wget-dev
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: wget2 | New option --dane (!522)


From: John Scott (@j0hns)
Subject: Re: wget2 | New option --dane (!522)
Date: Mon, 24 Apr 2023 20:37:36 +0000



John Scott commented on a discussion: 
https://gitlab.com/gnuwget/wget2/-/merge_requests/522#note_1365407811

As features are introduced to serve users, and as the person who requested this 
feature in the first place, I would like to voice my defense of the feature in 
its current state.

> If the certificate usage is PKIX-TA(0) or PKIX-EE(1), require both DANE and 
> WebPKI validation success.
If the certificate usage is DANE-TA(2) or DANE-EE(3), require DANE validation 
success.

The right place to implement this logic is in GnuTLS, which currently doesn't 
check the certificate usage information. That's why we document that this is 
not how it works, the feature is experimental, and it's opt-in. Note that the 
vast, vast majority of TLSA records in practice set the certificate usage to 
DANE-EE, so from a practical point of view this issue is moot.

> Note that for "opportunistic" DANE to offer meaningful security, TLSA lookup 
> failure (i.e. anything other than NOERROR, NODATA or NXDOMAIN) needs to be a 
> hard fail, otherwise a MiTM attacker can trivially downgrade DANE to absence 
> of records.

I'm no expert on DNS, I confess I learned about it only two weeks ago, but I 
thought DNSSEC wasn't vulnerable to downgrade attacks owing to the way it works?

Bottom line, this feature is not only useful for people like me who actually 
want it, but it builds excitement for DANE to get more folks implementing it 
and get more eyeballs before it can be deemed "stable." I'm using this feature 
today and I love it.

-- 
Reply to this email directly or view it on GitLab: 
https://gitlab.com/gnuwget/wget2/-/merge_requests/522#note_1365407811
You're receiving this email because of your account on gitlab.com.




reply via email to

[Prev in Thread] Current Thread [Next in Thread]