Rehaul of firmware TPM tests

Add new files: two tests, one common file, one program to set things up.

Change-Id: I4c9b7a937103f3978cbed6629ee4057018b80eae

More cleanup.  Also allow some tests to run even when TPM is already started.

Change-Id: I23558b96a1de55bbeca42dbf2e44f6802a0ec85b

Reorganize and standardize behavior of tests.

Change-Id: Id32fd09211a72deaa66a3dd0f973d35506ff96f2

BUG=433
TEST=ran all the tests I could run without TPM-free BIOS

[gauravsh, 2010-09-15 08:42:13]

What is PROF(). Doesn't seem descriptive enough.

http://codereview.chromium.org/3389004/diff/4001/5008
File tests/tpm_lite/enable.c (right):

http://codereview.chromium.org/3389004/diff/4001/5008#newcode28
tests/tpm_lite/enable.c:28: if (disable == 1 || deactivated == 1) {
slightly confusing. SetEnable doesn't take an option but SetDeactivated does.
Why?

http://codereview.chromium.org/3389004/diff/4001/5017
File tests/tpm_lite/timing.c (right):

http://codereview.chromium.org/3389004/diff/4001/5017#newcode21
tests/tpm_lite/timing.c:21: /* Run OP and ensure it returns success and doesn't
run longer than TIME_LIMIT
nit: |op| or [op]

[Luigi Semenzato, 2010-09-15 16:27:51]

Thanks.  I have changed PROF (Print and Return On Failure, as the comment said)
to the more descriptive PEROF (Print Error and Return On Failure).  This is used
everywhere so I'd rather keep it short.  I know it doesn't scale, but it doesn't
have to.

http://codereview.chromium.org/3389004/diff/4001/5008
File tests/tpm_lite/enable.c (right):

http://codereview.chromium.org/3389004/diff/4001/5008#newcode28
tests/tpm_lite/enable.c:28: if (disable == 1 || deactivated == 1) {
On 2010/09/15 08:42:13, gauravsh wrote:
> slightly confusing. SetEnable doesn't take an option but SetDeactivated does.
> Why?

For transparency, I am usually staying as close as possible to the TPM commands.
 Also notice this other difference: disable (adjective) and deactivated (past
participle).  That's all in the TCG specs.

http://codereview.chromium.org/3389004/diff/4001/5017
File tests/tpm_lite/timing.c (right):

http://codereview.chromium.org/3389004/diff/4001/5017#newcode21
tests/tpm_lite/timing.c:21: /* Run OP and ensure it returns success and doesn't
run longer than TIME_LIMIT
On 2010/09/15 08:42:13, gauravsh wrote:
> nit: |op| or [op]

Thanks, done.

[gauravsh, 2010-09-15 22:09:46]

LGTM

not too fond of either of PROF or PEROF. Not in common use, so sounds exotic. I
don't have a better alternative though, so it's fine to keep it as is.

On 2010/09/15 16:27:51, Luigi Semenzato wrote:
> Thanks.  I have changed PROF (Print and Return On Failure, as the comment
said)
> to the more descriptive PEROF (Print Error and Return On Failure).  This is
used
> everywhere so I'd rather keep it short.  I know it doesn't scale, but it
doesn't
> have to.
> 
> http://codereview.chromium.org/3389004/diff/4001/5008
> File tests/tpm_lite/enable.c (right):
> 
> http://codereview.chromium.org/3389004/diff/4001/5008#newcode28
> tests/tpm_lite/enable.c:28: if (disable == 1 || deactivated == 1) {
> On 2010/09/15 08:42:13, gauravsh wrote:
> > slightly confusing. SetEnable doesn't take an option but SetDeactivated
does.
> > Why?
> 
> For transparency, I am usually staying as close as possible to the TPM
commands.
>  Also notice this other difference: disable (adjective) and deactivated (past
> participle).  That's all in the TCG specs.
> 
> http://codereview.chromium.org/3389004/diff/4001/5017
> File tests/tpm_lite/timing.c (right):
> 
> http://codereview.chromium.org/3389004/diff/4001/5017#newcode21
> tests/tpm_lite/timing.c:21: /* Run OP and ensure it returns success and
doesn't
> run longer than TIME_LIMIT
> On 2010/09/15 08:42:13, gauravsh wrote:
> > nit: |op| or [op]
> 
> Thanks, done.

[Luigi Semenzato, 2010-09-15 22:44:41]

You're perfectly right.  I renamed it TPM_CHECK.

On 2010/09/15 22:09:46, gauravsh wrote:
> LGTM
> 
> not too fond of either of PROF or PEROF. Not in common use, so sounds exotic.
I
> don't have a better alternative though, so it's fine to keep it as is.
> 
> On 2010/09/15 16:27:51, Luigi Semenzato wrote:
> > Thanks.  I have changed PROF (Print and Return On Failure, as the comment
> said)
> > to the more descriptive PEROF (Print Error and Return On Failure).  This is
> used
> > everywhere so I'd rather keep it short.  I know it doesn't scale, but it
> doesn't
> > have to.
> > 
> > http://codereview.chromium.org/3389004/diff/4001/5008
> > File tests/tpm_lite/enable.c (right):
> > 
> > http://codereview.chromium.org/3389004/diff/4001/5008#newcode28
> > tests/tpm_lite/enable.c:28: if (disable == 1 || deactivated == 1) {
> > On 2010/09/15 08:42:13, gauravsh wrote:
> > > slightly confusing. SetEnable doesn't take an option but SetDeactivated
> does.
> > > Why?
> > 
> > For transparency, I am usually staying as close as possible to the TPM
> commands.
> >  Also notice this other difference: disable (adjective) and deactivated
(past
> > participle).  That's all in the TCG specs.
> > 
> > http://codereview.chromium.org/3389004/diff/4001/5017
> > File tests/tpm_lite/timing.c (right):
> > 
> > http://codereview.chromium.org/3389004/diff/4001/5017#newcode21
> > tests/tpm_lite/timing.c:21: /* Run OP and ensure it returns success and
> doesn't
> > run longer than TIME_LIMIT
> > On 2010/09/15 08:42:13, gauravsh wrote:
> > > nit: |op| or [op]
> > 
> > Thanks, done.