Critical Security Release for GitLab 8.2 through 8.7

https://about.gitlab.com/2016/05/02/cve-2016-4340-patches/

by iMerNibor • 10 years ago

95 35 10 years ago

about.gitlab.com

beefhash• 10 years ago

Is it just me or has gitlab issued a lot of critical security releases lately?

I'm not sure if this is because they're particularly open about these things or because their product might be particularly insecure.

JBReefer• 10 years ago

Having worked in a large set of codebases as a consultant, I'm pretty sure it's the former. The code that runs the world is mostly garbage (the proprietary stuff that is, because they can hide it)

nitrogen• 10 years ago

Similarly, never count on a compromise of your data to get reported unless local law absolutely requires it, and someone who cares notices. If your data was compromised in a way not covered by local law, companies will hide the compromise.

dalanmiller• 10 years ago

Other commentor already nailed it I think. How many critical security issues do you think Github has had that you've never heard of? Not knocking Github, it's just the nature of extreme open source I think.

danielsamuels• 10 years ago

Github isn't open source.

CameronBanga• 10 years ago

Finding, fixing, and openly discussing software issues makes me more confident in GitLab, not less. I see this as a positive thing.

sytse3• 10 years ago

Thanks! We're working on a postmortem blog post about our process for releasing these fixes.

JamesSwift• 10 years ago

Your comment is weirdly similar to this other comment that I saw in the initial announcement:

https://news.ycombinator.com/item?id=11593362

unethical_ban• 10 years ago

I had to check that it wasn't the same user. I know groupthink is a thing, but the cynic in me wondered if this comment wasn't an upvote grab.

sdesol2• 10 years ago

Since we are doing deja vu, I uploaded some new metrics that better highlights how insane GitLab's churn rate is.

http://imgur.com/a/4uaSR

What's really interesting is the number of contributors.

neandrake• 10 years ago

You should also factor in their software model. GitLab has a free install anyone can run locally without having to register. A competitor like GitHub has enterprise install but otherwise most public traffic is on their own hosted install. If there's a security issue they likely fix their hosted install and inform their enterprise customers only -- as well they probably shouldn't make it public otherwise it could be considered some level of exposure to their customers. Conversely GitLab essentially needs to publicize this to maintain the trust of their customers.

viraptor• 10 years ago

There are other reasons too. Once someone finds an issue in an open-source project, other people realise it's not super secure and start to dig deeper. Same thing happened with image decoders (libpng), asn1 parsers, syscalls fuzzing, digging into openssl, etc. Bugs in big products and common protocols come in waves.

ashitlerferad• 10 years ago

I think that makes them more trustworthy. If some open source software has never seen a CVE, then that is because no-one actually looked at it in depth enough. All software has bugs and security issues.

educar• 10 years ago

Not been tracking GitLab closely but how many critical issues have there been in the last 6 months?

kentonv4• 10 years ago

Gitlab can be run on Sandstorm.io (of which I am tech lead / co-founder). Sandstorm claims to mitigate most vulnerabilities in apps:

https://docs.sandstorm.io/en/latest/using/security-non-event...

Let's see how it scores here...

For background, on Sandstorm, each Gitlab project is placed in a separate grain (container), isolated from all others. In order to communicate with a grain at all, you must have been granted some level of access to it by its owner -- Sandstorm does not let you send requests to private grains to which you haven't been given access. So, private Gitlab repos hosted on Sandstorm are basically not vulnerable to any vulnerability.

Of course, Gitlab is the kind of thing you might intentionally make public to all, e.g. to host an open source project. Therefore, it makes sense to analyze whether a public repository would be exploitable.

    Privilege escalation via "impersonate" feature

On Sandstorm, authentication is handled by Sandstorm. The app receives an unspoofable header indicating which user the request came from, and what permissions they have. A well-written app uses this header on every request to authenticate the user.

Unfortunately, our Gitlab package currently uses this information only when a session first opens, then relies on the session cookie going forward. This pattern is sometimes used as a "hack" on Sandstorm to more easily integrate with existing login code designed to do upfront / one-time authentication. As such, public Gitlab repositories hosted on Sandstorm would be vulnerable.

Had Gitlab on Sandstorm been implemented "properly", it would not be vulnerable. That said, the "impersonate user" feature would not have worked at all. That's probably for the best: a feature like this really ought to be implemented by Sandstorm itself, which would have the ability to implement it (securely) across all apps at once, rather than have each app implement its own version.

Somewhat embarrassingly, the Sandstorm package of Gitlab actually predates this feature being added, therefore Gitlab instances on Sandstorm today actually aren't vulnerable. (Generally, if the upstream app author does not directly maintain the Sandstorm package, then the Sandstorm package will tend to fall behind. This should get better as Sandstorm gains popularity and upstream authors target it explicitly.)

    Privilege escalation via notes API
    Privilege escalation via project webhook API
    Information disclosure via project labels
    Information disclosure via new merge request page

These vulnerabilities allow a user to manipulate a private project to which they aren't supposed to have access. On Sandstorm, a private project would live in its own grain, and if you hadn't been given access then Sandstorm would deny you the ability to talk to the grain at all. Therefore, these cannot be exploited.

    XSS vulnerability via branch and tag names
    XSS vulnerability via custom issue tracker URL
    XSS vulnerability via label drop-down

These vulnerabilities require that you have write access to one project on the server in order to launch an attack. On Sandstorm, since every repository is its own instance, the attack would be limited to the repository on which you have write access -- you would not be able to use this attack to damage someone else's Gitlab repository on which you lack write access.

    XSS vulnerability via window.opener

This is a subtle phishing issue that is very widespread. However, it mostly doesn't work when the opener is a sandstorm app: the app lives inside an iframe which is prohibited by Content-Security-Policy from browsing away from the server.

    Information disclosure via milestone API
    Information disclosure via snippet API

A public project hosted on Sandstorm would be vulnerable to these (leaking confidential issues attached to public milestones, and leaking private snippets attached to a public project). Sandstorm can only enforce access control at the grain level; anything finer than that is up to the app, and is subject to app bugs. I generally recommend that Sandstorm users try to put confidential data in separate grains from public data -- e.g. creating a separate issue tracker for confidential issues.

(The Sandstorm packaging again predates the milestone bug's introduction, though may be affected by the snippet issue.)

We will update the Gitlab package tomorrow. Once an update is pushed, every Sandstorm user will receive a notification within 24 hours and can apply the update with one click.

Conclusion: Sandstorm mitigated 8/11 issues by design, 2/11 by accident, and is vulnerable to 1/11. The biggest issue was only mitigated by accident in this case, although a well-behaved Sandstorm app normally wouldn't have this kind of issue by design. Overall, though, I'm disappointed in Sandstorm's performance here -- it's much worse than the usual 95% mitigation rate.

forgotpwtomain• 10 years ago

Wouldn't this be better in a blog post than as an HN comment? It seems interesting enough - but at least personally I would prefer that the HN comment thread is a discussion of the topic, rather than marketing of another product.

mserdarsanli• 10 years ago

From https://sandstorm.io/install

Run this in a terminal:
curl https://install.sandstorm.io | bash

kthxbye

luch• 10 years ago

I get what you're trying to say, but how is it more insecure than doing "git clone; make; sudo make install" or double-clicking on a NSIS installer ?

corobo• 10 years ago

and then you read on into the next few lines for

"Wondering what this script will do? [Watch this 30 second screencast] or [read the install documentation].

If you prefer more validation, you can [download the install script from GitHub] or [build from source]."

I do get the point you're trying to make but as far as 'curl|bash' installations go these guys at least seem be doing it as right as they can

egil• 10 years ago

At least it is https, so no mitm attack. You have to trust the Sandstorm developers, even if you download the source and apply all the steps manually. If the site gets compromised, the attackers can modify the source and the website documentation for the PGP verification.

zackp302• 10 years ago

No need for the rudeness.

I always get bothered when people bring up the `curl [...] | bash` “argument” (which is usually less of an argument and more of a rude dismissal of a good product).

Sure, the script downloaded with curl should be validated.

Not sure, but I’m pretty sure you don’t validate every tarball you download, and even if you do validate them, you certainly wouldn’t look through each line of code making sure there isn’t a backdoor or something.

Any malicious person could break into a tarball download site, and replace the checksums as well (only way to mitigate that is using a signature that is associated with the tarball maintainer), and even if something like PGP is used, the attacker would be able to replace the instructions to obtain the PGP key with their own generated one, or even remove the signature completely.

This is, of course, the same for curl piped into bash, but my point is that it’s only slightly worse than an unsecured package site (which I’m sure there are many of). The only thing worse about it is that it’s less noticeable when the script has been replaced.

Also, some scripts (such as the rvm installatio script) /require/ PGP to be used for the installation to even start.

In conclusion, please don’t be rude and dismiss an entire product because of a single installation method, which isn’t actually worse than a malicious tarball.

sytse3• 10 years ago

Thanks Kenton, very interesting. And thanks for supporting GitLab on Sandstorm.

daxelrod• 10 years ago

Since every project is isolated, how do you handle operations that require data from multiple projects, such as the activity feed or merge requests?

I'm also curious, how extensively do you need to modify GitLab for Sandstorm? Is that https://github.com/dwrensha/gitlab-sandstorm ?

kentonv4• 10 years ago

Since every project is isolated, how do you handle operations that require data from multiple projects, such as the activity feed or merge requests?

Right now: Poorly.

But in a few months: - We're adding an activity feed API to Sandstorm, so that you can get a unified feed and notifications across all apps. - The Powerbox (https://sandstorm.io/how-it-works#powerbox) allows connecting grains to each other. We will define a git API which each of the git apps (Gitlab, Gitweb, Gogs) can implement, so that you can even do cross-app merge requests.

gramakri• 10 years ago

Thanks for the release! We updated the cloudron.io app to use 8.7.1 we well. Cloudron users will get GitLab auto-updated tonight.

dewiz• 10 years ago

Note for admins using Apache 2.4, regarding the manual quick fix described, take into consideration the Apache auth config changes described here https://httpd.apache.org/docs/current/upgrading.html

Critical Security Release for GitLab 8.2 through 8.7

Related Stories