Improve kselftest comparison with smart base branch detection

- Add checkout step with full history (fetch-depth: 0) to enable git operations
- Implement intelligent base branch detection:
  * For PRs: use the base branch directly from github.base_ref
  * For direct pushes: find closest common ancestor using git merge-base
  * Iterate through all remote branches to find best match
- Compare against detected base branch instead of same branch
- Add outputs (comparison_status, comparison_message) for downstream jobs
- Improve error handling and warning messages
- Skip comparison gracefully if base branch cannot be determined

This makes the comparison much more useful by comparing against the
branch you're merging into, rather than the previous run of the same
branch.

Signed-off-by: Shreeya Patel <spatel@ciq.com>

Author: shreeya-patel98

.github/workflows/kernel-build-and-test-x86_64.yml

@@ -199,39 +199,117 @@ jobs:
     if: success() || failure()
 
     steps:
+    - name: Checkout kernel source
+      uses: actions/checkout@v4
+      with:
+        fetch-depth: 0
+
     - name: Download current kselftest logs
       uses: actions/download-artifact@v4
       with:
         name: kselftest-logs-x86_64
         path: output-current
 
-    - name: Download previous kselftest logs
+    - name: Determine base branch for comparison
+      id: base_branch
+      run: |
+        BASE_BRANCH=""
+
+        # For PRs, use the base branch directly
+        if [ -n "${{ github.base_ref }}" ]; then
+          BASE_BRANCH="${{ github.base_ref }}"
+          echo "Using PR base branch: $BASE_BRANCH"
+        else
+          # For direct pushes, find the common ancestor branch
+          echo "Not a PR, finding base branch via merge-base"
+
+          # Get all remote branches and find the one with closest common ancestor
+          CURRENT_COMMIT=$(git rev-parse HEAD)
+          BEST_BRANCH=""
+          CLOSEST_DISTANCE=999999
+
+          for remote_branch in $(git for-each-ref --format='%(refname:short)' refs/remotes/origin/ | grep -v 'HEAD\|PR-CHECKER'); do
+            branch_name=$(echo "$remote_branch" | sed 's|^origin/||')
+
+            # Skip if it's the current branch
+            if [ "$branch_name" = "${{ github.ref_name }}" ]; then
+              continue
+            fi
+
+            MERGE_BASE=$(git merge-base HEAD "$remote_branch" 2>/dev/null || echo "")
+
+            # Check if this branch shares history and isn't the current commit
+            if [ -n "$MERGE_BASE" ] && [ "$MERGE_BASE" != "$CURRENT_COMMIT" ]; then
+              # Calculate distance (number of commits) from merge-base to current HEAD
+              DISTANCE=$(git rev-list --count ${MERGE_BASE}..HEAD 2>/dev/null || echo "999999")
+
+              # Find the branch with the shortest distance (most recent common ancestor)
+              if [ "$DISTANCE" -lt "$CLOSEST_DISTANCE" ]; then
+                CLOSEST_DISTANCE=$DISTANCE
+                BEST_BRANCH=$branch_name
+              fi
+            fi
+          done
+
+          if [ -n "$BEST_BRANCH" ]; then
+            BASE_BRANCH=$BEST_BRANCH
+            echo "Found common ancestor with $BASE_BRANCH (distance: $CLOSEST_DISTANCE commits)"
+          fi
+        fi
+
+        if [ -z "$BASE_BRANCH" ]; then
+          echo "::warning::Could not determine base branch for comparison - no common ancestor found"
+          echo "::warning::Kselftest comparison will be skipped"
+          echo "base_branch=" >> $GITHUB_OUTPUT
+          exit 0
+        fi
+
+        echo "base_branch=$BASE_BRANCH" >> $GITHUB_OUTPUT
+        echo "Base branch for comparison: $BASE_BRANCH"
+
+    - name: Download baseline kselftest logs from base branch
+      if: steps.base_branch.outputs.base_branch != ''
       uses: dawidd6/action-download-artifact@v3
       with:
         workflow: kernel-build-and-test-x86_64.yml
         name: kselftest-logs-x86_64
         path: output-previous
-        branch: ${{ github.ref_name }}
+        branch: ${{ steps.base_branch.outputs.base_branch }}
         workflow_conclusion: success
         search_artifacts: true
         skip_unpack: false
         if_no_artifact_found: warn
+        # Only search the last 5 successful runs for better performance
+        run_number: ${{ github.run_number }}
+        search_depth: 5
       continue-on-error: true
+      timeout-minutes: 3
 
     - name: Compare test results
+      id: comparison
       run: |
-        if [ -f output-previous/kselftests-*.log ]; then
+        # Check if we have a base branch to compare against
+        if [ -z "${{ steps.base_branch.outputs.base_branch }}" ]; then
+          echo "::warning::No base branch found for comparison"
+          echo "::warning::Kselftest comparison will be skipped"
+          echo "comparison_status=skipped" >> $GITHUB_OUTPUT
+          echo "comparison_message=No base branch found - unable to determine merge target" >> $GITHUB_OUTPUT
+          exit 0
+        fi
+
+        # Check if baseline logs exist
+        if ls output-previous/kselftests-*.log 1> /dev/null 2>&1; then
           # Compare passing tests (ok)
-          BEFORE_PASS=$(grep -a '^ok' output-previous/kselftests-*.log | wc -l)
-          AFTER_PASS=$(grep -a '^ok' output-current/kselftests-*.log | wc -l)
+          BEFORE_PASS=$(grep -a '^ok' output-previous/kselftests-*.log | wc -l || echo "0")
+          AFTER_PASS=$(grep -a '^ok' output-current/kselftests-*.log | wc -l || echo "0")
 
           # Compare failing tests (not ok)
-          BEFORE_FAIL=$(grep -a '^not ok' output-previous/kselftests-*.log | wc -l)
-          AFTER_FAIL=$(grep -a '^not ok' output-current/kselftests-*.log | wc -l)
+          BEFORE_FAIL=$(grep -a '^not ok' output-previous/kselftests-*.log | wc -l || echo "0")
+          AFTER_FAIL=$(grep -a '^not ok' output-current/kselftests-*.log | wc -l || echo "0")
 
-          echo "### Kselftest Comparison (Branch: ${{ github.ref_name }})"
-          echo "Passing tests: $BEFORE_PASS -> $AFTER_PASS"
-          echo "Failing tests: $BEFORE_FAIL -> $AFTER_FAIL"
+          echo "### Kselftest Comparison"
+          echo "Baseline (from ${{ steps.base_branch.outputs.base_branch }}): $BEFORE_PASS passing, $BEFORE_FAIL failing"
+          echo "Current (${{ github.ref_name }}): $AFTER_PASS passing, $AFTER_FAIL failing"
 
           # Calculate differences
           PASS_DIFF=$((AFTER_PASS - BEFORE_PASS))
@@ -255,10 +333,18 @@ jobs:
 
           if [ $REGRESSION -eq 1 ]; then
             echo "::error::Test regression exceeds acceptable threshold of 3 tests"
+            echo "comparison_status=failed" >> $GITHUB_OUTPUT
+            echo "comparison_message=Regression detected: Pass diff: $PASS_DIFF, Fail diff: $FAIL_DIFF (threshold: ±3)" >> $GITHUB_OUTPUT
             exit 1
           else
             echo "::notice::Test results within acceptable range (threshold: ±3 tests)"
+            echo "comparison_status=passed" >> $GITHUB_OUTPUT
+            echo "comparison_message=Baseline: $BEFORE_PASS passing, $BEFORE_FAIL failing | Current: $AFTER_PASS passing, $AFTER_FAIL failing" >> $GITHUB_OUTPUT
           fi
         else
-          echo "::warning::No previous successful test results found for branch ${{ github.ref_name }}, skipping comparison"
+          echo "::warning::No baseline test results found for branch ${{ steps.base_branch.outputs.base_branch }}"
+          echo "::notice::Cannot compare against base branch - artifacts may not exist or have expired (7-day retention)"
+          echo "::notice::Skipping comparison - PR will still be created with warning"
+          echo "comparison_status=skipped" >> $GITHUB_OUTPUT
+          echo "comparison_message=No baseline results available from ${{ steps.base_branch.outputs.base_branch }}" >> $GITHUB_OUTPUT
         fi