fix(search): bound glob/grep tree-walk with cooperative + watchdog timeout

crates/loopal-backend/src/limits.rs

@@ -22,6 +22,11 @@ pub struct ResourceLimits {
     pub max_fetch_bytes: usize,
     /// Default shell command timeout.
     pub default_timeout: Duration,
+    /// Cooperative deadline checked between glob/grep walk entries: bounds
+    /// slow-but-responsive trees and returns partial results. A syscall stuck
+    /// on a dead mount can't be interrupted here — that case is bounded by the
+    /// runtime per-tool watchdog instead.
+    pub walk_timeout: Duration,
     /// HTTP fetch timeout.
     pub fetch_timeout: Duration,
     /// Maximum image file size in bytes.
@@ -40,6 +45,7 @@ impl Default for ResourceLimits {
             max_grep_matches: 500,
             max_fetch_bytes: 5 * 1024 * 1024,          // 5 MB
             default_timeout: Duration::from_secs(300), // 5 min
+            walk_timeout: Duration::from_secs(30),
             fetch_timeout: Duration::from_secs(30),
             image_max_bytes: IMAGE_MAX_BYTES,
             image_max_pixels: IMAGE_MAX_PIXELS,

crates/loopal-backend/src/search/glob.rs

@@ -1,17 +1,18 @@
-//! Glob pattern search with file-type filtering and modification time.
-
 use std::path::Path;
-use std::time::UNIX_EPOCH;
+use std::sync::Arc;
+use std::sync::atomic::{AtomicBool, Ordering};
+use std::time::{Instant, UNIX_EPOCH};
 
 use globset::Glob;
+use ignore::WalkState;
 use loopal_error::ToolIoError;
 use loopal_tool_api::backend_types::{GlobEntry, GlobOptions, GlobSearchResult};
 use loopal_tool_api::save_to_overflow_file;
+use parking_lot::Mutex;
 
 use crate::limits::ResourceLimits;
 use crate::search::{overflow_fmt, walker};
 
-/// Execute a glob search and return matching entries.
 pub fn glob_search(
     opts: &GlobOptions,
     cwd: &Path,
@@ -25,50 +26,75 @@ pub fn glob_search(
 
     let glob =
         Glob::new(&opts.pattern).map_err(|e| ToolIoError::Other(format!("invalid glob: {e}")))?;
-    let matcher = glob.compile_matcher();
+    let max = opts.max_results.min(limits.max_glob_results).max(1);
 
-    let max = opts.max_results.min(limits.max_glob_results);
-    let Some(walker) = walker::build_walker(&search_path, opts.type_filter.as_deref()) else {
+    let Some(w) = walker::build_walker(&search_path, opts.type_filter.as_deref()) else {
         return Ok(GlobSearchResult {
             entries: Vec::new(),
             truncated: false,
+            timed_out: false,
             overflow_path: None,
         });
     };
 
-    let mut entries = Vec::new();
-    let mut truncated = false;
-
-    for entry in walker.build().flatten() {
-        if !entry.file_type().is_some_and(|ft| ft.is_file()) {
-            continue;
-        }
-        let path = entry.path();
-        let rel = match path.strip_prefix(&search_path) {
-            Ok(r) => r,
-            Err(_) => continue,
-        };
-        if !matcher.is_match(rel) {
-            continue;
-        }
-        let modified_secs = entry
-            .metadata()
-            .ok()
-            .and_then(|m| m.modified().ok())
-            .and_then(|t| t.duration_since(UNIX_EPOCH).ok())
-            .map(|d| d.as_secs());
-
-        entries.push(GlobEntry {
-            path: path.to_string_lossy().into_owned(),
-            modified_secs,
-        });
+    let deadline = Instant::now() + limits.walk_timeout;
+    let done = Arc::new(AtomicBool::new(false));
+    let timed_out = Arc::new(AtomicBool::new(false));
+    let entries: Arc<Mutex<Vec<GlobEntry>>> = Arc::new(Mutex::new(Vec::new()));
+    let search_path = Arc::new(search_path);
+    let matcher = Arc::new(glob.compile_matcher());
 
-        if entries.len() >= max {
-            truncated = true;
-            break;
-        }
-    }
+    w.build_parallel().run(|| {
+        let done = Arc::clone(&done);
+        let timed_out = Arc::clone(&timed_out);
+        let entries = Arc::clone(&entries);
+        let search_path = Arc::clone(&search_path);
+        let matcher = Arc::clone(&matcher);
+        Box::new(move |entry| {
+            if done.load(Ordering::Relaxed) {
+                return WalkState::Quit;
+            }
+            if Instant::now() >= deadline {
+                done.store(true, Ordering::Relaxed);
+                timed_out.store(true, Ordering::Relaxed);
+                return WalkState::Quit;
+            }
+            let Ok(entry) = entry else {
+                return WalkState::Continue;
+            };
+            if !entry.file_type().is_some_and(|ft| ft.is_file()) {
+                return WalkState::Continue;
+            }
+            let Ok(rel) = entry.path().strip_prefix(search_path.as_path()) else {
+                return WalkState::Continue;
+            };
+            if !matcher.is_match(rel) {
+                return WalkState::Continue;
+            }
+            let modified_secs = entry
+                .metadata()
+                .ok()
+                .and_then(|m| m.modified().ok())
+                .and_then(|t| t.duration_since(UNIX_EPOCH).ok())
+                .map(|d| d.as_secs());
+            let n = {
+                let mut guard = entries.lock();
+                guard.push(GlobEntry {
+                    path: entry.path().to_string_lossy().into_owned(),
+                    modified_secs,
+                });
+                guard.len()
+            };
+            if n >= max {
+                done.store(true, Ordering::Relaxed);
+                return WalkState::Quit;
+            }
+            WalkState::Continue
+        })
+    });
 
+    let entries = Arc::try_unwrap(entries).unwrap().into_inner();
+    let truncated = entries.len() >= max;
     let overflow_path = if truncated {
         Some(save_to_overflow_file(
             &overflow_fmt::serialize_glob_results(&entries),
@@ -81,6 +107,7 @@ pub fn glob_search(
     Ok(GlobSearchResult {
         entries,
         truncated,
+        timed_out: timed_out.load(Ordering::Relaxed),
         overflow_path,
     })
 }

crates/loopal-backend/src/search/grep.rs

@@ -1,6 +1,7 @@
 use std::path::Path;
 use std::sync::Arc;
 use std::sync::atomic::{AtomicBool, AtomicUsize, Ordering};
+use std::time::Instant;
 
 use globset::Glob;
 use ignore::WalkState;
@@ -55,12 +56,14 @@ pub fn grep_search(
         None => None,
     };
 
-    let max = opts.max_matches.min(limits.max_grep_matches);
+    let max = opts.max_matches.min(limits.max_grep_matches).max(1);
     let ctx_before = opts.context_before;
     let ctx_after = opts.context_after;
     let multiline = opts.multiline;
     let total = Arc::new(AtomicUsize::new(0));
     let done = Arc::new(AtomicBool::new(false));
+    let timed_out = Arc::new(AtomicBool::new(false));
+    let deadline = Instant::now() + limits.walk_timeout;
     let results: Arc<Mutex<Vec<FileMatchResult>>> = Arc::new(Mutex::new(Vec::new()));
     let search_path = Arc::new(search_path);
     let glob_matcher = Arc::new(glob_matcher);
@@ -74,11 +77,17 @@ pub fn grep_search(
         let search_path = Arc::clone(&search_path);
         let total = Arc::clone(&total);
         let done = Arc::clone(&done);
+        let timed_out = Arc::clone(&timed_out);
         let results = Arc::clone(&results);
         Box::new(move |entry| {
             if done.load(Ordering::Relaxed) {
                 return WalkState::Quit;
             }
+            if Instant::now() >= deadline {
+                done.store(true, Ordering::Relaxed);
+                timed_out.store(true, Ordering::Relaxed);
+                return WalkState::Quit;
+            }
             let entry = match entry {
                 Ok(e) => e,
                 Err(_) => return WalkState::Continue,
@@ -100,9 +109,10 @@ pub fn grep_search(
     });
 
     let file_matches = Arc::try_unwrap(results).unwrap().into_inner();
-    let truncated = done.load(Ordering::Relaxed);
+    let truncated = total.load(Ordering::Relaxed) >= max;
     Ok(GrepSearchResult {
         total_match_count: total.load(Ordering::Relaxed),
+        timed_out: timed_out.load(Ordering::Relaxed),
         overflow_path: maybe_save_overflow(truncated, &file_matches),
         file_matches,
     })

crates/loopal-backend/src/search/grep_file.rs

@@ -43,6 +43,7 @@ pub(crate) fn empty_result() -> GrepSearchResult {
     GrepSearchResult {
         file_matches: Vec::new(),
         total_match_count: 0,
+        timed_out: false,
         overflow_path: None,
     }
 }
@@ -90,6 +91,7 @@ pub(crate) fn search_single_file(
     let overflow_path = maybe_save_overflow(truncated, &file_matches);
     Ok(GrepSearchResult {
         total_match_count: count,
+        timed_out: false,
         file_matches,
         overflow_path,
     })

crates/loopal-backend/src/search/walker.rs

@@ -7,15 +7,15 @@ use ignore::types::TypesBuilder;
 
 /// Build a `WalkBuilder` with shared defaults.
 ///
-/// * Follows symlinks.
+/// * Does not follow symlinks (ripgrep default) — avoids cross-mount escape and traversal cycles.
 /// * Respects `.gitignore` (ignore crate default).
 /// * Applies file-type filtering when `type_filter` is given.
 ///
 /// Returns `None` when `type_filter` names an unknown file type — the
 /// caller should short-circuit with an empty result.
 pub fn build_walker(search_path: &Path, type_filter: Option<&str>) -> Option<WalkBuilder> {
     let mut builder = WalkBuilder::new(search_path);
-    builder.follow_links(true);
+    builder.follow_links(false);
 
     if let Some(ty) = type_filter {
         let mut tb = TypesBuilder::new();

crates/loopal-backend/tests/suite.rs

@@ -5,6 +5,8 @@ mod approved_paths_test;
 mod batch_test;
 #[path = "suite/fetch_headers_test.rs"]
 mod fetch_headers_test;
+#[path = "suite/glob_parallel_test.rs"]
+mod glob_parallel_test;
 #[path = "suite/image_test.rs"]
 mod image_test;
 #[path = "suite/log_file_test.rs"]
@@ -15,5 +17,7 @@ mod path_approval_test;
 mod process_group_test;
 #[path = "suite/resolve_checked_test.rs"]
 mod resolve_checked_test;
+#[path = "suite/search_timeout_test.rs"]
+mod search_timeout_test;
 #[path = "suite/tmp_cleanup_test.rs"]
 mod tmp_cleanup_test;

crates/loopal-backend/tests/suite/glob_parallel_test.rs

@@ -0,0 +1,76 @@
+use loopal_backend::ResourceLimits;
+use loopal_backend::search::glob_search;
+use loopal_tool_api::backend_types::GlobOptions;
+
+fn glob_opts(pattern: &str) -> GlobOptions {
+    GlobOptions {
+        pattern: pattern.to_string(),
+        path: None,
+        type_filter: None,
+        max_results: 10_000,
+    }
+}
+
+#[test]
+fn parallel_glob_finds_all_nested_matches() {
+    let tmp = tempfile::tempdir().unwrap();
+    std::fs::write(tmp.path().join("a.rs"), "").unwrap();
+    let sub = tmp.path().join("sub/deep");
+    std::fs::create_dir_all(&sub).unwrap();
+    std::fs::write(tmp.path().join("sub/b.rs"), "").unwrap();
+    std::fs::write(sub.join("c.rs"), "").unwrap();
+    std::fs::write(tmp.path().join("skip.txt"), "").unwrap();
+
+    let res = glob_search(
+        &glob_opts("**/*.rs"),
+        tmp.path(),
+        &ResourceLimits::default(),
+    )
+    .unwrap();
+
+    let paths: Vec<&str> = res.entries.iter().map(|e| e.path.as_str()).collect();
+    assert_eq!(res.entries.len(), 3);
+    assert!(paths.iter().any(|p| p.ends_with("a.rs")));
+    assert!(paths.iter().any(|p| p.ends_with("b.rs")));
+    assert!(paths.iter().any(|p| p.ends_with("c.rs")));
+    assert!(!res.truncated);
+    assert!(!res.timed_out);
+}
+
+#[test]
+fn parallel_glob_truncates_at_max_results_with_tolerance() {
+    let tmp = tempfile::tempdir().unwrap();
+    for i in 0..50 {
+        std::fs::write(tmp.path().join(format!("f{i:02}.rs")), "").unwrap();
+    }
+    let limits = ResourceLimits {
+        max_glob_results: 10,
+        ..ResourceLimits::default()
+    };
+
+    let res = glob_search(&glob_opts("**/*.rs"), tmp.path(), &limits).unwrap();
+
+    assert!(res.truncated);
+    assert!(!res.timed_out);
+    assert!(res.entries.len() >= 10);
+    assert!(res.entries.len() < 50);
+}
+
+#[cfg(unix)]
+#[test]
+fn glob_does_not_follow_symlinks() {
+    let tmp = tempfile::tempdir().unwrap();
+    let root = tmp.path().join("root");
+    let outside = tmp.path().join("outside");
+    std::fs::create_dir_all(&root).unwrap();
+    std::fs::create_dir_all(&outside).unwrap();
+    std::fs::write(root.join("real.rs"), "").unwrap();
+    std::fs::write(outside.join("secret.rs"), "").unwrap();
+    std::os::unix::fs::symlink(&outside, root.join("link")).unwrap();
+
+    let res = glob_search(&glob_opts("**/*.rs"), &root, &ResourceLimits::default()).unwrap();
+
+    let paths: Vec<&str> = res.entries.iter().map(|e| e.path.as_str()).collect();
+    assert!(paths.iter().any(|p| p.ends_with("real.rs")));
+    assert!(!paths.iter().any(|p| p.ends_with("secret.rs")));
+}