Age | Commit message (Collapse) | Author |
|
Add common platform components to be used by all
supported Marvell platforms.
Change-Id: Ie5f70fc7ff668c8f8073d6dd936458f35890ebcf
Signed-off-by: Haim Boot <hayim@marvell.com>
Signed-off-by: Konstantin Porotchkin <kostap@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/38133
|
|
Add Marvell flash image builder "doimage"
The "doimage" tool is used for building a boot image
with all headers and extensions acceptable by the SoC
BootROM.
It also is able to sign and encrypt the image for
trusted boot mode as far as verify and print the image
content information.
This "doimage" release supports Marvell A7K/A8K SoCs.
Change-Id: I5ad7377abdcbccb77cb4d71e5e4e4cd4358303e7
Signed-off-by: Haim Boot <hayim@marvell.com>
Signed-off-by: Yehuda Yitschak <yehuday@marvell.com>
Signed-off-by: Konstantin Porotchkin <kostap@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/37917
|
|
Add Marvell drivers for A7K and A8K SoC families
Change-Id: I7fb530e543e4f64782f41f8a058b6aabd82af9e8
Signed-off-by: Haim Boot <hayim@marvell.com>
Signed-off-by: Konstantin Porotchkin <kostap@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/37916
|
|
Add BLE stage that runs in BootROM context and used by
Marvell platforms for DRAM initialization and training.
Change-Id: I989b7298c47446824b25c4b68d27bd2379462035
Signed-off-by: Haim Boot <hayim@marvell.com>
Signed-off-by: Yehuda Yitschak <yehuday@marvell.com>
Signed-off-by: Konstantin Porotchkin <kostap@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/37915
|
|
Allow zero-address assertion bypass if requested by platform
compilation flag PLAT_ALLOW_ZERO_ADDR_COPY.
This patch allows loading BL33 (u-boot) to address 0x0
Change-Id: I10518db13466017110358437790ce5212d52d9e6
Signed-off-by: Haim Boot <hayim@marvell.com>
Signed-off-by: Konstantin Porotchkin <kostap@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/37914
|
|
Add ability to get timer value in micro- and milliseconds.
Change-Id: I864619c23fe8a2a01cef86795e39d1b0bfd2e3d3
Signed-off-by: Victor Axelrod <victora@marvell.com>
Signed-off-by: Ofir Fedida <ofedida@marvell.com>
Signed-off-by: Konstantin Porotchkin <kostap@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/37913
Reviewed-by: Haim Boot <hayim@marvell.com>
|
|
Please do not push this patch to the mainline, it is
intended to be used with Marvell CI only.
Add checkpatch.pl script to the source tree and modify
the checkpatch settings provided by the mainline
(checkpatch.conf)
The mainline configuration is created for unmodified
checkpatch from latest kernel sources. It includes
option --showfile which is not supported by the regular
u-boot/ATF checkpatch script.
Change-Id: Id36a49b7844e08751214b55db602b31621ef0b3c
Signed-off-by: Konstantin Porotchkin <kostap@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/37920
Reviewed-by: Haim Boot <hayim@marvell.com>
|
|
This reverts commit 79627dc37259781e578c47e1e63856dd0424b2a2.
It is required to allow build of ATF with MGCC and Yocto toolchain.
Change-Id: I7bfcf46e977c05cc988d93a2d1a22a9d0cc512d3
Reviewed-on: http://vgitil04.il.marvell.com:8080/37435
Reviewed-by: Hua Jing <jinghua@marvell.com>
Reviewed-by: Haim Boot <hayim@marvell.com>
Tested-by: Haim Boot <hayim@marvell.com>
|
|
Final updates for v1.3 release
|
|
Release v1.3: Minor updates to user guide
|
|
Updated the user guide to clarify building FIP for AArch32.
The instructions were previously specific to building a FIP for AArch64.
Change-Id: I7bd1a6b8e810cfda411f707e04f479006817858e
Signed-off-by: David Cunado <david.cunado@arm.com>
|
|
Update the release notes (readme.md) for the ARM Trusted Firmware
v1.3 release.
Change-Id: Ia1f4eb1897e63eeab7d69a593ba0ad91d50043f5
Signed-off-by: Dan Handley <dan.handley@arm.com>
|
|
Change-Id: I05991543d28e70b67be600b714990af6a8d7ba29
|
|
Updated change-log.md with summary of changes since release v1.2.
Change-Id: Ia1e18ff4b0da567cf12dfcb53e6317e995100bdf
|
|
PMF: Add documentation
|
|
Add a Performance Measurement Framework (PMF) section
to the firmware design document.
Change-Id: I5953bd3b1067501f190164c8827d2b0d8022fc0b
Signed-off-by: dp-arm <dimitris.papastamos@arm.com>
|
|
Fix documentation of bootwrapper boot on juno
|
|
The user guide incorrectly claimed that it is possible to load a
bootwrapped kernel over JTAG on Juno in the same manner as an EL3
payload. In the EL3 payload boot flow, some of the platform
initialisations in BL2 are modified. In particular, the TZC settings
are modified to allow unrestricted access to DRAM. This in turn allows
the debugger to access the DRAM and therefore to load the image there.
In the BL33-preloaded boot flow though, BL2 uses the default TZC
programming, which prevent access to most of the DRAM from secure state.
When execution reaches the SPIN_ON_BL1_EXIT loop, the MMU is disabled
and thus DS-5 presumably issues secure access transactions while trying
to load the image, which fails.
One way around it is to stop execution at the end of BL2 instead. At
this point, the MMU is still enabled and the DRAM is mapped as
non-secure memory. Therefore, the debugger is allowed to access this
memory in this context and to sucessfully load the bootwrapped kernel in
DRAM. The user guide is updated to suggest this alternative method.
Co-Authored-By: Sandrine Bailleux <sandrine.bailleux@arm.com>
Signed-off-by: Dan Handley <dan.handley@arm.com>
Change-Id: I537ea1c6d2f96edc06bc3f512e770c748bcabe94
|
|
AArch32: Update firmware-design.md
|
|
AArch32: Update user-guide and add DTBs
|
|
Fix GICv3 DT to include psci system off/reset
|
|
Docs: Rename duplicate title in porting guide
|
|
This patch adds necessary updates for building and running Trusted
Firmware for AArch32 to user-guide.md. The instructions for running
on both `FVP_Base_AEMv8A-AEMv8A` in AArch32 mode and
`FVP_Base_Cortex-A32x4` models are added. The device tree files for
AArch32 Linux kernel are also added in the `fdts` folder.
Change-Id: I0023b6b03e05f32637cb5765fdeda8c8df2d0d3e
|
|
This patch updates the firmware-design.md for AArch32 related changes.
Change-Id: Idf392a44861ab9c1f59f3de4f3435f508b17c678
|
|
Fix one of the two titles that ended up being the same, although both
describe different things.
Change-Id: I66ecf369643709898ee4c014659d8f85c0480643
|
|
The `fvp-base-gicv3-psci` and `fvp-foundation-gicv3-psci` device tree source
files did not have psci node entries for `system off` and `system reset`.
Also the DTS files included `rtsm_ve-motherboard-no_psci.dtsi` instead of
`rtsm_ve-motherboard.dtsi`. As a result, the Linux kernel failed to invoke
the PSCI_SYSTEM_OFF/RESET API when being shutdown/reset. This patch corrects
this problem and also updates the corresponding DTB files.
This patch also removes `rtsm_ve-motherboard-no_psci.dtsi` and
`fvp-foundation-motherboard-no_psci.dtsi` files as they are no longer used.
Change-Id: I8ba61a1323035f7508cae663bb490ac0e8a64618
|
|
AArch32: Add `memcpy4` function in assembly
|
|
Docs: Add the PSCI library integration guide
|
|
Drop requirement for CLA in contribution.md
|
|
This patch adds the PSCI library integration guide for AArch32 ARMv8-A
systems `psci-lib-integration-guide.md` to the documentation. The
patch also adds appropriate reference to the new document in
the `firmware-design.md` document.
Change-Id: I2d5b5c6b612452371713399702e318e3c73a8ee0
|
|
At present the `el3_entrypoint_common` macro uses `memcpy`
function defined in lib/stdlib/mem.c file, to copy data
from ROM to RAM for BL1. Depending on the compiler being
used the stack could potentially be used, in `memcpy`,
for storing the local variables. Since the stack is
initialized much later in `el3_entrypoint_common` it
may result in unknown behaviour.
This patch adds `memcpy4` function definition in assembly so
that it can be used before the stack is initialized and it
also replaces `memcpy` by `memcpy4` in `el3_entrypoint_common`
macro, to copy data from ROM to RAM for BL1.
Change-Id: I3357a0e8095f05f71bbbf0b185585d9499bfd5e0
|
|
It is no longer necessary for contributors to send a CLA to ARM
before making contributions. Contributors must instead add a
"Signed-off-by:" line to each commit, which certifies that the
contribution is made under the Developer Certificate of Origin
(DCO).
Update contributing.md to reflect this new policy and add a copy of
the DCO to the repository.
Change-Id: I7ca98bffc3bf57e8bdd51d763c24f13e415a328b
Signed-off-by: Dan Handley <dan.handley@arm.com>
|
|
Upgrade Linaro release, FVPs and mbed TLS versions
|
|
This patch updates the User Guide to recommend the latest version
of some of the software dependencies of ARM Trusted Firmware.
- Upgrade Linaro release: 16.02 -> 16.06
- Upgrade FVPs
- Foundation v8 FVP: 9.5 -> 10.1
- Base FVPs: 7.6 -> 7.7
- Upgrade mbed TLS library: 2.2.0 -> 2.2.1
Note that the latest release of mbed TLS as of today is 2.3.0 but it has
compilations issues with the set of library configuration options that
Trusted Firmware uses. 2.2.1 is the next most recent release known to
build with TF.
This patch also fixes the markdown formatting of a link in the
User Guide.
Change-Id: Ieb7dd336f4d3110fba060afec4ad580ae707a8f1
|
|
Whitelist version 9.6 of Foundation FVP
|
|
AArch32: Fix detection of virtualization support
|
|
The Virtualization field in the ID_PFR1 register has only 2
valid values (0 or 1) but it was incorrectly checked against
unrelated value tied to the SPSR register instead.
This patch fixes the detection of virtualization support by
using the valid values in BL1 context management code.
Change-Id: If12592e343770e1da90f0f5fecf0a3376047ac29
|
|
This prevents a warning being emitted in the console during FVP
configuration setup when using the Foundation FVP 9.6 onwards.
Change-Id: I685b8bd0dbd0119af4b0cb3f7d708fcc08e99561
|
|
Introduce PSCI Library argument structure
|
|
This patch moves the invocation of `psci_setup()` from BL31 and SP_MIN
into `std_svc_setup()` as part of ARM Standard Service initialization.
This allows us to consolidate ARM Standard Service initializations which
will be added to in the future. A new function `get_arm_std_svc_args()`
is introduced to get arguments corresponding to each standard service.
This function must be implemented by the EL3 Runtime Firmware and both
SP_MIN and BL31 implement it.
Change-Id: I38e1b644f797fa4089b20574bd4a10f0419de184
|
|
This patch introduces a `psci_lib_args_t` structure which must be
passed into `psci_setup()` which is then used to initialize the PSCI
library. The `psci_lib_args_t` is a versioned structure so as to enable
compatibility checks during library initialization. Both BL31 and SP_MIN
are modified to use the new structure.
SP_MIN is also modified to add version string and build message as part
of its cold boot log just like the other BLs in Trusted Firmware.
NOTE: Please be aware that this patch modifies the prototype of
`psci_setup()`, which breaks compatibility with EL3 Runtime Firmware
(excluding BL31 and SP_MIN) integrated with the PSCI Library.
Change-Id: Ic3761db0b790760a7ad664d8a437c72ea5edbcd6
|
|
Add basic AArch32 support for BL1 & BL2
|
|
This patch adds ARM Cortex-A32 MPCore Processor support
in the CPU specific operations framework. It also includes
this support for the Base FVP port.
Change-Id: If3697b88678df737c29f79cf3fa1ea2cb6fa565d
|
|
This patch adds support in SP_MIN to receive generic and
platform specific arguments from BL2.
The new signature is as following:
void sp_min_early_platform_setup(void *from_bl2,
void *plat_params_from_bl2);
ARM platforms have been modified to use this support.
Note: Platforms may break if using old signature.
Default value for RESET_TO_SP_MIN is changed to 0.
Change-Id: I008d4b09fd3803c7b6231587ebf02a047bdba8d0
|
|
This patch uses the `el3_entrypoint_common` macro to initialize
CPU registers, in SP_MIN entrypoint.s file, in both cold and warm
boot path. It also adds conditional compilation, in cold and warm
boot entry path, based on RESET_TO_SP_MIN.
Change-Id: Id493ca840dc7b9e26948dc78ee928e9fdb76b9e4
|
|
This patch adds ARM platform changes in BL2 for AArch32 state.
It instantiates a descriptor array for ARM platforms describing
image and entrypoint information for `SCP_BL2`, `BL32` and `BL33`.
It also enables building of BL2 for ARCH=aarch32.
Change-Id: I60dc7a284311eceba401fc789311c50ac746c51e
|
|
This patch adds generic changes in BL2 to support AArch32 state.
New AArch32 specific assembly/C files are introduced and
some files are moved to AArch32/64 specific folders.
BL2 for AArch64 is refactored but functionally identical.
BL2 executes in Secure SVC mode in AArch32 state.
Change-Id: Ifaacbc2a91f8640876385b953adb24744d9dbde3
|
|
This patch adds ARM platform changes in BL1 for AArch32 state.
It also enables building of BL1 for ARCH=aarch32.
Change-Id: I079be81a93d027f37b0f7d8bb474b1252bb4cf48
|
|
This patch adds generic changes in BL1 to support AArch32 state.
New AArch32 specific assembly/C files are introduced and
some files are moved to AArch32/64 specific folders.
BL1 for AArch64 is refactored but functionally identical.
BL1 executes in Secure Monitor mode in AArch32 state.
NOTE: BL1 in AArch32 state ONLY handles BL1_RUN_IMAGE SMC.
Change-Id: I6e2296374c7efbf3cf2aa1a0ce8de0732d8c98a5
|
|
This patch adds common changes to support AArch32 state in
BL1 and BL2. Following are the changes:
* Added functions for disabling MMU from Secure state.
* Added AArch32 specific SMC function.
* Added semihosting support.
* Added reporting of unhandled exceptions.
* Added uniprocessor stack support.
* Added `el3_entrypoint_common` macro that can be
shared by BL1 and BL32 (SP_MIN) BL stages. The
`el3_entrypoint_common` is similar to the AArch64
counterpart with the main difference in the assembly
instructions and the registers that are relevant to
AArch32 execution state.
* Enabled `LOAD_IMAGE_V2` flag in Makefile for
`ARCH=aarch32` and added check to make sure that
platform has not overridden to disable it.
Change-Id: I33c6d8dfefb2e5d142fdfd06a0f4a7332962e1a3
|