6 ms·

There is nothing there that says every service must use specific certificates, just that browsers should accept certain ones. So this in no way breaks encryptio

There is nothing there that says every service must use specific certificates, just that browsers should accept certain ones. So this in no way breaks encryption for apps who care, this only reduces security on apps that wants to reduce security. For example, if you use private "e2echat.com" it can still use safe certs and be safe, the risk is only that "governmentchat.com" will use bad certs, which was already a risk.

no_time 3y ago
If "e2echat.com" has no method to explicitly forbid your browser from accepting eIDAS certs (via a DNS record or something) then your browser will just blindly accept the compromised cert when attacked. This is still very bad.
- Jensson 3y ago
  > This is still very bad. Yes, potentially, but it isn't "another kind of chat control".
  - g-b-r 3y ago
    It's another side of the efforts of going around encryption, chat controls deals with communication services, this one with browsers
    - Jensson 3y ago
      But this doesn't force browsers and sites to use weak encryption. It is very different.
      - g-b-r 3y ago
        This forces browsers to accept all the CAs approved by the EU states, and you can be certain that some of them will be used for decrypting (and if needed modifying) the traffic
        Jensson 3y ago
        And then you can just tell the browser to not trust those CAs and you are safe. This is nothing like "chat control". This only lets the government spy on people who don't care if the government spies on them.
        yaris 3y ago
        IIRC one cannot tell the browser to not trust root CAs, that's why all the fuss.
        Jensson 3y ago
        Why shouldn't you be able to do that? Seems like a simple thing to implement. I get why they want a hardcoded list, but I don't get why you can't add a way to block parts of that hardcoded list.
        g-b-r 3y ago
        web-browsers shall ensure
        Jensson 3y ago
        The only requirement is that browsers displays the data. The browser can add "warning, this certificate is potentially compromised" when it displays it, nothing in the current document says browsers aren't allowed to say that, just that the browser has to be aware of the certificate. It is similar to how Chrome displays a warning when you visit some sites. You can visit the site anyway, but you get a warning since Google thinks it is bad.
        yaris 3y ago
        Technically correct. But if Firefox displays a big red warning when someone's grandma goes to her favourite recipe website, and Safari (or Chrome) just display the website to grandma (and to the officer on duty, but who cares) - how long will Firefox survive?
  - raverbashing 3y ago
    Yes, I agree. The crying wolf is too much sometimes. Accepting certificates from a given issuer does not give them the issuer the right to impersonate others
    - g-b-r 3y ago
      All root CAs can issue certificates for any site (except those with CAA records etc.)
  - no_time 3y ago
    Luckily, I never said anything like this anywhere.
- gallexme 3y ago
  Wouldn't a client certificate from e2echat protect that kind of attack ? Since even when a man in the middle offers u a server cert u accept, the e2echat servers can't validate the client certificate from you anymore (Still bad but would at least protect connections from ever talking to e2echats servers)
  - Filligree 3y ago
    Nobody uses client certs.
g-b-r 3y ago
Oh yeah if encryption is broken only for browsers no big deal right
- Jensson 3y ago
  Governments still can't see your requests to servers under normal circumstances with this law. The weakness is only if someone controls your internet connection and can use a compromised certification process to trick you into thinking you are at "e2e.com" when you are on another site, and in those cases the only difference from now is that your browser will display "secure" instead of "invalid cert". There is no other difference. So to orchestrate an attack they would need to build an webbapp that is sufficient similar for you not to notice, take over your internet connection and break the certification process.
  - calgoo 3y ago
    "The weakness is only if someone controls your internet connection and can use a compromised certification process to trick you into thinking you are at e2e.com" That will be (or already is) done at ISP level. It will probably be fully automated, where they just put a court order number into a form, and it automatically just catches all your traffic in gear that's installed at the ISP.
    - Jensson 3y ago
      It is only undetectable if the site actually uses the vulnerable certificates. Otherwise you can see that the government is spying on you since the browser tells you what certificate it got (Telling you what certificate was used is a part of eIDAS). There is no way the government will replace certificates like that on an automated basis, it is too easy for people to notice and make a big deal about.
      - supriyo-biswas 3y ago
        If a nonprofit like Let’s Encrypt can perform automated certificate renewal with a few API calls, so can the government. Also, MITMs are a thing and getting the EIDAS certs in the root store will show that the certs in question are trusted, which is all that really matters because there is no way for users to know what certificates were actually installed by the website owner.
isilofi 3y ago
There is no way for e2echat.com to make sure that the client will insist on a certain safe CA. Sure, in case e2echat.com controls all clients this would be possible, but this is a rare case. In the general case, any CA can sign any website certificate. So all those new government CAs can sign all the man-in-the-middle certificates they like, and browsers are obliged to accept them. Nothing the website can do about that. There are ways to pin certain CAs via DNSSEC and TLSA resource records in DNS. But browsers ignore those, and even if they didn't, the same EU proposal also specifies government DNS manipulation. So the gist is: EIDAS must die.
- Jensson 3y ago
  You still wont be able to break the end to end encryption of a site. You can only intercept traffic that the server can read, you can't intercept traffic that are encrypted end to end. And if the site can see your data assume the government can see it as well, they can get it with a warrant.
  - isilofi 3y ago
    Website-based end-to-end encryption isn't usually. In most cases, the "e2e-encrypting" website will deliver the Javascript that does the "e2e-encryption", which can easily be manipulated to provide a copy of all messages to some convenient third location. A warrant will maybe warn the site and the user that something is going on. A man-in-the-middle attack without a warrant delivered to either party is more likely to go undetected.
    - Jensson 3y ago
      > which can easily be manipulated to provide a copy of all messages to some convenient third location. Updating others javascript as a proxy isn't "easily". Also if the government goes all this way to tell each internet provider to spy on people, why do you think they couldn't tell certificate authorities to spy on people? It is the same level. I wouldn't be surprised if many CA's in USA already does this.

no_time 3y ago

If "e2echat.com" has no method to explicitly forbid your browser from accepting eIDAS certs (via a DNS record or something) then your browser will just blindly accept the compromised cert when attacked. This is still very bad.

Jensson 3y ago
> This is still very bad. Yes, potentially, but it isn't "another kind of chat control".
- g-b-r 3y ago
  It's another side of the efforts of going around encryption, chat controls deals with communication services, this one with browsers
  - Jensson 3y ago
    But this doesn't force browsers and sites to use weak encryption. It is very different.
    - g-b-r 3y ago
      This forces browsers to accept all the CAs approved by the EU states, and you can be certain that some of them will be used for decrypting (and if needed modifying) the traffic
      - Jensson 3y ago
        And then you can just tell the browser to not trust those CAs and you are safe. This is nothing like "chat control". This only lets the government spy on people who don't care if the government spies on them.
        yaris 3y ago
        IIRC one cannot tell the browser to not trust root CAs, that's why all the fuss.
        Jensson 3y ago
        Why shouldn't you be able to do that? Seems like a simple thing to implement. I get why they want a hardcoded list, but I don't get why you can't add a way to block parts of that hardcoded list.
        g-b-r 3y ago
        web-browsers shall ensure
        Jensson 3y ago
        The only requirement is that browsers displays the data. The browser can add "warning, this certificate is potentially compromised" when it displays it, nothing in the current document says browsers aren't allowed to say that, just that the browser has to be aware of the certificate. It is similar to how Chrome displays a warning when you visit some sites. You can visit the site anyway, but you get a warning since Google thinks it is bad.
        yaris 3y ago
        Technically correct. But if Firefox displays a big red warning when someone's grandma goes to her favourite recipe website, and Safari (or Chrome) just display the website to grandma (and to the officer on duty, but who cares) - how long will Firefox survive?
- raverbashing 3y ago
  Yes, I agree. The crying wolf is too much sometimes. Accepting certificates from a given issuer does not give them the issuer the right to impersonate others
  - g-b-r 3y ago
    All root CAs can issue certificates for any site (except those with CAA records etc.)
- no_time 3y ago
  Luckily, I never said anything like this anywhere.
gallexme 3y ago
Wouldn't a client certificate from e2echat protect that kind of attack ? Since even when a man in the middle offers u a server cert u accept, the e2echat servers can't validate the client certificate from you anymore (Still bad but would at least protect connections from ever talking to e2echats servers)
- Filligree 3y ago
  Nobody uses client certs.

g-b-r 3y ago

Oh yeah if encryption is broken only for browsers no big deal right

Jensson 3y ago
Governments still can't see your requests to servers under normal circumstances with this law. The weakness is only if someone controls your internet connection and can use a compromised certification process to trick you into thinking you are at "e2e.com" when you are on another site, and in those cases the only difference from now is that your browser will display "secure" instead of "invalid cert". There is no other difference. So to orchestrate an attack they would need to build an webbapp that is sufficient similar for you not to notice, take over your internet connection and break the certification process.
- calgoo 3y ago
  "The weakness is only if someone controls your internet connection and can use a compromised certification process to trick you into thinking you are at e2e.com" That will be (or already is) done at ISP level. It will probably be fully automated, where they just put a court order number into a form, and it automatically just catches all your traffic in gear that's installed at the ISP.
  - Jensson 3y ago
    It is only undetectable if the site actually uses the vulnerable certificates. Otherwise you can see that the government is spying on you since the browser tells you what certificate it got (Telling you what certificate was used is a part of eIDAS). There is no way the government will replace certificates like that on an automated basis, it is too easy for people to notice and make a big deal about.
    - supriyo-biswas 3y ago
      If a nonprofit like Let’s Encrypt can perform automated certificate renewal with a few API calls, so can the government. Also, MITMs are a thing and getting the EIDAS certs in the root store will show that the certs in question are trusted, which is all that really matters because there is no way for users to know what certificates were actually installed by the website owner.

isilofi 3y ago

There is no way for e2echat.com to make sure that the client will insist on a certain safe CA. Sure, in case e2echat.com controls all clients this would be possible, but this is a rare case. In the general case, any CA can sign any website certificate. So all those new government CAs can sign all the man-in-the-middle certificates they like, and browsers are obliged to accept them. Nothing the website can do about that. There are ways to pin certain CAs via DNSSEC and TLSA resource records in DNS. But browsers ignore those, and even if they didn't, the same EU proposal also specifies government DNS manipulation. So the gist is: EIDAS must die.

Jensson 3y ago
You still wont be able to break the end to end encryption of a site. You can only intercept traffic that the server can read, you can't intercept traffic that are encrypted end to end. And if the site can see your data assume the government can see it as well, they can get it with a warrant.
- isilofi 3y ago
  Website-based end-to-end encryption isn't usually. In most cases, the "e2e-encrypting" website will deliver the Javascript that does the "e2e-encryption", which can easily be manipulated to provide a copy of all messages to some convenient third location. A warrant will maybe warn the site and the user that something is going on. A man-in-the-middle attack without a warrant delivered to either party is more likely to go undetected.
  - Jensson 3y ago
    > which can easily be manipulated to provide a copy of all messages to some convenient third location. Updating others javascript as a proxy isn't "easily". Also if the government goes all this way to tell each internet provider to spy on people, why do you think they couldn't tell certificate authorities to spy on people? It is the same level. I wouldn't be surprised if many CA's in USA already does this.