WIP: hotfixing invalidated login links #158

Closed
linus wants to merge 3 commits from hotfix-invalidated-login-links into master
Owner

This fixes #129 and also fsfe-system-hackers/forms#12. It's quick and dirty and I'll see whether I can find a better solution.

I wonder though whether this meaningfully decreases security against the (quite insecure) status quo and if so how.

This fixes https://git.fsfe.org/fsfe-system-hackers/fsfe-cd/issues/129 and also https://git.fsfe.org/fsfe-system-hackers/forms/issues/12. It's quick and dirty and I'll see whether I can find a better solution. I wonder though whether this meaningfully decreases security against the (quite insecure) status quo and if so how.
linus added 3 commits 2022-03-09 14:45:26 +00:00
ff7b7b47ee remove token deletion on successful authentication
This removes the invalidation of the login token upon successful
authentication via passwordless e-mail
continuous-integration/drone/push Build is passing Details
continuous-integration/drone/pr Build is passing Details
fb40ae4bc3
remove misleading e-mail text
Owner

Thanks for coming up with this hot fix for a known bug.

My perspective is that it does not worsen the situation. If mails are intercepted, it does not really matter IMHO whether a link can be used only once. It's just a bit uncommon I think.

The only possible additional attack vector would be that someone who cannot read the email but the user's traffic (e.g. by sniffing) could extract the GET parameter of the one-time login and use it themselves. However, whoever can sniff network traffic could also read the unencrypted email. The user would just lose a "warning sign" if they actually would care about it – which I doubt for 99% of the users.

Of course, this is no silver bullet to the more fundamental rework of the FSFE-wide authentication system.

Thanks for coming up with this hot fix for a known bug. My perspective is that it does not worsen the situation. If mails are intercepted, it does not really matter IMHO whether a link can be used only once. It's just a bit uncommon I think. The only possible additional attack vector would be that someone who cannot read the email but the user's traffic (e.g. by sniffing) could extract the GET parameter of the one-time login and use it themselves. However, whoever can sniff network traffic could also read the unencrypted email. The user would just lose a "warning sign" if they actually would care about it – which I doubt for 99% of the users. Of course, this is no silver bullet to the more fundamental rework of the FSFE-wide authentication system.
Owner

I think this fix is acceptable. I don't see how it would fix fsfe-system-hackers/forms#12, though.

I think this fix is acceptable. I don't see how it would fix fsfe-system-hackers/forms#12, though.
linus changed title from WIP: hotfixing invalidated login tokens by link-followers to WIP: hotfixing invalidated login links 2022-03-09 16:27:08 +00:00
Author
Owner

I think this fix is acceptable. I don't see how it would fix fsfe-system-hackers/forms#12, though.

It would do so by the link not being invalidated upon successful authentication with it. But I already openened a PR with a more comprehensive fix for this issue: #159

> I think this fix is acceptable. I don't see how it would fix fsfe-system-hackers/forms#12, though. It would do so by the link not being invalidated upon successful authentication with it. But I already openened a PR with a more comprehensive fix for this issue: https://git.fsfe.org/fsfe-system-hackers/fsfe-cd/pulls/159
Owner

I'll have a look at #159 soon, but I think regarding fsfe-system-hackers/forms#12 there's a misunderstanding: that issue is not about the login links, but about the confirmation links for double opt-in, completely handled in fsfe-forms.

I'll have a look at #159 soon, but I think regarding fsfe-system-hackers/forms#12 there's a misunderstanding: that issue is not about the login links, but about the confirmation links for double opt-in, completely handled in fsfe-forms.
Author
Owner

I'll have a look at #159 soon, but I think regarding fsfe-system-hackers/forms#12 there's a misunderstanding: that issue is not about the login links, but about the confirmation links for double opt-in, completely handled in fsfe-forms.

Yes! I'm sorry for the confusion. I don't know a lot about forms, so I made unfounded assumptions.

> I'll have a look at #159 soon, but I think regarding fsfe-system-hackers/forms#12 there's a misunderstanding: that issue is not about the login links, but about the confirmation links for double opt-in, completely handled in fsfe-forms. Yes! I'm sorry for the confusion. I don't know a lot about `forms`, so I made unfounded assumptions.
linus added this to the (deleted) milestone 2022-03-14 09:59:06 +00:00
linus self-assigned this 2022-03-15 13:05:43 +00:00
Author
Owner

@reinhard @max.mehl Let me know if I should close this in favour of #159

@reinhard @max.mehl Let me know if I should close this in favour of #159
Owner

Yes, I think #159 is the by far better solution.

Yes, I think #159 is the by far better solution.
linus closed this pull request 2022-04-04 14:34:27 +00:00
All checks were successful
continuous-integration/drone/push Build is passing
continuous-integration/drone/pr Build is passing

Pull request closed

Sign in to join this conversation.
No description provided.