Bug 6485 - Potential issue parsing EDID in firmware
Summary: Potential issue parsing EDID in firmware
Status: RESOLVED FIXED
Alias: None
Product: MinnowBoard MAX
Classification: Hardware Platforms
Component: hw-minnowmax (show other bugs)
Version: unspecified
Hardware: MinnowBoard Max Multiple
: Medium major
Target Milestone: Production Release
Assignee: He, Tim
QA Contact:
URL:
Whiteboard:
Depends on:
Blocks:
 
Reported: 2014-06-26 21:10 UTC by John 'Warthog9' Hawley
Modified: 2014-08-14 23:17 UTC (History)
1 user (show)

See Also:
OS type for building Yocto: ---
Type of Regression: ---
Verified:
Documentation change: No (bug/feature does not impact docs)


Attachments
Lilliput xrandr output (8.31 KB, text/plain)
2014-07-09 21:47 UTC, John 'Warthog9' Hawley
no flags Details
BenQ (8.97 KB, text/plain)
2014-07-09 21:58 UTC, John 'Warthog9' Hawley
no flags Details

Note You need to log in before you can comment on or make changes to this bug.
Description John 'Warthog9' Hawley 2014-06-26 21:10:35 UTC
EFI Shell version 2.40 [1.0]
Current running mode 1.1.2

Minnowboard Max board version: A1

Certain monitors EDID seem to have issues with the MinnowBoard MAX, in particular with the Lilliput screens.  Larger, more normal, monitors (eg: Dell), seem to be fine, but the Lilliputs very often won't give a display at all, sometimes requiring odd re-plugging in, or VT switching to resolve.

It was thought this might get fixed with the A1 from the A0 (also seen there), but this has come up on the A1 now as well.

This bug is meant as a tracking bug with respect to that.
Comment 1 John 'Warthog9' Hawley 2014-06-27 01:28:13 UTC
Additional information:

Darren and I both have a Lilliput 7in screen, looks to be 619A

http://lilliputweb.net/resistive-touch-screen-lcd-monitors/7-inch-monitors/619a.html

Darren noted that with his A0 board it worked fine, with his A1 board it did not.

I just tested mine (with a cable borrowed from Pidge), and it works more or less fine.

I took my board, and tested it with Darren's monitor, and was able to replicate what he's seeing.

If I use my board, Darren's screen and the cable I borrowed - I get display on Darren's screen from my board.

If I use my A0 board, my screen and Pidge's cable: I get no display

If I use my A0 board, my Dell screen, and a different cable (hdmi <-> dvi w/ adapter), it does work
Comment 2 John 'Warthog9' Hawley 2014-07-09 21:47:24 UTC
Created attachment 2034 [details]
Lilliput xrandr output
Comment 3 John 'Warthog9' Hawley 2014-07-09 21:58:16 UTC
Created attachment 2035 [details]
BenQ
Comment 4 John 'Warthog9' Hawley 2014-07-09 21:59:44 UTC
(In reply to comment #3)
> Created attachment 2035 [details]
> BenQ

Model: Q24WS manufactured circa 2007
Comment 5 John 'Warthog9' Hawley 2014-07-09 22:28:13 UTC
Ok a More detailed problem description:

With some monitors the minnow just won't display anything, even at firmware boot time.  It's not clear what the exact problem is.  We have noticed an oddity with even specific cables which I'll detail:

Monitors Known to work regardless of cable used seemingly:

- Dell U2312HM

Monitors Known to have issues:

- Lilliput 7in screen, looks to be 619A
  http://lilliputweb.net/resistive-touch-screen-lcd-monitors/7-inch-monitors/619a.html
  EDID information from xrandr attached

- BENQ Q24WS
  EDID information from xrandr attached

Cables used:
- micro-hdmi to full hdmi cable pulled from FRI2 box: works (marked A, in John's posession)
- micro-hdmi to full hdmi cable pulled from FRI2 box: doesn't work (marked B, on Darren's desk)
- full hdmi to dvi cable w/ micro-hdmi to hdmi adapter: works
- micro-hdmi to full hdmi adapter attached to hdmi cable: doesn't work

Most testing was done with the Lilliput screens, which we have two of and act consistently between the two.  Using cable marked B, no MinnowBoard MAX A1 to Lilliput ever works.  Only changing the cable from the B to the A cable, all tests worked.
Comment 6 John 'Warthog9' Hawley 2014-08-14 23:17:22 UTC
The work-around seems to work, so I'm going to close this.  We can track the issue not being "fixed" in the GOP driver in 6599