-----Original Message----- From: Oliver Hartkopp [mailto:[email protected]] Sent: Monday, September 07, 2009 11:29 AM To: Gribov, Vladislav Cc: [email protected]; [email protected] Subject: Re: at91_can bus off handling
[email protected] wrote: > Hi, > >> Yes, the question is if we should support auto-recovery at all, e.g. > as >> proposed in >> > https://lists.berlios.de/pipermail/socketcan-core/2009-September/002971. > html >> I tend to disallow it, also to have a common behavior of auto and > manual >> recovery. > > restart-ms == 0 : > Stop the interface on BUS_OFF (all CAN controllers) > restart-ms > 0 : > SJA1000 & friends: Restart the controller after the given time. > MCP2510 & friends: Do nothing & wait for the controller > self-restart. > > > I like this solution. It uses the auto recovery feature of the HW, which > supports it and do it's best to simulate the auto recovery in driver for > the CAN controller without auto recovery feature. > > Me too ;-) > > IMO the question is, if we would like to disable the auto recovery feature on > recovery capable chips always (by setting them into reset-mode at BUS-OFF). > > Then we've just CAN controllers like the SJA1000 & friends left, and on > restart-ms > 0 also these auto recovery chips are initialized after the give time. IMHO, no > > OR > > if we allow recovery capable CAN controllers to recover themselves on > restart-ms > 0 I do like this option better the previous one. > IMHO letting a auto recovery capable CAN controller do the recovery would be > ok, even when we get a different behaviour on the bus ... Yes, in this case different behavior is something what we have to accept. If somebody want to have the same behavior for all types of HW, the restart-ms == 0 will do the job. Here the application has complete control over the reaction, timings etc. Regards, Vladislav _______________________________________________ Socketcan-core mailing list [email protected] https://lists.berlios.de/mailman/listinfo/socketcan-core
