Skip to content

Conversation

@fused0
Copy link

@fused0 fused0 commented Nov 23, 2025

Here's the start of configurable subvolume names, to support non-ubuntu layout (at least layouts that only have different names, not nested ones).

It's functional, but I'm sure some (small?) issues are remaining.

Advice and suggestions appreciated.

See issues #368, #405, #266, etc

Best,
fused0

@ygerlach
Copy link
Contributor

ygerlach commented Nov 23, 2025

Nice! i just ran into this issue today, as i tried to run timeshift on debian in btrfs mode (I had to manually rename @rootfs to @ and fix my grub for it to work)

I think in the long term we should allow an arbitrary amount of subvolumes for backups. I believe some distros allow /boot or /etc to be their own subvolume. But i guess that would be a bigger topic.

I think there is some room to improve add_opt_btrfs_subvolume_names :

  • it seems the code for creating home and root text box is quite similar. maybe we could create a function like create_btrfs_subvolume_selection to de-duplicate this.
  • There should be some kind of guard against entering a wrong subvolume names. Maybe even use a select box instead of a text box
  • the boxes should be hidden or greyed-out when btrfs mode is inactive
  • i think the home subvolume selection could be moved to the thing where it asks to backup @home at all (see screenshot). I am thinking a selection box with the options: "Disabled", @home, @boot, ...)
  • selecting the same subvolume twice should also be guarded against.
grafik

Some other thing: i don't know if we need to care about changes of names in the config. What if i configure @rootfs, make some snapshots, rename it to @ maybe fuck up my system some time later, want to restore my old @rootfs snapshot, but because i changed the name in the timeshift config its now failing.

The question is: should we use the configured btrfs subvolume, or just restore all subvolumes in the snapshot folder or use the information from info.json? Im not sure how it is currently handled. (Enable and Disable backup of @home could have a similar problem)

@fused0
Copy link
Author

fused0 commented Nov 23, 2025

Thanks for the quick review, I'll get to the improvement suggestions soon :)

Regarding:

I am thinking a selection box with the options: "Disabled", @home, @boot, ...)

What would "Disabled" do?

selecting the same subvolume twice should also be guarded against.

Thanks, definitely, that seems like it could fail spectacularly.

@ygerlach
Copy link
Contributor

What would "Disabled" do?

the same as not selecting the checkbox seen in the screenshot. (I would replace the checkbox with the dropdown)

@fused0
Copy link
Author

fused0 commented Nov 23, 2025

I put the option into the backup type page, because it depends on the type of backup used. Selecting rsync is the equivalent of the option being disabled.

To me, it would be a little confusing to have it on the users page, but I'm also not super familiar with the full functionality of this page, because I saw it allows selecting home directories in cases where multiple users exist on the system.

Please let me know what you think.

Also deduplicated the subvolume ui selection code
@fused0
Copy link
Author

fused0 commented Nov 24, 2025

I've turned the subvolume name selection into a combo box, which should address almost all of the issues you had, except the one about the location of the option

Not entirely sure what the common options are (please let me know) and if there should be a "custom" option as well. The custom option would make it tricky in terms of making sure the config is valid.

@fused0
Copy link
Author

fused0 commented Nov 24, 2025

Actually, I went and change the two comboboxes into a single one, to select a predefined layout. Hopefully I left some room to add a "custom" option (implementation would be similiar how the custom date format is done, presumably).

The current state would be more than sufficient for my - and probably most users - needs.

Still not sure about the debian layout/subvolume names.

Also, I'd like to change some of the messages in the UI, but I'm not sure how to proceed in regards to the translation, example:

status_details = _("Select BTRFS system disk with root subvolume" + @" ($(App.root_subvolume_name))");

@mtwebster mtwebster changed the title Configurable Btrfs subvol names [WIP] Configurable Btrfs subvol names Nov 27, 2025
@fused0
Copy link
Author

fused0 commented Nov 30, 2025

@ygerlach apologies for the ping, but any advice or suggestions how to proceed from here?

@ygerlach
Copy link
Contributor

apologies for the ping

@fused0 no problem, the PR looks fine to me.

Not entirely sure what the common options are (please let me know) and if there should be a "custom" option as well.

i think we will get issues if people miss some layout. so we can add it then, if we miss it here.

Actually, I went and change the two comboboxes into a single one, to select a predefined layout.

i think that is reasonable

Just checked the ui again on debian. Here how it looks (for others interested in this PR):

grafik I think it looks nice. Making it "disabled" when rsync is active is nice.

Just a thought: do we really need a selection? Can't timeshift find the correct layout itself? Because usually there should only be one valid layout possible and if the user fucks up the selection, timeshift behaves badly. So best keep the user out of this decision? And if a user has a layout that matches multiple layouts, with subvolumes like @rootfs, @, @homefs and @home, the user will be proficient enough to edit the timeshift config himself.

So if we have some detection what subvolumes are existing and choose the correct layout and maybe display the chosen selection like: "Using subvolume layout: Ubuntu (@, @home)" would be fine i guess.

Do all systems always have a seperate home subvolume? (EDIT: No, my debian test vm aparently only has @rootfs)

When i try to create a snapshot i get:
grafik
(timeshift did not exit)

Maybe you should run some tests with your branch on some different systems/vms (debian, fedora, linuxmint, ...)

P.S: im not a timeshift maintainer, just an contributor, so my opinion is not binding :)

@fused0
Copy link
Author

fused0 commented Nov 30, 2025

@fused0 no problem, the PR looks fine to me.

Thanks, but after your reply, there is more to consider.

Just a thought: do we really need a selection?

The layout actually is auto detected for the intial config values, based on the distro name. I wondered the same, when I added it. I suppose one could parse the fstab to find what subovlumes (if any) root and home map to... Seems like a simple enough solution for the entire problem. It would catch all reasonable user-defined layouts.

the user will be proficient enough to edit the timeshift config himself.

Unfortunately, due to how the timeshift handles the settings and wizard dialogs - more or less immediately overwriting the config - there can only be two options: Having no option in the UI at all, or supporting "custom" layouts with settings in the UI. Otherwise, the edited config would be overwritten when you accidentally (or not) open the settings or wizard.

I think I'll go for it and add the custom layout option.

Do all systems always have a seperate home subvolume? (EDIT: No, my debian test vm aparently only has @rootfs)

Now I know why I couldn't find anything on the name of the home subvolume in Debian. :)

I'm wondering if timeshift would work if you configure it to not backup the home subvolume and leave the home subvolume option blank or wiht a dummy value - or if supporting layouts without a separate home would imply more work. I would very much appreciate if you could give it a try.

Maybe you should run some tests with your branch on some different systems/vms (debian, fedora, linuxmint, ...)

Unfortunately I'm on a rather limited and flaky internet connection at the moment.

P.S: im not a timeshift maintainer, just an contributor, so my opinion is not binding :)

Seems I wrongfully assumed you to be a maintainer :) Thank you for your detailed reply and suggestions, much appreciated.

I suppose it would be good to get a maintainer to weigh in on the issue.

Thanks,
fused0

empty home subvolume name and force include_btrfs_home_for_backup to false
@fused0
Copy link
Author

fused0 commented Nov 30, 2025

@ygerlach Interestingly, timeshitft handles leaving the home subvolume name empty somewhat gravefully. I've made some more tweaks to make it work on debian without a separate home subvolume - it feels a bit hacky, but it works!

Now I'm wondering if this solution is adequate, or if parsing fstab and adding a custom input group is even needed.

@fused0 fused0 marked this pull request as ready for review December 8, 2025 06:48
@fused0 fused0 changed the title [WIP] Configurable Btrfs subvol names Configurable Btrfs subvol names Dec 8, 2025
@mtwebster mtwebster requested a review from Copilot December 8, 2025 15:36
@mtwebster mtwebster changed the title Configurable Btrfs subvol names [Next] Configurable Btrfs subvol names Dec 8, 2025
Copy link

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

This PR implements configurable Btrfs subvolume names to support different Linux distribution layouts beyond Ubuntu's default @ and @home convention. The implementation adds a UI for selecting preset layouts (Ubuntu, Debian, Fedora) or configuring custom subvolume names, addressing issues #368, #405, and #266.

Key Changes:

  • Added root_subvolume_name and home_subvolume_name properties to Main.vala with distro-specific defaults
  • Created UI in SnapshotBackendBox.vala for subvolume layout selection with preset and custom options
  • Replaced all hardcoded "@" and "@home" references throughout the codebase with configurable names

Reviewed changes

Copilot reviewed 8 out of 8 changed files in this pull request and generated 15 comments.

Show a summary per file
File Description
src/Core/Main.vala Added subvolume name properties, distro detection logic, config persistence, and updated all references to use configurable names
src/Gtk/SnapshotBackendBox.vala Implemented UI for subvolume layout selection with combo box and custom entry fields
src/Gtk/UsersBox.vala Added logic to disable home subvolume backup checkbox when home_subvolume_name is empty
src/Gtk/SnapshotListBox.vala Updated size calculation displays to use configurable subvolume names
src/Gtk/RestoreWindow.vala Updated restore logic to check for configurable home subvolume name
src/Core/Subvolume.vala Updated subvolume removal logic to use configurable names
src/Core/SnapshotRepo.vala Updated mount path handling to use configurable subvolume names
src/Core/Snapshot.vala Updated snapshot reading logic to use configurable subvolume names

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Comment on lines +4279 to +4282
&& sys_subvolumes.has_key(home_subvolume_name)
&& (sys_subvolumes[home_subvolume_name].id == subvol_id)){

subvol = sys_subvolumes["@home"];
subvol = sys_subvolumes[home_subvolume_name];
Copy link

Copilot AI Dec 8, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Similar to the issue on lines 4186-4189, this code is vulnerable to null pointer exceptions when home_subvolume_name is empty. If home_subvolume_name is an empty string (as it is for Debian), attempting to access sys_subvolumes[home_subvolume_name].id could fail.

Add an empty string check:

else if ((sys_subvolumes.size > 0)
    && (home_subvolume_name != "")
    && sys_subvolumes.has_key(home_subvolume_name)
    && (sys_subvolumes[home_subvolume_name].id == subvol_id)){

Copilot uses AI. Check for mistakes.
Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm not sure this follows, as it checks if sys_subvolumes has the key home_subvolume_name - sys_subvolumes itself is generated from fstab and crypttab, so doesn't depend on home_subvolume_name or root_subvolume_name.


if (bak.subvolumes.has_key("@home")){
size += bak.subvolumes["@home"].total_bytes;
if (bak.subvolumes.has_key(App.home_subvolume_name)){
Copy link

Copilot AI Dec 8, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Potential null pointer exception when App.home_subvolume_name is empty. If home_subvolume_name is an empty string (as configured for Debian), bak.subvolumes.has_key("") could match unintended entries or bak.subvolumes[""] could cause issues.

Consider adding an empty string check:

if ((App.home_subvolume_name != "") && bak.subvolumes.has_key(App.home_subvolume_name)){
    size += bak.subvolumes[App.home_subvolume_name].total_bytes;
}
Suggested change
if (bak.subvolumes.has_key(App.home_subvolume_name)){
if ((App.home_subvolume_name != "") && bak.subvolumes.has_key(App.home_subvolume_name)){

Copilot uses AI. Check for mistakes.
Comment on lines 152 to 156
Gtk.TreeIter strore_iter;
int active = -1;
for (int idx = 0; idx < possible_layouts.length[0]; idx++) {
list_store.append(out strore_iter);
list_store.set(strore_iter,
Copy link

Copilot AI Dec 8, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Typo in variable name: 'strore_iter' should be 'store_iter'.

Suggested change
Gtk.TreeIter strore_iter;
int active = -1;
for (int idx = 0; idx < possible_layouts.length[0]; idx++) {
list_store.append(out strore_iter);
list_store.set(strore_iter,
Gtk.TreeIter store_iter;
int active = -1;
for (int idx = 0; idx < possible_layouts.length[0]; idx++) {
list_store.append(out store_iter);
list_store.set(store_iter,

Copilot uses AI. Check for mistakes.
custom_home_subvol_entry.focus_out_event.connect((entry1, event1) => {
App.home_subvolume_name = custom_home_subvol_entry.text;

// If home subolume name is empty, do not backup home.
Copy link

Copilot AI Dec 8, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Typo in comment: 'subolume' should be 'subvolume'.

Suggested change
// If home subolume name is empty, do not backup home.
// If home subvolume name is empty, do not backup home.

Copilot uses AI. Check for mistakes.
}
if (bak.subvolumes.has_key("@home")){
size += bak.subvolumes["@home"].unshared_bytes;
if (bak.subvolumes.has_key(App.home_subvolume_name)){
Copy link

Copilot AI Dec 8, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Same issue as lines 438-440. When App.home_subvolume_name is empty (e.g., for Debian), accessing bak.subvolumes[""] could cause issues.

Add an empty string check:

if ((App.home_subvolume_name != "") && bak.subvolumes.has_key(App.home_subvolume_name)){
    size += bak.subvolumes[App.home_subvolume_name].unshared_bytes;
}
Suggested change
if (bak.subvolumes.has_key(App.home_subvolume_name)){
if ((App.home_subvolume_name != "") && bak.subvolumes.has_key(App.home_subvolume_name)){

Copilot uses AI. Check for mistakes.
}
else if (name == "@home"){
path = path_combine(App.mount_point_app + "/backup-home", "@home");
else if (name == App.home_subvolume_name){
Copy link

Copilot AI Dec 8, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Potential issue when App.home_subvolume_name is empty. For distributions like Debian that set home_subvolume_name = "", this condition name == "" will match any empty name, which could cause unintended behavior.

Consider adding an explicit check:

else if ((App.home_subvolume_name != "") && (name == App.home_subvolume_name)){
Suggested change
else if (name == App.home_subvolume_name){
else if ((App.home_subvolume_name != "") && (name == App.home_subvolume_name)){

Copilot uses AI. Check for mistakes.
fused0 and others added 6 commits December 8, 2025 17:31
When root subvolume name is misconfigured to empty string, don't early out, but treat it the same as empty home subvolume name
Co-authored-by: Copilot <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Development

Successfully merging this pull request may close these issues.

2 participants