Minor fixes for protocol perf tests

Minor fixes for protocol perf tests

1. Avoid base::RandDouble() in fake sockets. On posix base::RandBytes()
   reads from /dev/urandom. On linux this causes significant CPU
   consumption in rngd, comparable to CPU time consumed by the test
   itself. The test doesn't need crypto-quality random numbers,
   changed it to use std::rand().
2. Removed some outdated code from protocol_perftest.cc.
3. Added 2Mpbs tests.

Committed: https://crrev.com/6206df6c1bea743a8cdc01be1fc0c43ce15d3e60
Cr-Commit-Position: refs/heads/master@{#421337}

[Sergey Ulanov, 2016-09-23 21:09:41]

Description was changed from

==========
Minor fixes for protocol perf tests

1. Avoid base::RandDouble() in fake sockets. On posix base::RandBytes()
   reads from /dev/urandom. On linux this causes significant CPU
   consumption in rngd, comparable to CPU time consumed by the test
   itself.
2. Removed some outdated code from protocol_perftest.cc.
3. Added 2Mpbs tests.
==========

to

==========
Minor fixes for protocol perf tests

1. Avoid base::RandDouble() in fake sockets. On posix base::RandBytes()
   reads from /dev/urandom. On linux this causes significant CPU
   consumption in rngd, comparable to CPU time consumed by the test
   itself. The test doesn't need crypt-quality random numbers, so 
   the test uses std::rand() now.
2. Removed some outdated code from protocol_perftest.cc.
3. Added 2Mpbs tests.
==========

[Sergey Ulanov, 2016-09-23 21:09:41]

sergeyu@chromium.org changed reviewers:
+ isheriff@chromium.org

[Irfan, 2016-09-27 15:38:35]

https://codereview.chromium.org/2354263007/diff/1/remoting/test/fake_socket_f...
File remoting/test/fake_socket_factory.cc (right):

https://codereview.chromium.org/2354263007/diff/1/remoting/test/fake_socket_f...
remoting/test/fake_socket_factory.cc:19: #include "base/rand_util.h"
Do you still need the base/rand_util ? Also, what is the reason for moving awway
from base::Rand

[Irfan, 2016-09-27 15:39:45]

https://codereview.chromium.org/2354263007/diff/1/remoting/test/fake_socket_f...
File remoting/test/fake_socket_factory.cc (right):

https://codereview.chromium.org/2354263007/diff/1/remoting/test/fake_socket_f...
remoting/test/fake_socket_factory.cc:19: #include "base/rand_util.h"
On 2016/09/27 15:38:34, Irfan wrote:
> Do you still need the base/rand_util ? Also, what is the reason for moving
awway
> from base::Rand

duh, saw your commit message. The first Q still holds

[Sergey Ulanov, 2016-09-27 17:15:07]

https://codereview.chromium.org/2354263007/diff/1/remoting/test/fake_socket_f...
File remoting/test/fake_socket_factory.cc (right):

https://codereview.chromium.org/2354263007/diff/1/remoting/test/fake_socket_f...
remoting/test/fake_socket_factory.cc:19: #include "base/rand_util.h"
On 2016/09/27 15:38:34, Irfan wrote:
> Do you still need the base/rand_util ? Also, what is the reason for moving
awway
> from base::Rand

Done.

[Sergey Ulanov, 2016-09-27 17:15:54]

Description was changed from

==========
Minor fixes for protocol perf tests

1. Avoid base::RandDouble() in fake sockets. On posix base::RandBytes()
   reads from /dev/urandom. On linux this causes significant CPU
   consumption in rngd, comparable to CPU time consumed by the test
   itself. The test doesn't need crypt-quality random numbers, so 
   the test uses std::rand() now.
2. Removed some outdated code from protocol_perftest.cc.
3. Added 2Mpbs tests.
==========

to

==========
Minor fixes for protocol perf tests

1. Avoid base::RandDouble() in fake sockets. On posix base::RandBytes()
   reads from /dev/urandom. On linux this causes significant CPU
   consumption in rngd, comparable to CPU time consumed by the test
   itself. The test doesn't need crypt-quality random numbers,
   changed it to use std::rand().
2. Removed some outdated code from protocol_perftest.cc.
3. Added 2Mpbs tests.
==========

[Sergey Ulanov, 2016-09-27 17:16:07]

Description was changed from

==========
Minor fixes for protocol perf tests

1. Avoid base::RandDouble() in fake sockets. On posix base::RandBytes()
   reads from /dev/urandom. On linux this causes significant CPU
   consumption in rngd, comparable to CPU time consumed by the test
   itself. The test doesn't need crypt-quality random numbers,
   changed it to use std::rand().
2. Removed some outdated code from protocol_perftest.cc.
3. Added 2Mpbs tests.
==========

to

==========
Minor fixes for protocol perf tests

1. Avoid base::RandDouble() in fake sockets. On posix base::RandBytes()
   reads from /dev/urandom. On linux this causes significant CPU
   consumption in rngd, comparable to CPU time consumed by the test
   itself. The test doesn't need crypto-quality random numbers,
   changed it to use std::rand().
2. Removed some outdated code from protocol_perftest.cc.
3. Added 2Mpbs tests.
==========

[Irfan, 2016-09-27 20:54:38]

lgtm

[Sergey Ulanov, 2016-09-27 21:00:35]

The CQ bit was checked by sergeyu@chromium.org

[commit-bot: I haz the power, 2016-09-27 21:00:48]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-09-27 21:24:05]

Description was changed from

==========
Minor fixes for protocol perf tests

1. Avoid base::RandDouble() in fake sockets. On posix base::RandBytes()
   reads from /dev/urandom. On linux this causes significant CPU
   consumption in rngd, comparable to CPU time consumed by the test
   itself. The test doesn't need crypto-quality random numbers,
   changed it to use std::rand().
2. Removed some outdated code from protocol_perftest.cc.
3. Added 2Mpbs tests.
==========

to

==========
Minor fixes for protocol perf tests

1. Avoid base::RandDouble() in fake sockets. On posix base::RandBytes()
   reads from /dev/urandom. On linux this causes significant CPU
   consumption in rngd, comparable to CPU time consumed by the test
   itself. The test doesn't need crypto-quality random numbers,
   changed it to use std::rand().
2. Removed some outdated code from protocol_perftest.cc.
3. Added 2Mpbs tests.
==========

[commit-bot: I haz the power, 2016-09-27 21:24:06]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2016-09-27 21:25:50]

Description was changed from

==========
Minor fixes for protocol perf tests

1. Avoid base::RandDouble() in fake sockets. On posix base::RandBytes()
   reads from /dev/urandom. On linux this causes significant CPU
   consumption in rngd, comparable to CPU time consumed by the test
   itself. The test doesn't need crypto-quality random numbers,
   changed it to use std::rand().
2. Removed some outdated code from protocol_perftest.cc.
3. Added 2Mpbs tests.
==========

to

==========
Minor fixes for protocol perf tests

1. Avoid base::RandDouble() in fake sockets. On posix base::RandBytes()
   reads from /dev/urandom. On linux this causes significant CPU
   consumption in rngd, comparable to CPU time consumed by the test
   itself. The test doesn't need crypto-quality random numbers,
   changed it to use std::rand().
2. Removed some outdated code from protocol_perftest.cc.
3. Added 2Mpbs tests.

Committed: https://crrev.com/6206df6c1bea743a8cdc01be1fc0c43ce15d3e60
Cr-Commit-Position: refs/heads/master@{#421337}
==========

[commit-bot: I haz the power, 2016-09-27 21:25:51]

Patchset 2 (id:??) landed as
https://crrev.com/6206df6c1bea743a8cdc01be1fc0c43ce15d3e60
Cr-Commit-Position: refs/heads/master@{#421337}