Merge pull request #9385 from mr-raj12/darwin-f-fullsync

platform: use F_FULLSYNC on macOS for SyncFile data durability, fixes #9383

Author: ThomasWaldmann

src/borg/platform/__init__.py

@@ -50,6 +50,7 @@
     from .darwin import acl_get, acl_set
     from .darwin import is_darwin_feature_64_bit_inode, _get_birthtime_ns
     from .darwin import set_flags
+    from .darwin import fdatasync, sync_dir  # type: ignore[no-redef]
     from .base import get_flags
     from .base import SyncFile
     from .posix import process_alive, local_pid_alive

src/borg/platform/base.py

@@ -151,7 +151,6 @@ class SyncFile:
 
     Calling SyncFile(path) for an existing path will raise FileExistsError. See the comment in __init__.
 
-    TODO: Use F_FULLSYNC on macOS.
     TODO: A Windows implementation should use CreateFile with FILE_FLAG_WRITE_THROUGH.
     """
 

src/borg/platform/darwin.pyx

@@ -262,3 +262,33 @@ def set_flags(path, bsd_flags, fd=None):
         path_bytes = os.fsencode(path)
         if lchflags(path_bytes, c_flags) == -1:
             raise OSError(errno.errno, os.strerror(errno.errno), os.fsdecode(path_bytes))
+
+
+import errno as errno_mod
+import fcntl as fcntl_mod
+
+
+def fdatasync(fd):
+    """macOS fdatasync using F_FULLFSYNC for true data durability.
+
+    On macOS, os.fsync() only flushes to the drive's write cache.
+    fcntl F_FULLFSYNC flushes to persistent storage.
+    Falls back to os.fsync() if F_FULLFSYNC is not supported.
+    """
+    try:
+        fcntl_mod.fcntl(fd, fcntl_mod.F_FULLFSYNC)
+    except OSError:
+        # F_FULLFSYNC not supported (e.g. network filesystem), fall back
+        os.fsync(fd)
+
+
+def sync_dir(path):
+    """Sync a directory to persistent storage on macOS using F_FULLFSYNC."""
+    fd = os.open(str(path), os.O_RDONLY)
+    try:
+        fdatasync(fd)
+    except OSError as os_error:
+        if os_error.errno != errno_mod.EINVAL:
+            raise
+    finally:
+        os.close(fd)

src/borg/testsuite/platform/darwin_test.py

@@ -2,6 +2,7 @@
 import tempfile
 
 from ...platform import acl_get, acl_set
+from ...platform import fdatasync, sync_dir
 from .platform_test import skipif_not_darwin, skipif_fakeroot_detected, skipif_acls_not_working
 
 # Set module-level skips
@@ -46,3 +47,65 @@ def test_extended_acl():
         b"group:ABCDEFAB-CDEF-ABCD-EFAB-CDEF00000000::0:allow:read"
         in get_acl(file2.name, numeric_ids=True)["acl_extended"]
     )
+
+
+def test_fdatasync_uses_f_fullfsync(monkeypatch):
+    """Verify fcntl F_FULLFSYNC is called."""
+    import fcntl as fcntl_mod
+    from ...platform import darwin
+
+    calls = []
+    original_fcntl = fcntl_mod.fcntl
+
+    def mock_fcntl(fd, cmd, *args):
+        calls.append((fd, cmd))
+        return original_fcntl(fd, cmd, *args)
+
+    monkeypatch.setattr(fcntl_mod, "fcntl", mock_fcntl)
+
+    with tempfile.NamedTemporaryFile() as tmp:
+        tmp.write(b"test data")
+        tmp.flush()
+        darwin.fdatasync(tmp.fileno())
+
+    assert any(cmd == fcntl_mod.F_FULLFSYNC for _, cmd in calls), "fdatasync should call fcntl with F_FULLFSYNC"
+
+
+def test_fdatasync_falls_back_to_fsync(monkeypatch):
+    """Verify os.fsync fallback when F_FULLFSYNC fails."""
+    import fcntl as fcntl_mod
+    from ...platform import darwin
+
+    fsync_calls = []
+
+    def mock_fcntl(fd, cmd, *args):
+        if cmd == fcntl_mod.F_FULLFSYNC:
+            raise OSError("F_FULLFSYNC not supported")
+        return 0
+
+    def mock_fsync(fd):
+        fsync_calls.append(fd)
+
+    monkeypatch.setattr(fcntl_mod, "fcntl", mock_fcntl)
+    monkeypatch.setattr(os, "fsync", mock_fsync)
+
+    with tempfile.NamedTemporaryFile() as tmp:
+        tmp.write(b"test data")
+        tmp.flush()
+        darwin.fdatasync(tmp.fileno())
+
+    assert len(fsync_calls) == 1, "Should fall back to os.fsync when F_FULLFSYNC fails"
+
+
+def test_fdatasync_basic():
+    """Integration: fdatasync completes on a real file without error."""
+    with tempfile.NamedTemporaryFile() as tmp:
+        tmp.write(b"test data for fdatasync")
+        tmp.flush()
+        fdatasync(tmp.fileno())
+
+
+def test_sync_dir_basic():
+    """Integration: sync_dir completes on a real directory without error."""
+    with tempfile.TemporaryDirectory() as tmpdir:
+        sync_dir(tmpdir)