Add script to validate upstream references in PR branch commits

This script scans each commit in a PR branch for upstream Linux kernel
commit references and validates those references. If a commit references
an upstream commit, the script checks that the referenced commit exists
in mainline and reports if it has been marked with a Fixes: tag in the
upstream kernel.

Usage:
python3 check_kernel_commits.py <repo_path> <pr_branch> <base_branch> [--pretty]

By default, the script outputs results for terminal display. Use the
--pretty flag to format output for GitHub PR comments.

Author: bmastbergen

check_kernel_commits.py

@@ -0,0 +1,161 @@
+#!/usr/bin/env python3
+
+import argparse
+import subprocess
+import re
+import sys
+import textwrap
+
+def run_git(repo, args):
+    """Run a git command in the given repository and return its output as a string."""
+    result = subprocess.run(['git', '-C', repo] + args, text=True, capture_output=True, check=False)
+    if result.returncode != 0:
+        raise RuntimeError(f"Git command failed: {' '.join(args)}\n{result.stderr}")
+    return result.stdout
+
+def ref_exists(repo, ref):
+    """Return True if the given ref exists in the repository, False otherwise."""
+    result = subprocess.run(
+        ['git', '-C', repo, 'rev-parse', '--verify', '--quiet', ref],
+        stdout=subprocess.DEVNULL,
+        stderr=subprocess.DEVNULL
+    )
+    return result.returncode == 0
+
+def get_pr_commits(repo, pr_branch, base_branch):
+    """Get a list of commit SHAs that are in the PR branch but not in the base branch."""
+    output = run_git(repo, ['rev-list', f'{base_branch}..{pr_branch}'])
+    return output.strip().splitlines()
+
+def get_commit_message(repo, sha):
+    """Get the commit message for a given commit SHA."""
+    return run_git(repo, ['log', '-n', '1', '--format=%B', sha])
+
+def get_short_hash_and_subject(repo, sha):
+    """Get the abbreviated commit hash and subject for a given commit SHA."""
+    output = run_git(repo, ['log', '-n', '1', '--format=%h%x00%s', sha]).strip()
+    short_hash, subject = output.split('\x00', 1)
+    return short_hash, subject
+
+def hash_exists_in_mainline(repo, upstream_ref, hash_):
+    """Check if a commit hash exists in the upstream reference."""
+    result = subprocess.run(
+        ['git', '-C', repo, 'cat-file', '-e', f'{upstream_ref}^{hash_}'],
+        stdout=subprocess.DEVNULL,
+        stderr=subprocess.DEVNULL
+    )
+    return result.returncode == 0
+
+def find_fixes_in_mainline(repo, upstream_ref, hash_):
+    """Return commits in upstream_ref that have Fixes: <first 12 chars of hash_> in their message."""
+    short_hash = hash_[:12]
+    output = run_git(repo, [
+        'log', upstream_ref, '--grep', f'Fixes: {short_hash}', '--format=%h %s (%an)'
+    ])
+    return output.strip()
+
+def wrap_paragraph(text, width=80, initial_indent='', subsequent_indent=''):
+    """Wrap a paragraph of text to the specified width and indentation."""
+    wrapper = textwrap.TextWrapper(width=width,
+                                   initial_indent=initial_indent,
+                                   subsequent_indent=subsequent_indent,
+                                   break_long_words=False,
+                                   break_on_hyphens=False)
+    return wrapper.fill(text)
+
+def main():
+    parser = argparse.ArgumentParser(description="Check upstream references and Fixes: tags in PR branch commits.")
+    parser.add_argument("repo", help="Path to the git repo")
+    parser.add_argument("pr_branch", help="Name of the PR branch")
+    parser.add_argument("base_branch", help="Name of the base branch")
+    parser.add_argument("--pretty", action='store_true', help="Output in Markdown, suitable for GitHub PR comments")
+    args = parser.parse_args()
+
+    upstream_ref = 'origin/kernel-mainline'
+
+    # Validate that all required refs exist before continuing
+    missing_refs = []
+    for refname, refval in [('upstream reference', upstream_ref),
+                            ('PR branch', args.pr_branch),
+                            ('base branch', args.base_branch)]:
+        if not ref_exists(args.repo, refval):
+            missing_refs.append((refname, refval))
+    if missing_refs:
+        for refname, refval in missing_refs:
+            print(f"ERROR: The {refname} '{refval}' does not exist in the given repo.")
+        print("Please fetch or create the required references before running this script.")
+        sys.exit(1)
+
+    pr_commits = get_pr_commits(args.repo, args.pr_branch, args.base_branch)
+    if not pr_commits:
+        if args.pretty:
+            print("> ℹ️ **No commits found in PR branch that are not in base branch.**")
+        else:
+            print("No commits found in PR branch that are not in base branch.")
+        sys.exit(0)
+
+    any_findings = False
+    out_lines = []
+
+    for sha in reversed(pr_commits):  # oldest first
+        short_hash, subject = get_short_hash_and_subject(args.repo, sha)
+        pr_commit_desc = f"{short_hash} ({subject})"
+        msg = get_commit_message(args.repo, sha)
+        upstream_hashes = re.findall(r'^commit\s+([0-9a-fA-F]{12,40})', msg, re.MULTILINE)
+        for uhash in upstream_hashes:
+            short_uhash = uhash[:12]
+            exists = hash_exists_in_mainline(args.repo, upstream_ref, uhash)
+            if not exists:
+                any_findings = True
+                if args.pretty:
+                    out_lines.append(
+                        f"- ❗ PR commit `{pr_commit_desc}` references upstream commit  \n"
+                        f"  `{short_uhash}` which does **not** exist in the upstream Linux kernel.\n"
+                    )
+                else:
+                    header = (f"[NOTFOUND] PR commit {pr_commit_desc} references upstream commit "
+                              f"{short_uhash}, which does not exist in kernel-mainline.")
+                    out_lines.append(
+                        wrap_paragraph(header, width=80, initial_indent='',
+                                       subsequent_indent='           ')  # 11 spaces for '[NOTFOUND] '
+                    )
+                    out_lines.append("")  # blank line
+                continue
+            fixes = find_fixes_in_mainline(args.repo, upstream_ref, uhash)
+            if fixes:
+                any_findings = True
+                if args.pretty:
+                    fixes_block = "    " + fixes.replace("\n", "\n    ")
+                    out_lines.append(
+                        f"- ⚠️ PR commit `{pr_commit_desc}` references upstream commit  \n"
+                        f"  `{short_uhash}` which has been referenced by a `Fixes:` tag in the upstream  \n"
+                        f"  Linux kernel:\n\n"
+                        f"```text\n{fixes_block}\n```\n"
+                    )
+                else:
+                    header = (f"[FIXES] PR commit {pr_commit_desc} references upstream commit "
+                              f"{short_uhash}, which has Fixes tags:")
+                    out_lines.append(
+                        wrap_paragraph(header, width=80, initial_indent='',
+                                       subsequent_indent='        ')  # 8 spaces for '[FIXES] '
+                    )
+                    out_lines.append("")  # blank line after 'Fixes tags:'
+                    for line in fixes.splitlines():
+                        out_lines.append('    ' + line)
+                    out_lines.append("")  # blank line
+
+    if any_findings:
+        if args.pretty:
+            print("## :mag: Upstream Linux Kernel Commit Check\n")
+            print('\n'.join(out_lines))
+            print("*This is an automated message from the kernel commit checker workflow.*")
+        else:
+            print('\n'.join(out_lines))
+    else:
+        if args.pretty:
+            print("> ✅ **All referenced commits exist upstream and have no Fixes: tags.**")
+        else:
+            print("All referenced commits exist upstream and have no Fixes: tags.")
+
+if __name__ == "__main__":
+    main()