Closed
Bug 845126
Opened 13 years ago
Closed 13 years ago
Consider exposed preferences that allow FHR to be disabled
Categories
(Firefox Health Report Graveyard :: Client: Desktop, defect)
Firefox Health Report Graveyard
Client: Desktop
Tracking
(Not tracked)
RESOLVED
WONTFIX
People
(Reporter: gps, Unassigned)
Details
Currently there are a number of preferences exposed to FHR that allow it to be disabled while the application is running. It's come to my attention that some nefarious add-ons do things like disable the blocklist ping. Presumably they could disable FHR as well. We should consider what level of control we really want to give to the process itself.
I'm really not sure what SOP is here. Wait and react?
One idea would be to allow preferences to be set by a system administrator policy but not user set from within the application. Is there a way we could enforce that?
Are there any other ideas? I guess add-ons are still pretty much unsandboxed, so there is little we can do once a bad one has been installed :/
I don't have to remind anyone that we need to balance user control against concerns for nefarious add-ons.
Comment 1•13 years ago
|
||
Curtis, some fodder for sec review here.
Updated•13 years ago
|
Flags: sec-review?
We don't see a need to have a security review of this particular part of the feature. We are of the opinion that a pref should be publicly available for users to disable this if they so choose and said pref should be easy to find and use.
Flags: sec-review?
Comment 3•13 years ago
|
||
The summary makes it sound like you want to add a clearly defined UI checkbox so users can disable FHR like they can disable Telemetry. Comment 0 sounds like you want to get rid of such preferences altogether lest "nefarious" apps disable it.
Which one is this bug?
If people want to turn off FHR we let them turn it off. It should be easy for them to do so -- hiding these pings only makes MOZILLA appear to be the nefarious ones.
As to apps hacking you... is it really "nefarious" ones? There's no malicious reason, no advantage, to turn those off. If apps are doing it it means there's some number of people who do have those privacy concerns. You don't know who they are, maybe they're dissidents who are being tracked by the timing of their update pings or some equally farfetched scheme. Not for us to judge.
| Reporter | ||
Comment 4•13 years ago
|
||
(In reply to Daniel Veditz [:dveditz] from comment #3)
> The summary makes it sound like you want to add a clearly defined UI
> checkbox so users can disable FHR like they can disable Telemetry. Comment 0
> sounds like you want to get rid of such preferences altogether lest
> "nefarious" apps disable it.
We already have a checkbox in prefs linked to a pref. We're talking about hidden preferences in about:config.
> If people want to turn off FHR we let them turn it off. It should be easy
> for them to do so -- hiding these pings only makes MOZILLA appear to be the
> nefarious ones.
We already have all the controls required (essentially "allow data submission."). We implemented additional switches because we figured some people (like system administrators) may ask for them.
> As to apps hacking you... is it really "nefarious" ones? There's no
> malicious reason, no advantage, to turn those off. If apps are doing it it
> means there's some number of people who do have those privacy concerns. You
> don't know who they are, maybe they're dissidents who are being tracked by
> the timing of their update pings or some equally farfetched scheme. Not for
> us to judge.
Apparently some "nefarious" extensions disable the blocklist ping so they can't be remotely disabled by said ping. Presumably, these same extensions may also target FHR because FHR submits add-on info and being cloaked in a layer of invisibility is a good defense against your add-on being blocklisted :)
Comment 5•13 years ago
|
||
(In reply to Gregory Szorc [:gps] from comment #4)
> Apparently some "nefarious" extensions disable the blocklist ping so they
> can't be remotely disabled by said ping. Presumably, these same extensions
> may also target FHR because FHR submits add-on info and being cloaked in a
> layer of invisibility is a good defense against your add-on being
> blocklisted :)
Yes, this has always been a potential issue with the blocklist - though I don't know of any actual example of it happening in the wild (maybe Jorge does - CCed). There's not a lot we can currently do that would be effective. The strategies we can currently use either removes choice from the user, or creates an arms race (one where both we and users would lose). However, saying that, we do have a plan that would make all these types of issues go away: https://wiki.mozilla.org/Features/Desktop/Improve_Add-on_Monitoring
Comment 6•13 years ago
|
||
I've never seen an example of an extension doing this in the wild.
| Reporter | ||
Comment 7•13 years ago
|
||
(In reply to Jorge Villalobos [:jorgev] from comment #6)
> I've never seen an example of an extension doing this in the wild.
Interesting. Perhaps I heard incorrectly or the person I heard from was misinformed.
Perhaps we should WONTFIX this then?
Comment 8•13 years ago
|
||
I'll add my two bits here. :)
I don't want to see anything done in the short term to try to mitigate this risk. In my opinion, the chances of stepping on the toes of a user trying to exert control outweigh the possible benefit of preventing a bad actor from doing something without the user's consent.
In the long term though, I believe it would be useful to have access controls built in to the add-on framework, similar to how app permissions for iOS or Android work. And if an add-on tried to take certain actions such as changing the user's search engine or home page or changing a core preference, the user should be notified before the add-on is allowed to make such a change.
It is still a bit of an arms race, but I think it would be a solid first step toward preventing this type of bad actor.
| Reporter | ||
Comment 9•13 years ago
|
||
OK. The bug was meant to inspire a conversation (it has "consider" in its summary after all). I think we had a meaningful discussion and we decided not to do anything.
So, WONTFIX.
Thanks for chiming in, everybody.
Status: NEW → RESOLVED
Closed: 13 years ago
Resolution: --- → WONTFIX
Updated•13 years ago
|
Component: Metrics and Firefox Health Report → Client: Desktop
Product: Mozilla Services → Firefox Health Report
Updated•8 years ago
|
Product: Firefox Health Report → Firefox Health Report Graveyard
You need to log in
before you can comment on or make changes to this bug.
Description
•