i am re investigating whether the initial granted that was signaled, was probably due to another resource acquiring and then releasing the resource. In the meantime i am working on another strange behavior i noticed
If ResourceDefaultOwner is the current owner, a call to call ArbiterInfo.inUse() returns FALSE, however if another module has requested the resource, a call to ArbiterInfo.inUse() returns TRUE even though ResourceDefaultOwner has not yet released the resource. I have been discussing this with Eric, so Eric might have some more insight On Mon, Sep 15, 2014 at 3:36 AM, Martin Cerveny <mar...@c-home.cz> wrote: > Hello. > > (this may be irrelevant hint) > > On Sun, 14 Sep 2014, Roadstar Runner wrote: > >> I have noticed that resource.granted() is signaled at startup. >> > > What kind of startup & what kind of module ? > There is still unresolved bug for modules, that call request in > Software.Init(): > > https://github.com/tinyos/tinyos-main/blob/master/tos/ > chips/rf230/RF230DriverHwAckP.nc#L247 > > This is buggy, that it can be called to uninitialized resource (you have > only 50% to work), the initialization order of modules > (SoftwareInit.init()) is undefined :-) > > https://github.com/tinyos/tinyos-main/issues/134 > > If this is your issue, try simple fix: > > https://www.mail-archive.com/tinyos-help%40millennium. > berkeley.edu/msg40413.html > > M.C> > > > If ResourceRequested.requested() is signaled to a >> module that does not own the resource, i think that is a bug >> >> On Sun, Sep 14, 2014 at 3:32 PM, Eric Decker <cire...@gmail.com> wrote: >> >> >> On Sun, Sep 14, 2014 at 11:16 AM, Roadstar Runner < >> redstripe...@gmail.com> wrote: >> I have also noticed some strange behavior with >> ResourceDefaultOwner. If a module calls >> Resource.request() in Boot.booted(), >> ResourceDefaultOwner.requested() >> is signaled before >> the resource is granted to the default owner >> >> >> by default on power up, after initialization, the default owner owns the >> resource, by default. It doesn't >> actually get granted on the way up. >> >> [cleardot.gif] >> >> On Sun, Sep 14, 2014 at 12:59 PM, Eric Decker <cire...@gmail.com> wrote: >> >> >> On Sun, Sep 14, 2014 at 7:28 AM, Roadstar Runner < >> redstripe...@gmail.com> wrote: >> But if thee is no default owner for the resource, then the >> system could be >> locked because module 2 will never get the resource. I was >> under the >> impression that if a call to Resouce.request() returns >> SUCCESS , then either >> granted or requested will have to be signaled so that the >> resource is >> actually shared. >> >> >> It is my understanding that if Resource.request() returns SUCCESS than >> one is guaranteed that >> .granted will eventually be signalled. (Not either .granted or >> .requested). >> >> .requested will only be delivered after a module owns a resource and >> another module requests the >> resource. >> >> Also if module2 has the resource, releases it and immediately >> requests it again >> before module 1 is granted the resource...we will lock up the system >> >> >> I have to verify this (currently working on it), but what currently >> happens (with FCFSArbiter) >> is Module2 will be granted the resource again. So the system doesn't >> lock up but the proper >> behaviour doesn't occur. Queue discipline isn't properly honored. >> >> >> -- >> Eric B. Decker >> Senior (over 50 :-) Researcher >> >> >> >> >> >> -- >> Eric B. Decker >> Senior (over 50 :-) Researcher >> >> >> >>
_______________________________________________ Tinyos-help mailing list Tinyos-help@millennium.berkeley.edu https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help
