Fix SAFT to work on all x86 platforms using crossystem.

Fix SAFT to work on all x86 platforms using crossystem.

This is the second and final CL restoring SAFT functionality 
on X86 platforms. The first one was

http://codereview.chromium.org/6849042/

This change 

- builds up the VECTOR_MAP to cover all legacy boot vector
  values

- adds the provision to modify the map at run time on Mario
  (because Mario firmware always reports the recovery switch 
  state as 'ON' if booted in recovery mode)

- adds test cases to verify this map adjustment on Mario 
  and to verify proper reporting of failed shel commands
 
Change-Id: I9982f53ba8a5cc286eb8767686e714e22140f892

BUG=chrome-os-partner:2617
TEST=manual

- run SAFT on Mario, Alex and ZGB, follow the directions
  (unplug and plug back in the USB flash stick a required).
- Observe SAFT complete successfully (the /tmp/saft.log file
  contains 'We are done!' as the last line.

Committed: http://chrome-svn/viewvc/chromeos?view=rev&revision=1bf1d8a

[Randall Spangler, 2011-04-19 19:40:32]

LGTM

http://codereview.chromium.org/6881019/diff/4002/chromeos_interface.py
File chromeos_interface.py (right):

http://codereview.chromium.org/6881019/diff/4002/chromeos_interface.py#newcod...
chromeos_interface.py:162: partially built list, which is an indication of a
problem for the
'a partially built list'

[vb, 2011-04-19 21:15:22]

I had to make yet another adjustment for Mario case (see the diffs). tested
again, the test ran to completion on both Mario and ZGB.

PLA

http://codereview.chromium.org/6881019/diff/4002/chromeos_interface.py
File chromeos_interface.py (right):

http://codereview.chromium.org/6881019/diff/4002/chromeos_interface.py#newcod...
chromeos_interface.py:162: partially built list, which is an indication of a
problem for the
On 2011/04/19 19:40:32, Randall Spangler wrote:
> 'a partially built list'

Done.

[Randall Spangler, 2011-04-19 22:40:21]

LGTM again