Re: [tboot-devel] LCP index definitions

2009-07-21 Thread Shane Wang
Hi Martin, Sorry for that. We are looking at different versions. I can't find it in version 2008-June, either. For what you mentioned, it is not clear in the current spec. We will mention the definitions according to what you pointed out in the next release of the spec, which will be publish

Re: [tboot-devel] Bug in find_mle_hdr(), mlehash.c

2009-07-21 Thread Shane Wang
Hi Michael, Thank you for pointing out the potential issue. Here is the fix for it. Fix the potential segmentation fault in find_mle_hdr, when size%sizeof(uuid_t)!=0 where size is unsigned long. Signed-off-by: Shane Wang diff -r ad96c7e8bf5a lcptools/mlehash.c --- a/lcptools/mlehash.cTu

Re: [tboot-devel] Intel's P-MAPS research project

2009-07-21 Thread Lil Evil
There are many different projects with similar goals out there: BitVisor(sourcecode available somewhere) or Daonity and of course flickr, probably more that I am not aware of. They all seem to target a particular use case and scenario. Cutting out Operating System is certainly an elegant and inte

Re: [tboot-devel] LCP index definitions

2009-07-21 Thread Martin Pirker
Shane Wang wrote: > The index values are the protocol with the SINIT module, consulting with > the SINIT. The tools in tboot write launch control policy to those > addresses in TPM NV, and during machine bootup, the SINIT reads policies > from those addresses by default. You can get some info about

[tboot-devel] Calculation of PCR 18 - Error in doc

2009-07-21 Thread Michael Gissing
Hi! I tried to calculate the final value of PCR 18 by paper and pen, it seems that tboot README is wrong about that. In section "PCR Usage" it says that tboot policy will also be extended to PCR 18, that's wrong. PCR 18 is calculated only by: 1) extend hash of tboot (as measured by lcp_mlehash)

[tboot-devel] Bug in find_mle_hdr(), mlehash.c

2009-07-21 Thread Michael Gissing
Hi! This is just a minor issue, but I want to share it with you ;-) file mlehash.c, line 311: size is a size_t (typedefed unsigned long), so if "size%sizeof(uuid_t) != 0", size will _always_ be >0, the loop won't exit and you'll get a segfault. I've got a question too: How do you ensure that t