/tech/ - Technology★

muh heat

Nothing is stopping you. But if you did something like that then TLA's would fork your project and use it for black box making. The only thing of worth to put on the ME chip would be power management related drivers or clocks anyways. Or just running the whole OS on it's ring like a linux instance that keeps kernel mode in the ME and userland on the main processor or something stupid like that.

Halo is shit.

MINIX tbh fam

>>828900

License it as GPLv3 and proceed to sue the fuck out of everyone.

fill it with botnet and sell it on ebay

>>829080

>minix

>not gnu/hurd

>not even templeOS

pathetic

>>828877 (OP)

OpenWRT/LEDE would fit the bill nicely as a distro. With the right software stack It could potentially be used to add full lights-out capability to boards that didn't come with this feature.

>>829135

Librecmc.

>>829135

you could even use it as a nice utility to be able to control the computer over the network and troubleshoot problems even when the OS doesn't post.

Put GNU Hurd on it.

>useful software

>Linux

Pick one.

>>829169

This, if you choose linux as opposed to something useful like Plan9 you're doing something wrong.

>>829207

>Plan 9

>useful

lel enjoy your no drivers

>>828877 (OP)

nothing

>>828900 is wrong, there is something stopping you

public key cryptography. chip requires signed firmware. if you dont have the signing key, you are fucked

>>829161

That's what I meant by lights-out capacity.

Intel AMT out-of-band management is a premium feature that commands a price. Investigations of ME have revealed that the necessary hardware is present in all Intel systems. To take control of it, remove the potential for alphabet agency implants, and use it to add functionality that Intel would charge for would be a satisfying "fuck you" to Intel for kikeing our systems in the first place.

OpenWRT/LEDE is about the most minimal Linux distro possible, totally user-configurable and would fit in the SPI flash. If we could manage to get it (or the GNU de-blobbed version, Librecmc, as >>829156 suggested) installed in place of the ME kikeware, then the next challenge would be finding or writing the software to redirect keyboard, mouse and video to be served over the network as some remote protocol. I suspect SPICE would be most suitable for this application.

>>829217

>being too lazy to make your own drivers

>>829247

There may be a way around that. See >>818210

>>829324

What are you going to do to close the door you enter through? Any good hacker closes the backdoor as he enters. Would be nice to burn fuses in order to store your own private key.

>>829324

Doesn't include chipsets before skylake yet. Aside from persistent ring -3 rootkits in Core2 hardware.

>>829434

Would have to know more about the exploit to answer that. They're doing a public presentation in a week or so, until then this is just speculation.

>>829435

Well, at least it will be nice if we're able to de-kike current hardware rather than ancient stuff.

>>829135

This is actually a good idea, this and the "power management drivers or clocks".

Does Intel ME even have 4MB of storage? (openwrt/lede require 4MB). If it does you won't even have to use the previously botnet features to pull the OS from the hard drive.

>>829713

>Does Intel ME even have 4MB of storage?

I believe most boards have an 8MB SPI flash which contains both the UEFI code and the ME code. ME was about 5MB, so big enough. Presumably would be replacing the UEFI with Coreboot/Libreboot, which might free up more room.