In current Linux kernel Tegra DT files, 64-bit addresses are represented
in unit addresses as a pair of comma-separated 32-bit values. Apparently
this is no longer the correct representation for simple busses, and the
unit address should be represented as a single 64-bit value. If this is
changed in the DTs, arm/arm/mach-tegra/board2.c:ft_system_setup() will no
longer be able to find and enable the GPU node, since it looks up the node
by name.
Fix that function to enable nodes based on their compatible value rather
than their node name. This will work no matter what the node name is, i.e
for DTs both before and after any rename operation.
Cc: Thierry Reding <treding@nvidia.com>
Cc: Alexandre Courbot <acourbot@nvidia.com>
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Remove the old PWM code. Remove calls to CONFIG_LCD functions now that we
are using driver model for video.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Warren <twarren@nvidia.com>
While we transition to using driver model for video, we need to support both
options.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Warren <twarren@nvidia.com>
We can skip this manual init when using driver model for the PWM.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Warren <twarren@nvidia.com>
When loading U-Boot into RAM over USB protocols using tools such as
tegrarcm or L4T's exec-uboot.sh/tegraflash.py, Tegra's USB device
mode controller is initialized and enumerated by the host PC running
the tool. Unfortunately, these tools do not shut down the USB
controller before executing the downloaded code, and so the host PC
does not "de-enumerate" the USB device. This patch implements optional
code to shut down the USB controller when U-Boot boots to avoid leaving
a stale USB device present.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Adjust the Tegra PCI driver to support driver model and move all boards over
at the same time. This can make use of some generic driver model code, such
as the range-decoding logic.
Signed-off-by: Simon Glass <sjg@chromium.org>
Tested-by: Stephen Warren <swarren@nvidia.com>
Rename GPU functions to less generic names to avoid potential name
collisions.
Signed-off-by: Alexandre Courbot <acourbot@nvidia.com>
Acked-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Enable the GPU node in the system-wide ft_system_setup() hook instead of
the board-specific ft_board_hook(). This allows us to enable GPU per SoC
generation instead of per-board as we did initially.
Reported-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Alexandre Courbot <acourbot@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Represent all available RAM in either one or two banks. The first bank
describes any RAM below 4GB. The second bank describes any RAM above 4GB.
This split is driven by the following requirements:
- The NVIDIA L4T kernel requires separate entries in the DT /memory/reg
property for memory below and above the 4GB boundary. The layout of that
DT property is directly driven by the entries in the U-Boot bank array.
- On systems with RAM beyond a physical address of 4GB, the potential
existence of a carve-out at the end of RAM below 4GB can only be
represented using multiple banks, since usable RAM is not contiguous.
While making this change, add a lot more comments re: how and why RAM is
represented in banks, and implement a few more "semantic" functions that
define (and perhaps later detect at run-time) the size of any carve-out.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Warren <twarren@nvidia.com>
U-boot is responsible for enabling the GPU DT node after all necessary
configuration (VPR setup for T124) is performed. In order to be able to
check whether this configuration has been performed right before booting
the kernel, make it happen during board_init().
Also move VPR configuration into the more generic gpu.c file, which will
also host other GPU-related functions, and let boards specify
individually whether they need VPR setup or not.
Signed-off-by: Alexandre Courbot <acourbot@nvidia.com>
Cc: Stephen Warren <swarren@nvidia.com>
Cc: Tom Warren <twarren@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Additionally, ARM64 devices typically run a secure monitor in EL3 and
U-Boot in EL2, and set up some secure RAM carve-outs to contain the EL3
code and data. These carve-outs are located at the top of 32-bit address
space. Restrict U-Boot's RAM usage to well below the location of those
carve-outs. Ideally, we would the secure monitor would inform U-Boot of
exactly which RAM it could use at run-time. However, I'm not sure how to
do that at present (and even if such a mechanism does exist, it would
likely not be generic across all forms of secure monitor).
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
CPU board (E2530) has a fan - turn it on via GPIO to keep
the SoC cool.
Acked-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
A subsequent patch will enable the use of the architected timer on
ARMv8. Doing so implies that udelay() will be backed by this timer
implementation, and hence the architected timer must be ready when
udelay() is first called. The first time udelay() is used is while
resetting the debug UART, which happens very early. Make sure that
arch_timer_init() is called before that.
Signed-off-by: Thierry Reding <treding@nvidia.com>
Acked-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Most peripherals on Tegra can do DMA only to the lower 32-bit
address space, even on 64-bit SoCs. This limitation is
typically overcome by the use of an IOMMU. Since the IOMMU is
not entirely trivial to set up and serves no other purpose
(I/O protection, ...) in U-Boot, restrict 64-bit Tegra SoCs to
the lower 32-bit address space for RAM. This ensures that the
physical addresses of buffers that are programmed into the
various DMA engines are valid and don't alias to lower addresses.
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Somehow this change was dropped in the various merges. I noticed when I
came to turn off the non-driver-model support for Tegra. We need to make
this change (and deal with any problems) before going further.
Change-Id: Ib9389a0d41008014eb0df0df98c27be65bc79ce6
Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Marek Vasut <marex@denx.de>
Add a hook to allows boards to add their own init to board_init().
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Warren <twarren@nvidia.com>
All the Tegra boards borrow the files from board/nvidia/common/
directory, i.e., board/nvidia/common/* are not vendor-common files,
but SoC-common files.
Move NVIDIA common files to arch/arm/mach-tegra/ to clean up
Makefiles.
As arch/arm/mach-tegra/board.c already exists, this commit renames
board/nvidia/common/board.c to arch/arm/mach-tegra/board2.c,
expecting they will be consolidated as a second step.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
Cc: Stephen Warren <swarren@nvidia.com>
Cc: Tom Warren <twarren@nvidia.com>
Cc: Simon Glass <sjg@chromium.org>
Acked-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Stephen Warren