Commit bfb33f0bc4 ("sunxi: mctl_mem_matches: Add missing memory
barrier") broke compilation for the Pine64, as dram_helper.c now
includes <asm/armv7.h>, which does not compile on arm64.
Fix this by moving all barrier instructions into a separate header
file, which can easily be shared between arm and arm64.
Also extend the inline assembly to take the "sy" argument, which is
optional for ARMv7, but mandatory for v8.
This fixes compilation for 64-bit sunxi boards (Pine64).
Acked-by: Ian Campbell <ijc@hellion.org.uk>
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
This is mandatory, otherwise the USB does not work.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Stefan Roese <sr@denx.de>
Cc: Dinh Nguyen <dinguyen@kernel.org>
According PL310 TRM, Auxiliary Control Register
"
The register must be written to using a secure access, and it can be
read using either a secure or a NS access. If you write to this register
with a NS access, it results in a write response with a DECERR response,
and the register is not updated. Writing to this register with the L2
cache enabled, that is, bit[0] of L2 Control Register set to 1,
results in a SLVERR.
"
So If L2 cache is already enabled by ROM, chaning value of ACR
will cause SLVERR and uboot hang.
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Cc: Stefano Babic <sbabic@denx.de>
Cc: Fabio Estevam <fabio.estevam@nxp.com>
The current delays in the DDR initialization routines for am33xx
architectures are sometimes not running long enough leading to DDR
init errors. On am437x, this shows up as an L3 NOC error after the
kernel boots. This is due to the timer not being initialized
properly, but instead still containing the timer init values from
the boot ROM which cause timers to expire in 1/4th the time
required.
timer_init is typically not called until board_init_r, however on
am33xx/am43xx udelay is required in sdram_init which is called
from board_init_f, so a call to timer_init is required earlier.
Note that this issue introduced in v2015.01 by:
b352dde "am33xx: Drop timer_init call from s_init".
Although this could instead fixed by reverting said commit, it
would cause timer_init to be called twice in both SPL and non-SPL
cases. This gives a little more fine grained control and also
matches what is being done on omap-command and fsl-layerscape.
Signed-off-by: Russ Dill <russ.dill@ti.com>
Commit 724219a65f "ARM: always perform per-CPU GIC init" removed some
ifdefs to unify the MULTIENTRY-vs-non-MULTIENTRY paths. However, the
wrong endif was removed. This patch adds back that missing endif, and
adds a new ifdef to match the endif the now-correctly-terminated block
used to match against. Use "git show -U25 724219a65f55" to see enough
context to make the original issue clear.
In practical terms, this makes no difference to runtime behaviour. The
code that was incorrectly compiled into the binary when ifndef MULTIENTRY
is a no-op for other cases, since branch_if_master evaluates to a hard-
coded jump. The only issues were:
- A few extra instructions were added to the binary.
- The comment on the endif at the very end of the function, indicating
which ifdef it matched, were wrong.
An alternative might be to simply fix the comment on that trailing ifdef,
but that only addresses the second point above, not the first.
Fixes: 724219a65f ("ARM: always perform per-CPU GIC init")
Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Reviewed-by: Masahiro Yamada <yamada.masahiro@socionext.com>
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>
This reverts commit b5788dc0dd.
Ram size is incorrectly reported as 512MB on a firefly-rk3288 board
with 2GB of ram. Reverting this patch displays the full amount of ram.
Signed-off-by: Vagrant Cascadian <vagrant@debian.org>
Acked-by: Simon Glass <sjg@chromium.org>
Unfortunately with this change we now are unable to do FS mode boots
from MMC1 as with the way the code works today we will always load and
assume that the hard-coded raw location contains U-Boot. Further, we
cannot fix this by just changing other logic to try FS-then-RAW as it
would also make us have to ignore what order the ROM is telling us to
try.
This reverts commit 22d90d560a.
Signed-off-by: Tom Rini <trini@konsulko.com>
The code in uniphier_sld3_sbc_init() is pin-muxing, so it would
be a better fit in uniphier_sld3_early_pin_init().
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
The System Bus is not available by default on the ROM boot mode of
PH1-LD20. To use devices connected to the System Bus, such as the
Micro Support Card, it is necessary to set up pin-muxing and some
System Bus Controller register.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
PH1-LD20 does not have the dedicated boot swap select latch.
Instead, it is controlled from the boot mode select.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
We are running with the caches disabled when mctl_mem_matches gets called,
but the cpu's write buffer is still there and can still get in the way,
add a memory barrier to fix this.
This avoids mctl_mem_matches always returning false in some cases, which
was resulting in:
U-Boot SPL 2015.07 (Apr 14 2016 - 18:47:26)
DRAM: 1024 MiB
U-Boot 2015.07 (Apr 14 2016 - 18:47:26 +0200) Allwinner Technology
CPU: Allwinner A23 (SUN8I)
DRAM: 512 MiB
Where 512 MiB is the right amount, but the DRAM controller would be
initialized for 1024 MiB.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
ABB should be initialized for all required domains voltage domain
for DRA7: IVA, GPU, EVE in addition to the existing MPU domain. If
we do not do this, kernel configuring just the frequency using the
default boot loader configured voltage can fail on many corner lot
units and has been hard to debug. This specifically is a concern with
DRA7 generation of SoCs since other than VDD_MPU, all other domains
are only permitted to setup the voltages to required OPP only at boot.
Reported-by: Richard Woodruff <r-woodruff2@ti.com>
Signed-off-by: Nishanth Menon <nm@ti.com>
Since we setup the voltage and frequency for the MM domain, we *must*
setup the ABB configuration needed for the domain as well. If we do not
do this, kernel configuring just the frequency using the default boot
loader configured voltage can fail on many corner lot units.
Reported-by: Richard Woodruff <r-woodruff2@ti.com>
Signed-off-by: Nishanth Menon <nm@ti.com>
ABB TX_DONE mask will vary depending on ABB module. For example,
3630 never had ABB on IVA domain, while OMAP5 does use ABB on MM domain,
DRA7 has it on all domains with the exception of CORE, RTC.
Hence, move the txdone mask definition over to structure describing
voltage domain.
Signed-off-by: Nishanth Menon <nm@ti.com>
This information is already available under vcores->volts.efuse.reg.
There is no reason for duplicating the information since AVS Class 0
definitions are common for OMAP5 and DRA7 and defined with
STD_FUSE_OPP_* macros. This allows a central location of defining
the ABB and voltage definitions especially since they are reused.
This also makes it simpler to prevent mistakes involved when changing
the boot OPP for the device.
Signed-off-by: Nishanth Menon <nm@ti.com>
This is a binding which only exists in U-Boot, but is
required to get working serial in U-Boot.
Signed-off-by: Peter Griffin <peter.griffin@linaro.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
Currently only the serial pl01x driver is using DT,
and the other drivers still use platform data but
as more DT lands in the upstream kernel the aim is
to migrate the other drivers over to DT as well to
have a fully DT configured hikey u-boot.
Signed-off-by: Peter Griffin <peter.griffin@linaro.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
Import the upstream kernel dts into U-Boot. Currently
only serial is supported, but a lot more DT changes are
queued for v4.7.
Signed-off-by: Peter Griffin <peter.griffin@linaro.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
This syncs up the current cmd/Kconfig and include/configs/ files with the
only exception being CMD_NAND. Due to how we have used this historically
we need to take further care here when converting.
Signed-off-by: Tom Rini <trini@konsulko.com>
PH1-LD20 does not support 1.8V signaling for SD card; only Default
Speed and High Speed (up to 50MHz) with 3.3V signaling is supported.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
The current table look-up for the DTB file name turned out bothersome
in terms of maintainability; I ended up adding a new entry every time
a new board is supported.
There is a common pattern between the DT compatible string and the
corresponding file name; drop the vendor prefix "socionext," and
prefix it with "uniphier-" and suffix it with ".dtb".
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
No need to stop booting U-Boot even if boot mode is unknown.
Setting the "bootmode" environment is only useful for booting
Linux Kernel. Anyway, U-Boot has already booted by this point.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
For PH1-LD20 or later, per-pin input-enable control is supported,
that is, we need to set-up IECTRL registers for a group of pins.
This helper function will be useful for a bunch of register settings.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Currently, fdtgrep would not accept uniphier-ph1-ld20-ref.dtb
and uniphier-ph1-ld11-ref.dtb unless the aliases node comes
the first in the root node.
$ make -s uniphier_pxs2_ld6b_defconfig
$ make CROSS_COMPILE=arm-linux-gnueabi- DEVICE_TREE=uniphier-ph1-ld20-ref
[snip]
LDS spl/u-boot-spl.lds
LD spl/u-boot-spl
OBJCOPY spl/u-boot-spl-nodtb.bin
FDTGREP spl/u-boot-spl.dtb
Error at 'fdt_find_regions': FDT_ERR_BADLAYOUT
/aliases node must come before all other nodes
Error: FDT_ERR_BADMAGIC
make[1]: *** [spl/u-boot-spl.dtb] Error 1
make: *** [spl/u-boot-spl] Error 2
This commit moves the aliases node as the error message from the
fdtgrep tool suggests, although this requirement does not sound
reasonable to me.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Because DT properties are 4-byte aligned, the pointer access
*(fdt64_t *) in this code causes unaligned access.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Our own ACPI implementation (when CONFIG_QEMU_ACPI_TABLE is not set)
does not build anymore after x86 has been fully converted to DM PCI.
Instead of trying to fix the build errors, given we now have the ACPI
support via QEMU's fw_cfg interface, which is a more reliable way to
generate correct ACPI tables than by ourselves, hence drop our own
ACPI implementation.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
flush_dcache_all() is used in the very end of U-Boot self relocation
to write back all copied and then patched code and data to their
new location in the very end of available memory space.
Since that has nothing to do with IO (i.e. no external DMA happens
here) IOC won't help here and we need to write back data cache contents
manually.
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
This enables full ethernet usage, including U-Boot to write the board
specific MAC address (ethaddr) into the DT blob before passing it to
Linux.
Without this, the ethaddr is not detected in U-Boot at all, resulting
in this error upon bootup:
...
Model: EBV SOCrates
Net:
Error: ethernet@ff702000 address not set.
No ethernet found.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Marek Vasut <marex@denx.de>
Cc: Dinh Nguyen <dinguyen@opensource.altera.com>
Cc: Chin Liang See <clsee@altera.com>
Fix a small typo in some of the SoCFPGA dts files that has spread via
copy-and-paste.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Marek Vasut <marex@denx.de>
Cc: Dinh Nguyen <dinguyen@opensource.altera.com>
Cc: Chin Liang See <clsee@altera.com>
Add support for Technexion's PICO-IMX6UL-EMMC board.
For information about this board, please visit:
http://www.technexion.com/products/pico/pico-som/pico-imx6-emmc
Signed-off-by: Richard Hu <richard.hu@technexion.com>
Signed-off-by: Fabio Estevam <fabio.estevam@nxp.com>
Due to incorrect placement of the clock gate cell in the ldb_di[x]_clk tree,
the glitchy parent mux of ldb_di[x]_clk can cause a glitch to enter the
ldb_di_ipu_div divider. If the divider gets locked up, no ldb_di[x]_clk is
generated, and the LVDS display will hang when the ipu_di_clk is sourced from
ldb_di_clk.
To fix the problem, both the new and current parent of the ldb_di_clk should
be disabled before the switch. This patch ensures that correct steps are
followed when ldb_di_clk parent is switched in the beginning of boot.
This patch was ported from the 3.10.17 NXP kernel
http://git.freescale.com/git/cgit.cgi/imx/linux-2.6-imx.git/commit/?h=imx_3.10.17_1.0.1_ga&id=eecbe9a52587cf9eec30132fb9b8a6761f3a1e6d
NXP errata number: ERR009219, EB821
Signed-off-by: Akshay Bhat <akshay.bhat@timesys.com>
Cc: Stefano Babic <sbabic@denx.de>
Cc: Fabio Estevam <fabio.estevam@nxp.com>
