board_inspector/extractors: framework to support extractors

This patch makes the `run.py` enumerate and invoke all extractors (whose
name should be `##-<name>.py` where `##` is a decimal number for ordering)
under the extractors/ directory. Only some helper subroutines are added in
this patch; the actual extractors will be added in the subsequent patches
in this series.

v1 -> v2:
 - Allow an extractor to be classified as advanced by defining the variable
   `advanced` to True. Advanced extractors are not enabled by default and
   can be invoked by passing `--advanced` to the board inspector.

Tracked-On: #5922
Signed-off-by: Junjie Mao <junjie.mao@intel.com>

misc/config_tools/board_inspector/cli.py

@@ -15,7 +15,7 @@ from importlib import import_module
 script_dir = os.path.dirname(os.path.realpath(__file__))
 sys.path.append(os.path.join(script_dir))
 
-def main(board_name, board_xml):
+def main(board_name, board_xml, args):
     try:
         # First invoke the legacy board parser to create the board XML ...
         legacy_parser = os.path.join(script_dir, "legacy", "board_parser.py")
@@ -24,12 +24,28 @@ def main(board_name, board_xml):
 
         # ... then load the created board XML and append it with additional data by invoking the extractors.
         board_etree = lxml.etree.parse(board_xml)
+        root_node = board_etree.getroot()
 
         # Clear the whitespaces between adjacent children under the root node
-        board_etree.getroot().text = None
-        for elem in board_etree.getroot():
+        root_node.text = None
+        for elem in root_node:
             elem.tail = None
 
+        # Create nodes for each kind of resource
+        root_node.append(lxml.etree.Element("processors"))
+        root_node.append(lxml.etree.Element("caches"))
+        root_node.append(lxml.etree.Element("memory"))
+        root_node.append(lxml.etree.Element("devices"))
+
+        extractors_path = os.path.join(script_dir, "extractors")
+        extractors = [f for f in os.listdir(extractors_path) if f[:2].isdigit()]
+        for extractor in sorted(extractors):
+            module_name = os.path.splitext(extractor)[0]
+            module = import_module(f"extractors.{module_name}")
+            if not args.advanced and getattr(module, "advanced", False):
+                continue
+            module.extract(board_etree)
+
         # Finally overwrite the output with the updated XML
         board_etree.write(board_xml, pretty_print=True)
 
@@ -42,7 +58,8 @@ if __name__ == "__main__":
     parser = argparse.ArgumentParser()
     parser.add_argument("board_name", help="the name of the board that runs the ACRN hypervisor")
     parser.add_argument("--out", help="the name of board info file")
+    parser.add_argument("--advanced", action="store_true", default=False, help="extract advanced information such as ACPI namespace")
     args = parser.parse_args()
 
     board_xml = args.out if args.out else f"{args.board_name}.xml"
-    main(args.board_name, board_xml)
+    main(args.board_name, board_xml, args)

misc/config_tools/board_inspector/extractors/helpers.py

@@ -0,0 +1,19 @@
+# Copyright (C) 2021 Intel Corporation. All rights reserved.
+#
+# SPDX-License-Identifier: BSD-3-Clause
+#
+
+import lxml
+
+def add_child(element, tag, text=None, **kwargs):
+    child = lxml.etree.Element(tag)
+    child.text = text
+    for k,v in kwargs.items():
+        child.set(k, v)
+    element.append(child)
+    return child
+
+def get_node(etree, xpath):
+    result = etree.xpath(xpath)
+    assert len(result) <= 1, "Internal error: cannot get texts from multiple nodes at a time"
+    return result[0] if len(result) == 1 else None