netdev - Re: [RFC v2 net-next 00/10] Time based packet transmission

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Tue, 23 Jan 2018 10:07:23 -0800
From:   Jesus Sanchez-Palencia <jesus.sanchez-palencia@...el.com>
To:     Richard Cochran <richardcochran@...il.com>
Cc:     netdev@...r.kernel.org, jhs@...atatu.com, xiyou.wangcong@...il.com,
        jiri@...nulli.us, vinicius.gomes@...el.com,
        intel-wired-lan@...ts.osuosl.org, anna-maria@...utronix.de,
        henrik@...tad.us, tglx@...utronix.de, john.stultz@...aro.org,
        andre.guedes@...el.com, ivan.briano@...el.com,
        levi.pearson@...man.com
Subject: Re: [RFC v2 net-next 00/10] Time based packet transmission

Hi,


On 01/22/2018 09:26 PM, Richard Cochran wrote:
> On Mon, Jan 22, 2018 at 09:23:27PM -0800, Richard Cochran wrote:
>> On Wed, Jan 17, 2018 at 03:06:11PM -0800, Jesus Sanchez-Palencia wrote:
>>> First, a baseline test was ran for 10 minutes with the plain kernel only:
>>>
>>> |                 | plain kernel @ 1ms |
>>> |-----------------+--------------------+
>>> | min (ns):       |    +4.820000e+02   |
>>> | max (ns):       |    +9.999300e+05   |
>>> | pk-pk:          |    +9.994480e+05   |
>>
>> ...
>>
>>> |                 |    tbs SW @ 1ms   |  tbs HW @ 1ms  | tbs HW @ 250 us |
>>> |-----------------+-------------------+----------------+-----------------|
>>> | min (ns):       |    +1.510000e+02  |  +4.420000e+02 |   +4.260000e+02 |
>>> | max (ns):       |    +9.977030e+05  |  +5.060000e+02 |   +5.060000e+02 |
>>> | pk-pk:          |    +9.975520e+05  |  +6.400000e+01 |   +8.000000e+01 |
>>
>> I wonder about these worst case measurements of 999 and 998
>> milliseconds.  It almost looks like you missed one entire period.
>   ^^^^
> microseconds
> 
>> Could this simply be a bug in the test setup?


Yes. From the data set of the tbs SW:

offset |      timestamp
-------+---------------------
(...)  |
10639  | 1516117448.058010639
9503   | 1516117448.059009503
10167  | 1516117448.060010167
9823   | 1516117448.061009823
9567   | 1516117448.062009567
997703 | 1516117448.062997703 ****
911719 | 1516117448.063911719
12655  | 1516117448.065012655
12399  | 1516117448.066012399
(...)

Since the period was 1ms, the highlighted entry should have arrived within the
[1516117448.063000000, 1516117448.063999999] range, so in this case it was
early. For the next runs, I will modify the test setup so the txtime is sent as
part of the packet payload and later taken into account by the post-processing
script that is calculating the offsets.


Thanks,
Jesus


>>
>> Thanks,
>> Richard