Hi, On 9/24/19 8:45 PM, Jeroen Hofstee wrote: > * Marc Kleine-Budde suggested to release the rx mailboxes a bit earlier, > that is included here. It is preceded by a small change in rx-offload > to make it clear that all mailboxes are read. It is not a functional > change. > * The CAN protocol kernel (CPK) should be stopped when the interface > goes down. > * The tx path has a (harmless) race for HECC_CANMIM. Just keep it set. > * The interface state keeping can lack resulting in wrong interface states, > the state is now based on the interrupt flags and are updated directly. > * The driver lacked a manner to lower the state again. This is added > based on the error counters. > > Jeroen Hofstee (7): > can: rx-offload: continue on error > can: ti_hecc: release the mailbox a bit earlier > can: ti_hecc: stop the CPK on down > can: ti_hecc: keep MIM and MD set > can: ti_hecc: add fifo underflow error reporting > can: ti_hecc: properly report state changes > can: ti_hecc: add missing state changes > > drivers/net/can/rx-offload.c | 2 +- > drivers/net/can/ti_hecc.c | 208 +++++++++++++++++++++-------------- > 2 files changed, 126 insertions(+), 84 deletions(-) > I forgot to mention that I added the fifo underflow counter... Regards, Jeroen
