node *sched_timer;
> -	u32 rate;
> -
> -	sched_timer = of_find_matching_node(NULL, sptimer_ids);
> -	if (!sched_timer)
> -		panic("No RTC for sched clock to use");
> +/*
> +   You don't have to use dw_apb_timer for scheduler clock,
> +   this should also work fine on arm:
>  
> -	timer_get_base_and_rate(sched_timer, &sched_io_base, &rate);
> -	of_node_put(sched_timer);
> +  twd_local_timer_of_register();
> +  arch_timer_of_register();                 
> +  arch_timer_sched_clock_init();
> +*/
>  
> -	setup_sched_clock(read_sched_clock, 32, rate);
> -}
> -
> -static const struct of_device_id osctimer_ids[] __initconst = {
> -	{ .compatible = "picochip,pc3x2-timer" },
> -	{ .compatible = "snps,dw-apb-timer-osc" },
> -	{},
> -};
>  
>  void __init dw_apb_timer_init(void)
>  {
> @@ -121,7 +116,6 @@ void __init dw_apb_timer_init(void)
>  		panic("No timer for clocksource");
>  	add_clocksource(source_timer);
>  
> +	of_node_put(event_timer);
>  	of_node_put(source_timer);
> -
> -	init_sched_clock();
>  }
> diff --git a/include/linux/dw_apb_timer.h b/include/linux/dw_apb_timer.h
> index dd755ce..a64c987 100644
> --- a/include/linux/dw_apb_timer.h
> +++ b/include/linux/dw_apb_timer.h
> @@ -17,6 +17,12 @@
>  #include <linux/clocksource.h>
>  #include <linux/interrupt.h>
>  
> +#define APBTMR_N_LOAD_COUNT		0x00
> +#define APBTMR_N_CURRENT_VALUE		0x04
> +#define APBTMR_N_CONTROL		0x08
> +#define APBTMR_N_EOI			0x0c
> +#define APBTMR_N_INT_STATUS		0x10
> +
>  #define APBTMRS_REG_SIZE       0x14
>  
>  struct dw_apb_timer {
> 
> 
> -- 
> (english) http://www.livejournal.com/~pavelmachek
> (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
> 
>
John Stultz Aug. 21, 2013, 8:39 p.m. UTC | #2
On 08/21/2013 01:02 PM, Dinh Nguyen wrote:
> On Wed, 2013-08-21 at 21:07 +0200, ZY - pavel wrote:
>> On Wed 2013-08-21 09:04:07, Linus Walleij wrote:
>>> On Thu, Aug 15, 2013 at 1:29 AM,  <dinguyen@altera.com> wrote:
>>>
>>>> -       sched_io_base = iobase + 0x04;
>>>> +       sched_io_base = iobase;
>>>>         sched_rate = rate;
>>>>  }
>>>>
>>>>  static u32 read_sched_clock(void)
>>>>  {
>>>> -       return __raw_readl(sched_io_base);
>>>> +       return ~__raw_readl(sched_io_base + 0x4);
>>> So what about #define what 0x04 is?
>>>
>>> #define MY_FOO_REGISTER_OFFSET 0x04
>>>
>>> raw_readl(sched_io_base + MY_FOO_REGISTER_OFFSET);
>> That define is already there, #define APBTMR_N_CURRENT_VALUE 0x04, but
>> it is in .c file, not in header.
> I'll send a V2 that will address this comment and Stephen Warren's.
Please be sure to CC Daniel Lezcano <daniel.lezcano@linaro.org> on
future versions as he's now maintaining the drivers/clocksource directory.

thanks
-john
diff mbox

Patch

diff --git a/arch/arm/mach-picoxcell/common.h b/arch/arm/mach-picoxcell/common.h
index 481b42a..237fb3b 100644
--- a/arch/arm/mach-picoxcell/common.h
+++ b/arch/arm/mach-picoxcell/common.h
@@ -12,6 +12,4 @@ 
 
 #include <asm/mach/time.h>
 
-extern void dw_apb_timer_init(void);
-
 #endif /* __PICOXCELL_COMMON_H__ */
diff --git a/drivers/clocksource/dw_apb_timer.c b/drivers/clocksource/dw_apb_timer.c
index 8c2a35f..460ac99 100644
--- a/drivers/clocksource/dw_apb_timer.c
+++ b/drivers/clocksource/dw_apb_timer.c
@@ -21,12 +21,6 @@ 
 #define APBT_MIN_PERIOD			4
 #define APBT_MIN_DELTA_USEC		200
 
-#define APBTMR_N_LOAD_COUNT		0x00
-#define APBTMR_N_CURRENT_VALUE		0x04
-#define APBTMR_N_CONTROL		0x08
-#define APBTMR_N_EOI			0x0c
-#define APBTMR_N_INT_STATUS		0x10
-
 #define APBTMRS_INT_STATUS		0xa0
 #define APBTMRS_EOI			0xa4
 #define APBTMRS_RAW_INT_STATUS		0xa8
diff --git a/drivers/clocksource/dw_apb_timer_of.c b/drivers/clocksource/dw_apb_timer_of.c
index ab09ed3..0fa3104 100644
--- a/drivers/clocksource/dw_apb_timer_of.c
+++ b/drivers/clocksource/dw_apb_timer_of.c
@@ -57,6 +57,15 @@  static void add_clockevent(struct device_node *event_timer)
 	dw_apb_clockevent_register(ced);
 }
 
+static void __iomem *sched_io_base;
+
+/* This is actually same as __apbt_read_clocksource(), but with
+   different interface */
+static u32 read_sched_clock_sptimer(void)
+{
+	return ~__raw_readl(sched_io_base + APBTMR_N_CURRENT_VALUE);
+}
+
 static void add_clocksource(struct device_node *source_timer)
 {
 	void __iomem *iobase;
@@ -71,41 +80,27 @@  static void add_clocksource(struct device_node *source_timer)
 
 	dw_apb_clocksource_start(cs);
 	dw_apb_clocksource_register(cs);
-}
 
-static void __iomem *sched_io_base;
-
-static u32 read_sched_clock(void)
-{
-	return __raw_readl(sched_io_base);
+	sched_io_base = iobase;
+	setup_sched_clock(read_sched_clock_sptimer, 32, rate);
 }
 
-static const struct of_device_id sptimer_ids[] __initconst = {
-	{ .compatible = "picochip,pc3x2-rtc" },
+static const struct of_device_id osctimer_ids[] __initconst = {
+	{ .compatible = "picochip,pc3x2-timer" },
+	{ .compatible = "snps,dw-apb-timer-osc" },
 	{ .compatible = "snps,dw-apb-timer-sp" },
-	{ /* Sentinel */ },
+	{  /* Sentinel */ },
 };
 
-static void init_sched_clock(void)
-{
-	struct device_node *sched_timer;
-	u32 rate;
-
-	sched_timer = of_find_matching_node(NULL, sptimer_ids);
-	if (!sched_timer)
-		panic("No RTC for sched clock to use");
+/*
+   You don't have to use dw_apb_timer for scheduler clock,
+   this should also work fine on arm:
 
-	timer_get_base_and_rate(sched_timer, &sched_io_base, &rate);
-	of_node_put(sched_timer);
+  twd_local_timer_of_register();
+  arch_timer_of_register();                 
+  arch_timer_sched_clock_init();
+*/
 
-	setup_sched_clock(read_sched_clock, 32, rate);
-}
-
-static const struct of_device_id osctimer_ids[] __initconst = {
-	{ .compatible = "picochip,pc3x2-timer" },
-	{ .compatible = "snps,dw-apb-timer-osc" },
-	{},
-};
 
 void __init dw_apb_timer_init(void)
 {
@@ -121,7 +116,6 @@  void __init dw_apb_timer_init(void)
 		panic("No timer for clocksource");
 	add_clocksource(source_timer);
 
+	of_node_put(event_timer);
 	of_node_put(source_timer);
-
-	init_sched_clock();
 }
diff --git a/include/linux/dw_apb_timer.h b/include/linux/dw_apb_timer.h
index dd755ce..a64c987 100644
--- a/include/linux/dw_apb_timer.h
+++ b/include/linux/dw_apb_timer.h
@@ -17,6 +17,12 @@ 
 #include <linux/clocksource.h>
 #include <linux/interrupt.h>
 
+#define APBTMR_N_LOAD_COUNT		0x00
+#define APBTMR_N_CURRENT_VALUE		0x04
+#define APBTMR_N_CONTROL		0x08
+#define APBTMR_N_EOI			0x0c
+#define APBTMR_N_INT_STATUS		0x10
+
 #define APBTMRS_REG_SIZE       0x14
 
 struct dw_apb_timer {