U-Boot now stashes its bootstage buffer into a reserved memory region
whenever CONFIG_BOOTSTAGE_STASH is enabled, just before exiting to the
kernel. This allows a post boot parser to read a unified timeline
(SPL→U-Boot→Kernel→MCU/DSP) directly from DDR, enabling standardized
and repeatable boot-time profiling across releases and SoCs.
Change summary:
Call bootstage_stash_default() in announce_and_cleanup()
when CONFIG_BOOTSTAGE_STASH is set.
Reference boot-time parser utility:
https://github.com/v-singh1/boot-time-parser
Sample boot time report:
+--------------------------------------------------------------------+
am62xx-evm Boot Time Report
+--------------------------------------------------------------------+
Device Power On : 0 ms
SPL Time : 843 ms
U-Boot Time : 2173 ms
Kernel handoff time : 462 ms
Kernel Time : 2522 ms
Total Boot Time : 6000 ms
+--------------------------------------------------------------------+
+--------------------------------------------------------------------+
Bootloader and Kernel Boot Records
+--------------------------------------------------------------------+
BOOTSTAGE_AWAKE = 0 ms (+ 0 ms)
BOOTSTAGE_START_UBOOT_F = 843 ms (+ 0 ms)
BOOTSTAGE_ACCUM_DM_F = 843 ms (+ 0 ms)
BOOTSTAGE_START_UBOOT_R = 1951 ms (+1108 ms)
BOOTSTAGE_ACCUM_DM_R = 1951 ms (+ 0 ms)
BOOTSTAGE_NET_ETH_START = 2032 ms (+ 81 ms)
BOOTSTAGE_NET_ETH_INIT = 2053 ms (+ 21 ms)
BOOTSTAGE_MAIN_LOOP = 2055 ms (+ 2 ms)
BOOTSTAGE_START_MCU = 2661 ms (+606 ms)
BOOTSTAGE_BOOTM_START = 2959 ms (+298 ms)
BOOTSTAGE_RUN_OS = 3016 ms (+ 57 ms)
BOOTSTAGE_BOOTM_HANDOFF = 3016 ms (+ 0 ms)
BOOTSTAGE_KERNEL_START = 3478 ms (+462 ms)
BOOTSTAGE_KERNEL_END = 6000 ms (+2522 ms)
+--------------------------------------------------------------------+
+--------------------------------------------------------------------+
MCU Boot Records
+--------------------------------------------------------------------+
MCU_AWAKE = 2661 ms (+ 0 ms)
BOARD_PERIPHERALS_INIT = 2661 ms (+ 0 ms)
MAIN_TASK_CREATE = 2661 ms (+ 0 ms)
FIRST_TASK = 2662 ms (+ 1 ms)
DRIVERS_OPEN = 2662 ms (+ 0 ms)
BOARD_DRIVERS_OPEN = 2662 ms (+ 0 ms)
IPC_SYNC_FOR_LINUX = 6636 ms (+3974 ms)
IPC_REGISTER_CLIENT = 6636 ms (+ 0 ms)
IPC_SUSPEND_TASK = 6636 ms (+ 0 ms)
IPC_RECEIVE_TASK = 6636 ms (+ 0 ms)
IPC_SYNC_ALL = 6787 ms (+151 ms)
+--------------------------------------------------------------------+
Signed-off-by: Vishnu Singh <v-singh1@ti.com>
In order for the U-Boot specific tests we've added (along with the long
line test) to be run on ".env" files as well, we need to update the line
in the process function that starts to limit the file extensions that we
test on.
Signed-off-by: Tom Rini <trini@konsulko.com>
Rasmus Villemoes <ravi@prevas.dk> says:
This all started from me wondering "why does this standalone app end
up being so huge"? Oh, it essentially links in its own standard C
library to get strlen() and snprintf(). Which then led to asking "why
don't we export all those standard C functions when we have them
anyway?".
CI has chewed on these and seem happy - it was CI which told me about
the necessity of [1/9]: https://github.com/u-boot/u-boot/pull/813
Link: https://lore.kernel.org/r/20250919101002.568488-1-ravi@prevas.dk
There have been quite a few changes to _exports.h since the last
update of XF_VERSION, also before the previous patches in this series.
I doubt the mechanism is actually being used in practice, it is simply
too fragile: Not only does the list of exported functions depend on
.config, so with the same XF_VERSION the jump table entries could have
different offsets. But getting to the jump table itself from gd to
even call the ->get_version() is fragile, since offsetof(gd_t, jt)
can, and does, change. For example, as recently as commit
d990210702 ("global_data: Remove jump table in SPL").
One really must build one's standalone app against the proper U-Boot
version and config.h. But for good measure, do bump it now.
Signed-off-by: Rasmus Villemoes <ravi@prevas.dk>
Reviewed-by: Tom Rini <trini@konsulko.com>
The current list of exported functions lacks quite a few bog-standard C
library functions that we might as well expose, since U-Boot certainly
has them implemented anyway. There's no reason a standalone
application should have its own strlen() implementation or link in a
copy from some tiny libc.
For a customer's standalone app, this means it goes from 95K to 10K.
More importantly, we can ditch the custom toolchain including a
newlibc used to build the standalone app and just use the same
toolchain as used to build u-boot itself.
Signed-off-by: Rasmus Villemoes <ravi@prevas.dk>
Reviewed-by: Tom Rini <trini@konsulko.com>
The current list of exported functions is somewhat of a
mess. Reorganize them so that related functionality is kept together:
- console I/O: move vprintf next to printf and the getc/putc functions
- integer parsing: move the *strto* functions together
- standard string.h stuff: move memset() and strcpy() next to strcmp()
- time: move mdelay() next to udelay() and get_timer()
Signed-off-by: Rasmus Villemoes <ravi@prevas.dk>
Reviewed-by: Tom Rini <trini@konsulko.com>
After finishing a later patch in this series, I discovered I had
neglected to update the list of declarations in exports.h to
match. But then I realized I wasn't the first to do that.
Use the existing mechanism and DRY it out.
Signed-off-by: Rasmus Villemoes <ravi@prevas.dk>
Reviewed-by: Tom Rini <trini@konsulko.com>
The !CONFIG_PHY_AQUANTIA defines were already superfluous since
_exports.h does have a CONFIG_PHY_AQUANTIA, so the entries never
existed. In fact, it couldn't have worked, because the defines would
affect both occurences of the mdio_get_current_dev identifier in the
EXPORT_FUNC(mdio_get_current_dev, struct mii_dev *,
mdio_get_current_dev, void)
so the C code would end up containing four copies of
gd->jt->dummy = dummy
but struct jt_funcs would not and does not have any 'dummy' member.
Now that nothing in _exports.h refers to dummy(), remove the empty
function.
Signed-off-by: Rasmus Villemoes <ravi@prevas.dk>
Reviewed-by: Tom Rini <trini@konsulko.com>
As for the i2c functions, drop the dummy entries that, if ever used,
would just have the standalone app get some random content in the
return register.
While deprecated, the spi_{setup,free}_slave functions do exist even
with CONFIG_DM_SPI - and a standalone app can't really do anything but
refer to a spi device via a (bus, cs) pair.
Eventually, one should probably export some function that could allow
a standalone app to get a struct udevice* corresponding to either a
full DT path, an alias, or perhaps a label (provided one builds with
-@), and then export functions that can operate on that.
Signed-off-by: Rasmus Villemoes <ravi@prevas.dk>
Reviewed-by: Tom Rini <trini@konsulko.com>
There's really no good reason to create stub entries that would call a
function that doesn't even return anything sensible.
The existence of these two i2c_* functions depends on
CONFIG_IS_ENABLED(SYS_I2C_LEGACY), which does depend on !DM_I2C, but
is not equivalent to it. They are probably rather hard to use unless
CMD_I2C and something in U-Boot has called "i2c dev foo" to set the
current i2c bus before calling the standalone app, so keep that
dependency.
Signed-off-by: Rasmus Villemoes <ravi@prevas.dk>
Reviewed-by: Tom Rini <trini@konsulko.com>
Currrently, malloc and free are function-like macros, while calloc,
realloc and memalign are object-like macros.
Usually, this doesn't matter, but it does when the identifiers appear
without a following open parenthesis, such as when their address is
taken for building the export table. Adding calloc or realloc to that
table breaks the build on sandbox due to this inconsistency.
Signed-off-by: Rasmus Villemoes <ravi@prevas.dk>
Reviewed-by: Tom Rini <trini@konsulko.com>
Add PCIE1 and PCIE2 ECAM space and outbound space to MMU pagetable,
so A55 can access them.
Signed-off-by: Ye Li <ye.li@nxp.com>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
Add i.MX8ULP flexspi compatible string and driver data.
The flexspi on i.MX8ULP only has 16 LUT sequences and uses 1KB RX FIFO.
Signed-off-by: Ye Li <ye.li@nxp.com>
According to i.MX8DXL A1 errata ERR050601, concurrent read accesses
from the A35 cores to the peripherals within the LSIO subsystem
(region 0_5DXX_XXXX) and address spaces in the regions
[0_0000_0000 – 0_1BFF_FFFF] and [4_0000_0000 – 4_3FFF_FFFF] can collide
and cause data corruption in the returned data, with no failure report.
Even a single A35 core accessing both these regions can trigger the issue
because an A35 core can have more than one parallel read operation in
progress.
The flexspi0 AHB memory is in LSIO region mentioned in above errata.
So we can't use AHB read, only can read data from FIFO.
Add the compatible string for 8DXL and use a flag for the IPS read.
Signed-off-by: Ye Li <ye.li@nxp.com>
Use a dedicated LUT (second last) for AHB read command, so we can
directly read from the AHB memory-mapped address and booting M core
for XIP on Flexspi NOR.
Signed-off-by: Ye Li <ye.li@nxp.com>
Set ethprime to eth0 since FEC interface is considered the primary IF on
phyCORE-i.MX93 SoM based boards. This comes from the fact that the same
bootloader is reused for both carrier boards, that is phyBOARD-Segin and
phyBOARD-Nash which both use different Ethernet PHYs on the EQOS (eth1)
interface and thus eth1 cannot be used as a prime.
Signed-off-by: Primoz Fiser <primoz.fiser@norik.com>
Add board-level code and defconfig for the i.MX943 EVK board, supporting
multiple SOM variants: 19x19 LPDDR5, 19x19 LPDDR4 and 15x15 LPDDR4.
Signed-off-by: Ye Li <ye.li@nxp.com>
Signed-off-by: Alice Guo <alice.guo@nxp.com>
Acked-by: Peng Fan <peng.fan@nxp.com>
Reviewed-by: Jacky Bai <ping.bai@nxp.com>
Introduce the base dtsi files for the i.MX943 EVK board. These files
define the essential components such as messaging units, uSDHC, GPIOs
and lpuart for board bring-up.
Signed-off-by: Ye Li <ye.li@nxp.com>
Signed-off-by: Alice Guo <alice.guo@nxp.com>
Acked-by: Peng Fan <peng.fan@nxp.com>
Reviewed-by: Jacky Bai <ping.bai@nxp.com>
The register reflects lifecycle and some lifecycle-derived state of
i.MX95 has new offset address and layout, so display_life_cycle() is
added specifically for it.
Signed-off-by: Ye Li <ye.li@nxp.com>
Signed-off-by: Alice Guo <alice.guo@nxp.com>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
Due to i.MX95 has reserved first 256MB DDR, change to use the DDR
start address in u-boot as the container header buffer.
Signed-off-by: Ye Li <ye.li@nxp.com>
Signed-off-by: Alice Guo <alice.guo@nxp.com>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
Define the daisy register offset for i.MX94 at 0x608 within the iomuxc
register space. This enables correct pad selection for daisy chain
configuration on i.MX94 platforms.
Signed-off-by: Jacky Bai <ping.bai@nxp.com>
Signed-off-by: Alice Guo <alice.guo@nxp.com>
Acked-by: Peng Fan <peng.fan@nxp.com>
- Add base addresses for WDG3, WDG4, GPIO6, and GPIO7 for i.MX94.
- Introduce common.h with macros of clock IDs, power domains, and CPU
types for platform-specific replacement (e.g., i.MX94, i.MX95).
- Extend imx_get_mac_from_fuse() to support i.MX94.
Signed-off-by: Ye Li <ye.li@nxp.com>
Signed-off-by: Alice Guo <alice.guo@nxp.com>
Acked-by: Peng Fan <peng.fan@nxp.com>
Reviewed-by: Jacky Bai <ping.bai@nxp.com>
Update get_reset_reason() to support i.MX94 to send message to the
System Manager to retrieve the LM/system last booted/shutdown reasons.
Signed-off-by: Ye Li <ye.li@nxp.com>
Signed-off-by: Alice Guo <alice.guo@nxp.com>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
System Manager provides the last booted and shutdown reasons of the
logical machines (LM) and system using the SCMI misc protocol (Protocol
ID: 0x84, Message ID: 0xA). This path adds get_reset_reason() to query
and print these reasons in SPL and U-Boot.
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Signed-off-by: Alice Guo <alice.guo@nxp.com>
Reviewed-by: Ye Li <ye.li@nxp.com>
Add a case for i.MX94 to return the correct string identifier in the
get_imx_type_str() function. This ensures proper CPU type reporting for
i.MX94 platforms.
Signed-off-by: Alice Guo <alice.guo@nxp.com>
Introduce support for the new i.MX94 processor, including its CPU type
and SoC-level Kconfig entry.
The i.MX94 is a new member of the i.MX9 family. It uses a System Manager
to handle system-level functions such as power, clock, sensor and pin
control. The System Manager runs on a Cortex-M processor, while the
Cortex-A processor communicates with it via the ARM SCMI protocol and a
messaging unit.
Signed-off-by: Ye Li <ye.li@nxp.com>
Signed-off-by: Alice Guo <alice.guo@nxp.com>
Acked-by: Peng Fan <peng.fan@nxp.com>
Reviewed-by: Jacky Bai <ping.bai@nxp.com>
A few of the USB musb-new host glue drivers cannot build without access
to some platform specific header files. Express those requirements in
Kconfig as well.
Reviewed-by: Marek Vasut <marek.vasut@mailbox.org>
Signed-off-by: Tom Rini <trini@konsulko.com>
A few of the USB host drivers cannot build without access to some
platform specific header files. Express those requirements in Kconfig as
well.
Reviewed-by: Marek Vasut <marek.vasut@mailbox.org>
Signed-off-by: Tom Rini <trini@konsulko.com>
This driver was never enabled by any platforms after being added to the
tree over 5 years ago. Remove it.
Signed-off-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Marek Vasut <marek.vasut@mailbox.org>
This driver is unused since the removal of the bcm28155_ap board in
commit 0f6807e77b ("arm: Remove bcm28155_ap board"). Remove it.
Fixes: 0f6807e77b ("arm: Remove bcm28155_ap board")
Signed-off-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Marek Vasut <marek.vasut@mailbox.org>
This driver requires some mach-at91 specific header files in order to build.
Express that requirement in Kconfig as well.
Reviewed-by: Marek Vasut <marek.vasut@mailbox.org>
Signed-off-by: Tom Rini <trini@konsulko.com>
A few of the platform specific DWC3 host glue drivers cannot build
without access to some platform specific header files. Express those
requirements in Kconfig as well.
Reviewed-by: Marek Vasut <marek.vasut@mailbox.org>
Signed-off-by: Tom Rini <trini@konsulko.com>
This was apparently not built for several years: Since a2535243e0,
optee_copy_fdt_nodes implicitly works against the U-Boot dt. We
therefore have to tweak its reference before using the function and
restore things afterwards.
If it had been built, actually trying it out would have failed next: We
need CONFIG_OPTEE_LIB to actually build the function that is primarily
being tested here. And we need to re-initialize target fdt, now that the
tests may run in random order.
Fixes: a2535243e0 ("lib: optee: migration optee_copy_fdt_nodes for OF_LIVE support")
Fixes: ba2feaf414 ("test: Split optee tests into three functions")
Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
The sandbox time offset is intended to be controlled via the U-Boot
environment, not the host process environment. Update os_get_time_offset()
to use env_get() instead of the libc getenv().
Leave other getenv() uses (e.g. U_BOOT_PERSISTENT_DATA_DIR,
UBOOT_SB_FUZZ_TEST) unchanged, since those refer to host environment
variables needed by sandbox tests.
Signed-off-by: Osama Abdelkader <osama.abdelkader@gmail.com>
The android boot header is page aligned but the current code made the
assumption that the header was always smaller than the current header
format.
When the page_size is defined as 2048, as this is the case with the
cuttlefish target, the current code sets the end of the header in the
middle of it as the v3 and v4 headers are respectively 2112 and 2128
bytes long.
Fix that by aligning to page_size
Fixes: 1115027d2f ("android: boot: update android_image_get_data to support v3, v4")
Signed-off-by: Guillaume Ranquet <ranquet.guillaume@gmail.com>
Fix a couple of style issues reported by checkpatch.pl:
- Replace `#ifdef CONFIG_MMC_TRACE` with `#if IS_ENABLED(CONFIG_MMC_TRACE)`
to follow the preferred kernel style for config-dependent branches.
- Drop explicit zero initialization of a static variable.
No functional change intended.
Signed-off-by: Bhimeswararao Matsa <bhimeswararao.matsa@gmail.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
In regulator_list_autoset there is a test for ret being non-zero and
error being zero but it uses the binary '&' instead of the logical '&&'
which could well lead to unexpected results. Correct this to use the
logical '&&' instead.
This issue found by Smatch.
Signed-off-by: Andrew Goodbody <andrew.goodbody@linaro.org>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
This reverts commit 4d84fa1261, reversing
changes made to b82a1fa7dd.
I had missed some feedback on this series from earlier, and we have
since had reports of regressions due to this as well. For now, revert
this.
Signed-off-by: Tom Rini <trini@konsulko.com>
Aristo Chen <jj251510319013@gmail.com> says:
This patch series enhances FIT image robustness by adding **memory
region overlap detection** to `mkimage` and fixing existing overlaps
in DTS files and `binman` tests.
The primary goal is to prevent runtime memory corruption from
conflicting load addresses in FIT images.
Key Changes:
1. `mkimage` Overlap Detection: A new validation in
`tools/fit_image.c` checks for overlapping load addresses
within FIT configurations. `mkimage` now errors out with
detailed info on conflicts, preventing bad FIT image creation.
2. New Test Case: A Python test verifies the new detection.
It intentionally creates an overlap (kernel and FDT)
to confirm correct error handling.
3. Fixes for Existing Overlaps:
* Board DTS (k3-am6xx): Adjusted load addresses for TI
firmware stubs to prevent conflicts. This resolves
previously undetected overlaps.
* `binman` Tests: Fixed several tests. U-Boot load
addresses were shifted to avoid ATF conflicts. A new
linker script for TEE ELF sections ensures distinct
memory layouts.
4. Documentation: Added guidance for developers on how to
determine ELF load addresses using readelf, linker scripts,
and objdump when working with binman FIT images.
Impact:
This series improves FIT image reliability by catching overlaps
at build time, helping developers resolve issues before runtime
failures.
Link: https://lore.kernel.org/r/20250914110021.4103-1-aristo.chen@canonical.com
Vishnu SinghandTom Rini
Marek VasutandTom Rini
Michal SimekandTom Rini
Rasmus VillemoesandTom Rini
Ye LiandFabio Estevam
Primoz FiserandFabio Estevam
Alice GuoandFabio Estevam
Peng FanandFabio Estevam
Jan KiszkaandTom Rini
Osama AbdelkaderandTom Rini
Guillaume RanquetandTom Rini
Bhimeswararao MatsaandPeng Fan
Andrew GoodbodyandPeng Fan