Re: [PATCH v1 2/2] arm: dts: mt2701: add nor flash node

[Boris Brezillon]

On Sun, 15 Jan 2017 01:23:48 +0100
Marek Vasut <marek.vasut@xxxxxxxxx> wrote:

> On 01/14/2017 09:29 AM, Boris Brezillon wrote:
> > On Fri, 13 Jan 2017 18:33:40 +0100
> > Marek Vasut <marek.vasut@xxxxxxxxx> wrote:
> >   
> >> On 01/13/2017 05:56 PM, Boris Brezillon wrote:  
> >>> On Fri, 13 Jan 2017 17:44:12 +0100
> >>> Marek Vasut <marek.vasut@xxxxxxxxx> wrote:
> >>>     
> >>>> On 01/13/2017 05:28 PM, Boris Brezillon wrote:    
> >>>>> On Fri, 13 Jan 2017 17:13:55 +0100
> >>>>> Marek Vasut <marek.vasut@xxxxxxxxx> wrote:
> >>>>>       
> >>>>>> On 01/13/2017 04:12 PM, Matthias Brugger wrote:      
> >>>>>>>
> >>>>>>>
> >>>>>>> On 13/01/17 15:17, Boris Brezillon wrote:        
> >>>>>>>> On Fri, 13 Jan 2017 15:13:29 +0800
> >>>>>>>> Guochun Mao <guochun.mao@xxxxxxxxxxxx> wrote:
> >>>>>>>>        
> >>>>>>>>> Add Mediatek nor flash node.
> >>>>>>>>>
> >>>>>>>>> Signed-off-by: Guochun Mao <guochun.mao@xxxxxxxxxxxx>
> >>>>>>>>> ---
> >>>>>>>>>  arch/arm/boot/dts/mt2701-evb.dts |   25 +++++++++++++++++++++++++
> >>>>>>>>>  arch/arm/boot/dts/mt2701.dtsi    |   12 ++++++++++++
> >>>>>>>>>  2 files changed, 37 insertions(+)
> >>>>>>>>>
> >>>>>>>>> diff --git a/arch/arm/boot/dts/mt2701-evb.dts
> >>>>>>>>> b/arch/arm/boot/dts/mt2701-evb.dts
> >>>>>>>>> index 082ca88..85e5ae8 100644
> >>>>>>>>> --- a/arch/arm/boot/dts/mt2701-evb.dts
> >>>>>>>>> +++ b/arch/arm/boot/dts/mt2701-evb.dts
> >>>>>>>>> @@ -24,6 +24,31 @@
> >>>>>>>>>      };
> >>>>>>>>>  };
> >>>>>>>>>
> >>>>>>>>> +&nor_flash {
> >>>>>>>>> +    pinctrl-names = "default";
> >>>>>>>>> +    pinctrl-0 = <&nor_pins_default>;
> >>>>>>>>> +    status = "okay";
> >>>>>>>>> +    flash@0 {
> >>>>>>>>> +        compatible = "jedec,spi-nor";
> >>>>>>>>> +        reg = <0>;
> >>>>>>>>> +    };
> >>>>>>>>> +};
> >>>>>>>>> +
> >>>>>>>>> +&pio {
> >>>>>>>>> +    nor_pins_default: nor {
> >>>>>>>>> +        pins1 {
> >>>>>>>>> +            pinmux = <MT2701_PIN_240_EXT_XCS__FUNC_EXT_XCS>,
> >>>>>>>>> +                 <MT2701_PIN_241_EXT_SCK__FUNC_EXT_SCK>,
> >>>>>>>>> +                 <MT2701_PIN_239_EXT_SDIO0__FUNC_EXT_SDIO0>,
> >>>>>>>>> +                 <MT2701_PIN_238_EXT_SDIO1__FUNC_EXT_SDIO1>,
> >>>>>>>>> +                 <MT2701_PIN_237_EXT_SDIO2__FUNC_EXT_SDIO2>,
> >>>>>>>>> +                 <MT2701_PIN_236_EXT_SDIO3__FUNC_EXT_SDIO3>;
> >>>>>>>>> +            drive-strength = <MTK_DRIVE_4mA>;
> >>>>>>>>> +            bias-pull-up;
> >>>>>>>>> +        };
> >>>>>>>>> +    };
> >>>>>>>>> +};
> >>>>>>>>> +
> >>>>>>>>>  &uart0 {
> >>>>>>>>>      status = "okay";
> >>>>>>>>>  };
> >>>>>>>>> diff --git a/arch/arm/boot/dts/mt2701.dtsi
> >>>>>>>>> b/arch/arm/boot/dts/mt2701.dtsi
> >>>>>>>>> index bdf8954..1eefce4 100644
> >>>>>>>>> --- a/arch/arm/boot/dts/mt2701.dtsi
> >>>>>>>>> +++ b/arch/arm/boot/dts/mt2701.dtsi
> >>>>>>>>> @@ -227,6 +227,18 @@
> >>>>>>>>>          status = "disabled";
> >>>>>>>>>      };
> >>>>>>>>>
> >>>>>>>>> +    nor_flash: spi@11014000 {
> >>>>>>>>> +        compatible = "mediatek,mt2701-nor",
> >>>>>>>>> +                 "mediatek,mt8173-nor";        
> >>>>>>>>
> >>>>>>>> Why define both here? Is "mediatek,mt8173-nor" really providing a
> >>>>>>>> subset of the features supported by "mediatek,mt2701-nor"?
> >>>>>>>>        
> >>>>>>>
> >>>>>>> I think even if the ip block is the same, we should provide both
> >>>>>>> bindings, just in case in the future we find out that mt2701 has some
> >>>>>>> hidden bug, feature or bug-feature. This way even if we update the
> >>>>>>> driver, we stay compatible with older device tree blobs in the wild.
> >>>>>>>
> >>>>>>> We can drop the mt2701-nor in the bindings definition if you want.       
> >>>>>
> >>>>> Oh, sorry, I misunderstood. What I meant is that if you want to
> >>>>> list/support all possible compatibles, maybe you should just put one
> >>>>> compatible in your DT and patch your driver (+ binding doc) to define
> >>>>> all of them.      
> >>>>
> >>>> Uh, what ? I lost you here :-)  
> > 
> > I mean adding a new entry in the mtk_nor_of_ids table (in
> > mtk-quadspi.c) so that the mediatek,mt2701-nor compatible string can be
> > matched directly, and you won't need to define 2 compatible strings in
> > your device tree.  
> 
> But then you grow the table in the driver, is that what we want if we
> can avoid that ?

The space you save by not growing the mtk_nor_of_ids table is lost in
your dtbs, so I'm not sure the size argument is relevant here. Also,
note that distros are shipping a lot of dtbs, and you're likely to have
several boards based on the mt2701 SoC, so, for this specific use case,
it's better to make the in-driver of-id table grow than specifying 2
compatibles in the DT. But as I said, I'm not sure we should rely on
this argument to decide which approach to choose (we're talking about a
few bytes here).

> 
> >>>>    
> >>>>>> This exactly. We should have a DT compat in the form:
> >>>>>> compatible = "vendor,<soc>-block", "vendor,<oldest-compat-soc>-block";
> >>>>>> Then if we find a problem in the future, we can match on the
> >>>>>> "vendor,<soc>-block" and still support the old DTs.      
> >>>>>
> >>>>> Not sure it's only in term of whose IP appeared first. My understanding
> >>>>> is that it's a way to provide inheritance. For example:
> >>>>>
> >>>>> 	"<soc-vendor>,<ip-version>", "<ip-vendor>,<ip-version>";
> >>>>>
> >>>>> or
> >>>>>
> >>>>> 	"<soc-vendor>,<full-featured-ip-version>","<soc-vendor>,<basic-feature-ip-version>";
> >>>>>
> >>>>> BTW, which one is the oldest between mt8173 and mt2701? :-)      
> >>>>
> >>>> And that's another thing and I agree with you, but I don't think that's
> >>>> what we're discussing in this thread. But (!), OT, I think we should
> >>>> codify the rules in Documentation/ . This discussion came up multiple
> >>>> times recently.
> >>>>
> >>>> And my question still stands, what do we put into the DT here, IMO
> >>>> compatible = "mediatek,mt2701-nor", "mediatek,mt8173-nor";    
> >>>
> >>> I'd say
> >>>
> >>> 	compatible = "mediatek,mt8173-nor";
> >>>
> >>> because both compatible are referring to very specific IP version. It's
> >>> not the same as    
> >>
> >> But then you don't have the ability to handle a block in this particular
> >> SoC in case there's a bug found in it in the future,
> >> so IMO it should be:
> >>
> >> compatible = "mediatek,mt2701-nor", "mediatek,mt8173-nor";  
> > 
> > Sorry again, I meant
> > 
> > 	compatible = "mediatek,mt2701-nor";
> >   
> >>  
> >>> 	compatible = "mediatek,mt8173-nor", "mediatek,mt81xx-nor";    
> >>
> >> This doesn't look right, since here we add two new compatibles ...  
> > 
> > That was just an example to describe how compatible inheritance works
> > (at least that's my understanding of it), it does not apply to this
> > particular use case.  
> 
> Well this is OK I guess, but then you can also use "mediatek,mt8173-nor"
> as the oldest supported compatible and be done with it, no ? It looks a
> bit crappy though, I admit that ...
> 

Let's stop bikeshedding and wait for DT maintainers feedback
before taking a decision ;-).

Rob, Mark, any opinion?