u-boot/tools/binman/entry.py

678 lines
24 KiB
Python
Raw Normal View History

# SPDX-License-Identifier: GPL-2.0+
# Copyright (c) 2016 Google, Inc
#
# Base class for all entries
#
from __future__ import print_function
from collections import namedtuple
# importlib was introduced in Python 2.7 but there was a report of it not
# working in 2.7.12, so we work around this:
# http://lists.denx.de/pipermail/u-boot/2016-October/269729.html
try:
import importlib
have_importlib = True
except:
have_importlib = False
import os
import sys
import fdt_util
import state
import tools
import tout
modules = {}
our_path = os.path.dirname(os.path.realpath(__file__))
# An argument which can be passed to entries on the command line, in lieu of
# device-tree properties.
EntryArg = namedtuple('EntryArg', ['name', 'datatype'])
# Information about an entry for use when displaying summaries
EntryInfo = namedtuple('EntryInfo', ['indent', 'name', 'etype', 'size',
'image_pos', 'uncomp_size', 'offset',
'entry'])
class Entry(object):
"""An Entry in the section
An entry corresponds to a single node in the device-tree description
of the section. Each entry ends up being a part of the final section.
Entries can be placed either right next to each other, or with padding
between them. The type of the entry determines the data that is in it.
This class is not used by itself. All entry objects are subclasses of
Entry.
Attributes:
section: Section object containing this entry
node: The node that created this entry
offset: Offset of entry within the section, None if not known yet (in
which case it will be calculated by Pack())
size: Entry size in bytes, None if not known
uncomp_size: Size of uncompressed data in bytes, if the entry is
compressed, else None
contents_size: Size of contents in bytes, 0 by default
align: Entry start offset alignment, or None
align_size: Entry size alignment, or None
align_end: Entry end offset alignment, or None
pad_before: Number of pad bytes before the contents, 0 if none
pad_after: Number of pad bytes after the contents, 0 if none
data: Contents of entry (string of bytes)
compress: Compression algoithm used (e.g. 'lz4'), 'none' if none
orig_offset: Original offset value read from node
orig_size: Original size value read from node
"""
def __init__(self, section, etype, node, read_node=True, name_prefix=''):
self.section = section
self.etype = etype
self._node = node
self.name = node and (name_prefix + node.name) or 'none'
self.offset = None
self.size = None
self.uncomp_size = None
self.data = None
self.contents_size = 0
self.align = None
self.align_size = None
self.align_end = None
self.pad_before = 0
self.pad_after = 0
self.offset_unset = False
self.image_pos = None
self._expand_size = False
self.compress = 'none'
if read_node:
self.ReadNode()
@staticmethod
def Lookup(node_path, etype):
"""Look up the entry class for a node.
Args:
node_node: Path name of Node object containing information about
the entry to create (used for errors)
etype: Entry type to use
Returns:
The entry class object if found, else None
"""
# Convert something like 'u-boot@0' to 'u_boot' since we are only
# interested in the type.
module_name = etype.replace('-', '_')
if '@' in module_name:
module_name = module_name.split('@')[0]
module = modules.get(module_name)
# Also allow entry-type modules to be brought in from the etype directory.
# Import the module if we have not already done so.
if not module:
old_path = sys.path
sys.path.insert(0, os.path.join(our_path, 'etype'))
try:
if have_importlib:
module = importlib.import_module(module_name)
else:
module = __import__(module_name)
except ImportError as e:
raise ValueError("Unknown entry type '%s' in node '%s' (expected etype/%s.py, error '%s'" %
(etype, node_path, module_name, e))
finally:
sys.path = old_path
modules[module_name] = module
# Look up the expected class name
return getattr(module, 'Entry_%s' % module_name)
@staticmethod
def Create(section, node, etype=None):
"""Create a new entry for a node.
Args:
section: Section object containing this node
node: Node object containing information about the entry to
create
etype: Entry type to use, or None to work it out (used for tests)
Returns:
A new Entry object of the correct type (a subclass of Entry)
"""
if not etype:
etype = fdt_util.GetString(node, 'type', node.name)
obj = Entry.Lookup(node.path, etype)
# Call its constructor to get the object we want.
return obj(section, etype, node)
def ReadNode(self):
"""Read entry information from the node
This reads all the fields we recognise from the node, ready for use.
"""
if 'pos' in self._node.props:
self.Raise("Please use 'offset' instead of 'pos'")
self.offset = fdt_util.GetInt(self._node, 'offset')
self.size = fdt_util.GetInt(self._node, 'size')
self.orig_offset = self.offset
self.orig_size = self.size
# These should not be set in input files, but are set in an FDT map,
# which is also read by this code.
self.image_pos = fdt_util.GetInt(self._node, 'image-pos')
self.uncomp_size = fdt_util.GetInt(self._node, 'uncomp-size')
self.align = fdt_util.GetInt(self._node, 'align')
if tools.NotPowerOfTwo(self.align):
raise ValueError("Node '%s': Alignment %s must be a power of two" %
(self._node.path, self.align))
self.pad_before = fdt_util.GetInt(self._node, 'pad-before', 0)
self.pad_after = fdt_util.GetInt(self._node, 'pad-after', 0)
self.align_size = fdt_util.GetInt(self._node, 'align-size')
if tools.NotPowerOfTwo(self.align_size):
binman: Convert Image to a subclass of Entry When support for sections (and thus hierarchical images) was added to binman, the decision was made to create a new Section class which could be used by both Image and an Entry_section class. The decision between using inheritance and composition was tricky to make, but in the end it was decided that Image was different enough from Entry that it made sense to put the implementation of sections in an entirely separate class. It also has the advantage that core Image code does have to rely on an entry class in the etype directory. This work was mostly completed in commit: 8f1da50ccc "binman: Refactor much of the image code into 'section' As a result of this, the Section class has its own version of things like offset and size and these must be kept in sync with the parent Entry_section class in some cases. In the last year it has become apparent that the cost of keeping things in sync is larger than expected, since more and more code wants to access these properties. An alternative approach, previously considered and rejected, now seems better. Adjust Image to be a subclass of Entry_section. Move the code from Section (in bsection.py) to Entry_section and delete Section. Update all tests accordingly. This requires substantial changes to Image. Overall the changes reduce code size by about 240 lines. While much of that is just boilerplate from Section, there are quite a few functions in Entry_section which now do not need to be overiden from Entry. This suggests the change is beneficial even without further functionality being added. A side benefit is that the properties of sections are now consistent with other entries. This fixes a problem in testListCmd() where some properties are missing for sections. Unfortunately this is a very large commit since it is not feasible to do the migration piecemeal. Given the substantial tests available and the 100% code coverage of binman, we should be able to do this safely. Signed-off-by: Simon Glass <sjg@chromium.org>
2019-07-08 20:25:47 +00:00
self.Raise("Alignment size %s must be a power of two" %
self.align_size)
self.align_end = fdt_util.GetInt(self._node, 'align-end')
self.offset_unset = fdt_util.GetBool(self._node, 'offset-unset')
self.expand_size = fdt_util.GetBool(self._node, 'expand-size')
def GetDefaultFilename(self):
return None
def GetFdtSet(self):
"""Get the set of device trees used by this entry
Returns:
Set containing the filename from this entry, if it is a .dtb, else
an empty set
"""
return set()
def ExpandEntries(self):
pass
def AddMissingProperties(self):
"""Add new properties to the device tree as needed for this entry"""
for prop in ['offset', 'size', 'image-pos']:
if not prop in self._node.props:
state.AddZeroProp(self._node, prop)
if self.compress != 'none':
state.AddZeroProp(self._node, 'uncomp-size')
err = state.CheckAddHashProp(self._node)
if err:
self.Raise(err)
def SetCalculatedProperties(self):
"""Set the value of device-tree properties calculated by binman"""
state.SetInt(self._node, 'offset', self.offset)
state.SetInt(self._node, 'size', self.size)
binman: Convert Image to a subclass of Entry When support for sections (and thus hierarchical images) was added to binman, the decision was made to create a new Section class which could be used by both Image and an Entry_section class. The decision between using inheritance and composition was tricky to make, but in the end it was decided that Image was different enough from Entry that it made sense to put the implementation of sections in an entirely separate class. It also has the advantage that core Image code does have to rely on an entry class in the etype directory. This work was mostly completed in commit: 8f1da50ccc "binman: Refactor much of the image code into 'section' As a result of this, the Section class has its own version of things like offset and size and these must be kept in sync with the parent Entry_section class in some cases. In the last year it has become apparent that the cost of keeping things in sync is larger than expected, since more and more code wants to access these properties. An alternative approach, previously considered and rejected, now seems better. Adjust Image to be a subclass of Entry_section. Move the code from Section (in bsection.py) to Entry_section and delete Section. Update all tests accordingly. This requires substantial changes to Image. Overall the changes reduce code size by about 240 lines. While much of that is just boilerplate from Section, there are quite a few functions in Entry_section which now do not need to be overiden from Entry. This suggests the change is beneficial even without further functionality being added. A side benefit is that the properties of sections are now consistent with other entries. This fixes a problem in testListCmd() where some properties are missing for sections. Unfortunately this is a very large commit since it is not feasible to do the migration piecemeal. Given the substantial tests available and the 100% code coverage of binman, we should be able to do this safely. Signed-off-by: Simon Glass <sjg@chromium.org>
2019-07-08 20:25:47 +00:00
base = self.section.GetRootSkipAtStart() if self.section else 0
state.SetInt(self._node, 'image-pos', self.image_pos - base)
if self.uncomp_size is not None:
state.SetInt(self._node, 'uncomp-size', self.uncomp_size)
state.CheckSetHashValue(self._node, self.GetData)
def ProcessFdt(self, fdt):
"""Allow entries to adjust the device tree
Some entries need to adjust the device tree for their purposes. This
may involve adding or deleting properties.
Returns:
True if processing is complete
False if processing could not be completed due to a dependency.
This will cause the entry to be retried after others have been
called
"""
return True
def SetPrefix(self, prefix):
"""Set the name prefix for a node
Args:
prefix: Prefix to set, or '' to not use a prefix
"""
if prefix:
self.name = prefix + self.name
def SetContents(self, data):
"""Set the contents of an entry
This sets both the data and content_size properties
Args:
data: Data to set to the contents (bytes)
"""
self.data = data
self.contents_size = len(self.data)
def ProcessContentsUpdate(self, data):
"""Update the contents of an entry, after the size is fixed
This checks that the new data is the same size as the old. If the size
has changed, this triggers a re-run of the packing algorithm.
Args:
data: Data to set to the contents (bytes)
Raises:
ValueError if the new data size is not the same as the old
"""
size_ok = True
new_size = len(data)
if state.AllowEntryExpansion():
if new_size > self.contents_size:
tout.Debug("Entry '%s' size change from %#x to %#x" % (
self._node.path, self.contents_size, new_size))
# self.data will indicate the new size needed
size_ok = False
elif new_size != self.contents_size:
self.Raise('Cannot update entry size from %d to %d' %
(self.contents_size, new_size))
self.SetContents(data)
return size_ok
def ObtainContents(self):
"""Figure out the contents of an entry.
Returns:
True if the contents were found, False if another call is needed
after the other entries are processed.
"""
# No contents by default: subclasses can implement this
return True
def ResetForPack(self):
"""Reset offset/size fields so that packing can be done again"""
self.offset = self.orig_offset
self.size = self.orig_size
def Pack(self, offset):
"""Figure out how to pack the entry into the section
Most of the time the entries are not fully specified. There may be
an alignment but no size. In that case we take the size from the
contents of the entry.
If an entry has no hard-coded offset, it will be placed at @offset.
Once this function is complete, both the offset and size of the
entry will be know.
Args:
Current section offset pointer
Returns:
New section offset pointer (after this entry)
"""
if self.offset is None:
if self.offset_unset:
self.Raise('No offset set with offset-unset: should another '
'entry provide this correct offset?')
self.offset = tools.Align(offset, self.align)
needed = self.pad_before + self.contents_size + self.pad_after
needed = tools.Align(needed, self.align_size)
size = self.size
if not size:
size = needed
new_offset = self.offset + size
aligned_offset = tools.Align(new_offset, self.align_end)
if aligned_offset != new_offset:
size = aligned_offset - self.offset
new_offset = aligned_offset
if not self.size:
self.size = size
if self.size < needed:
self.Raise("Entry contents size is %#x (%d) but entry size is "
"%#x (%d)" % (needed, needed, self.size, self.size))
# Check that the alignment is correct. It could be wrong if the
# and offset or size values were provided (i.e. not calculated), but
# conflict with the provided alignment values
if self.size != tools.Align(self.size, self.align_size):
self.Raise("Size %#x (%d) does not match align-size %#x (%d)" %
(self.size, self.size, self.align_size, self.align_size))
if self.offset != tools.Align(self.offset, self.align):
self.Raise("Offset %#x (%d) does not match align %#x (%d)" %
(self.offset, self.offset, self.align, self.align))
return new_offset
def Raise(self, msg):
"""Convenience function to raise an error referencing a node"""
raise ValueError("Node '%s': %s" % (self._node.path, msg))
def GetEntryArgsOrProps(self, props, required=False):
"""Return the values of a set of properties
Args:
props: List of EntryArg objects
Raises:
ValueError if a property is not found
"""
values = []
missing = []
for prop in props:
python_prop = prop.name.replace('-', '_')
if hasattr(self, python_prop):
value = getattr(self, python_prop)
else:
value = None
if value is None:
value = self.GetArg(prop.name, prop.datatype)
if value is None and required:
missing.append(prop.name)
values.append(value)
if missing:
self.Raise('Missing required properties/entry args: %s' %
(', '.join(missing)))
return values
def GetPath(self):
"""Get the path of a node
Returns:
Full path of the node for this entry
"""
return self._node.path
def GetData(self):
return self.data
def GetOffsets(self):
"""Get the offsets for siblings
Some entry types can contain information about the position or size of
other entries. An example of this is the Intel Flash Descriptor, which
knows where the Intel Management Engine section should go.
If this entry knows about the position of other entries, it can specify
this by returning values here
Returns:
Dict:
key: Entry type
value: List containing position and size of the given entry
type. Either can be None if not known
"""
return {}
def SetOffsetSize(self, offset, size):
"""Set the offset and/or size of an entry
Args:
offset: New offset, or None to leave alone
size: New size, or None to leave alone
"""
if offset is not None:
self.offset = offset
if size is not None:
self.size = size
def SetImagePos(self, image_pos):
"""Set the position in the image
Args:
image_pos: Position of this entry in the image
"""
self.image_pos = image_pos + self.offset
def ProcessContents(self):
"""Do any post-packing updates of entry contents
This function should call ProcessContentsUpdate() to update the entry
contents, if necessary, returning its return value here.
Args:
data: Data to set to the contents (bytes)
Returns:
True if the new data size is OK, False if expansion is needed
Raises:
ValueError if the new data size is not the same as the old and
state.AllowEntryExpansion() is False
"""
return True
binman: Support accessing binman tables at run time Binman construct images consisting of multiple binary files. These files sometimes need to know (at run timme) where their peers are located. For example, SPL may want to know where U-Boot is located in the image, so that it can jump to U-Boot correctly on boot. In general the positions where the binaries end up after binman has finished packing them cannot be known at compile time. One reason for this is that binman does not know the size of the binaries until everything is compiled, linked and converted to binaries with objcopy. To make this work, we add a feature to binman which checks each binary for symbol names starting with '_binman'. These are then decoded to figure out which entry and property they refer to. Then binman writes the value of this symbol into the appropriate binary. With this, the symbol will have the correct value at run time. Macros are used to make this easier to use. As an example, this declares a symbol that will access the 'u-boot-spl' entry to find the 'pos' value (i.e. the position of SPL in the image): binman_sym_declare(unsigned long, u_boot_spl, pos); This converts to a symbol called '_binman_u_boot_spl_prop_pos' in any binary that includes it. Binman then updates the value in that binary, ensuring that it can be accessed at runtime with: ulong u_boot_pos = binman_sym(ulong, u_boot_spl, pos); This assigns the variable u_boot_pos to the position of SPL in the image. Signed-off-by: Simon Glass <sjg@chromium.org>
2017-11-14 01:55:01 +00:00
def WriteSymbols(self, section):
binman: Support accessing binman tables at run time Binman construct images consisting of multiple binary files. These files sometimes need to know (at run timme) where their peers are located. For example, SPL may want to know where U-Boot is located in the image, so that it can jump to U-Boot correctly on boot. In general the positions where the binaries end up after binman has finished packing them cannot be known at compile time. One reason for this is that binman does not know the size of the binaries until everything is compiled, linked and converted to binaries with objcopy. To make this work, we add a feature to binman which checks each binary for symbol names starting with '_binman'. These are then decoded to figure out which entry and property they refer to. Then binman writes the value of this symbol into the appropriate binary. With this, the symbol will have the correct value at run time. Macros are used to make this easier to use. As an example, this declares a symbol that will access the 'u-boot-spl' entry to find the 'pos' value (i.e. the position of SPL in the image): binman_sym_declare(unsigned long, u_boot_spl, pos); This converts to a symbol called '_binman_u_boot_spl_prop_pos' in any binary that includes it. Binman then updates the value in that binary, ensuring that it can be accessed at runtime with: ulong u_boot_pos = binman_sym(ulong, u_boot_spl, pos); This assigns the variable u_boot_pos to the position of SPL in the image. Signed-off-by: Simon Glass <sjg@chromium.org>
2017-11-14 01:55:01 +00:00
"""Write symbol values into binary files for access at run time
Args:
section: Section containing the entry
binman: Support accessing binman tables at run time Binman construct images consisting of multiple binary files. These files sometimes need to know (at run timme) where their peers are located. For example, SPL may want to know where U-Boot is located in the image, so that it can jump to U-Boot correctly on boot. In general the positions where the binaries end up after binman has finished packing them cannot be known at compile time. One reason for this is that binman does not know the size of the binaries until everything is compiled, linked and converted to binaries with objcopy. To make this work, we add a feature to binman which checks each binary for symbol names starting with '_binman'. These are then decoded to figure out which entry and property they refer to. Then binman writes the value of this symbol into the appropriate binary. With this, the symbol will have the correct value at run time. Macros are used to make this easier to use. As an example, this declares a symbol that will access the 'u-boot-spl' entry to find the 'pos' value (i.e. the position of SPL in the image): binman_sym_declare(unsigned long, u_boot_spl, pos); This converts to a symbol called '_binman_u_boot_spl_prop_pos' in any binary that includes it. Binman then updates the value in that binary, ensuring that it can be accessed at runtime with: ulong u_boot_pos = binman_sym(ulong, u_boot_spl, pos); This assigns the variable u_boot_pos to the position of SPL in the image. Signed-off-by: Simon Glass <sjg@chromium.org>
2017-11-14 01:55:01 +00:00
"""
pass
def CheckOffset(self):
"""Check that the entry offsets are correct
This is used for entries which have extra offset requirements (other
than having to be fully inside their section). Sub-classes can implement
this function and raise if there is a problem.
"""
pass
@staticmethod
def GetStr(value):
if value is None:
return '<none> '
return '%08x' % value
@staticmethod
def WriteMapLine(fd, indent, name, offset, size, image_pos):
print('%s %s%s %s %s' % (Entry.GetStr(image_pos), ' ' * indent,
Entry.GetStr(offset), Entry.GetStr(size),
name), file=fd)
def WriteMap(self, fd, indent):
"""Write a map of the entry to a .map file
Args:
fd: File to write the map to
indent: Curent indent level of map (0=none, 1=one level, etc.)
"""
self.WriteMapLine(fd, indent, self.name, self.offset, self.size,
self.image_pos)
def GetEntries(self):
"""Return a list of entries contained by this entry
Returns:
List of entries, or None if none. A normal entry has no entries
within it so will return None
"""
return None
def GetArg(self, name, datatype=str):
"""Get the value of an entry argument or device-tree-node property
Some node properties can be provided as arguments to binman. First check
the entry arguments, and fall back to the device tree if not found
Args:
name: Argument name
datatype: Data type (str or int)
Returns:
Value of argument as a string or int, or None if no value
Raises:
ValueError if the argument cannot be converted to in
"""
value = state.GetEntryArg(name)
if value is not None:
if datatype == int:
try:
value = int(value)
except ValueError:
self.Raise("Cannot convert entry arg '%s' (value '%s') to integer" %
(name, value))
elif datatype == str:
pass
else:
raise ValueError("GetArg() internal error: Unknown data type '%s'" %
datatype)
else:
value = fdt_util.GetDatatype(self._node, name, datatype)
return value
@staticmethod
def WriteDocs(modules, test_missing=None):
"""Write out documentation about the various entry types to stdout
Args:
modules: List of modules to include
test_missing: Used for testing. This is a module to report
as missing
"""
print('''Binman Entry Documentation
===========================
This file describes the entry types supported by binman. These entry types can
be placed in an image one by one to build up a final firmware image. It is
fairly easy to create new entry types. Just add a new file to the 'etype'
directory. You can use the existing entries as examples.
Note that some entries are subclasses of others, using and extending their
features to produce new behaviours.
''')
modules = sorted(modules)
# Don't show the test entry
if '_testing' in modules:
modules.remove('_testing')
missing = []
for name in modules:
if name.startswith('__'):
continue
module = Entry.Lookup(name, name)
docs = getattr(module, '__doc__')
if test_missing == name:
docs = None
if docs:
lines = docs.splitlines()
first_line = lines[0]
rest = [line[4:] for line in lines[1:]]
hdr = 'Entry: %s: %s' % (name.replace('_', '-'), first_line)
print(hdr)
print('-' * len(hdr))
print('\n'.join(rest))
print()
print()
else:
missing.append(name)
if missing:
raise ValueError('Documentation is missing for modules: %s' %
', '.join(missing))
def GetUniqueName(self):
"""Get a unique name for a node
Returns:
String containing a unique name for a node, consisting of the name
of all ancestors (starting from within the 'binman' node) separated
by a dot ('.'). This can be useful for generating unique filesnames
in the output directory.
"""
name = self.name
node = self._node
while node.parent:
node = node.parent
if node.name == 'binman':
break
name = '%s.%s' % (node.name, name)
return name
def ExpandToLimit(self, limit):
"""Expand an entry so that it ends at the given offset limit"""
if self.offset + self.size < limit:
self.size = limit - self.offset
# Request the contents again, since changing the size requires that
# the data grows. This should not fail, but check it to be sure.
if not self.ObtainContents():
self.Raise('Cannot obtain contents when expanding entry')
def HasSibling(self, name):
"""Check if there is a sibling of a given name
Returns:
True if there is an entry with this name in the the same section,
else False
"""
return name in self.section.GetEntries()
def GetSiblingImagePos(self, name):
"""Return the image position of the given sibling
Returns:
Image position of sibling, or None if the sibling has no position,
or False if there is no such sibling
"""
if not self.HasSibling(name):
return False
return self.section.GetEntries()[name].image_pos
@staticmethod
def AddEntryInfo(entries, indent, name, etype, size, image_pos,
uncomp_size, offset, entry):
"""Add a new entry to the entries list
Args:
entries: List (of EntryInfo objects) to add to
indent: Current indent level to add to list
name: Entry name (string)
etype: Entry type (string)
size: Entry size in bytes (int)
image_pos: Position within image in bytes (int)
uncomp_size: Uncompressed size if the entry uses compression, else
None
offset: Entry offset within parent in bytes (int)
entry: Entry object
"""
entries.append(EntryInfo(indent, name, etype, size, image_pos,
uncomp_size, offset, entry))
def ListEntries(self, entries, indent):
"""Add files in this entry to the list of entries
This can be overridden by subclasses which need different behaviour.
Args:
entries: List (of EntryInfo objects) to add to
indent: Current indent level to add to list
"""
self.AddEntryInfo(entries, indent, self.name, self.etype, self.size,
self.image_pos, self.uncomp_size, self.offset, self)
def ReadData(self, decomp=True):
"""Read the data for an entry from the image
This is used when the image has been read in and we want to extract the
data for a particular entry from that image.
Args:
decomp: True to decompress any compressed data before returning it;
False to return the raw, uncompressed data
Returns:
Entry data (bytes)
"""
# Use True here so that we get an uncompressed section to work from,
# although compressed sections are currently not supported
data = self.section.ReadData(True)
tout.Info('%s: Reading data from offset %#x-%#x, size %#x (avail %#x)' %
(self.GetPath(), self.offset, self.offset + self.size,
self.size, len(data)))
return data[self.offset:self.offset + self.size]