[GH-ISSUE #4020] Revisit profiles allowing netlink protocol #2508

Closed
opened 2026-05-05 09:11:28 -06:00 by gitea-mirror · 2 comments
Owner

Originally created by @ghost on GitHub (Mar 1, 2021).
Original GitHub issue: https://github.com/netblue30/firejail/issues/4020

As recently pointed out by @kris7t here the netlink protocol potentially exposes a huge kernel API sandbox escape surface. In that regard it might be wise to revisit our profiles containing that option.

I've been running several profiles without the netlink protocol via local overrides without issues for quite some time (e.g. aria2c, chromium-common-hardened, firefox-common, mocp, mpg123, mplayer, mpv, nodejs-common, qnapi, sqlitebrowser, sysprof) but that's always dependant on personal workflow and specific use-case. To be clear, I'm not aware of any real-world issues but wanted to file this more as a reminder here. A thorough check of all relevant profiles will certainly take considerable time. Still, IMO it's worth looking at it again.

Originally created by @ghost on GitHub (Mar 1, 2021). Original GitHub issue: https://github.com/netblue30/firejail/issues/4020 As recently pointed out by @kris7t [here](https://github.com/netblue30/firejail/issues/4013#issuecomment-786780504) the `netlink` protocol `potentially` exposes a huge kernel API sandbox escape surface. In that regard it might be wise to revisit our profiles containing that option. I've been running several profiles without the netlink protocol via local overrides without issues for quite some time (e.g. aria2c, chromium-common-hardened, firefox-common, mocp, mpg123, mplayer, mpv, nodejs-common, qnapi, sqlitebrowser, sysprof) but that's always dependant on personal workflow and specific use-case. To be clear, I'm not aware of any real-world issues but wanted to file this more as a reminder here. A thorough check of all relevant profiles will certainly take considerable time. Still, IMO it's worth looking at it again.
gitea-mirror 2026-05-05 09:11:28 -06:00
Author
Owner

@rusty-snake commented on GitHub (Mar 5, 2021):

Removing netlink from chromium-common-hardened beaks WebRTC, likely the same for firefox-common. mpv has a lot of features, so it's likely needed (maybe for hw related streams), sysprof likely needs it for some features I guess. sqlitebrowser I can not thing any case where it is needed.

Quoting myself for the record (https://github.com/netblue30/firejail/issues/4013#issuecomment-786762799):

It's always better to not add / remove netlink as long as nothing is broken. But how do you know that nothing is? If the program complains about it, it misses something it wants and if the profile contains it, it's often for a reason.

<!-- gh-comment-id:791727248 --> @rusty-snake commented on GitHub (Mar 5, 2021): Removing `netlink` from chromium-common-hardened beaks WebRTC, likely the same for firefox-common. mpv has a lot of features, so it's likely needed (maybe for hw related streams), sysprof likely needs it for some features I guess. sqlitebrowser I can not thing any case where it is needed. Quoting myself for the record (https://github.com/netblue30/firejail/issues/4013#issuecomment-786762799): > It's always better to not add _/ remove `netlink`_ as long as nothing is broken. But how do you know that nothing is? If the program complains about it, it misses something it wants _and if the profile contains it, it's often for a reason._
Author
Owner

@rusty-snake commented on GitHub (Jun 10, 2021):

Should we close?

<!-- gh-comment-id:858440143 --> @rusty-snake commented on GitHub (Jun 10, 2021): Should we close?
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#2508
No description provided.