1
=======================================
2
LLVM's Optional Rich Disassembly Output
3
=======================================
11
LLVM's default disassembly output is raw text. To allow consumers more ability
12
to introspect the instructions' textual representation or to reformat for a more
13
user friendly display there is an optional rich disassembly output.
15
This optional output is sufficient to reference into individual portions of the
16
instruction text. This is intended for clients like disassemblers, list file
17
generators, and pretty-printers, which need more than the raw instructions and
18
the ability to print them.
20
To provide this functionality the assembly text is marked up with annotations.
21
The markup is simple enough in syntax to be robust even in the case of version
22
mismatches between consumers and producers. That is, the syntax generally does
23
not carry semantics beyond "this text has an annotation," so consumers can
24
simply ignore annotations they do not understand or do not care about.
26
After calling ``LLVMCreateDisasm()`` to create a disassembler context the
27
optional output is enable with this call:
31
LLVMSetDisasmOptions(DC, LLVMDisassembler_Option_UseMarkup);
33
Then subsequent calls to ``LLVMDisasmInstruction()`` will return output strings
34
with the marked up annotations.
36
Instruction Annotations
37
=======================
39
.. _contextual markups:
44
Annoated assembly display will supply contextual markup to help clients more
45
efficiently implement things like pretty printers. Most markup will be target
46
independent, so clients can effectively provide good display without any target
49
Annotated assembly goes through the normal instruction printer, but optionally
50
includes contextual tags on portions of the instruction string. An annotation
51
is any '<' '>' delimited section of text(1).
55
annotation: '<' tag-name tag-modifier-list ':' annotated-text '>'
57
tag-modifier-list: comma delimited identifier list
59
The tag-name is an identifier which gives the type of the annotation. For the
60
first pass, this will be very simple, with memory references, registers, and
61
immediates having the tag names "mem", "reg", and "imm", respectively.
63
The tag-modifier-list is typically additional target-specific context, such as
66
Clients should accept and ignore any tag-names or tag-modifiers they do not
67
understand, allowing the annotations to grow in richness without breaking older
70
For example, a possible annotation of an ARM load of a stack-relative location
71
might be annotated as:
75
ldr <reg gpr:r0>, <mem regoffset:[<reg gpr:sp>, <imm:#4>]>
78
1: For assembly dialects in which '<' and/or '>' are legal tokens, a literal token is escaped by following immediately with a repeat of the character. For example, a literal '<' character is output as '<<' in an annotated assembly string.
83
The intended consumers of this information use the C API, therefore the new C
84
API function for the disassembler will be added to provide an option to produce
85
disassembled instructions with annotations, ``LLVMSetDisasmOptions()`` and the
86
``LLVMDisassembler_Option_UseMarkup`` option (see above).