On 02/18/2017 01:26 PM, Jonathan Neuschäfer wrote:
On Fri, Feb 17, 2017 at 08:24:49PM -0500, Taiidan@gmx.com wrote: [...]
I will be greatly impressed if the 30 minute bullshit is somehow bypassed.
I guess you haven't seen Nicola Corna's recent work on the me_cleaner tool, then:
- https://review.coreboot.org/cgit/coreboot.git/tree/util/me_cleaner
- https://github.com/corna/me_cleaner/wiki/How-does-it-work?
Jonathan
Nerfing and disabling are different things. So are disabled, and removed all together (remove the special ARC core from the CPU package)
FYI so everyone knows despite popular belief (again spread by purism types) the MEI device in lspci/device manager being present or not has no bearing on if it is on, off or disabled. There isn't any way to tell for real if it is actually disabled and not DMA capable without removing the ME firmware and somehow analyzing the ME processor to see if it does anything.
I for one don't think that it is possible to truly disable it (permanently) without a team of experts and millions in equipment and research (per platform), at this point the only real options for libre devices is POWER and ARM (RISC-V doesn't have any hardware with real juice yet). This hasn't happened yet because even linux people are addicted to x86-64.
Conspiracy section: The 30 minute thing begs the question of why does intel care so much about making sure people have ME functional? the corporate manageability excuse goes out the window when you consider the fact that it is present on the non-vPro series processors as well. It makes no sense to me.