Hello, Am 30.11.2018 um 16:12 schrieb kernel@xxxxxxxxxxxxxxxx: > >> On 30.11.2018, at 15:20, Wolfgang Grandegger <wg@xxxxxxxxxxxxxx> wrote: >> >> Hello, >> >> Am 29.11.2018 um 16:57 schrieb kernel@xxxxxxxxxxxxxxxx: >>> Hi Wolfgang! >>>> On 28.11.2018, at 21:55, Wolfgang Grandegger <wg@xxxxxxxxxxxxxx> wrote: [...snip...] >>>> The symptoms are similar to what we get with bus-errors starving the CPU >>>> on some low-end systems (where we cannot disable the creation of- >>>> bus-errors, e.g. if no cable is connected). >>> >>> I can not comment on that... >> >> >> It's a different issue then. > > but I heard a comment that the old version of the driver will get high CPU > load when not connected to a CAN bus (Bus-off or no peer to do ACKs). Sending a message without connection to the bus results in ACK errors. Even if the MCP2515 manual states that the error interrupt will only be triggered if the error state of the transmitter or receiver has changed, it triggers much more often... on every bus error, I suspect. Do you see that problem with the MCP2517 as well? Wolfgang.
