u-boot/tools/binman/image.py

391 lines
15 KiB
Python
Raw Normal View History

# SPDX-License-Identifier: GPL-2.0+
# Copyright (c) 2016 Google, Inc
# Written by Simon Glass <sjg@chromium.org>
#
# Class for an image, the output of binman
#
from collections import OrderedDict
import fnmatch
from operator import attrgetter
import os
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
import re
import sys
from binman.entry import Entry
from binman.etype import fdtmap
from binman.etype import image_header
from binman.etype import section
from dtoc import fdt
from dtoc import fdt_util
from patman import tools
from patman import tout
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
class Image(section.Entry_section):
"""A Image, representing an output from binman
An image is comprised of a collection of entries each containing binary
data. The image size must be large enough to hold all of this data.
This class implements the various operations needed for images.
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
Attributes:
filename: Output filename for image
image_node: Name of node containing the description for this image
fdtmap_dtb: Fdt object for the fdtmap when loading from a file
fdtmap_data: Contents of the fdtmap when loading from a file
allow_repack: True to add properties to allow the image to be safely
repacked later
test_section_timeout: Use a zero timeout for section multi-threading
(for testing)
Args:
copy_to_orig: Copy offset/size to orig_offset/orig_size after reading
from the device tree
test: True if this is being called from a test of Images. This this case
there is no device tree defining the structure of the section, so
we create a section manually.
ignore_missing: Ignore any missing entry arguments (i.e. don't raise an
exception). This should be used if the Image is being loaded from
a file rather than generated. In that case we obviously don't need
the entry arguments since the contents already exists.
use_expanded: True if we are updating the FDT wth entry offsets, etc.
and should use the expanded versions of the U-Boot entries.
Any entry type that includes a devicetree must put it in a
separate entry so that it will be updated. For example. 'u-boot'
normally just picks up 'u-boot.bin' which includes the
devicetree, but this is not updateable, since it comes into
binman as one piece and binman doesn't know that it is actually
an executable followed by a devicetree. Of course it could be
taught this, but then when reading an image (e.g. 'binman ls')
it may need to be able to split the devicetree out of the image
in order to determine the location of things. Instead we choose
to ignore 'u-boot-bin' in this case, and build it ourselves in
binman with 'u-boot-dtb.bin' and 'u-boot.dtb'. See
Entry_u_boot_expanded and Entry_blob_phase for details.
"""
def __init__(self, name, node, copy_to_orig=True, test=False,
ignore_missing=False, use_expanded=False):
super().__init__(None, 'section', node, test=test)
self.copy_to_orig = copy_to_orig
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.name = 'main-section'
self.image_name = name
self._filename = '%s.bin' % self.image_name
self.fdtmap_dtb = None
self.fdtmap_data = None
self.allow_repack = False
self._ignore_missing = ignore_missing
self.use_expanded = use_expanded
self.test_section_timeout = False
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
if not test:
self.ReadNode()
def ReadNode(self):
super().ReadNode()
filename = fdt_util.GetString(self._node, 'filename')
if filename:
self._filename = filename
self.allow_repack = fdt_util.GetBool(self._node, 'allow-repack')
@classmethod
def FromFile(cls, fname):
"""Convert an image file into an Image for use in binman
Args:
fname: Filename of image file to read
Returns:
Image object on success
Raises:
ValueError if something goes wrong
"""
data = tools.ReadFile(fname)
size = len(data)
# First look for an image header
pos = image_header.LocateHeaderOffset(data)
if pos is None:
# Look for the FDT map
pos = fdtmap.LocateFdtmap(data)
if pos is None:
raise ValueError('Cannot find FDT map in image')
# We don't know the FDT size, so check its header first
probe_dtb = fdt.Fdt.FromData(
data[pos + fdtmap.FDTMAP_HDR_LEN:pos + 256])
dtb_size = probe_dtb.GetFdtObj().totalsize()
fdtmap_data = data[pos:pos + dtb_size + fdtmap.FDTMAP_HDR_LEN]
fdt_data = fdtmap_data[fdtmap.FDTMAP_HDR_LEN:]
out_fname = tools.GetOutputFilename('fdtmap.in.dtb')
tools.WriteFile(out_fname, fdt_data)
dtb = fdt.Fdt(out_fname)
dtb.Scan()
# Return an Image with the associated nodes
root = dtb.GetRoot()
image = Image('image', root, copy_to_orig=False, ignore_missing=True)
image.image_node = fdt_util.GetString(root, 'image-node', 'image')
image.fdtmap_dtb = dtb
image.fdtmap_data = fdtmap_data
image._data = data
image._filename = fname
image.image_name, _ = os.path.splitext(fname)
return image
def Raise(self, msg):
"""Convenience function to raise an error referencing an image"""
raise ValueError("Image '%s': %s" % (self._node.path, msg))
def PackEntries(self):
"""Pack all entries into the image"""
super().Pack(0)
def SetImagePos(self):
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
# This first section in the image so it starts at 0
super().SetImagePos(0)
def ProcessEntryContents(self):
"""Call the ProcessContents() method for each entry
This is intended to adjust the contents as needed by the entry type.
Returns:
True if the new data size is OK, False if expansion is needed
"""
return super().ProcessContents()
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):
"""Write symbol values into binary files for access at run time"""
super().WriteSymbols(self)
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
def BuildImage(self):
"""Write the image to a file"""
fname = tools.GetOutputFilename(self._filename)
tout.Info("Writing image to '%s'" % fname)
with open(fname, 'wb') as fd:
data = self.GetPaddedData()
fd.write(data)
tout.Info("Wrote %#x bytes" % len(data))
def WriteMap(self):
"""Write a map of the image to a .map file
Returns:
Filename of map file written
"""
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
filename = '%s.map' % self.image_name
fname = tools.GetOutputFilename(filename)
with open(fname, 'w') as fd:
print('%8s %8s %8s %s' % ('ImagePos', 'Offset', 'Size', 'Name'),
file=fd)
super().WriteMap(fd, 0)
return fname
def BuildEntryList(self):
"""List the files in an image
Returns:
List of entry.EntryInfo objects describing all entries in the image
"""
entries = []
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.ListEntries(entries, 0)
return entries
def FindEntryPath(self, entry_path):
"""Find an entry at a given path in the image
Args:
entry_path: Path to entry (e.g. /ro-section/u-boot')
Returns:
Entry object corresponding to that past
Raises:
ValueError if no entry found
"""
parts = entry_path.split('/')
entries = self.GetEntries()
parent = '/'
for part in parts:
entry = entries.get(part)
if not entry:
raise ValueError("Entry '%s' not found in '%s'" %
(part, parent))
parent = entry.GetPath()
entries = entry.GetEntries()
return entry
def ReadData(self, decomp=True):
tout.Debug("Image '%s' ReadData(), size=%#x" %
(self.GetPath(), len(self._data)))
return self._data
def GetListEntries(self, entry_paths):
"""List the entries in an image
This decodes the supplied image and returns a list of entries from that
image, preceded by a header.
Args:
entry_paths: List of paths to match (each can have wildcards). Only
entries whose names match one of these paths will be printed
Returns:
String error message if something went wrong, otherwise
3-Tuple:
List of EntryInfo objects
List of lines, each
List of text columns, each a string
List of widths of each column
"""
def _EntryToStrings(entry):
"""Convert an entry to a list of strings, one for each column
Args:
entry: EntryInfo object containing information to output
Returns:
List of strings, one for each field in entry
"""
def _AppendHex(val):
"""Append a hex value, or an empty string if val is None
Args:
val: Integer value, or None if none
"""
args.append('' if val is None else '>%x' % val)
args = [' ' * entry.indent + entry.name]
_AppendHex(entry.image_pos)
_AppendHex(entry.size)
args.append(entry.etype)
_AppendHex(entry.offset)
_AppendHex(entry.uncomp_size)
return args
def _DoLine(lines, line):
"""Add a line to the output list
This adds a line (a list of columns) to the output list. It also updates
the widths[] array with the maximum width of each column
Args:
lines: List of lines to add to
line: List of strings, one for each column
"""
for i, item in enumerate(line):
widths[i] = max(widths[i], len(item))
lines.append(line)
def _NameInPaths(fname, entry_paths):
"""Check if a filename is in a list of wildcarded paths
Args:
fname: Filename to check
entry_paths: List of wildcarded paths (e.g. ['*dtb*', 'u-boot*',
'section/u-boot'])
Returns:
True if any wildcard matches the filename (using Unix filename
pattern matching, not regular expressions)
False if not
"""
for path in entry_paths:
if fnmatch.fnmatch(fname, path):
return True
return False
entries = self.BuildEntryList()
# This is our list of lines. Each item in the list is a list of strings, one
# for each column
lines = []
HEADER = ['Name', 'Image-pos', 'Size', 'Entry-type', 'Offset',
'Uncomp-size']
num_columns = len(HEADER)
# This records the width of each column, calculated as the maximum width of
# all the strings in that column
widths = [0] * num_columns
_DoLine(lines, HEADER)
# We won't print anything unless it has at least this indent. So at the
# start we will print nothing, unless a path matches (or there are no
# entry paths)
MAX_INDENT = 100
min_indent = MAX_INDENT
path_stack = []
path = ''
indent = 0
selected_entries = []
for entry in entries:
if entry.indent > indent:
path_stack.append(path)
elif entry.indent < indent:
path_stack.pop()
if path_stack:
path = path_stack[-1] + '/' + entry.name
indent = entry.indent
# If there are entry paths to match and we are not looking at a
# sub-entry of a previously matched entry, we need to check the path
if entry_paths and indent <= min_indent:
if _NameInPaths(path[1:], entry_paths):
# Print this entry and all sub-entries (=higher indent)
min_indent = indent
else:
# Don't print this entry, nor any following entries until we get
# a path match
min_indent = MAX_INDENT
continue
_DoLine(lines, _EntryToStrings(entry))
selected_entries.append(entry)
return selected_entries, lines, widths
def LookupImageSymbol(self, sym_name, optional, msg, base_addr):
"""Look up a symbol in an ELF file
Looks up a symbol in an ELF file. Only entry types which come from an
ELF image can be used by this function.
This searches through this image including all of its subsections.
At present the only entry properties supported are:
offset
image_pos - 'base_addr' is added if this is not an end-at-4gb image
size
Args:
sym_name: Symbol name in the ELF file to look up in the format
_binman_<entry>_prop_<property> where <entry> is the name of
the entry and <property> is the property to find (e.g.
_binman_u_boot_prop_offset). As a special case, you can append
_any to <entry> to have it search for any matching entry. E.g.
_binman_u_boot_any_prop_offset will match entries called u-boot,
u-boot-img and u-boot-nodtb)
optional: True if the symbol is optional. If False this function
will raise if the symbol is not found
msg: Message to display if an error occurs
base_addr: Base address of image. This is added to the returned
image_pos in most cases so that the returned position indicates
where the targeted entry/binary has actually been loaded. But
if end-at-4gb is used, this is not done, since the binary is
already assumed to be linked to the ROM position and using
execute-in-place (XIP).
Returns:
Value that should be assigned to that symbol, or None if it was
optional and not found
Raises:
ValueError if the symbol is invalid or not found, or references a
property which is not supported
"""
entries = OrderedDict()
entries_by_name = {}
self._CollectEntries(entries, entries_by_name, self)
return self.LookupSymbol(sym_name, optional, msg, base_addr,
entries_by_name)