Migrate tests to RPyC (#97)

## Description/Motivation/Screenshots

This PR is the GEF-Extras counter-part of hugsy/gef#1040 

Also since `kallsyms` was removed from GEF (in preparation for the
kernel module), it was added here with its tests

## How Has This Been Tested ?

"Tested" indicates that the PR works *and* the unit test (i.e. `make
test`) run passes without issue.

*  [x] x86-32
*  [x] x86-64
*  [ ] ARM
*  [x] AARCH64
*  [ ] MIPS
*  [ ] POWERPC
*  [ ] SPARC
*  [ ] RISC-V

## Checklist

<!-- N.B.: Your patch won't be reviewed unless fulfilling the following
base requirements: -->
<!--- Put an `x` in all the boxes that are complete, or that don't apply
-->
*  [x] My code follows the code style of this project.
*  [x] My change includes a change to the documentation, if required.
*  [x] If my change adds new code,
[adequate tests](https://hugsy.github.io/gef/testing) have been added.
*  [x] I have read and agree to the

[CONTRIBUTING](https://github.com/hugsy/gef/blob/main/.github/CONTRIBUTING.md)
document.

---------

Co-authored-by: Grazfather <grazfather@gmail.com>

Author: hugsy

.github/workflows/tests.yml

@@ -2,6 +2,7 @@ name: CI Test for GEF-EXTRAS
 
 env:
   BRANCH: main
+  NB_CPU: 1
 
 on:
   push:
@@ -66,30 +67,18 @@ jobs:
 
     - name: Checkout GEF
       run: |
-        mkdir -p ${{ env.GEF_PATH_DIR }}
-        wget -O ${{ env.GEF_PATH }} https://raw.githubusercontent.com/hugsy/gef/${{ env.BRANCH }}/gef.py
+        git clone -b ${{ env.BRANCH }} https://github.com/hugsy/gef ${{ env.GEF_PATH_DIR }}
         echo "source ${{ env.GEF_PATH }}" > ~/.gdbinit
         gdb -q -ex 'gef missing' -ex 'gef help' -ex 'gef config' -ex start -ex continue -ex quit /bin/pwd
 
-    - name: Build config file
+    - name: Setup Tests
       run: |
-        gdb -q \
-          -ex "gef config pcustom.struct_path '$(pwd)/structs'" \
-          -ex "gef config syscall-args.path '$(pwd)/syscall-tables'" \
-          -ex "gef config context.libc_args True" \
-          -ex "gef config context.libc_args_path '$(pwd)/glibc-function-args'" \
-          -ex 'gef save' \
-          -ex quit
+        make -C tests/binaries -j ${{ env.NB_CPU }}
 
     - name: Run Tests
       run: |
-        make -C tests/binaries -j ${{ env.NB_CPU }}
         python${{ env.PY_VER }} -m pytest --forked -n ${{ env.NB_CPU }} -v -k "not benchmark" tests/
 
-    - name: Run linter
-      run: |
-        python${{ env.PY_VER }} -m pylint --rcfile=$(pwd)/.pylintrc gef.py tests/*/*.py
-
   standalone:
     runs-on: ubuntu-latest
     name: "Verify GEF-Extras install from gef/scripts"

.github/workflows/validate.yml

@@ -2,15 +2,18 @@ name: Validation
 
 on:
   pull_request:
-
+    branches:
+      - main
 
 jobs:
   pre_commit:
     name: Check formatting
     runs-on: ubuntu-latest
     steps:
     - uses: actions/checkout@v3
-    - uses: actions/setup-python@v3
+    - uses: actions/setup-python@v5.0.0
+      with:
+        python-version: "3.8"
     - uses: pre-commit/action@v3.0.0
 
   docs_link_check:
@@ -20,9 +23,9 @@ jobs:
       contents: read
     steps:
       - name: checkout
-        uses: actions/checkout@v2
+        uses: actions/checkout@v4
       - name: Check links
-        uses: lycheeverse/lychee-action@v1.4.1
+        uses: lycheeverse/lychee-action@v1.9.1
         env:
           GITHUB_TOKEN: ${{secrets.GITHUB_TOKEN}}
         with:

scripts/__init__.pyi

@@ -814,7 +814,7 @@ def register_external_context_pane(pane_name: str, display_pane_function: Callab
     def pane_title() -> str:  ...
 
 
-def register(cls: Type["GenericCommand"]) -> Type["GenericCommand"]: ...
+def register(cls: Union[Type["GenericCommand"], Type["GenericFunction"]]) -> Union[Type["GenericCommand"], Type["GenericFunction"]]: ...
 
 
 class GenericCommandBase:

scripts/assemble.py

@@ -1,4 +1,3 @@
-
 __AUTHOR__ = "hugsy"
 __VERSION__ = 0.2
 __LICENSE__ = "MIT"

scripts/kernel/__init__.py

@@ -0,0 +1,57 @@
+"""
+Collection of functions and commands to manipulate kernel symbols
+"""
+
+__AUTHOR__ = "hugsy"
+__VERSION__ = 0.1
+__LICENSE__ = "MIT"
+
+import argparse
+from typing import TYPE_CHECKING, Any, List
+
+if TYPE_CHECKING:
+    from .. import *  # this will allow linting for GEF and GDB
+
+
+@register
+class SolveKernelSymbolCommand(GenericCommand):
+    """Solve kernel symbols from kallsyms table."""
+
+    _cmdline_ = "ksymaddr"
+    _syntax_ = f"{_cmdline_} SymbolToSearch"
+    _example_ = f"{_cmdline_} prepare_creds"
+
+    @parse_arguments({"symbol": ""}, {})
+    def do_invoke(self, _: List[str], **kwargs: Any) -> None:
+        def hex_to_int(num):
+            try:
+                return int(num, 16)
+            except ValueError:
+                return 0
+
+        args: argparse.Namespace = kwargs["arguments"]
+        if not args.symbol:
+            self.usage()
+            return
+        sym = args.symbol
+        with open("/proc/kallsyms", "r") as f:
+            syms = [line.strip().split(" ", 2) for line in f]
+        matches = [
+            (hex_to_int(addr), sym_t, " ".join(name.split()))
+            for addr, sym_t, name in syms
+            if sym in name
+        ]
+        for addr, sym_t, name in matches:
+            if sym == name.split()[0]:
+                ok(f"Found matching symbol for '{name}' at {addr:#x} (type={sym_t})")
+            else:
+                warn(
+                    f"Found partial match for '{sym}' at {addr:#x} (type={sym_t}): {name}"
+                )
+        if not matches:
+            err(f"No match for '{sym}'")
+        elif matches[0][0] == 0:
+            err(
+                "Check that you have the correct permissions to view kernel symbol addresses"
+            )
+        return

scripts/kernel/symbols.py

@@ -31,7 +31,6 @@ class GlibcFunctionArguments:
     @staticmethod
     def load_libc_args() -> bool:
         """Load the LIBC function arguments. Returns `True` on success, `False` or an Exception otherwise."""
-        global gef
 
         # load libc function arguments' definitions
         path = pathlib.Path(

scripts/libc_function_args/__init__.py

@@ -0,0 +1,88 @@
+import os
+import pathlib
+import random
+import subprocess
+import tempfile
+import time
+import unittest
+
+import rpyc
+
+from .utils import GEF_EXTRAS_SCRIPTS_PATH, debug_target
+
+COVERAGE_DIR = os.getenv("COVERAGE_DIR", "")
+GEF_PATH = pathlib.Path(os.getenv("GEF_PATH", "../gef/gef.py")).absolute()
+RPYC_GEF_PATH = GEF_PATH.parent / "scripts/remote_debug.py"
+RPYC_HOST = "localhost"
+RPYC_PORT = 18812
+RPYC_SPAWN_TIME = 1.0
+
+
+class RemoteGefUnitTestGeneric(unittest.TestCase):
+    """
+    The base class for GEF test cases. This will create the `rpyc` environment to programmatically interact with
+    GDB and GEF in the test.
+    """
+
+    def setUp(self) -> None:
+        self._coverage_file = None
+        if not hasattr(self, "_target"):
+            setattr(self, "_target", debug_target("default"))
+        else:
+            assert isinstance(self._target, pathlib.Path)  # type: ignore pylint: disable=E1101
+            assert self._target.exists()  # type: ignore pylint: disable=E1101
+        self._port = random.randint(1025, 65535)
+        self._commands = ""
+
+        if COVERAGE_DIR:
+            self._coverage_file = pathlib.Path(COVERAGE_DIR) / os.getenv(
+                "PYTEST_XDIST_WORKER", "gw0"
+            )
+            self._commands += f"""
+pi import coverage
+pi cov = coverage.Coverage(data_file="{self._coverage_file}", auto_data=True, branch=True)
+pi cov.start()
+"""
+
+        self._commands += f"""
+source {GEF_PATH}
+gef config gef.debug True
+gef config gef.propagate_debug_exception True
+gef config gef.disable_color True
+
+gef config gef.extra_plugins_dir {GEF_EXTRAS_SCRIPTS_PATH}
+
+source {RPYC_GEF_PATH}
+pi start_rpyc_service({self._port})
+"""
+
+        self._initfile = tempfile.NamedTemporaryFile(mode="w", delete=False)
+        self._initfile.write(self._commands)
+        self._initfile.flush()
+        self._command = [
+            "gdb",
+            "-q",
+            "-nx",
+            "-ex",
+            f"source {self._initfile.name}",
+            "--",
+            str(self._target.absolute()),  # type: ignore pylint: disable=E1101
+        ]
+        self._process = subprocess.Popen(self._command)
+        assert self._process.pid > 0
+        time.sleep(RPYC_SPAWN_TIME)
+        self._conn = rpyc.connect(
+            RPYC_HOST,
+            self._port,
+        )
+        self._gdb = self._conn.root.gdb
+        self._gef = self._conn.root.gef
+        return super().setUp()
+
+    def tearDown(self) -> None:
+        if COVERAGE_DIR:
+            self._gdb.execute("pi cov.stop()")
+            self._gdb.execute("pi cov.save()")
+        self._conn.close()
+        self._process.terminate()
+        return super().tearDown()

tests/base.py

@@ -3,39 +3,55 @@
 """
 
 import pytest
+from tests.base import RemoteGefUnitTestGeneric
 
-from tests.utils import (ARCH, GefUnitTestGeneric, gdb_run_cmd,
-                         gdb_start_silent_cmd, removeuntil)
+from tests.utils import (
+    ARCH,
+    ERROR_INACTIVE_SESSION_MESSAGE,
+    removeuntil,
+)
 
 
-@pytest.mark.skipif(ARCH in ("mips64el", "ppc64le", "riscv64"), reason=f"Skipped for {ARCH}")
-class CapstoneDisassembleCommand(GefUnitTestGeneric):
+@pytest.mark.skipif(
+    ARCH in ("mips64el", "ppc64le", "riscv64"), reason=f"Skipped for {ARCH}"
+)
+class CapstoneDisassembleCommand(RemoteGefUnitTestGeneric):
     """`capstone-disassemble` command test module"""
 
     def setUp(self) -> None:
         try:
             import capstone  # pylint: disable=W0611
         except ImportError:
-            pytest.skip("capstone-engine not available",
-                        allow_module_level=True)
+            pytest.skip("capstone-engine not available", allow_module_level=True)
         return super().setUp()
 
     def test_cmd_capstone_disassemble(self):
-        self.assertFailIfInactiveSession(gdb_run_cmd("capstone-disassemble"))
-        res = gdb_start_silent_cmd("capstone-disassemble")
-        self.assertNoException(res)
-        self.assertTrue(len(res.splitlines()) > 1)
-
-        self.assertFailIfInactiveSession(
-            gdb_run_cmd("capstone-disassemble --show-opcodes"))
-        res = gdb_start_silent_cmd(
-            "capstone-disassemble --show-opcodes --length 5 $pc")
-        self.assertNoException(res)
-        self.assertTrue(len(res.splitlines()) >= 5)
+        gdb = self._gdb
+        cmd = "capstone-disassemble"
+
+        self.assertEqual(
+            ERROR_INACTIVE_SESSION_MESSAGE, gdb.execute(cmd, to_string=True)
+        )
+
+        gdb.execute("start")
+        res = gdb.execute("capstone-disassemble", to_string=True) or ""
+        assert res
+
+        cmd = "capstone-disassemble --show-opcodes"
+        res = gdb.execute(cmd, to_string=True) or ""
+        assert res
+
+        cmd = "capstone-disassemble --show-opcodes --length 5 $pc"
+        res = gdb.execute(cmd, to_string=True) or ""
+        assert res
+
+        lines = res.splitlines()
+        self.assertGreaterEqual(len(lines), 5)
+
         # jump to the output buffer
         res = removeuntil("→  ", res, included=True)
-        addr, opcode, symbol, *_ = [x.strip()
-                                    for x in res.splitlines()[2].strip().split()]
+        addr, opcode, symbol, *_ = [x.strip() for x in lines[2].strip().split()]
+
         # match the correct output format: <addr> <opcode> [<symbol>] mnemonic [operands,]
         # gef➤  cs --show-opcodes --length 5 $pc
         # →    0xaaaaaaaaa840 80000090    <main+20>        adrp   x0, #0xaaaaaaaba000
@@ -49,6 +65,7 @@ def test_cmd_capstone_disassemble(self):
         self.assertTrue(int(opcode, 16))
         self.assertTrue(symbol.startswith("<") and symbol.endswith(">"))
 
-        res = gdb_start_silent_cmd("cs --show-opcodes main")
-        self.assertNoException(res)
-        self.assertTrue(len(res.splitlines()) > 1)
+        cmd = "cs --show-opcodes main"
+        res = gdb.execute(cmd, to_string=True) or ""
+        assert res
+        self.assertGreater(len(res.splitlines()), 1)

tests/commands/capstone_disassemble.py

@@ -0,0 +1,17 @@
+"""
+`ksymaddr` command test module
+"""
+
+
+from tests.base import RemoteGefUnitTestGeneric
+
+
+class KsymaddrCommand(RemoteGefUnitTestGeneric):
+    """`ksymaddr` command test module"""
+
+    cmd = "ksymaddr"
+
+    def test_cmd_ksymaddr(self):
+        gdb = self._gdb
+        res = gdb.execute(f"{self.cmd} prepare_kernel_cred", to_string=True)
+        self.assertIn("Found matching symbol for 'prepare_kernel_cred'", res)