At the beginning of April, the Fedora Security Team (FST) started on a journey to close all critical and important CVEs in Fedora and EPEL that had originated in 2014 and before. Now that we’re two-thirds the way through I figured it would be a good time to see what we’ve accomplished so far.
Of the 38 CVEs (37 important and 1 critical) we originally identified: 14 have been closed, 1 is currently on QA, and 23 remain open. The 14 closed CVEs represent around a third of all the identified CVEs. So, not bad but also not great; there is still work to be done.
If you want to help get some of these CVEs cleaned up here’s a list of the target packages. We need to make sure that upstream has fixed the problem and that the packagers are pushing these fixes into the repos.
I hope to come back to you at the end of the month with a report on how all of the CVEs were fixed and who helped fix them!
Earlier this month the Fedora Security Team started a 90-day challenge to close all critical and important CVEs in Fedora that came out in 2014 and before. These bugs include packages affected in both Fedora and EPEL repositories. Since we started the process we’ve made some good progress.
Of the thirty-eight Important CVE bugs, six have been closed, three are on QA, and the rest are open. The one critical bug, rubygems-activesupport in EPEL, still remains but maybe fixed as early as this week.
Want to help? Please join us in helping make Fedora (and EPEL) and safer place and pitch in to help close these security bugs.
The Fedora Security Team (FST) has uncovered an interesting problem. Many packages in Fedora aren’t being actively maintained meaning they are unofficially orphaned. This is likely not a problem since at least some of these packages will happily sit there and be well behaved. The ones we worry about are the ones that pick up CVEs along the way, warning of unscrupulous behaviour.
The FST has been plugging away at trying to help maintainers update their packages when security flaws are known to exist. So far we’ve almost hit the 250 bug level. Unfortunately we forced a policy that still isn’t perfect. What do you do with a package that is no longer is supported and has a known vulnerability in it? Unless you can recruit someone to adopt the package the only responsible choice you have is to retire the package and remove it from the repositories.
This, of course, leads to other problems, specifically that someone has that package installed and they know not that the package is no longer supported nor do they know it contains a security vulnerability. This morning, during the FST meeting, we discussed the problem a bit and I had an idea that I’ll share here in hopes of starting a discussion.
Create a file containing all the packages that have been retired from a repository and perhaps a short reason for why this package has been retired. Then have yum/dnf consume this information regularly and notify the user/admin when a package that is installed is added to this list. This allows the system admin to become aware of the unsupported nature of the package and allows them to make a decision as to whether or not to keep the package on the system.
Wow, I had no idea that people would care about the start of this project. There seems to be a few questions out there that I’d like to address here to clarify what we are doing and why.
OMG! Fedora is just getting a security team? Does this mean Fedora has been insecure this entire time?!?
Umm, no, it doesn’t mean that Fedora has been insecure this entire time. In all actuality Fedora is in pretty good shape overall. There is always room for improvement and so we’re organizing a team to help facilitate that improvement.
What exactly is the security team responsible for?
We here to help packagers get the patches or new releases that fix vulnerabilities into the Fedora repositories faster. Most of our packagers are very good at shipping fixes for bugs when upstream rolls a new version of their software. Bug fixes can usually wait a few days, though, as most aren’t critical. Security vulnerabilities are a bit different and fixes should be made available as soon as possible. A little helping hand is never a bad thing and that’s what we’re here to do… help.
Can the security team audit package x?
No. This may become a service a different team (also falling under the Security SIG) can provide but I/we haven’t gotten there yet.
I read where Fedora has 566 vulnerabilities! How can you say that Fedora isn’t insecure?
Well, it’s actually 573 right this second. That’s down from 577 last week. 566 was Monday’s number. It’s important to not get caught up in the numbers cause they are, well, just numbers. The numbers only deal specifically with the number of tickets open. Many of the tickets are duplicates in that the same vulnerability might have several tickets opened for it if the finding is in only certain Fedora versions and EPEL versions. Since the same packager is likely responsible for all versions and the same fix can be made we can likely close several bugs at a time with minimal work.
I should also point out that the majority of these bugs fall well below the “world is on fire” level of Critical and the “this isn’t good” level of Important. This doesn’t mean we should just ignore these lower vulnerabilities but rather we should understand that they aren’t something that is likely to be exploited without many other bad things happening. Should they be fixed? Yes, but we should probably be more concerned with the Critical and Important vulnerabilities first. If you’d like to know more about the process for coming up with the severity rating my friend Vincent wrote an excellent article that you should read.
“6. Close bug when vulnerability is shipped in Fedora repos.”
Yeah, that isn’t correct. This is what happens when I try to multi-task. Glad I don’t get paid to write…. err… never mind. Luckily it’s a wiki and someone fixed it for me. Whew!
(We try to not deliberately release a package with a vulnerability. It seems people don’t appreciate vulnerabilities in the same way they like other features. Who would have thought?)
I’d like to help! How can I join up?
Go to the Security Team wiki page and look for the link to the mailing list and IRC channels, sign up, join up, and use the work flow to start digging in. Questions? Feel free to ask in the IRC channel or on the mailing list. You can also contact me directly if can’t otherwise find the answer to your question.
Do you hate security vulnerabilities?
Do you want to help make Fedora more secure?
Do you have a little extra time in your week to do a little work (no coding required)?
If you answered yes to the questions above I want you for a beta test of an idea I have to help make Fedora more secure. I’m looking for just a few people (maybe five) to sort through security bugs and work with upstream and packagers to get patches or new releases into Fedora and help make everyone’s computing experience a little safer. If you’re interested please contact me (firstname.lastname@example.org 0x024BB3D1) and let me know you’re interested.
It’s been a while since anyone has really gone through the Fedora Security Guide. In the few places I’ve poked at I’ve found a massive build-up of… old stuff. I’m quite sure that many of those commands don’t do what is said here any longer.
There are ~50 files that need to be looked at to make sure that the information is current and actually tells people the proper thing to do to secure their Linux system. In support of this mission (that I hope to complete before the release of Fedora 21) I’ve created a wiki page that outlines the goals, explains how to help, and lists the pages that need to be audited and/or updated. I’d appreciate some assistance if only to have more eyes looking over the pages and making comments on the wiki about things that need to be fixed.