Android: uses settask to enable CPU affinity for ADB.

Android: uses settask to enable CPU affinity for ADB.

There are some flakyness in ADB that can potentially be worked around
by using a single CPU affinity.
This is an experiment for sharded perf tests, it'll probably be reverted soon.

BUG=268450
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=216685

[bulach, 2013-08-08 14:59:33]

ptal

[navabi, 2013-08-08 15:40:03]

On 2013/08/08 14:59:33, bulach wrote:
> ptal

LGTM.

I'm interested to see if this helps with adb connectivity issues. Is there any
reason you have to believe that the adb connectivity problems are caused by not
having CPU affinity?

[bulach, 2013-08-08 15:46:52]

yep.. please see the discussion in the bug, and the linked docs from there :)

[frankf, 2013-08-08 21:38:09]

lgtm

https://codereview.chromium.org/22680002/diff/1/build/android/bb_run_sharded_...
File build/android/bb_run_sharded_steps.py (right):

https://codereview.chromium.org/22680002/diff/1/build/android/bb_run_sharded_...
build/android/bb_run_sharded_steps.py:180: cmd_helper.RunCmd(['taskset', '-c',
'1', 'adb', 'start-server'])
why not -c 0?

[bulach, 2013-08-09 09:15:46]

thanks! rebased after reverting the adb trace, addressed the comment, landing.

https://codereview.chromium.org/22680002/diff/1/build/android/bb_run_sharded_...
File build/android/bb_run_sharded_steps.py (right):

https://codereview.chromium.org/22680002/diff/1/build/android/bb_run_sharded_...
build/android/bb_run_sharded_steps.py:180: cmd_helper.RunCmd(['taskset', '-c',
'1', 'adb', 'start-server'])
On 2013/08/08 21:38:09, frankf wrote:
> why not -c 0?

my bad, I was confused by "man taskset": it says for mask, 0x00000001 is the CPU
#0... that's not the case for -c, thanks for pointing out.

[commit-bot: I haz the power, 2013-08-09 09:17:07]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/bulach@chromium.org/22680002/8001

[commit-bot: I haz the power, 2013-08-09 14:32:13]

Change committed as 216685