|
@depletionmode | |||||
|
My latest post: Understanding modern UEFI-based platform boot depletionmode.com/uefi-boot.html
Complete with some possibly half-baked rambling thoughts on DRTM at the end pic.twitter.com/5kpOGNUJGs
|
||||||
|
||||||
|
Peter Bosch
@peterbjornx
|
15. kol |
|
@qrs and I actually did discover TOCTOUs in the IBB phase :conference.hitb.org/hitbsecconf201… github.com/tianocore/edk2…
|
||
|
|
||
|
David Kaplan
@depletionmode
|
15. kol |
|
Well there we go! Great work. I'll add the link to your research to the blog post.
|
||
|
|
||
|
Graham Sutherland [Polynomial^DSS]
@gsuberland
|
15. kol |
|
Nice article, although is there any chance you could bump up the contrast on the text on your site? My eyes aren't what they used to be.
|
||
|
|
||
|
David Kaplan
@depletionmode
|
15. kol |
|
Thanks!
Ye so sorry. Reading on mobile is a disaster (unless you're speaking about PC?). I am going to address this asap
|
||
|
|
||
|
vincent zimmer
@vincentzimmer
|
14. kol |
|
Nice write up. I’m a fan of modeling systems using Clark Wilson. We have some examples of this regarding the boot chain in legacy.gitbook.com/book/edk2-docs…. You can find other docs on EDKII security at github.com/tianocore/tian… including some threat modeling in legacy.gitbook.com/book/edk2-docs…
|
||
|
|
||
|
David Kaplan
@depletionmode
|
14. kol |
|
Thanks so much Vincent.
I stole the Clark Wilson thing from you of course. :) (From somewhere in the Intel BGT flow docs).
Thanks for the resources. Super helpful.
|
||
|
|
||
|
chris
@hugelgupf
|
15. kol |
|
Agree on why DRTM is suboptimal. Disagree on the SRTM reasoning:
"The initial trust is rooted in the CPU+chipset vendor."
That doesn't have to be true for SRTM - see Microsoft's Cerberus, or Google's Titan (ai.google/research/pubs/… et al)
|
||
|
|
||
|
David Kaplan
@depletionmode
|
15. kol |
|
Thanks for pointing this out. I agree with you.
I was speaking about SRTM as implemented in the run of the mill off-the-shelf Windows-based machine. I will amend the blog to make it clear that the root-of-trust is not necessarily the cpu vendor.
|
||
|
|
||
|
ralph waldo cybersyn
@atomicthumbs
|
15. kol |
|
i absolutely love the implication in the opening sentence that most of us use voodoo extensively
|
||
|
|
||
|
David Kaplan
@depletionmode
|
15. kol |
|
Not far off the truth here.. :)
|
||
|
|
||