[Intel-gfx] [PATCH RESEND 9/9] drm/i915: set proper N/M in modeset

[Yang, Libin]

> -----Original Message-----
> From: Ville Syrjälä [mailto:ville.syrjala at linux.intel.com]
> Sent: Thursday, October 20, 2016 8:03 PM
> To: Nikula, Jani <jani.nikula at intel.com>
> Cc: Yang, Libin <libin.yang at intel.com>; Lin, Mengdong
> <mengdong.lin at intel.com>; intel-gfx at lists.freedesktop.org; Zhang, Keqiao
> <keqiao.zhang at intel.com>; libin.yang at linux.intel.com; Pandiyan,
> Dhinakaran <dhinakaran.pandiyan at intel.com>; Zhao, Juan J
> <juan.j.zhao at intel.com>
> Subject: Re: [Intel-gfx] [PATCH RESEND 9/9] drm/i915: set proper N/M in
> modeset
> 
> On Thu, Oct 20, 2016 at 02:34:34PM +0300, Jani Nikula wrote:
> > On Thu, 20 Oct 2016, "Yang, Libin" <libin.yang at intel.com> wrote:
> > >> >> Any particular reason these M/N values are half of what they're
> > >> >> in table
> > >> >> 2-104 of DP 1.4 spec? (Admittedly the table is an informative
> > >> >> example.)
> > >> >
> > >> > For HDMI, we found only set N is enough. HW then can handle the
> > >> remaining.
> > >>
> > >> I meant, the M and N values in this part of the dp_aud_n_m table
> > >> are 1/2 of what they are in the DP spec table. Why?
> > >
> > > Which table are you meaning? I calculate the values myself. I didn't
> > > find the full table in DP spec. I only find the table for 270MHz and
> > > 162MHz in Table 2-50: Examples of Maud and Naud Values
> >
> > Table 2-104 in DP 1.4. Maybe you're looking at an older version of the
> > spec.
> 
> So it looks like they used the same M value for all link frequencies in that
> example, which for 540M ends up being double what the minimal accurate
> value is. But as both M and N are doubled the ratio is still exactly the same.

There is a formula in DP 1.2 spec:
Maud/Naud = 512 * fs / f_LS_Clk

Suppose all the values meeting the formula is acceptable.

> 
> And there is indeed a 64k and 128k bitrates added as well. Those are not
> classified under HBR audio. But as those can't be expressed via short audio
> descriptors in the EDID, I don't think they would be accepted by ALSA. Or am
> I wrong? I do wonder how they are supposed to be used though since there
> must be some way for the sink to advertise them, otherwise I can't see the
> point in adding them.

Oh, sorry, I was thinking 128k is HBR. I didn't see 64k and 128k description
in DP 1.2 spec. 

Regards,
Libin

> 
> --
> Ville Syrjälä
> Intel OTC