31 Jul
2018
31 Jul
'18
1:51 p.m.
On Tue, Jul 31, 2018 at 03:29:35PM +0200, Takashi Iwai wrote:
Mark Brown wrote:
However since it's not supposed to be providing any DMA a CPU DAI really shouldn't be doing this...
Well, if so, the CPU dai also cannot get the exact base delay corresponding to the reported position, either, no?
It can know how much delay it's adding internally between its input and output, which feeds into the overall delay experienced by the user.
But isn't it merely the additional delay that should be applied on top of the existing runtime->delay?
Yes. I'm saying that if the CPU DAI thinks it can figure out the base delay something is confused.