When mainline kernels reboot TK1 they use SW_RESET,
that reset mode does not reset PMIC. Some rails
need to be off for RAM Re-repair to work correctly.
Reviewed-by: Oleksandr Suvorov <oleksandr.suvorov@toradex.com>
Signed-off-by: Dominik Sliwa <dominik.sliwa@toradex.com>
Signed-off-by: Igor Opaniuk <igor.opaniuk@toradex.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
By default, CONFIG_SYS_INIT_SP_BSS_OFFSET was made invisible by not
giving a prompt to it.
The only way to define it is to hard-code an extra entry in SoC/board
Kconfig, like arch/arm/mach-tegra/tegra{186,210}/Kconfig.
Add a prompt to it in order to allow defconfig files to specify the
value of CONFIG_SYS_INIT_SP_BSS_OFFSET.
With this, CONFIG_SYS_INIT_SP_BSS_OFFSET would become always visible.
So, we need a new bool option to turn it off by default.
I move the 'default 524288' to the common place. This value is not too
big, but is big enough to avoid the overwrap of DT in most platforms.
If 512KB is not a suitable choice for your platform, you can change it
from your defconfig or menuconfig etc.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Acked-by: Stephen Warren <swarren@nvidia.com>
Note that U-Boot is always chainloaded from cboot starting with L4T
release 28. cboot always loads U-Boot to a fixed address, so making
the builds position independent isn't strictly necessary. However,
position independent builds can be convenient because if U-Boot is
ever loaded to an address different from its link address, it will
still be able to boot.
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Read the boot arguments passed by cboot via the /chosen/bootargs
property and store it in the cbootargs environment variable.
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
This function will attempt to look up an Ethernet address in the DTB
that was passed in from cboot. It does so by first trying to locate the
default Ethernet device for the board (identified by the "ethernet"
alias) and if found, reads the "local-mac-address" property. If the
"ethernet" alias does not exist, or if it points to a device tree node
that doesn't exist, or if the device tree node that it points to does
not have a "local-mac-address" property or if the value is invalid, it
will fall back to the legacy mechanism of looking for the MAC address
stored in the "nvidia,ethernet-mac" or "nvidia,ether-mac" properties of
the "/chosen" node.
The MAC address is then written to the default Ethernet device for the
board (again identified by the "ethernet" alias) in U-Boot's control
DTB. This allows the device driver for that device to read the MAC
address from the standard location in device tree.
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
This is easier to deal with and works just as well for this simple
function.
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Tegra186 build are currently dealt with in very special ways, which is
because Tegra186 is fundamentally different in many respects. It is no
longer necessary to do many of the low-level programming because early
boot firmware will already have taken care of it.
Unfortunately, separating Tegra186 builds from the rest in this way
makes it difficult to share code with prior generations of Tegra. With
all of the low-level programming code behind Kconfig guards, the build
for Tegra186 can again be unified.
As a side-effect, and partial reason for this change, other Tegra SoC
generations can now make use of the code that deals with taking over a
boot from earlier bootloaders. This used to be nvtboot, but has been
replaced by cboot nowadays. Rename the files and functions related to
this to avoid confusion. The implemented protocols are unchanged.
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Resetting the USB device controller on boot is only necessary if the SoC
actually has a UDC controller and U-Boot enables support for it. All the
Tegra boards support UDC via the ChipIdea UDC driver, so make the UDC on
boot workaround depend on the ChipIdea UDC driver.
This prevents a crash on Tegra186 which does not have the ChipIdea UDC.
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Some devices may restrict access to the PMC to TrustZone software only.
Non-TZ software can detect this and use SMC calls to the firmware that
runs in the TrustZone to perform accesses to PMC registers.
Note that this also fixes reset_cpu() and the enterrcm command on
Tegra186 where they were previously trying to access the PMC at a wrong
physical address.
Based on work by Kalyani Chidambaram <kalyanic@nvidia.com> and Tom
Warren <twarren@nvidia.com>.
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
The save_boot_params() function takes as its first four arguments the
first four registers. On 32-bit ARM these are r0, r1, r2 and r3, all of
which are 32 bits wide. However, on 64-bit ARM thene registers are x0,
x1, x2 and x3, all of which are 64 bits wide. In order to allow reusing
the save_boot_params() implementation on 64-bit ARM, change it to take
unsigned long parameters rather than the fixed size 32-bit integers.
This ensures that the correct values are passed.
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Powergate code is not relevant on all Tegra SoC generations, so guard it
with a Kconfig symbol that can be selected by the generations that need
it.
This is in preparation for unifying Tegra186 code with the code used on
older generations.
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Pin controller code is not relevant on all Tegra SoC generations, so
guard it with a Kconfig symbol that can be selected by the generations
that need it.
This is in preparation for unifying Tegra186 code with the code used on
older generations.
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Memory controller code is not relevant on all Tegra SoC generations, so
guard it with a Kconfig symbol that can be selected by the generations
that need it.
This is in preparation for unifying Tegra186 code with the code used on
older generations.
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
The GP pad control code is not relevant on all Tegra SoC generations, so
guard it with a Kconfig symbol that can be selected by the generations
that need it.
This is in preparation for unifying Tegra186 code with the code used on
older generations.
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Clock code is not relevant on all Tegra SoC generations, so guard it
with a Kconfig symbol that can be selected by the generations that need
it.
This is in preparation for unifying Tegra186 code with the code used on
older generations.
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
There's no need to replicate the pmu.h header file for every Tegra SoC
generation. Use a single header that is shared across generations.
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
pll_c is not a valid parent for the disp1 clock, so trying to set it
will fail. Given that display is not used in U-Boot, remove the init
table entry so that disp1 will keep its default parent (clk_m).
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
On Tegra210 the parents for the disp1 and disp2 clocks are slightly
different from earlier chips. Only pll_p, pll_d_out0, pll_d2_out0 and
clk_m are valid parents (technically pll_d_out is as well, but U-Boot
doesn't know anything about it). Fix up the type name and the mux
definition.
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
This function enables a peripheral clock and then immediately sets its
divider. Add a delay to allow the clock to settle first. This matches the
delay in other places which do a similar thing.
Without this, the I2S device on Nyan does not init properly.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Warren <twarren@nvidia.com>
The first clock type appears to have and incorrect setting for out of the
mux outputs. It should be CLK_M, not OSC. Fix it and its only user.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Warren <twarren@nvidia.com>
While converting CONFIG_SYS_[DI]CACHE_OFF to Kconfig, there are instances
where these configuration items are conditional on SPL. This commit adds SPL
variants of these configuration items, uses CONFIG_IS_ENABLED(), and updates
the configurations as required.
Acked-by: Alexey Brodkin <abrodkin@synopsys.com>
Signed-off-by: Trevor Woerner <trevor@toganlabs.com>
[trini: Make the default depend on the setting for full U-Boot, update
more zynq hardware]
Signed-off-by: Tom Rini <trini@konsulko.com>
Similar changes was done for Zynq in past and this patch just follow
this pattern to separate cpu code from SoC code.
Move arch/arm/cpu/armv8/zynqmp/* -> arch/arm/mach-zynqmp/*
And also fix references to these files.
Based on
"ARM: zynq: move SoC sources to mach-zynq"
(sha1: 0107f24036)
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
The kernel added SZ_4G macro in commit f2b9ba871b (arm64/kernel: kaslr:
reduce module randomization range to 4 GB).
Include linux/const.h for the _AC macro.
Drop a local SZ_4G definition in tegra code.
Cc: Tom Warren <twarren@nvidia.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Stefan Roese <sr@denx.de>
Tegra U-Boot ensures that board_get_usable_ram_top() never returns a value
over 4GB, since some peripherals can't access such addresses. However, on
systems with more than 2GB of RAM, RAM bank 1 does describe this extra
RAM, so that Linux (or whatever OS) can use it, subject to DMA
limitations. Since board_get_usable_ram_top() points at the top of RAM
bank 0, the memory locations describes by RAM bank 1 are not mapped by
U-Boot's MMU configuration, and so cannot be used for anything.
For some completely inexplicable reason, U-Boot's EFI support ignores the
value returned by board_get_usable_ram_top(), and EFI memory allocation
routines will return values above U-Boot's RAM top. This causes U-Boot to
crash when it accesses that RAM, since it isn't mapped by the MMU. One
use-case where this happens is TFTP download of a file on Jetson TX1
(p2371-2180).
This change explicitly tells the EFI code that this extra RAM should not
be used, thus avoiding the crash.
A previous attempt to make EFI honor board_get_usable_ram_top() was
rejected. So, this patch will need to be replicated for any board that
implements board_get_usable_ram_top().
Fixes: aa909462d0 ("efi_loader: efi_allocate_pages is too restrictive")
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
A secure monitor that runs before U-Boot, and hence causes U-Boot to run
in non-secure world, must implement a few operations that U-Boot
otherwise implements when running in secure world. Fix U-Boot to skip
these operations when running in non-secure world. In particular:
- The secure monitor must provide the LP0 resume code and own LP0
configuration in order to maintain security, so must initialize all
the PMC scratch registers used by the boot ROM during LP0 resume.
Consequently, U-Boot should not attempt to clear those registers,
since the register accesses will fail or cause an error.
- The secure monitor owns system security, and so is responsible for
configuring security-related items such as the VPR.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Align the size of the carveout region to 2M. This ensures that the size
can be accurately represented by an LPAE page table that uses sections.
This solves a bug (hang at boot time soon after printing the DRAM size)
that only shows up when the following two commits are merged together:
d32e86bde8 ARM: HYP/non-sec: enable ARMV7_LPAE if HYP mode is supported
6e584e633d ARM: tegra: avoid using secure carveout RAM
Cc: Mark Kettenis <kettenis@openbsd.org>
Cc: Alexander Graf <agraf@suse.de>
Acked-by: Tom Warren <twarren@nvidia.com>
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Another round of sorting Kconfig entries aplhabetically.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Acked-by: Alexey Brodkin <abrodkin@synopsys.com>
Reviewed-by: Joe Hershberger <joe.hershberger@ni.com>
CMD_DM is used for debug purpose and it shouldn't be enabled by default
via Kconfig. Unfortunately this is in the tree for quite a long time
that's why solution is to use imply DM for all targets which are
enabling DM.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Joe Hershberger <joe.hershberger@ni.com>
Fix Kconfig bool, default, select and imply options to be
alphabetically sorted.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
If a secure carveout exists, U-Boot cannot use that memory. Fix
carveout_size() to reflect this, and hence transitively fix
usable_ram_size_below_4g() and board_get_usable_ram_top(). This change
ensures that when U-Boot copies the secure monitor code to install it, the
copy target is not in-use for U-Boot code/data.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
RAM repair has a few pre-requisites:
1) PMIC power supply (rail) enabled.
2) PMC CRAIL power partition powered.
3) Fuse clock active (it's the default).
4) PLLP reshift branch enabled (it's the default, when PLLP is active).
RAM repair also only need run whenever specific partitions are powered
(main SoC and CCPLEX respectively); RAM repair does not need to be
triggered when any other partition changes state.
start_cpu() needs to be re-ordered slightly to match these requirements.
Note that C0NC and CE0 aren't required for RAM repair to
operate, but they also do no harm, so the entire of powerup_cpus() is
moved rather than splitting it up. The call to remove_cpu_resets() is
moved last to ensure that all other actions complete before releasing
reset; since the PMC power partitions are now enabled early, releasing
reset is what causes the CPUs to start executing code, and RAM repair must
complete before the CPU boots.
Note that this commit is the result of squashing a numbmer of commits
in NVIDIA's downstream L4T branch, hence the multiple signoffs below.
Signed-off-by: Bibek Basu <bbasu@nvidia.com>
Signed-off-by: Sandipan Patra <spatra@nvidia.com>
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
This reverts commit 701b7b1d2c. It will
be immediately replaced by a different implementation that is more
complete and runs are more targetted times.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Currently CPU_V7 kconfig symbol supports only ARMv7A architectures under
armv7 folder. This led to a misconception of creating separate folders
for armv7m and armv7r. There is no reason to create separate folder for
other armv7 based architectures when it can co-exist with few Kconfig
symbols.
As a first step towards a common folder, rename CPU_V7 as CPUV7A. Later
separate Kconfig symbols can be added for CPU_V7R and CPU_V7M and
can co exist in the same folder.
Reviewed-by: Tom Rini <trini@konsulko.com>
Tested-by: Michal Simek <michal.simek@xilinx.com>
Suggested-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Replace the psci_save_target_pc call by the new function
psci_save(cpu, pc,context_id)
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
When U-Boot started using SPDX tags we were among the early adopters and
there weren't a lot of other examples to borrow from. So we picked the
area of the file that usually had a full license text and replaced it
with an appropriate SPDX-License-Identifier: entry. Since then, the
Linux Kernel has adopted SPDX tags and they place it as the very first
line in a file (except where shebangs are used, then it's second line)
and with slightly different comment styles than us.
In part due to community overlap, in part due to better tag visibility
and in part for other minor reasons, switch over to that style.
This commit changes all instances where we have a single declared
license in the tag as both the before and after are identical in tag
contents. There's also a few places where I found we did not have a tag
and have introduced one.
Signed-off-by: Tom Rini <trini@konsulko.com>
We have a large number of places where while we historically referenced
gd in the code we no longer do, as well as cases where the code added
that line "just in case" during development and never dropped it.
Signed-off-by: Tom Rini <trini@konsulko.com>
This converts the following to Kconfig:
CONFIG_SPI
This partly involves updating code that assumes that CONFIG_SPI implies
things that are specific to the MPC8xx SPI driver. For now, just update
the CONFIG tests. This also involves reworking the default for
CONFIG_SYS_DEF_EEPROM_ADDR so that we don't set it when we cannot make a
reasonable default, as it does not cause any compile failures.
Signed-off-by: Adam Ford <aford173@gmail.com>
Signed-off-by: Tom Rini <trini@konsulko.com>
AES encryption in CBC mode, in most cases, must be used with random
initialization vector. Using the same key and initialization vector several
times is weak and must be avoided.
Added iv parameter to the aes_cbc_encrypt_blocks and aes_cbc_decrypt_blocks
functions for passing initialization vector.
Command 'aes' now also require the initialization vector parameter.
Signed-off-by: Andrey Mozzhuhin <amozzhuhin@yandex.ru>
In the presence of potentially fragemented memory, we cannot hard-code
addresses into environment variables such as kernel_addr_r. Instead, we
must calculate those addresses at run-time based on available memory
locations. Implement the code to perform such runtime calculation, based
on requirements described in environment variables, to allow the user
full control over the allocation.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Tegra186 currently restricts its DRAM usage to entries in the /memory node
in the DTB passed to it. However, the MMU configuration always maps the
entire first 2GB of RAM. This could allow the CPU to speculatively access
RAM that isn't part of the in-use banks. This patch switches to runtime
construction of the table that's used to construct the MMU translation
tables, and thus prevents access to RAM that's not part of a valid bank.
Note: This patch is intended to prevent access to RAM regions which U-Boot
does not need to access, with the primary purpose of avoiding theoretical
speculative access to physical regions for which the HW will throw errors
(e.g. carve-outs that the CPU has no permission to access at a bus level,
bad ECC pages, etc.). In particular, this patch is not deliberately
related to the speculation-related security issues that were recently
announced. The apparent similarity is a coincidence.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
In the future, the list of DRAM regions passed to U-Boot in the DTB may
be quite long and fragmented. Due to this, U-Boot must search through the
regions to find the best region to relocate into, rather than relying on
the current assumption that the top of bank 0 is a reasonable relocation
target. This change implements such searching.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Apply a few small fixes for the DTB /memory node parsing from NVIDIA's
downstream U-Boot:
- Allow arbitrary number of DRAM banks.
- Correctly calculate the number of DRAM banks.
- Clip PCIe memory in the same way as U-Boot CPU memory use.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Enable CONFIG_LINUX_KERNEL_IMAGE_HEADER for all 64-bit Tegra boards.
cboot (the boot SW that runs before U-Boot) will eventually use this
information.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Enable CONFIG_SYS_INIT_SP_BSS_OFFSET for all 64-bit Tegra boards. Place
the stack/... 512KiB from the end of the U-Boot binary. This should be
plenty to accommodate the current DTBs (max 64 KiB), early malloc region
(6KiB), stack usage, and plenty of slack, while still not placing it too
far away from the U-Boot binary.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Update tegra to use binman for image creation. This still includes the
current Makefile logic, but a later patch will remove this. Three output
files are created, all of which combine
SPL and U-Boot:
u-boot-tegra.bin - standard image
u-boot-dtb-tegra.bin - same as u-boot-tegra.bin
u-boot-nodtb-target.bin - includes U-Boot without the appended device tree
The latter is useful for build systems where the device is appended later,
perhaps after being modified.
Signed-off-by: Simon Glass <sjg@chromium.org>
U-Boot widely uses error() as a bit noisier variant of printf().
This macro causes name conflict with the following line in
include/linux/compiler-gcc.h:
# define __compiletime_error(message) __attribute__((error(message)))
This prevents us from using __compiletime_error(), and makes it
difficult to fully sync BUILD_BUG macros with Linux. (Notice
Linux's BUILD_BUG_ON_MSG is implemented by using compiletime_assert().)
Let's convert error() into now treewide-available pr_err().
Done with the help of Coccinelle, excluing tools/ directory.
The semantic patch I used is as follows:
// <smpl>
@@@@
-error
+pr_err
(...)
// </smpl>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
[trini: Re-run Coccinelle]
Signed-off-by: Tom Rini <trini@konsulko.com>
Add a convenience macro to iterate over subnodes of a node. Make use of
this where appropriate in the code.
Signed-off-by: Simon Glass <sjg@chromium.org>
We are now using an env_ prefix for environment functions. Rename these
two functions for consistency. Also add function comments in common.h.
Quite a few places use getenv() in a condition context, provoking a
warning from checkpatch. These are fixed up in this patch also.
Suggested-by: Wolfgang Denk <wd@denx.de>
Signed-off-by: Simon Glass <sjg@chromium.org>
We are now using an env_ prefix for environment functions. Rename these
commonly used functions, for consistency. Also add function comments in
common.h.
Suggested-by: Wolfgang Denk <wd@denx.de>
Signed-off-by: Simon Glass <sjg@chromium.org>
