ARM: tegra: support for Trusted Foundations
v3 cover ARM: tegra: support for Trusted Foundations
This is the long-overdue third version of this series, hopefully addressing the suggestions received in v2.
Since I could confirm that the SMC interface used by Trusted Foundations is not Tegra-specific, I have moved core TF support into arch/arm/firmware as a proposal to consolidate all implementations of Tomasz Figa's firmware interface into that directory.
Changes from v2:
- Moved core TF support into arch/arm/firmware
- Added TF version number in DT bindings
- Cleaned TF registration code
- Tegra kernel now fails if TF is declared in the device tree but support is not compiled in
- Tegra kernel now fails if attempt to set CPU reset vector through secure monitor returned an error
- Moved update to tegra_defconfig into its own patch
Alexandre Courbot (5):
ARM: add basic Trusted Foundations support
ARM: tegra: add support for Trusted Foundations
ARM: tegra: split setting of CPU reset handler
ARM: tegra: set CPU reset handler with firmware op
ARM: tegra: support Trusted Foundations by default
.../arm/firmware/tl,trusted-foundations.txt | 16 ++++++ Documentation/devicetree/bindings/arm/tegra.txt | 5 ++ .../devicetree/bindings/vendor-prefixes.txt | 1 + arch/arm/Kconfig | 2 + arch/arm/Makefile | 1 + arch/arm/configs/tegra_defconfig | 1 + arch/arm/firmware/Kconfig | 22 ++++++++ arch/arm/firmware/Makefile | 1 + arch/arm/firmware/trusted_foundations.c | 58 ++++++++++++++++++++++ arch/arm/include/asm/trusted_foundations.h | 36 ++++++++++++++ arch/arm/mach-tegra/Kconfig | 1 + arch/arm/mach-tegra/Makefile | 1 + arch/arm/mach-tegra/common.c | 2 + arch/arm/mach-tegra/firmware.c | 41 +++++++++++++++ arch/arm/mach-tegra/firmware.h | 19 +++++++ arch/arm/mach-tegra/reset.c | 40 +++++++++++---- 16 files changed, 236 insertions(+), 11 deletions(-) create mode 100644 Documentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt create mode 100644 arch/arm/firmware/Kconfig create mode 100644 arch/arm/firmware/Makefile create mode 100644 arch/arm/firmware/trusted_foundations.c create mode 100644 arch/arm/include/asm/trusted_foundations.h create mode 100644 arch/arm/mach-tegra/firmware.c create mode 100644 arch/arm/mach-tegra/firmware.h
Signature · 1 line
1.8.3.4
PATCH v3 1/5 ARM: add basic Trusted Foundations support
Trusted Foundations is a TrustZone-based secure monitor for ARM that can be invoked using a consistent smc-based API on all supported platforms. This patch adds initial basic support for Trusted Foundations using the ARM firmware API. Current features are limited to the ability to boot secondary processors.
--- .../arm/firmware/tl,trusted-foundations.txt | 16 ++++++ .../devicetree/bindings/vendor-prefixes.txt | 1 + arch/arm/Kconfig | 2 + arch/arm/Makefile | 1 + arch/arm/firmware/Kconfig | 22 ++++++++ arch/arm/firmware/Makefile | 1 + arch/arm/firmware/trusted_foundations.c | 58 ++++++++++++++++++++++ arch/arm/include/asm/trusted_foundations.h | 36 ++++++++++++++ 8 files changed, 137 insertions(+) create mode 100644 Documentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt create mode 100644 arch/arm/firmware/Kconfig create mode 100644 arch/arm/firmware/Makefile create mode 100644 arch/arm/firmware/trusted_foundations.c create mode 100644 arch/arm/include/asm/trusted_foundations.h
Documentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt new file +16 -0
@@ -0,0 +1,16 @@ +Trusted Foundations + +Boards that use the Trusted Foundations secure monitor can signal its +presence by declaring a node compatible with "tl,trusted-foundations" +(the name and location of the node does not matter). + +Required properties: +- compatible : "tl,trusted-foundations" +- version : Must contain the version number string of the Trusted Foundation + firmware. + +Example: + firmware { + compatible = "tl,trusted-foundations"; + version = "02.08"; + };
Documentation/devicetree/bindings/vendor-prefixes.txt +1 -0
@@ -63,6 +63,7 @@ ste ST-Ericsson stericsson ST-Ericsson toumaz Toumaz ti Texas Instruments +tl Trusted Logic toshiba Toshiba Corporation v3 V3 Semiconductor via VIA Technologies, Inc.
arch/arm/Kconfig +2 -0
@@ -1074,6 +1074,8 @@ config ARM_TIMER_SP804 select CLKSRC_MMIO select CLKSRC_OF if OF +source "arch/arm/firmware/Kconfig" + source arch/arm/mm/Kconfig config ARM_NR_BANKS
arch/arm/Makefile +1 -0
@@ -267,6 +267,7 @@ core-$(CONFIG_KVM_ARM_HOST) += arch/arm/kvm/ core-y += arch/arm/kernel/ arch/arm/mm/ arch/arm/common/ core-y += arch/arm/net/ core-y += arch/arm/crypto/ +core-y += arch/arm/firmware/ core-y += $(machdirs) $(platdirs) drivers-$(CONFIG_OPROFILE) += arch/arm/oprofile/
arch/arm/firmware/Kconfig new file +22 -0
@@ -0,0 +1,22 @@ +config ARCH_SUPPORTS_TRUSTED_FOUNDATIONS + bool + +menu "Firmware options" + depends on ARCH_SUPPORTS_TRUSTED_FOUNDATIONS + +config TRUSTED_FOUNDATIONS + bool "Trusted Foundations secure monitor support" + depends on ARCH_SUPPORTS_TRUSTED_FOUNDATIONS + help + Some devices are booted with the Trusted Foundations secure monitor + active, requiring some core operations to be performed by the secure + monitor instead of the kernel. + + This option allows the kernel to invoke the secure monitor whenever + required on devices using Trusted Foundations. + + Devices using Trusted Foundations should pass a device tree containing + a node compatible with "tl,trusted-foundations" to signal the presence + of the secure monitor. + +endmenu
arch/arm/firmware/Makefile new file +1 -0
@@ -0,0 +1 @@ +obj-$(CONFIG_TRUSTED_FOUNDATIONS) += trusted_foundations.o
arch/arm/firmware/trusted_foundations.c new file +58 -0
58 lines not on this page · raw message
arch/arm/include/asm/trusted_foundations.h new file +36 -0
@@ -0,0 +1,36 @@ +/* +* Copyright (c) 2013, NVIDIA Corporation. +* +* This program is free software; you can redistribute it and/or modify +* it under the terms of the GNU General Public License as published by +* the Free Software Foundation; either version 2 of the License, or +* (at your option) any later version. +* +* This program is distributed in the hope that it will be useful, but WITHOUT +* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or +* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for +* more details. +*/ + +#ifndef __ASM_ARM_TRUSTED_FOUNDATIONS_H +#define __ASM_ARM_TRUSTED_FOUNDATIONS_H + +#include <linux/kconfig.h> +#include <linux/printk.h> +#include <asm/bug.h> + +#if IS_ENABLED (CONFIG_TRUSTED_FOUNDATIONS ) + +void register_trusted_foundations (const char *version); + +#else + +static inline void register_trusted_foundations (const char *version) +{ +pr_crit ("No support for Trusted Foundations, stopping...\n" ); +BUG (); +} + +#endif /* IS_ENABLED(CONFIG_TRUSTED_FOUNDATIONS) */ + +#endif
Signature · 1 line
1.8.3.4
Re: [PATCH v3 1/5] ARM: add basic Trusted Foundations support
Quoted 9 lines · Alexandre Courbot
Trusted Foundations is a TrustZone-based secure monitor for ARM that can be invoked using a consistent smc-based API on all supported platforms. This patch adds initial basic support for Trusted Foundations using the ARM firmware API. Current features are limited to the ability to boot secondary processors.
Documentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt +1
-0+Trusted Foundations
Documentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt line 10
Quoted 10 lines
Documentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt +9
-0+ +Boards that use the Trusted Foundations secure monitor can signal its +presence by declaring a node compatible with "tl,trusted-foundations" +(the name and location of the node does not matter). + +Required properties: +- compatible : "tl,trusted-foundations" +- version : Must contain the version number string of the Trusted Foundation + firmware.
Can the version be queried at run-time from the firmware itself?
If not, I wonder if we shouldn't instead encode the version number into the compatible value.
Some comments on the exact format of the version property would be useful; from the example I assume it's "%02d.%02d" % (major_ver, minor_ver)?
arch/arm/firmware/Kconfig line 2
Quoted 4 lines
arch/arm/firmware/Kconfig +2
-0+config ARCH_SUPPORTS_TRUSTED_FOUNDATIONS + bool
Shouldn't that be "config ARCH_SUPPORTS_FIRMWARE", since presumably in the future there will be more entries in the menu, and hence we want the menu to appear if any of those entries are useful?
arch/arm/firmware/Kconfig line 5
Quoted 3 lines
arch/arm/firmware/Kconfig +3
-0+ +menu "Firmware options" + depends on ARCH_SUPPORTS_TRUSTED_FOUNDATIONS
Or perhaps that comment is more appropriate for that "depends".
Re: [PATCH v3 1/5] ARM: add basic Trusted Foundations support
Quoted 18 lines · Stephen Warren
Quoted 16 lines · Alexandre Courbot
Trusted Foundations is a TrustZone-based secure monitor for ARM that can be invoked using a consistent smc-based API on all supported platforms. This patch adds initial basic support for Trusted Foundations using the ARM firmware API. Current features are limited to the ability to boot secondary processors.
Documentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt +8
-0+Trusted Foundations + +Boards that use the Trusted Foundations secure monitor can signal its +presence by declaring a node compatible with "tl,trusted-foundations" +(the name and location of the node does not matter). + +Required properties: +- compatible : "tl,trusted-foundations"
Quoted 4 lines
Quoted 2 lines
+- version : Must contain the version number string of the Trusted Foundation + firmware.Can the version be queried at run-time from the firmware itself?
I'm afraid there is not, unfortunately. :/
Quoted 2 lines
If not, I wonder if we shouldn't instead encode the version number into the compatible value.
Why? This would make the node harder to find and would also complicate the case where a given firmware handler can support a given range of versions (which is what I expect will happen).
Quoted 2 lines
Some comments on the exact format of the version property would be useful; from the example I assume it's "%02d.%02d" % (
major_ver,minor_ver)?
Agreed.
Quoted 8 lines
arch/arm/firmware/Kconfig line 2
Quoted 4 lines
arch/arm/firmware/Kconfig +2
-0+config ARCH_SUPPORTS_TRUSTED_FOUNDATIONS + boolShouldn't that be "config
ARCH_SUPPORTS_FIRMWARE", since presumably in the future there will be more entries in the menu, and hence we want the menu to appear if any of those entries are useful?
The things is that because you support one firmware does not mean you will support then all. Only some set of architectures will support firmware X, and another set (maybe not inclusive) will support firmware Y. We do not want to allow selecting firmware Y if the kernel does not support any of the architectures that may make use of it.
Quoted 5 lines
arch/arm/firmware/Kconfig line 5
Quoted 3 lines
arch/arm/firmware/Kconfig +3
-0+ +menu "Firmware options" + depends on ARCH_SUPPORTS_TRUSTED_FOUNDATIONSOr perhaps that comment is more appropriate for that "depends".
Here the idea is to not show the "Firmware options" menu if the kernel does not include support for any architecture that uses them. As more firmwares get added, the depends line should expand into something like "depends on ARCH_SUPPORTS_FIRMWARE_X || ARCH_SUPPORTS_FIRMWARE_Y || ...
Maybe there's a better way to express this?
Thanks,
Alex.
Re: [PATCH v3 1/5] ARM: add basic Trusted Foundations support
Quoted 24 lines · Alexandre Courbot
Quoted 18 lines · Stephen Warren
Quoted 16 lines · Alexandre Courbot
Trusted Foundations is a TrustZone-based secure monitor for ARM that can be invoked using a consistent smc-based API on all supported platforms. This patch adds initial basic support for Trusted Foundations using the ARM firmware API. Current features are limited to the ability to boot secondary processors.
Documentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt +8
-0+Trusted Foundations + +Boards that use the Trusted Foundations secure monitor can signal its +presence by declaring a node compatible with "tl,trusted-foundations" +(the name and location of the node does not matter). + +Required properties: +- compatible : "tl,trusted-foundations"Quoted 4 lines
Quoted 2 lines
+- version : Must contain the version number string of the Trusted Foundation + firmware.Can the version be queried at run-time from the firmware itself?
Quoted 8 lines
I'm afraid there is not, unfortunately. :/
Quoted 2 lines
If not, I wonder if we shouldn't instead encode the version number into the compatible value.
Why? This would make the node harder to find and would also complicate the case where a given firmware handler can support a given range of versions (which is what I expect will happen).
I guess that would be painful. There are probably a lot more valid version numbers for a SW firmware interface than there are for a HW IP block interface, so it makes less sense to encoded the version into the compatible value, at least in cases where the interface is "similar enough"; we can add some kind of basic version number into the compatible values for radically different interfaces if they appear in the future.
Quoted 21 lines
Quoted 8 lines
arch/arm/firmware/Kconfig line 2
Quoted 4 lines
arch/arm/firmware/Kconfig +2
-0+config ARCH_SUPPORTS_TRUSTED_FOUNDATIONS + boolShouldn't that be "config
ARCH_SUPPORTS_FIRMWARE", since presumably in the future there will be more entries in the menu, and hence we want the menu to appear if any of those entries are useful?The things is that because you support one firmware does not mean you will support then all. Only some set of architectures will support firmware X, and another set (maybe not inclusive) will support firmware Y. We do not want to allow selecting firmware Y if the kernel does not support any of the architectures that may make use of it.
Quoted 5 lines
arch/arm/firmware/Kconfig line 5
Quoted 3 lines
arch/arm/firmware/Kconfig +3
-0+ +menu "Firmware options" + depends on ARCH_SUPPORTS_TRUSTED_FOUNDATIONSOr perhaps that comment is more appropriate for that "depends".
Quoted 7 lines
Here the idea is to not show the "Firmware options" menu if the kernel does not include support for any architecture that uses them. As more firmwares get added, the depends line should expand into something like "depends on
ARCH_SUPPORTS_FIRMWARE_X||ARCH_SUPPORTS_FIRMWARE_Y|| ...Maybe there's a better way to express this?
Is there a need for a menu at all? I suppose it might group things nicely. If so, how about:
config ARCH_SUPPORTS_FIRMWARE
bool
config ARCH_SUPPORTS_TRUSTED_FOUNDATIONS
bool
select ARCH_SUPPORTS_FIRMWARE
config ARCH_SUPPORTS_SOME_OTHER_FW
bool
select ARCH_SUPPORTS_FIRMWARE
menu "Firmware options"
depends on ARCH_SUPPORTS_FIRMWARE
That's probably slightly easier to read than a huge "depends" statement in the menu stanza, and ARCH_SUPPORTS_FIRMWARE might be a useful to build in support for any core firmware utility code.
Re: [PATCH v3 1/5] ARM: add basic Trusted Foundations support
In reply to Alexandre CourbotQuoted 29 lines · Alexandre Courbot
Trusted Foundations is a TrustZone-based secure monitor for ARM that can be invoked using a consistent smc-based API on all supported platforms. This patch adds initial basic support for Trusted Foundations using the ARM firmware API. Current features are limited to the ability to boot secondary processors.
Signed-off-by Alexandre Courbot <[email protected]>--- .../arm/firmware/tl,trusted-foundations.txt | 16 ++++++ .../devicetree/bindings/vendor-prefixes.txt | 1 + arch/arm/Kconfig | 2 + arch/arm/Makefile | 1 + arch/arm/firmware/Kconfig | 22 ++++++++ arch/arm/firmware/Makefile | 1 + arch/arm/firmware/trusted_foundations.c | 58 ++++++++++++++++++++++ arch/arm/include/asm/trusted_foundations.h | 36 ++++++++++++++ 8 files changed, 137 insertions(+) create mode 100644 Documentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt create mode 100644 arch/arm/firmware/Kconfig create mode 100644 arch/arm/firmware/Makefile create mode 100644 arch/arm/firmware/trusted_foundations.c create mode 100644 arch/arm/include/asm/trusted_foundations.hDocumentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt new file +0
-0
Documentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt line 5
Quoted 5 lines
Documentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt new file +5
-0@@ -0,0 +1,16 @@ +Trusted Foundations + +Boards that use the Trusted Foundations secure monitor can signal its +presence by declaring a node compatible with "tl,trusted-foundations" +(the name and location of the node does not matter).
We shouldn't encougrage people to arrange DT nodes at random.
Unless there's a good reason not to, this should just be a child of /, with the meaning that the firmware is callable (at least to the point of discoverability) on all CPUs.
The meaning of the node should be left unspecified if it is in any other location.
This is the convention already adopted for things like arm,psci (though to be fair, the arm,psci binding documentation fails to mention that also -- nonetheless, the node is placed in /).
Documentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt line 10
Quoted 5 lines
Documentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt new file +5
-0+ +Required properties: +- compatible : "tl,trusted-foundations" +- version : Must contain the version number string of the Trusted Foundation + firmware.
Are you sure there is no low-level way to probe vendor and version info? If there is, then the DT should describe nothing except the fact that the probe interface exists.
I also worry that two integrations on different SoCs might have the same version number, yet still be different due to vendor-specific features and options.
Quoted 97 lines
Documentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt new file +6
-0+ +Example: + firmware { + compatible = "tl,trusted-foundations"; + version = "02.08"; + };Documentation/devicetree/bindings/vendor-prefixes.txt +1
-0@@ -63,6 +63,7 @@ ste ST-Ericsson stericsson ST-Ericsson toumaz Toumaz ti Texas Instruments +tl Trusted Logic toshiba Toshiba Corporation v3 V3 Semiconductor via VIA Technologies, Inc.arch/arm/Kconfig +2
-0@@ -1074,6 +1074,8 @@ config ARM_TIMER_SP804 select CLKSRC_MMIO select CLKSRC_OF if OF +source "arch/arm/firmware/Kconfig" + source arch/arm/mm/Kconfig config ARM_NR_BANKSarch/arm/Makefile +1
-0@@ -267,6 +267,7 @@ core-$(CONFIG_KVM_ARM_HOST) += arch/arm/kvm/ core-y += arch/arm/kernel/ arch/arm/mm/ arch/arm/common/ core-y += arch/arm/net/ core-y += arch/arm/crypto/ +core-y += arch/arm/firmware/ core-y += $(machdirs) $(platdirs) drivers-$(CONFIG_OPROFILE) += arch/arm/oprofile/arch/arm/firmware/Kconfig new file +22
-0@@ -0,0 +1,22 @@ +config ARCH_SUPPORTS_TRUSTED_FOUNDATIONS + bool + +menu "Firmware options" + depends on ARCH_SUPPORTS_TRUSTED_FOUNDATIONS + +config TRUSTED_FOUNDATIONS + bool "Trusted Foundations secure monitor support" + depends on ARCH_SUPPORTS_TRUSTED_FOUNDATIONS + help + Some devices are booted with the Trusted Foundations secure monitor + active, requiring some core operations to be performed by the secure + monitor instead of the kernel. + + This option allows the kernel to invoke the secure monitor whenever + required on devices using Trusted Foundations. + + Devices using Trusted Foundations should pass a device tree containing + a node compatible with "tl,trusted-foundations" to signal the presence + of the secure monitor. + +endmenuarch/arm/firmware/Makefile new file +1
-0@@ -0,0 +1 @@ +obj-$(CONFIG_TRUSTED_FOUNDATIONS) += trusted_foundations.oarch/arm/firmware/trusted_foundations.c new file +13
-0@@ -0,0 +1,58 @@ +/* +* Trusted Foundations support for ARM CPUs +* +* Copyright (c) 2013, NVIDIA Corporation. +* +* This program is free software; you can redistribute it and/or modify +* it under the terms of the GNU General Public License as published by +* the Free Software Foundation; either version 2 of the License, or +* (at your option) any later version. +* +* This program is distributed in the hope that it will be useful, but WITHOUT +* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or +* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
arch/arm/firmware/trusted_foundations.c line 23
Quoted 10 lines
arch/arm/firmware/trusted_foundations.c new file +10
-0+* more details. +*/ + +#include <linux/kernel.h> +#include <linux/init.h> +#include <linux/of.h> +#include <asm/firmware.h> + +#define TF_SET_CPU_BOOT_ADDR_SMC 0xfffff200 +
Better to use __naked instead of an explicit attribute.
Quoted 23 lines
arch/arm/firmware/trusted_foundations.c new file +23
-0+static void __attribute__((naked))tf_generic_smc (u32 type,u32 subtype, +u32 arg) +{ + asmvolatile ( +".arch_extension sec\n\t" +"stmfd sp!, {r4 - r11, lr}\n\t" + __asmeq("%0", "r0") + __asmeq("%1", "r1") + __asmeq("%2", "r2") + "mov r3, #0\n\t" + "mov r4, #0\n\t" + "smc #0\n\t" + "ldmfd sp!, {r4 - r11, pc}" + : + : "r" (type), "r" (subtype), "r" (arg) + : "memory"); +} + +static int tf_set_cpu_boot_addr (int cpu,unsigned long boot_addr) +{ +tf_generic_smc (TF_SET_CPU_BOOT_ADDR_SMC , boot_addr,0 ); + +return 0 ;
arch/arm/firmware/trusted_foundations.c line 56
Quoted 10 lines
arch/arm/firmware/trusted_foundations.c new file +10
-0+} + +static const struct firmware_ops trusted_foundations_ops = { + .set_cpu_boot_addr = tf_set_cpu_boot_addr, +}; + +void register_trusted_foundations (const char *version) +{ +/* we are not using version information for now since currently +* supported SMCs are compatible with all TF releases */
I think it would be wise at least to sanity-check the version, even if there's only one major version supported for now.
Quoted 52 lines
arch/arm/firmware/trusted_foundations.c new file +2
-0+register_firmware_ops (&trusted_foundations_ops); +}arch/arm/include/asm/trusted_foundations.h new file +36
-0@@ -0,0 +1,36 @@ +/* +* Copyright (c) 2013, NVIDIA Corporation. +* +* This program is free software; you can redistribute it and/or modify +* it under the terms of the GNU General Public License as published by +* the Free Software Foundation; either version 2 of the License, or +* (at your option) any later version. +* +* This program is distributed in the hope that it will be useful, but WITHOUT +* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or +* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for +* more details. +*/ + +#ifndef __ASM_ARM_TRUSTED_FOUNDATIONS_H +#define __ASM_ARM_TRUSTED_FOUNDATIONS_H + +#include <linux/kconfig.h> +#include <linux/printk.h> +#include <asm/bug.h> + +#if IS_ENABLED (CONFIG_TRUSTED_FOUNDATIONS ) + +void register_trusted_foundations (const char *version); + +#else + +static inline void register_trusted_foundations (const char *version) +{ +pr_crit ("No support for Trusted Foundations, stopping...\n" ); +BUG (); +} + +#endif /* IS_ENABLED(CONFIG_TRUSTED_FOUNDATIONS) */ + +#endif Signature · 7 lines
1.8.3.4 _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel at lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
Re: [PATCH v3 1/5] ARM: add basic Trusted Foundations support
Quoted 18 lines · Dave Martin
Documentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt line 10
Quoted 12 lines · Alexandre Courbot
Trusted Foundations is a TrustZone-based secure monitor for ARM that can be invoked using a consistent smc-based API on all supported platforms. This patch adds initial basic support for Trusted Foundations using the ARM firmware API. Current features are limited to the ability to boot secondary processors.
Documentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt +4
-0+Required properties: +- compatible : "tl,trusted-foundations" +- version : Must contain the version number string of the Trusted Foundation + firmware.Are you sure there is no low-level way to probe vendor and version info? If there is, then the DT should describe nothing except the fact that the probe interface exists.
Quoted 3 lines
I also worry that two integrations on different SoCs might have the same version number, yet still be different due to vendor-specific features and options.
I would expect HW-specific compatible values also to be present in a DT. For example, perhaps:
compatible = "tl,trusted-foundations-nvidia-shield", "tl,trusted-foundations";
(nvidia vendor, shield board/implementation)
This would allow matching on the specific value "tl,trusted-foundations-nvidia-shield" in the future if some quirking was needed, but if this wasn't needed, drivers could just bind to the generic "tl,trusted-foundations".
Quoted 10 lines
Documentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt line 10
Quoted 2 lines
+- version : Must contain the version number string of the Trusted Foundation
Documentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt +1
-0+ firmware.Are you sure there is no low-level way to probe vendor and version info? If there is, then the DT should describe nothing except the fact that the probe interface exists.
I also worry that two integrations on different SoCs might have the same version number, yet still be different due to vendor-specific features and options.
Talking of the version - if we do need to represent this in the DT, how about 2 separate cells for major/minor version rather than encoding it into a string? Then, no parsing would be required.
Re: [PATCH v3 1/5] ARM: add basic Trusted Foundations support
Quoted 29 lines · Stephen Warren
Quoted 18 lines · Dave Martin
Documentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt line 10
Quoted 12 lines · Alexandre Courbot
Trusted Foundations is a TrustZone-based secure monitor for ARM that can be invoked using a consistent smc-based API on all supported platforms. This patch adds initial basic support for Trusted Foundations using the ARM firmware API. Current features are limited to the ability to boot secondary processors.
Documentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt +4
-0+Required properties: +- compatible : "tl,trusted-foundations" +- version : Must contain the version number string of the Trusted Foundation + firmware.Are you sure there is no low-level way to probe vendor and version info? If there is, then the DT should describe nothing except the fact that the probe interface exists.
Quoted 3 lines
I also worry that two integrations on different SoCs might have the same version number, yet still be different due to vendor-specific features and options.
I would expect HW-specific compatible values also to be present in a DT. For example, perhaps:
compatible = "tl,trusted-foundations-nvidia-shield", "tl,trusted-foundations";
Quoted 6 lines
(nvidia vendor, shield board/implementation)
This would allow matching on the specific value "tl,trusted-foundations-nvidia-shield" in the future if some quirking was needed, but if this wasn't needed, drivers could just bind to the generic "tl,trusted-foundations".
That seems reasonable *unless* there is a reliable way to obtain a vendor ID from the SMC ABI directly, in which case we should just use that.
One could debate whether the extra compatible string should have "nvidia," or "tl," but the fact that "nvidia" is in the name at all pretty much narrows it down.
Quoted 8 lines
Quoted 8 lines
Documentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt line 10
Quoted 3 lines
+- version : Must contain the version number string of the Trusted Foundation
Documentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt +1
-0+ firmware.Are you sure there is no low-level way to probe vendor and version info? If there is, then the DT should describe nothing except the fact that the probe interface exists.
Quoted 7 lines
Quoted 3 lines
I also worry that two integrations on different SoCs might have the same version number, yet still be different due to vendor-specific features and options.
Talking of the version - if we do need to represent this in the DT, how about 2 separate cells for major/minor version rather than encoding it into a string? Then, no parsing would be required.
I think the key thing here is to match whatever TF's native notion of version is.
If it's truly a string with specific comparison rules, we should leave it a string and write code to examine it. If it's a simple <major minor> pair, then putting it in the DT in this form makes sense.
And, as you previosly suggested, if there's a reliable ABI for getting the firmware to tell you its version number, we should just use that and not describe the version number in the DT at all.
I'm not in a position to answer that one.
If we really have no idea, we'll have to go with a best guess, and change the compatible string later if necessary. But it would be better to get it right if we can.
If TF has version/vendor probe interfaces but those might be missing/ broken/wrong, we could still specify "vendor" and "version" properties which allow the probed values to be overridden.
Cheers
---Dave
Re: [PATCH v3 1/5] ARM: add basic Trusted Foundations support
Quoted 61 lines · Dave Martin
Quoted 29 lines · Stephen Warren
Quoted 14 lines · Dave Martin
Documentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt line 10
Quoted 12 lines · Alexandre Courbot
Trusted Foundations is a TrustZone-based secure monitor for ARM that can be invoked using a consistent smc-based API on all supported platforms. This patch adds initial basic support for Trusted Foundations using the ARM firmware API. Current features are limited to the ability to boot secondary processors.
Documentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt +4
-0+Required properties: +- compatible : "tl,trusted-foundations" +- version : Must contain the version number string of the Trusted Foundation + firmware.Quoted 7 lines
Are you sure there is no low-level way to probe vendor and version info? If there is, then the DT should describe nothing except the fact that the probe interface exists.
I also worry that two integrations on different SoCs might have the same version number, yet still be different due to vendor-specific features and options.
I would expect HW-specific compatible values also to be present in a DT. For example, perhaps:
compatible = "tl,trusted-foundations-nvidia-shield", "tl,trusted-foundations";
Quoted 6 lines
(nvidia vendor, shield board/implementation)
This would allow matching on the specific value "tl,trusted-foundations-nvidia-shield" in the future if some quirking was needed, but if this wasn't needed, drivers could just bind to the generic "tl,trusted-foundations".
That seems reasonable *unless* there is a reliable way to obtain a vendor ID from the SMC ABI directly, in which case we should just use that.
One could debate whether the extra compatible string should have "nvidia," or "tl," but the fact that "nvidia" is in the name at all pretty much narrows it down.
Quoted 8 lines
Quoted 8 lines
Quoted 3 lines
+- version : Must contain the version number string of the Trusted Foundation + firmware.Are you sure there is no low-level way to probe vendor and version info? If there is, then the DT should describe nothing except the fact that the probe interface exists.
Quoted 7 lines
Quoted 3 lines
I also worry that two integrations on different SoCs might have the same version number, yet still be different due to vendor-specific features and options.
Talking of the version - if we do need to represent this in the DT, how about 2 separate cells for major/minor version rather than encoding it into a string? Then, no parsing would be required.
Quoted 6 lines
I think the key thing here is to match whatever TF's native notion of version is.
If it's truly a string with specific comparison rules, we should leave it a string and write code to examine it. If it's a simple <major minor> pair, then putting it in the DT in this form makes sense.
TF's native version is this major/minor pair, and indeed it should be sensible and harmless to turn it into a pair of cells.
Then there are other components that can be added to build a more precise "version string". You can have a look at this file to see how this is done:
The version string I used to far is the S_VERSION_MAIN macro, since it is the only one that seems to be platform-independant. S_VERSION_PLATFORM and S_VERSION_OS can be used to indentify if Android is used and which version. Then all these strings (plus a few others) are concatenated to build S_VERSION_STRING which is the ultimate version, but also considerably more complicated to handle.
At the current level of support, I don't think it makes sense to look further than S_VERSION_MAIN for the moment. If we switch it to a pair of integer cells as Stephen suggested, nothing prevents us to extend the bindings with other properties (for instance version-platform and version-os) if they become necessary. But at the moment I don't think it makes sense to overthink versions beyond the main version number, especially if it can be extended safely if needed in the future.
Thanks,
Alex.
Re: [PATCH v3 1/5] ARM: add basic Trusted Foundations support
In reply to Dave MartinQuoted 42 lines · Dave Martin
Quoted 29 lines · Alexandre Courbot
Trusted Foundations is a TrustZone-based secure monitor for ARM that can be invoked using a consistent smc-based API on all supported platforms. This patch adds initial basic support for Trusted Foundations using the ARM firmware API. Current features are limited to the ability to boot secondary processors.
Signed-off-by Alexandre Courbot <[email protected]>--- .../arm/firmware/tl,trusted-foundations.txt | 16 ++++++ .../devicetree/bindings/vendor-prefixes.txt | 1 + arch/arm/Kconfig | 2 + arch/arm/Makefile | 1 + arch/arm/firmware/Kconfig | 22 ++++++++ arch/arm/firmware/Makefile | 1 + arch/arm/firmware/trusted_foundations.c | 58 ++++++++++++++++++++++ arch/arm/include/asm/trusted_foundations.h | 36 ++++++++++++++ 8 files changed, 137 insertions(+) create mode 100644 Documentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt create mode 100644 arch/arm/firmware/Kconfig create mode 100644 arch/arm/firmware/Makefile create mode 100644 arch/arm/firmware/trusted_foundations.c create mode 100644 arch/arm/include/asm/trusted_foundations.hDocumentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt new file +0
-0Documentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt line 5
Quoted 5 lines
Documentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt new file +5
-0@@ -0,0 +1,16 @@ +Trusted Foundations + +Boards that use the Trusted Foundations secure monitor can signal its +presence by declaring a node compatible with "tl,trusted-foundations" +(the name and location of the node does not matter).We shouldn't encougrage people to arrange DT nodes at random.
Unless there's a good reason not to, this should just be a child of /, with the meaning that the firmware is callable (at least to the point of discoverability) on all CPUs.
Quoted 6 lines
The meaning of the node should be left unspecified if it is in any other location.
This is the convention already adopted for things like arm,psci (though to be fair, the arm,psci binding documentation fails to mention that also -- nonetheless, the node is placed in /).
I agree the node should be a child of /. At the moment this is not enforced at runtime though (because of the use of of_find_compatible_node()), is it ok if this is just specified as a rule in the bindings?
Quoted 9 lines
Documentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt line 10
Quoted 5 lines
Documentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt new file +5
-0+ +Required properties: +- compatible : "tl,trusted-foundations" +- version : Must contain the version number string of the Trusted Foundation + firmware.Are you sure there is no low-level way to probe vendor and version info? If there is, then the DT should describe nothing except the fact that the probe interface exists.
I have looked for such a probing method myself and asked some people who know better than me, and the answer was that there was none, unfortunately. :(
Quoted 3 lines
I also worry that two integrations on different SoCs might have the same version number, yet still be different due to vendor-specific features and options.
I also investigated that point (since it was raised for v2 as well) and the answer was that for TF as least, the interface should not be different across vendors.
Quoted 102 lines
Quoted 102 lines
Documentation/devicetree/bindings/arm/firmware/tl,trusted-foundations.txt new file +6
-0+ +Example: + firmware { + compatible = "tl,trusted-foundations"; + version = "02.08"; + };Documentation/devicetree/bindings/vendor-prefixes.txt +1
-0@@ -63,6 +63,7 @@ ste ST-Ericsson stericsson ST-Ericsson toumaz Toumaz ti Texas Instruments +tl Trusted Logic toshiba Toshiba Corporation v3 V3 Semiconductor via VIA Technologies, Inc.arch/arm/Kconfig +2
-0@@ -1074,6 +1074,8 @@ config ARM_TIMER_SP804 select CLKSRC_MMIO select CLKSRC_OF if OF +source "arch/arm/firmware/Kconfig" + source arch/arm/mm/Kconfig config ARM_NR_BANKSarch/arm/Makefile +1
-0@@ -267,6 +267,7 @@ core-$(CONFIG_KVM_ARM_HOST) += arch/arm/kvm/ core-y += arch/arm/kernel/ arch/arm/mm/ arch/arm/common/ core-y += arch/arm/net/ core-y += arch/arm/crypto/ +core-y += arch/arm/firmware/ core-y += $(machdirs) $(platdirs) drivers-$(CONFIG_OPROFILE) += arch/arm/oprofile/arch/arm/firmware/Kconfig new file +22
-0@@ -0,0 +1,22 @@ +config ARCH_SUPPORTS_TRUSTED_FOUNDATIONS + bool + +menu "Firmware options" + depends on ARCH_SUPPORTS_TRUSTED_FOUNDATIONS + +config TRUSTED_FOUNDATIONS + bool "Trusted Foundations secure monitor support" + depends on ARCH_SUPPORTS_TRUSTED_FOUNDATIONS + help + Some devices are booted with the Trusted Foundations secure monitor + active, requiring some core operations to be performed by the secure + monitor instead of the kernel. + + This option allows the kernel to invoke the secure monitor whenever + required on devices using Trusted Foundations. + + Devices using Trusted Foundations should pass a device tree containing + a node compatible with "tl,trusted-foundations" to signal the presence + of the secure monitor. + +endmenuarch/arm/firmware/Makefile new file +1
-0@@ -0,0 +1 @@ +obj-$(CONFIG_TRUSTED_FOUNDATIONS) += trusted_foundations.oarch/arm/firmware/trusted_foundations.c new file +17
-0@@ -0,0 +1,58 @@ +/* +* Trusted Foundations support for ARM CPUs +* +* Copyright (c) 2013, NVIDIA Corporation. +* +* This program is free software; you can redistribute it and/or modify +* it under the terms of the GNU General Public License as published by +* the Free Software Foundation; either version 2 of the License, or +* (at your option) any later version. +* +* This program is distributed in the hope that it will be useful, but WITHOUT +* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or +* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for +* more details. +*/ + +#include <linux/kernel.h>
Quoted 8 lines
arch/arm/firmware/trusted_foundations.c line 23
Quoted 6 lines
arch/arm/firmware/trusted_foundations.c new file +6
-0+#include <linux/init.h> +#include <linux/of.h> +#include <asm/firmware.h> + +#define TF_SET_CPU_BOOT_ADDR_SMC 0xfffff200 +Better to use
__nakedinstead of an explicit attribute.
Will fix, thanks.
Quoted 29 lines
Quoted 29 lines
arch/arm/firmware/trusted_foundations.c new file +28
-0+static void __attribute__((naked))tf_generic_smc (u32 type,u32 subtype, +u32 arg) +{ + asmvolatile ( +".arch_extension sec\n\t" +"stmfd sp!, {r4 - r11, lr}\n\t" + __asmeq("%0", "r0") + __asmeq("%1", "r1") + __asmeq("%2", "r2") +"mov r3, #0\n\t" +"mov r4, #0\n\t" +"smc #0\n\t" +"ldmfd sp!, {r4 - r11, pc}" + : + : "r" (type), "r" (subtype), "r" (arg) + : "memory"); +} + +static int tf_set_cpu_boot_addr (int cpu,unsigned long boot_addr) +{ +tf_generic_smc (TF_SET_CPU_BOOT_ADDR_SMC , boot_addr,0 ); + +return 0 ; +} + +static const struct firmware_ops trusted_foundations_ops = { + .set_cpu_boot_addr = tf_set_cpu_boot_addr, +};
Quoted 8 lines
arch/arm/firmware/trusted_foundations.c line 56
Quoted 5 lines
arch/arm/firmware/trusted_foundations.c new file +5
-0+ +void register_trusted_foundations (const char *version) +{ +/* we are not using version information for now since currently +* supported SMCs are compatible with all TF releases */ I think it would be wise at least to sanity-check the version, even if there's only one major version supported for now.
Ok.
Thanks,
Alex.
PATCH v3 2/5 ARM: tegra: add support for Trusted Foundations
In reply to Alexandre CourbotRegister the firmware operations for Trusted Foundations if the device tree indicates it is active on the device.
--- Documentation/devicetree/bindings/arm/tegra.txt | 5 +++ arch/arm/mach-tegra/Kconfig | 1 + arch/arm/mach-tegra/Makefile | 1 + arch/arm/mach-tegra/common.c | 2 ++ arch/arm/mach-tegra/firmware.c | 41 +++++++++++++++++++++++++ arch/arm/mach-tegra/firmware.h | 19 ++++++++++++ 6 files changed, 69 insertions(+) create mode 100644 arch/arm/mach-tegra/firmware.c create mode 100644 arch/arm/mach-tegra/firmware.h
Documentation/devicetree/bindings/arm/tegra.txt +5 -0
@@ -32,3 +32,8 @@ board-specific compatible values: nvidia,whistler toradex,colibri_t20-512 toradex,iris + +Trusted Foundations +------------------------------------------- +Tegra supports the Trusted Foundation secure monitor. See the +"tl,trusted-foundations" binding for more details.
arch/arm/mach-tegra/Kconfig +1 -0
@@ -2,6 +2,7 @@ config ARCH_TEGRA bool "NVIDIA Tegra" if ARCH_MULTI_V7 select ARCH_HAS_CPUFREQ select ARCH_REQUIRE_GPIOLIB + select ARCH_SUPPORTS_TRUSTED_FOUNDATIONS select CLKDEV_LOOKUP select CLKSRC_MMIO select CLKSRC_OF
arch/arm/mach-tegra/Makefile +1 -0
@@ -12,6 +12,7 @@ obj-y += pm.o obj-y += reset.o obj-y += reset-handler.o obj-y += sleep.o +obj-y += firmware.o obj-y += tegra.o obj-$(CONFIG_CPU_IDLE) += cpuidle.o obj-$(CONFIG_ARCH_TEGRA_2x_SOC) += tegra20_speedo.o
arch/arm/mach-tegra/common.c +2 -0
@@ -39,6 +39,7 @@#include "sleep.h" #include "pm.h" #include "reset.h" +#include "firmware.h" /* * Storage for debug-macro.S's state.
@@ -99,6 +100,7 @@ static void __init tegra_init_cache(void)void __init tegra_init_early (void ) { +tegra_init_firmware ();tegra_cpu_reset_handler_init ();tegra_apb_io_init ();tegra_init_fuse ();
arch/arm/mach-tegra/firmware.c new file +41 -0
41 lines not on this page · raw message
arch/arm/mach-tegra/firmware.h new file +19 -0
@@ -0,0 +1,19 @@ +/* +* Copyright (c) 2013, NVIDIA Corporation. All rights reserved. +* +* This program is free software; you can redistribute it and/or modify it +* under the terms and conditions of the GNU General Public License, +* version 2, as published by the Free Software Foundation. +* +* This program is distributed in the hope it will be useful, but WITHOUT +* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or +* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for +* more details. +*/ + +#ifndef __TEGRA_FIRMWARE_H +#define __TEGRA_FIRMWARE_H + +void tegra_init_firmware (void ); + +#endif
Signature · 1 line
1.8.3.4
Re: [PATCH v3 2/5] ARM: tegra: add support for Trusted Foundations
arch/arm/mach-tegra/firmware.c line 25
Quoted 11 lines · Alexandre Courbot
Register the firmware operations for Trusted Foundations if the device tree indicates it is active on the device.
arch/arm/mach-tegra/firmware.c +6
-0+void __init tegra_init_firmware (void ) +{ +struct device_node *node; + +if (!of_have_populated_dt ()) +return ;
That's never false on Tegra, at least upstream.
Quoted 5 lines
arch/arm/mach-tegra/firmware.c +5
-0+ node =of_find_compatible_node (NULL ,NULL ,"tl,trusted-foundations" ); +if (node) { +const char *tf_version; +int err; +
arch/arm/mach-tegra/firmware.c line 41
Quoted 10 lines
arch/arm/mach-tegra/firmware.c +10
-0+ err =of_property_read_string (node,"version" , &tf_version); +if (err !=0 ) { +pr_crit ("Cannot read Trusted Foundation version: %d\n" , + err); +BUG (); + }else { +register_trusted_foundations (tf_version); + } + } +}
I think most/all of that function will be common across all SoCs that support the TF monitor. I think you want to move the code into register_trusted_foundations() (or a function that calls that) in patch 1?
Re: [PATCH v3 2/5] ARM: tegra: add support for Trusted Foundations
Quoted 6 lines · Stephen Warren
Quoted 4 lines · Alexandre Courbot
Register the firmware operations for Trusted Foundations if the device tree indicates it is active on the device.
arch/arm/mach-tegra/firmware.c +0
-0
Quoted 8 lines
arch/arm/mach-tegra/firmware.c line 25
Quoted 6 lines
arch/arm/mach-tegra/firmware.c +6
-0+void __init tegra_init_firmware (void ) +{ +struct device_node *node; + +if (!of_have_populated_dt ()) +return ;That's never false on Tegra, at least upstream.
Will remove, thanks.
Quoted 11 lines
Quoted 11 lines
arch/arm/mach-tegra/firmware.c +11
-0+ node =of_find_compatible_node (NULL ,NULL ,"tl,trusted-foundations" ); +if (node) { +const char *tf_version; +int err; + + err =of_property_read_string (node,"version" , &tf_version); +if (err !=0 ) { +pr_crit ("Cannot read Trusted Foundation version: %d\n" , + err); +BUG (); + }else {
Quoted 8 lines
arch/arm/mach-tegra/firmware.c line 41
Quoted 4 lines
arch/arm/mach-tegra/firmware.c +4
-0+register_trusted_foundations (tf_version); + } + } +}I think most/all of that function will be common across all SoCs that support the TF monitor. I think you want to move the code into
register_trusted_foundations()(or a function that calls that) in patch 1?
Agreed. If we push the logic a bit further, we should also have some platform parameters for non-DT users, along with an of_register_trusted_foundations() for those who prefer DT.
Thanks,
Alex.
PATCH v3 3/5 ARM: tegra: split setting of CPU reset handler
In reply to Alexandre CourbotNot all Tegra devices can set the CPU reset handler in the same way. In particular, devices using a TrustZone secure monitor cannot set the reset handler directly and need to do it through a firmware operation.
This patch separates the act of setting the reset handler from its preparation, so the former can be implemented in a different way.
--- arch/arm/mach-tegra/reset.c | 27 +++++++++++++++++---------- 1 file changed, 17 insertions(+), 10 deletions(-)
arch/arm/mach-tegra/reset.c +17 -10
@@ -33,26 +33,18 @@static bool is_enabled; -static void __init tegra_cpu_reset_handler_enable (void ) +static void __init tegra_cpu_reset_handler_set (const u32 reset_address) { -void __iomem *iram_base =IO_ADDRESS (TEGRA_IRAM_RESET_BASE );void __iomem *evp_cpu_reset =IO_ADDRESS (TEGRA_EXCEPTION_VECTORS_BASE +0x100 );void __iomem *sb_ctrl =IO_ADDRESS (TEGRA_SB_BASE );u32 reg; -BUG_ON (is_enabled); -BUG_ON (tegra_cpu_reset_handler_size >TEGRA_IRAM_RESET_HANDLER_SIZE ); - -memcpy (iram_base, (void *)__tegra_cpu_reset_handler_start, - tegra_cpu_reset_handler_size); -/* * NOTE: This must be the one and only write to the EVP CPU reset * vector in the entire system. */ -writel (TEGRA_IRAM_RESET_BASE + tegra_cpu_reset_handler_offset, - evp_cpu_reset); +writel (reset_address, evp_cpu_reset);wmb (); reg =readl (evp_cpu_reset);
@@ -66,6 +58,21 @@ static void __init tegra_cpu_reset_handler_enable(void)writel (reg, sb_ctrl);wmb (); } +} + +static void __init tegra_cpu_reset_handler_enable (void ) +{ +void __iomem *iram_base =IO_ADDRESS (TEGRA_IRAM_RESET_BASE ); +const u32 reset_address =TEGRA_IRAM_RESET_BASE + + tegra_cpu_reset_handler_offset; + +BUG_ON (is_enabled); +BUG_ON (tegra_cpu_reset_handler_size >TEGRA_IRAM_RESET_HANDLER_SIZE ); + +memcpy (iram_base, (void *)__tegra_cpu_reset_handler_start, + tegra_cpu_reset_handler_size); + +tegra_cpu_reset_handler_set (reset_address); is_enabled = true; }
Signature · 1 line
1.8.3.4
PATCH v3 4/5 ARM: tegra: set CPU reset handler with firmware op
In reply to Alexandre CourbotUse a firmware operation to set the CPU reset handler and only resort to doing it ourselves if there is none defined.
This supports the booting of secondary CPUs on devices using a TrustZone secure monitor.
--- arch/arm/mach-tegra/reset.c | 17 ++++++++++++++--- 1 file changed, 14 insertions(+), 3 deletions(-)
arch/arm/mach-tegra/reset.c +14 -3
@@ -21,6 +21,7 @@#include <asm/cacheflush.h> #include <asm/hardware/cache-l2x0.h> +#include <asm/firmware.h> #include "iomap.h" #include "irammap.h"
@@ -65,6 +66,7 @@ static void __init tegra_cpu_reset_handler_enable(void)void __iomem *iram_base =IO_ADDRESS (TEGRA_IRAM_RESET_BASE );const u32 reset_address =TEGRA_IRAM_RESET_BASE + tegra_cpu_reset_handler_offset; +int err;BUG_ON (is_enabled);BUG_ON (tegra_cpu_reset_handler_size >TEGRA_IRAM_RESET_HANDLER_SIZE );
@@ -72,9 +74,18 @@ static void __init tegra_cpu_reset_handler_enable(void)memcpy (iram_base, (void *)__tegra_cpu_reset_handler_start, tegra_cpu_reset_handler_size); -tegra_cpu_reset_handler_set (reset_address); - - is_enabled = true; + err =call_firmware_op (set_cpu_boot_addr,0 , reset_address); +switch (err) { +case -ENOSYS : +tegra_cpu_reset_handler_set (reset_address); +/* pass-through */ +case 0 : + is_enabled = true; +break ; +default : +pr_crit ("Cannot set CPU reset handler: %d\n" , err); +BUG (); + } }void __init tegra_cpu_reset_handler_init (void )
Signature · 1 line
1.8.3.4
Re: [PATCH v3 4/5] ARM: tegra: set CPU reset handler with firmware op
Quoted 10 lines · Alexandre Courbot
Use a firmware operation to set the CPU reset handler and only resort to doing it ourselves if there is none defined.
This supports the booting of secondary CPUs on devices using a TrustZone secure monitor.
arch/arm/mach-tegra/reset.c +2
-0+ err =call_firmware_op (set_cpu_boot_addr,0 , reset_address); +switch (err) {
arch/arm/mach-tegra/reset.c static void __init tegra_cpu_reset_handler_enable(void) · line 88
Quoted 10 lines
arch/arm/mach-tegra/reset.c +10
-0+case -ENOSYS : +tegra_cpu_reset_handler_set (reset_address); +/* pass-through */ +case 0 : + is_enabled = true; +break ; +default : +pr_crit ("Cannot set CPU reset handler: %d\n" , err); +BUG (); + }
Instead of trying and failing, does it make sense to register tegra_cpu_reset_handler_set() as the set_cpu_boot_addr firmware op when there is no firmware present? That would simplify all call-sites of any firmware op.
Re: [PATCH v3 4/5] ARM: tegra: set CPU reset handler with firmware op
Quoted 22 lines · Stephen Warren
Quoted 10 lines · Alexandre Courbot
Use a firmware operation to set the CPU reset handler and only resort to doing it ourselves if there is none defined.
This supports the booting of secondary CPUs on devices using a TrustZone secure monitor.
arch/arm/mach-tegra/reset.c +2
-0+ err =call_firmware_op (set_cpu_boot_addr,0 , reset_address); +switch (err) {arch/arm/mach-tegra/reset.c static void __init tegra_cpu_reset_handler_enable(void) · line 88
Quoted 10 lines
arch/arm/mach-tegra/reset.c +10
-0+case -ENOSYS : +tegra_cpu_reset_handler_set (reset_address); +/* pass-through */ +case 0 : + is_enabled = true; +break ; +default : +pr_crit ("Cannot set CPU reset handler: %d\n" , err); +BUG (); + }
Quoted 4 lines
Instead of trying and failing, does it make sense to register
tegra_cpu_reset_handler_set()as theset_cpu_boot_addrfirmware op when there is no firmware present? That would simplify all call-sites of any firmware op.
We discussed that point in v2 already IIRC (that is some time ago). The reason I did it this way is because I wanted to follow the way Tomasz was using his interface with Exynos - it seemed appropriate that all users of an interface use it the same way. But if you prefer to use a "non-firmware" firmware_ops for Tegra, I have absolutely nothing against this.
Thanks,
Alex.
PATCH v3 5/5 ARM: tegra: support Trusted Foundations by default
In reply to Alexandre CourbotSupport for Trusted Foundations is light and allows the kernel to run on a wider range of devices, so enable it by default.
--- arch/arm/configs/tegra_defconfig | 1 + 1 file changed, 1 insertion(+)
arch/arm/configs/tegra_defconfig +1 -0
@@ -28,6 +28,7 @@ CONFIG_ARCH_TEGRA_3x_SOC=y CONFIG_ARCH_TEGRA_114_SOC=y CONFIG_TEGRA_PCI=y CONFIG_TEGRA_EMC_SCALING_ENABLE=y +CONFIG_TRUSTED_FOUNDATIONS=y CONFIG_SMP=y CONFIG_PREEMPT=y CONFIG_AEABI=y
Signature · 1 line
1.8.3.4