Memory region used by remote cores was set to non-cached region but
commit 7c9c6e1925 ("arm: mach-k3: Merge initial memory maps") makes
all memory region as cached, unified across K3 devices.
This causes inconsistency while booting remote cores on devices, due to
cache incoherency between remote core and boot code.
So to make this operation coherent, cache the address and len while
loading ELF program headers to memory and flush that region in the next
cycle of load.
Signed-off-by: Udit Kumar <u-kumar1@ti.com>
Signed-off-by: Beleswar Padhi <b-padhi@ti.com>
Defining CONFIG_MT8512 is unnecessary as now board for mediatek
target can be changed in config.
Use CONFIG_TARGET_MT8512 to replace CONFIG_MT8512.
Signed-off-by: Weijie Gao <weijie.gao@mediatek.com>
The PCS exists only in GMAC1 and relates to SGMII interface and
is used to control the SGMII PHY.
Signed-off-by: Jim Liu <JJLIU0@nuvoton.com>
[trini: Adjust slightly for white space and to move 'start' to within if
block]
First set of u-boot-at91 features for the 2025.07 cycle:
This feature set includes improvements on the atmel-quadspi driver, a
fix for the nand driver, and improvements on the pinctrl driver to be
able to use the Linux DT (also sync on the DT side as well).
When 3-wire mode is claimed on the bus, use the MOSI (output) pin to
receive data. In this mode, since the transfer can only be either TX
or RX, return -EINVAL if both are required at the same time.
Signed-off-by: Hironori KIKUCHI <kikuchan98@gmail.com>
Judith Mendez <jm@ti.com> says:
Some K3 devices like am62x and am64x have a M4 processor in the MCU
voltage domain. This patch series introduces remoteproc M4 driver which
will be used to load firmware into and start the M4 remote core.
This series also adds support for R5F cores on am64x SoCs in patch 2 and
sets up environment to load FW in remote cores in patch 3,4,5.
This patch series also enables remoteproc drivers by default as per what
remoteproc sybsystem is supported per SoC, thus all remoteproc options
are now deleted in configs/* since they are no longer required.
This patch series was tested on am64x EVM, am62x SK, am62ax SK,
am62px SK boards.
Any additional tested by's are welcome since I was not able to
test any additional boards.
Tested by running the following commands in u-boot prompt:
=> setenv dorprocboot 1
=> run boot_rprocs
Link: https://lore.kernel.org/r/20250210202944.1071931-1-jm@ti.com
If SYS_K3_SPL_ATF is enabled, for K3 ARCH enable the
remoteproc ARM64 driver by default so that it does not
have to be defined in each board defconfig file.
Signed-off-by: Judith Mendez <jm@ti.com>
Reviewed-by: Andrew Davis <afd@ti.com>
AM64x SoCs have two R5F clusters in the main power domain.
Extend support for R5F remote proc driver on AM64x with compatible
strings.
Signed-off-by: Hari Nagalla <hnagalla@ti.com>
Signed-off-by: Judith Mendez <jm@ti.com>
Reviewed-by: Daniel Schultz <d.schultz@phytec.de>
Reviewed-by: Andrew Davis <afd@ti.com>
Some K3 devices like AM64, AM62 devices have a M4 processor in MCU
voltage domain.
Add a remote proc driver to support this subsystem to be able to load
and boot the M4 core.
Signed-off-by: Hari Nagalla <hnagalla@ti.com>
[Ryan: Fix implicitly include warning]
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
[Judith: Cleanup driver, fix warnings, remove lreset logic]
Signed-off-by: Judith Mendez <jm@ti.com>
Tested-by: Daniel Schultz <d.schultz@phytec.de>
Reviewed-by: Andrew Davis <afd@ti.com>
Add CONFIG_BOOTCOUNT_ALTBOOTCMD so the developer is able to add
custom altbootcmd via Kconfig when they enable BOOTCOUNT. With this now
in Kconfig, we need to move it from environment files / config.h files
and in to the defconfig file.
This was done by generating u-boot-initial-env for all platforms before
the Kconfig change, to extract altbootcmd values and then again after to
compare the result.
[trini: Perform migration to defconfigs, reword commit message]
Signed-off-by: Tom Rini <trini@konsulko.com>
From reading the S34ML02G1 and the SAM9X60 datasheets again, it seems
like we have to wait tREA after rising RE# before sampling the data.
Thus pulse time must be at least tREA.
Without this fix we got PMECC errors when reading, after switching to
ONFI timing mode 3 on SAM9X60 SoC with S34ML02G1 raw NAND flash chip.
The approach to set timings used before worked on sam9g20 and sama5d2
with the same flash (S34ML02G1), probably because those have a slower
mck clock rate and thus the resolution of the timings setup is not as
tight as with sam9x60.
The approach to fix the issue was carried over from at91bootstrap, and
has been successfully tested in at91bootstrap, U-Boot and Linux.
Link: https://github.com/linux4sam/at91bootstrap/issues/174
Cc: Li Bin <bin.li@microchip.com>
Signed-off-by: Alexander Dahl <ada@thorsis.com>
U-Boot pinctrl driver expects a reg property explicitly unlike linux.
To align the DT of U-boot with the Linux, reg property is also arrvied
from child GPIO bank nodes when configured under the pinctrl node.
Signed-off-by: Manikandan Muralidharan <manikandan.m@microchip.com>
In Linux DT,the pinctrl node acts as parent nodes with all other
gpio banks as child nodes and a single driver in Linux handles both
pinctrl settings and gpio requests.Current U-Boot DT maintains both
pinctrl and gpio nodes as separate nodes and offers two different class
of U-Boot drivers: UCLASS_PINCTRL which handles pin functions and
UCLASS_GPIO which handles gpio requests. In order to align the DT
of U-Boot with the DT of Linux, a hook is been added in the pinctrl
driver to bind the gpio driver with the pinctrl driver so that
when adding gpio nodes as subnodes to pinctrl node (as per the Linux ABI),
the corresponding APIs will be redirected and handled by valid
drivers attached to the pinctrl driver.
Signed-off-by: Manikandan Muralidharan <manikandan.m@microchip.com>
The qspi controller on sama5d2 and sam9x60 supports "classic" SPI mode
without spi-mem enhancements and accelerations, very similar to the old
SPI controller on sam9g20 or the modern flexcom controllers of the same
SoC family.
Register interface differs somewhat, especially because only one
hardware controlled CS line is supported. Some fields are missing, some
are in different registers, but in principal it works similar. So code
is very much inspired by the old atmel-spi driver.
Tested on sam9x60 with a non-mainline driver to configure an FPGA.
Signed-off-by: Alexander Dahl <ada@thorsis.com>
Switching between Serial Memory Mode (SMM) and (classic) SPI mode is a
preparation for implementing .xfer() in the future.
Signed-off-by: Alexander Dahl <ada@thorsis.com>
The Serial Memory Mode (SMM) is enabled with atmel_qspi_set_cfg() on
each invocation of atmel_qspi_exec_op(). Setting SMM through
atmel_qspi_init() at probe time is redundant.
Removing the SMM setting at probe time should therefore 1) be safe to do
and 2) allows for setting it to a different value in a future
implementation of .xfer() which needs to disable SMM.
Signed-off-by: Alexander Dahl <ada@thorsis.com>
Port these commits:
- v6.11-rc5-90-g329ca3eed4a9a ("spi: atmel-quadspi: Avoid overwriting delay register settings")
- v6.12-rc1-1-g162d9b5d2308c ("spi: atmel-quadspi: Fix wrong register value written to MR").
- v6.13-rc2-27-gf663898d047a7 ("spi: atmel-quadspi: Factor out switching to Serial Memory Mode to function")
Cc: Csókás Bence <csokas.bence@prolan.hu>
Signed-off-by: Alexander Dahl <ada@thorsis.com>
Port changes from a 4 piece patch series from Linux kernel v5.10, merged
with v5.10-rc1-83-gc732b7567d869 ("Merge series "spi: atmel-quadspi: Fix
AHB memory accesses" from Tudor Ambarus …").
Port the single fix v5.15-rc1-14-g09134c5322df9 ("spi: Fixed division by
zero warning").
Reduces differences between linux and u-boot driver.
Cc: Tudor Ambarus <tudor.ambarus@microchip.com>
Cc: Yoshitaka Ikeda <ikeda@nskint.co.jp>
Signed-off-by: Alexander Dahl <ada@thorsis.com>
Most other spi-mem drivers also depend on SPI_MEM. Fixes this build
error:
arm-v5te-linux-gnueabi-ld.bfd: drivers/spi/atmel-quadspi.o: in function `atmel_qspi_supports_op':
/mnt/data/adahl/src/u-boot/drivers/spi/atmel-quadspi.c:460: undefined reference to `spi_mem_default_supports_op'
make[1]: *** [/mnt/data/adahl/src/u-boot/Makefile:1821: u-boot] Error 1
Signed-off-by: Alexander Dahl <ada@thorsis.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
This is inside of an 'if DM_SPI' block, and thus always true.
Signed-off-by: Alexander Dahl <ada@thorsis.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Show the time taken by each test suite with 'ut all' and the total time
for all suites.
Take care to remove any sandbox time-offset from the values.
Fix the comment-format on timer_test_add_offset() while we are here.
Signed-off-by: Simon Glass <sjg@chromium.org>
emul_rtc_probe() is not called from anywhere else that's why make it
static. Issue is reported by build with W=1.
Signed-off-by: Michal Simek <michal.simek@amd.com>
This patch add support to read register base address by name if
provided.
Also devfdt_get_addr_* is changed to dev_read_addr_* to support DT
live tree.
Signed-off-by: Weijie Gao <weijie.gao@mediatek.com>
The reg-names in mt7988.dtsi have no _base suffix. Remove the suffix
will also make it match upstream linux format.
Fixes: 8c2cb748ef (pinctrl: mediatek: mt7988: rename reg-names to upstream linux format)
Signed-off-by: Weijie Gao <weijie.gao@mediatek.com>
Heiko Schocher <hs@denx.de> says:
In linux we have the option to create the name of a led
optionally through the following properties:
- function
- color
- function-enumerator
This series adds support for parsing this properties if there
is no label property.
Link: https://lore.kernel.org/r/20250128135246.74838-1-hs@denx.de
[trini: Document name parameter in led.h]
in linux we have the option to create the name of a led
optionally through the following properties:
- function
- color
- function-enumerator
This patch adds support for parsing this properties if there
is no label property.
The led name is created in led_post_bind() and we need some
storage place for it. Currently this patch prevents to use
malloc() instead it stores the name in new member :
char name[LED_MAX_NAME_SIZE];
of struct led_uc_plat. While at it append led tests for the
new feature.
Signed-off-by: Heiko Schocher <hs@denx.de>
Reviewed-by: Tom Rini <trini@konsulko.com>
Fixes annoying warnings of the following type when built with W=1 (for
each file including fpga.h):
CC drivers/fpga/fpga.o
In file included from /mnt/data/adahl/src/u-boot/include/xilinx.h:7,
from /mnt/data/adahl/src/u-boot/drivers/fpga/fpga.c:10:
/mnt/data/adahl/src/u-boot/include/fpga.h:61:1: warning: type qualifiers ignored on function return type [-Wignored-qualifiers]
61 | const fpga_desc *const fpga_get_desc(int devnum);
| ^~~~~
/mnt/data/adahl/src/u-boot/include/fpga.h:81:1: warning: type qualifiers ignored on function return type [-Wignored-qualifiers]
81 | const fpga_desc *const fpga_validate(int devnum, const void *buf,
| ^~~~~
/mnt/data/adahl/src/u-boot/drivers/fpga/fpga.c:36:1: warning: type qualifiers ignored on function return type [-Wignored-qualifiers]
36 | const fpga_desc *const fpga_get_desc(int devnum)
| ^~~~~
/mnt/data/adahl/src/u-boot/drivers/fpga/fpga.c:53:1: warning: type qualifiers ignored on function return type [-Wignored-qualifiers]
53 | const fpga_desc *const fpga_validate(int devnum, const void *buf,
| ^~~~~
Do some type and cast cleanup on that fpga_desc type while at it.
Link: https://gcc.gnu.org/onlinedocs/gcc/Warning-Options.html#index-Wignored-qualifiers
Signed-off-by: Alexander Dahl <ada@thorsis.com>
Link: https://lore.kernel.org/r/20250121162213.1477506-1-ada@thorsis.com
Signed-off-by: Michal Simek <michal.simek@amd.com>
Update the debug() print, use __func__ to always print matching
function name, and also print bus name in case there are multiple
busses.
Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
Reviewed-by: Paul Barker <paul.barker.ct@bp.renesas.com>
Explicitly flush icache on the CR52 core before jumping to the next
stage software to make sure it does not contain any invalid content.
Explicitly flash and invalidate dcache on the CA76 core both over the
trampoline buffer and over the CR52 firmware, and then trigger full
system synchronization, to make sure the data surely land in DRAM,
from where the CR52 can surely pick them up.
Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
Christian Marangi <ansuelsmth@gmail.com> says:
This is the last batch of part to push actual support of
OF_UPSTREAM for the mediatek SoC.
The plan is to make the current downstream DTS on part with
upstream implementation so we can permit a gradual transition to
it while we don't cause any regression to any user.
This is to have the same node downstream and upstream.
Mediatek is working hard upstream to also push all the remaining
nodes.
All patch are the final changes after the pinctrl patch
merged previously.
All patch pass CI tests
Link: https://github.com/u-boot/u-boot/pull/731
Link: https://lore.kernel.org/r/20250127134046.26345-1-ansuelsmth@gmail.com
Rename reg-names to upstream linux format. Upstream linux drop the
"_base". To make use of upstream DTSI, align to the upstream naming.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Reviewed-by: Weijie Gao <weijie.gao@mediatek.com>
Tested-by: Weijie Gao <weijie.gao@mediatek.com>
VPL may want to use driver model for MMC even if TPL does not. Update
the rule in this driver to support that.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
This patch fixed the following coding style suggested by checkpatch.pl:
1. Use tab instead of space
2. Use BIT() instead of <<
3. Use mdelay for long time delay
4. Remove useless parenthesises
Signed-off-by: Weijie Gao <weijie.gao@mediatek.com>
The SCSI specification originally required that the second
Command Data Byte contain the LUN value in its high-order bits,
but this field has been marked as reserved since the SCSI-3 spec
from 1996.
Some vendors uses this byte to pass vendor specific data,
and specifying the LUN can trigger strange behaviors.
For the record, this happened on an UFS device where LUN0 was
working perfectly and reading the other LUNs would get the last
buffer data that was read for LUN0, making this issue very very
hard to debug.
It's sane to assume U-Boot will probably never encounter
an SCSI-2 multi-LUN device, if somehow it happens the enquiry
command would need to get the SCSI level to handle this case.
The Linux fix was added in [1] to fix the exact same issue.
[1] https://lore.kernel.org/all/Pine.LNX.4.44L0.1409021108380.2308-100000@iolanthe.rowland.org/
Signed-off-by: Neil Armstrong <neil.armstrong@linaro.org>
STM32 MPU:
- Remove dt-bindings headers available in dts/upstream
- Fixes for stm32prog
- Enable CONFIG_SYS_64BIT_LBA for STM32MP15/13/25 defconfigs
- Add upport of ck_usbo_48m in pre-reloc stage for STM32MP13
- Clean env_get_location() for STM32MP1
- Fix board_get_usable_ram_top() to fix infinite loop in cache
management for STM32MP2.
- Fix ck_flexgen_08 frequency for STM32MP2
STM32 MCU:
- Tune CYCLIC_MAX_CPU_TIME_US to avoid cyclic warning for STM32F469-Disco
- Tune CYCLIC_MAX_CPU_TIME_US to avoid cyclic warning for STM32F769-Disco
If CONFIG_SYS_64BIT_LBA is enable, following compilation warning is
triggered:
CC drivers/fastboot/fb_mmc.o
../drivers/fastboot/fb_mmc.c: In function 'fb_mmc_erase_mmc_hwpart':
../drivers/fastboot/fb_mmc.c:215:35: warning: format '%lu' expects
argument of type 'long unsigned int', but argument 2 has type
'long long unsigned int' [-Wformat=]
215 | printf("........ erased %lu bytes from mmc hwpart[%u]\n",
| ~~^
| |
| long unsigned int
| %llu
216 | dev_desc->lba * dev_desc->blksz, dev_desc->hwpart);
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
| |
| long long unsigned int
../drivers/fastboot/fb_mmc.c: In function 'fb_mmc_boot_ops':
../drivers/fastboot/fb_mmc.c:261:42: warning: format '%lu' expects
argument of type 'long unsigned int', but argument 2 has type
'long long unsigned int' [-Wformat=]
261 | printf("........ wrote %lu bytes to EMMC_BOOT%d\n",
| ~~^
| |
| long unsigned int
| %llu
262 | blkcnt * blksz, hwpart);
| ~~~~~~~~~~~~~~
| |
| long long unsigned int
Signed-off-by: Patrice Chotard <patrice.chotard@foss.st.com>
Reviewed-by: Mattijs Korpershoek <mkorpershoek@baylibre.com>
Acked-by: Mattijs Korpershoek <mkorpershoek@baylibre.com>
Reviewed-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
Some dt-binding headers mask the upstream ones which can lead to build
failures, or worse: super weird bugs, if they get out of sync.
Remove these headers so our devicetree and binding headers will both be
in sync with upstream.
Signed-off-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@foss.st.com>
The TI DP83822 does have support for configurable RGMII RX/TX clock
shift, add support for parsing DT properties which describe the RX/TX
clock shift configuration and configuration of the matching bits in
RCSR register.
The shift is only configurable on DP83822, the other PHYs supported
by this PHY driver, namely DP83825/DP83826 variants, do not implement
this functionality and the RCSR bits used to configure the clock shift
are missing from those PHYs.
The shift is configurable separately for RX and TX path. Each path can
either enable the shift or disable the shift using single bit. In case
the shift is disabled, a delay of 0ns is added to the path, otherwise
a delay of 3.5ns is added to the path.
Note that the two RCSR bits 11 and 12 have inverted logic, RCSR bit 12
enables RX internal shift when SET, while RCSR bit 11 enables TX shift
when UNSET.
Signed-off-by: Marek Vasut <marex@denx.de>
Weijie Gao <weijie.gao@mediatek.com> says:
This patch series add support for MediaTek MT7987 SoC with its reference
boards and related drivers.
This patch series add basic boot support on eMMC/SD/SPI-NOR/SPI-NAND for these
boards. The clock, pinctrl drivers and the SoC initializaton code are also
included.
Link: https://lore.kernel.org/r/cover.1737621362.git.weijie.gao@mediatek.com
