[Bug 100755] [HSW] linux next-20170407+ boots to a blank screen

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=100755

--- Comment #19 from Jani Nikula <jani.nikula at intel.com> ---
(In reply to Rafael Ristovski from comment #17)
> You can see a clear difference:

Yeah. The link M and N values are used to regenerate the pixel clock from the
link clock on the sink side. Regenrated pixel clock = M / N * link clock. They
are 24 bit values, and per the DP spec the source can choose them freely, as
long as they are within the error boundaries described in comment #12.

Except now we have sinks that fail on some values, and some other sinks that
fail on some other values. I am not amused.

Pixel clock 69300:

> Good:
> [1.710939] [drm:intel_dump_pipe_config] dp m_n: lanes: 2; gmch_m: 2422210,
> gmch_n: 8388608, link_m: 134567, link_n: 524288, tu: 64

Regenerated pixel clock 69299.869537354. We don't know whether the sink rounds
that to 69299 or 69300.

> Bad:
> [1.708518] [drm:intel_dump_pipe_config] dp m_n: lanes: 2; gmch_m: 302776,
> gmch_n: 1048576, link_m: 33642, link_n: 131072, tu: 64

Regenerated pixel clock with the rounding patch 69300.384521484 and without the
rounding patch 69298.324584961. They should be rounded to 69300 and 69298,
respectively.

I don't see a whole lot of difference here. Except that the sinks don't know
how to count.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
You are the QA Contact for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/intel-gfx-bugs/attachments/20170427/309efa47/attachment.html>