Docs: Add initial XML dovcument description. #2136.

Makefile.in

@@ -148,6 +148,7 @@ DISTFILES_INC = $(INFOS) .gitignore \
 	docs/install.adoc \
 	docs/internals.adoc \
 	docs/verilator_logo.png \
+	docs/xml.adoc \
 	install-sh configure *.pod \
 	include/*.[chv]* \
 	include/*.in \

docs/xml.adoc

@@ -0,0 +1,65 @@
+= Verilator XML Output
+:toc: right
+
+// Github doesn't render unless absolute URL
+image::https://www.veripool.org/img/verilator_256_200_min.png[Logo,256,200,role="right"]
+
+== Introduction
+
+This document describes Verilator's XML output. For more general information
+please see https://verilator.org[verilator.org].
+
+== General
+
+Verilator's XML output is enabled with the `--xml-only` flag. It contains
+limited information about the elaborated design including files, modules,
+instance hierarchy, logic and data types. There is no formal schema since part
+of the structure of the XML document matches the compiled code which would
+require the schema to describe legal SystemVerilog structure. The intended
+usage is to enable other downstream tools to take advantage of Verilator's
+parser.
+
+== Structure
+
+The XML document is consists of 4 sections within the top level `verilator_xml`
+element:
+
+`<files>`...`</files>`::
+
+This section contains a list of all design files read, including the
+built-in constructs and the command line as their own entries.  Each
+`<file>` has an attribute `id` which is a short ASCII string unique to that
+file. Other elements' `loc` attributes use this id to refer to a particular
+file.
+
+`<module_files>`...`</module_files>`::
+
+All files containing Verilog module definitions are listed in this section.
+This element's contents is a subset of the `<files>` element's contents.
+
+`<cells>`...`</cells>`::
+
+The cells section of the XML document contains the design instance
+hierarchy.  Each instance is represented with the `<cell>` element with the
+following attributes:
+
+* `fl` (deprecated): The file id and line number where the module was
+  instanced.  Use `loc` instead.
+
+* `loc`: The file id, first line number, last line number, first column
+   number and last column number of the identifier where the module was
+   instanced, separated by commas.
+
+* `name`: The instance name.
+
+* `submodname`: The module name uniquified with particular parameter values (if any).
+
+* `hier`: The full hierarchy path.
+
+`<netlist>`...`</netlist>`::
+
+The netlist section contains a number of `<module>`...`</module>` elements,
+each describing the contents of that module, and a single `<typetable>`...
+`</typetable>` element which lists all used types used within the
+modules. Each type has a numeric `id` attribute that is referred to by
+elements in the `<module>` elements using the `dtype_id` attribute.