Hi! On 14/05/17 17:17, Vinod Koul wrote: > On Sun, May 14, 2017 at 03:36:56PM +0200, Alexander Sverdlin wrote: >> Hello Vinod, >> >> On 14/05/17 15:33, Alexander Sverdlin wrote: >>>>> + if (m2p_channel_nextbuf(edmac) != 0) >>>>> + dev_warn(chan2dev(edmac), "M2P: Starting from BASE1\n"); >>>> But then you are actually not restarting from BASE1 as you reset, so the >>>> warn is wrong.. Perhaps a more meaningful msg would be to say "expected 0 >>>> but found 1, so resetting" >> Should it be something like "M2P: Channel has not been reset properly!\n"? > Better and with a word that you recovered and did reset of the state No, no, I do not touch HW at all, I make the driver state consistent. So the condition for warning is not expected at all. It could be even BUG_ON(); Only in case someone brakes the driver or changes the dmaengine core. The problem was: we are maintaining the shadow of the controller state in the driver. This "->buffer" member. It was not properly reset when the controller was. Alexander. -- To unsubscribe from this list: send the line "unsubscribe dmaengine" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
