Hi,

Il 16/02/2018 16:50, Maxime Ripard ha scritto:

On Thu, Feb 15, 2018 at 07:05:56PM +0100, Giulio Benetti wrote:

quoted

quoted

If so, and if remember the captures properly, the sampling would occur
right before the rise, and not really around the fall.

Would 2/3 be better here?

Yes, you're right, 2/3 phase is better:

1/3 phase: https://pasteboard.co/H4VehON.png
2/3 phase: https://pasteboard.co/H4Veq8a.png

Take a look at the bit in middle(yellow) sampled by clock(blue).

Rising edge is almost in the middle of D0 bit.

quoted

quoted

According to scope captures above on both A20 and A33.
Unfortunately I don't have other boards for the other SoCs to take captures.

What do you think?

I guess we can make that part applicable to all SoCs, we haven't seen
any significant differences on those part.

So let's keep:
- As normal(rising edge) => IO_POL_REG "0x2 => 2/3 phase"
- As inverted(falling edge) => IO_POL_REG "0x0 => normal phase"

I was actually thinking 1/3 for rising, 2/3 for falling.

1/3 is almost the same waveform as D0,
having rising an falling edges almost in sync with D0.
It's not so clear because DCLK has a bad figure,
but it is that way.

2/3 instead is almost in the middle of D0 as rising.

Summarizing:
- use 0/3 as falling, then DRM_BUS_FLAG_PIXDATA_NEGEDGE
- use 2/3 as rising, then DRM_BUS_FLAG_PIXDATA_POSEDGE

I follow with a new patch using clk_set_phase function.

Maxime

-- 
Giulio Benetti
CTO

MICRONOVA SRL
Sede: Via A. Niedda 3 - 35010 Vigonza (PD)
Tel. 049/8931563 - Fax 049/8931346
Cod.Fiscale - P.IVA 02663420285
Capitale Sociale ? 26.000 i.v.
Iscritta al Reg. Imprese di Padova N. 02663420285
Numero R.E.A. 258642