| Age | Commit message (Collapse) | Author |
|---|
|
If device_set_name is called on a device that already has
DM_FLAG_NAME_ALLOCED set, the old dynamically-allocated name is leaked.
Free it before assigning the new name.
See: drivers/net/mdio_gpio.c:mdio_gpio_bind(). There is
device_set_name() here, however dm_mdio_post_bind() will also call
device_set_name() if "device-name" exists.
Signed-off-by: Peng Fan <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
|
|
Every other calloc in drivers/core uses calloc(count, size) with count
first. Fix the two reversed calloc(sizeof, 1) calls for consistency.
Signed-off-by: Peng Fan <[email protected]>
Reviewed-by: Raphaël Gallais-Pou <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
|
|
The duplicate-name check uses strlen on the search name instead of the
child name, so a child named "trevor" would falsely match a search
for "trev". Fix by checking strlen of child->name.
Signed-off-by: Peng Fan <[email protected]>
Reviewed-by: Raphaël Gallais-Pou <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
|
|
If fdt_create_empty_tree fails, the malloc-ed fdt buffer is never freed.
Add free before the error return.
Signed-off-by: Peng Fan <[email protected]>
Reviewed-by: Raphaël Gallais-Pou <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
|
|
Set the acpi_ops structure as static const where applicable. The
The structure is not accessible from outside of drivers and is not
going to be modified at runtime. The structure may be unused in a
couple of drivers depending on their configuration, mark those
sites with __maybe_unused .
Signed-off-by: Marek Vasut <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
|
|
CLK is an optional dependency of simple-pm-bus. Remove the dependency.
Fixes: 447bd8f1e5cf ("simple-pm-bus: Make clocks optional")
Signed-off-by: Markus Schneider-Pargmann (TI) <[email protected]>
|
|
Similar to SIMPLE_BUS, create a SPL_SIMPLE_PM_BUS additional to the
SIMPLE_PM_BUS. Most boards will not need SIMPLE_PM_BUS in SPL.
This is currently needed to reduce the SPL size for beagle bone black
with OF_UPSTREAM enabled.
Reviewed-by: Simon Glass <[email protected]>
Reviewed-by: Kory Maincent <[email protected]>
Signed-off-by: Markus Schneider-Pargmann (TI) <[email protected]>
|
|
Update dev_read_addr_*_ptr() and dev_remap_addr_*() helpers to use
dev_read_addr_size_*() and pass the DT-provided region size to
map_sysmem() / map_physmem() instead of 0.
Ensure mappings are consistent with the size defined in the device tree
and avoids implicit or unbounded mappings. When the DT does not provide
a size, the behavior remains unchanged since size is initialized to 0.
Signed-off-by: Peng Fan <[email protected]>
|
|
Rasmus Villemoes <[email protected]> says:
There are quite a few places where we allocate X+1 bytes, initialize
the first X bytes via memcpy() and then set the last byte to 0.
The kernel has a helper for that, kmemdup_nul(). Introduce a similar
one, and start making use of it in a few places. Also the existing
memdup() helper can be put to more use.
There are lots more places one could modify. But for code shared with
host tools, one would need to do some refactoring, putting memdup()
and memdup_nul() in their own str-util.c TU which could then also be
included in the tools build.
Link: https://lore.kernel.org/r/[email protected]
|
|
Use memdup() instead of open-coding it.
In the dm_setup_inst() case, there was never any reason to use
calloc(), as the whole allocation is definitely initialized via the
immediately following memcpy().
Reviewed-by: Simon Glass <[email protected]>
Signed-off-by: Rasmus Villemoes <[email protected]>
|
|
Casey Connolly <[email protected]> says:
This series implements various improvements to Linux header
compatibility, largely in preparation for a full port of Linux CCF but
many of these changes would also be helpful when porting other drivers.
Beside the basic header/compat stuff there are a few larger patches:
Patch 1 adds the "%pOF" format specifier to vsprintf, this behaves the
same as it does in Linux printing the name of the ofnode, but notably it
expects an ofnode pointer rather than a device_node.
Patch 2 adds an option to skip doing a full DM scan pre-relocation.
Some platforms like Qualcomm don't actually need devices to be probed
prior to relocation, it is also quite slow to scan the entire FDT before
caches are up. This option gets us to main loop 30-50% faster.
Unfortunately it isn't possible to totally skip DM since U-Boot will
panic if it can't find a serial port, but the serial uclass code will
bind the serial port itself by reading /chosen/stdout-path, however any
dependencies like clocks won't be found so this should only be enabled
if the serial driver gracefully handles missing clocks.
Patch 3 adds [k]strdup_const(), this works the same as the Linux version
saving a small amount of memory by avoiding duplicating strings stored
in .rodata, this is particularly useful for CCF.
Patch 4 adds 64-bit versions of some 32-bit ofnode utilities functions,
making it possible to parse 64-bit arrays.
Patch 6 provides a simple implementation of kref, this will be used
by CCF.
Patch 9 adds devm_krealloc() support to devres, it relies on storing
allocation sizes in the devres struct which will add a small overhead.
Link: https://lore.kernel.org/r/[email protected]
|
|
These are similar to their Linux counterparts, adding helpers
for reading arrays of 64-bit values with of_access and fdtdec
implementations.
Signed-off-by: Casey Connolly <[email protected]>
|
|
Rework dev_phys_to_bus() and dev_bus_to_phys() to respect the size
of the area specified in dma-ranges DT property. The area outside
of ranges is remapped 1:1, while the area in the ranges is remapped
according to the description in the dma-ranges property.
Adjust the test to test the area within the remapped range, not area
outside the remapped range, which was incorrect.
Signed-off-by: Marek Vasut <[email protected]>
Reviewed-by: Markus Schneider-Pargmann <[email protected]>
|
|
As exposed by "make randconfig", we have an issue with the dependencies
for REGMAP (and xPL variants). As this is a library function, it should
always be selected and not depended on by other functionality. This is
largely done correctly today, so just correct the few outliers.
Acked-by: Anshul Dalal <[email protected]>
Signed-off-by: Tom Rini <[email protected]>
|
|
Simon Glass <[email protected]> says:
Each arch does something slightly different before booting the OS. Some
archs even do different things depending on the CPU type.
It is quite hard to know what actually happens in the final milliseconds
before the OS boot.
This series attempts to start cleaning up U-Boot in this area.
The basic intent is to create a new bootm_final() function which can be
called by all archs. It provides some flags for a couple of necessary
variations but otherwise it is generic.
All architectures are converted over to use this new function.
board_quiesce_devices() is moved into bootm_final() so that all archs
benefit from it.
This series fixes a bug in device_remove() is fixed where removing a
parent with specialised flags (e.g. DM_REMOVE_ACTIVE_ALL) could leave
children activated, since they do not match the flags. This fixes is
needed to avoid bootm_final() causing test failures on sandbox.
Future work could take this a little further:
- Convert EFI loader to use the same function
- Improve comments for cleanup_before_linux() across architectures
- Support fake-run tracing on all archs
Link: https://lore.kernel.org/r/[email protected]
|
|
When dm_remove_devices_active() removes devices using specialised flags
like DM_REMOVE_ACTIVE_ALL, a parent device may match (e.g. MMC has
DM_FLAG_OS_PREPARE) while its children do not. This deactivates the
parent but leaves children activated, an inconsistent state.
Later, when uclass_destroy() calls device_remove() with DM_REMOVE_NORMAL
on the already-deactivated parent, it returns early without touching the
children. The subsequent device_unbind() then fails because the children
are still activated.
Fix this by dropping only the DM_REMOVE_ACTIVE_ALL requirement for child
removal when the parent is being removed. This ensures children are
removed along with their parent, while still preserving other flags like
DM_REMOVE_NON_VITAL so that vital devices remain protected.
Signed-off-by: Simon Glass <[email protected]>
|
|
Move the flags_remove() call before device_chld_remove() and save the
result in a separate variable. This is just a refactoring with no
behaviour change, preparing for the next commit which needs to know
whether the parent will be removed before deciding how to remove its
children.
Signed-off-by: Simon Glass <[email protected]>
|
|
The ofnode_to_fdt() function may return a NULL pointer in multiple cases.
Or, this function's return value is often passed directly to functions such
as fdt_getprop() which end up dereferencing it, thus causing a NULL pointer
exception.
Don't allow ofnode_to_fdt() to return NULL, to avoid a NULL pointer
dereference.
Reviewed-by: Raphaël Gallais-Pou <[email protected]>
Signed-off-by: Romain Gantois <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
|
|
The loop in lists_bind_fdt uses an indented style for log messages
within the loop and normal messages for errors that lead to the exit of
the function. Due to the change of the previous patch that adds support
for continuation on -ENODEV returned by bind, the log message should be
indented.
Signed-off-by: Markus Schneider-Pargmann (TI) <[email protected]>
|
|
Currently once a driver matched the compatible string of a device, other
drivers are ignored. If the first matching driver returns -ENODEV, no
other possibly matching drivers are iterated with that compatible of the
device. Instead the next compatible in the list of compatibles is
selected, assuming only one driver matches one compatible at a time.
To be able to use the bind function to return -ENODEV and continue
matching other drivers with the same compatible, move the for loop a bit
to continue the for loop after -ENODEV was returned. The loop had to be
adjusted a bit to still support the 'drv' argument properly. Some
simplifications were done as well.
The modification will only add additional loop iterations if -ENODEV is
returned. Otherwise the exit and continue conditions for the loop stay
the same and do not cause any additional iterations and should not
impact performance.
This is required for ti-musb-host and ti-musb-peripheral which both
match on the same device but differ based on the dr_mode DT property.
Depending on this property, the driver is either UCLASS_USB or
UCLASS_USB_GADGET_GENERIC. By checking the DT property in the bind
function and returning -ENODEV the other driver can probe instead.
Reviewed-by: Simon Glass <[email protected]>
Acked-by: Dinesh Maniyam <[email protected]>
Signed-off-by: Markus Schneider-Pargmann (TI.com) <[email protected]>
|
|
'found' is only used at the end of the function to print a debug
message. No need to maintain a variable if we can just return 0
immediately when a driver was bound successfully.
Reviewed-by: Simon Glass <[email protected]>
Reviewed-by: Mattijs Korpershoek <[email protected]>
Signed-off-by: Markus Schneider-Pargmann (TI.com) <[email protected]>
|
|
'result' is unused in this function, remove it.
Reviewed-by: Simon Glass <[email protected]>
Reviewed-by: Mattijs Korpershoek <[email protected]>
Signed-off-by: Markus Schneider-Pargmann (TI.com) <[email protected]>
|
|
The devm alloc functions that we have may follow the Linux kernel model
where allocations are (almost always) automatically free()'d. However,
quite often we don't enable, in full U-Boot, the tracking and free()'ing
functionality. This in turn leads to memory leaks because the driver
author expects that since the functions have the same name as in the
Linux Kernel they have the same behavior. In turn we then get
functionally correct commits such as commit 00e1fed93c8c ("firmware:
ti_sci: Fix memory leaks in devm_ti_sci_get_of_resource") that manually
add these calls. Rather than manually tracking allocations and
implementing free()s, rework things so that we follow expectations by
enabling the DEVRES functionality (outside of xPL phases).
This turns DEVRES from a prompted symbol to a symbol that must be
select'd, and we now remove our non-managed alloc/free functions from
outside of xPL builds.
Reviewed-by: Michael Trimarchi <[email protected]>
Signed-off-by: Tom Rini <[email protected]>
|
|
simple-pm-bus binding requires either power-domains or clocks, not both.
Allow clk_get_bulk() to return -ENOENT.
When no clocks are present, bulk->count is set to 0, which works
correctly with clk_enable_bulk() and other clk functions used in this
driver.
Reviewed-by: Kory Maincent <[email protected]>
Signed-off-by: Markus Schneider-Pargmann (TI.com) <[email protected]>
|
|
The "Display Timings" in panel-common.yaml can be provided by 2 properties
- panel-timing: when display panels are restricted to a single resolution
the "panel-timing" node expresses the required timings.
- display-timings: several resolutions with different timings are supported
with several timing subnode of "display-timings" node
This patch update the parsing function to handle this 2 possibility
when index = 0.
Reviewed-by: Patrice Chotard <[email protected]>
Reviewed-by: Yannick Fertre <[email protected]>
Signed-off-by: Raphael Gallais-Pou <[email protected]>
|
|
Support for using "u-boot,dm-..." rather than "bootph-..." has been
deprecated since February 2023. Any platforms using this have had a
console message saying to migrate by 2023.07. Go and remove all support
here now, for the v2026.01 release.
The results of this change that aren't clear from the above are that we
still have a checkpatch.pl error message, and document in
doc/develop/spl.rst that they have been migrated since 2023. We also
change the key2dtsi.py tool to use the correct bootph phase rather than
the legacy phase.
Signed-off-by: Tom Rini <[email protected]>
|
|
root_info driver structure is not changing that's why mark them as const
which ensure that structure will be moved from .data section to .rodata
section.
Signed-off-by: Michal Simek <[email protected]>
Link: https://lore.kernel.org/r/18d92a93a9863ed0452e82a1f8e0ff9205adb4f9.1753444878.git.michal.simek@amd.com
|
|
uclass_find_next_device always returns 0, so instead make it a void and
update calling sites.
Signed-off-by: Andrew Goodbody <[email protected]>
|
|
Move the NULL check of bootscr_flash_size to before the first time it is
dereferenced to avoid any possible segment violations.
This issue found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
Reviewed-by: Tom Rini <[email protected]>
|
|
dev_read_s32_default is for getting an s32 from a Device Tree property
and allows to take a default value if that property is missing.
Considering it calls ofnode_read_u32_default which takes a u32 and
returns a u32, it should do the same instead of using an int, especially
considering that int size is typically architecture-specific, as opposed
to s32/u32.
s32 and u32 being the same size, dev_read_s32* functions calling
ofnode_read_u32_default shouldn't be an issue (at the type level at
least) as the information will be stored appropriately in 4B regardless
of the sign.
This incidentally matches all other dev_read_*_default functions.
Fixes: a1b17e4f4c82 ("dm: core: Add a function to read into a unsigned int")
Signed-off-by: Quentin Schulz <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
|
|
dev_read_u32_default is for getting a u32 from a Device Tree property
and allows to take a default value if that property is missing.
Considering it calls ofnode_read_u32_default which takes a u32 and
returns a u32, it should do the same instead of using an int, especially
considering that int size is typically architecture-specific, as opposed
to u32.
This incidentally matches all other dev_read_*_default functions (except
dev_read_s32_default which will be tackled in the next commit).
Fixes: 47a0fd3bad38 ("dm: core: Implement live tree 'read' functions")
Signed-off-by: Quentin Schulz <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
|
|
Tom Rini <[email protected]> says:
This series switches to always using $(PHASE_) in Makefiles when
building rather than $(PHASE_) or $(XPL_). It also starts on documenting
this part of the build, but as a follow-up we need to rename
doc/develop/spl.rst and expand on explaining things a bit.
Link: https://lore.kernel.org/r/[email protected]
|
|
It is confusing to have both "$(PHASE_)" and "$(XPL_)" be used in our
Makefiles as part of the macros to determine when to do something in our
Makefiles based on what phase of the build we are in. For consistency,
bring this down to a single macro and use "$(PHASE_)" only.
Signed-off-by: Tom Rini <[email protected]>
|
|
There are already several helpers to find a udevice based on its
position in a device tree, like getting a child or a node pointed by a
phandle, but there was no support for graph endpoints, which are very
common in display pipelines.
Add a new helper, named uclass_get_device_by_endpoint() which enters the
child graph reprensentation, looks for a specific port, then follows the
remote endpoint, and finally retrieves the first parent of the given
uclass_id.
This is a very handy and straightforward way to get a bridge or a panel
handle.
Reviewed-by: Simon Glass <[email protected]>
Signed-off-by: Miquel Raynal <[email protected]>
|
|
Naming between the parameter list, the prototype and the main comment do
not match. Fix the comment which seems the be the one that is incorrect.
Fixes: 9057077cf4e1 ("core: ofnode: add of_graph parsing helpers")
Signed-off-by: Miquel Raynal <[email protected]>
Reviewed-by: Svyatoslav Ryhel <[email protected]>
|
|
Add a mostly complete list of ofnode analogs of of_graph
parsing helpers.
Signed-off-by: Svyatoslav Ryhel <[email protected]>
|
|
Unfortunately this change was not safe as some devices are bound before
relocation, but we don't want to probe them.
It causes 'raise: Signal # 8 caught' on jerry.
Move the bootstage timer to after autoprobe in initf_dm() since the
trace test does not tolerate any variance.
This reverts commit 21dd873572a01d74bfdfceb7a30b056f8ccba187.
Signed-off-by: Simon Glass <[email protected]>
|
|
Simon Glass <[email protected]> says:
The current UPL spec[1] has been tidied up and improved over the last
year, since U-Boot's original UPL support was written.
This series includes some prerequisite patches needed for the real UPL
patches. It is split from [2]
[1] https://github.com/UniversalPayload/spec/tree/3f1450d
[2] https://patchwork.ozlabs.org/project/uboot/list/?series=438574&state=*
Link: https://lore.kernel.org/r/[email protected]
|
|
The ofnode_find_subnode() function currently processes things two
different ways, so the treatment of unit addresses differs depending on
whether OF_LIVE is enabled or not.
Add a new version which uses the ofnode API and add a test to check that
unit addresses can be matched correctly. Leave the old function in place
for the !OF_LIVE case, to avoid a code-size increase, e.g. on
firefly-rk3288
Signed-off-by: Simon Glass <[email protected]>
|
|
When a unit-address is provided, use it to match against the node
name.
Since this increases code size, put it into a separate function.
Signed-off-by: Simon Glass <[email protected]>
|
|
Update ofnode_add_subnode() and ofnode_add_prop() to return a suitable
error when space is exhausted in the FDT. This makes it easier to see
what is going wrong.
Signed-off-by: Simon Glass <[email protected]>
Reviewed-by: Tom Rini <[email protected]>
|
|
At some point it would be nice to have the ofnode API automatically
expand the tree as required, to accommodate new nodes. For now, expand
the default size so that UPL can be supported.
Signed-off-by: Simon Glass <[email protected]>
|
|
Rather than doing autoprobe within the driver model code, move it out to
the board-init code. This makes it clear that it is a separate step from
binding devices.
For now this is always done twice, before and after relocation, but we
should discuss whether it might be possible to drop the post-relocation
probe.
For boards with SPL, the autoprobe is still done there as well.
Note that with this change, autoprobe happens after the
EVT_DM_POST_INIT_R/F events are sent, rather than before.
Link: https://lore.kernel.org/u-boot/[email protected]/
Signed-off-by: Simon Glass <[email protected]>
|
|
There is no point in checking the pre_reloc flag, since devices not
marked as pre-reloc will not have been bound, so won't exist yet.
There doesn't seem to be any point in checking if the device has a
valid devicetree node either, so drop that too.
Signed-off-by: Simon Glass <[email protected]>
|
|
implementation"
Christian Marangi <[email protected]> says:
This series is split in 2 part.
While adapting the LED boot and activity code to the new property
accepted by Rob in dt-schema repository, a big BUG was discovered.
The reason wasn't clear at start and took me some days to figure it
out.
This was triggered by adding a new phandle in the test.dts to
introduce test for the new OPs.
This single addition caused the sandbox CI test to fail in the
dm_test_ofnode_phandle_ot test.
This doesn't make sense as reverting the change made the CI test
to correctly finish. Also moving the uboot node down
after the first phandle (in test.dts the gpio one) also made
the CI test to correctly finish.
A little bit of searching and debugging made me realize the
parse phandle OPs didn't support other.dts at all and they
were still referencing phandle index from test.dts.
(more info in the related commit)
In short the test was broken all along and was working by
pure luck. The first 4 patch address and fix the problem for good.
The other 4 patch expand and address the property change for
LED boot/activity.
Posting in a single series as changes are trivial and just
to speedup review process. (and also because the second
part depends on the first)
All CI tested with azure pipeline.
Link: https://lore.kernel.org/r/[email protected]
|
|
Implement ofnode_options phandle helper to get an ofnode from a phandle
option in /options/u-boot.
This helper can be useful since new DT yaml usually require to link a
phandle of a node instead of referencing it by name or other indirect
way.
Signed-off-by: Christian Marangi <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
|
|
Implement ofnode/tree_parse_phandle() helper as an equivalent of
of_parse_phandle to handle simple single value phandle.
Signed-off-by: Christian Marangi <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
|
|
Implement oftree variant of parse_phandle OPs.
There is currently a very hidden and laten BUG with parse_phandle OPs
that doesn't permit the support of multiple DTS in a system. One usage
example if sandbox with the usage of other.dts
The BUG is only present on live scenario where of_... OPs are used and
it's not present when fdt... OPs are used.
This is caused by an assumption made in __of_parse_phandle_with_args,
with the of_find_node_by_phandle call that pass the first arg as NULL.
This makes of_find_node_by_phandle use the default root node of the
system and doesn't permit the usage of alternative tree. This is correct
for normal system and also for the linux kernel where it's assumed a
single device tree.
It's problematic if other device tree needs to be used.
To fix this, introduce __of_root_parse_phandle_with_args to define a
root device tree for of_find_node_by_phandle.
Introduce all the variant OPs for this and in ofnode, the oftree OPs
following how it's done for other OPs with similar task.
For FDT scenario, ofnode_from_fdtdec_phandle_args is reworked to accept
a new variable, node and noffset_to_ofnode is used instead of
offset_to_ofnode. This is required to support multiple FDB blob to
calculate the correct of_offset of the ofnode.
Signed-off-by: Christian Marangi <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
|
|
This replaces dm_remove_devices_flags() calls in all boot
implementations to ensure non vital devices are consistently removed
first. All boot implementation except arch/arm/lib/bootm.c currently
just call dm_remove_devices_flags(DM_REMOVE_ACTIVE_ALL). This can result
in crashes when dependencies between devices exists. The driver model's
design document describes DM_FLAG_VITAL as "indicates that the device is
'vital' to the operation of other devices". Device removal at boot
should follow this.
Instead of adding dm_remove_devices_flags() with (DM_REMOVE_ACTIVE_ALL |
DM_REMOVE_NON_VITAL) everywhere add dm_remove_devices_active() which
does this.
Fixes a NULL pointer deref in the apple dart IOMMU driver during EFI
boot. The xhci-pci (driver which depends on the IOMMU to work) removes
its mapping on removal. This explodes when the IOMMU device was removed
first.
dm_remove_devices_flags() is kept since it is used for testing of
device_remove() calls in dm.
Signed-off-by: Janne Grunau <[email protected]>
|
|
All the uclass functions for finding a device end up creating a uclass
if it doesn't exist. Add a function which instead returns NULL in this
case.
This is useful when in the 'unbind' path, since we don't want to undo
any unbinding which has already happened.
Signed-off-by: Simon Glass <[email protected]>
|
