Em Seg, 2009-08-17 às 17:35 +0200, Daniel Mack escreveu:
0-1-0-1 should be 4 slots, with slots 2 and 4 active, but I think that what you want is 1-0-1-0 = set_tdm_slot(5, 5, 4, 16). But... See below.
Maybe we should first decide which TDM mode _should_ be correct for the mode I'm working on, and then amend the actual implementation? And provide some documentation once the API is decided.
The correct should be set_tdm_slots(5, 5, 4, 16), no doubt on this.
The problem is that "16" becomes "32" only on the active slots because of DMYSTOP. :(
- set_tdm_slot(5, 5, 4, 16) if pxa is slave of SFRM.
Wouldn't 1-0-1-0 rather be 0xa?
Well, I'm reading 1-0-1-0 as a time-line(1-2-3-4) from left to right, not as the binary value of the TTSA register. Activating slots 1 and 3 means 0x5(0101b) on TTSA value (right to left).
Confusing... ;)
Consider that you have real network mode, and you are using 2 slots of 16 bit mono audio. You want to ignore 16bit for each 16bit of data. ie, only one channel active. You really don't want 32bit DMA in this case.
I can just say that in case the DMA material is _not_ 32bit, audio will play at half speed. So the mode I'm using seems to need the double amout of PCM data. For whatever reason.
Its stereo, 2x16. So 32b DMA is correct. I use 32b DMA for 16bit stereo too, but with 1 slot of 32bits. :)
I agree that this should be (slot_width * number_of_active_slots) > 16 tough.
Ok. And that won't break your use case?
No, in my case number_of_active_slots is always 1. (Both with and without network mode on the wire)
sspsp |= SSPSP_DMYSTOP((slot_width * slots) / 2 - width - 1);sspsp |= SSPSP_SFRMWDTH((slot_width * slots) / 2);These two calculations are wrong my case. What works here is
sspsp |= SSPSP_DMYSTOP( (slot_width * slots * chn) / 2 - width - 1); sspsp |= SSPSP_SFRMWDTH((slot_width * slots * chn) / 2);
... which is another multiplication by factor 2. But I'm not sure if I got the wrong variable with value 2 :)
DMYSTOP itself should be accounted for slot_width, but this will only work if all slots are active otherwise you will get asynchronous SFRM. Something like:
[...]
Care to send modified patches? I'd be lucky to test them.
Unfortunately I would not be able to test them at the moment, not even compile test.. :/
I will send new patches when I'm back to my workstation.
This should be able to deal with I2S when pxa is slave, on both pxa2xx and pxa3xx, with set_tdm_slot(5, 5, 4, 16).
Or when pxa is master, on pxa3xx only, with set_tdm_slot(3, 3, 2, 16).
Again - shouldn't we care for a consistent calling convention here and handle the special cases in the implementation rather than feeding nebulous integer arguments to it, which vary depending on the clocking directions?
Yes, I think we can change the slots configuration in the case we set DMYSTOP.
This still violates the "DMYSTOP must be clear on network mode" rule, but as all slots are active its not really network mode. And it seems to work for you, so... ;)
Yeah, that makes no sense at all. I think we should ignore that comment. It seems to be simply wrong, or we're not in network mode afterall, even though we selected it. I don't know.
The comment is right for real network mode, when you want to ignore some timeslots. In your case, as all slots are active its not really network mode, its just a way to increase the frame width above the 32b limit.