vboot_reference: add creation of GBB blob to gbb_utility

vboot_reference: add creation of GBB blob to gbb_utility

Non-x86 platforms (ex, ARM) may need to create GBB from scratch.
The new "create" (-c) creates a GBB blob by given property maximum sizes.
The list must follow same order as specified in GBB header.

BUG=chromium-os:1302
TEST=gbb_utility -c 0x100,0x1000,0x03DE80,0x1000 gbb.blob
     gbb_utility -s --hwid='test' gbb_blob
     gbb_utility -s --rootkey='some_rootkey' gbb_blob
     gbb_utility -g --hwid --rootkey='rk.bin' gbb_blob # output hwid=test

Change-Id: Ic854609effa32020f9536bc2d3f8457dff1c3719

Committed: http://chrome-svn/viewvc/chromeos?view=rev&revision=a59a029

[Randall Spangler, 2011-01-07 21:47:34]

http://codereview.chromium.org/6173001/diff/1/utility/gbb_utility.cc
File utility/gbb_utility.cc (right):

http://codereview.chromium.org/6173001/diff/1/utility/gbb_utility.cc#newcode487
utility/gbb_utility.cc:487: static string create_new_gbb(const
std::vector<uint32_t> &create_param) {
The current implementation of the GBB is done via the GoogleBinaryBlockUtil
class above.  It'd be cleaner to extend that with a create_new() method than to
have this standalone function here.

[Che-Liang Chiou, 2011-01-10 02:55:19]

LGTM with nits.  You also need Randall's approval.

http://codereview.chromium.org/6173001/diff/1/utility/gbb_utility.cc
File utility/gbb_utility.cc (right):

http://codereview.chromium.org/6173001/diff/1/utility/gbb_utility.cc#newcode548
utility/gbb_utility.cc:548: myopts.get_mode = myopts.set_mode = false;
I'm not sure, but doesn't myopts.create_mode have to be initialized here?

http://codereview.chromium.org/6173001/diff/1/utility/gbb_utility.cc#newcode580
utility/gbb_utility.cc:580: if (!parse_creation_param(param,
myopts.create_param)) {
Isn't this equivalent to "if (!optarg || !parse_creation_param(optarg,
myopts.create_param))"?

I mean, why pass a syntax incorrect argument "(empty)" to
parse_creation_param()?

[Hung-Te, 2011-01-10 03:40:57]

Thanks for the comments! Please take a look again.

http://codereview.chromium.org/6173001/diff/1/utility/gbb_utility.cc
File utility/gbb_utility.cc (right):

http://codereview.chromium.org/6173001/diff/1/utility/gbb_utility.cc#newcode487
utility/gbb_utility.cc:487: static string create_new_gbb(const
std::vector<uint32_t> &create_param) {
On 2011/01/07 21:47:34, Randall Spangler wrote:
> The current implementation of the GBB is done via the GoogleBinaryBlockUtil
> class above.  It'd be cleaner to extend that with a create_new() method than
to
> have this standalone function here.

Done.

http://codereview.chromium.org/6173001/diff/1/utility/gbb_utility.cc#newcode548
utility/gbb_utility.cc:548: myopts.get_mode = myopts.set_mode = false;
My bad. fixed.

http://codereview.chromium.org/6173001/diff/1/utility/gbb_utility.cc#newcode580
utility/gbb_utility.cc:580: if (!parse_creation_param(param,
myopts.create_param)) {
On 2011/01/10 02:55:20, Che-Liang Chiou wrote:
> I mean, why pass a syntax incorrect argument "(empty)" to
> parse_creation_param()?
 Because in that case I need to put something in error message when printed
below. Blank may be confusing so I choose "(empty)".

 However, writing in this way seems tricky because it relies on "(empty)" as
invalid input. I'll fix it.

[Che-Liang Chiou, 2011-01-10 05:05:18]

LGTM.

On 2011/01/10 03:40:57, Hung-Te wrote:
> Thanks for the comments! Please take a look again.
> 
> http://codereview.chromium.org/6173001/diff/1/utility/gbb_utility.cc
> File utility/gbb_utility.cc (right):
> 
>
http://codereview.chromium.org/6173001/diff/1/utility/gbb_utility.cc#newcode487
> utility/gbb_utility.cc:487: static string create_new_gbb(const
> std::vector<uint32_t> &create_param) {
> On 2011/01/07 21:47:34, Randall Spangler wrote:
> > The current implementation of the GBB is done via the GoogleBinaryBlockUtil
> > class above.  It'd be cleaner to extend that with a create_new() method than
> to
> > have this standalone function here.
> 
> Done.
> 
>
http://codereview.chromium.org/6173001/diff/1/utility/gbb_utility.cc#newcode548
> utility/gbb_utility.cc:548: myopts.get_mode = myopts.set_mode = false;
> My bad. fixed.
> 
>
http://codereview.chromium.org/6173001/diff/1/utility/gbb_utility.cc#newcode580
> utility/gbb_utility.cc:580: if (!parse_creation_param(param,
> myopts.create_param)) {
> On 2011/01/10 02:55:20, Che-Liang Chiou wrote:
> > I mean, why pass a syntax incorrect argument "(empty)" to
> > parse_creation_param()?
>  Because in that case I need to put something in error message when printed
> below. Blank may be confusing so I choose "(empty)".
> 
>  However, writing in this way seems tricky because it relies on "(empty)" as
> invalid input. I'll fix it.

[Randall Spangler, 2011-01-12 22:48:55]

LGTM