| Message ID | 20241211060936.57452-1-naoki@radxa.com (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | dt-bindings: arm: rockchip: add Radxa ROCK 5C Lite | expand |
Hello Fukaumi, On 2024-12-11 07:09, FUKAUMI Naoki wrote: > The Radxa ROCK 5C Lite uses a different SoC (RK3582) compared to the > Radxa ROCK 5C (RK3588S2), but the two are compatible from a software > perspective. > > Fixes: df4e08a5eed1 ("dt-bindings: arm: rockchip: add Radxa ROCK 5C") > Signed-off-by: FUKAUMI Naoki <naoki@radxa.com> > --- > Documentation/devicetree/bindings/arm/rockchip.yaml | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/Documentation/devicetree/bindings/arm/rockchip.yaml > b/Documentation/devicetree/bindings/arm/rockchip.yaml > index 753199a12923..2254ee079094 100644 > --- a/Documentation/devicetree/bindings/arm/rockchip.yaml > +++ b/Documentation/devicetree/bindings/arm/rockchip.yaml > @@ -895,7 +895,7 @@ properties: > - const: radxa,rock-5b > - const: rockchip,rk3588 > > - - description: Radxa ROCK 5C > + - description: Radxa ROCK 5C/5C Lite > items: > - const: radxa,rock-5c > - const: rockchip,rk3588s I think it would be better to use "rockchip,rk3582" here, to allow us to possibly use that information later. For example, we might want to be able to recognize RK3582-based boards in U-Boot without the need to look into the e-fuses at some point, for which purpose having a clear designator in the DT would fit perfectly. As a reminder, using "rockchip,rk3582" would also require a small addition to drivers/irqchip/irq-gic-v3-its.c.
Hi Dragan, On 12/11/24 15:36, Dragan Simic wrote: > Hello Fukaumi, > > On 2024-12-11 07:09, FUKAUMI Naoki wrote: >> The Radxa ROCK 5C Lite uses a different SoC (RK3582) compared to the >> Radxa ROCK 5C (RK3588S2), but the two are compatible from a software >> perspective. >> >> Fixes: df4e08a5eed1 ("dt-bindings: arm: rockchip: add Radxa ROCK 5C") >> Signed-off-by: FUKAUMI Naoki <naoki@radxa.com> >> --- >> Documentation/devicetree/bindings/arm/rockchip.yaml | 2 +- >> 1 file changed, 1 insertion(+), 1 deletion(-) >> >> diff --git a/Documentation/devicetree/bindings/arm/rockchip.yaml >> b/Documentation/devicetree/bindings/arm/rockchip.yaml >> index 753199a12923..2254ee079094 100644 >> --- a/Documentation/devicetree/bindings/arm/rockchip.yaml >> +++ b/Documentation/devicetree/bindings/arm/rockchip.yaml >> @@ -895,7 +895,7 @@ properties: >> - const: radxa,rock-5b >> - const: rockchip,rk3588 >> >> - - description: Radxa ROCK 5C >> + - description: Radxa ROCK 5C/5C Lite >> items: >> - const: radxa,rock-5c >> - const: rockchip,rk3588s > > I think it would be better to use "rockchip,rk3582" here, to allow > us to possibly use that information later. For example, we might > want to be able to recognize RK3582-based boards in U-Boot without > the need to look into the e-fuses at some point, for which purpose > having a clear designator in the DT would fit perfectly. It may be okay to introduce "rockchip,rk3582", but reading e-fuse is still required in U-Boot because which unit (cpu coreX, gpu, etc) is broken cannot be determined without reading e-fuse at run-time. Best regards, -- FUKAUMI Naoki Radxa Computer (Shenzhen) Co., Ltd. > As a reminder, using "rockchip,rk3582" would also require a small > addition to drivers/irqchip/irq-gic-v3-its.c. >
just FYI, https://lore.kernel.org/linux-rockchip/?q=e52c -- FUKAUMI Naoki Radxa Computer (Shenzhen) Co., Ltd. On 12/11/24 15:44, FUKAUMI Naoki wrote: > Hi Dragan, > > On 12/11/24 15:36, Dragan Simic wrote: >> Hello Fukaumi, >> >> On 2024-12-11 07:09, FUKAUMI Naoki wrote: >>> The Radxa ROCK 5C Lite uses a different SoC (RK3582) compared to the >>> Radxa ROCK 5C (RK3588S2), but the two are compatible from a software >>> perspective. >>> >>> Fixes: df4e08a5eed1 ("dt-bindings: arm: rockchip: add Radxa ROCK 5C") >>> Signed-off-by: FUKAUMI Naoki <naoki@radxa.com> >>> --- >>> Documentation/devicetree/bindings/arm/rockchip.yaml | 2 +- >>> 1 file changed, 1 insertion(+), 1 deletion(-) >>> >>> diff --git a/Documentation/devicetree/bindings/arm/rockchip.yaml >>> b/Documentation/devicetree/bindings/arm/rockchip.yaml >>> index 753199a12923..2254ee079094 100644 >>> --- a/Documentation/devicetree/bindings/arm/rockchip.yaml >>> +++ b/Documentation/devicetree/bindings/arm/rockchip.yaml >>> @@ -895,7 +895,7 @@ properties: >>> - const: radxa,rock-5b >>> - const: rockchip,rk3588 >>> >>> - - description: Radxa ROCK 5C >>> + - description: Radxa ROCK 5C/5C Lite >>> items: >>> - const: radxa,rock-5c >>> - const: rockchip,rk3588s >> >> I think it would be better to use "rockchip,rk3582" here, to allow >> us to possibly use that information later. For example, we might >> want to be able to recognize RK3582-based boards in U-Boot without >> the need to look into the e-fuses at some point, for which purpose >> having a clear designator in the DT would fit perfectly. > > It may be okay to introduce "rockchip,rk3582", but reading e-fuse is > still required in U-Boot because which unit (cpu coreX, gpu, etc) is > broken cannot be determined without reading e-fuse at run-time. > > Best regards, > > -- > FUKAUMI Naoki > Radxa Computer (Shenzhen) Co., Ltd. > >> As a reminder, using "rockchip,rk3582" would also require a small >> addition to drivers/irqchip/irq-gic-v3-its.c. >> > >
On 2024-12-11 07:44, FUKAUMI Naoki wrote: > On 12/11/24 15:36, Dragan Simic wrote: >> On 2024-12-11 07:09, FUKAUMI Naoki wrote: >>> The Radxa ROCK 5C Lite uses a different SoC (RK3582) compared to the >>> Radxa ROCK 5C (RK3588S2), but the two are compatible from a software >>> perspective. >>> >>> Fixes: df4e08a5eed1 ("dt-bindings: arm: rockchip: add Radxa ROCK 5C") >>> Signed-off-by: FUKAUMI Naoki <naoki@radxa.com> >>> --- >>> Documentation/devicetree/bindings/arm/rockchip.yaml | 2 +- >>> 1 file changed, 1 insertion(+), 1 deletion(-) >>> >>> diff --git a/Documentation/devicetree/bindings/arm/rockchip.yaml >>> b/Documentation/devicetree/bindings/arm/rockchip.yaml >>> index 753199a12923..2254ee079094 100644 >>> --- a/Documentation/devicetree/bindings/arm/rockchip.yaml >>> +++ b/Documentation/devicetree/bindings/arm/rockchip.yaml >>> @@ -895,7 +895,7 @@ properties: >>> - const: radxa,rock-5b >>> - const: rockchip,rk3588 >>> >>> - - description: Radxa ROCK 5C >>> + - description: Radxa ROCK 5C/5C Lite >>> items: >>> - const: radxa,rock-5c >>> - const: rockchip,rk3588s >> >> I think it would be better to use "rockchip,rk3582" here, to allow >> us to possibly use that information later. For example, we might >> want to be able to recognize RK3582-based boards in U-Boot without >> the need to look into the e-fuses at some point, for which purpose >> having a clear designator in the DT would fit perfectly. > > It may be okay to introduce "rockchip,rk3582", but reading e-fuse is > still required in U-Boot because which unit (cpu coreX, gpu, etc) is > broken cannot be determined without reading e-fuse at run-time. Sure, but I think it would be rather bad to miss this opportunity to introduce a clear DT designator for RK3582-based boards. It's better to have the designator unused, than to bang our heads later, if we conclude that we need it at some point. :) >> As a reminder, using "rockchip,rk3582" would also require a small >> addition to drivers/irqchip/irq-gic-v3-its.c.
On 2024-12-11 08:02, FUKAUMI Naoki wrote: > just FYI, > > https://lore.kernel.org/linux-rockchip/?q=e52c AFAICT, it isn't too late to introduce "rockchip,rk3582". :) > On 12/11/24 15:44, FUKAUMI Naoki wrote: >> On 12/11/24 15:36, Dragan Simic wrote: >>> On 2024-12-11 07:09, FUKAUMI Naoki wrote: >>>> The Radxa ROCK 5C Lite uses a different SoC (RK3582) compared to the >>>> Radxa ROCK 5C (RK3588S2), but the two are compatible from a software >>>> perspective. >>>> >>>> Fixes: df4e08a5eed1 ("dt-bindings: arm: rockchip: add Radxa ROCK >>>> 5C") >>>> Signed-off-by: FUKAUMI Naoki <naoki@radxa.com> >>>> --- >>>> Documentation/devicetree/bindings/arm/rockchip.yaml | 2 +- >>>> 1 file changed, 1 insertion(+), 1 deletion(-) >>>> >>>> diff --git a/Documentation/devicetree/bindings/arm/rockchip.yaml >>>> b/Documentation/devicetree/bindings/arm/rockchip.yaml >>>> index 753199a12923..2254ee079094 100644 >>>> --- a/Documentation/devicetree/bindings/arm/rockchip.yaml >>>> +++ b/Documentation/devicetree/bindings/arm/rockchip.yaml >>>> @@ -895,7 +895,7 @@ properties: >>>> - const: radxa,rock-5b >>>> - const: rockchip,rk3588 >>>> >>>> - - description: Radxa ROCK 5C >>>> + - description: Radxa ROCK 5C/5C Lite >>>> items: >>>> - const: radxa,rock-5c >>>> - const: rockchip,rk3588s >>> >>> I think it would be better to use "rockchip,rk3582" here, to allow >>> us to possibly use that information later. For example, we might >>> want to be able to recognize RK3582-based boards in U-Boot without >>> the need to look into the e-fuses at some point, for which purpose >>> having a clear designator in the DT would fit perfectly. >> >> It may be okay to introduce "rockchip,rk3582", but reading e-fuse >> is still required in U-Boot because which unit (cpu coreX, gpu, etc) >> is broken cannot be determined without reading e-fuse at run-time. >> >>> As a reminder, using "rockchip,rk3582" would also require a small >>> addition to drivers/irqchip/irq-gic-v3-its.c.
diff --git a/Documentation/devicetree/bindings/arm/rockchip.yaml b/Documentation/devicetree/bindings/arm/rockchip.yaml index 753199a12923..2254ee079094 100644 --- a/Documentation/devicetree/bindings/arm/rockchip.yaml +++ b/Documentation/devicetree/bindings/arm/rockchip.yaml @@ -895,7 +895,7 @@ properties: - const: radxa,rock-5b - const: rockchip,rk3588 - - description: Radxa ROCK 5C + - description: Radxa ROCK 5C/5C Lite items: - const: radxa,rock-5c - const: rockchip,rk3588s
The Radxa ROCK 5C Lite uses a different SoC (RK3582) compared to the Radxa ROCK 5C (RK3588S2), but the two are compatible from a software perspective. Fixes: df4e08a5eed1 ("dt-bindings: arm: rockchip: add Radxa ROCK 5C") Signed-off-by: FUKAUMI Naoki <naoki@radxa.com> --- Documentation/devicetree/bindings/arm/rockchip.yaml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)