<urja>
The current set of patches i have around (minus some documentation ones that no longer applied so i just dropped them) are the result of my last attempt at that lol.
<urja>
and yeah "standard" 1080p displays should work without, almost anything else (1280x1024, 1920x1200, etc..) wont really work right
<urja>
so... not really, i wasnt able to manage to get upstream to accept that it could be okay to just say it in the DTS that this device needs to dedicate a whole PLL to HDMI for flexible rates...
<urja>
(like, it would be trouble for a device with both the outputs on external connectors... it could only be flexible on one of them, and support only the fixed set of divisors on the other...)
<urja>
so the upstream default of one more fixed PLL does give slightly larger set of fixed divisors on both, but ehh... that's not enough for what people would expect of a HDMI port
<macc24>
hm
<urja>
maybe one could write new code to detect this dynamically? but it still would have many a kernel development complications (but in the style of hey a fixed panel, hey a port, okay that one is the flexible one...)
<urja>
and also if you want to be officially EMC compliant (not that i think it really matters) you should still only allow the Google-tested list of HDMI frequencies (and associated PLL configurations) on the device, *shrug*
<urja>
(they're pushing the internal PLL frequency above the limits in the public Rockchip datasheet, for some of the rates, to get a more accurate output frequency match...)
<urja>
(is kinda wild to me, that the RK3288 has a PLL that can safely hit over 5 Ghz iirc? was the "limit" that... i seem to remember so anyways)