IPC between Tasks

Tobias Fink tobias.fink at stud.h-da.de
Tue Jun 12 15:12:21 CEST 2012


Hi Björn,

that helps to understand it a lot yes. The background of this is, that
I want to show what low level communication possibilities there would
be for two instances of l4linux for example without having to use
sockets or similar technologies. If this would be in principle
possible due to the fact that everything is a thread, then the example
that Vasily linked to is enough for me.
Thank you!

Tobi


2012/6/12 Björn Döbel <doebel at os.inf.tu-dresden.de>:
> On 12.06.2012 14:41, Tobias Fink wrote:
>> Hi,
>>
>> I'm planning on building a small example of a producer/consumer build
>> on top of L4Re. From the documentation I know about the shared memory
>> inteface which will help to pass information from the producer to the
>> consumer but I am a bit clueless on how I notify the consumer that
>> there is something new to read. The semaphore is only to be used in
>> between threads so there is no way to use them for IPC between Tasks?!
>> Is there another possibility to make this work other than to use
>> polling?
>
> There is no communication between tasks, because tasks in terms of L4 /
> Fiasco.OC are only address spaces (plus some other state, such as the
> capability index table). The units of execution are always threads and
> hence communication always happens between those. Does that help you?
>
> Bjoern
> --
> Dipl.-Inf. Bjoern Doebel    Mail:  doebel at tudos.org
> TU Dresden, OS Chair        Phone: +49 351 463 38 799
> Noethnitzer Str. 46         Fax:   +49 351 463 38 284
> 01187 Dresden, Germany      WWW:   http://www.tudos.org/~doebel
> --
> "Futbol no es en simple juego. Es una arma de la revolucion!"
>
> _______________________________________________
> l4-hackers mailing list
> l4-hackers at os.inf.tu-dresden.de
> http://os.inf.tu-dresden.de/mailman/listinfo/l4-hackers




More information about the l4-hackers mailing list