Hi Adam,
On Wed, 2024-11-13 at 07:06 +0000, Yannic Moog wrote:
On Mon, 2024-11-11 at 04:49 -0600, Adam Ford wrote:
On Mon, Nov 11, 2024 at 1:48 AM Yannic Moog Y.Moog@phytec.de wrote:
Hi,
On Fri, 2024-11-08 at 10:05 -0800, Tim Harvey wrote:
On Fri, Nov 8, 2024 at 5:49 AM Adam Ford aford173@gmail.com wrote:
On Thu, Nov 7, 2024 at 2:42 AM Yannic Moog y.moog@phytec.de wrote:
Add tee node in SoC u-boot device trees. Use a kconfig entry to specify load and entry addresses for the op-tee image in the respective defconfig.
Signed-off-by: Yannic Moog y.moog@phytec.de
arch/arm/dts/imx8mm-u-boot.dtsi | 17 ++++++++++++++++- arch/arm/dts/imx8mn-u-boot.dtsi | 17 ++++++++++++++++- arch/arm/dts/imx8mp-u-boot.dtsi | 17 ++++++++++++++++- arch/arm/mach-imx/imx8m/Kconfig | 8 ++++++++ 4 files changed, 56 insertions(+), 3 deletions(-)
diff --git a/arch/arm/dts/imx8mm-u-boot.dtsi b/arch/arm/dts/imx8mm-u-boot.dtsi index d31bc822532..ecc2319279e 100644 --- a/arch/arm/dts/imx8mm-u-boot.dtsi +++ b/arch/arm/dts/imx8mm-u-boot.dtsi @@ -164,6 +164,21 @@ }; #endif
+ tee: tee { + description = "OP-TEE"; + type = "tee"; + arch = "arm64"; + compression = "none"; + os = "tee"; + load = <CONFIG_IMX8M_OPTEE_LOAD_ADDR>; + entry = <CONFIG_IMX8M_OPTEE_LOAD_ADDR>;
+ tee-os { + filename = "tee.bin"; + optional; + }; + };
binman_fip: fip { arch = "arm64"; compression = "none"; @@ -192,7 +207,7 @@ fdt = "fdt-SEQ"; firmware = "uboot"; #ifndef CONFIG_ARMV8_PSCI - loadables = "atf"; + loadables = "atf", "tee"; #endif }; }; diff --git a/arch/arm/dts/imx8mn-u-boot.dtsi b/arch/arm/dts/imx8mn-u-boot.dtsi index 6875c6d44ff..f9108cb75c7 100644 --- a/arch/arm/dts/imx8mn-u-boot.dtsi +++ b/arch/arm/dts/imx8mn-u-boot.dtsi @@ -235,6 +235,21 @@ }; #endif
+ tee: tee { + description = "OP-TEE"; + type = "tee"; + arch = "arm64"; + compression = "none"; + os = "tee"; + load = <CONFIG_IMX8M_OPTEE_LOAD_ADDR>; + entry = <CONFIG_IMX8M_OPTEE_LOAD_ADDR>;
+ tee-os { + filename = "tee.bin"; + optional; + }; + };
binman_fip: fip { arch = "arm64"; compression = "none"; @@ -263,7 +278,7 @@ fdt = "fdt-SEQ"; firmware = "uboot"; #ifndef CONFIG_ARMV8_PSCI - loadables = "atf"; + loadables = "atf", "tee"; #endif }; }; diff --git a/arch/arm/dts/imx8mp-u-boot.dtsi b/arch/arm/dts/imx8mp-u-boot.dtsi index 56749ccacd2..9ede98a11e4 100644 --- a/arch/arm/dts/imx8mp-u-boot.dtsi +++ b/arch/arm/dts/imx8mp-u-boot.dtsi @@ -185,6 +185,21 @@ }; #endif
+ tee: tee { + description = "OP-TEE"; + type = "tee"; + arch = "arm64"; + compression = "none"; + os = "tee"; + load = <CONFIG_IMX8M_OPTEE_LOAD_ADDR>; + entry = <CONFIG_IMX8M_OPTEE_LOAD_ADDR>;
+ tee-os { + filename = "tee.bin"; + optional; + }; + };
@fdt-SEQ { description = "NAME"; type = "flat_dt"; @@ -204,7 +219,7 @@ fdt = "fdt-SEQ"; firmware = "uboot"; #ifndef CONFIG_ARMV8_PSCI - loadables = "atf"; + loadables = "atf", "tee"; #endif }; }; diff --git a/arch/arm/mach-imx/imx8m/Kconfig b/arch/arm/mach-imx/imx8m/Kconfig index b254a50b091..14f14db6a35 100644 --- a/arch/arm/mach-imx/imx8m/Kconfig +++ b/arch/arm/mach-imx/imx8m/Kconfig @@ -37,6 +37,14 @@ config SYS_HAS_ARMV8_SECURE_BASE If enabled, please also define the value for ARMV8_SECURE_BASE, for i.MX8M, it could be some address in OCRAM.
+config IMX8M_OPTEE_LOAD_ADDR + hex "Load address of OPTEE image" + default 0x56000000 + help + The load and entry address for the OPTEE image. This value defaults to + 0x56000000 if not provided in the board defconfig file.
I think this would need to vary based on the platform. I think the addresses may be different depending on Mini, Nano and Plus and also dependent on the available SDRAM.
Indeed the address differs for platform (as the DRAM base address does) and is also based on the DRAM size of the board. Here are my notes for IMX8M:
The address is the link/load address for TEE that should match BL32_BASE for the ATF (for imx8m) and CFG_TZDRAM_START for OPTEE. The BL32_BASE is the same for imx8mm/n/p but it does depend on the size of DRAM. The address should be the top 32MiB of DRAM but because U-Boot is loading tee.bin from a FIT image into DRAM it needs to be a 32bit address and thus needs to be adjusted down if you are on a board with 4GiB: 1GiB DRAM use 0x7e000000 2GiB DRAM use 0xbe000000 3GiB or larger use 0xfe000000
I agree the addresses may be different. To my knowledge you can take whatever (32bit) address you want within addressable RAM as load address. I assume you are looking for sensible defaults here. It appears to me the challenge is agreeing on a sensible value. Do you have any suggestions?
On my Mini with 2GB of RAM, there are several definitions: PHYS_SDRAM 0x40000000 PHYS_SDRAM_SIZE 0x80000000 /* 2GB DDR */
What we if did a calculation like: PHYS_SDRAM + MIN(PHYS_SDRAM_SIZE, 0xc0000000) - 0x2000000
With 2 GB of RAM, it would be 0xbe000000
My Nano Board has the following:
PHYS_SDRAM 0x40000000 PHYS_SDRAM_SIZE 0x40000000
That calculation should return 7E000000.
Math looks good to me, however I don't understand where we would put this in the source code. How can we set this as default while having access to the macros?
As I don't think it is possible to implement the mechanic as you suggested, I would like to keep it simple.
trusted-firmware-a also has default for BL32_BASE. I would like to set the OPTEE_LOAD_ADDR as they are set in tf-a. This has the advantage that for users who don't change the default, they can rely on it being consistent across projects.
Yannic
Yannic
I am not sure if we can do MIN and MAX functions in the precompiler, but it would then cap the value to 0xfe000000
adam
Yannic
Best Regards,
Tim
upstream mailing list upstream@lists.phytec.de http://lists.phytec.de/cgi-bin/mailman/listinfo/upstream