There exists shims that are signed by Microsoft and were not revoked. Normally this would be fine but these shims had weaknesses that allowes hackers to load any code using them. Normally the shims should only run other signed/trusted code. These vulnerable shims can be used to bypass secure boot by replacing your existing bootloader with the shim and then running rootkits/hackerOS/whatever and bypass bitlocker using TPM or just running a level 0 virus that can’t be detected by the OS on any PC which trusts Microsoft’s keys (99% of all PCs)

To prevent this you’d have to not trust the vulnerable shims by either adding them manually to the exclusions list or using your own secure boot keys which would only trust the few bootloader files your pc uses and no other files.

Worst case: it behaves as if secure boot wasn’t on. Without secure boot you wouldn’t need this exploit cause then you can replace the bootloader with whatever you want anyways. With or without secure boot you need administrative permission to replace the bootloader so this is only an issue after your PC is already compromised or if someone had physical access to your PC.

tldr: Either use your own keys or don’t trust secure boot.

I have a Server with ~16 podman services, each their own user, network namespace and uids. This is managed using NixOS and Home manager (which supports quadlets) but I am changing my setup to a single node k3s cluster with user namespaces because that seems simpler to manage. Here a snippet for how the subuids/subuids are defined:

users.users.<username> = {
        subUidRanges = [{
            startUid = 100000+65536*( config.users.users.<username>.uid - 999);
            count = 65536;
        }];
        subGidRanges = [{
            startGid = 100000+65536*( config.users.users.<username>.uid - 999);
            count = 65536;
        }];
        home = "[...]";
        isNormalUser = true;
        linger = true;
        group = "users";
        openssh.authorizedKeys.keys = config.users.users.root.openssh.authorizedKeys.keys;
    };

This is the year of the OpenAI bankruptcy.

The fix is upstream commit 3bfdc63936dd, landed in April, and distributions are rolling it out now.

Just found a recording of the last meeting he attended: https://youtu.be/I08EpZU8z8M

https://youtu.be/tU48ouWNr8E

Now is the time to move fast and break things, do it!

Fair enough, I meant consumer demand as in not ai and data centers

Ram demand expected to drop another 40-50% in Q3?

How are you running podman? As a rootful systemd service or as a rootless user systemd services?

If you have an openwrt compatible router then you can use that as the tang server.

Also you can tell clevis to use both tpm and tang and require both to work.

So when the bubble pops they’ll just stop selling while the other sellers continue?

Yeah I don’t think so, unless all sellers “lock in” their prices for 5 years it isn’t happening and they’re just saying this to get more companies to buy at high prices. If all of them do “lock in” their prices then this will probably be the second time the memory cartel is fined.

All things are solved by lowering your standards far enough.

Oh, cry me an ocean

Tbf, he said no one should go in there for a few hours, can he really be blamed for their actions?

…where do you think the salt goes when the water evaporates?

You sure it wasn’t cat?

I know you can use a root shell with cat image.ISO > /dev/[...] for disk images/ISO in disk image Format (so most ISO that aren’t Windows) but using the cp command would be news for me.