Use st_nlink in hardlink tracking (#39328)

Only add potential hardlinks to a set/dict, instead of each file. This
should be much cheaper, since hardlinks are very rare.

2 files changed, 12 insertions, 11 deletions

diff --git a/lib/spack/spack/binary_distribution.py b/lib/spack/spack/binary_distribution.py
index 50043af762..fc6056e6be 100644
--- a/lib/spack/spack/binary_distribution.py
+++ b/lib/spack/spack/binary_distribution.py
@@ -647,8 +647,7 @@ class BuildManifestVisitor(BaseDirectoryVisitor):
     directories."""
 
     def __init__(self):
-        # Save unique identifiers of files to avoid
-        # relocating hardlink files for each path.
+        # Save unique identifiers of hardlinks to avoid relocating them multiple times
         self.visited = set()
 
         # Lists of files we will check
@@ -657,6 +656,8 @@ class BuildManifestVisitor(BaseDirectoryVisitor):
 
     def seen_before(self, root, rel_path):
         stat_result = os.lstat(os.path.join(root, rel_path))
+        if stat_result.st_nlink == 1:
+            return False
         identifier = (stat_result.st_dev, stat_result.st_ino)
         if identifier in self.visited:
             return True
@@ -1581,9 +1582,10 @@ def dedupe_hardlinks_if_necessary(root, buildinfo):
         for rel_path in buildinfo[key]:
             stat_result = os.lstat(os.path.join(root, rel_path))
             identifier = (stat_result.st_dev, stat_result.st_ino)
-            if identifier in visited:
-                continue
-            visited.add(identifier)
+            if stat_result.st_nlink > 1:
+                if identifier in visited:
+                    continue
+                visited.add(identifier)
             new_list.append(rel_path)
         buildinfo[key] = new_list
 
diff --git a/lib/spack/spack/hooks/drop_redundant_rpaths.py b/lib/spack/spack/hooks/drop_redundant_rpaths.py
index 4cbbf5359d..a32d435e38 100644
--- a/lib/spack/spack/hooks/drop_redundant_rpaths.py
+++ b/lib/spack/spack/hooks/drop_redundant_rpaths.py
@@ -79,8 +79,7 @@ class ElfFilesWithRPathVisitor(BaseDirectoryVisitor):
     """Visitor that collects all elf files that have an rpath"""
 
     def __init__(self):
-        # Map from (ino, dev) -> path. We need 1 path per file, if there are hardlinks,
-        # we don't need to store the path multiple times.
+        # Keep track of what hardlinked files we've already visited.
         self.visited = set()
 
     def visit_file(self, root, rel_path, depth):
@@ -89,10 +88,10 @@ class ElfFilesWithRPathVisitor(BaseDirectoryVisitor):
         identifier = (s.st_ino, s.st_dev)
 
         # We're hitting a hardlink or symlink of an excluded lib, no need to parse.
-        if identifier in self.visited:
-            return
-
-        self.visited.add(identifier)
+        if s.st_nlink > 1:
+            if identifier in self.visited:
+                return
+            self.visited.add(identifier)
 
         result = drop_redundant_rpaths(filepath)