tablet connected to monitor shows black screen - Idek Iiyama PL2409HD

Bug #1574777 reported by ch
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Canonical System Image
Confirmed
Undecided
Stephen M. Webb
Mir
New
Undecided
Unassigned
mir (Ubuntu)
New
Undecided
Unassigned
qtmir (Ubuntu)
New
Undecided
Unassigned
unity8 (Ubuntu)
New
Undecided
Unassigned

Bug Description

When connecting my M10 device to my monitor (Idek Iiyama PL2409HD), the monitor shows that there is a signal on hdmi, but the screen stays black.

Revision history for this message
ch (ch-w) wrote :
summary: - tablet connected to monitor shows black screen
+ tablet connected to monitor shows black screen - Idek Iiyama PL2409HD
Revision history for this message
ch (ch-w) wrote :

So, is there any log file (or something like that) where I can see what's going on or rather going wrong when attaching the monitor?

Revision history for this message
Pat McGowan (pat-mcgowan) wrote :

/home/phablet/.cache/upstart/unity8.log
/var/log/syslog

Changed in canonical-devices-system-image:
status: New → Incomplete
Revision history for this message
ch (ch-w) wrote :
Revision history for this message
ch (ch-w) wrote :
ch (ch-w)
Changed in canonical-devices-system-image:
status: Incomplete → New
Changed in canonical-devices-system-image:
assignee: nobody → Michał Sawicz (saviq)
Revision history for this message
Michał Sawicz (saviq) wrote :

Here are the relevant parts of the log:

[2016-04-28 23:02:17.359724] mirserver: New display configuration:
[2016-04-28 23:02:17.370686] mirserver: 0.1: LVDS 10,0" 135x216mm
[2016-04-28 23:02:17.370872] mirserver: Current mode 1200x1920 57,07Hz
[2016-04-28 23:02:17.370953] mirserver: Preferred mode 1200x1920 57,07Hz
[2016-04-28 23:02:17.371005] mirserver: Logical position +0+0
[2016-04-28 23:02:17.381173] mirserver: 0.2: DisplayPort 9,2" 203x114mm
[2016-04-28 23:02:17.381298] mirserver: Current mode 1920x1080 0,60Hz
[2016-04-28 23:02:17.381354] mirserver: Preferred mode 1920x1080 0,60Hz
[2016-04-28 23:02:17.381406] mirserver: Logical position +1200+0
[2016-04-28 23:02:17.381464] mirserver: 0.3: unused LVDS

qtmir.screens: Screen(0x800400) - id: 1 geometry: QRect(0,0 1200x1920) window: 0x14e6e90 type 7
qtmir.screens: Screen(0x12a4760) - id: 2 geometry: QRect(1200,0 1920x1080) window: 0x11ac1a0 type 10

[2016-04-28 23:02:43.188113] mirserver: New display configuration:
[2016-04-28 23:02:43.198915] mirserver: 0.1: LVDS 10,0" 135x216mm
[2016-04-28 23:02:43.199087] mirserver: Current mode 1200x1920 57,07Hz
[2016-04-28 23:02:43.199148] mirserver: Preferred mode 1200x1920 57,07Hz
[2016-04-28 23:02:43.199199] mirserver: Logical position +0+0
[2016-04-28 23:02:43.199257] mirserver: 0.2: unused DisplayPort
[2016-04-28 23:02:43.199310] mirserver: 0.3: unused LVDS

qtmir.screens: Screen(0x800400) - id: 1 geometry: QRect(0,0 1200x1920) window: 0x11ac1a0 type 7

I'm afraid the "Current mode 1920x1080 0,60Hz" part is the culprit. It seems the monitor tricks Mir into thinking it wants a refresh rate of... 0,6 times per second...

Looking at the EDID data from Xorg.log no such mode is advertised - there is a lot of "The EDID for Idek Iiyama PL2409HD (DFP-1) contradicts itself" messages. Mir's likely less able to work around such quirks still.

Changed in canonical-devices-system-image:
assignee: Michał Sawicz (saviq) → Stephen M. Webb (bregma)
tags: added: black-screen
Revision history for this message
Daniel van Vugt (vanvugt) wrote :

Mir does have some code that will throttle proportionally to the advertised refresh rate. That's the 'predictive bypass' optimization. Although even with that you should see a frames every second or two -- something would be on screen. So the monitor reporting (or Mir detecting?) an incorrect frame rate is not the main problem.

As shown in comment #6, Mir did detect the external monitor but it was only connected for 26 seconds. The log appears to show Mir notifying Unity8 of this successfully so I'm not sure who is to blame for the image remaining black.

One thing you can test right now is the cable itself. The biggest problem we have with external monitors on the phablet is faulty display cables. They are very common. So if you can, please test the cable on a different device, or test a different cable.

Revision history for this message
Daniel van Vugt (vanvugt) wrote :

I take it back. On Mir's android platform the predictive bypass optimization will never delay rendering by more than 10ms at most. So an incorrect EDID won't stop Mir from compositing. Although there might be deeper Android code that is confused by the invalid refresh rate reported by the monitor...

Equally possible is a bug in Unity8/QtMir causing the black screen...

tags: added: multimonitor
Revision history for this message
Daniel van Vugt (vanvugt) wrote :

ch: Please try/check:

1. The same HDMI cable plugged into a different tablet/phone.

2. The same tablet+cable plugged into a different monitor.

3. Does the M10's screen change when you plug in the cable? I thought it was meant to...?

Changed in canonical-devices-system-image:
status: New → Incomplete
Changed in mir:
status: New → Incomplete
Changed in mir (Ubuntu):
status: New → Incomplete
Changed in qtmir (Ubuntu):
status: New → Incomplete
Changed in unity8 (Ubuntu):
status: New → Incomplete
Revision history for this message
Michał Sawicz (saviq) wrote :

Well, I suppose the problem is that the monitor simply rejects refreshing at 0,60Hz. According to unity8 it does create a second screen etc., so I suspect the device screen changes to "virtual touchpad", but nothing's displayed on the external screen because of the crazy refresh rate.

Revision history for this message
ch (ch-w) wrote :

Because Stephen mentioned on the mailing list that it could depend on the
cable, I bought a second one, but no luck with that either.

1. sorry, the M10 is my only tablet/phone with hdmi

2. tablet+cables work on my LCD-TV. It seems my tv also wants a refresh rate of 0,60Hz

[2016-05-10 22:53:54.491656] mirserver: New display configuration:
[2016-05-10 22:53:54.500558] mirserver: 0.1: LVDS 10,0" 135x216mm
[2016-05-10 22:53:54.500700] mirserver: Current mode 1200x1920 57,07Hz
[2016-05-10 22:53:54.500730] mirserver: Preferred mode 1200x1920 57,07Hz
[2016-05-10 22:53:54.500753] mirserver: Logical position +0+0
[2016-05-10 22:53:54.510085] mirserver: 0.2: DisplayPort 9,2" 203x114mm
[2016-05-10 22:53:54.510203] mirserver: Current mode 1920x1080 0,60Hz
[2016-05-10 22:53:54.510259] mirserver: Preferred mode 1920x1080 0,60Hz
[2016-05-10 22:53:54.510330] mirserver: Logical position +1200+0
[2016-05-10 22:53:54.510371] mirserver: 0.3: unused LVDS
[2016-05-10 22:54:49.797556] mirserver: New display configuration:
[2016-05-10 22:54:49.832026] mirserver: 0.1: LVDS 10,0" 135x216mm
[2016-05-10 22:54:49.841870] mirserver: Current mode 1200x1920 57,07Hz
[2016-05-10 22:54:49.844096] mirserver: Preferred mode 1200x1920 57,07Hz
[2016-05-10 22:54:49.846243] mirserver: Logical position +0+0
[2016-05-10 22:54:49.849306] mirserver: 0.2: unused DisplayPort
[2016-05-10 22:54:49.850798] mirserver: 0.3: unused LVDS

3. M10's screen changes to the grey virtual touchpad immediately, after a few seconds the OSD of the monitor shows that there is something on hdmi, the backlights go on, but only black pixels.

In the attach unity8.log:
I made a "service lightdm restart", plugged in the tv , saw the desktop on the tv-screen, unplugged the tv, plugged in the monitor, waited a few 30 seconds, unplugged the monitor.

Changed in canonical-devices-system-image:
status: Incomplete → New
Revision history for this message
Daniel van Vugt (vanvugt) wrote :

Can I assume other M10 owners also see "0,60Hz" or "0.60Hz"?.. In that case we could ignore that bit.

Changed in mir:
status: Incomplete → New
Changed in mir (Ubuntu):
status: Incomplete → New
Changed in qtmir (Ubuntu):
status: Incomplete → New
Changed in unity8 (Ubuntu):
status: Incomplete → New
Revision history for this message
Daniel van Vugt (vanvugt) wrote :

Interesting you say the same tablet and cable works on your TV. That suggests it might be a signal/connection incompatibility in the Idek Iiyama PL2409HD rather than a software problem.

Revision history for this message
ch (ch-w) wrote :

Sorry, I'd attached the wrong unity8.log. Now this one is the one I described above.

Hmm, and why can you ignore that the mirserver logs the same wrong information about the display for two different monitors?

Revision history for this message
Daniel van Vugt (vanvugt) wrote :

Yes 0.6Hz sounds bad but I'm hoping we can ignore it for a couple of reasons:

1. Mir does not use that information to change its behaviour. Only prints it out. Instead Mir will sync to vblank (60Hz or whatever).

1.1. Android drivers on the other hand might use that information, which would be bad but not if #2 is proven...

2. If other M10 owners also see "0.6Hz" but can get external displays to work fine then the 0.6Hz message is benign and is a minor bug that can be ignored. In that case it's _not_ an indication of the speed Mir is trying to render at.

Revision history for this message
ch (ch-w) wrote :

Ok. As said, I see "0.6Hz" when connecting to my tv, which works.

Can this bug report "merged" with

https://bugs.launchpad.net/canonical-devices-system-image/+bug/1574635

Seems to me the very same problem.

Revision history for this message
Daniel van Vugt (vanvugt) wrote :

Thanks for the clarification.

That bug sounds similar but when dealing with individuals' experiences with different device types, it's best to not merge such bugs. Because we're often surprised when we find they have different root causes. And then people get upset when their duplicate bug isn't resolved with the main bug's resolution.

Changed in canonical-devices-system-image:
status: New → Confirmed
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.