On 5/20/19 17:03, Stephen Boyd wrote: > Quoting Jorge Ramirez (2019-05-20 07:58:54) >> On 5/20/19 16:56, Jorge Ramirez wrote: >>> >>> yeah, semantically confusing msm_reset_dm_count is what really matters: >>> it tells the hardware to only take n bytes (in this case only one) so >>> the others will be ignored >> >> um after I said this, maybe iowrite32_rep should only be applied to >> uartdm ... what do you think? >> > > Probably. The uartdm hardware typically required words everywhere while > the pre-dm hardware didn't. It's an if condition so it should be OK. > > It may be time to remove non-uartdm support from this driver > all-together. From what I recall the only devices that are upstream are > the uartdm ones, so it may be easier to just remove the legacy stuff > that nobody has tested in many years. > > should this be merged before removing the non-uartdm support or after? I also have some other changes - in particular with respect to the usage of fifosize which according to the documentation is in words not in bytes.
