| Age | Commit message (Collapse) | Author |
|---|
|
The default load addresses for the various payloads (kernel, DT,
ramdisk) on systems with just 32MB of DRAM have some issues:
For a start the preceding comment doesn't match the actual values:
apparently they were copied from the 64MB S3 layout, then halved, but
since 0x5 is NOT the half of 0x10, they don't match up.
Also those projected maximum sizes are quite restrictive: it's not easy
to build a compressed kernel image with just 4MB. The only defconfig in
mainline Linux that supports the F1C100s (the only 32MB user so far)
creates a 6MB compressed / 15MB uncompressed kernel.
Rearrange the default load addresses to accommodate such a kernel: we
allow an 7MB/16MB kernel, and up to 5MB of ramdisk, stuffing the smaller
binaries like the DTB towards the end, just before the relocated U-Boot.
Shrink the size for DTB and scripts on the way, there is no need for
allowing up to 512K for them.
Signed-off-by: Andre Przywara <[email protected]>
Reviewed-by: Jernej Skrabec <[email protected]>
|
|
Traditionally we assumed that every Allwinner board would come with at
least 256 MB of DRAM, and set our DRAM layout accordingly. This affected
both the default load addresses, but also U-Boot's own address
expectations (like being loaded at 160 MB).
Some SoCs come with co-packaged DRAM, but only provide 32 or 64MB. So
far we special-cased those *chips*, as there was only one chip per DRAM
size. However new chips force us to take a more general approach.
Introduce a Kconfig symbol, which provides the minimum DRAM size of the
board. If nothing else is specified, we use 256 MB, and default to
smaller values for those co-packaged SoCs.
Then select the different DRAM maps according to this new symbol, so
that different SoCs with the same DRAM size can share those definitions.
Inspired by an idea from Icenowy.
This is just refactoring: compiled for all boards before and after this
patch: the binaries were identical.
Signed-off-by: Andre Przywara <[email protected]>
Reviewed-by: Jernej Skrabec <[email protected]>
|
|
As the compile error when D-Cache is enabled is gone, we can have
D-Cache enabled now.
Signed-off-by: Icenowy Zheng <[email protected]>
Reviewed-by: Andre Przywara <[email protected]>
Signed-off-by: Andre Przywara <[email protected]>
|
|
The enable_caches function in architecture-specific board code is only
necessary for V7A CPUs, code for both V8A and ARM926 have already
declared this function.
Only provide our implementation of enable_caches() for V7A CPUs.
Signed-off-by: Icenowy Zheng <[email protected]>
Reviewed-by: Andre Przywara <[email protected]>
Signed-off-by: Andre Przywara <[email protected]>
|
|
The X96 Mate TV box is a TV box with the Allwinner H616 SoC. It is
available with up to 4GB of DRAM and 64GB eMMC.
The DRAM chips require a different configuration when compared to the
OrangePi Zero2, we must not use read/write training and write leveling.
Add a defconfig for the box, so that we can easily build U-Boot for it.
We synced the .dts file already from the kernel tree.
Signed-off-by: Andre Przywara <[email protected]>
Reviewed-by: Jernej Skrabec <[email protected]>
|
|
prepare for rk3566 based board
Signed-off-by: FUKAUMI Naoki <[email protected]>
Reviewed-by: Kever Yang <[email protected]>
|
|
At present the display does not work since it needs the reset driver to
operate. Fix this by enabling it.
Signed-off-by: Simon Glass <[email protected]>
Fixes: cd529f7ad62 ("rockchip: video: edp: Add missing reset support")
Fixes: 9749d2ea29e ("rockchip: video: vop: Add reset support")
Reviewed-by: Kever Yang <[email protected]>
|
|
Now that a single binary containing TPL/SPL correctly formatted for SPI
flashes and U-Boot proper, can be generated by binman, let's do it.
Also update the documentation to tell the user to use this newly
generated file instead of manually generating and flashing the binaries.
Cc: Quentin Schulz <[email protected]>
Signed-off-by: Quentin Schulz <[email protected]>
Reviewed-by: Kever Yang <[email protected]>
|
|
The offset of the SPL payload on Puma is different than for other
Rockchip devices in that it is stored at offset 256K instead of much
further away in the MMC.
Flashing one binary instead of two at different offsets is much more
user friendly so let's migrate to it by modifying the offset in the Puma
specific Device Tree.
Cc: Quentin Schulz <[email protected]>
Signed-off-by: Quentin Schulz <[email protected]>
Reviewed-by: Kever Yang <[email protected]>
|
|
Depending on the toolchain used to compile the SPL for Puma RK3399-Q7
module, the board does not boot because the resulting binary is too big
to fit in SRAM.
Let's add a TPL so that there's no need to fiddle with or hack the
defconfig to have a working bootloader.
This follows what's been done for the majority of other RK3399-based
boards.
See the original commit for the first migrations:
bdc00080111f "rockchip: rk3399: update defconfig for TPL"
Unfortunately, the offset in SPI-NOR for U-Boot proper needs to be
modified, since the move from SPL to TPL+SPL for idbloader.img (and the
"only the first 2KB per 4KB blocks are written" "hack" for rkspi format)
increased the size above 256KB. Let's move it to 512KB to, hopefully, be
safe.
Cc: Quentin Schulz <[email protected]>
Signed-off-by: Quentin Schulz <[email protected]>
Reviewed-by: Kever Yang <[email protected]>
|
|
The SPL payload offset when booting from SPI defaults to
CONFIG_SYS_SPI_U_BOOT_OFFS but can be overridden by
u-boot,spl-payload-offset. The Device Tree for Puma Haikou has this
property so there's no need to have this one option in the defconfig,
especially since they are not in sync and therefore confusing.
Cc: Quentin Schulz <[email protected]>
Signed-off-by: Quentin Schulz <[email protected]>
Reviewed-by: Kever Yang <[email protected]>
|
|
U-Boot proper
Chances are when one boots U-Boot proper from a given storage medium,
they want the same medium to be used to load and store the environment.
This basically allows to have completely separate U-Boot (TPL/SPL/U-Boot
proper/environment) per storage medium which is convenient when working
with recovery from SD-Card as one would just need to insert a properly
configured SD-Card into the device to have access to their whole debug
setup.
No fallback mechanism is provided as to not dirty other storage medium
environment by mistake. However, since arch_env_get_location() is called
by env_init() which is part of the pre-relocation process, a valid,
non-ENVL_UNKNOWN, value shall be returned otherwise the relocation fails
with the following message:
initcall sequence 00000000002866c0 failed at call 0000000000256b34 (err=-19)
This valid, non-ENVL_UNKNOWN, value is ENVL_NOWHERE which requires to
always select CONFIG_ENV_IS_NOWHERE otherwise this work-around does not
work.
Cc: Quentin Schulz <[email protected]>
Signed-off-by: Quentin Schulz <[email protected]>
Reviewed-by: Kever Yang <[email protected]>
|
|
There's a SPI-NOR flash available from which SPL and U-Boot proper can
be booted, it makes sense to also allow this medium to store U-Boot
environment so let's enable it.
The Device Tree advertises a max frequency of 50MHz so let's set the
config option appropriately.
Cc: Quentin Schulz <[email protected]>
Signed-off-by: Quentin Schulz <[email protected]>
Reviewed-by: Kever Yang <[email protected]>
|
|
U-Boot proper
Automatically detect which MMC device (SD-Card or eMMC) was used to load
U-Boot proper and load the environment from that MMC device instead of
a hardcoded one.
Cc: Quentin Schulz <[email protected]>
Signed-off-by: Quentin Schulz <[email protected]>
Reviewed-by: Kever Yang <[email protected]>
|
|
CONFIG_ENV_OFFSET is set in the defconfig to a different value already
so this isn't used. Let's remove it as to not confuse users.
Cc: Quentin Schulz <[email protected]>
Signed-off-by: Quentin Schulz <[email protected]>
Reviewed-by: Kever Yang <[email protected]>
|
|
from SD-Card
Trying to load U-Boot proper from SPL when SPL was not loaded from
SD-Card is currently not working because the SDMMC pins aren't muxed
correctly. It is assumed the BootROM is doing this for us when booting
from SD-Card hence why it's not needed when booting TPL/SPL from
SD-Card.
The pinctrl properties are removed from the SPL DT property removal list
and the pinctrl configuration nodes made available in the SPL DT, in
addition to the pull-up configurations to allow loading U-Boot proper
from SD-Card as a fallback mechanism for SPI-NOR and eMMC.
Cc: Quentin Schulz <[email protected]>
Signed-off-by: Quentin Schulz <[email protected]>
Reviewed-by: Kever Yang <[email protected]>
|
|
eMMC/SPI-NOR
On Haikou devkit, it is possible to disable eMMC and SPI-NOR to force
booting from SD card or USB via rkdeveloptool by toggling a switch. This
switch needs to be overridden in software to be able to access eMMC and
SPI-NOR once the device has booted from SD Card. Puma SoM can override
this pin via gpio3_d5.
Until now, fixed regulator device was abused to model this, but since
there's now support for GPIO hogs, let's use it.
Since we want to be able to boot the SPL from SD Card but give it the
ability to load U-Boot proper from a fallback medium such as eMMC and
SPI-NOR, SPL support for GPIO hogs needs to be enabled too. Support for
other kinds of regulators are not needed anymore, so let's disable them.
Cc: Quentin Schulz <[email protected]>
Signed-off-by: Quentin Schulz <[email protected]>
Reviewed-by: Kever Yang <[email protected]>
|
|
medium
distroboot should try first on the same MMC medium as the one the SPL
loaded U-Boot proper from. This was the case when the introducing commit
was merged because the default order was eMMC first and then SD card.
The check was therefore made only on whether we booted from SD card,
because otherwise the order was the expected one.
However, in commit b212ad24a604 ("rockchip: Fix MMC boot order"), the
order was swapped. Meaning our simple check is now useless.
Let's fix that by accounting for all scenarii: default boot_targets has
mmc0 first but booting from SD Card, mmc1 first but booting from eMMC.
Fixes: b212ad24a604 ("rockchip: Fix MMC boot order")
Cc: Quentin Schulz <[email protected]>
Signed-off-by: Quentin Schulz <[email protected]>
Reviewed-by: Kever Yang <[email protected]>
|
|
Specification
- Rockchip RK3399
- LPDDR3 4GB
- TF sd scard slot
- eMMC
- AP6255 for WiFi + BT
- Gigabit ethernet
- HDMI out
- 40 pin header
- USB 2.0 x 2
- USB 3.0 x 1
- USB 3.0 Type-C x 1 work in otg mode
- 12V DC Power supply
The dts file is sync from linux-next[0].
[0]:https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/tree/arch/arm64/boot/dts/rockchip/rk3399-eaidk-610.dts
Signed-off-by: Andy Yan <[email protected]>
Reviewed-by: Kever Yang <[email protected]>
|
|
CONFIG_SERIAL_TAG is not selectable for ARM64 machines. While
get_board_serial is weakly defined if ENV_VARS_UBOOT_RUNTIME_CONFIG is
defined, it is only called when CONFIG_SUPPORT_PASSING_ATAGS is defined,
which also is not selectable for ARM64 machines. Therefore this is dead
code so let's remove it.
Cc: Quentin Schulz <[email protected]>
Signed-off-by: Quentin Schulz <[email protected]>
Reviewed-by: Kever Yang <[email protected]>
|
|
RAM_PX30_DDR4 is for DDR4 support and not DDR3 so let's fix the typo.
Fixes: 2db36c64bd5a ("ram: rockchip: px30: add a config-based ddr selection")
Cc: Quentin Schulz <[email protected]>
Signed-off-by: Quentin Schulz <[email protected]>
Reviewed-by: Kever Yang <[email protected]>
|
|
UART0 can obviously also be used for debug uart in U-Boot, so let's add
its support.
Cc: Quentin Schulz <[email protected]>
Signed-off-by: Quentin Schulz <[email protected]>
Reviewed-by: Kever Yang <[email protected]>
|
|
The check to perform is on CONFIG_SPL_GPIO and not SPL_GPIO.
Because this was never compiled in, it missed an include of cru.h that
was not detected before. Let's include it too.
Also switch to IS_ENABLED in-code check as it is the preferred
inclusion/exclusion mechanism.
Fixes: 07586ee4322a ("rockchip: rk3399: Support common spl_board_init")
Cc: Quentin Schulz <[email protected]>
Signed-off-by: Quentin Schulz <[email protected]>
Reviewed-by: Kever Yang <[email protected]>
|
|
The check to perform is on CONFIG_SPL_DM_REGULATOR and not
SPL_DM_REGULATOR. Also switch to in-code check instead of ifdefs.
Fixes: 07586ee4322a ("rockchip: rk3399: Support common spl_board_init")
Cc: Quentin Schulz <[email protected]>
Signed-off-by: Quentin Schulz <[email protected]>
Reviewed-by: Jagan Teki <[email protected]>
Reviewed-by: Peter Robinson <[email protected]>
Tested-by: Peter Robinson <[email protected]> # Rock960
Reviewed-by: Kever Yang <[email protected]>
|
|
The original code set up the DDR clock to 48 MHz, not 50MHz as
requested, and did it in a way that didn't satisfy the Application
Notes in RK3399 TRM [1]. 2.9.2.B says:
PLL frequency range requirement
[...]
FOUTVCO: 800MHz to 3.2GHz
2.9.2.A :
PLL output frequency configuration
[...]
FOUTVCO = FREF / REFDIV * FBDIV
FOUTPOSTDIV = FOUTVCO / POSTDIV1 / POSTDIV2
FREF = 24 MHz
The original code gives FOUTVCO: 24MHz/1 * 12 = 288MHz < 800MHz
And the resulting FOUTPOSTDIV is 288MHz / 3 / 2 = 48MHz
but the requested frequency was 50MHz
Note:
2.7.2 Detail Register Description
PMUCRU_PPLL_CON0 says
fbdiv
Feedback Divide Value
Valid divider settings are:
[16, 3200] in integer mode
So .fbdiv = 12 wouldn't be right. But 2.9.2.C says:
PLL setting consideration
[...]
The following settings are valid for FBDIV:
DSMPD=1 (Integer Mode):
12,13,14,16-4095 (practical value is limited to 3200, 2400, or 1600
(FVCOMAX / FREFMIN))
[...]
So .fbdiv = 12 would be right.
In any case FOUTVCO is still wrong. I thank YouMin Chen for
confirmation and explanation.
Despite documentation, I don't seem to be able to reproduce a
practical problem with the wrong FOUTVCO. When I initially found it I
thought some problems with detecting the RAM capacity in my Rock Pi 4B
could be related to it and my patch seemed to help. But since I'm no
longer able to reproduce the issue, it works with or without this
patch. And meanwhile a patch[2] by Lee Jones and YouMin Chen addresses
this issue. Btw, shouldn't that be commited?
So this patches solves no visible problem. Yet, to prevent future
problems, I think it'd be best to stick to spec.
An alternative to this patch could be
{.refdiv = 1, .fbdiv = 75, .postdiv1 = 6, .postdiv2 = 6};
This would theoretically consume more power and yield less jitter,
according to 2.9.2.C :
PLL setting consideration
[...]
For lowest power operation, the minimum VCO and FREF frequencies
should be used. For minimum jitter operation, the highest VCO and
FREF frequencies should be used.
[...]
But I haven't tried it because I don't think it matters much. 50MHz
for DDR is only shortly used by TPL at RAM init. Normal operation is
at 800MHz. Maybe it's better to use less power until later when more
complex software can control batteries or charging or whatever ?
Cc: Simon Glass <[email protected]>
Cc: Philipp Tomsich <[email protected]>
Cc: Kever Yang <[email protected]>
Cc: Lukasz Majewski <[email protected]>
Cc: Sean Anderson <[email protected]>
Link: [1] https://opensource.rock-chips.com/images/e/ee/Rockchip_RK3399TRM_V1.4_Part1-20170408.pdf
Link: [2] https://patchwork.ozlabs.org/project/uboot/list/?series=305766
Signed-off-by: Xavier Drudis Ferran <[email protected]>
Tested-by: Michal Suchánek <[email protected]>
Reviewed-by: Kever Yang <[email protected]>
|
|
Partial sync of rk3288.dtsi from Linux version 5.18
Changed:
only properties and functions that are not yet included
swap some clocks positions
fix some irq numbers
style and sort nodes
Signed-off-by: Johan Jonker <[email protected]>
Reviewed-by: Kever Yang <[email protected]>
|
|
In order to better compare the Linux rk3288.dtsi version
with the u-boot version update the cpu and gpu nodes.
Changed:
use operating-points-v2
update thermal for all cpus
add labels to all cpus
change gpu compatible
change gpu interrupt names
Signed-off-by: Johan Jonker <[email protected]>
Reviewed-by: Kever Yang <[email protected]>
|
|
In order to better compare the Linux rk3288.dtsi version
with the u-boot version move thermal sub nodes to the dtsi
file and remove rk3288-thermal.dtsi
Changed:
replace underscore in nodename
remove comments about sensor and ID
use gpu phandle
add #cooling-cells to gpu node
lower critical temparature
remove linux,hwmon property
Signed-off-by: Johan Jonker <[email protected]>
Reviewed-by: Kever Yang <[email protected]>
|
|
This is not needed or used, and adds code size. Drop it.
Signed-off-by: Simon Glass <[email protected]>
Reviewed-by: Kever Yang <[email protected]>
|
|
The phycore_rk3288 board has a SPL size problem,
so remove phycore_init() function to stay within the limits.
Signed-off-by: Johan Jonker <[email protected]>
Reviewed-by: Wadim Egorov <[email protected]>
Reviewed-by: Kever Yang <[email protected]>
|
|
- Assorted fixes and improvements to some TI platforms
|
|
All of the required values for using the omap_wdt.c driver are found in
<asm/ti-common/omap_wdt.h> and this is what is indirectly pulled in via
<asm/arch/hardware.h> when it exists.
Signed-off-by: Tom Rini <[email protected]>
|
|
For some unknown reason GNU assembler version 2.31.1 (arm-linux-gnueabi-as
from Debian Buster) cannot compile following code from located in file
board/nokia/rx51/lowlevel_init.S:
kernoffs:
.word KERNEL_OFFSET - (. - CONFIG_SYS_TEXT_BASE)
when CONFIG_SYS_TEXT_BASE is set to 0x80008000. It throws strange compile
error which is even without line number:
AS board/nokia/rx51/lowlevel_init.o
{standard input}: Assembler messages:
{standard input}: Error: attempt to get value of unresolved symbol `L0'
make[2]: *** [scripts/Makefile.build:293: board/nokia/rx51/lowlevel_init.o] Error 1
I have no idea about this error and my experiments showed that ARM GNU
assembler is happy with negation of that number. So changing code to:
kernoffs:
.word . - CONFIG_SYS_TEXT_BASE - KERNEL_OFFSET
and then replacing mathematical addition by substraction of "kernoffs"
value (so calculation of address does not change) compiles assembler file
without any error now.
There should be not any functional change.
Signed-off-by: Pali Rohár <[email protected]>
Reviewed-by: Tom Rini <[email protected]>
|
|
These hardware register definitions are common for all K3, remove
duplicate data them by moving them to hardware.h.
While here do some minor whitespace cleanup + grouping.
Signed-off-by: Andrew Davis <[email protected]>
|
|
This matches how this would be done in Linux and these functions
do the alignment for us which makes the code look cleaner.
Signed-off-by: Andrew Davis <[email protected]>
|
|
DMA operations should function on DMA addresses, not virtual addresses.
Although these are usually the same in U-Boot, it is more correct
to be explicit with our types here.
Signed-off-by: Andrew Davis <[email protected]>
|
|
We should clean the caches before any DMA operation and clean+invalidate
after. This matches what the DMA framework does for us already but adds
it to the two functions here in this driver that don't yet go through the
new DMA framework.
Signed-off-by: Andrew Davis <[email protected]>
|
|
The DMA'd memory area needs cleaned and invalidated after the DMA
write so that any stale cache lines do not mask new data.
Signed-off-by: Andrew Davis <[email protected]>
|
|
This matches what we did for pre-K3 devices. This allows us to build
boot commands that can check for our device type at runtime.
Signed-off-by: Andrew Davis <[email protected]>
|
|
Add support for j784s4-wiz-10g device which has two core reference
clocks (e.g core_ref_clk, core_ref1_clk) which requires an additional
mux selection option.
Signed-off-by: Matt Ranostay <[email protected]>
|
|
Move to latest DDR4 1600MT/s for k3-am64-evm based on EMIF tool
v0.08.40.
Signed-off-by: Dave Gerlach <[email protected]>
Signed-off-by: Anand Gadiyar <[email protected]>
|
|
Move k3-am64-sk to use 1600MT/s LPDDR4 configuration and update to latest EMIF
tool v0.08.40.
Signed-off-by: Dave Gerlach <[email protected]>
Signed-off-by: Anand Gadiyar <[email protected]>
|
|
https://source.denx.de/u-boot/custodians/u-boot-dm
Update uclass iterators to work better when devices fail to probe
Support VBE OS requests / fixups
Minor error-handling tweaks to bootm command
|
|
https://source.denx.de/u-boot/custodians/u-boot-stm
- stm32mp: fix compilation issue with DEBUG_UART
- DT update :
- Remove buck3 regulator-always-on on AV96
- Enable btrfs support on DHSOM
- Drop extra newline from AV96 U-Boot extras DT
- Add DHCOR based Testbench board
- Fix and expand PLL configuration comments
- update SCMI dedicated file
|
|
The OrangePi Zero 2 board comes with 2MB of SPI flash, from which the
BROM is able to boot from. Please note that the fuse setup requires
PC5 (BOOT_SEL3) to be pulled to GND for that to actually work.
Enable the SPL code responsible for finding and loading U-Boot proper and
friends, so that u-boot-sunxi-with-spl.bin can be written into the flash.
Signed-off-by: Andre Przywara <[email protected]>
Reviewed-by: Jagan Teki <[email protected]>
Tested-by: Ivan Shishkin <[email protected]>
|
|
The H616 SoC uses the same SPI IP as the H6, also shares the same clocks
and reset bits.
The only real difference is a slight change in the pin assignment: the
H6 uses PC5, the H616 PC4 instead. This makes for a small change in
our spi0_pinmux_setup() routine.
Apart from that, just extend the H6 #ifdef guards to also cover the H616,
using the shared CONFIG_SUN50I_GEN_H6 symbol.
Also use this symbol for the Kconfig dependency.
Signed-off-by: Andre Przywara <[email protected]>
Reviewed-by: Jagan Teki <[email protected]>
Tested-by: Ivan Shishkin <[email protected]>
|
|
When some configuration symbols were converted from header files to
Kconfig, their values were placed into *every* defconfig file.
Since we now have sensible per-SoC defaults defined in Kconfig, those
values are now redundant, and can just be removed.
This affects CONFIG_SPL_STACK, CONFIG_SYS_PBSIZE, CONFIG_SPL_MAX_SIZE,
and CONFIG_SYS_BOOTM_LEN.
Signed-off-by: Andre Przywara <[email protected]>
|
|
Some configuration symbols formerly defined in header files were
recently converted to Kconfig symbols. This moved their value definition
into *every* defconfig file, even though those values are hardly board
choices.
Use the new Kconfig option to define per-SoC default values, in just one
place, which makes the definition in each defconfig file redundant.
We refrain from setting a sunxi specific value for CONFIG_SYS_BOOTM_LEN,
so this defaults to a much better 64MB for uncompressed arm64 kernels.
Signed-off-by: Andre Przywara <[email protected]>
Reviewed-by: Tom Rini <[email protected]>
|
|
Update the devicetree files from the Linux kernel, version v6.0-rc4.
This is covering the 32-bit SoCs, from arch/arm/boot/dts/.
This avoids the not backwards-compatible r_intc binding change, to allow
older kernels to boot, but the other nodes are updated.
Not much change here, the vast majority is actually cosmetic: node names
and using symbolic names for the the RTC clocks.
The R40 boards gain DVFS support.
Some A23/A33 tablet DTs are unified into a single file.
Signed-off-by: Andre Przywara <[email protected]>
Reviewed-by: Jernej Skrabec <[email protected]>
|
|
Update the devicetree files from the Linux kernel, version v6.0-rc4.
This is covering the 64-bit SoCs, from arch/arm64/boot/dts/allwinner.
This avoids the not backwards-compatible r_intc binding change, to allow
older kernels to boot, but the other nodes are updated.
Not much change here, the vast majority is actually cosmetic: node names
and using symbolic names for the the RTC clocks.
Some A64 boards gain some audio nodes.
The H616 DTs are now switched to the version finally merged into the
kernel, which brings some changes, but none affecting U-Boot.
Signed-off-by: Andre Przywara <[email protected]>
Reviewed-by: Jernej Skrabec <[email protected]>
|
