* Re: [PATCH] arm64: dts: Add/fix /memory node unit-addresses
2024-04-30 19:18 [PATCH] arm64: dts: Add/fix /memory node unit-addresses Rob Herring (Arm)
@ 2024-04-30 21:49 ` Florian Fainelli
2024-05-01 10:31 ` Heiko Stübner
` (4 subsequent siblings)
5 siblings, 0 replies; 8+ messages in thread
From: Florian Fainelli @ 2024-04-30 21:49 UTC (permalink / raw)
To: Rob Herring (Arm), soc, Andreas Färber,
Manivannan Sadhasivam, Krzysztof Kozlowski, Conor Dooley,
Khuong Dinh, Ray Jui, Scott Branden,
Broadcom internal kernel review list, Robert Richter, Chanho Min,
Avi Fishman, Tomer Maimon, Tali Perry, Patrick Venture,
Nancy Yuen, Benjamin Fair, Thierry Reding, Jonathan Hunter,
Bjorn Andersson, Konrad Dybcio, Heiko Stuebner, Orson Zhai,
Baolin Wang, Chunyan Zhang
Cc: linux-arm-kernel, linux-actions, devicetree, linux-kernel,
openbmc, linux-tegra, linux-arm-msm, linux-rockchip
On 4/30/24 12:18, Rob Herring (Arm) wrote:
> '/memory' nodes always have a 'reg' property, and therefore should have
> a unit-address with just plain hex (i.e. no commas). Fix all the arm64
> '/memory' nodes.
>
> It's possible that some bootloader depends on /memory (arm32 ATAG to DT
> code does for example). If so, the memory node should be commented with
> that requirement.
>
> Signed-off-by: Rob Herring (Arm) <robh@kernel.org>
> ---
> SoC maintainers, please take this directly.
>
> arm32 is much worse, but the platforms using ATAG to DT code can't be
> changed and I don't know which ones those are.
>
> arch/arm64/boot/dts/actions/s700-cubieboard7.dts | 2 +-
> arch/arm64/boot/dts/apm/apm-merlin.dts | 2 +-
> arch/arm64/boot/dts/apm/apm-mustang.dts | 2 +-
> arch/arm64/boot/dts/broadcom/northstar2/ns2-svk.dts | 2 +-
> arch/arm64/boot/dts/broadcom/northstar2/ns2-xmc.dts | 2 +-
For the above two:
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
--
Florian
^ permalink raw reply [flat|nested] 8+ messages in thread* Re: [PATCH] arm64: dts: Add/fix /memory node unit-addresses
2024-04-30 19:18 [PATCH] arm64: dts: Add/fix /memory node unit-addresses Rob Herring (Arm)
2024-04-30 21:49 ` Florian Fainelli
@ 2024-05-01 10:31 ` Heiko Stübner
2024-05-01 10:44 ` Bryan O'Donoghue
` (3 subsequent siblings)
5 siblings, 0 replies; 8+ messages in thread
From: Heiko Stübner @ 2024-05-01 10:31 UTC (permalink / raw)
To: soc, Andreas Färber, Manivannan Sadhasivam,
Krzysztof Kozlowski, Conor Dooley, Khuong Dinh, Ray Jui,
Scott Branden, Broadcom internal kernel review list,
Robert Richter, Chanho Min, Avi Fishman, Tomer Maimon, Tali Perry,
Patrick Venture, Nancy Yuen, Benjamin Fair, Thierry Reding,
Jonathan Hunter, Bjorn Andersson, Konrad Dybcio, Orson Zhai,
Baolin Wang, Chunyan Zhang, Rob Herring (Arm)
Cc: linux-arm-kernel, linux-actions, devicetree, linux-kernel,
openbmc, linux-tegra, linux-arm-msm, linux-rockchip
Am Dienstag, 30. April 2024, 21:18:54 CEST schrieb Rob Herring (Arm):
> '/memory' nodes always have a 'reg' property, and therefore should have
> a unit-address with just plain hex (i.e. no commas). Fix all the arm64
> '/memory' nodes.
>
> It's possible that some bootloader depends on /memory (arm32 ATAG to DT
> code does for example). If so, the memory node should be commented with
> that requirement.
>
> Signed-off-by: Rob Herring (Arm) <robh@kernel.org>
> ---
> arch/arm64/boot/dts/rockchip/rk3368-evb.dtsi | 2 +-
> arch/arm64/boot/dts/rockchip/rk3368-orion-r68-meta.dts | 2 +-
> arch/arm64/boot/dts/rockchip/rk3368-r88.dts | 2 +-
For the Rockchip boards
Reviewed-by: Heiko Stuebner <heiko@sntech.de>
Other rk3368-based boards like the Geekbox and PX5-evb already use
the correct memory@0 node names, so it looks like the the bootloaders
used on rk3368 don't have special requirements for the node-naming.
^ permalink raw reply [flat|nested] 8+ messages in thread* Re: [PATCH] arm64: dts: Add/fix /memory node unit-addresses
2024-04-30 19:18 [PATCH] arm64: dts: Add/fix /memory node unit-addresses Rob Herring (Arm)
2024-04-30 21:49 ` Florian Fainelli
2024-05-01 10:31 ` Heiko Stübner
@ 2024-05-01 10:44 ` Bryan O'Donoghue
2024-05-02 5:00 ` Chanho Min
` (2 subsequent siblings)
5 siblings, 0 replies; 8+ messages in thread
From: Bryan O'Donoghue @ 2024-05-01 10:44 UTC (permalink / raw)
To: Rob Herring (Arm), soc, Andreas Färber,
Manivannan Sadhasivam, Krzysztof Kozlowski, Conor Dooley,
Khuong Dinh, Ray Jui, Scott Branden,
Broadcom internal kernel review list, Robert Richter, Chanho Min,
Avi Fishman, Tomer Maimon, Tali Perry, Patrick Venture,
Nancy Yuen, Benjamin Fair, Thierry Reding, Jonathan Hunter,
Bjorn Andersson, Konrad Dybcio, Heiko Stuebner, Orson Zhai,
Baolin Wang, Chunyan Zhang
Cc: linux-arm-kernel, linux-actions, devicetree, linux-kernel,
openbmc, linux-tegra, linux-arm-msm, linux-rockchip
On 30/04/2024 20:18, Rob Herring (Arm) wrote:
> +++ b/arch/arm64/boot/dts/qcom/ipq8074-hk10.dtsi
> @@ -16,7 +16,7 @@ chosen {
> stdout-path = "serial0";
> };
>
> - memory {
> + memory@40000000 {
> device_type = "memory";
> reg = <0x0 0x40000000 0x0 0x20000000>;
> };
Reviewed-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
^ permalink raw reply [flat|nested] 8+ messages in thread* Re: [PATCH] arm64: dts: Add/fix /memory node unit-addresses
2024-04-30 19:18 [PATCH] arm64: dts: Add/fix /memory node unit-addresses Rob Herring (Arm)
` (2 preceding siblings ...)
2024-05-01 10:44 ` Bryan O'Donoghue
@ 2024-05-02 5:00 ` Chanho Min
2024-05-02 6:34 ` Krzysztof Kozlowski
2024-06-28 15:06 ` Thierry Reding
5 siblings, 0 replies; 8+ messages in thread
From: Chanho Min @ 2024-05-02 5:00 UTC (permalink / raw)
To: Rob Herring (Arm), soc, Andreas Färber,
Manivannan Sadhasivam, Krzysztof Kozlowski, Conor Dooley,
Khuong Dinh, Ray Jui, Scott Branden,
Broadcom internal kernel review list, Robert Richter, Avi Fishman,
Tomer Maimon, Tali Perry, Patrick Venture, Nancy Yuen,
Benjamin Fair, Thierry Reding, Jonathan Hunter, Bjorn Andersson,
Konrad Dybcio, Heiko Stuebner, Orson Zhai, Baolin Wang,
Chunyan Zhang
Cc: linux-arm-kernel, linux-actions, devicetree, linux-kernel,
openbmc, linux-tegra, linux-arm-msm, linux-rockchip
On 5/1/24 4:18, Rob Herring (Arm) wrote:
> '/memory' nodes always have a 'reg' property, and therefore should have
> a unit-address with just plain hex (i.e. no commas). Fix all the arm64
> '/memory' nodes.
>
> It's possible that some bootloader depends on /memory (arm32 ATAG to DT
> code does for example). If so, the memory node should be commented with
> that requirement.
> arch/arm64/boot/dts/lg/lg1312-ref.dts | 2 +-
> arch/arm64/boot/dts/lg/lg1313-ref.dts | 2 +-
Reviewed-by: Chanho Min <chanho.min@lge.com>
^ permalink raw reply [flat|nested] 8+ messages in thread* Re: [PATCH] arm64: dts: Add/fix /memory node unit-addresses
2024-04-30 19:18 [PATCH] arm64: dts: Add/fix /memory node unit-addresses Rob Herring (Arm)
` (3 preceding siblings ...)
2024-05-02 5:00 ` Chanho Min
@ 2024-05-02 6:34 ` Krzysztof Kozlowski
2024-06-28 15:06 ` Thierry Reding
5 siblings, 0 replies; 8+ messages in thread
From: Krzysztof Kozlowski @ 2024-05-02 6:34 UTC (permalink / raw)
To: Rob Herring (Arm), soc, Andreas Färber,
Manivannan Sadhasivam, Krzysztof Kozlowski, Conor Dooley,
Khuong Dinh, Ray Jui, Scott Branden,
Broadcom internal kernel review list, Robert Richter, Chanho Min,
Avi Fishman, Tomer Maimon, Tali Perry, Patrick Venture,
Nancy Yuen, Benjamin Fair, Thierry Reding, Jonathan Hunter,
Bjorn Andersson, Konrad Dybcio, Heiko Stuebner, Orson Zhai,
Baolin Wang, Chunyan Zhang
Cc: linux-arm-kernel, linux-actions, devicetree, linux-kernel,
openbmc, linux-tegra, linux-arm-msm, linux-rockchip
On 30/04/2024 21:18, Rob Herring (Arm) wrote:
> '/memory' nodes always have a 'reg' property, and therefore should have
> a unit-address with just plain hex (i.e. no commas). Fix all the arm64
> '/memory' nodes.
>
> It's possible that some bootloader depends on /memory (arm32 ATAG to DT
> code does for example). If so, the memory node should be commented with
> that requirement.
>
> Signed-off-by: Rob Herring (Arm) <robh@kernel.org>
> ---
> SoC maintainers, please take this directly.
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Best regards,
Krzysztof
^ permalink raw reply [flat|nested] 8+ messages in thread* Re: [PATCH] arm64: dts: Add/fix /memory node unit-addresses
2024-04-30 19:18 [PATCH] arm64: dts: Add/fix /memory node unit-addresses Rob Herring (Arm)
` (4 preceding siblings ...)
2024-05-02 6:34 ` Krzysztof Kozlowski
@ 2024-06-28 15:06 ` Thierry Reding
2024-07-01 14:29 ` Diogo Ivo
5 siblings, 1 reply; 8+ messages in thread
From: Thierry Reding @ 2024-06-28 15:06 UTC (permalink / raw)
To: Rob Herring (Arm)
Cc: soc, Krzysztof Kozlowski, Conor Dooley, Diogo Ivo,
linux-arm-kernel, devicetree, linux-kernel, linux-tegra
[-- Attachment #1: Type: text/plain, Size: 1135 bytes --]
On Tue, Apr 30, 2024 at 02:18:54PM GMT, Rob Herring (Arm) wrote:
[...]
> diff --git a/arch/arm64/boot/dts/nvidia/tegra210-smaug.dts b/arch/arm64/boot/dts/nvidia/tegra210-smaug.dts
> index 9ebb7369256e..2e5b6b2c1f56 100644
> --- a/arch/arm64/boot/dts/nvidia/tegra210-smaug.dts
> +++ b/arch/arm64/boot/dts/nvidia/tegra210-smaug.dts
> @@ -25,7 +25,7 @@ chosen {
> stdout-path = "serial0:115200n8";
> };
>
> - memory {
> + memory@80000000 {
> device_type = "memory";
> reg = <0x0 0x80000000 0x0 0xc0000000>;
> };
[trimming the recipient list and adding Diogo]
Sorry I just noticed this as I was doing a cleanup path of patchwork.
For tegra210-smaug, unfortunately we can't do this. The problem is that
the firmware for this device looks for a hard-coded /memory device when
updating with the proper memory size and this firmware can't always be
updated.
Diogo, you're one of the few remaining people that I know of that still
use this device (with an upstream kernel). Do you have any more details
about this? Is there any way the firmware can be safely updated on these
devices?
Thierry
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
^ permalink raw reply [flat|nested] 8+ messages in thread* Re: [PATCH] arm64: dts: Add/fix /memory node unit-addresses
2024-06-28 15:06 ` Thierry Reding
@ 2024-07-01 14:29 ` Diogo Ivo
0 siblings, 0 replies; 8+ messages in thread
From: Diogo Ivo @ 2024-07-01 14:29 UTC (permalink / raw)
To: Thierry Reding
Cc: Rob Herring (Arm), soc, Krzysztof Kozlowski, Conor Dooley,
linux-arm-kernel, devicetree, linux-kernel, linux-tegra
Hi Thierry,
On Fri, Jun 28, 2024 at 05:06:10PM GMT, Thierry Reding wrote:
> On Tue, Apr 30, 2024 at 02:18:54PM GMT, Rob Herring (Arm) wrote:
> [...]
> > diff --git a/arch/arm64/boot/dts/nvidia/tegra210-smaug.dts b/arch/arm64/boot/dts/nvidia/tegra210-smaug.dts
> > index 9ebb7369256e..2e5b6b2c1f56 100644
> > --- a/arch/arm64/boot/dts/nvidia/tegra210-smaug.dts
> > +++ b/arch/arm64/boot/dts/nvidia/tegra210-smaug.dts
> > @@ -25,7 +25,7 @@ chosen {
> > stdout-path = "serial0:115200n8";
> > };
> >
> > - memory {
> > + memory@80000000 {
> > device_type = "memory";
> > reg = <0x0 0x80000000 0x0 0xc0000000>;
> > };
>
> [trimming the recipient list and adding Diogo]
>
> Sorry I just noticed this as I was doing a cleanup path of patchwork.
>
> For tegra210-smaug, unfortunately we can't do this. The problem is that
> the firmware for this device looks for a hard-coded /memory device when
> updating with the proper memory size and this firmware can't always be
> updated.
>
> Diogo, you're one of the few remaining people that I know of that still
> use this device (with an upstream kernel). Do you have any more details
> about this? Is there any way the firmware can be safely updated on these
> devices?
To be honest I am not quite sure. I do know that at least there is a part
of the firmware that is locked down and cannot be updated easily, as you
have to disconnect a ribbon cable inside of the device. Do you know of
any firmware updates for this device? Currently I am not aware of any.
Diogo
^ permalink raw reply [flat|nested] 8+ messages in thread