linux-kernel - Re: [PATCH 2/6] dt-bindings: rtc: stm32: add alarm A out property to select output

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:   Fri, 22 Jul 2022 18:02:51 +0200
From:   Alexandre Belloni <alexandre.belloni@...tlin.com>
To:     Valentin CARON <valentin.caron@...s.st.com>
Cc:     Alessandro Zummo <a.zummo@...ertech.it>,
        Rob Herring <robh+dt@...nel.org>,
        Krzysztof Kozlowski <krzk+dt@...nel.org>,
        Alexandre Torgue <alexandre.torgue@...s.st.com>,
        Gabriel Fernandez <gabriel.fernandez@...s.st.com>,
        Amelie Delaunay <amelie.delaunay@...s.st.com>,
        linux-rtc@...r.kernel.org, devicetree@...r.kernel.org,
        linux-stm32@...md-mailman.stormreply.com,
        linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 2/6] dt-bindings: rtc: stm32: add alarm A out property to
 select output

On 23/05/2022 14:34:22+0200, Valentin CARON wrote:
> Hi Alexandre,
> 
> On 5/4/22 22:27, Alexandre Belloni wrote:
> > Hello,
> > 
> > On 04/05/2022 15:06:13+0200, Valentin Caron wrote:
> > > STM32 RTC can pulse some SOC pins when an alarm of RTC expires.
> > > 
> > > This patch adds property to activate alarm A output. The pulse can
> > > output on three pins RTC_OUT1, RTC_OUT2, RTC_OUT2_RMP
> > > (PC13, PB2, PI8 on stm32mp15) (PC13, PB2, PI1 on stm32mp13).
> > > 
> > > Signed-off-by: Valentin Caron <valentin.caron@...s.st.com>
> > > ---
> > >   .../devicetree/bindings/rtc/st,stm32-rtc.yaml | 19 ++++++++++++++++++-
> > >   1 file changed, 18 insertions(+), 1 deletion(-)
> > > 
> > > diff --git a/Documentation/devicetree/bindings/rtc/st,stm32-rtc.yaml b/Documentation/devicetree/bindings/rtc/st,stm32-rtc.yaml
> > > index 56d46ea35c5d..71e02604e8de 100644
> > > --- a/Documentation/devicetree/bindings/rtc/st,stm32-rtc.yaml
> > > +++ b/Documentation/devicetree/bindings/rtc/st,stm32-rtc.yaml
> > > @@ -59,6 +59,13 @@ properties:
> > >         Refer to <include/dt-bindings/rtc/rtc-stm32.h> for the supported values.
> > >         Pinctrl state named "default" may be defined to reserve pin for RTC output.
> > > +  st,alarm:
> > > +    $ref: "/schemas/types.yaml#/definitions/uint32"
> > > +    description: |
> > > +      To select and enable RTC Alarm A output.
> > > +      Refer to <include/dt-bindings/rtc/rtc-stm32.h> for the supported values.
> > > +      Pinctrl state named "default" may be defined to reserve pin for RTC output.
> > > +
> > >   allOf:
> > >     - if:
> > >         properties:
> > > @@ -75,6 +82,9 @@ allOf:
> > >           st,lsco:
> > >             maxItems: 0
> > > +        st,alarm:
> > > +          maxItems: 0
> > > +
> > >           clock-names: false
> > >         required:
> > > @@ -95,6 +105,9 @@ allOf:
> > >           st,lsco:
> > >             maxItems: 0
> > > +        st,alarm:
> > > +          maxItems: 0
> > > +
> > >         required:
> > >           - clock-names
> > >           - st,syscfg
> > > @@ -117,6 +130,9 @@ allOf:
> > >           st,lsco:
> > >             maxItems: 1
> > > +        st,alarm:
> > > +          maxItems: 1
> > > +
> > >         required:
> > >           - clock-names
> > > @@ -153,8 +169,9 @@ examples:
> > >         clocks = <&rcc RTCAPB>, <&rcc RTC>;
> > >         clock-names = "pclk", "rtc_ck";
> > >         interrupts = <GIC_SPI 3 IRQ_TYPE_LEVEL_HIGH>;
> > > +      st,alarm = <RTC_OUT1>;
> > >         st,lsco = <RTC_OUT2_RMP>;
> > Shouldn't that be exactly the opposite? You have two pins that can
> > output different functions. The property should be the pin and the value
> > the function. I'd go even further and I would say this is actually
> > pinmuxing.
> > 
> You're right, if the property is the pin and the value the function, this
> looks like a pinctrl node.
> We choose to develop theses functionalities in the reverse order, to avoid
> the complexity of adding
> the pinctrl framework to our driver. Moreover, LSCO and AlarmA may haven't a
> peripheral client and
> this would probably require to also implement pinctrl hogging.
> 
> Is the implementation that we have proposed is acceptable regarding theses
> elements ?
> 


I still think that the pin has to be the property and the function the value.

Or we could find a generic name and provide an array of pin, function
pair

Or, go for pinmuxing

My point here is that this is a common feature an RTCs and I don't want
every vendor to come up with their own properties.

Regards,

-- 
Alexandre Belloni, co-owner and COO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com