On 06/20/2010 12:45 PM, [email protected] wrote: > Hello Wolfgang, > >>> I expect to get the bus error recovery event also over the > CAN_ERR_CRTL >>> event class, because we deal here with CAN controller state change >>> (currently I am pretty sure that I do not get anything on error >>> recovery, >>> not even CAN_ERR_RESTARTED). >> >> I'm pretty sure that you get CAN_ERR_RESTARTED with a recent version of >> Socket-CAN. If not, it's a bug. > > OK, will recheck that. > > > >>> CAN_ERR_PROT_ACTIVE and CAN_ERR_RESTARTED are another classes of > events, >>> so >>> I think it is conceptually not the right way to signal error recovery >>> over >>> that events. >> >> The CAN_ERR_RESTARTED was introduced for that purpose but I agree that >> the CAN error state change should be signaled to the user as well. What >> do you think about signaling *any* state change to the user, not just >> when the state gets worse? > > Well, I don't see any reason to tell the user only the bad news 8) > Our CANopen Stack (as an example of the user application) cares also > about > resuming from error state.
OK, that's something I would like to change anyway. Going to prepare a RFC patch for the SJA1000 as soon as time permits. Nevertheless, for the for bus error recovery CAN_ERR_RESTARTED should be used. A better name would be CAN_ERR_RECOVERED, but historically it was used in a more general context. Wolfgang. Wolfgang. _______________________________________________ Socketcan-core mailing list [email protected] https://lists.berlios.de/mailman/listinfo/socketcan-core
