pkg/vcs: make git fetch calls more specific

This should make syzkaller only fetch the commits relevant for further
processing. Also, specifying the exact commit/branch name to fetch
allows us to access commits from custom refs.

Test the new behaviour and double-check that remote tags fetch was not
broken.

3 files changed, 63 insertions, 4 deletions

diff --git a/pkg/vcs/git.go b/pkg/vcs/git.go
index 71fcac97a..0bd72d9a6 100644
--- a/pkg/vcs/git.go
+++ b/pkg/vcs/git.go
@@ -134,17 +134,21 @@ func (git *git) CheckoutCommit(repo, commit string) (*Commit, error) {
 	if err := git.repair(); err != nil {
 		return nil, err
 	}
-	if err := git.fetchRemote(repo); err != nil {
+	if err := git.fetchRemote(repo, commit); err != nil {
 		return nil, err
 	}
 	return git.SwitchCommit(commit)
 }
 
-func (git *git) fetchRemote(repo string) error {
+func (git *git) fetchRemote(repo, commit string) error {
 	repoHash := hash.String([]byte(repo))
 	// Ignore error as we can double add the same remote and that will fail.
 	git.git("remote", "add", repoHash, repo)
-	_, err := git.git("fetch", "--force", "--tags", repoHash)
+	fetchArgs := []string{"fetch", "--force", "--tags", repoHash}
+	if commit != "" {
+		fetchArgs = append(fetchArgs, commit)
+	}
+	_, err := git.git(fetchArgs...)
 	if err != nil {
 		var verbose *osutil.VerboseError
 		if errors.As(err, &verbose) &&
diff --git a/pkg/vcs/git_test.go b/pkg/vcs/git_test.go
index 264cb54d8..5e9d30fa7 100644
--- a/pkg/vcs/git_test.go
+++ b/pkg/vcs/git_test.go
@@ -6,10 +6,12 @@ package vcs
 import (
 	"os"
 	"reflect"
+	"sort"
 	"testing"
 	"time"
 
 	"github.com/google/go-cmp/cmp"
+	"github.com/stretchr/testify/assert"
 )
 
 func TestGitParseCommit(t *testing.T) {
@@ -371,3 +373,56 @@ func TestMergeBase(t *testing.T) {
 		t.Fatalf("expected base commit, got %v", mergeCommits)
 	}
 }
+
+func TestGitCustomRefs(t *testing.T) {
+	remoteRepoDir := t.TempDir()
+	remote := MakeTestRepo(t, remoteRepoDir)
+	remote.Git("commit", "--no-edit", "--allow-empty", "-m", "base commit")
+	remote.Git("checkout", "-b", "base_branch")
+	remote.Git("tag", "base_tag")
+
+	// Create a commit non reachable from any branch or tag.
+	remote.Git("checkout", "base_branch")
+	remote.Git("checkout", "-b", "temp_branch")
+	remote.Git("commit", "--no-edit", "--allow-empty", "-m", "detached commit")
+	// Add a ref to prevent the commit from getting garbage collected.
+	remote.Git("update-ref", "refs/custom/test", "temp_branch")
+	refCommit, _ := remote.repo.HeadCommit()
+
+	// Remove the branch, let the commit stay only in refs.
+	remote.Git("checkout", "base_branch")
+	remote.Git("branch", "-D", "temp_branch")
+
+	// Create a local repo.
+	localRepoDir := t.TempDir()
+	local := newGit(localRepoDir, nil, nil)
+
+	// Fetch the commit from the custom ref.
+	_, err := local.CheckoutCommit(remoteRepoDir, refCommit.Hash)
+	assert.NoError(t, err)
+}
+
+func TestGitRemoteTags(t *testing.T) {
+	remoteRepoDir := t.TempDir()
+	remote := MakeTestRepo(t, remoteRepoDir)
+	remote.Git("commit", "--no-edit", "--allow-empty", "-m", "base commit")
+	remote.Git("checkout", "-b", "base_branch")
+	remote.Git("tag", "v1.0")
+
+	// Diverge sub_branch and add a tag.
+	remote.Git("commit", "--no-edit", "--allow-empty", "-m", "sub-branch")
+	remote.Git("checkout", "-b", "sub_branch")
+	remote.Git("tag", "v2.0")
+
+	// Create a local repo.
+	localRepoDir := t.TempDir()
+	local := newGit(localRepoDir, nil, nil)
+
+	// Ensure all tags were fetched.
+	commit, err := local.CheckoutCommit(remoteRepoDir, "sub_branch")
+	assert.NoError(t, err)
+	tags, err := local.previousReleaseTags(commit.Hash, true, false, false)
+	assert.NoError(t, err)
+	sort.Strings(tags)
+	assert.Equal(t, []string{"v1.0", "v2.0"}, tags)
+}
diff --git a/pkg/vcs/linux.go b/pkg/vcs/linux.go
index 62319096d..98ae07cda 100644
--- a/pkg/vcs/linux.go
+++ b/pkg/vcs/linux.go
@@ -195,7 +195,7 @@ func (ctx *linux) PrepareBisect() error {
 	if ctx.vmType != "gvisor" {
 		// Some linux repos we fuzz don't import the upstream release git tags. We need tags
 		// to decide which compiler versions to use. Let's fetch upstream for its tags.
-		err := ctx.git.fetchRemote("https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git")
+		err := ctx.git.fetchRemote("https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git", "")
 		if err != nil {
 			return fmt.Errorf("fetching upstream linux failed: %w", err)
 		}