[Telemetry] Improve elevate_privilege handling in LaunchApplication() + Fix killing of power metri…

[Telemetry] Improve elevate_privilege handling in LaunchApplication()  + Fix killing of power metrics process on bots.

1. Fix case where calling LauchApplication() twice in a row with elevate_privelege=True will not prepend sudo on the second invocation.  This behavior occurs because sudo is already authenticated therefore IsElevated() returns True causing CanRunElevatedWithoutSudo() to return True.
2. Moving functions that handle sudo and setuid to a new utility function.
3. Do not prompt to set sticky bit on binaries run with elevate_privilege=True
4. Re-enable power and smoke unit tests on Mac 10.9.

BUG=420615, 423688
Committed: https://crrev.com/c045fc3103d04ab47f1cf0e55457979ee203a940
Cr-Commit-Position: refs/heads/master@{#305691}

[jeremy, 2014-11-23 15:14:05]

jeremy@chromium.org changed reviewers:
+ dtu@chromium.org, skyostil@chromium.org

[jeremy, 2014-11-23 15:14:06]

Sami, dtu: This is important for the Fixit, just need a review from one of you -
thanks!

[Sami, 2014-11-24 15:05:57]

lgtm with one question/suggestion for _KillPowerMetricsProcess.

https://codereview.chromium.org/720413006/diff/20001/tools/telemetry/telemetr...
File tools/telemetry/telemetry/core/platform/posix_sudo_util.py (right):

https://codereview.chromium.org/720413006/diff/20001/tools/telemetry/telemetr...
tools/telemetry/telemetry/core/platform/posix_sudo_util.py:26: return
os.stat(path).st_mode & stat.S_ISUID == stat.S_ISUID
nit: the "== stat.S_ISUID" part isn't strictly necessary.

https://codereview.chromium.org/720413006/diff/20001/tools/telemetry/telemetr...
File
tools/telemetry/telemetry/core/platform/power_monitor/powermetrics_power_monitor.py
(right):

https://codereview.chromium.org/720413006/diff/20001/tools/telemetry/telemetr...
tools/telemetry/telemetry/core/platform/power_monitor/powermetrics_power_monitor.py:235:
if posix_sudo_util.IsSetUID(self.binary_path):
Interesting. Looks like the process's real UID is different depending on whether
you use sudo or a suid bit to run it.

I wonder if tying this to the suid bit is a little fragile and whether we should
just try terminate first and fall back to /bin/kill if it fails?

[jeremy, 2014-11-25 14:20:27]

Sami: I made this a smaller change, could you take another look please?

https://codereview.chromium.org/720413006/diff/20001/tools/telemetry/telemetr...
File
tools/telemetry/telemetry/core/platform/power_monitor/powermetrics_power_monitor.py
(right):

https://codereview.chromium.org/720413006/diff/20001/tools/telemetry/telemetr...
tools/telemetry/telemetry/core/platform/power_monitor/powermetrics_power_monitor.py:235:
if posix_sudo_util.IsSetUID(self.binary_path):
On 2014/11/24 15:05:57, Sami wrote:
> Interesting. Looks like the process's real UID is different depending on
whether
> you use sudo or a suid bit to run it.
> 
> I wonder if tying this to the suid bit is a little fragile and whether we
should
> just try terminate first and fall back to /bin/kill if it fails?

Done.

[Sami, 2014-11-25 14:28:25]

Thanks, lgtm.

[jeremy, 2014-11-25 15:23:39]

Thanks Sami!! Waiting on chromelabs to add pkill to sudoers file before landing.

[jeremy, 2014-11-25 19:21:55]

The CQ bit was checked by jeremy@chromium.org

[commit-bot: I haz the power, 2014-11-25 19:23:14]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/720413006/60001

[commit-bot: I haz the power, 2014-11-25 20:43:36]

Committed patchset #4 (id:60001)

[commit-bot: I haz the power, 2014-11-25 20:44:21]

Patchset 4 (id:??) landed as
https://crrev.com/c045fc3103d04ab47f1cf0e55457979ee203a940
Cr-Commit-Position: refs/heads/master@{#305691}