Please review. Thanks Xiaokai
-----Original Message----- From: Wang, Xiaokai Sent: Monday, March 25, 2013 4:58 PM To: 'Kent Yoder' Subject: RE: [TrouSerS-tech] patches about tpm_getpermanentflags and tpm_nvdefine commands 1) Initializing askOwnerPass to TURE. Done 2) logically & UNIT32 with its flags. Done. Work well on x86 processor. Having no power architecture processor,so,test is not be done on which. Sorry. 3) Adding a condition to judge whether hPcrsRead object exist. Please review. Thanks again. Regards Xiaokai -----Original Message----- From: Kent Yoder [mailto:[email protected]] Sent: Saturday, March 23, 2013 5:13 AM To: Wang, Xiaokai Subject: Re: [TrouSerS-tech] patches about tpm_getpermanentflags and tpm_nvdefine commands Hi Xiaokai, On Mon, Mar 18, 2013 at 12:45 AM, Wang, Xiaokai <[email protected]> wrote: > Hi kent, > > The patches are ok? A couple things: 1) when run w/ no arguments, tpm_getpermaentflags segfaults. It looks like initializing askOwnerPass to TRUE will fix this 2) bitfields like tpm_perm_flags_t and tpm_stclear_flags_t do not work on the power architecture. The fix is to logically & the UINT32 with its flags. Please convert these. These flags are in tpm.h already in trousers: #define TPM_PF_DISABLE ((UINT32)0x00000001) #define TPM_PF_OWNERSHIP ((UINT32)0x00000002) #define TPM_PF_DEACTIVATED ((UINT32)0x00000003) ... and so on.. 3) I have a questino on tpm_nvdefine's changes. It looks like the new code (after your patch) will overwrite the hPcrsRead/hPcrsWrite objects. For instance if the user uses -f to pass in a list of pcr values, the hPcrsRead object will get created in parseNVPermsFile(), then its object handle will be overwritten by contextCreateObject() just before you test "if(w_loc_flag == 1)". Try to restructure the code so that the localities can be set regardless of how the pcr values are set )if at all). Thanks, Kent > > Thanks, > Xiaokai > > -----Original Message----- > From: Kent Yoder [mailto:[email protected]] > Sent: Tuesday, March 05, 2013 4:04 AM > To: Wang, Xiaokai > Cc: Stefan Berger ([email protected]); > [email protected] > Subject: Re: patches about tpm_getpermanentflags and tpm_nvdefine > commands > > Hi Xiaokai, > > On Mon, Mar 4, 2013 at 4:15 AM, Wang, Xiaokai <[email protected]> wrote: >> Hi all, >> >> >> >> Attach files are the patches I made about tpm-tools. > > Thanks for the patches. It looks like they didn't get formatted correctly > for git though. You can use git format-patch on your local repository to > export patches. Please try this and resubmit. > >> I change the name tpm_getcap to tpm_getpermanentflags,so that don't >> need any other arguments to display TPM_PERMANENTFLAGS. >> >> >> >> One probem is below: >> >> + if (genericOptHandler >> >> + (argc, argv, "o:", hOpts, >> >> + sizeof(hOpts) / sizeof(struct option), parse, >> + help) != >> 0) >> >> + goto out; >> >> >> >> "o:" cann't get rid of ":", otherwise the argument ownerpassword >> value will be 0x0 and the cmd will implement abnormal. >> >> The reason I am not very clear,please help debug that. > > Please take a look at how tpm_nvdefine.c handles the owner password. > The argument for ownerpass should be optional. > >> >> >> If the patches have no other problems I hope you can apply them. Thank you! > > Your patch to man/man8/tpm_nvdefine.8 made me realize that the .8 manpages > with .pod extensions don't even need to be in git, since they're regenerated > at build time. I've removed them entirely in the "1.3.9" branch in git. All > you need to do is patch the .pod file here. > > Thanks, > Kent > >> >> >> >> >> Regards >> >> Xiaokai >> >> ------------------------------------------------------------------------------ Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis & visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter _______________________________________________ TrouSerS-tech mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/trousers-tech
