Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1660228 > unrolled thread

Re: [PATCH 46/58] clocksource/drivers: Add a new driver for the Atmel ARM TC blocks

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  Re: [PATCH 46/58] clocksource/drivers: Add a new driver for the  Atmel ARM TC blocks Daniel Lezcano <daniel.lezcano@linaro.org> - 2017-06-07 23:10 +0200
    Re: [PATCH 46/58] clocksource/drivers: Add a new driver for the  Atmel ARM TC blocks Alexandre Belloni <alexandre.belloni@free-electrons.com> - 2017-06-08 01:20 +0200
      Re: [PATCH 46/58] clocksource/drivers: Add a new driver for the  Atmel ARM TC blocks Boris Brezillon <boris.brezillon@free-electrons.com> - 2017-06-08 07:50 +0200
        Re: [PATCH 46/58] clocksource/drivers: Add a new driver for the  Atmel ARM TC blocks Daniel Lezcano <daniel.lezcano@linaro.org> - 2017-06-08 09:50 +0200
          Re: [PATCH 46/58] clocksource/drivers: Add a new driver for the  Atmel ARM TC blocks Alexandre Belloni <alexandre.belloni@free-electrons.com> - 2017-06-08 10:00 +0200
            Re: [PATCH 46/58] clocksource/drivers: Add a new driver for the  Atmel ARM TC blocks Daniel Lezcano <daniel.lezcano@linaro.org> - 2017-06-08 10:30 +0200
              Re: [PATCH 46/58] clocksource/drivers: Add a new driver for the  Atmel ARM TC blocks Boris Brezillon <boris.brezillon@free-electrons.com> - 2017-06-08 10:40 +0200
              Re: [PATCH 46/58] clocksource/drivers: Add a new driver for the  Atmel ARM TC blocks Alexandre Belloni <alexandre.belloni@free-electrons.com> - 2017-06-08 10:50 +0200
          Re: [PATCH 46/58] clocksource/drivers: Add a new driver for the  Atmel ARM TC blocks Boris Brezillon <boris.brezillon@free-electrons.com> - 2017-06-08 10:20 +0200
            Re: [PATCH 46/58] clocksource/drivers: Add a new driver for the  Atmel ARM TC blocks Daniel Lezcano <daniel.lezcano@linaro.org> - 2017-06-08 10:50 +0200
              Re: [PATCH 46/58] clocksource/drivers: Add a new driver for the  Atmel ARM TC blocks Boris Brezillon <boris.brezillon@free-electrons.com> - 2017-06-08 11:00 +0200
              Re: [PATCH 46/58] clocksource/drivers: Add a new driver for the Atmel  ARM TC blocks Nicolas Ferre <nicolas.ferre@microchip.com> - 2017-06-12 15:00 +0200
                Re: [PATCH 46/58] clocksource/drivers: Add a new driver for the Atmel  ARM TC blocks Daniel Lezcano <daniel.lezcano@linaro.org> - 2017-06-12 15:30 +0200
                  Re: [PATCH 46/58] clocksource/drivers: Add a new driver for the Atmel  ARM TC blocks Nicolas Ferre <nicolas.ferre@microchip.com> - 2017-06-12 17:30 +0200
    Re: [PATCH 46/58] clocksource/drivers: Add a new driver for the  Atmel ARM TC blocks Boris Brezillon <boris.brezillon@free-electrons.com> - 2017-06-08 08:30 +0200

#1660228 — Re: [PATCH 46/58] clocksource/drivers: Add a new driver for the Atmel ARM TC blocks

On Wed, Jun 07, 2017 at 05:27:50PM +0200, Alexandre Belloni wrote:
> On 07/06/2017 at 16:17:35 +0200, Daniel Lezcano wrote:
> > > You mean you don't like how it is used? Or you don't think having the
> > > timer full name is useful?
> > 
> > The field is not needed, the only place where it is used is where we affect it.
> >  
> 
> It is used in tcb_clksrc_register:
> 
> tc.clksrc.name = tc.name;

Yes, but tc.name is only in the scope of the function, so there is no need to
include this field in the structure, just a working buffer in the function is
enough.
 
> > > The main reason is that RobH doesn't want to have the use (clocksource
> > > or clockevent) of the timer in the DT so when probing a timer, I need to
> > > know whether I already have a clocksource to decide when it is time to
> > > register a clockevent.
> > 
> > Yes, we had this discussion some weeks ago.
> > 
> > This registered hack forces the DT to define first the clocksource, then the
> > clockevent.
> > 
> > So, I suggest you fold the timer definition into a single one like the other
> > drivers.
> 
> I was going to agree but this is not flexible enough because the
> quadrature decoder always uses the first two channels. So on some
> products, we may have:
>  - TCB0:
>    o channels 0,1: qdec
>    o channel 2: clocksource
> 
>  - TCB1:
>    o channels 0,1: qdec
>    o channel 2: clockevent
> 
> This avoids wasting TCB channels.

Ok. In this case you can check if the interrupt is specified for the node, if
yes, then it is a clockevent.

-- 

 <http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs

Follow Linaro:  <http://www.facebook.com/pages/Linaro> Facebook |
<http://twitter.com/#!/linaroorg> Twitter |
<http://www.linaro.org/linaro-blog/> Blog

[toc] | [next] | [standalone]

#1660381

On 07/06/2017 at 23:08:48 +0200, Daniel Lezcano wrote:
> > I was going to agree but this is not flexible enough because the
> > quadrature decoder always uses the first two channels. So on some
> > products, we may have:
> >  - TCB0:
> >    o channels 0,1: qdec
> >    o channel 2: clocksource
> > 
> >  - TCB1:
> >    o channels 0,1: qdec
> >    o channel 2: clockevent
> > 
> > This avoids wasting TCB channels.
> 
> Ok. In this case you can check if the interrupt is specified for the node, if
> yes, then it is a clockevent.
> 

But currently it is always specified in the SoC's dtsi. I don't find
that too practical to push that to the board's dts. Also, lying by
omission (the IRQ is always wired) in the DT is not different from
having a property selecting which timer is the clocksource and which is
the clockevent.

-- 
Alexandre Belloni, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com

[toc] | [prev] | [next] | [standalone]

#1660760

Le Thu, 8 Jun 2017 01:17:15 +0200,
Alexandre Belloni <alexandre.belloni@free-electrons.com> a écrit :

> On 07/06/2017 at 23:08:48 +0200, Daniel Lezcano wrote:
> > > I was going to agree but this is not flexible enough because the
> > > quadrature decoder always uses the first two channels. So on some
> > > products, we may have:
> > >  - TCB0:
> > >    o channels 0,1: qdec
> > >    o channel 2: clocksource
> > > 
> > >  - TCB1:
> > >    o channels 0,1: qdec
> > >    o channel 2: clockevent
> > > 
> > > This avoids wasting TCB channels.  
> > 
> > Ok. In this case you can check if the interrupt is specified for the node, if
> > yes, then it is a clockevent.
> >   
> 
> But currently it is always specified in the SoC's dtsi. I don't find
> that too practical to push that to the board's dts. Also, lying by
> omission (the IRQ is always wired) in the DT is not different from
> having a property selecting which timer is the clocksource and which is
> the clockevent.
> 

I agree with Alexandre here. Really, there's not much we can do to
detect which timer should be used as a clockevent and which one should
be used as a clocksource except explicitly specifying it in the DT.
Having an interrupt defined in one case (clockevent) and undefined in
the other case (clocksource), is just as hack-ish as the detection logic
Alexandre developed to avoid explicitly specifying the function
assigned to a specific timer.

Can we please find a solution that makes everyone happy (DT,
clocksoure/clockevent and at91 maintainers)?

How about adding a linux,timer-function property to specify which
function this timer is providing?

Something like that for example:

	tcb0: timer@fff7c000 {
		compatible = "atmel,at91rm9200-tcb", "simple-mfd", "syscon";
		#address-cells = <1>;
		#size-cells = <0>;
		reg = <0xfff7c000 0x100>;
		interrupts = <18 4>;
		clocks = <&tcb0_clk>, <&clk32k>;
		clock-names = "t0_clk", "slow_clk";

		timer@0 {
			compatible = "atmel,tcb-timer";
			reg = <0>, <1>;
			linux,timer-function = "clocksource";
		};

		timer@2 {
			compatible = "atmel,tcb-timer";
			reg = <2>;
			linux,timer-function = "clockevent";
		};
	};

Alternatively, we could have a property or a node in chosen describing which
timer should be used:

	chosen {
		clockevent {
			timer = <&timer2>;
		};

		clocksource {
			timer = <&timer0>;
		};

		/*
		 * or
		 *
		 * clockevent = <&timer2>;
		 * clocksource = <&timer0>;
		 *
		 * but I think the clocksource/clockevent node approach
		 * is more future proof in case we need to add extra
		 * information like the expected resolution/precision or
		 * anything that could be tweakable.
		 */
	};

	tcb0: timer@fff7c000 {
		compatible = "atmel,at91rm9200-tcb", "simple-mfd", "syscon";
		#address-cells = <1>;
		#size-cells = <0>;
		reg = <0xfff7c000 0x100>;
		interrupts = <18 4>;
		clocks = <&tcb0_clk>, <&clk32k>;
		clock-names = "t0_clk", "slow_clk";

		timer0: timer@0 {
			compatible = "atmel,tcb-timer";
			reg = <0>, <1>;
		};

		timer2: timer@2 {
			compatible = "atmel,tcb-timer";
			reg = <2>;
		};
	};

[toc] | [prev] | [next] | [standalone]

#1660874

+Mark Rutland, +Rob Herring


Alexandre, Boris, have a look at https://www.spinics.net/lists/arm-kernel/msg572652.html

That will tell you the story.

On Thu, Jun 08, 2017 at 07:42:36AM +0200, Boris Brezillon wrote:
> Le Thu, 8 Jun 2017 01:17:15 +0200,
> Alexandre Belloni <alexandre.belloni@free-electrons.com> a écrit :
> 
> > On 07/06/2017 at 23:08:48 +0200, Daniel Lezcano wrote:
> > > > I was going to agree but this is not flexible enough because the
> > > > quadrature decoder always uses the first two channels. So on some
> > > > products, we may have:
> > > >  - TCB0:
> > > >    o channels 0,1: qdec
> > > >    o channel 2: clocksource
> > > > 
> > > >  - TCB1:
> > > >    o channels 0,1: qdec
> > > >    o channel 2: clockevent
> > > > 
> > > > This avoids wasting TCB channels.  
> > > 
> > > Ok. In this case you can check if the interrupt is specified for the node, if
> > > yes, then it is a clockevent.
> > >   
> > 
> > But currently it is always specified in the SoC's dtsi. I don't find
> > that too practical to push that to the board's dts. Also, lying by
> > omission (the IRQ is always wired) in the DT is not different from
> > having a property selecting which timer is the clocksource and which is
> > the clockevent.
> > 
> 
> I agree with Alexandre here. Really, there's not much we can do to
> detect which timer should be used as a clockevent and which one should
> be used as a clocksource except explicitly specifying it in the DT.
> Having an interrupt defined in one case (clockevent) and undefined in
> the other case (clocksource), is just as hack-ish as the detection logic
> Alexandre developed to avoid explicitly specifying the function
> assigned to a specific timer.
> 
> Can we please find a solution that makes everyone happy (DT,
> clocksoure/clockevent and at91 maintainers)?
> 
> How about adding a linux,timer-function property to specify which
> function this timer is providing?
> 
> Something like that for example:
> 
> 	tcb0: timer@fff7c000 {
> 		compatible = "atmel,at91rm9200-tcb", "simple-mfd", "syscon";
> 		#address-cells = <1>;
> 		#size-cells = <0>;
> 		reg = <0xfff7c000 0x100>;
> 		interrupts = <18 4>;
> 		clocks = <&tcb0_clk>, <&clk32k>;
> 		clock-names = "t0_clk", "slow_clk";
> 
> 		timer@0 {
> 			compatible = "atmel,tcb-timer";
> 			reg = <0>, <1>;
> 			linux,timer-function = "clocksource";
> 		};
> 
> 		timer@2 {
> 			compatible = "atmel,tcb-timer";
> 			reg = <2>;
> 			linux,timer-function = "clockevent";
> 		};
> 	};
> 
> Alternatively, we could have a property or a node in chosen describing which
> timer should be used:
> 
> 	chosen {
> 		clockevent {
> 			timer = <&timer2>;
> 		};
> 
> 		clocksource {
> 			timer = <&timer0>;
> 		};
> 
> 		/*
> 		 * or
> 		 *
> 		 * clockevent = <&timer2>;
> 		 * clocksource = <&timer0>;
> 		 *
> 		 * but I think the clocksource/clockevent node approach
> 		 * is more future proof in case we need to add extra
> 		 * information like the expected resolution/precision or
> 		 * anything that could be tweakable.
> 		 */
> 	};
> 
> 	tcb0: timer@fff7c000 {
> 		compatible = "atmel,at91rm9200-tcb", "simple-mfd", "syscon";
> 		#address-cells = <1>;
> 		#size-cells = <0>;
> 		reg = <0xfff7c000 0x100>;
> 		interrupts = <18 4>;
> 		clocks = <&tcb0_clk>, <&clk32k>;
> 		clock-names = "t0_clk", "slow_clk";
> 
> 		timer0: timer@0 {
> 			compatible = "atmel,tcb-timer";
> 			reg = <0>, <1>;
> 		};
> 
> 		timer2: timer@2 {
> 			compatible = "atmel,tcb-timer";
> 			reg = <2>;
> 		};
> 	};

-- 

 <http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs

Follow Linaro:  <http://www.facebook.com/pages/Linaro> Facebook |
<http://twitter.com/#!/linaroorg> Twitter |
<http://www.linaro.org/linaro-blog/> Blog

[toc] | [prev] | [next] | [standalone]

#1660898

On 08/06/2017 at 09:44:46 +0200, Daniel Lezcano wrote:
> 
> +Mark Rutland, +Rob Herring
> 
> 
> Alexandre, Boris, have a look at https://www.spinics.net/lists/arm-kernel/msg572652.html
> 
> That will tell you the story.
> 

Ok, so is the solution putting the driver back in mach-at91 were we can
do whatever we want like mach-omap2 is doing?

> On Thu, Jun 08, 2017 at 07:42:36AM +0200, Boris Brezillon wrote:
> > Le Thu, 8 Jun 2017 01:17:15 +0200,
> > Alexandre Belloni <alexandre.belloni@free-electrons.com> a écrit :
> > 
> > > On 07/06/2017 at 23:08:48 +0200, Daniel Lezcano wrote:
> > > > > I was going to agree but this is not flexible enough because the
> > > > > quadrature decoder always uses the first two channels. So on some
> > > > > products, we may have:
> > > > >  - TCB0:
> > > > >    o channels 0,1: qdec
> > > > >    o channel 2: clocksource
> > > > > 
> > > > >  - TCB1:
> > > > >    o channels 0,1: qdec
> > > > >    o channel 2: clockevent
> > > > > 
> > > > > This avoids wasting TCB channels.  
> > > > 
> > > > Ok. In this case you can check if the interrupt is specified for the node, if
> > > > yes, then it is a clockevent.
> > > >   
> > > 
> > > But currently it is always specified in the SoC's dtsi. I don't find
> > > that too practical to push that to the board's dts. Also, lying by
> > > omission (the IRQ is always wired) in the DT is not different from
> > > having a property selecting which timer is the clocksource and which is
> > > the clockevent.
> > > 
> > 
> > I agree with Alexandre here. Really, there's not much we can do to
> > detect which timer should be used as a clockevent and which one should
> > be used as a clocksource except explicitly specifying it in the DT.
> > Having an interrupt defined in one case (clockevent) and undefined in
> > the other case (clocksource), is just as hack-ish as the detection logic
> > Alexandre developed to avoid explicitly specifying the function
> > assigned to a specific timer.
> > 
> > Can we please find a solution that makes everyone happy (DT,
> > clocksoure/clockevent and at91 maintainers)?
> > 
> > How about adding a linux,timer-function property to specify which
> > function this timer is providing?
> > 
> > Something like that for example:
> > 
> > 	tcb0: timer@fff7c000 {
> > 		compatible = "atmel,at91rm9200-tcb", "simple-mfd", "syscon";
> > 		#address-cells = <1>;
> > 		#size-cells = <0>;
> > 		reg = <0xfff7c000 0x100>;
> > 		interrupts = <18 4>;
> > 		clocks = <&tcb0_clk>, <&clk32k>;
> > 		clock-names = "t0_clk", "slow_clk";
> > 
> > 		timer@0 {
> > 			compatible = "atmel,tcb-timer";
> > 			reg = <0>, <1>;
> > 			linux,timer-function = "clocksource";
> > 		};
> > 
> > 		timer@2 {
> > 			compatible = "atmel,tcb-timer";
> > 			reg = <2>;
> > 			linux,timer-function = "clockevent";
> > 		};
> > 	};
> > 
> > Alternatively, we could have a property or a node in chosen describing which
> > timer should be used:
> > 
> > 	chosen {
> > 		clockevent {
> > 			timer = <&timer2>;
> > 		};
> > 
> > 		clocksource {
> > 			timer = <&timer0>;
> > 		};
> > 
> > 		/*
> > 		 * or
> > 		 *
> > 		 * clockevent = <&timer2>;
> > 		 * clocksource = <&timer0>;
> > 		 *
> > 		 * but I think the clocksource/clockevent node approach
> > 		 * is more future proof in case we need to add extra
> > 		 * information like the expected resolution/precision or
> > 		 * anything that could be tweakable.
> > 		 */
> > 	};
> > 
> > 	tcb0: timer@fff7c000 {
> > 		compatible = "atmel,at91rm9200-tcb", "simple-mfd", "syscon";
> > 		#address-cells = <1>;
> > 		#size-cells = <0>;
> > 		reg = <0xfff7c000 0x100>;
> > 		interrupts = <18 4>;
> > 		clocks = <&tcb0_clk>, <&clk32k>;
> > 		clock-names = "t0_clk", "slow_clk";
> > 
> > 		timer0: timer@0 {
> > 			compatible = "atmel,tcb-timer";
> > 			reg = <0>, <1>;
> > 		};
> > 
> > 		timer2: timer@2 {
> > 			compatible = "atmel,tcb-timer";
> > 			reg = <2>;
> > 		};
> > 	};
> 
> -- 
> 
>  <http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs
> 
> Follow Linaro:  <http://www.facebook.com/pages/Linaro> Facebook |
> <http://twitter.com/#!/linaroorg> Twitter |
> <http://www.linaro.org/linaro-blog/> Blog

-- 
Alexandre Belloni, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com

[toc] | [prev] | [next] | [standalone]

#1660930

On Thu, Jun 08, 2017 at 09:59:01AM +0200, Alexandre Belloni wrote:
> On 08/06/2017 at 09:44:46 +0200, Daniel Lezcano wrote:
> > 
> > +Mark Rutland, +Rob Herring
> > 
> > 
> > Alexandre, Boris, have a look at https://www.spinics.net/lists/arm-kernel/msg572652.html
> > 
> > That will tell you the story.
> > 
> 
> Ok, so is the solution putting the driver back in mach-at91 were we can
> do whatever we want like mach-omap2 is doing?

No. And putting a driver in mach-<whatever> does not give the permission to do
whatever you want. I won't tell you how OSS works, but moving code around or
using another tree to circumvent a code review is just the best way to upset
maintainers in general and hurt your karma.

That said, I think you misunderstood my comment (or I was not clear). In the
discussion given in the link above, I am in favor, somehow, to distinguish
clockevent and clocksource to solve exactly what you are facing.

Rob Herring told me it could be acceptable to have a property to tell if it is
a clockevent or a clocksource.

Mark Rutland disagreed on this.

I was alone in the discussion, no consensus have been found.

Now, you have a particular use case and I would like to resurrect the
discussion in order to find a solution which can apply to all DT drivers.

> > On Thu, Jun 08, 2017 at 07:42:36AM +0200, Boris Brezillon wrote:
> > > Le Thu, 8 Jun 2017 01:17:15 +0200,
> > > Alexandre Belloni <alexandre.belloni@free-electrons.com> a écrit :
> > > 
> > > > On 07/06/2017 at 23:08:48 +0200, Daniel Lezcano wrote:
> > > > > > I was going to agree but this is not flexible enough because the
> > > > > > quadrature decoder always uses the first two channels. So on some
> > > > > > products, we may have:
> > > > > >  - TCB0:
> > > > > >    o channels 0,1: qdec
> > > > > >    o channel 2: clocksource
> > > > > > 
> > > > > >  - TCB1:
> > > > > >    o channels 0,1: qdec
> > > > > >    o channel 2: clockevent
> > > > > > 
> > > > > > This avoids wasting TCB channels.  
> > > > > 
> > > > > Ok. In this case you can check if the interrupt is specified for the node, if
> > > > > yes, then it is a clockevent.
> > > > >   
> > > > 
> > > > But currently it is always specified in the SoC's dtsi. I don't find
> > > > that too practical to push that to the board's dts. Also, lying by
> > > > omission (the IRQ is always wired) in the DT is not different from
> > > > having a property selecting which timer is the clocksource and which is
> > > > the clockevent.
> > > > 
> > > 
> > > I agree with Alexandre here. Really, there's not much we can do to
> > > detect which timer should be used as a clockevent and which one should
> > > be used as a clocksource except explicitly specifying it in the DT.
> > > Having an interrupt defined in one case (clockevent) and undefined in
> > > the other case (clocksource), is just as hack-ish as the detection logic
> > > Alexandre developed to avoid explicitly specifying the function
> > > assigned to a specific timer.
> > > 
> > > Can we please find a solution that makes everyone happy (DT,
> > > clocksoure/clockevent and at91 maintainers)?
> > > 
> > > How about adding a linux,timer-function property to specify which
> > > function this timer is providing?
> > > 
> > > Something like that for example:
> > > 
> > > 	tcb0: timer@fff7c000 {
> > > 		compatible = "atmel,at91rm9200-tcb", "simple-mfd", "syscon";
> > > 		#address-cells = <1>;
> > > 		#size-cells = <0>;
> > > 		reg = <0xfff7c000 0x100>;
> > > 		interrupts = <18 4>;
> > > 		clocks = <&tcb0_clk>, <&clk32k>;
> > > 		clock-names = "t0_clk", "slow_clk";
> > > 
> > > 		timer@0 {
> > > 			compatible = "atmel,tcb-timer";
> > > 			reg = <0>, <1>;
> > > 			linux,timer-function = "clocksource";
> > > 		};
> > > 
> > > 		timer@2 {
> > > 			compatible = "atmel,tcb-timer";
> > > 			reg = <2>;
> > > 			linux,timer-function = "clockevent";
> > > 		};
> > > 	};
> > > 
> > > Alternatively, we could have a property or a node in chosen describing which
> > > timer should be used:
> > > 
> > > 	chosen {
> > > 		clockevent {
> > > 			timer = <&timer2>;
> > > 		};
> > > 
> > > 		clocksource {
> > > 			timer = <&timer0>;
> > > 		};
> > > 
> > > 		/*
> > > 		 * or
> > > 		 *
> > > 		 * clockevent = <&timer2>;
> > > 		 * clocksource = <&timer0>;
> > > 		 *
> > > 		 * but I think the clocksource/clockevent node approach
> > > 		 * is more future proof in case we need to add extra
> > > 		 * information like the expected resolution/precision or
> > > 		 * anything that could be tweakable.
> > > 		 */
> > > 	};
> > > 
> > > 	tcb0: timer@fff7c000 {
> > > 		compatible = "atmel,at91rm9200-tcb", "simple-mfd", "syscon";
> > > 		#address-cells = <1>;
> > > 		#size-cells = <0>;
> > > 		reg = <0xfff7c000 0x100>;
> > > 		interrupts = <18 4>;
> > > 		clocks = <&tcb0_clk>, <&clk32k>;
> > > 		clock-names = "t0_clk", "slow_clk";
> > > 
> > > 		timer0: timer@0 {
> > > 			compatible = "atmel,tcb-timer";
> > > 			reg = <0>, <1>;
> > > 		};
> > > 
> > > 		timer2: timer@2 {
> > > 			compatible = "atmel,tcb-timer";
> > > 			reg = <2>;
> > > 		};
> > > 	};
> > 
> > -- 
> > 
> >  <http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs
> > 
> > Follow Linaro:  <http://www.facebook.com/pages/Linaro> Facebook |
> > <http://twitter.com/#!/linaroorg> Twitter |
> > <http://www.linaro.org/linaro-blog/> Blog
> 
> -- 
> Alexandre Belloni, Free Electrons
> Embedded Linux and Kernel engineering
> http://free-electrons.com

-- 

 <http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs

Follow Linaro:  <http://www.facebook.com/pages/Linaro> Facebook |
<http://twitter.com/#!/linaroorg> Twitter |
<http://www.linaro.org/linaro-blog/> Blog

[toc] | [prev] | [next] | [standalone]

#1660940

On Thu, 8 Jun 2017 10:24:17 +0200
Daniel Lezcano <daniel.lezcano@linaro.org> wrote:

> On Thu, Jun 08, 2017 at 09:59:01AM +0200, Alexandre Belloni wrote:
> > On 08/06/2017 at 09:44:46 +0200, Daniel Lezcano wrote:  
> > > 
> > > +Mark Rutland, +Rob Herring
> > > 
> > > 
> > > Alexandre, Boris, have a look at https://www.spinics.net/lists/arm-kernel/msg572652.html
> > > 
> > > That will tell you the story.
> > >   
> > 
> > Ok, so is the solution putting the driver back in mach-at91 were we can
> > do whatever we want like mach-omap2 is doing?  
> 
> No. And putting a driver in mach-<whatever> does not give the permission to do
> whatever you want. I won't tell you how OSS works, but moving code around or
> using another tree to circumvent a code review is just the best way to upset
> maintainers in general and hurt your karma.
> 
> That said, I think you misunderstood my comment (or I was not clear). In the
> discussion given in the link above, I am in favor, somehow, to distinguish
> clockevent and clocksource to solve exactly what you are facing.
> 
> Rob Herring told me it could be acceptable to have a property to tell if it is
> a clockevent or a clocksource.
> 
> Mark Rutland disagreed on this.
> 
> I was alone in the discussion, no consensus have been found.

Indeed, I misunderstood your point.

> 
> Now, you have a particular use case and I would like to resurrect the
> discussion in order to find a solution which can apply to all DT drivers.

Ok, glad to see we're on the same page.

Mark, can we re-open the discussion?

Thanks,

Boris

[toc] | [prev] | [next] | [standalone]

#1660943

On 08/06/2017 at 10:24:17 +0200, Daniel Lezcano wrote:
> On Thu, Jun 08, 2017 at 09:59:01AM +0200, Alexandre Belloni wrote:
> > On 08/06/2017 at 09:44:46 +0200, Daniel Lezcano wrote:
> > > 
> > > +Mark Rutland, +Rob Herring
> > > 
> > > 
> > > Alexandre, Boris, have a look at https://www.spinics.net/lists/arm-kernel/msg572652.html
> > > 
> > > That will tell you the story.
> > > 
> > 
> > Ok, so is the solution putting the driver back in mach-at91 were we can
> > do whatever we want like mach-omap2 is doing?
> 
> No. And putting a driver in mach-<whatever> does not give the permission to do
> whatever you want. I won't tell you how OSS works, but moving code around or
> using another tree to circumvent a code review is just the best way to upset
> maintainers in general and hurt your karma.
> 

I know that but some SoCs will not be able to even boot until we have a
solution. And it has been one year since we raised the issue without any
solution coming from the DT maintainers.


> That said, I think you misunderstood my comment (or I was not clear). In the
> discussion given in the link above, I am in favor, somehow, to distinguish
> clockevent and clocksource to solve exactly what you are facing.
> 
> Rob Herring told me it could be acceptable to have a property to tell if it is
> a clockevent or a clocksource.
> 

Ok, then let's do it!


> Mark Rutland disagreed on this.
> 
> I was alone in the discussion, no consensus have been found.
> 
> Now, you have a particular use case and I would like to resurrect the
> discussion in order to find a solution which can apply to all DT drivers.
> 

Again, it has been one year and it seems nobody actually cares. There is
a whole family of SoCs that can't boot because of that. Should I resort
to the evil vendor tree argument again?




> > > On Thu, Jun 08, 2017 at 07:42:36AM +0200, Boris Brezillon wrote:
> > > > Le Thu, 8 Jun 2017 01:17:15 +0200,
> > > > Alexandre Belloni <alexandre.belloni@free-electrons.com> a écrit :
> > > > 
> > > > > On 07/06/2017 at 23:08:48 +0200, Daniel Lezcano wrote:
> > > > > > > I was going to agree but this is not flexible enough because the
> > > > > > > quadrature decoder always uses the first two channels. So on some
> > > > > > > products, we may have:
> > > > > > >  - TCB0:
> > > > > > >    o channels 0,1: qdec
> > > > > > >    o channel 2: clocksource
> > > > > > > 
> > > > > > >  - TCB1:
> > > > > > >    o channels 0,1: qdec
> > > > > > >    o channel 2: clockevent
> > > > > > > 
> > > > > > > This avoids wasting TCB channels.  
> > > > > > 
> > > > > > Ok. In this case you can check if the interrupt is specified for the node, if
> > > > > > yes, then it is a clockevent.
> > > > > >   
> > > > > 
> > > > > But currently it is always specified in the SoC's dtsi. I don't find
> > > > > that too practical to push that to the board's dts. Also, lying by
> > > > > omission (the IRQ is always wired) in the DT is not different from
> > > > > having a property selecting which timer is the clocksource and which is
> > > > > the clockevent.
> > > > > 
> > > > 
> > > > I agree with Alexandre here. Really, there's not much we can do to
> > > > detect which timer should be used as a clockevent and which one should
> > > > be used as a clocksource except explicitly specifying it in the DT.
> > > > Having an interrupt defined in one case (clockevent) and undefined in
> > > > the other case (clocksource), is just as hack-ish as the detection logic
> > > > Alexandre developed to avoid explicitly specifying the function
> > > > assigned to a specific timer.
> > > > 
> > > > Can we please find a solution that makes everyone happy (DT,
> > > > clocksoure/clockevent and at91 maintainers)?
> > > > 
> > > > How about adding a linux,timer-function property to specify which
> > > > function this timer is providing?
> > > > 
> > > > Something like that for example:
> > > > 
> > > > 	tcb0: timer@fff7c000 {
> > > > 		compatible = "atmel,at91rm9200-tcb", "simple-mfd", "syscon";
> > > > 		#address-cells = <1>;
> > > > 		#size-cells = <0>;
> > > > 		reg = <0xfff7c000 0x100>;
> > > > 		interrupts = <18 4>;
> > > > 		clocks = <&tcb0_clk>, <&clk32k>;
> > > > 		clock-names = "t0_clk", "slow_clk";
> > > > 
> > > > 		timer@0 {
> > > > 			compatible = "atmel,tcb-timer";
> > > > 			reg = <0>, <1>;
> > > > 			linux,timer-function = "clocksource";
> > > > 		};
> > > > 
> > > > 		timer@2 {
> > > > 			compatible = "atmel,tcb-timer";
> > > > 			reg = <2>;
> > > > 			linux,timer-function = "clockevent";
> > > > 		};
> > > > 	};
> > > > 
> > > > Alternatively, we could have a property or a node in chosen describing which
> > > > timer should be used:
> > > > 
> > > > 	chosen {
> > > > 		clockevent {
> > > > 			timer = <&timer2>;
> > > > 		};
> > > > 
> > > > 		clocksource {
> > > > 			timer = <&timer0>;
> > > > 		};
> > > > 
> > > > 		/*
> > > > 		 * or
> > > > 		 *
> > > > 		 * clockevent = <&timer2>;
> > > > 		 * clocksource = <&timer0>;
> > > > 		 *
> > > > 		 * but I think the clocksource/clockevent node approach
> > > > 		 * is more future proof in case we need to add extra
> > > > 		 * information like the expected resolution/precision or
> > > > 		 * anything that could be tweakable.
> > > > 		 */
> > > > 	};
> > > > 
> > > > 	tcb0: timer@fff7c000 {
> > > > 		compatible = "atmel,at91rm9200-tcb", "simple-mfd", "syscon";
> > > > 		#address-cells = <1>;
> > > > 		#size-cells = <0>;
> > > > 		reg = <0xfff7c000 0x100>;
> > > > 		interrupts = <18 4>;
> > > > 		clocks = <&tcb0_clk>, <&clk32k>;
> > > > 		clock-names = "t0_clk", "slow_clk";
> > > > 
> > > > 		timer0: timer@0 {
> > > > 			compatible = "atmel,tcb-timer";
> > > > 			reg = <0>, <1>;
> > > > 		};
> > > > 
> > > > 		timer2: timer@2 {
> > > > 			compatible = "atmel,tcb-timer";
> > > > 			reg = <2>;
> > > > 		};
> > > > 	};
> > > 
> > > -- 
> > > 
> > >  <http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs
> > > 
> > > Follow Linaro:  <http://www.facebook.com/pages/Linaro> Facebook |
> > > <http://twitter.com/#!/linaroorg> Twitter |
> > > <http://www.linaro.org/linaro-blog/> Blog
> > 
> > -- 
> > Alexandre Belloni, Free Electrons
> > Embedded Linux and Kernel engineering
> > http://free-electrons.com
> 
> -- 
> 
>  <http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs
> 
> Follow Linaro:  <http://www.facebook.com/pages/Linaro> Facebook |
> <http://twitter.com/#!/linaroorg> Twitter |
> <http://www.linaro.org/linaro-blog/> Blog

-- 
Alexandre Belloni, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com

[toc] | [prev] | [next] | [standalone]

#1660917

On Thu, 8 Jun 2017 09:44:46 +0200
Daniel Lezcano <daniel.lezcano@linaro.org> wrote:

> +Mark Rutland, +Rob Herring

Mark doesn't seem to be CCed.

> 
> 
> Alexandre, Boris, have a look at https://www.spinics.net/lists/arm-kernel/msg572652.html
> 
> That will tell you the story.

Then we're in a deadlock situation here. I'm tired of hearing this kind
of argument "DT is only supposed to describe HW, not configuration, bla
bla". The truth is, we already have plenty of bindings that do not
strictly describe HW.

A simple example: ECC configuration on NAND devices. This is clearly a
configuration thing, the NAND controller is usually able to support
several kind of strength+ECC-block-size config, but we are able to
overload this with the nand-ecc-xxx properties. Another example, still
MTD related: MTD partitions, this is purely a software configuration,
still we allow users to pass this information in the DT. You want
another one? What about the linux,code and linux,input-type properties
described here [1]?

So please, let's not use these "this is not decribing HW" or "this is
linux specific" arguments every time someone tries to encode something
that can be considered a configuration detail.

Let's be pragmatic. How you want to use your timer counter blocks (I'm
talking about atmel TCBs) is clearly board specific. Whether you want
to use the PIT for your clocksource or use one or 2 channels of a TCB
at a specific resolution is again board specific. We need a solution to
assign timer channels to a linux function, and I'm not convinced
passing this information through the command line makes much more sense
than specifying it in the DT (and it's definitely less intuitive, since
you have to reference something defined in the DT from the cmdline).

Now, in his review, Mark says:

"
To me it sounds like what we need is Linux infrastructure that allows
one to register a device as having both clockevent/clocksource
functionality.

That way, we can choose to do something sane at boot time, and if the
user really wants specific devices used in specific ways, they can
request that.
"

Does that mean that, after adding this "HW timer" infrastructure, we
would have a standard way to assign a function to a specific timer
block from the DT? How is this different from what I suggest below
(note the linux, prefix on my linux,timer-function property, which
clearly shows that this is Linux specific)?

> 
> On Thu, Jun 08, 2017 at 07:42:36AM +0200, Boris Brezillon wrote:
> > Le Thu, 8 Jun 2017 01:17:15 +0200,
> > Alexandre Belloni <alexandre.belloni@free-electrons.com> a écrit :
> >   
> > > On 07/06/2017 at 23:08:48 +0200, Daniel Lezcano wrote:  
> > > > > I was going to agree but this is not flexible enough because the
> > > > > quadrature decoder always uses the first two channels. So on some
> > > > > products, we may have:
> > > > >  - TCB0:
> > > > >    o channels 0,1: qdec
> > > > >    o channel 2: clocksource
> > > > > 
> > > > >  - TCB1:
> > > > >    o channels 0,1: qdec
> > > > >    o channel 2: clockevent
> > > > > 
> > > > > This avoids wasting TCB channels.    
> > > > 
> > > > Ok. In this case you can check if the interrupt is specified for the node, if
> > > > yes, then it is a clockevent.
> > > >     
> > > 
> > > But currently it is always specified in the SoC's dtsi. I don't find
> > > that too practical to push that to the board's dts. Also, lying by
> > > omission (the IRQ is always wired) in the DT is not different from
> > > having a property selecting which timer is the clocksource and which is
> > > the clockevent.
> > >   
> > 
> > I agree with Alexandre here. Really, there's not much we can do to
> > detect which timer should be used as a clockevent and which one should
> > be used as a clocksource except explicitly specifying it in the DT.
> > Having an interrupt defined in one case (clockevent) and undefined in
> > the other case (clocksource), is just as hack-ish as the detection logic
> > Alexandre developed to avoid explicitly specifying the function
> > assigned to a specific timer.
> > 
> > Can we please find a solution that makes everyone happy (DT,
> > clocksoure/clockevent and at91 maintainers)?
> > 
> > How about adding a linux,timer-function property to specify which
> > function this timer is providing?
> > 
> > Something like that for example:
> > 
> > 	tcb0: timer@fff7c000 {
> > 		compatible = "atmel,at91rm9200-tcb", "simple-mfd", "syscon";
> > 		#address-cells = <1>;
> > 		#size-cells = <0>;
> > 		reg = <0xfff7c000 0x100>;
> > 		interrupts = <18 4>;
> > 		clocks = <&tcb0_clk>, <&clk32k>;
> > 		clock-names = "t0_clk", "slow_clk";
> > 
> > 		timer@0 {
> > 			compatible = "atmel,tcb-timer";
> > 			reg = <0>, <1>;
> > 			linux,timer-function = "clocksource";
> > 		};
> > 
> > 		timer@2 {
> > 			compatible = "atmel,tcb-timer";
> > 			reg = <2>;
> > 			linux,timer-function = "clockevent";
> > 		};
> > 	};
> > 
> > Alternatively, we could have a property or a node in chosen describing which
> > timer should be used:
> > 
> > 	chosen {
> > 		clockevent {
> > 			timer = <&timer2>;
> > 		};
> > 
> > 		clocksource {
> > 			timer = <&timer0>;
> > 		};
> > 
> > 		/*
> > 		 * or
> > 		 *
> > 		 * clockevent = <&timer2>;
> > 		 * clocksource = <&timer0>;
> > 		 *
> > 		 * but I think the clocksource/clockevent node approach
> > 		 * is more future proof in case we need to add extra
> > 		 * information like the expected resolution/precision or
> > 		 * anything that could be tweakable.
> > 		 */
> > 	};
> > 
> > 	tcb0: timer@fff7c000 {
> > 		compatible = "atmel,at91rm9200-tcb", "simple-mfd", "syscon";
> > 		#address-cells = <1>;
> > 		#size-cells = <0>;
> > 		reg = <0xfff7c000 0x100>;
> > 		interrupts = <18 4>;
> > 		clocks = <&tcb0_clk>, <&clk32k>;
> > 		clock-names = "t0_clk", "slow_clk";
> > 
> > 		timer0: timer@0 {
> > 			compatible = "atmel,tcb-timer";
> > 			reg = <0>, <1>;
> > 		};
> > 
> > 		timer2: timer@2 {
> > 			compatible = "atmel,tcb-timer";
> > 			reg = <2>;
> > 		};
> > 	};  
> 

[1]http://elixir.free-electrons.com/linux/latest/source/Documentation/devicetree/bindings/input/gpio-keys.txt

[toc] | [prev] | [next] | [standalone]

#1660946

On Thu, Jun 08, 2017 at 10:13:34AM +0200, Boris Brezillon wrote:
> On Thu, 8 Jun 2017 09:44:46 +0200
> Daniel Lezcano <daniel.lezcano@linaro.org> wrote:
> 
> > +Mark Rutland, +Rob Herring
> 
> Mark doesn't seem to be CCed.

Ah, yes. Thanks for fixing this.

> > 
> > 
> > Alexandre, Boris, have a look at https://www.spinics.net/lists/arm-kernel/msg572652.html
> > 
> > That will tell you the story.
> 
> Then we're in a deadlock situation here. I'm tired of hearing this kind
> of argument "DT is only supposed to describe HW, not configuration, bla
> bla". The truth is, we already have plenty of bindings that do not
> strictly describe HW.
> 
> A simple example: ECC configuration on NAND devices. This is clearly a
> configuration thing, the NAND controller is usually able to support
> several kind of strength+ECC-block-size config, but we are able to
> overload this with the nand-ecc-xxx properties. Another example, still
> MTD related: MTD partitions, this is purely a software configuration,
> still we allow users to pass this information in the DT. You want
> another one? What about the linux,code and linux,input-type properties
> described here [1]?
> 
> So please, let's not use these "this is not decribing HW" or "this is
> linux specific" arguments every time someone tries to encode something
> that can be considered a configuration detail.
> 
> Let's be pragmatic. How you want to use your timer counter blocks (I'm
> talking about atmel TCBs) is clearly board specific. Whether you want
> to use the PIT for your clocksource or use one or 2 channels of a TCB
> at a specific resolution is again board specific. We need a solution to
> assign timer channels to a linux function, and I'm not convinced
> passing this information through the command line makes much more sense
> than specifying it in the DT (and it's definitely less intuitive, since
> you have to reference something defined in the DT from the cmdline).
> 
> Now, in his review, Mark says:
> 
> "
> To me it sounds like what we need is Linux infrastructure that allows
> one to register a device as having both clockevent/clocksource
> functionality.
> 
> That way, we can choose to do something sane at boot time, and if the
> user really wants specific devices used in specific ways, they can
> request that.
> "
> 
> Does that mean that, after adding this "HW timer" infrastructure, we
> would have a standard way to assign a function to a specific timer
> block from the DT? How is this different from what I suggest below
> (note the linux, prefix on my linux,timer-function property, which
> clearly shows that this is Linux specific)?

I like the 'chosen' approach with the nodes you are proposing below. Thanks for
the constructive suggestion. The binding description matches perfectly what we
are trying to achieve.

Rob? what do you think?

> > On Thu, Jun 08, 2017 at 07:42:36AM +0200, Boris Brezillon wrote:
> > > Le Thu, 8 Jun 2017 01:17:15 +0200,
> > > Alexandre Belloni <alexandre.belloni@free-electrons.com> a écrit :
> > >   
> > > > On 07/06/2017 at 23:08:48 +0200, Daniel Lezcano wrote:  
> > > > > > I was going to agree but this is not flexible enough because the
> > > > > > quadrature decoder always uses the first two channels. So on some
> > > > > > products, we may have:
> > > > > >  - TCB0:
> > > > > >    o channels 0,1: qdec
> > > > > >    o channel 2: clocksource
> > > > > > 
> > > > > >  - TCB1:
> > > > > >    o channels 0,1: qdec
> > > > > >    o channel 2: clockevent
> > > > > > 
> > > > > > This avoids wasting TCB channels.    
> > > > > 
> > > > > Ok. In this case you can check if the interrupt is specified for the node, if
> > > > > yes, then it is a clockevent.
> > > > >     
> > > > 
> > > > But currently it is always specified in the SoC's dtsi. I don't find
> > > > that too practical to push that to the board's dts. Also, lying by
> > > > omission (the IRQ is always wired) in the DT is not different from
> > > > having a property selecting which timer is the clocksource and which is
> > > > the clockevent.
> > > >   
> > > 
> > > I agree with Alexandre here. Really, there's not much we can do to
> > > detect which timer should be used as a clockevent and which one should
> > > be used as a clocksource except explicitly specifying it in the DT.
> > > Having an interrupt defined in one case (clockevent) and undefined in
> > > the other case (clocksource), is just as hack-ish as the detection logic
> > > Alexandre developed to avoid explicitly specifying the function
> > > assigned to a specific timer.
> > > 
> > > Can we please find a solution that makes everyone happy (DT,
> > > clocksoure/clockevent and at91 maintainers)?
> > > 
> > > How about adding a linux,timer-function property to specify which
> > > function this timer is providing?
> > > 
> > > Something like that for example:
> > > 
> > > 	tcb0: timer@fff7c000 {
> > > 		compatible = "atmel,at91rm9200-tcb", "simple-mfd", "syscon";
> > > 		#address-cells = <1>;
> > > 		#size-cells = <0>;
> > > 		reg = <0xfff7c000 0x100>;
> > > 		interrupts = <18 4>;
> > > 		clocks = <&tcb0_clk>, <&clk32k>;
> > > 		clock-names = "t0_clk", "slow_clk";
> > > 
> > > 		timer@0 {
> > > 			compatible = "atmel,tcb-timer";
> > > 			reg = <0>, <1>;
> > > 			linux,timer-function = "clocksource";
> > > 		};
> > > 
> > > 		timer@2 {
> > > 			compatible = "atmel,tcb-timer";
> > > 			reg = <2>;
> > > 			linux,timer-function = "clockevent";
> > > 		};
> > > 	};
> > > 
> > > Alternatively, we could have a property or a node in chosen describing which
> > > timer should be used:
> > > 
> > > 	chosen {
> > > 		clockevent {
> > > 			timer = <&timer2>;
> > > 		};
> > > 
> > > 		clocksource {
> > > 			timer = <&timer0>;
> > > 		};
> > > 
> > > 		/*
> > > 		 * or
> > > 		 *
> > > 		 * clockevent = <&timer2>;
> > > 		 * clocksource = <&timer0>;
> > > 		 *
> > > 		 * but I think the clocksource/clockevent node approach
> > > 		 * is more future proof in case we need to add extra
> > > 		 * information like the expected resolution/precision or
> > > 		 * anything that could be tweakable.
> > > 		 */
> > > 	};
> > > 
> > > 	tcb0: timer@fff7c000 {
> > > 		compatible = "atmel,at91rm9200-tcb", "simple-mfd", "syscon";
> > > 		#address-cells = <1>;
> > > 		#size-cells = <0>;
> > > 		reg = <0xfff7c000 0x100>;
> > > 		interrupts = <18 4>;
> > > 		clocks = <&tcb0_clk>, <&clk32k>;
> > > 		clock-names = "t0_clk", "slow_clk";
> > > 
> > > 		timer0: timer@0 {
> > > 			compatible = "atmel,tcb-timer";
> > > 			reg = <0>, <1>;
> > > 		};
> > > 
> > > 		timer2: timer@2 {
> > > 			compatible = "atmel,tcb-timer";
> > > 			reg = <2>;
> > > 		};
> > > 	};  
> > 
> 
> [1]http://elixir.free-electrons.com/linux/latest/source/Documentation/devicetree/bindings/input/gpio-keys.txt

-- 

 <http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs

Follow Linaro:  <http://www.facebook.com/pages/Linaro> Facebook |
<http://twitter.com/#!/linaroorg> Twitter |
<http://www.linaro.org/linaro-blog/> Blog

[toc] | [prev] | [next] | [standalone]

#1660959

On Thu, 8 Jun 2017 10:40:26 +0200
Daniel Lezcano <daniel.lezcano@linaro.org> wrote:

> > > 
> > > Alexandre, Boris, have a look at https://www.spinics.net/lists/arm-kernel/msg572652.html
> > > 
> > > That will tell you the story.  
> > 
> > Then we're in a deadlock situation here. I'm tired of hearing this kind
> > of argument "DT is only supposed to describe HW, not configuration, bla
> > bla". The truth is, we already have plenty of bindings that do not
> > strictly describe HW.
> > 
> > A simple example: ECC configuration on NAND devices. This is clearly a
> > configuration thing, the NAND controller is usually able to support
> > several kind of strength+ECC-block-size config, but we are able to
> > overload this with the nand-ecc-xxx properties. Another example, still
> > MTD related: MTD partitions, this is purely a software configuration,
> > still we allow users to pass this information in the DT. You want
> > another one? What about the linux,code and linux,input-type properties
> > described here [1]?
> > 
> > So please, let's not use these "this is not decribing HW" or "this is
> > linux specific" arguments every time someone tries to encode something
> > that can be considered a configuration detail.
> > 
> > Let's be pragmatic. How you want to use your timer counter blocks (I'm
> > talking about atmel TCBs) is clearly board specific. Whether you want
> > to use the PIT for your clocksource or use one or 2 channels of a TCB
> > at a specific resolution is again board specific. We need a solution to
> > assign timer channels to a linux function, and I'm not convinced
> > passing this information through the command line makes much more sense
> > than specifying it in the DT (and it's definitely less intuitive, since
> > you have to reference something defined in the DT from the cmdline).
> > 
> > Now, in his review, Mark says:
> > 
> > "
> > To me it sounds like what we need is Linux infrastructure that allows
> > one to register a device as having both clockevent/clocksource
> > functionality.
> > 
> > That way, we can choose to do something sane at boot time, and if the
> > user really wants specific devices used in specific ways, they can
> > request that.
> > "
> > 
> > Does that mean that, after adding this "HW timer" infrastructure, we
> > would have a standard way to assign a function to a specific timer
> > block from the DT? How is this different from what I suggest below
> > (note the linux, prefix on my linux,timer-function property, which
> > clearly shows that this is Linux specific)?  
> 
> I like the 'chosen' approach with the nodes you are proposing below. Thanks for
> the constructive suggestion. The binding description matches perfectly what we
> are trying to achieve.

Actually, this is Alexandre who initially suggested the chosen
approach (I thought it was important to mention that ;-)).

[toc] | [prev] | [next] | [standalone]

#1663480 — Re: [PATCH 46/58] clocksource/drivers: Add a new driver for the Atmel ARM TC blocks

Le 08/06/2017 à 10:40, Daniel Lezcano a écrit :
> On Thu, Jun 08, 2017 at 10:13:34AM +0200, Boris Brezillon wrote:
>> On Thu, 8 Jun 2017 09:44:46 +0200
>> Daniel Lezcano <daniel.lezcano@linaro.org> wrote:
>>
>>> +Mark Rutland, +Rob Herring
>>
>> Mark doesn't seem to be CCed.
> 
> Ah, yes. Thanks for fixing this.
> 
>>>
>>>
>>> Alexandre, Boris, have a look at https://www.spinics.net/lists/arm-kernel/msg572652.html
>>>
>>> That will tell you the story.
>>
>> Then we're in a deadlock situation here. I'm tired of hearing this kind
>> of argument "DT is only supposed to describe HW, not configuration, bla
>> bla". The truth is, we already have plenty of bindings that do not
>> strictly describe HW.
>>
>> A simple example: ECC configuration on NAND devices. This is clearly a
>> configuration thing, the NAND controller is usually able to support
>> several kind of strength+ECC-block-size config, but we are able to
>> overload this with the nand-ecc-xxx properties. Another example, still
>> MTD related: MTD partitions, this is purely a software configuration,
>> still we allow users to pass this information in the DT. You want
>> another one? What about the linux,code and linux,input-type properties
>> described here [1]?
>>
>> So please, let's not use these "this is not decribing HW" or "this is
>> linux specific" arguments every time someone tries to encode something
>> that can be considered a configuration detail.
>>
>> Let's be pragmatic. How you want to use your timer counter blocks (I'm
>> talking about atmel TCBs) is clearly board specific. Whether you want
>> to use the PIT for your clocksource or use one or 2 channels of a TCB
>> at a specific resolution is again board specific. We need a solution to
>> assign timer channels to a linux function, and I'm not convinced
>> passing this information through the command line makes much more sense
>> than specifying it in the DT (and it's definitely less intuitive, since
>> you have to reference something defined in the DT from the cmdline).
>>
>> Now, in his review, Mark says:
>>
>> "
>> To me it sounds like what we need is Linux infrastructure that allows
>> one to register a device as having both clockevent/clocksource
>> functionality.
>>
>> That way, we can choose to do something sane at boot time, and if the
>> user really wants specific devices used in specific ways, they can
>> request that.
>> "
>>
>> Does that mean that, after adding this "HW timer" infrastructure, we
>> would have a standard way to assign a function to a specific timer
>> block from the DT? How is this different from what I suggest below
>> (note the linux, prefix on my linux,timer-function property, which
>> clearly shows that this is Linux specific)?
> 
> I like the 'chosen' approach with the nodes you are proposing below. Thanks for
> the constructive suggestion. The binding description matches perfectly what we
> are trying to achieve.
> 
> Rob? what do you think?

I'm following this work from a distance but as we've just celebrated the
1st anniversary for this patch series (11 June 2016), I propose that we
now make up our mind quickly. Everybody seem to be on the same page and
willing to make this rework move forward.

In Microchip/Atmel we would like to actually use this TCB rework both
internally and in our mainline work to avoid having to rely on our own
out-of-tree implementation.

The newly-added samv7 cortex-M can't boot without this series and a use
of our current cortex-A SoCs with TrustZone in Secure World (SWd) is not
possible with current mainline code only. On these two examples, the
current timer on which we rely, the PIT, is not present.

So you probably understand that more than one year without real progress
begins to be a little bit frustrating for the AT91 users...

Regards,


>>> On Thu, Jun 08, 2017 at 07:42:36AM +0200, Boris Brezillon wrote:
>>>> Le Thu, 8 Jun 2017 01:17:15 +0200,
>>>> Alexandre Belloni <alexandre.belloni@free-electrons.com> a écrit :
>>>>   
>>>>> On 07/06/2017 at 23:08:48 +0200, Daniel Lezcano wrote:  
>>>>>>> I was going to agree but this is not flexible enough because the
>>>>>>> quadrature decoder always uses the first two channels. So on some
>>>>>>> products, we may have:
>>>>>>>  - TCB0:
>>>>>>>    o channels 0,1: qdec
>>>>>>>    o channel 2: clocksource
>>>>>>>
>>>>>>>  - TCB1:
>>>>>>>    o channels 0,1: qdec
>>>>>>>    o channel 2: clockevent
>>>>>>>
>>>>>>> This avoids wasting TCB channels.    
>>>>>>
>>>>>> Ok. In this case you can check if the interrupt is specified for the node, if
>>>>>> yes, then it is a clockevent.
>>>>>>     
>>>>>
>>>>> But currently it is always specified in the SoC's dtsi. I don't find
>>>>> that too practical to push that to the board's dts. Also, lying by
>>>>> omission (the IRQ is always wired) in the DT is not different from
>>>>> having a property selecting which timer is the clocksource and which is
>>>>> the clockevent.
>>>>>   
>>>>
>>>> I agree with Alexandre here. Really, there's not much we can do to
>>>> detect which timer should be used as a clockevent and which one should
>>>> be used as a clocksource except explicitly specifying it in the DT.
>>>> Having an interrupt defined in one case (clockevent) and undefined in
>>>> the other case (clocksource), is just as hack-ish as the detection logic
>>>> Alexandre developed to avoid explicitly specifying the function
>>>> assigned to a specific timer.
>>>>
>>>> Can we please find a solution that makes everyone happy (DT,
>>>> clocksoure/clockevent and at91 maintainers)?
>>>>
>>>> How about adding a linux,timer-function property to specify which
>>>> function this timer is providing?
>>>>
>>>> Something like that for example:
>>>>
>>>> 	tcb0: timer@fff7c000 {
>>>> 		compatible = "atmel,at91rm9200-tcb", "simple-mfd", "syscon";
>>>> 		#address-cells = <1>;
>>>> 		#size-cells = <0>;
>>>> 		reg = <0xfff7c000 0x100>;
>>>> 		interrupts = <18 4>;
>>>> 		clocks = <&tcb0_clk>, <&clk32k>;
>>>> 		clock-names = "t0_clk", "slow_clk";
>>>>
>>>> 		timer@0 {
>>>> 			compatible = "atmel,tcb-timer";
>>>> 			reg = <0>, <1>;
>>>> 			linux,timer-function = "clocksource";
>>>> 		};
>>>>
>>>> 		timer@2 {
>>>> 			compatible = "atmel,tcb-timer";
>>>> 			reg = <2>;
>>>> 			linux,timer-function = "clockevent";
>>>> 		};
>>>> 	};
>>>>
>>>> Alternatively, we could have a property or a node in chosen describing which
>>>> timer should be used:
>>>>
>>>> 	chosen {
>>>> 		clockevent {
>>>> 			timer = <&timer2>;
>>>> 		};
>>>>
>>>> 		clocksource {
>>>> 			timer = <&timer0>;
>>>> 		};
>>>>
>>>> 		/*
>>>> 		 * or
>>>> 		 *
>>>> 		 * clockevent = <&timer2>;
>>>> 		 * clocksource = <&timer0>;
>>>> 		 *
>>>> 		 * but I think the clocksource/clockevent node approach
>>>> 		 * is more future proof in case we need to add extra
>>>> 		 * information like the expected resolution/precision or
>>>> 		 * anything that could be tweakable.
>>>> 		 */
>>>> 	};
>>>>
>>>> 	tcb0: timer@fff7c000 {
>>>> 		compatible = "atmel,at91rm9200-tcb", "simple-mfd", "syscon";
>>>> 		#address-cells = <1>;
>>>> 		#size-cells = <0>;
>>>> 		reg = <0xfff7c000 0x100>;
>>>> 		interrupts = <18 4>;
>>>> 		clocks = <&tcb0_clk>, <&clk32k>;
>>>> 		clock-names = "t0_clk", "slow_clk";
>>>>
>>>> 		timer0: timer@0 {
>>>> 			compatible = "atmel,tcb-timer";
>>>> 			reg = <0>, <1>;
>>>> 		};
>>>>
>>>> 		timer2: timer@2 {
>>>> 			compatible = "atmel,tcb-timer";
>>>> 			reg = <2>;
>>>> 		};
>>>> 	};  
>>>
>>
>> [1]http://elixir.free-electrons.com/linux/latest/source/Documentation/devicetree/bindings/input/gpio-keys.txt
> 


-- 
Nicolas Ferre

[toc] | [prev] | [next] | [standalone]

#1663508 — Re: [PATCH 46/58] clocksource/drivers: Add a new driver for the Atmel ARM TC blocks

On 12/06/2017 14:54, Nicolas Ferre wrote:

[ ... ]

>> I like the 'chosen' approach with the nodes you are proposing below. Thanks for
>> the constructive suggestion. The binding description matches perfectly what we
>> are trying to achieve.
>>
>> Rob? what do you think?
> 
> I'm following this work from a distance but as we've just celebrated the
> 1st anniversary for this patch series (11 June 2016), I propose that we
> now make up our mind quickly. Everybody seem to be on the same page and
> willing to make this rework move forward.
> 
> In Microchip/Atmel we would like to actually use this TCB rework both
> internally and in our mainline work to avoid having to rely on our own
> out-of-tree implementation.
> 
> The newly-added samv7 cortex-M can't boot without this series and a use
> of our current cortex-A SoCs with TrustZone in Secure World (SWd) is not
> possible with current mainline code only. On these two examples, the
> current timer on which we rely, the PIT, is not present.
> 
> So you probably understand that more than one year without real progress
> begins to be a little bit frustrating for the AT91 users...

Nicolas,

who are you exactly blaming?

Are you surprised a 58 patches series, with a gazillion of Cc'ed people
posted awhile ago, is ignored?




-- 
 <http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs

Follow Linaro:  <http://www.facebook.com/pages/Linaro> Facebook |
<http://twitter.com/#!/linaroorg> Twitter |
<http://www.linaro.org/linaro-blog/> Blog

[toc] | [prev] | [next] | [standalone]

#1663609 — Re: [PATCH 46/58] clocksource/drivers: Add a new driver for the Atmel ARM TC blocks

Le 12/06/2017 à 15:25, Daniel Lezcano a écrit :
> On 12/06/2017 14:54, Nicolas Ferre wrote:
> 
> [ ... ]
> 
>>> I like the 'chosen' approach with the nodes you are proposing below. Thanks for
>>> the constructive suggestion. The binding description matches perfectly what we
>>> are trying to achieve.
>>>
>>> Rob? what do you think?
>>
>> I'm following this work from a distance but as we've just celebrated the
>> 1st anniversary for this patch series (11 June 2016), I propose that we
>> now make up our mind quickly. Everybody seem to be on the same page and
>> willing to make this rework move forward.
>>
>> In Microchip/Atmel we would like to actually use this TCB rework both
>> internally and in our mainline work to avoid having to rely on our own
>> out-of-tree implementation.
>>
>> The newly-added samv7 cortex-M can't boot without this series and a use
>> of our current cortex-A SoCs with TrustZone in Secure World (SWd) is not
>> possible with current mainline code only. On these two examples, the
>> current timer on which we rely, the PIT, is not present.
>>
>> So you probably understand that more than one year without real progress
>> begins to be a little bit frustrating for the AT91 users...
> 
> Nicolas,
> 
> who are you exactly blaming?
> 
> Are you surprised a 58 patches series, with a gazillion of Cc'ed people
> posted awhile ago, is ignored?

Daniel,

Well, I'm talking about the only patch about DT bindings here, not the
other (large number) of soc/board dts changes.
Note that Alexandre tried to extract this discussion from the other
patches without coming to a conclusion either ("[PATCH v3] ARM: at91:
Document new TCB bindings", for instance).

I had the feeling that this email thread was about to be fading out on a
DT binding discussion, as it had done a couple of times during the last
months... Both you and me produced "ping" messages to make this
discussion progress with no real success so far... This is why I was a
little worried.

Regards,
-- 
Nicolas Ferre

[toc] | [prev] | [next] | [standalone]

#1660785

Le Wed, 7 Jun 2017 23:08:48 +0200,
Daniel Lezcano <daniel.lezcano@linaro.org> a écrit :

> On Wed, Jun 07, 2017 at 05:27:50PM +0200, Alexandre Belloni wrote:
> > On 07/06/2017 at 16:17:35 +0200, Daniel Lezcano wrote:  
> > > > You mean you don't like how it is used? Or you don't think having the
> > > > timer full name is useful?  
> > > 
> > > The field is not needed, the only place where it is used is where we affect it.
> > >    
> > 
> > It is used in tcb_clksrc_register:
> > 
> > tc.clksrc.name = tc.name;  
> 
> Yes, but tc.name is only in the scope of the function, so there is no need to
> include this field in the structure, just a working buffer in the function is
> enough.

Hm, do you mean allocating the buffer dynamically and leaking the
resource (not a real leak here, since we're talking about something
that cannot be unregistered) or defining a 'static char name[]' variable
in the function and passing this pointer to clksrc.name? Note
that putting the name var on the stack won't work, because the core
does not seem to duplicate the name, and uses clksrc->name after
tcb_clksrc_register() has returned.

Anyway, I'm not sure dynamically generating the name is really
needed because we only accept a single tcb-clocksource device. We can
just set tc.clksrc.name to "atmel-tcb-clocksource" and we should be
good.

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web