In commit 6e4675b8e5 ("lmb: replace the lmb_alloc() and
lmb_alloc_base() API's") an additional allocation was mistakenly
introduced resulting in ${kernel_comp_size} containing the address of a
second 64mb region rather than the actual value of KERNEL_COMP_SIZE.
Additionally, in commit b40d7b8f72 ("Merge patch series "lmb: use
a single API for all allocations"") merge conflict resulted in an
additional 128mb allocation for ${loadaddr} when CONFIG_FASTBOOT
is enabled, where it should actually be set to the same value as
${fastboot_addr_r} to respect size constraints (and since it doesn't
seem to interfer with any bootflows).
Fixup both of these, freeing up 192mb of memory.
Fixes: 6e4675b8e5 ("lmb: replace the lmb_alloc() and lmb_alloc_base() API's")
Fixes: b40d7b8f72 ("Merge patch series "lmb: use a single API for all allocations"")
Link: https://lore.kernel.org/r/20250811104710.1896382-1-casey.connolly@linaro.org
Signed-off-by: Casey Connolly <casey.connolly@linaro.org>
In mci_send_cmd the pointer 'data' is optional so guard its use with a
NULL check to prevent any attempt to dereference it when not provided.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <andrew.goodbody@linaro.org>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
Remove duplicate checks on status from mci_data_read and mci_data_write
which are guaranteed to be true as exiting the above do..while loop
above requires that to be so.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <andrew.goodbody@linaro.org>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
If the test in the for loop is never matched then the variable 'match'
will never be assigned to. Provide an initial value so this cannot be a
problem.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <andrew.goodbody@linaro.org>
The use of the unsigned variable 'i' as a loop index leads to the test
for i being non-negative always being true. Instead declare 'i' as an
int so that the for loop will terminate as expected.
If the original for loop completes 'i' will be 1 past the end of the
array so decrement it in the subsequent error path to prevent an out of
bounds access occurring.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <andrew.goodbody@linaro.org>
The variable 'index' is declared as unsigned but used to receive the
return value of a function returning 'int'. This value is then tested
for being less than zero to detect an error condition but as index is
unsigned this can never be true. Change the variable 'index' to be an
int so that the error condition can be detected.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <andrew.goodbody@linaro.org>
In atmel_qspi_transfer(), the status register is polled with:
imr = QSPI_SR_INSTRE | QSPI_SR_CSR;
return readl_poll_timeout(aq->regs + QSPI_SR, sr,
(sr & imr) == imr,
ATMEL_QSPI_TIMEOUT);
However, this is racy: QSPI_SR_INSTRE can be set before QSPI_SR_CSR,
and will then be cleared by the read. If that happens, the condition
"(sr & imr) == imr" can never be true, and the function times out.
This race condition is avoided in at91bootstrap by accumulating the
status bits across reads until both bits have been observed:
/* Poll INSTruction End and Chip Select Rise flags. */
imr = (QSPI_SR_INSTRE | QSPI_SR_CSR);
sr = 0;
do {
udelay(1);
sr |= qspi_readl(qspi, QSPI_SR) & imr;
} while ((--timeout) && (sr != imr));
Update U-Boot's atmel_qspi_transfer() to use the same pattern,
ensuring that both flags are observed even if they are not set
simultaneously.
Signed-off-by: Ramin Moussavi <lordrasmus@gmail.com>
[eugen.hristev@linaro.org: remove 'sr' and fix commit msg]
Signed-off-by: Eugen Hristev <eugen.hristev@linaro.org>
These changes are mostly smaller fixes, and some additions that were
dependent on recent DT updates: We switch the S3/V3/V3s over to
OF_UPSTREAM, the DTs were already identical anyway. While at it, Paul
sent more fixes for this SoC, I am pulling in two easy fixes that were
ready and low risk. Apart from other assorted fixes, this PR also
enables Ethernet on the new A527/T527 boards, made possible by the DT
update from the DT rebasing repository.
This passed the CI, and was boot tested on boards with Allwinner A10,
A20, A33, V40, A80, A83T, T113s3, F1C100s, H3, A64, H6, H616, A133 and
T527 SoCs.
The JEDEC LPDDR3 spec defines mode register 0 (MR0) as being read-only,
so there is no point in trying to set its value.
Also the H616 memory controller encodes the mode register index to be
written starting from bit 8 in MRCTRL1 (for LPDDR3 and LPDDR4 chips), so
we need to OR in that number to tell the controller which MR to program.
On top of that, the mode registers between DDR3 and LPDDR3 are
completely different, so writing values crafted for DDR3 into a LPDDR3
chip is just wrong. Due to the above mentioned bugs the writes for
MR0-MR2 did not have any effect (as they were all trying to set the
read-only MR0), so the mode registers just stayed unchanged.
Looking at the LPDDR3 spec and the BSP code, let's write the proper MR
values into LPDDR3 chips, using the proper addressing mode.
Use the opportunity to document the LPDDR3 mode register bits written.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Jernej Skrabec <jernej.skrabec@gmail.com>
58e9502e6 "arm: sunxi: Enable SPL/SPI boot for Olinuxino Lime2-eMMC
boards" enabled SPI boot for the eMMC variant. Olimex offers the
"s16MB" variant with SPI flash populated but without eMMC populated.
Tested on board rev L.
Signed-off-by: Eric Anderson <ejona86@gmail.com>
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Recent changes in the H6 clock code added delay() calls into the SPL clock
setup routine, which requires the timers to work. When compiling for
AArch64, we are always using the Arm Generic Timer (aka. arch timer),
which does not require further setup, hence having an empty timer_init()
routine.
However for 32-bit SoCs we use the Allwinner timers, which require some
setup routine, and hence we need timer_init() to be called before
clock_init().
Swap the order of the two calls, to be more robust when compiling the H6
clock code for AArch32 or when using the Allwinner timers for whatever
reason.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Jernej Skrabec <jernej.skrabec@gmail.com>
The pinecube has an ethernet connector which uses the EMAC and internal
PHY of the chip. Enable it in the config.
Reviewed-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Paul Kocialkowski <contact@paulk.fr>
Variables holding addresses are typically using the "long" C type in
U-Boot, to be easily compatible with both 32-bit and 64-bit builds.
The A133 DRAM driver is typically compiled for AArch64, so u64 is the
same type as unsigned long, but that breaks when compiling the DRAM
driver in AArch32 (for some experiments).
Fix the type to make the code more portable.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Jernej Skrabec <jernej.skrabec@gmail.com>
The first of the two Ethernet controllers in the Allwinner A527/T527 is
compatible to the MAC from the previous SoCs. Consequently the recent DT
update brought use the MAC node, using the A64 compatible string as the
fallback, which works out of the box.
Enable the sun8i-emac Ethernet driver in the defconfig for the Radxa and
the Avaota boards, so that kernels and other data can be loaded via
TFTP. Please note that only one of the Ethernet sockets will work, the
second MAC is not compatible, and needs a new U-Boot driver. The
X96QPro+ TV box unfortunately uses only this second EMAC, so Ethernet
cannot be used there at the moment.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Jernej Skrabec <jernej.skrabec@gmail.com>
The name of the pin function was changed last minute in the DT, from
emac0 to gmac0.
Adjust the name we use in the pinctrl driver accordingly.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Jernej Skrabec <jernej.skrabec@gmail.com>
The V3s (using co-packaged DRAM) runs at 360 MHz, which is specified in
the common platform Kconfig file. However the value for MACH_SUN8I will
be picked up instead due to ordering.
Re-order the defaults to have MACH_SUN8I_V3S before MACH_SUN8I and let
it select the correct default. Also update the LicheePi Zero Dock
defconfig to remove the value, which is now correctly selected.
Signed-off-by: Paul Kocialkowski <contact@paulk.fr>
Reviewed-by: Andre Przywara <andre.przywara@arm.com>
There is nothing special for u-boot in the V3/V3s device-tree files,
they are just copies of the upstream ones.
Remove the copies and switch to OF_UPSTREAM for supported boards.
Signed-off-by: Paul Kocialkowski <contact@paulk.fr>
Reviewed-by: Andre Przywara <andre.przywara@arm.com>
The Linux pinctrl-sx150 was originally written as a GPIO driver
and fully rewritten by me as a Pinctrl driver and extended by
other contributors.
Fixup the Copyright header style and correctly report the
Copyright headers from the Linux driver.
Signed-off-by: Neil Armstrong <neil.armstrong@linaro.org>
In gpio_dwapb_bind plat is used to reference memory allocated by
devm_kcalloc but it is attempted to be freed using kfree. Instead free
this memory using the correct devm_kfree function.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <andrew.goodbody@linaro.org>
Acked-by: Quentin Schulz <quentin.schulz@cherry.de>
The function dev_read_u32_default does not return an error and the
variable 'val' is unsigned so testing for >= 0 will always be true. It
looks like the code was attempting to return -1 if xtal-load-pf was not
present but that cannot work. Instead use dev_read_u32 which returns an
error code separately from writing the value into the passed pointer.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <andrew.goodbody@linaro.org>
Fixes: 260777fc23 ("clk: cdce9xx: add support for cdce9xx clock synthesizer")
Acked-by: Quentin Schulz <quentin.schulz@cherry.de>
This adds more common functions found on other architectures that will
allow for more compile-testing of drivers. These are either dummy
functions as we do not need them or mappings to existing functions,
similar to how other architectures handle it.
Signed-off-by: Tom Rini <trini@konsulko.com>
A large number of PCI controllers cannot build without access to some
platform specific header files. Express those requirements in Kconfig as
well.
Signed-off-by: Tom Rini <trini@konsulko.com>
This driver references the SZ_ macros while relying on an indirection
inclusion of <linux/sizes.h>. Add the missing include directly.
Signed-off-by: Tom Rini <trini@konsulko.com>
Allows run-time control over watchdog auto-start and the timeout via
setting the environment variable watchdog_timeout_ms. A value of zero
means "do not start". Use CONFIG_WATCHDOG_TIMEOUT_MSECS as initial value.
Users can enable the watchdog to monitor the boot process until userspace
or OS takes over to serve the watchdog.
Signed-off-by: Wadim Egorov <w.egorov@phytec.de>
We switched towards standard boot with still keeping a fallback
using legacy boot command alive. Add a deprecation warning to
make it more clear that we will remove it in future versions.
Signed-off-by: Wadim Egorov <w.egorov@phytec.de>
Allows run-time control over watchdog auto-start and the timeout via
setting the environment variable watchdog_timeout_ms. A value of zero
means "do not start". Use CONFIG_WATCHDOG_TIMEOUT_MSECS as initial value.
Users can enable the watchdog to monitor the boot process until userspace
or OS takes over to serve the watchdog.
Signed-off-by: Wadim Egorov <w.egorov@phytec.de>
Allows run-time control over watchdog auto-start and the timeout via
setting the environment variable watchdog_timeout_ms. A value of zero
means "do not start". Use CONFIG_WATCHDOG_TIMEOUT_MSECS as initial value.
Users can enable the watchdog to monitor the boot process until userspace
or OS takes over to serve the watchdog.
Signed-off-by: Wadim Egorov <w.egorov@phytec.de>
The watchdog driver probes all available watchdog devices.
This causes SMP boot errors when bringing up secondary CPUs.
In our setup, only a single watchdog is needed to monitor the
boot process until userspace or the OS takes over. Disable all
unnecessary watchdog devices in U-Boot to avoid conflicts
during CPU bring-up.
Signed-off-by: Wadim Egorov <w.egorov@phytec.de>
The watchdog driver probes all available watchdog devices.
This causes SMP boot errors when bringing up secondary CPUs.
In our setup, only a single watchdog is needed to monitor the
boot process until userspace or the OS takes over. Disable all
unnecessary watchdog devices in U-Boot to avoid conflicts
during CPU bring-up.
Signed-off-by: Wadim Egorov <w.egorov@phytec.de>
The watchdog driver probes all available watchdog devices.
This causes SMP boot errors when bringing up secondary CPUs.
In our setup, only a single watchdog is needed to monitor the
boot process until userspace or the OS takes over. Disable all
unnecessary watchdog devices in U-Boot to avoid conflicts
during CPU bring-up.
Signed-off-by: Wadim Egorov <w.egorov@phytec.de>
In order to set the BOOTFILE symbol we first need to have USE_BOOTFILE
be set, or some of the logic might not work as expected later on when
building. Second, defaults like this belong with the symbol itself.
Fixes: da595d236b ("include: configs: soc64: Use CONFIG_SPL_ATF to differentiate bootfile")
Signed-off-by: Tom Rini <trini@konsulko.com>
This pull request introduces initial U-Boot support for Agilex7 M-series, along
with several enhancements and cleanups across existing Agilex platforms. Key
changes include new board support, DDR driver additions, updated device trees,
and broader SoCFPGA SPL improvements.
Highlights:
- Agilex7 M-series bring-up:
- Basic DT support and board initialization for Agilex7 M-series SoC and
SoCDK.
- New sdram_agilex7m DDR driver with UIBSSM mailbox support and HBM support.
- Clock driver support for Agilex7 M-series.
- New defconfig: socfpga_agilex7m_defconfig.
- Agilex and Agilex5 enhancements:
- Improved SPL support: ASYNC interrupt enabling, system manager init
refactor, and cold scratch register usage.
- Updated firewall probing and watchdog support in SPL.
- Cleaned up DDR code, added secure region support for ATF, and improved warm
reset handling.
- Device Tree and config updates:
- Migration to upstream Linux DT layout for Agilex platforms.
- Consolidated socfpga_agilex_defconfig and removed deprecated configs.
- Platform-specific environment variables for Distro Boot added.
- Driver fixes and cleanups:
- dwc_eth_xgmac and clk-agilex cleanup and improvements.
- Several coverity and style fixes.
Contributions in this PR are from Alif Zakuan Yuslaimi, Tingting Meng, and
Andrew Goodbody. This patch set has been tested on Agilex 5 devkit, Agilex
devkit and Agilex7m devkit.
Passing all pipeline tests at SoCFPGA U-boot custodian
https://source.denx.de/u-boot/custodians/u-boot-socfpga/-/pipelines/27318
Pull request efi-2025-10-rc2
Documentation:
* Describe running U-Boot as a pflash image on RISC-V QEMU
* Correct the description of the bootph-verify device-tree property
UEFI:
* In mkeficapsule fix a resource leak in read_bin_file()
* Support loading a Linux initial RAM disk with the bootefi command.
use the new email address for community work.
While at it, cleanup git shortlog output, by adding
fixes in .mailmap
Signed-off-by: Heiko Schocher <hs@nabladev.com>
CI: https://source.denx.de/u-boot/custodians/u-boot-imx/-/pipelines/27314
- Several Smatch reported fixes.
- Enable the temperature command on imx8ulp-evk.
- Fix mx8mm_fracpll_tbl.
- Make optee packaging optional for imx8m.
- Reuse and export low_drive_freq_update() on imx9.
- Enable USB OTG ID pin pull up in SPL on dh-imx6.
This is for new platform enablement for Agilex7 M-series.
Add DDR driver for Agilex7 M-series. This driver is designed to support
DDR and HBM memory. The official HBM handoff is not ready yet, therefore
hardcoded handoff is used for HBM driver validation on mUDV board.
Signed-off-by: Tingting Meng <tingting.meng@altera.com>
Signed-off-by: Alif Zakuan Yuslaimi <alif.zakuan.yuslaimi@altera.com>
Reviewed-by: Tien Fong Chee <tien.fong.chee@altera.com>
