| Message ID | 20220408081216.57146-1-tony@atomide.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | ARM: dts: Drop unused ti,am4372-timer compatible | expand |
On 08/04/2022 10:12, Tony Lindgren wrote: > Drop unused (and undocumented) ti,am4372-timer related compatible > properties to avoid dtb warnings when updating the binding to yaml. > > The dual-mode timer instances on am43 are compatible with the am3 > timers. Are you sure this is correct approach? Maybe some other user of DTS references the am43 compatible, so instead it should be documented in the bindings? See also recommendation "DO use fallback compatibles...": Documentation/devicetree/bindings/writing-bindings.rst which maybe applies here? Best regards, Krzysztof
* Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> [220408 08:22]: > On 08/04/2022 10:12, Tony Lindgren wrote: > > Drop unused (and undocumented) ti,am4372-timer related compatible > > properties to avoid dtb warnings when updating the binding to yaml. > > > > The dual-mode timer instances on am43 are compatible with the am3 > > timers. > > Are you sure this is correct approach? Maybe some other user of DTS > references the am43 compatible, so instead it should be documented in > the bindings? No users. The driver does not parse these. The am4 SoCs boot using the am3 related compatible properties. > See also recommendation "DO use fallback compatibles...": > Documentation/devicetree/bindings/writing-bindings.rst > which maybe applies here? Yes I'd agree if there were users for this, or if this was already documented :) But I don't see a need to keep these am4 specific properties. Regards, Tony
On 08/04/2022 10:29, Tony Lindgren wrote: > * Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> [220408 08:22]: >> On 08/04/2022 10:12, Tony Lindgren wrote: >>> Drop unused (and undocumented) ti,am4372-timer related compatible >>> properties to avoid dtb warnings when updating the binding to yaml. >>> >>> The dual-mode timer instances on am43 are compatible with the am3 >>> timers. >> >> Are you sure this is correct approach? Maybe some other user of DTS >> references the am43 compatible, so instead it should be documented in >> the bindings? > > No users. You sure you check all operating systems and all out-of-tree users? I did not talk about in-kernel DTS... > The driver does not parse these. The am4 SoCs boot using > the am3 related compatible properties. > >> See also recommendation "DO use fallback compatibles...": >> Documentation/devicetree/bindings/writing-bindings.rst >> which maybe applies here? > > Yes I'd agree if there were users for this, or if this was already > documented :) But I don't see a need to keep these am4 specific > properties. Best regards, Krzysztof
On 08/04/2022 11:33, Krzysztof Kozlowski wrote: > On 08/04/2022 10:29, Tony Lindgren wrote: >> * Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> [220408 08:22]: >>> On 08/04/2022 10:12, Tony Lindgren wrote: >>>> Drop unused (and undocumented) ti,am4372-timer related compatible >>>> properties to avoid dtb warnings when updating the binding to yaml. >>>> >>>> The dual-mode timer instances on am43 are compatible with the am3 >>>> timers. >>> >>> Are you sure this is correct approach? Maybe some other user of DTS >>> references the am43 compatible, so instead it should be documented in >>> the bindings? >> >> No users. > > You sure you check all operating systems and all out-of-tree users? I > did not talk about in-kernel DTS... > >> The driver does not parse these. The am4 SoCs boot using >> the am3 related compatible properties. >> >>> See also recommendation "DO use fallback compatibles...": >>> Documentation/devicetree/bindings/writing-bindings.rst >>> which maybe applies here? >> >> Yes I'd agree if there were users for this, or if this was already >> documented :) But I don't see a need to keep these am4 specific >> properties. > I'd prefer to drop it if not used and not documented. conversation to yaml is good time for clean up. if somebody complains - can always return back and document it.
* Grygorii Strashko <grygorii.strashko@ti.com> [220408 09:31]: > > > On 08/04/2022 11:33, Krzysztof Kozlowski wrote: > > On 08/04/2022 10:29, Tony Lindgren wrote: > > > * Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> [220408 08:22]: > > > > On 08/04/2022 10:12, Tony Lindgren wrote: > > > > > Drop unused (and undocumented) ti,am4372-timer related compatible > > > > > properties to avoid dtb warnings when updating the binding to yaml. > > > > > > > > > > The dual-mode timer instances on am43 are compatible with the am3 > > > > > timers. > > > > > > > > Are you sure this is correct approach? Maybe some other user of DTS > > > > references the am43 compatible, so instead it should be documented in > > > > the bindings? > > > > > > No users. > > > > You sure you check all operating systems and all out-of-tree users? I > > did not talk about in-kernel DTS... Yeah you can never be quite sure :) > > > The driver does not parse these. The am4 SoCs boot using > > > the am3 related compatible properties. > > > > > > > See also recommendation "DO use fallback compatibles...": > > > > Documentation/devicetree/bindings/writing-bindings.rst > > > > which maybe applies here? > > > > > > Yes I'd agree if there were users for this, or if this was already > > > documented :) But I don't see a need to keep these am4 specific > > > properties. > > > > I'd prefer to drop it if not used and not documented. > conversation to yaml is good time for clean up. > if somebody complains - can always return back and document it. Well this will produce warnings for Krzysztof for the dt bindings branch until it's dropped for the dts files. And since it's possible to discuss this with arguments both ways.. Let's just keep it and document it. I just noticed we already have few redundant compatibles with dm814 and dm816 that are probably compatible with omap4 if rememember correctly. Regards, Tony
diff --git a/arch/arm/boot/dts/am437x-l4.dtsi b/arch/arm/boot/dts/am437x-l4.dtsi --- a/arch/arm/boot/dts/am437x-l4.dtsi +++ b/arch/arm/boot/dts/am437x-l4.dtsi @@ -361,7 +361,7 @@ SYSC_OMAP2_SOFTRESET | ranges = <0x0 0x31000 0x1000>; timer1: timer@0 { - compatible = "ti,am4372-timer-1ms","ti,am335x-timer-1ms"; + compatible = "ti,am335x-timer-1ms"; reg = <0x0 0x400>; interrupts = <GIC_SPI 67 IRQ_TYPE_LEVEL_HIGH>; ti,timer-alwon; @@ -906,7 +906,7 @@ timer2_target: target-module@40000 { /* 0x48040000, ap 18 1e.0 */ ranges = <0x0 0x40000 0x1000>; timer2: timer@0 { - compatible = "ti,am4372-timer","ti,am335x-timer"; + compatible = "ti,am335x-timer"; reg = <0x0 0x400>; interrupts = <GIC_SPI 68 IRQ_TYPE_LEVEL_HIGH>; clocks = <&timer2_fck>; @@ -933,7 +933,7 @@ target-module@42000 { /* 0x48042000, ap 20 24.0 */ ranges = <0x0 0x42000 0x1000>; timer3: timer@0 { - compatible = "ti,am4372-timer","ti,am335x-timer"; + compatible = "ti,am335x-timer"; reg = <0x0 0x400>; interrupts = <GIC_SPI 69 IRQ_TYPE_LEVEL_HIGH>; status = "disabled"; @@ -959,7 +959,7 @@ target-module@44000 { /* 0x48044000, ap 22 26.0 */ ranges = <0x0 0x44000 0x1000>; timer4: timer@0 { - compatible = "ti,am4372-timer","ti,am335x-timer"; + compatible = "ti,am335x-timer"; reg = <0x0 0x400>; interrupts = <GIC_SPI 92 IRQ_TYPE_LEVEL_HIGH>; ti,timer-pwm; @@ -986,7 +986,7 @@ target-module@46000 { /* 0x48046000, ap 24 28.0 */ ranges = <0x0 0x46000 0x1000>; timer5: timer@0 { - compatible = "ti,am4372-timer","ti,am335x-timer"; + compatible = "ti,am335x-timer"; reg = <0x0 0x400>; interrupts = <GIC_SPI 93 IRQ_TYPE_LEVEL_HIGH>; ti,timer-pwm; @@ -1013,7 +1013,7 @@ target-module@48000 { /* 0x48048000, ap 26 1a.0 */ ranges = <0x0 0x48000 0x1000>; timer6: timer@0 { - compatible = "ti,am4372-timer","ti,am335x-timer"; + compatible = "ti,am335x-timer"; reg = <0x0 0x400>; interrupts = <GIC_SPI 94 IRQ_TYPE_LEVEL_HIGH>; ti,timer-pwm; @@ -1040,7 +1040,7 @@ target-module@4a000 { /* 0x4804a000, ap 71 48.0 */ ranges = <0x0 0x4a000 0x1000>; timer7: timer@0 { - compatible = "ti,am4372-timer","ti,am335x-timer"; + compatible = "ti,am335x-timer"; reg = <0x0 0x400>; interrupts = <GIC_SPI 95 IRQ_TYPE_LEVEL_HIGH>; ti,timer-pwm; @@ -1549,7 +1549,7 @@ target-module@c1000 { /* 0x481c1000, ap 94 68.0 */ ranges = <0x0 0xc1000 0x1000>; timer8: timer@0 { - compatible = "ti,am4372-timer","ti,am335x-timer"; + compatible = "ti,am335x-timer"; reg = <0x0 0x400>; interrupts = <GIC_SPI 131 IRQ_TYPE_LEVEL_HIGH>; status = "disabled"; @@ -2263,7 +2263,7 @@ target-module@3d000 { /* 0x4833d000, ap 102 6e.0 */ ranges = <0x0 0x3d000 0x1000>; timer9: timer@0 { - compatible = "ti,am4372-timer","ti,am335x-timer"; + compatible = "ti,am335x-timer"; reg = <0x0 0x400>; interrupts = <GIC_SPI 132 IRQ_TYPE_LEVEL_HIGH>; status = "disabled"; @@ -2289,7 +2289,7 @@ target-module@3f000 { /* 0x4833f000, ap 104 5c.0 */ ranges = <0x0 0x3f000 0x1000>; timer10: timer@0 { - compatible = "ti,am4372-timer","ti,am335x-timer"; + compatible = "ti,am335x-timer"; reg = <0x0 0x400>; interrupts = <GIC_SPI 133 IRQ_TYPE_LEVEL_HIGH>; status = "disabled"; @@ -2315,7 +2315,7 @@ target-module@41000 { /* 0x48341000, ap 106 76.0 */ ranges = <0x0 0x41000 0x1000>; timer11: timer@0 { - compatible = "ti,am4372-timer","ti,am335x-timer"; + compatible = "ti,am335x-timer"; reg = <0x0 0x400>; interrupts = <GIC_SPI 134 IRQ_TYPE_LEVEL_HIGH>; status = "disabled";
Drop unused (and undocumented) ti,am4372-timer related compatible properties to avoid dtb warnings when updating the binding to yaml. The dual-mode timer instances on am43 are compatible with the am3 timers. Cc: Daniel Lezcano <daniel.lezcano@linaro.org> Cc: Keerthy <j-keerthy@ti.com> Cc: Krzysztof Kozlowski <krzk+dt@kernel.org> Cc: Nishanth Menon <nm@ti.com> Cc: Rob Herring <robh+dt@kernel.org> Cc: Vignesh Raghavendra <vigneshr@ti.com> Signed-off-by: Tony Lindgren <tony@atomide.com> --- arch/arm/boot/dts/am437x-l4.dtsi | 22 +++++++++++----------- 1 file changed, 11 insertions(+), 11 deletions(-)