[ns] control flow thru 802.11 mac
M.Gopi Krishna
mgopi@csa.iisc.ernet.in
Tue Apr 16 12:40:04 2002
Hi,
Thanks for the reply. I think you are correct.
So what happens when a mac packet is pending due to backoff and another
mac packet comes? Can this happen? If so, won't it corrupt the pktRTS_ and
pktDATA_ variables?
Thanks,
Gopi
On Tue, 16 Apr 2002, Sorav Bansal wrote:
>
> Hi,
> No I dont think thats true. The RTS packet is stored in pktRTS_ until
> it gets a chance to be sent on the channel. The RTS packet is created only
> once and once it is sent successfully (after multiple retransmits), it is
> deleted. Till then it is stored in pktRTS_.
> On sensing the carrier to be busy, the backoff is started and the
> backoffHandler when backoff expires (and hence the pktRTS_ is safe).
> regards,
> --sorav
>
>
>
>
> "M.Gopi
> Krishna" To: ns-users@ISI.EDU
> <mgopi@csa.iisc cc:
> .ernet.in> Subject: [ns] control flow thru 802.11 mac
> Sent by:
> ns-users-admin@
> ISI.EDU
>
>
> 04/16/02 04:16
> PM
> Please respond
> to "M.Gopi
> Krishna"
>
>
>
>
>
>
> Hello,
> I've some doubt reg packet flow thru 802.11 mac.
> According to me, link layer schedules the packet handler of mac using some
> delay. After this delay, mac handler is activated and recv(Packet *p) of
> 802_11 mac is called. This checks whether is call is upcall or
> downcall. Let's assume this is downcall, this calls
> Mac802_11::send(...). This initializes mac header and creates (allocs) an
> RTS pkt. It then checks for medium busy. If the medium is busy, it does
> the same as done by link layer (schedule handler).
> But where does it free the RTSpacket alloc'd.
>
>
> Since this may happen multiple times, this may create multiple RTS packets
> unfreed.
> Is this true?
>
> Thanks,
> Gopi
>
>
>
>
>