From 09099cc9a309050d1fe73663df141e25eb2b51d6 Mon Sep 17 00:00:00 2001
From: Tim Ledbetter <timledbetter@gmail.com>
Date: Tue, 26 Sep 2023 18:28:15 +0100
Subject: [PATCH] du: Prevent multiple counting of the same file

Previously, the same file would be counted more than once if its
containing directory was visited multiple times, or there were
multiple hard links pointing to it.

We now keep track of every visited inode to ensure that files aren't
evaluated multiple times. This matches the behavior of `du` on FreeBSD
and Linux.
---
 Userland/Utilities/du.cpp | 27 +++++++++++++++++++++++++++
 1 file changed, 27 insertions(+)

diff --git a/Userland/Utilities/du.cpp b/Userland/Utilities/du.cpp
index 829e98a3e4..737ac83dff 100644
--- a/Userland/Utilities/du.cpp
+++ b/Userland/Utilities/du.cpp
@@ -36,6 +36,26 @@ struct DuOption {
     size_t max_depth = SIZE_MAX;
 };
 
+struct VisitedFile {
+    dev_t device;
+    ino_t inode;
+};
+
+template<>
+struct AK::Traits<VisitedFile> : public GenericTraits<VisitedFile> {
+    static unsigned hash(VisitedFile const& visited_file)
+    {
+        return pair_int_hash(u64_hash(visited_file.device), u64_hash(visited_file.inode));
+    }
+
+    static bool equals(VisitedFile const& a, VisitedFile const& b)
+    {
+        return a.device == b.device && a.inode == b.inode;
+    }
+};
+
+static HashTable<VisitedFile> s_visited_files;
+
 static ErrorOr<void> parse_args(Main::Arguments arguments, Vector<DeprecatedString>& files, DuOption& du_option);
 static u64 print_space_usage(DeprecatedString const& path, DuOption const& du_option, size_t current_depth, bool inside_dir = false);
 
@@ -144,6 +164,13 @@ u64 print_space_usage(DeprecatedString const& path, DuOption const& du_option, s
     }
 
     auto path_stat = path_stat_or_error.release_value();
+
+    VisitedFile visited_file { path_stat.st_dev, path_stat.st_ino };
+    if (s_visited_files.contains(visited_file)) {
+        return 0;
+    }
+    s_visited_files.set(visited_file);
+
     bool const is_directory = S_ISDIR(path_stat.st_mode);
     if (is_directory) {
         auto di = Core::DirIterator(path, Core::DirIterator::SkipParentAndBaseDir);