The ksmbd changes for Linux 7.3 include opt-in Apple Time Machine interoperability. This enables a new use for the kernel SMB server, but a successful share connection alone does not establish that a backup can be restored.
What the kernel change adds
The maintainer's 7.3 pull request describes AAPL negotiation and related compatibility work for macOS backupd. Support is explicitly opt-in. Per-share encryption and request replay are separate additions in the same server update.
The feature belongs to the server path. An ordinary Linux SMB client mounting another server is a different component. Also check the userspace configuration tooling packaged with the server: the presence of a kernel feature does not establish that every distribution exposes it through the same setting.
Evaluate the recovery path
Use expendable sample data for an initial rehearsal. Create a small folder with a document, a nested directory and a filename containing accented characters. Record their contents locally. Complete a backup, change the document and complete another backup.
Then restore the earlier document to a separate location and compare its contents with the recorded original. Restore the directory as well, checking names and the properties your workflow relies on. This tests a different path from merely browsing the backup destination.
Next, rehearse an interruption and reconnection with the same sample data. Record whether the client resumes, starts a new operation or reports an error. A completed first backup does not predict how a later interrupted one behaves.
This is a proposed acceptance exercise, not a Time Machine certification or a test performed by PeopleAreGeek. It does not establish bare-metal recovery, every macOS version or every backing filesystem.
Separate availability from protection
A network share, encrypted transport and a recoverable backup answer three different questions: can the client reach the server, can the chosen connection protect data in transit, and can the required historical data be recovered? Success on one does not prove the others.
Keep the kernel version, server tooling, client version and destination storage configuration with the rehearsal result. That makes the result useful after a future update. The new interoperability is worth evaluating precisely because it expands the available server choices; it should be assessed on the complete backup and recovery workflow.
Explain opt-in AAPL support and a restore-based acceptance workflow, without pretending to test a backup.