Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Netgear Signed TLS Cert Private Key Disclosure

gist.github.com

Post reply on HN
Print view View original

151–158 of 158 posts

Re: Netgear Signed TLS Cert Private Key Disclosure

#151

Post by dlgeek »

cjbprime wrote: post #147
dlgeek wrote: post #112

Earlier quoted context omitted.

Anyone with access to the private key can submit a problem report to the CA, they're obligated to revoke if the key is exposed to a non-subscriber. I have submitted a report - they are obligated to revoke within 24 hours.

Looks like they've failed to perform the revocation in time -- other users reported to Entrust >24h ago.

They revoked it 7 minutes shy of the 24 hour deadline after my report.

If other users reported it and they failed to act, that's a BR violation and they're required to provide an incident reports to the root programs via Mozilla's mechanisms.

If they don't, and you have evidence of those reports, you should provide them to the root programs via the mozilla.dev.security.policy mailing list/group. There's already a thread about this issue, though not claiming that EnTrust was notified.

(Based on timestamps, it's also possible they revoked in response to that thread).

Re: Netgear Signed TLS Cert Private Key Disclosure

#152

Post by cjbprime »

dlgeek wrote: post #151
cjbprime wrote: post #147

Earlier quoted context omitted.

Looks like they've failed to perform the revocation in time -- other users reported to Entrust >24h ago.

They revoked it 7 minutes shy of the 24 hour deadline after my report. If other users reported it and they failed to act, that's a BR violation and they're required to provide an incident reports to the root programs via Mozilla's mechanisms. If they don't, and you have evidence of those reports, you should provide them to the root programs via the mozilla.dev.security.policy mailing list/group. There's already a thr…

Thanks. Here is the timestamp showing that it took more than 32 hours from reporting to revocation.

https://twitter.com/FiloSottile/status/1219147543667453953?s...

Re: Netgear Signed TLS Cert Private Key Disclosure

#153

Post by mkeedlinger »

This reminds me of something I have been looking for for a long time. If you know any reasonable way to fix it please let me know!

This "any CA can authenticate any domain" system is ridiculous. I manage my own CA, which is fine for my own self-hosted sites, but the issue is that it doesn't protect me from a cert made valid by some other CA.

Is there any way I can whitelist my own CA such that IT ALONE will work for my domains? (note: this is a me-only thing. Don't need anyone to be able to validate these domains).

PS, I've asked this question elsewhere [0] before. I did not find an adequate answer.

[0]: https://security.stackexchange.com/questions/211401/can-you-...

Re: Netgear Signed TLS Cert Private Key Disclosure

#154

Post by zzzcpan »

mkeedlinger wrote: post #153

This reminds me of something I have been looking for for a long time. If you know any reasonable way to fix it please let me know! This "any CA can authenticate any domain" system is ridiculous. I manage my own CA, which is fine for my own self-hosted sites, but the issue is that it doesn't protect me from a cert made valid by some other CA. Is there any way I can whitelist my own CA such that IT ALONE will work for…

See DNS CAA https://en.wikipedia.org/wiki/DNS_Certification_Authority_Au...

Re: Netgear Signed TLS Cert Private Key Disclosure

#155

Post by mkeedlinger »

zzzcpan wrote: post #154
mkeedlinger wrote: post #153

This reminds me of something I have been looking for for a long time. If you know any reasonable way to fix it please let me know! This "any CA can authenticate any domain" system is ridiculous. I manage my own CA, which is fine for my own self-hosted sites, but the issue is that it doesn't protect me from a cert made valid by some other CA. Is there any way I can whitelist my own CA such that IT ALONE will work for…

See DNS CAA https://en.wikipedia.org/wiki/DNS_Certification_Authority_Au...

Thanks for the reply!

This is very interesting, I will surely take a look. Only issue I see is that it seems to use DNS, and that's not exactly secure either.

In my mind I am thinking of something that would be more theoretically secure from any remote attack (closer to a form of key-pinning maybe?)

Re: Netgear Signed TLS Cert Private Key Disclosure

#156

Post by gdm85 »

air7 wrote: post #117
m-p-3 wrote: post #97

Earlier quoted context omitted.

Agreed, the key was already publicly available, and it was just a matter of time for someone with malicious intent to find it if that's not already the case. Exposing something already public to speed up the resolution and make sure the impact is kept to a minimum considering the circumstances.

Well, one can argue that every vulnerability is "already publicly available, and just a matter of time for someone with malicious intent to find it"

I didn't know that Netgear already has proven to not be following coordinated disclosure; generally I think it's fair to give some time to the vendor.

I like how Google Project Zero does it: give time, but don't extend indefinitely.

Re: Netgear Signed TLS Cert Private Key Disclosure

#157

Post by willhschmid »

tialaramex wrote: post #101
vbezhenar wrote: post #73

Earlier quoted context omitted.

AFAIK CA can't issue certificate for more than two years. So if you've bought the device two years after it was manufactured, you're getting error and it's even worse than HTTP for user.

825 days is the limit. It's enforced in, for example, Chrome, if a certificate claims to have been issued after this rule changed and it lasts longer than 825 days Chrome considers the certificate bogus immediately. Certificate Authorities use the difference between 825 days and two years to offer certificate renewal weeks prior to the expiry date while keeping your "extra" days. e.g. your certificate expires 17 Febr…

Can you please provide a reference to the "825 days" limitation.

Re: Netgear Signed TLS Cert Private Key Disclosure

#158

Post by willhschmid »

creeble wrote: post #123
wbl wrote: post #121

Earlier quoted context omitted.

Plain HTTP is fine because the IP is a site-local.

Except for the browser complaint when filling out the form. I have to agree with the calls for allowing warning-less HTTP for LAN connections. Edit: Add last sentence.

>I have to agree with the calls for allowing warning-less HTTP

Completely flawed assumption, see the following for details:

https://realtimelogic.com/articles/How-Anyone-Can-Hack-Your-...

Post reply on HN