With of-platdata-inst we want to set up a reference to each devices'
parent device, if there is one. If we find that the device has a parent
(i.e. is not a root node) but it is not in the list of devices being
written, then we cannot create the reference.
Report an error in this case, since it indicates that the parent node
is either missing a compatible string, is disabled, or perhaps does not
have any properties because it was not tagged for SPL.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present each invocation of run_steps() updates OUTPUT_FILES_COMMON,
since it does not make a copy of the dict. This is fine for a single
invocation, but for tests, run_steps() is invoked many times.
As a result it may include unwanted items from the previous run, if it
happens that a test runs twice on the same CPU. The problem has not been
noticied previously, as there are few enough tests and enough CPUs that
is is rare for the 'wrong' combination of tests to run together.
Fix this by making a copy of the dict, before updating it. Update the
tests to suit, taking account of the files that are no-longer generated.
With this fix, we no-longer generate files which are not needed for a
particular state of OF_PLATDATA_INST, so the check_instantiate() function
is not needed anymore. It has become dead code and so fails the
code-coverage test (dtoc -T). Remove it.
Signed-off-by: Simon Glass <sjg@chromium.org>
This existing code assumes that a reg property is larger than one cell,
but this is not always the case. Fix this assumption.
Also if a node's parent is missing the #address-cells and #size-cells
properties we use 2 as a default for each. But this should not happen in
practice. More likely the properties were removed for SPL due to there
being no 'u-boot,dm-pre-reloc' property, or similar. Add a warning for
this as the failure can be very confusing.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present an empty size is considered to be a 64-bit value. This does not
seem useful and wastes space. Limit the 64-bit detection to where one or
both of the addr/size is two cells or more.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present warnings are shown as soon as they are discovered in the
source scannner. But the function that detects them may be called multiple
times.
Collect all the warnings and show them at the end.
Signed-off-by: Simon Glass <sjg@chromium.org>
Add a few more internal checks to make sure offsets are correct, before
updating the dtb.
To make this easier, update the functions which add a property to return
that property,.
Signed-off-by: Simon Glass <sjg@chromium.org>
So far we have only needed to add subnodes to empty notds, so have not
had to deal with ordering. However this feature is needed for binman's
expanded nodes, since there may be another node in the same section.
While libfdt adds new properties after existing properties, it adds new
subnodes before existing subnodes. This means that we must reorder the
nodes in the cached version, so that the ordering remains consistent.
Update the sync implementation to sync existing subnodes first, then
add new ones, then tidy up the ordering in the cached version. Update the
test to cover this behaviour.
Also improve the comment about property syncing while we are here.
Signed-off-by: Simon Glass <sjg@chromium.org>
Add a new test that adds a subnode alongside an existing one, as well as
adding properties to a subnode. This will expand to adding multiple
subnodes in future patches. Put a node after the one we are adding to so
we can check that things sync correctly.
The testAddNode() test should be in the TestNode class since it is a node
test, so move it.
Signed-off-by: Simon Glass <sjg@chromium.org>
Once the tree has been synced, thus potentially moving things around in the
fdt, we set _cached_offsets to False so that a refresh will happen next
time a property is accessed.
This 'lazy' refresh doesn't really save much time, since refresh is a very
fast operation, just a single walk of the tree. Also, having the refresh
happen in the bowels of property access it makes it harder to figure out
what is going on.
Simplify the code by always doing a refresh before and after a sync. Set
_cached_offsets to True immediately after this, in the Refresh() function,
since this makes more sense than doing it in the caller.
Signed-off-by: Simon Glass <sjg@chromium.org>
If a property does not yet have an offset, then that means it exists in
the cache'd fdt but has not yet been synced back to the flat tree. Use
the dirty flag for this so we don't need to check the offset too. Improve
the comments for Prop and Node to make it clear what an offset of None
means.
Also clear the dirty flag after the property is synced.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present all possible files are generated, even if some of them just
have a header and an empty body. It is better to generate only the files
that are needed, so that the two types of build (based on the setting of
OF_PLATDATA_INST) can be mutually exclusive.
This is intended to fix a strange problem sometimes found with CI:
Building current source for 1 boards (1 thread, 40 jobs per thread)
sandbox: + sandbox_spl
+drivers/built-in.o: In function `dm_setup_inst':
+drivers/core/root.c:135: undefined reference to
`_u_boot_list_2_udevice_2_root'
+dts/dt-uclass.o:(.u_boot_list_2_uclass_2_serial+0x10): undefined
reference to `_u_boot_list_2_udevice_2_serial'
...
This likely happens when switching from !OF_PLATDATA_INST to
OF_PLATDATA_INST since running 'make xxx_defconfig" does not currently
cause any change in which files are generated. With !OF_PLATDATA_INST
the dt-device.c file has no declarations and this is assumed to be the
starting state. The error above seems to indicate that, after changing
to OF_PLATDATA_INST, the dt-uclass.c file is regenerated but the
dt-device.c files is not. This does not seem possible from the relevant
Makefile.spl rule:
u-boot-spl-platdata := $(obj)/dts/dt-plat.o $(obj)/dts/dt-uclass.o
$(obj)/dts/dt-device.o
cmd_dtoc = $(DTOC_ARGS) -c $(obj)/dts -C include/generated all
include/generated/dt-structs-gen.h $(u-boot-spl-platdata_c) &: \
$(obj)/$(SPL_BIN).dtb
@[ -d $(obj)/dts ] || mkdir -p $(obj)/dts
$(call if_changed,dtoc)
It seems that this cannot regenerate dt-uclass.c without dt-device.c since
'dtoc all' is used. So here the trail ends for now.
In any case it seems better to generate files that are uses and not bother
with those that serve no purpose. So update dtoc to do this automatically.
Signed-off-by: Simon Glass <sjg@chromium.org>
We can use extern instead, so let's drop these macros. It adds one more
thing to learn about and doesn't make the code any clearer.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
Add support for generating a file containing udevice instances. This
avoids the need to create these at run time.
Update a test uclass to include a 'per_device_plat_auto' member, to
increase test coverage.
Add another tab to the driver_info output so it lines up nicely like the
device-instance output.
Signed-off-by: Simon Glass <sjg@chromium.org>
Add support for generating a file containing uclass instances. This avoids
the need to create these at run time.
Update a test uclass to include a 'priv_auto' member, to increase test
coverage.
Signed-off-by: Simon Glass <sjg@chromium.org>
Add a summary to the top of the generated code, to make it easier to see
what the file contains.
Also add a tab to .plat so that its value lines up with the others.
Signed-off-by: Simon Glass <sjg@chromium.org>
For now dtoc only supports a hard-coded list of phandle properties, to
avoid any situation where it makes a mistake in its determination.
Make this into a constant dict, recording both the phandle property name
and the associated #cells property in the target node. This makes it
easier to find and modify.
Signed-off-by: Simon Glass <sjg@chromium.org>
This file is not used when instantiating devices. Update dtoc to skip
generating its contents and just add a comment instead.
Also it is useful to see the driver name and parent for each device.
Update the file to show that information, to avoid updating the same
tests twice.
Signed-off-by: Simon Glass <sjg@chromium.org>
Add an option to generate the declaration file, which declares all
drivers and uclasses, so references can be used in the code.
Signed-off-by: Simon Glass <sjg@chromium.org>
Add an option to instantiate devices at build time. For now this just
parses the option and sets up a few parameters.
Signed-off-by: Simon Glass <sjg@chromium.org>
The device for the root node is normally bound by driver model on init.
With devices being instantiated at build time, we must handle the root
device also.
Add support for processing the root node, which may not have a compatible
string.
Signed-off-by: Simon Glass <sjg@chromium.org>
We only care about uclasses that are actually used. This is determined by
the drivers that use them. Check all the used drivers and build a list of
'valid' uclasses.
Also add references to the uclasses so we can generate C code that uses
them. Attach a uclass to each valid driver.
For the tests, now that we have uclasses we must create an explicit test
for the case where a node does not have one. This should only happen if
the source code does not build, or the source-code scanning fails to find
it.
Signed-off-by: Simon Glass <sjg@chromium.org>
Now that we have the alias information we can assign a sequence number
to each device in the uclass. Store this in the node associated with each
device.
This requires renaming the sandbox test drivers to have the right name.
Note that test coverage is broken with this patch, but fixed in the next
one.
Signed-off-by: Simon Glass <sjg@chromium.org>
If a driver declaration is included in a comment, dtoc currently gets
confused. Update the parser to only consider declarations that begin at
the start of a line. Since multi-line comments begin with an asterisk,
this avoids the problem.
Signed-off-by: Simon Glass <sjg@chromium.org>
Scan the aliases in the device tree to establish the number of devices
within each uclass, and the sequence number of each.
Signed-off-by: Simon Glass <sjg@chromium.org>
If drivers have the same name then we cannot distinguish them. This only
matters if the driver is actually used by dtoc, but in that case, issue
a warning.
Signed-off-by: Simon Glass <sjg@chromium.org>
Instead of using a separate step for this processing, handle it while
scanning its associated driver. This allows us to drop the code coverage
exception in this case.
Note that only files containing drivers are scanned by dtoc, so aliases
declared in a file that doesn't hold a driver will not be noticed. It
would be confusing to put them anywhere other than in the driver that they
relate to, but update the documentation to say this explicitly, just in
case.
Signed-off-by: Simon Glass <sjg@chromium.org>
Typically dtoc can detect the header file needed for a driver by looking
for the structs that it uses. For example, if a driver as a .priv_auto
that uses 'struct serial_priv', then dtoc can search header files for the
definition of that struct and use the file.
In some cases, enums are used in drivers, typically with the .data field
of struct udevice_id. Since dtoc does not support searching for these,
add a way to tell dtoc which header to use. This works as a macro included
in the driver definition.
Signed-off-by: Simon Glass <sjg@chromium.org>
U-Boot operates in several phases, typically TPL, SPL and U-Boot proper.
The latter does not use dtoc.
In some rare cases different drivers are used for two phases. For example,
in TPL it may not be necessary to use the full PCI subsystem, so a simple
driver can be used instead.
This works in the build system simply by compiling in one driver or the
other (e.g. PCI driver + uclass for SPL; simple_bus for TPL). But dtoc has
no way of knowing which code is compiled in for which phase, since it does
not inspect Makefiles or dependency graphs.
So to make this work for dtoc, we need to be able to explicitly mark
drivers with their phase. This is done by adding an empty macro to the
driver. Add support for this in dtoc.
Signed-off-by: Simon Glass <sjg@chromium.org>
Add logic to assign property values to nodes as required by dtoc. The
references allow nodes to refer to each other in C code. The macros used
by dtoc are not yet defined in driver model. They will be added along
with the actual driver model implementation.
Signed-off-by: Simon Glass <sjg@chromium.org>
It is convenient to attach drivers, etc. to nodes so that we can use the
Node object as the main data structure in this module.
Add a function which adds the new properties, along with documentation.
Signed-off-by: Simon Glass <sjg@chromium.org>
These have '_test' suffixes which are not present on the drivers in the
source code. Drop the suffixes to avoid a mismatch when scanning.
Signed-off-by: Simon Glass <sjg@chromium.org>
It is confusing to have the test files in the same places as the
implementation. Move them into a separate directory.
Add a helper function for test_dtoc, to avoid repeating the same
path.
Signed-off-by: Simon Glass <sjg@chromium.org>
Drivers can have private / platform data contained in structs and these
struct definitions are generally kept in header files. In order to
generate build-time devices, dtoc needs to generate code that declares
the data contained in those structs. This generated code must include the
relevant header file, to avoid a build error.
We need a way for dtoc to scan header files for struct definitions. Then,
when it wants to generate code that uses a struct, it can make sure it
includes the correct header file, first.
Add a parser for struct information, similar to drivers. Keep a dict of
the structs that were found.
Signed-off-by: Simon Glass <sjg@chromium.org>
Uclasses can have per-device private / platform data so dtoc needs to
scan these drivers. This allows it to find out the size of this data so
it can be allocated a build time.
Add a parser for uclass information, similar to drivers. Keep a dict of
the uclasses that were found.
Signed-off-by: Simon Glass <sjg@chromium.org>
In order to output variables to hold the priv/plat information used by
each device, dtoc needs to know the struct for each. With this, it can
declare this at build time:
u8 xxx_priv [sizeof(struct <name>)];
Collect the various struct names from the drivers.
Signed-off-by: Simon Glass <sjg@chromium.org>
We should ignore anything in the .git directory or any of the
build-sandbox, etc. directories created by 'make check'. These can confuse
dtoc. Update the code to ignore these.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present most of the tests scan the U-Boot source tree as part of their
run. This information does not change across tests, so we can save time
by remembering it.
Add a way to set up this information and use it for each test, taking a
copy first, so as not to mess up the original.
This reduces the run time from about 1.6 seconds to 1.5 seconds on my
machine. For code coverage (which cannot run in parallel), it reduces from
33 seconds to 5.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present we simply record the name of a driver parsed from its
implementation file. We also need to get the uclass and a few other
things so we can instantiate devices at build time. Add support for
collecting this information. This requires parsing each driver file.
Signed-off-by: Simon Glass <sjg@chromium.org>
When packing files it is sometimes useful to align the start of each file,
e.g. if the flash driver can only access 32-bit-aligned data. Provides a
new property to support this.
Signed-off-by: Simon Glass <sjg@chromium.org>
Some of these tests don't actually check anything. Add a few more checks
to complete the tests.
Also add a simple scan test that does the basics.
Signed-off-by: Simon Glass <sjg@chromium.org>
Move the tests related to scanning into their own class, updating them
to avoid using dtb_platdata as a pass-through.
Signed-off-by: Simon Glass <sjg@chromium.org>
Before expanding the scanning features any more, move this into a separate
file. This will make it easier to maintain in the future. In particular,
it reduces the size of dtb_platdata.py and allows us to add tests
specifically for scanning, without going through that file.
The pieces moved are the Driver class, the scanning code and the various
naming functions, since they mostly depend on the scanning results.
So far there is are no separate tests for src_scan. These will be added
as new functionality appears.
This introduces no functional change.
Signed-off-by: Simon Glass <sjg@chromium.org>
This has not been needed since parent information was added and we started
using indicies for references to other drivers instead of pointers. It was
kept around in the expectation that it might be needed later.
However with the latest updates, it doesn't seem likely that we'll need
this in the foreseeable future.
Drop dm_populate_phandle_data() from dtoc and driver model.
Signed-off-by: Simon Glass <sjg@chromium.org>
Previously we had to worry about nodes being output before those that they
depended on, thus causing build errors. So the current algorithm is
careful to output nodes in the right order.
We now use a different method for outputting phandles that does not
involve pointers. Also we plan to add a 'declarations' header file to
declare all drivers as 'extern'.
Update the code to drop the dependency checking and output in a simple
loop. This makes the output easier to follow since drivers are in order of
thier indices (0, 1, ...), which is also the order it appears in in the
linker list.
Signed-off-by: Simon Glass <sjg@chromium.org>
The base directory of U-Boot, where the source is, it currently calculated
from the directory of the dtb_platdata.py script. If this is installed
elsewhere that will not work. Also it is inconvenient for tests.
Add a parameter to allow specifying this base directory.
To test this, pass a temporary directory with some files in it and check
that they are passed to scan_driver().
Signed-off-by: Simon Glass <sjg@chromium.org>
Rather than the if/else construct, update OutputFile with the method to
call to process each command. This is easier to maintain as the number of
commands increases.
Rename generate_tables to generate_plat since it better describes what is
being generated ('plat' is the U-Boot name for platform data).
With this, each output method needs to have the same signature. Store the
output structures in a member variable instead of using parameters, to
accomplish this.
Signed-off-by: Simon Glass <sjg@chromium.org>
