←back to thread

F-Droid site certificate expired

(gitlab.com)
155 points kxxt | 2 comments | 31 Aug 25 12:16 UTC | HN request time: 0.653s | source
Show context
gethly ◴[31 Aug 25 14:32 UTC] No.45083427[source]▶
Because those ephemeral LE certificates are such a great idea...
replies(6): >>45083455 #>>45083516 #>>45083798 #>>45083991 #>>45084464 #>>45088393 #
hkt ◴[31 Aug 25 15:35 UTC] No.45083991[source]▶
DANE would be better than LE, but weirdly the massive companies building browsers don't want to provide support. Spooky!

https://en.wikipedia.org/wiki/DNS-based_Authentication_of_Na...

replies(2): >>45084450 #>>45084457 #
aaomidi ◴[31 Aug 25 16:30 UTC] No.45084450[source]▶
You’re just moving your root of trust to DNS then?

With certificates we’re doing multi perspective validation.

DNS root of trust is silly. DNSSEC is not a proper root of trust

replies(1): >>45085033 #
ocdtrekkie ◴[31 Aug 25 17:29 UTC] No.45085033[source]▶
DNS is already the root of trust, certificates are domain-validated. We currently just depend on both DNS and an unelected group of random companies Google has decided jump through their arbitrary hoops often enough.

If your domain register or DNS provider is compromised in any way, all of the bullcrud the CA/B demands of certificates is entirely meaningless, the bad actor can legitimately request certificates.

replies(1): >>45086328 #
aaomidi ◴[31 Aug 25 19:35 UTC] No.45086328[source]▶
This is what multi perspective helps with. It doesn’t mitigate every single attack.

But think about what DANE is for a second. If a bad actor is MITMing your connection to some endpoint, they certainly can MITM your DNS queries too.

replies(1): >>45086412 #
1. ocdtrekkie ◴[31 Aug 25 19:45 UTC] No.45086412[source]▶
Multi-perspective helps prevent MITM, it doesn't provide any better security than your domain and DNS provider's security. It's just another layer to patch over the bad idea of CAs in the first place.
replies(1): >>45090641 #
2. crote ◴[01 Sep 25 08:11 UTC] No.45090641[source]▶
It's a completely different attack scenario.

DANE isn't going to be of any value when an attacker is sitting between the end user and their ISP - which was already the requirement for compromising the TLS connection in the first place - as they could just strip DNSSEC and fake the DANE records.