On Tue, 13 Jul 2004, Jan Kiszka wrote: > Bret Yen-Ting Lin wrote: > >>>There is also another thing.. if I have RTmac insmoded, but say the TDMA > >>>has failed to operate properly in the ./rtnet start. How much of RTmac is > >>>still working ? can I still connect to network with stations without RTAI > >>>+ RTnet and work as expected just without TDMA ? > >>> > >> > >>Not really. The station which failed to start up will not be able to > >>transmit any data as long as you don't issue an "rtifconfig rtethX mac > >>down". The server, although lacking at least one expected station, will > >>work if at least one other client went up. > >> > > > > > > sorry I am not following you .. what do you mean by "will not be able to > > transmit any data as long as you don't issue an "rtifconfig rtethX mac > > The TDMA plugin simply blocks any outgoing traffic of a station until > that node has been accepted by the TDMA master. This blocking can be > resolved by shutting down the RTmac/TDMA support of the station. > > > down" ". But what if the client does not return from prompt after 'rtcfg > > rteth0 ready', which i manually ^C it out. TDMA is definitely not > > synchronised properly.. will RTmac still work between the those two > > statiosn? > > > > When RTcfg comes into play, things currently get a bit more complicated. > As you start the RTcfg server with a list of all expected stations, all > nodes will wait for this stations at any stage, also at this one > (ready). You may add timeouts to the rtcfg commands, but be warned, the > timeout support of RTcfg is not fully tested yet. This is because of the > current TDMA implementation which generally requires a clean startup of > all stations anyhow. > > This all sounds ugly, and it actually is. For now, it is definitely > better to avoid that a station fails to start up. Generally, this should > be possible, or do you have any special scenario where this error occurs > regularly? > > TDMA V2 shall solve this - someday...
unfortunately.. sometimes RTmac/TDMA just fails for no reason .. further, i know that wireless driver is not yet implemented..(don't know if it will be? ) but i m hoping to replace the cable with a wireless modem and see how it goes.. but in this case, I am sure it would even harder for RTmac/TDMA for function properly > > >>>Sorry .. forgot to add.. how has it been done for you guys for the > >>>server to wait for signals ? > >>> > >> > >>Well, didn't get this yet. What kind of signals do you mean? > >> > > > > > > say if you want to check whether the sockets has data to read ? > > (this in Unix is dealt with poll or select). What would really happen if i > > used poll with rtnet APIs such as recvfrom_rt ? or mixing the two just > > doesn't really make sense. > > > > No, makes absolutely no sense. The file descriptors used by RTnet (which > are now managed by the RTDM module) are in an other domain than the > standard Linux descriptors. Mixing them up in the RTnet implementation > would easily break real-time constraints. That's why we need a dedicated > poll_rt or select_rt... > As i thought... > Jan > I will probably do a bit more testings on the server/client I am writing now.. and see how it turned out first. Thanks Bret ------------------------------------------------------- This SF.Net email sponsored by Black Hat Briefings & Training. Attend Black Hat Briefings & Training, Las Vegas July 24-29 - digital self defense, top technical experts, no vendor pitches, unmatched networking opportunities. Visit www.blackhat.com _______________________________________________ RTnet-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/rtnet-users
