x86_64 and ARM 64 memory leak drama

[PDP-8]

Ok, like many others, I'm keeping my eye on the whole major design flaw of the X86_64 cpu thing which seems to encompass Intel and now even possibly the ARM V8 chips....

While the patches are flying fast and furious, and it seems inevitable that performance issues are cropping up, my major question is:

Can this issue be avoided in the short run by running a 32-bit kernel on the 64-bit hardware? -- or is that too simplistic a solution?...

[vinceASPECT]

hmm.....i heard it on the news.  Memory leaks hu?

a flaw that's been dormant for over a decade....gosh

V

[PDP-8]

Well, we've been through this before and my hope is that it will spur new research to break the typical "intrude/patch" cycle endlessly.

This is something the Roger Schell had to deal with at the NSA, Novell, Multics, and other earlier computers in the cold-war era militarized computers.

Real interest in computer security was lost long ago for various reasons.  For anyone interested, do a web grep for:

"Oral history interview with Roger R Schell"

It is from 2012 and is a fascinating read, albeit long.  I sure wish computer vendors  / researchers would read it rather than be satisified with the intrude/patch cycle.

[curaga]

No, running a 32-bit kernel on 64-bit hw won't protect from that. But even if you have an Intel cpu, you don't have to get the patch and drop your performance if you only run trusted software.

[vinceASPECT]

They say the patch gives maybe  a typical 15 percent performance loss of your cpu.

hmmm

V

[PDP-8]

And once that's fixed, then there is the internal "management engine / AMT" (which we and the OS don't see).  I forgot where, but someone said that it is actually Minix inside that..