Rietveld Code Review Tool
Help | Bug tracker | Discussion group | Source code | Sign in
(1506)

Issue 4890044: Monitor Mode support for ns-3

Can't Edit
Can't Publish+Mail
Start Review
Created:
12 years, 7 months ago by ashwin.narayan89
Modified:
12 years, 6 months ago
Reviewers:
Lalith Suresh, Gary Pei
CC:
ns-3-reviews_googlegroups.com
Visibility:
Public.

Description

This work involves full radiotap header support for ns-3's Wifi model. It includes the provision of a MonitorWifiMac abstraction and support for a monitor mode. A click-based script was also developed for the purpose of testing the working of the monitor mode by means of setting TX parameters in the radiotap header for the packet through the click script. Frame injection is also supported and the parameters that can be controlled are TX power, RTS retries and DATA retries

Patch Set 1 #

Total comments: 5

Patch Set 2 : Updated RELEASE_NOTES #

Patch Set 3 : Changed GetDataMode to GetDataTxVector in airtime-metric #

Patch Set 4 : Change in YansWifiPhy behaviour for control frames #

Patch Set 5 : Checking for mgt and data headers for monitor devices in YansWifiPhy #

Unified diffs Side-by-side diffs Delta from patch set Stats (+4 lines, -1 line) Patch
M src/wifi/model/yans-wifi-phy.cc View 1 2 3 4 1 chunk +4 lines, -1 line 0 comments Download

Messages

Total messages: 6
Gary Pei
Hi, Here are some comments that need to be addressed before I can do more ...
12 years, 6 months ago (2011-09-15 21:12:17 UTC) #1
Lalith Suresh
Hello Gary, I'll ask Ashwin to merge with ns-3-dev and update the review issue. Here's ...
12 years, 6 months ago (2011-09-15 21:16:55 UTC) #2
Gary Pei
Hi Lalith, Thanks for pointing me to the ns-3-click-mac-dev/ repo. I went through several examples. ...
12 years, 6 months ago (2011-09-16 05:33:57 UTC) #3
ashwin.narayan89
Hello Gary, Thank you for your comments. I'll attempt to fix the problem as soon ...
12 years, 6 months ago (2011-09-17 23:02:50 UTC) #4
Lalith Suresh
Hi Gary, On Fri, Sep 16, 2011 at 7:33 AM, <guangyu.pei@boeing.com> wrote: > > http://codereview.appspot.com/4890044/diff/1/src/wifi/model/yans-wifi-phy.cc#newcode836 ...
12 years, 6 months ago (2011-09-19 19:47:17 UTC) #5
Gary Pei
12 years, 6 months ago (2011-09-20 16:33:30 UTC) #6
On 2011/09/19 19:47:17, Lalith Suresh wrote:

> 
> Are you suggesting that real time MAC frames that go even into the
> pcap traces should not have the radiotap header on them at all?
> 

Yes. I never see these frames from Linux monitor mode. And, some specialized
commercial sniffer tools can see these raw packets but without any radioTap
header. 

> The radiotap header is actually added higher up in the stack in
> MonitorWifiMac. Note that in the section you've pointed to in
> YansWifiPhy, the header is added to a copy of the packet which is sent
> to the pcap sniff sources. This is basically a refactoring of the pcap
> tracing code wherein the header was initially added within
> YansWifiPhyHelper.
> 
> 

That is good. So there should be no over the air frames with radioTap headers.
The radioTap headers should be local to monitor device itself. To this end, all
real time control frames should not be visible and pcap trace on these monitor
device should not report them with radioTap headers.
Sign in to reply to this message.

Powered by Google App Engine
RSS Feeds Recent Issues | This issue
This is Rietveld f62528b