[GH-ISSUE #2155] Add firejail.config option to prevent loading of user profiles. #1464

Open
opened 2026-05-05 08:07:41 -06:00 by gitea-mirror · 5 comments
Owner

Originally created by @crass on GitHub (Oct 9, 2018).
Original GitHub issue: https://github.com/netblue30/firejail/issues/2155

On systems where the unprivileged user does not have access to root, the sysadmin should have the ability to disallow custom profiles. This will lower the attack surface on firejail itself. In order for this to be effective, when user profiles are disabled, the --profile option should only allow loading files from ${CFG}. What should this option be named in firejail.config? Perhaps user_profiles? And I suggest having them off by default (secure default).

Originally created by @crass on GitHub (Oct 9, 2018). Original GitHub issue: https://github.com/netblue30/firejail/issues/2155 On systems where the unprivileged user does not have access to root, the sysadmin should have the ability to disallow custom profiles. This will lower the attack surface on firejail itself. In order for this to be effective, when user profiles are disabled, the `--profile` option should only allow loading files from `${CFG}`. What should this option be named in `firejail.config`? Perhaps `user_profiles`? And I suggest having them off by default (secure default).
gitea-mirror added the
enhancement
label 2026-05-05 08:07:41 -06:00
Author
Owner

@smitsohu commented on GitHub (Oct 9, 2018):

This will lower the attack surface on firejail itself.

There is still always the command line, though.

<!-- gh-comment-id:428271417 --> @smitsohu commented on GitHub (Oct 9, 2018): > This will lower the attack surface on firejail itself. There is still always the command line, though.
Author
Owner

@reinerh commented on GitHub (Oct 9, 2018):

Right, when the users can still override everything on the command line, but no longer in their config files, that's an even bigger reason to keep it on by default.

<!-- gh-comment-id:428272941 --> @reinerh commented on GitHub (Oct 9, 2018): Right, when the users can still override everything on the command line, but no longer in their config files, that's an even bigger reason to keep it on by default.
Author
Owner

@crass commented on GitHub (Oct 9, 2018):

@smitsohu very good point... ☹️ As I see it now, either we should not do this or it needs to disable any options which can loosen the container. Options that can tighten up the container should be ok. This will require a more thorough evaluation of the options...

<!-- gh-comment-id:428305648 --> @crass commented on GitHub (Oct 9, 2018): @smitsohu very good point... :frowning_face: As I see it now, either we should not do this or it needs to disable any options which can loosen the container. Options that can tighten up the container should be ok. This will require a more thorough evaluation of the options...
Author
Owner

@crass commented on GitHub (Oct 10, 2018):

So I'm thinking that the pull request shouldn't be merged as-is because it gives the illusion of locking down the users. I was thinking of abandoning the rest of the work because it seemed like a PITA. But, it looks like it could be easily implemented in check_arg using checkcfg and a list of options we want to disallow.

So now the PITA part is getting the list of options that we should disallow. If someone or the collective wants to come up with such a list, I could implement it. I'm not interested in doing that part myself though.

<!-- gh-comment-id:428469658 --> @crass commented on GitHub (Oct 10, 2018): So I'm thinking that the pull request shouldn't be merged as-is because it gives the illusion of locking down the users. I was thinking of abandoning the rest of the work because it seemed like a PITA. But, it looks like it could be easily implemented in `check_arg` using `checkcfg` and a list of options we want to disallow. So now the PITA part is getting the list of options that we should disallow. If someone or the collective wants to come up with such a list, I could implement it. I'm not interested in doing that part myself though.
Author
Owner

@chiraag-nataraj commented on GitHub (May 21, 2019):

Are we interested in moving forward with this or should we close this?

<!-- gh-comment-id:494192782 --> @chiraag-nataraj commented on GitHub (May 21, 2019): Are we interested in moving forward with this or should we close this?
Sign in to join this conversation.
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference: github-starred/firejail#1464
No description provided.