Now that PIE works when U-Boot is started from ROM, let's enable
CONFIG_POSITION_INDEPENDENT, which allows to load U-Boot also via
ARM Trusted-Firmware's fip.bin to DRAM, without tweaking the
configuration.
To get a writable initial stack, we need to keep the fixed initial
stack pointer, which points to DRAM in our case.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
CONFIG_ARCH_SUPPORT_TFABOOT was used on the qemu-arm64 platform to
guard a tweak to the flash bank configuration. U-Boot now reads the
current flash setup from the devicetree, so there is no need for
this option anymore.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Currently we hard-code the number and initial addresses of QEMU's flash
banks, even though our code is perfectly able to gather the same
information from the DTB provided by QEMU.
This is especially annoying, since we have two slightly different
U-Boot configurations ("bare-metal" vs. loaded via Arm Trusted
Firmware), which need to be selected at build time.
Drop the two hard coded alternatives, and use
CONFIG_SYS_MAX_FLASH_BANKS_DETECT instead, which relies on the DTB to
figure out the actual flash configuration at runtime.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Currently selecting CONFIG_POSITION_INDEPENDENT also forces us to use an
initial stack pointer relative to the beginning of the BSS section.
This makes some sense, because this should be writable memory anyway.
However the BSS section is not cleared or used until later in the
setup process (after relocation), so memory nearby might not be
available early enough to host the initial stack. This is an issue if
U-Boot is loaded from (Flash-)ROM, for instance.
Allow CONFIG_INIT_SP_RELATIVE to be turned off by a board's config, to
be able to select a fixed stack pointer, for instance in known good
DRAM.
This will help QEMU utilising PIE, when it's loaded to (Flash-)ROM.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
When the actual offset between link and runtime address is zero, there
is no need for patching up U-Boot early when running with
CONFIG_POSITION_INDEPENDENT.
Skip the whole routine when the distance is 0.
This helps when U-Boot is loaded into ROM, or in otherwise sensitive
memory locations.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
When we build an arm64 target and enable POSITION_INDEPENDENT, we were
skipping our build-time dynamic relocation fixup routine (STATIC_RELA).
This was probably done because we didn't need it in this case, as the
PIE fixup routine in start.S would take care of that at runtime.
However when we now skip this routine (upon detecting that the fixup
offset is 0), this might lead to uninitialised pointers.
Remove the exception, so that we always do the build-time relocation.
NOTE: GNU binutils starting with v2.27.1 do this build-time relocation
automatically, to be in-line with other architecures. So on newer
toolchains our manual fixup is actually not needed. It doesn't hurt to
have it, though, so that we keep compatibility with the popular Linaro
toolchains, which lack this feature.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
The CONFIG_TFABOOT option is more about what U-Boot DOES NOT need to do
than to support some features.
Explain a bit more in the Kconfig help text to avoid misunderstandings.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Most users don't need the standalone API examples. Distributions like SUSE
do not supply libgcc for cross-compiling and we cannot do without on ARMv8
for building examples/.
Make examples selectable via symbol CONFIG_EXAMPLES. It defaults to
yes on ARCH_QEMU to ensure that we compile the API as part of our
continuous integration.
Cc: Matthias Brugger <mbrugger@suse.com>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Matthias Brugger <mbrugger@suse.com>
The timeout calculation is based on the clk being in KHz but
the clk api returns the clk value in Hz. Convert this to KHz
to calculate the correct timeout value.
Signed-off-by: Jack Mitchell <ml@embed.me.uk>
Add a device tree binding for the BOOT button on the Maix board.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Sean Anderson <seanga2@gmail.com>
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Rick Chen <rick@andestech.com>
This patch adds the necessary device tree bindings.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Acked-by: Rick Chen <rick@andestech.com>
This modifies the existing led test to check for default led naming as
added in the previous patch.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
This more closely mirrors Linux's behaviour, and will make it easier to
transition to using function+color in the future.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
dm_gpio_ops.get_value can be called when the gpio is either input or
output. The current dw code always returns the input value, which is
invalid if the direction is set to out.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Ley Foon Tan <ley.foon.tan@intel.com>
Previously, if there was no bank-name property, it was easy to have
confusing gpio names like "gpio1@08", instead of "gpio1@0_8". This patch
follows the example of the sifive gpio driver.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Change the type of gpio_dwabp_platdata.base from fdt_addr_t to a void
pointer, since we pass it to readl.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Ley Foon Tan <ley.foon.tan@intel.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
The Fully-Programmable Input/Output Array (FPIOA) device controls pin
multiplexing on the K210. The FPIOA can remap any supported function to any
multifunctional IO pin. It can also perform basic GPIO functions, such as
reading the current value of a pin. However, GPIO functionality remains
largely unimplemented (in favor of the dedicated GPIO peripherals).
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
This extends the pinctrl-sandbox driver to support pin muxing, and adds a
test for that behaviour. The test is done in C and not python (like the
existing tests for the pinctrl uclass) because it needs to call
pinctrl_select_state. Another option could be to add a command that
invokes pinctrl_select_state and then test everything in
test/py/tests/test_pinmux.py.
The pinctrl-sandbox driver now mimics the way that many pinmux devices
work. There are two groups of pins which are muxed together, as well as
four pins which are muxed individually. I have tried to test all normal
paths. However, very few error cases are explicitly checked for.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
This normalizes the documentation to conform to kernel-doc style [1]. It
also moves the documentation for pinctrl_ops inline, and adds argument and
return-value documentation. I have kept the usual function style for these
comments. I could not find any existing examples of function documentation
inside structs.
[1] https://www.kernel.org/doc/html/latest/doc-guide/kernel-doc.html
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
The pinmux property allows for smaller and more compact device trees,
especially when there are many pins which need to be assigned individually.
Instead of specifying an array of strings to be parsed as pins and a
function property, the pinmux property contains an array of integers
representing pinmux groups. A pinmux group consists of the pin identifier
and mux settings represented as a single integer or an array of integers.
Each individual pin controller driver specifies the exact format of a
pinmux group. As specified in the Linux documentation, a pinmux group may
be multiple integers long. However, no existing drivers use multi-integer
pinmux groups, so I have chosen to omit this feature. This makes the
implementation easier, since there is no need to allocate a buffer to do
endian conversions.
Support for the pinmux property is done differently than in Linux. As far
as I can tell, inversion of control is used when implementing support for
the pins and groups properties to avoid allocating. This results in some
duplication of effort; every property in a config node is parsed once for
each pin in that node. This is not such an overhead with pins and groups
properties, since having multiple pins in one config node does not occur
especially often. However, the semantics of the pinmux property make such a
configuration much more appealing. A future patch could parse all config
properties at once and store them in an array. This would make it easier to
create drivers which do not function solely as callbacks from
pinctrl-generic.
This commit increases the size of the sandbox build by approximately 48
bytes. However, it also decreases the size of the K210 device tree by 2
KiB from the previous version of this series.
The documentation has been updated from the last Linux commit before it was
split off into yaml files.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
the aspeed ddr sdram controller needs to know if the memory chip mounted on
the board is dual x8 die or not. Or it may get the wrong size of the
memory space.
Signed-off-by: Dylan Hung <dylan_hung@aspeedtech.com>
Reviewed-by: Ryan Chen <ryan_chen@aspeedtech.com>
to improve the maintainability. It is more easier to modify and add
configurations of the driver in the centralized ram driver directory.
Signed-off-by: Dylan Hung <dylan_hung@aspeedtech.com>
Reviewed-by: Ryan Chen <ryan_chen@aspeedtech.com>
The cfi-flash driver uses an open-coded version of the generic
algorithm to decode and translate multiple frames of a "reg" property.
This starts off the wrong foot by using the address-cells and size-cells
properties of *this* very node, and not of the parent. This somewhat
happened to work back when we were using a wrong default size of 2,
but broke about a year ago with commit 0ba41ce1b7 ("libfdt: return
correct value if #size-cells property is not present").
Instead of fixing the reinvented wheel, just use the generic function
that does all of this properly.
This fixes U-Boot on QEMU (-arm64), which was crashing due to decoding
a wrong flash base address:
DRAM: 1 GiB
Flash: "Synchronous Abort" handler, esr 0x96000044
elr: 00000000000211dc lr : 00000000000211b0 (reloc)
elr: 000000007ff5e1dc lr : 000000007ff5e1b0
x0 : 00000000000000f0 x1 : 000000007ff5e1d8
x2 : 000000007edfbc48 x3 : 0000000000000000
x4 : 0000000000000000 x5 : 00000000000000f0
x6 : 000000007edfbc2c x7 : 0000000000000000
x8 : 000000007ffd8d70 x9 : 000000000000000c
x10: 0400000000000003 x11: 0000000000000055
^^^^^^^^^^^^^^^^
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Stefan Roese <sr@denx.de>
When possible use DMA for reading from CFI flash, this provides upto 5x
improvement in read performance with high speed CFI compliant flashes
like HyperFlash.
Code will gracefully fallback to CPU copy when DMA is unavailable.
Signed-off-by: Vignesh Raghavendra <vigneshr@ti.com>
Reviewed-by: Stefan Roese <sr@denx.de>
Caller would need gracefully handle failures of dma_get_device(),
therefore reduce pr_err() to pr_debug() when DMA device is not found.
Signed-off-by: Vignesh Raghavendra <vigneshr@ti.com>
Reviewed-by: Stefan Roese <sr@denx.de>
- mips: octeon: add support for DWC3 USB
- mips: octeon: add support for booting Linux
-----BEGIN PGP SIGNATURE-----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=v+p3
-----END PGP SIGNATURE-----
Merge tag 'mips-pull-2020-10-07' of https://gitlab.denx.de/u-boot/custodians/u-boot-mips
- mips: octeon: add support for DDR4 memory controller
- mips: octeon: add support for DWC3 USB
- mips: octeon: add support for booting Linux
Octeon needs a platform specific cmd to boot the Linux kernel, as
specific parameters need to be passed and special handling for the
multiple cores (SMP) is needed.
Co-developed-by: Stefan Roese <sr@denx.de>
Signed-off-by: Aaron Williams <awilliams@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>
[use gd->ram_base instead of gd->bd->bi_memstart]
Signed-off-by: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
This is needed for Linux booting, as the memory infos need to be passed
in this bootmem format to the Linux kernel.
Signed-off-by: Aaron Williams <awilliams@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>
Add header to handle bootinfo support, needed for Octeon Linux kernel
booting.
Signed-off-by: Aaron Williams <awilliams@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>
This header includes the Octeon feature detection used in many Octeon
drivers.
Signed-off-by: Aaron Williams <awilliams@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>
This header includes common register defines and accessor functions.
Signed-off-by: Aaron Williams <awilliams@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>
This patch adds the necessary lowlevel init code, to enable SMP Linux
booting. This code will be used with the platform specific Octeon Linux
boot command "bootoctlinux", which starts a configurable number of cores
into Linux.
Additionally some erratas and lowlevel register initializations are
copied from the original Cavium / Marvell U-Boot source code, enabling
booting into the Linux kernel.
Signed-off-by: Stefan Roese <sr@denx.de>
As noticed while working on the USB xHCI support, Octeon needs to flush
all pending writes so that the values are present in the memory. Add
this "syncw" instruction (twice) to flush_dcache_range().
Signed-off-by: Stefan Roese <sr@denx.de>
Import platform specific mangle-port.h header, allowing a area specific
swapping, which is needed on Octeon for USB & PCI areas.
Imported from Linux v5.7.
Signed-off-by: Stefan Roese <sr@denx.de>
Import octeon_should_swizzle_table[] which is needed for the area
specific swapping. It will be used by the platform specific
mangle-port.h header.
Imported from Linux v5.7.
Signed-off-by: Stefan Roese <sr@denx.de>
This patch adds the glue layer for the MIPS Octeon SoCs. It's ported
mainly from the Linux code.
Signed-off-by: Stefan Roese <sr@denx.de>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Cc: Bin Meng <bmeng.cn@gmail.com>
Cc: Marek Vasut <marex@denx.de>
Octeon uses mapped addresses for virtual and physical memory. It's not
that easy to calculate the resulting addresses here. So let's remove
this BUG_ON() completely, as it's not really helpful.
Please also note, that BUG_ON() is not recommended any more in the Linux
kernel.
Signed-off-by: Stefan Roese <sr@denx.de>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Cc: Bin Meng <bmeng.cn@gmail.com>
Cc: Marek Vasut <marex@denx.de>
On MIPS platforms, mapping of the base address is needed. This patch
switches from dev_get_addr() to dev_remap_addr() to get the mapped base
address of the xHCI controller.
Signed-off-by: Stefan Roese <sr@denx.de>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Cc: Bin Meng <bmeng.cn@gmail.com>
Cc: Marek Vasut <marex@denx.de>
This patch adds the board specific configuration (struct) for the
Octeon 3 EBB7304 EVK. This struct is ported from the 2013er Cavium /
Marvell U-Boot repository. Also, the Octeon RAM driver is enabled in
the board defconfig for its usage.
Tested with one and two DIMMs on the EBB7304 EVK (8 & 16 GiB).
Signed-off-by: Stefan Roese <sr@denx.de>
This patch adds the initialization call for the Octeon RAM driver to
the Octeon platforms code. So if enabled via Kconfig, the DDR driver
will be called and the RAM will be configured and used. If the RAM
driver is not enabled, the L2 cache is still used as RAM.
Signed-off-by: Stefan Roese <sr@denx.de>
This Octeon 3 DDR driver is ported from the 2013 Cavium / Marvell U-Boot
repository. It currently supports DDR4 on Octeon 3. It can be later
extended to support also DDR3 and Octeon 2 platforms.
Part 3 includes the DIMM SPD handling code and the Kconfig / Makefile
integration.
Signed-off-by: Aaron Williams <awilliams@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>
This Octeon 3 DDR driver is ported from the 2013 Cavium / Marvell U-Boot
repository. It currently supports DDR4 on Octeon 3. It can be later
extended to support also DDR3 and Octeon 2 platforms.
Part 2 includes the very complex Octeon 3 DDR4 configuration
Signed-off-by: Aaron Williams <awilliams@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>
