<html>
<head>
<base href="https://bugs.freedesktop.org/">
</head>
<body>
<p>
<div>
<b><a class="bz_bug_link
bz_status_NEW "
title="NEW - [SKL] Can we support YCbCr 420 output in HDMI 2.0 port?"
href="https://bugs.freedesktop.org/show_bug.cgi?id=97774#c4">Comment # 4</a>
on <a class="bz_bug_link
bz_status_NEW "
title="NEW - [SKL] Can we support YCbCr 420 output in HDMI 2.0 port?"
href="https://bugs.freedesktop.org/show_bug.cgi?id=97774">bug 97774</a>
from <span class="vcard"><a class="email" href="mailto:ville.syrjala@linux.intel.com" title="Ville Syrjala <ville.syrjala@linux.intel.com>"> <span class="fn">Ville Syrjala</span></a>
</span></b>
<pre>(In reply to Ville Syrjala from <a href="show_bug.cgi?id=97774#c3">comment #3</a>)
<span class="quote">> (In reply to Jun Zhao from <a href="show_bug.cgi?id=97774#c0">comment #0</a>)
> > We also found some TVs just support 4:2:0 YCbCr. Has Linux driver not
> > support it yet?
>
> No support as of now.</span >
Just to elaborat ea bit. For SKL and whatnot the only way to get it would be
via an external DP->HDMI 2.0 dongle/cable, or via LSPCON (which is the same
thing pretty much).
Adding the support would involve several parts:
1) Add all the generic HDMI forum VSDB parsing and whatnot to the EDID parser
This is needed so that we'll know whether the sink supports YCbCr 4:2:0. There
have been some patches towards HDMI 2.0 support for the EDID parser, but I
can't recall how much was still missing from them. Not much merged so far.
3) Implement YCbCr 4:4:4 output support
This would involve doing RGB->YCbCr conversion with the pipe CSC. I started to
implement this at some point, but got sidetracked and never actually finished.
I should probably have a branch with some half finished code somewhere.
2) Extend out DP->HDMI dongle support to allow YCbCr 4:2:0 output.
Looking at the DP 1.4 spec, I see a PROTOCOL_CONVERTER_CONTROL_1 register that
would allow you to enable YCbCr 4:2:0 conversion. If I'm reading things
correctly we need to output YCbCr 4:4:4 to the dongle, and then dongle will
downsample that to 4:2:0 when instructed to do so via this register. We
probably have to parse a few other things from the DPCD as well so that we'll
know whether the dongle even supports this.</pre>
</div>
</p>
<hr>
<span>You are receiving this mail because:</span>
<ul>
<li>You are the assignee for the bug.</li>
<li>You are on the CC list for the bug.</li>
<li>You are the QA Contact for the bug.</li>
</ul>
</body>
</html>