Get desktop application:
View/edit binary Protocol Buffers messages
No longer written. This is here so that BinDiff can work with older BinExport files.
Rich comment index used for BinDiff's comment porting.
Used in:
Implicit: instruction sequence
This is a space optimization. The instructions for an individual basic block will usually be in a continuous index range. Thus it is more efficient to store the range instead of individual indices. However, this does not hold true for all basic blocks, so we need to be able to store multiple index ranges per block.
Used in:
These work like begin and end iterators, i.e. the sequence is [begin_index, end_index). If the sequence only contains a single element end_index will be omitted.
Used in:
vertices == functions in the call graph. Important: Most downstream tooling (notably BinDiff), need these to be sorted by `Vertex::address` (ascending). For C++, the `BinExport2Writer` class enforces this invariant.
edges == calls in the call graph.
Used in:
source and target index into the vertex repeated field.
Used in:
The function's entry point address. Messages need to be sorted, see comment below on `vertex`.
If the function has a user defined, real name it will be given here. main() is a proper name, sub_BAADF00D is not (auto generated dummy name).
Demangled name if the function is a mangled C++ function and we could demangle it.
If this is a library function, what is its index in library arrays.
If module name, such as class name for DEX files, is present - index in module table.
Used in:
Regular function with full disassembly.
This function is a well known library function.
Imported from a dynamic link library (e.g. dll).
A thunk function, forwarding its work via an unconditional jump.
An invalid function (a function that contained invalid code or was considered invalid by some heuristics).
Used in:
Index into the global instruction table. This is here to enable comment processing without having to iterate over all instructions. There is an N:M mapping of instructions to comments.
Index into the operand array local to an instruction.
Index into the expression array local to an operand, like in Reference. This is not currently used, but allows to implement expression substitutions.
Index into the global string table.
Comment is propagated to all locations that reference the original location.
Used in:
A regular instruction comment. Typically displayed next to the instruction disassembly.
A comment line that is typically displayed before (above) the instruction it refers to.
Like ANTERIOR, but a typically displayed after (below).
Similar to an ANTERIOR comment, but applies to the beginning of an identified function. Programs displaying the proto may choose to render these differently (e.g. above an inferred function signature).
Named constants, bitfields and similar.
Named locations, usually the target of a jump.
Data cross references.
Local/stack variables.
Used in:
Index into the global instruction table.
Address being referred.
An operand consists of 1 or more expressions, linked together as a tree.
Used in:
IMMEDIATE_INT is by far the most common type and thus we can save some space by omitting it as the default.
Symbol for this expression. Interpretation depends on type. Examples include: "eax", "[", "+"
If the expression can be interpreted as an integer value (IMMEDIATE_INT) the value is given here.
The parent expression. Example expression tree for the second operand of: mov eax, b4 [ebx + 12] "b4" --- "[" --- "+" --- "ebx" \ "12"
true if the expression has entry in relocation table
Used in:
Used in:
Basic blocks are sorted by address.
The flow graph's entry point address is the first instruction of the entry_basic_block.
Used in:
Source instruction will always be the last instruction of the source basic block, target instruction the first instruction of the target basic block.
Indicates whether this is a loop edge as determined by Lengauer-Tarjan.
Used in:
Used in:
This will only be filled for instructions that do not just flow from the immediately preceding instruction. Regular instructions will have to calculate their own address by adding raw_bytes.size() to the previous instruction's address.
If this is a call instruction and call targets could be determined they'll be given here. Note that we may or may not have a flow graph for the target and thus cannot use an index into the flow graph table here. We could potentially use call graph nodes, but linking instructions to the call graph directly does not seem a good choice.
Index into the mnemonic array of strings. Used for de-duping the data. The default value is used for the most common mnemonic in the executable.
Indices into the operand tree. On X86 this can be 0, 1 or 2 elements long, 3 elements with VEX/EVEX. Implicit: operand sequence
The unmodified input bytes corresponding to this instruction.
Implicit: comment sequence
Used in:
If this library is statically linked.
Address where this library was loaded, 0 if unknown.
Name of the library (format is platform-dependent).
Used in:
Input binary filename including file extension but excluding file path. example: "insider_gcc.exe"
Application defined executable id. Often the SHA256 hash of the input binary.
Input architecture name, e.g. x86-32.
When did this file get created? Unix time. This may be used for some primitive versioning in case the file format ever changes.
An instruction has exactly 1 mnemonic.
Used in:
Literal representation of the mnemonic, e.g.: "mov".
Used in:
Name, such as Java class name. Platform-dependent.
An instruction may have 0 or more operands.
Used in:
Contains all expressions constituting this operand. All expressions should be linked into a single tree, i.e. there should only be one expression in this list with parent_index == NULL and all others should descend from that. Rendering order for expressions on the same tree level (siblings) is implicitly given by the order they are referenced in this repeated field. Implicit: expression sequence
Generic reference class used for address comments (deprecated), string references and expression substitutions. It allows referencing from an instruction, operand, expression subtree tuple to a de-duped string in the string table.
Used in:
Index into the global instruction table.
Index into the operand array local to an instruction.
Index into the expression array local to an operand.
Index into the global string table.
Used in:
Section start address.
Section size.
Read flag of the section, True when section is readable.
Write flag of the section, True when section is writable.
Execute flag of the section, True when section is executable.