Am 04.01.19 um 12:21 schrieb Andreas Färber: > Currently there's still some bugs to be investigated, with communication > stalling on one device and another reading the radio versions wrong > (07 / 1f instead of 21, also observed on spi once).
Reproducable 100% on SPI by setting REGCACHE_RBTREE in sx130x.c. Since this serdev driver was using REGCACHE_NONE still and I don't spot a register missing as volatile either, I guess it may be a timing issue? My earlier locking patch is applied, to rule out any non-determinism in the register access order due to radio vs. concentrator interactions. Regards, Andreas -- SUSE Linux GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany GF: Felix Imendörffer, Jane Smithard, Graham Norton HRB 21284 (AG Nürnberg)
