Hi everybody, bit of a warning here: The recovery key generated during the installation of Ubuntu 23.10 (if you select tpm-backed fde) cannot be used to unlock the disk outside of boot, as in any ‘cryptsetup’ command and so on will not accept the recovery key. unlocking when accessed from different system does not work etc.

You can use it to unlock the disk while booting if your tpm somehow fails, but ONLY in that specific situation.

I kind of purposefully broke my tpm keys to see if it could be restored with 23.10 and ended up having to reinstal, as I ended up having to enter the recovery key at boot every time and no way of adding additional unlock options to the volume, as cryptsetup would not accept the recovery key as passphrase.

This bug could be very bad for new users.

See this bug report: https://bugs.launchpad.net/ubuntu-desktop-installer/+bug/2039741

  • @mvirts
    link
    167 months ago

    Yet another reason why I wouldn’t put anything important on a tpm encrypted volume 😹 I just don’t trust it

      • @ichbinjasokreativOP
        link
        37 months ago

        That’s true, but the issue is that cryptsetup does not accept the recovery key as a passphrase for the disk. Once the tpm gets reset, the user has to always enter the recovery key and cannot implement a new key to luks and the tpm.

          • @ichbinjasokreativOP
            link
            27 months ago

            I wasn’t blaming cryptsetup, my mistake if it came across as though I did.

            Thank you for taking the time to look into a possible fix, I might just reinstall with tpm-backed fde again to see if this really works.

          • @MAFoElffen
            link
            1
            edit-2
            7 months ago

            @Skull giver – I mentioned this above, but couldn’t link it to you.

            I like the code but the go run recover.go 2> key.text does not redirect the key to a file. It does not get input of the recovery key, so errors. Could you please add a few lines to write it directly to that file, instead of displaying it? (or output to both?)

            As someone thought, what is displayed onscreen is jibberish, because console cannot display raw hex characters… I’m thinking the LUKS key in the keyslot is raw or hex.

            As I now know, it is translated. And we can get the recovery key (hopefully) translated the other way around…

            If I can just find the valid key value, and write it to a file, then I can help @inchbinjasokreativ to write it back to his TPM. I’ve already written a BASH script to do that, but am just missing that key-file. I have the problem replicated to a VM, so can test it on that first.

              • @MAFoElffen
                link
                37 months ago

                Success! I have a working key file…

                root@ubuntu:/home/ubuntu/Downloads# cryptsetup -v luksOpen /dev/sda3 luks-01 --key-file ./key-file.key
                No usable token is available.
                Warning: keyslot operation could fail as it requires more than available memory.
                Key slot 1 unlocked.
                Command successful.
                root@ubuntu:/home/ubuntu/Downloads# cryptsetup -v luksOpen /dev/sda4 luks-02 --key-file ./key-file.key
                No usable token is available.
                Warning: keyslot operation could fail as it requires more than available memory.
                Key slot 1 unlocked.
                Command successful.
                

                Success on the first volume, which I picked as first because it was only 53M in size. Mounted it to /mnt… And guess what I found inside it?

                root@ubuntu:/home/ubuntu/Downloads# ls -l /mnt/device/private-keys-v1/
                total 4
                -rw------- 1 root root 2459 Oct 18 18:29 O8CbAEpnfm7jGKkMqnokmdMBlE1oV6Xma_bUNudlshDYPxE4aJNhbhiGnF360Ze4
                

                That is a key, but not connected to either LUKS container there… I dumped the headers of both LUKS. There are 2 key-slots, and the key translated from the recovery key is in slot one of both containers, The second key-slot’s key must be the TPM’s key, which is unknown if that is stored anywhere except the TPM…

                But is shouldn’t matter now… Because that key-file did work to add a new passphrase to both LUKS containers.

                Thank you @Skull giver.

      • @kautau
        link
        1
        edit-2
        7 months ago

        Yeah I’m hybrid Windows/Linux user, but many of my drives are Bitlocker encrypted. I need to install a bios update. To do so, it requires me to decrypt every bitlocker drive and not just the OS drive. Because the TPM keystore is based on the OS key store for each bitlocker drive. It’s frustrating, but it makes sense, so I’m all for additional security

  • @MAFoElffen
    link
    12
    edit-2
    7 months ago

    As you can see by my UserName, I am the person who filed that Bug as a danger to users… Specifically brought up by @ichbinsokreativ, me trying to help him…

    Thank you Skull giver for the golang code, I had to debug your code, to get it to run, because of the forum transposing HTML char codes for some characters… Unfortunatley, displaying the result onto console displays ‘jibberish’, as you thought might happen. Console cannot display raw hex characters.

    Redirecting the output as you posted doesn’t work, as the script then doesn’t get the input of the recovery key, so errors.

    I have a request… I know a lot of languages, but GO isn’t one of them. Please… Could you please add a few lines to write the result directly to a file called recovery.key? Then if raw or hex, it would get to a key-file… Then I can test if that is going to work to add additional keys to the LUKS containers, and to be able to help people re-enroll the TPM key.

    If you could, they we would have a recovery work-around.

    Yes. They did something similar for past ZFS encryptions i their canned installs, but used native ZFS encryption, with a locked encrytped keyfile stored within a LUKS container, that had to be unlocked and mounted before unlocking the ZFS pools. Sometimes I don’t follow the logic behind some things.

    It seems like they often add complication to what should be simpler.

  • @[email protected]
    link
    fedilink
    27 months ago

    As a new user, TPM -backed FDE is above my head, as concept alone. I doubt newbros are getting hung up on this.

    Still, nice to know.

    • @ichbinjasokreativOP
      link
      17 months ago

      the tpm basically gets registered as a valid way to decrypt the luks partition that all your data is in. It does so by first checking if the system is authorized to receive the keys. If the key in the tpm gets deleted or invalidated, or if the system signature changes because of a firmware update (shouldn’t really happen), then the tpm will no longer unlock the partition. I’m no expert either though, just how I understand it. TPM-backed Full Disk Encryption has become more mainstream after microsoft made it a requirement for windows 11 and the option is right there in the installer, it’s also one of the selling points of 23.10. Yes, it’s clearly marked as experimental, but people will do what people do and thus they should know of the recovery key issue. Currently, if your tpm does not unlock your partition, you HAVE to enter the recovery key every time you boot, or reinstall Ubuntu.

  • @stealthnerd
    link
    17 months ago

    Okay so I just read up on this. It’s it true that TPM backed FDE only allows snaps?!?

    Debs are completely unsupported?

    • @ichbinjasokreativOP
      link
      17 months ago

      The ubuntu implementation is based on snap, but if you want to purge snapd (just install another distro at that point though), you can implement encryption as usual and add the tpm later on. Snap is only neccessary if you enable tpm-backed fde from setup.

  • @erik1984
    link
    1
    edit-2
    3 months ago

    deleted by creator

    • @ichbinjasokreativOP
      link
      27 months ago

      it’s not, because to to so you need to enter a valid passphrase to cryptsetup, which the recovery key is NOT