channel switching support for wifi

[Mathieu Lacage, 2009-08-19 08:19:11]

http://codereview.appspot.com/96109/diff/1/6
File src/devices/wifi/dcf-manager.cc (right):

http://codereview.appspot.com/96109/diff/1/6#newcode626
Line 626: DcfManager::NotifySwitchingStartNow (Time duration)
I would like to see an associated test in dcf-manager-test.cc for this function
to exercise its many code paths.

http://codereview.appspot.com/96109/diff/1/6#newcode633
Line 633: {
indent

http://codereview.appspot.com/96109/diff/1/12
File src/devices/wifi/mac-low.cc (right):

http://codereview.appspot.com/96109/diff/1/12#newcode297
Line 297: m_phyMacLowListener = new PhyMacLowListener (this);
delete in DoDispose ?

http://codereview.appspot.com/96109/diff/1/13
File src/devices/wifi/mac-low.h (right):

http://codereview.appspot.com/96109/diff/1/13#newcode284
Line 284: void SetupPhyMacLowListener (Ptr<WifiPhy> phy);
make this private

http://codereview.appspot.com/96109/diff/1/17
File src/devices/wifi/yans-wifi-phy.cc (right):

http://codereview.appspot.com/96109/diff/1/17#newcode355
Line 355: m_channelNumber = nch;
I am curious: do you ever use m_channelNumber anywhere ? it seems all you do in
this patchset is keep track of this number, but never use it to implement an
actual model of receiving packets on different channels ?

http://codereview.appspot.com/96109/diff/1/17#newcode359
Line 359: * This function wouldn't be really needed because channel switching
delay is
so, why is it needed ?

http://codereview.appspot.com/96109/diff/1/17#newcode412
Line 412: NS_LOG_DEBUG ("drop packet because of channel switching (power="<<
is this debug message really relevant ? I don't see how rx power is used here.

[Ramon Bauza, 2009-09-08 11:36:07]

Hi Mathieu, all,

After almost a month (due to summer holidays and a lot of work :-), I have been
able now to update the patch related to channel switching support for wifi. I
have uploaded a new patch set with additions in dcf-manager-test
(http://codereview.appspot.com/96109/show).

Comments and suggestions are welcome.
Thank you very much.
Ramon.

http://codereview.appspot.com/96109/diff/1/6
File src/devices/wifi/dcf-manager.cc (right):

http://codereview.appspot.com/96109/diff/1/6#newcode626
Line 626: DcfManager::NotifySwitchingStartNow (Time duration)
On 2009/08/19 08:19:12, Mathieu Lacage wrote:
> I would like to see an associated test in dcf-manager-test.cc for this
function
> to exercise its many code paths.

done.

http://codereview.appspot.com/96109/diff/1/6#newcode633
Line 633: {
On 2009/08/19 08:19:12, Mathieu Lacage wrote:
> indent

done.

http://codereview.appspot.com/96109/diff/1/12
File src/devices/wifi/mac-low.cc (right):

http://codereview.appspot.com/96109/diff/1/12#newcode297
Line 297: m_phyMacLowListener = new PhyMacLowListener (this);
On 2009/08/19 08:19:12, Mathieu Lacage wrote:
> delete in DoDispose ?

Not before. done.

http://codereview.appspot.com/96109/diff/1/13
File src/devices/wifi/mac-low.h (right):

http://codereview.appspot.com/96109/diff/1/13#newcode284
Line 284: void SetupPhyMacLowListener (Ptr<WifiPhy> phy);
On 2009/08/19 08:19:12, Mathieu Lacage wrote:
> make this private

done.

http://codereview.appspot.com/96109/diff/1/17
File src/devices/wifi/yans-wifi-phy.cc (right):

http://codereview.appspot.com/96109/diff/1/17#newcode355
Line 355: m_channelNumber = nch;
On 2009/08/19 08:19:12, Mathieu Lacage wrote:
> I am curious: do you ever use m_channelNumber anywhere ? it seems all you do
in
> this patchset is keep track of this number, but never use it to implement an
> actual model of receiving packets on different channels ?

Yes, m_channelNumber is used (through YansWifiPhy::GetChannelNumber) in
YansWifiChannel::Send.

http://codereview.appspot.com/96109/diff/1/17#newcode359
Line 359: * This function wouldn't be really needed because channel switching
delay is
On 2009/08/19 08:19:12, Mathieu Lacage wrote:
> so, why is it needed ?

removed.

http://codereview.appspot.com/96109/diff/1/17#newcode412
Line 412: NS_LOG_DEBUG ("drop packet because of channel switching (power="<<
On 2009/08/19 08:19:12, Mathieu Lacage wrote:
> is this debug message really relevant ? I don't see how rx power is used here.

This message allows tracking those packets received during a switching period
(YansWifiPhy::m_channelSwitchDelay).