| Age | Commit message (Collapse) | Author |
|---|
|
- XHCI DMA bugfix
|
|
Sparrow Hawk
The 9FGV0441 PCIe clock generator can operate in autonomous mode, which
is the default mode. U-Boot currently does not have a driver for this
PCIe clock generator, but Linux 6.17 DT does describe the clock generator
in Sparrow Hawk board DT and this DT is included in U-Boot since commit
eea470fd7f6a ("Subtree merge tag 'v6.17-dts' of dts repo [1] into dts/upstream").
Reinstate basic PCIe clock description which matches the behavior of
Linux DT before Linux 6.17.y release in in U-Boot DT extras to allow
PCIe to be used on Sparrow Hawk board in U-Boot until the 9FGV0441
driver gets implemented or ported from Linux.
Signed-off-by: Marek Vasut <[email protected]>
|
|
When aborting a Transfer Descriptor (TD), the xHCI driver updates the
device dequeue pointer by converting the virtual enqueue TRB pointer
into a DMA address.
Previously, the code OR-ed the ring's Dequeue Cycle State (DCS) bit into
the virtual TRB pointer before passing it to xhci_trb_virt_to_dma().
This produced an unaligned virtual address (e.g. ending in 0x...1).
Inside xhci_trb_virt_to_dma(), the offset calculation:
segment_offset = trb - seg->trbs;
operated on this unaligned pointer, resulting in an incorrect TRB index.
In wraparound cases, this caused the bounds check to fail and the
function to return 0.
As a result, a SET_DEQ_PTR command was issued with a DMA address of 0x0,
leading to controller hangs and transfer timeouts, most commonly when
aborting TDs near the end of a ring segment (e.g. index 63).
Fix this by translating the aligned virtual TRB pointer to a DMA address
first, and only then applying the DCS bit to the resulting physical
address.
Reviewed-by: Marek Vasut <[email protected]>
Signed-off-by: ANANDHAKRISHNAN S <[email protected]>
|
|
When building with a toolchain that uses a modified sysroot (such as a
Yocto-generated SDK) that does not include libyaml, on a host that does
have libyaml, building dtc will fail with errors like:
HOSTLD scripts/dtc/dtc
/usr/lib/gcc/x86_64-pc-linux-gnu/14/../../../../x86_64-pc-linux-gnu/bin/ld:
scripts/dtc/yamltree.o: in function `yaml_propval_int':
yamltree.c:(.text+0x167): undefined reference to
`yaml_sequence_start_event_initialize'
/usr/lib/gcc/x86_64-pc-linux-gnu/14/../../../../x86_64-pc-linux-gnu/bin/ld:
yamltree.c:(.text+0x172): undefined reference to `yaml_emitter_emit'
/usr/lib/gcc/x86_64-pc-linux-gnu/14/../../../../x86_64-pc-linux-gnu/bin/ld:
yamltree.c:(.text+0x1e8): undefined reference to
`yaml_scalar_event_initialize'
/usr/lib/gcc/x86_64-pc-linux-gnu/14/../../../../x86_64-pc-linux-gnu/bin/ld:
yamltree.c:(.text+0x1f5): undefined reference to `yaml_emitter_emit'
(... rest of errors truncated ...)
This happens because the test looks for the file in the default path but
uses pkg-config, which is affected by changing sysroot, to determine the
correct linker arguments. This does not happen when building entirely
within yocto, as pseudo will intercept and rewrite the file path when
trying to test for /usr/include/yaml.h to match the sysroot and thus
generate consistent behavior.
This commit adds the PKG_CONFIG_SYSROOT_DIR prefix to the file path
in order to test against the same conditions that are used to resolve
the build flags for libyaml.
In linux commit ef8795f3f1c ("dt-bindings: kbuild: Use DTB files for
validation"), including yaml is disabled again anyway because of other
problems that it causes, so this problem can also be addressed by
partially backporting that commit instead and simply disabling the yaml
support.
Fixes: 0535e46d55d7 ("scripts/dtc: Update to upstream version v1.7.2-35-g52f07dcca47c")
Signed-off-by: Greg Malysa <[email protected]>
|
|
A TrueType font for U-Boot should fulfill the following requirements:
* mono spaced
* support full code page 437
* easily readable
Unfortunately none of the fonts provided with U-Boot fulfills all of these
requirements.
Let's add the DejaVu Mono font. To reduce the code size the characters are
limited to code page 437.
Signed-off-by: Heinrich Schuchardt <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
|
|
Greg Malysa <[email protected]> says:
This series adds the final pieces to enable mainline U-Boot to build and
boot all Analog Devices SC5xx SoCs and supports the associated carrier
board options. At this point it should be viable for new users for these
platforms to start with the latest version of U-Boot rather than our
vendor fork, however some features (such as OSPI support and falcon
boot) remain unavailable until we are able to unify our implementations
with the mainline implementations.
Link: https://lore.kernel.org/r/[email protected]
[trini: Rebuild CI containers to have new tools]
Signed-off-by: Tom Rini <[email protected]>
|
|
This adds missing maintainers entries for the ADI SC5xx defconfigs and
for a device tree binding file that was previously missed.
Signed-off-by: Greg Malysa <[email protected]>
|
|
This adds support for the Analog Devices SC594 SOM and configurations
for using it with both the SOMCRR-EZKIT and SOMCRR-EZLITE.
Signed-off-by: Vasileios Bimpikas <[email protected]>
Signed-off-by: Utsav Agarwal <[email protected]>
Signed-off-by: Arturs Artamonovs <[email protected]>
Signed-off-by: Nathan Barrett-Morrison <[email protected]>
Signed-off-by: Caleb Ethridge <[email protected]>
Signed-off-by: Philip Molloy <[email protected]>
Signed-off-by: Greg Malysa <[email protected]>
|
|
This adds support for the Analog Devices SC598-SOM and configurations
for using it with both the SOMCRR-EZKIT and SOMCRR-EZLITE. This adds
dtsis for both Rev D (including older revisions) and Rev E SOMs, which
are not compatible due to BOM changes. Although no new Rev D SOMs are
produced as of 2025, many are in circulation, so the RevD dtsi is
included to facilitate use for existing customers.
Signed-off-by: Vasileios Bimpikas <[email protected]>
Signed-off-by: Utsav Agarwal <[email protected]>
Signed-off-by: Arturs Artamonovs <[email protected]>
Signed-off-by: Nathan Barrett-Morrison <[email protected]>
Signed-off-by: Caleb Ethridge <[email protected]>
Signed-off-by: Philip Molloy <[email protected]>
Signed-off-by: Greg Malysa <[email protected]>
|
|
This adds support for the Analog Devices SC584-EZKIT.
Signed-off-by: Vasileios Bimpikas <[email protected]>
Signed-off-by: Utsav Agarwal <[email protected]>
Signed-off-by: Arturs Artamonovs <[email protected]>
Signed-off-by: Nathan Barrett-Morrison <[email protected]>
Signed-off-by: Caleb Ethridge <[email protected]>
Signed-off-by: Greg Malysa <[email protected]>
|
|
This adds support for the Analog Devices SC589-EZKIT and SC589-mini.
Signed-off-by: Vasileios Bimpikas <[email protected]>
Signed-off-by: Utsav Agarwal <[email protected]>
Signed-off-by: Arturs Artamonovs <[email protected]>
Signed-off-by: Nathan Barrett-Morrison <[email protected]>
Signed-off-by: Caleb Ethridge <[email protected]>
Signed-off-by: Philip Molloy <[email protected]>
Signed-off-by: Greg Malysa <[email protected]>
|
|
This adds support for the Analog Devices SC573 EZKIT.
Signed-off-by: Vasileios Bimpikas <[email protected]>
Signed-off-by: Utsav Agarwal <[email protected]>
Signed-off-by: Arturs Artamonovs <[email protected]>
Signed-off-by: Nathan Barrett-Morrison <[email protected]>
Signed-off-by: Caleb Ethridge <[email protected]>
Signed-off-by: Philip Molloy <[email protected]>
Signed-off-by: Greg Malysa <[email protected]>
|
|
This adds infrastructure and shared library code for building targets
that use the ADI SOMCRR-EZKIT and SOMCRR-EZLITE carrier boards. These
are not used directly as board targets in their own right.
Co-developed-by: Nathan Barrett-Morrison <[email protected]>
Signed-off-by: Nathan Barrett-Morrison <[email protected]>
Signed-off-by: Vasileios Bimpikas <[email protected]>
Signed-off-by: Utsav Agarwal <[email protected]>
Signed-off-by: Arturs Artamonovs <[email protected]>
Signed-off-by: Greg Malysa <[email protected]>
|
|
The boot ROM on Analog Devices ADSP-SC5xx SoCs requires code packaged
in the LDR format. Normally this is available as part of
our yocto-derived toolchain but, it is not a part of any other pre-made
toolchain anymore, so it is otherwise unavailable in the docker image
for CI. This patch adds a source build from the ADI maintained github
repository. In the future, a package available for install via apt will
be available, but currently there is no arm64 build upstream, so we must
build from source for the time being to support CI on both amd64 and
arm64 runners. The same ldr tool is used for arm and arm64 for all of
our boards with names adjusted to match the expected $(CROSS_COMPILE)
for these boards.
Signed-off-by: Greg Malysa <[email protected]>
|
|
Binman is optionally supported for Analog Devices sc5xx SoCs if Yocto is
not being used to create and assemble system images. The spl LDR is
generated locally but other artifacts such as kernel FIT image and root
file system are built externally and must be supplied to binman if used.
Binman is enabled by selecting the SC5XX_USE_BINMAN config symbol and
the image structure is included in the shared sc5xx device tree.
Signed-off-by: Greg Malysa <[email protected]>
Signed-off-by: Utsav Agarwal <[email protected]>
|
|
Support for the SOM-CRR variants introduces library level changes that
are not modelled in the device tree. As a result they cannot both be
selected at the same time, so this updates the dependency in Kconfig to
prevent them from being enabled together.
Reported-by: Philip Molloy <[email protected]>
Signed-off-by: Greg Malysa <[email protected]>
|
|
The symbol SC_BOOT_MODE was named incorrectly and inconsistently with
its usage. The selected boot mode is set only by hardware and cannot be
adjusted through software (apart from the use of FORCE_BMODE to instruct
the boot rom to ignore the hardware setting when loading uboot proper,
but this cannot change how SPL is loaded).
This symbol actually controlled the BCODE (easily confused with BMODE,
shorthand for boot mode), so this renames it to SC_BCODE and updates the
help text to reflect its actual usage: the BCODE is an SoC- and boot
mode-specific setting that affects how the boot rom configures QSPI or
OSPI in order to read an LDR file from the associated peripheral.
Reported-by: Philip Molloy <[email protected]>
Signed-off-by: Greg Malysa <[email protected]>
|
|
Add Kconfig symbols to parameterize the SPI flash layout used in a
default-ish configuration. This adds more flexibility to the default ADI
environment, enabling customers with boards based on but not identical
to an ezkit to reuse more of the infrastructure. Furthermore it allows
for yocto (the expected default) or binman (to be introduced in this
series) to configure or use the flash layout based on a single
definition of all of the parameters when creating an image.
Signed-off-by: Greg Malysa <[email protected]>
Signed-off-by: Utsav Agarwal <[email protected]>
|
|
https://source.denx.de/u-boot/custodians/u-boot-mmc
CI: https://source.denx.de/u-boot/custodians/u-boot-mmc/-/pipelines/29066
- mmc: Fix sdhci-cadence6 license
- mmc: Fix sd_get_capabilities retry logic
- mmc: use max-frequency from device tree
- Clean up regulator and build fix
|
|
In sd_get_capabilities an ACMD is sent (SD_CMD_APP_SEND_SCR),
which requires sending APP_CMD (MMC_CMD_APP_CMD) before.
Currently, the ACMD is retried on error, however APP_CMD isn't.
In this case, when the ACMD fails and it is tried again,
the retry attempts will not be handled as ACMD, which is wrong.
The fix performs the retry attempts on the sequence of
APP_CMD and the ACMD together.
Signed-off-by: Yanir Levin <[email protected]>
Reviewed-by: Eran Moshe <[email protected]>
Signed-off-by: Peng Fan <[email protected]>
|
|
The license of the file is not valid. Fix it to GPL-2.0+.
Fixes: fe11aa0b8ca3 ("mmc: sdhci-cadence: Add support for Cadence sdmmc v6")
Reported-by: Quentin Schulz <[email protected]>
Reported-by: oliver Fendt <[email protected]>
Closes: https://lore.kernel.org/all/CAFoF8fC4foffYJgYm9CkViET83gDu05noVRxLxgs+KWXN_-LBQ@mail.gmail.com/
Signed-off-by: Hal Feng <[email protected]>
Reviewed-by: Quentin Schulz <[email protected]>
Reviewed-by: Tanmay Kathpalia <[email protected]>
Acked-by: Peng Fan <[email protected]>
Signed-off-by: Peng Fan <[email protected]>
|
|
Use dev_read_u32_default() instead of dev_read_u32() to read the
"max-frequency" property from device tree. This preserves the driver-set
cfg->f_max value when the optional "max-frequency" property is not
present, ensuring the controller's default frequency is used as fallback
rather than being overwritten.
Suggested-by: Marek Vasut <[email protected]>
Signed-off-by: Tanmay Kathpalia <[email protected]>
Signed-off-by: Peng Fan <[email protected]>
|
|
If CONFIG_DM_GPIO is not enabled, compilation fails with the following
errors:
aarch64-none-linux-gnu-ld: drivers/power/regulator/regulator_common.o: in function `regulator_common_of_to_plat':
<...>/u-boot/drivers/power/regulator/regulator_common.c:30: undefined reference to `gpio_request_by_name'
aarch64-none-linux-gnu-ld: drivers/power/regulator/regulator_common.o: in function `regulator_common_get_enable':
<...>/u-boot/drivers/power/regulator/regulator_common.c:57: undefined reference to `dm_gpio_get_value'
aarch64-none-linux-gnu-ld: drivers/power/regulator/regulator_common.o: in function `regulator_common_set_enable':
<...>/u-boot/drivers/power/regulator/regulator_common.c:92: undefined reference to `dm_gpio_set_value'
make: *** [Makefile:2029: u-boot] Error 139
Since the enable gpio is optional we can conditionally skip these calls.
Reviewed-by: Tanmay Kathpalia <[email protected]>
Reviewed-by: Peng Fan <[email protected]>
Signed-off-by: Julien Stephan <[email protected]>
Reviewed-by: Quentin Schulz <[email protected]>
Signed-off-by: Peng Fan <[email protected]>
|
|
Use dm_gpio_is_valid() helper function instead of manually checking the
gpio.
Reviewed-by: Peng Fan <[email protected]>
Reviewed-by: Tanmay Kathpalia <[email protected]>
Reviewed-by: Neil Armstrong <[email protected]>
Signed-off-by: Julien Stephan <[email protected]>
Signed-off-by: Peng Fan <[email protected]>
|
|
Drop the ftrace like debug() that checkpatch --strict complains about:
WARNING: Unnecessary ftrace-like logging - prefer using ftrace
Reviewed-by: Peng Fan <[email protected]>
Reviewed-by: Tanmay Kathpalia <[email protected]>
Reviewed-by: Neil Armstrong <[email protected]>
Signed-off-by: Julien Stephan <[email protected]>
Signed-off-by: Peng Fan <[email protected]>
|
|
Add pull-up/pull-down (PUPD) register definitions for mt8365.
Signed-off-by: David Lechner <[email protected]>
|
|
Replace casting with long to IS_ERR_VALUE() macro to check for error
return values from rate calculation functions. This is the recommended
way to check the return value from clock rate functions.
Signed-off-by: David Lechner <[email protected]>
|
|
Replace uclass_get_device_by_ofnode() with syscon_regmap_lookup_by_phandle()
to get the "mediatek,pctl-regmap" syscon device.
Depending on probe order, uclass_get_device_by_ofnode() may fail, but
syscon_regmap_lookup_by_phandle() has logic in it to handle that case
correctly.
The previous implementation could read more than one syscon if the
"mediatek,pctl-regmap" property had more than one phandle, but the one
board with a devicetree that does that is not supported in U-Boot yet,
so we can save that for later (it may never be needed).
Fixes: 424ceba18bfb ("pinctrl: mediatek: support mediatek,pctl-regmap property")
Signed-off-by: David Lechner <[email protected]>
|
|
Pass the unmapped parent ID when setting parent for mux clocks.
For technical reasons, some Mediatek clock driver have a mapping between
the clock IDs used in the devicetree and ID used in the generic clock
framework.
The mtk_clk_mux_set_parent() function is comparing the passed mapped
parent ID against the unmapped IDs in the chip-specific data structures.
Before this change, we were passing the mapped parent ID. When there is
a mapping, this resulted in buggy behavior (usually just incorrectly
failing to find a match and returning an error). We need to pass the
unmapped ID of the parent clock instead for the matching to work
correctly.
Since the reverse lookup is a bit verbose, a helper function is added to
keep the code clean.
Fixes: b1358915728b ("clk: mediatek: add of_xlate ops")
Signed-off-by: David Lechner <[email protected]>
|
|
Fix typo with correct spelling of "purposes".
Also change "debug" to "debugging" while touching this since that is
the more common phrasing.
Reviewed-by: Quentin Schulz <[email protected]>
Signed-off-by: David Lechner <[email protected]>
|
|
Ignore the error return value from mtk_pinconf_bias_set_pupd_r1_r0().
The PUPD/R1/R0 registers only include a small subset of the pins, so
it is normal for this function to return an error for most pins.
Therefore, this error should not be propagated.
This fixes not all pins in a pinmux group being configured in some
cases because the propagated error caused the configuration loop to
exit early.
The rest of the function is refactored to return early on errors to
improve readability.
Signed-off-by: David Lechner <[email protected]>
|
|
Set the size of the reg_cals arrays to PINCTRL_PIN_REG_MAX to in all
affected mediatek pinctrl drivers. This is needed to avoid potential
out-of-bounds accesses when they is used in mtk_hw_pin_field_get().
All array members need to be initialized since the code loops from 0
to PINCTRL_PIN_REG_MAX - 1. mt7622_reg_cals was already defined this
way, but the others were not.
Signed-off-by: David Lechner <[email protected]>
|
|
David Lechner <[email protected]> says:
I needed to debug some clock issues and found some places where pointer
addresses were being printed when names were available. The addresses
are not very helpful, but the names are. So here a couple of patches to
improve that.
Link: https://lore.kernel.org/r/[email protected]
|
|
Consistently use the device name in debug messages. The clk-uclass file
previously had a mix of printing the dev pointer and the device name.
Changing all to use the device name makes the debug messages more
useful.
Signed-off-by: David Lechner <[email protected]>
|
|
Change the format specifier from %p to %s when printing the ofnode name
so that the actual name is printed instead of the pointer address.
Signed-off-by: David Lechner <[email protected]>
|
|
With commit 5663b137e682 ("arm: Remove edminiv2 board") the last
ARCH_ORION5X platform was removed. Remove the rest of the architecture
code which is now unused.
Reviewed-by: Tony Dinh <[email protected]>
Signed-off-by: Tom Rini <[email protected]>
|
|
PHYTEC maintains an actively monitored mailing list for upstream
activities: [email protected]. Add it to the MAINTAINERS
entries for PHYTEC boards we actively develop and contribute to.
Signed-off-by: Wadim Egorov <[email protected]>
Acked-by: Teresa Remmet <[email protected]>
|
|
This series from Wadim Egorov <[email protected]> performs some
improvements to some of the phycore am6xx series devices.
Link: https://lore.kernel.org/r/[email protected]
|
|
BOOTP does not support dynamic lease expiration. Using random MAC
addresses on R5 network boot binaries would result in continuously
allocated IPs without proper release.
Since only one interface is enabled for network boot, rely on the
MAC address provided by the efuses instead.
Signed-off-by: Daniel Schultz <[email protected]>
Signed-off-by: Wadim Egorov <[email protected]>
|
|
BOOTP does not support dynamic lease expiration. Using random MAC
addresses on R5 network boot binaries would result in continuously
allocated IPs without proper release.
Since only one interface is enabled for network boot, rely on the
MAC address provided by the efuses instead.
Signed-off-by: Daniel Schultz <[email protected]>
Signed-off-by: Wadim Egorov <[email protected]>
|
|
Enable command for verifying DDRSS inline ECC features.
Signed-off-by: Wadim Egorov <[email protected]>
|
|
correctly aligned"
Tom Rini <[email protected]> says:
This series builds on top of what Beleswar Padhi did in [1]. While
there's still discussion about the mkimage related parts, the linker
portion appears to the reliable path forward. An alternative that I had
mentioned before, and was part of previous discussions on this topic[2]
is in the end I believe not reliable enough. While we can take an output
file and pad it to where we think it needs to be, ultimately the linker
needs to place the symbol where we want it and if that isn't where we
pad to, we have a different problem. So what this series does (but each
commit message elaborates on the arch-specific linker scripts being
inconsistent) is make sure the linker script will place the required
symbol at 8-byte alignment, and then also use an ASSERT to fail the
build if this would not be true due to some unforseen event.
[1]: https://lore.kernel.org/u-boot/[email protected]/
[2]: https://source.denx.de/u-boot/u-boot/-/issues/30
Link: https://lore.kernel.org/r/[email protected]
|
|
With commit 0535e46d55d7 ("scripts/dtc: Update to upstream version
v1.7.2-35-g52f07dcca47c") it is now a fatal error to U-Boot if our
device tree is not 8-byte aligned. In commit 85f586035d75 ("ARM: OMAP2+:
Pad SPL binary to 8-byte alignment before DTB") Beleswar Padhi explains
that we must have ALIGN(x) statements inside of a section to ensure that
padding is included and not simply that the linker address counter is
incremented. To that end, this patch:
- Rewrite the '.rel.dyn' (u-boot.lds) to follow modern practices, and
include the 8-byte alignment at the end of the section.
- Expands the '.dynamic' section (u-boot-64.lds) to be more readable
when adding a second statement to the section.
- Aligns the final section before _end (for U-Boot) or _image_binary_end
or __bss_end (for xPL phases) by 8-bytes by adding '. = ALIGN(8);' to
the final section before the symbol or changing an existing ALIGN(4)
statement.
- Ensure that we do have alignment by adding an ASSERT so that when not
aligned we fail to link (and explain why).
Signed-off-by: Tom Rini <[email protected]>
|
|
With commit 0535e46d55d7 ("scripts/dtc: Update to upstream version
v1.7.2-35-g52f07dcca47c") it is now a fatal error to U-Boot if our
device tree is not 8-byte aligned. In commit 85f586035d75 ("ARM: OMAP2+:
Pad SPL binary to 8-byte alignment before DTB") Beleswar Padhi explains
that we must have ALIGN(x) statements inside of a section to ensure that
padding is included and not simply that the linker address counter is
incremented. To that end, this patch:
- Aligns the final section before _image_binary_end (for xPL phases) by
8-bytes by adding '. = ALIGN(8);' to the final section before the
symbol.
- Ensure that we do have alignment by adding an ASSERT so that when not
aligned we fail to link (and explain why).
Signed-off-by: Tom Rini <[email protected]>
|
|
With commit 0535e46d55d7 ("scripts/dtc: Update to upstream version
v1.7.2-35-g52f07dcca47c") it is now a fatal error to U-Boot if our
device tree is not 8-byte aligned. In commit 85f586035d75 ("ARM: OMAP2+:
Pad SPL binary to 8-byte alignment before DTB") Beleswar Padhi explains
that we must have ALIGN(x) statements inside of a section to ensure that
padding is included and not simply that the linker address counter is
incremented. To that end, this patch:
- Aligns the final section before _end (for U-Boot) or _image_binary_end
(for xPL phases) by 8-bytes by adding '. = ALIGN(8);' to the final
section before the symbol.
- Remove a now-spurious '. = ALIGN(x);' statement that was intended to
provide the above alignments.
- Ensure that we do have alignment by adding an ASSERT so that when not
aligned we fail to link (and explain why).
Reviewed-by: Leo Yu-Chi Liang <[email protected]>
Reviewed-by: Ilias Apalodimas <[email protected]>
Signed-off-by: Tom Rini <[email protected]>
|
|
With commit 0535e46d55d7 ("scripts/dtc: Update to upstream version
v1.7.2-35-g52f07dcca47c") it is now a fatal error to U-Boot if our
device tree is not 8-byte aligned. In commit 85f586035d75 ("ARM: OMAP2+:
Pad SPL binary to 8-byte alignment before DTB") Beleswar Padhi explains
that we must have ALIGN(x) statements inside of a section to ensure that
padding is included and not simply that the linker address counter is
incremented. To that end, this patch:
- Aligns the final section before _end by 8-bytes by adding '. =
ALIGN(8);' or changing an existing ALIGN(4) statement.
- Ensure that we do have alignment by adding an ASSERT so that when not
aligned we fail to link (and explain why).
Signed-off-by: Tom Rini <[email protected]>
|
|
With commit 0535e46d55d7 ("scripts/dtc: Update to upstream version
v1.7.2-35-g52f07dcca47c") it is now a fatal error to U-Boot if our
device tree is not 8-byte aligned. In commit 85f586035d75 ("ARM: OMAP2+:
Pad SPL binary to 8-byte alignment before DTB") Beleswar Padhi explains
that we must have ALIGN(x) statements inside of a section to ensure that
padding is included and not simply that the linker address counter is
incremented. To that end, this patch:
- Aligns the final section before _end 8-bytes by adding '. = ALIGN(8);'
to the final section before the symbol.
- Remove a now-spurious '. = ALIGN(x);' statement that was intended to
provide the above alignments.
- Ensure that we do have alignment by adding an ASSERT so that when not
aligned we fail to link (and explain why).
Acked-by: Ilias Apalodimas <[email protected]>
Signed-off-by: Tom Rini <[email protected]>
|
|
With commit 0535e46d55d7 ("scripts/dtc: Update to upstream version
v1.7.2-35-g52f07dcca47c") it is now a fatal error to U-Boot if our
device tree is not 8-byte aligned. In commit 85f586035d75 ("ARM: OMAP2+:
Pad SPL binary to 8-byte alignment before DTB") Beleswar Padhi explains
that we must have ALIGN(x) statements inside of a section to ensure that
padding is included and not simply that the linker address counter is
incremented. To that end, this patch:
- Aligns the final section before _end (for U-Boot) or _image_binary_end
(for xPL phases) by 8-bytes by adding '. = ALIGN(8);' to the final
section before the symbol. For SPL we need this in two places to cover
all build configurations.
- Remove now-spurious '. = ALIGN(x);' statements that were intended to
provide the above alignments.
- Ensure that we do have alignment by adding an ASSERT so that when not
aligned we fail to link (and explain why).
Reviewed-by: Daniel Schwierzeck <[email protected]>
Signed-off-by: Tom Rini <[email protected]>
|
|
With commit 0535e46d55d7 ("scripts/dtc: Update to upstream version
v1.7.2-35-g52f07dcca47c") it is now a fatal error to U-Boot if our
device tree is not 8-byte aligned. In commit 85f586035d75 ("ARM: OMAP2+:
Pad SPL binary to 8-byte alignment before DTB") Beleswar Padhi explains
that we must have ALIGN(x) statements inside of a section to ensure that
padding is included and not simply that the linker address counter is
incremented. To that end, this patch:
- Aligns the final section before _end (for U-Boot) or _image_binary_end
(for xPL phases) by 8-bytes by adjusting the ALIGN(4) statement to be
ALIGN(8) in the final section before the symbol.
- Ensure that we do have alignment by adding an ASSERT so that when not
aligned we fail to link (and explain why).
Tested-by: Michal Simek <[email protected]>
Reviewed-by: Ilias Apalodimas <[email protected]>
Signed-off-by: Tom Rini <[email protected]>
|
|
With commit 0535e46d55d7 ("scripts/dtc: Update to upstream version
v1.7.2-35-g52f07dcca47c") it is now a fatal error to U-Boot if our
device tree is not 8-byte aligned. In commit 85f586035d75 ("ARM: OMAP2+:
Pad SPL binary to 8-byte alignment before DTB") Beleswar Padhi explains
that we must have ALIGN(x) statements inside of a section to ensure that
padding is included and not simply that the linker address counter is
incremented. To that end, this patch:
- Remove part of what Marek Vasut did in commit 9ed99e2eeadb ("m68k:
Assure end of U-Boot is at 8-byte aligned offset") as we now better
understand what can trigger failure and check for it.
- Rewrite the '.dynsym' section to follow modern practices, and include
the 8-byte alignment at the end of the section.
- Ensure that we do have alignment by adding an ASSERT so that when not
aligned we fail to link (and explain why).
Signed-off-by: Tom Rini <[email protected]>
|
