Not all monitors are supported on 11.06 Panda LEB

[Patrik Ryd]

The first release candidate for the Linaro Android 11.06 release (linaro-android-11.06-release branch on the manifest) has problems with some monitors for the LEB build.

There is no official build yet. Use https://android-build.linaro.org/builds/~patrik-ryd/dev1106lebp/ to test.

The same problem exists on the lt-panda build https://android-build.linaro.org/builds/~patrik-ryd/lt-panda/, but not on the daily LEB-panda build (until the TILT kernel was added).

So the problem is in the TILT kernels support for different monitors.

We know that TILT kernel works on agreens monitor (and probably on Zachs)
We know that TILT kernel works on patrikryds and fabos TVs
We know that TILT kernel does not work on patrikryds HP LP2065 monitor
We know that TILT kernel does not work on fabos viewsonic vx2255x monitor

Both fabo and patrikryd uses a hdmi2dvi adaptor

[Paul Sokolovsky (pfalcon)]

I have LG Flatron L1952TR monitor here, 1280*1024, connected via HDMI-DVI cable. I see teh picture, but it's distorted: there's empty fields of ~1/5 screen width on the left, the rest of picture is flattened to fit the rest of width (fit on the screen though). There's also broken-sync noise appear in the left field (alot) and in main picture.

[Paul Sokolovsky (pfalcon)]

The above happened with https://android-build.linaro.org/builds/~patrik-ryd/dev1106lebp/#build=6

[warmcat (andy-warmcat)]

I noticed there are multiple 1080p timings listed in omaplfb, I wonder if this is happening when the "wrong" one gets selected according to EDID content.

[Paul Sokolovsky (pfalcon)]

FWIW, I tried ~patrik-ryd/dev1106genericp #1, and it spots good picture, but with 640*480 resolution (~linaro-android/leb-panda #98 used native resolution).

[Zach Pfeffer (pfefferz)]

> Android Panda LEB
> https://android-build.linaro.org/builds/~patrik-ryd/dev1106lebp/

https://android-build.linaro.org/builds/~patrik-ryd/dev1106lebp/#build=6

trybuild.sh https://android-build.linaro.org/jenkins/job/patrik-ryd_dev1106lebp/6/artifact/build/out/target/product/pandaboard
/dev/sdd

Booted up to full Android with monitor. 1920x1080P @60Hz HDMI 1080P on
Dynex DX-24L230A12, plugged into the P2 HDMI-1080P port

> Android Generic Panda
> https://android-build.linaro.org/builds/~patrik-ryd/dev1106genericp/

trybuild.sh https://android-build.linaro.org/jenkins/job/patrik-ryd_dev1106genericp/1/artifact/build/out/target/product/pandaboard
/dev/sdd

Booted up to full Android with monitor. 640x480 @60Hz DVI on Dynex
DX-24L230A12, plugged into the P2 HDMI-1080P port

> Android Generic Beagle
> https://android-build.linaro.org/builds/~patrik-ryd/dev1106genericb/

trybuild.sh https://android-build.linaro.org/jenkins/job/patrik-ryd_dev1106genericb/1/artifact/build/out/target/product/beagleboard
/dev/sdd

Does not boot on C4 beagle
Does not boot on xM beagle

[Zach Pfeffer (pfefferz)]

Scratch the beagle comments...testing now

[Zach Pfeffer (pfefferz)]

The 11.06 Panda LEB works with at least:

1920x1080P @60Hz HDMI 1080P on Dynex DX-24L230A12, plugged into the P2 HDMI-1080P port
1920x1080P @60Hz DVI to HDMI connector on HP Compaq LA 2405wg, plugged into the P2 HDMI-1080P port

I'm going to close this bug since monitors will vary.

Reported not to work on:

HP LP2065 monitor (1600 × 1200 - VGA - DVI)
Viewsonic vx2255x monitor (1680 x 1050 - VGA - DVI)

Distorted on:

LG Flatron L1952TR monitor here, 1280*1024, connected via HDMI-DVI

[Paul Sokolovsky (pfalcon)]

I now tested http://releases.linaro.org/platform/linaro-n/android/11.06/leb-panda/ (still LG Flatron L1952TR) and have following: on the first book, the monitor is completely blank (but not in power save). On the second boot, it's back to previously described behavior with blank field on the left and narrowed image. There're two implications: 1) it starts to look similar to lp:801108 ; 2) do guys who reported "doesn't work" tried to boot second time?

About bug closure - it's clear this is "won't fix" for 11.06, but how will er track this issue for 11.07?

[Paul Sokolovsky (pfalcon)]

Well, worse - it's not 100% reproducible at all. I must have been lucky to have my monitor detected at all

Now, can't get it working over 10 boots. dmesg snippet:

<3>[ 4.754302] ENTER hdmi_panel_enable
<3>[ 4.758117] ENTER hdmi_display_enable
<3>[ 4.762359] hdmi_power_on x_res= 640 y_res = 480
<3>[ 4.767303] Read EDID as no EDID is not set on poweron
<3>[ 4.774902] **** hdmi_core_ddc_edid in
<3>[ 4.798339] 00
<3>[ 4.802093] FF
<3>[ 4.805816] FF
<3>[ 4.809417] FF
<3>[ 4.812927] FF
<3>[ 4.816375] FF
<3>[ 4.819915] FF
<3>[ 4.823333] 00
<3>[ 4.826660] 1E
<3>[ 4.829956] 6D
<3>[ 4.833099] 25
<3>[ 4.836273] 4B
<3>[ 4.839416] D2
<3>[ 4.842468] 39
<3>[ 4.845458] 02
<3>[ 4.848358] 00
<3>[ 4.851287] 0C
<3>[ 4.854125] 10
<3>[ 4.856872] 01
<7>[ 4.859588] usb 1-1.1: udev 3, busnum 1, minor = 2
<6>[ 4.859588] usb 1-1.1: New USB device found, idVendor=0424, idProduct=ec00
<3>[ 4.867797] 03
<3>[ 4.870452] EE
<3>[ 4.873016] 26
<6>[ 4.875549] usb 1-1.1: New USB device strings: Mfr=0, Product=0, SerialNumber=0
<3>[ 4.884155] 1E
<3>[ 4.886749] 78
<3>[ 4.889282] EA
<3>[ 4.892242] EC
<7>[ 4.894775] usb 1-1.1: usb_probe_device
<7>[ 4.894805] usb 1-1.1: configuration #1 chosen from 1 choice
<3>[ 4.894836] 54
<7>[ 4.897430] usb 1-1.1: adding 1-1.1:1.0 (config #1, interface 0)
<3>[ 4.897674] A5
<7>[ 4.900238] smsc95xx 1-1.1:1.0: usb_probe_interface
<7>[ 4.900238] smsc95xx 1-1.1:1.0: usb_probe_interface - got id
<6>[ 4.900329] smsc95xx v1.0.4
<3>[ 4.904052] 58
<3>[ 4.907196] 4A
<3>[ 4.909790] 9A
<3>[ 4.912384] 26
<3>[ 4.914978] 21
<3>[ 4.917572] 50
<3>[ 4.920166] 54
<3>[ 4.922760] A5
<3>[ 4.925354] 6A
<3>[ 4.927917] 80
<3>[ 4.930816] 31
<3>[ 4.933441] 4F
<3>[ 4.936004] 45
<3>[ 4.938629] 4F
<3>[ 4.941223] 61
<3>[ 4.943817] 4F
<3>[ 4.946319] 81
<3>[ 4.948760] 80
<3>[ 4.951171] 01
<3>[ 4.953613] 01
<3>[ 4.956024] 01
<3>[ 4.958251] 01
<3>[ 4.960479] 01
<3>[ 4.962615] 01
<3>[ 4.964752] 01
<3>[ 4.966888] 01
<3>[ 4.969024] 30
<3>[ 4.971160] 2A
<3>[ 4.973236] 00
<3>[ 4.975402] 98
<3>[ 4.977508] 51
<3>[ 4.979583] 00
<3>[ 4.981689] 2A
<3>[ 4.983856] 40
<3>[ 4.985961] 30
<3>[ 4.988159] 70
<3>[ 4.990325] 13
<3>[ 4.992431] 00
<3>[ 4.994537] 78
<3>[ 4.996612] 2D
<3>[ 4.998718] 11
<3>[ 5.000885] 00
<3>[ 5.002960] 00
<3>[ 5.005065] 1E
<3>[ 5.007232] 00
<3>[ 5.009429] 00
<3>[ 5.011596] 00
<3>[ 5.013763] FD
<3>[ 5.015960] 00
<3>[ 5.018157] 38
<3>[ 5.020355] 4B
<3>[ 5.022552] 1E
<3>[ 5.024749] 47
<3>[ 5.026947] 0B
<3>[ 5.029144] 00
<3>[ 5.031341] 0A
<3>[ 5.033538] 20
<3>[ 5.035736] 20
<3>[ 5.037933] 20
<3>[ 5.040130] 20
<3>[ 5.042297] 20
<3>[ 5.044464] 20
<3>[ 5.046661] 00
<3>[ 5.048828] 00
<3>[ 5.050994] 00
<3>[ 5.053741] FC
<3>[ 5.055908] 00
<3>[ 5.058074] 4C
<3>[ 5.060211] 31
<3>[ 5.062316] 39
<3>[ 5.064422] 35
<3>[ 5.066497] 32
<3>[ 5.068572] 54
<3>[ 5.070709] 52
<3>[ 5.072784] 0A
<3>[ 5.074890] 20
<3>[ 5.077026] 20
<3>[ 5.079223...

[Zach Pfeffer (pfefferz)]

I think the best thing is to file a per-monitor bug. The fixes will be per monitor so it'll be easier to track.

[Paul Sokolovsky (pfalcon)]

> I think the best thing is to file a per-monitor bug. The fixes will be
> per monitor so it'll be easier to track.

I hope that's a joke ;-). TI LT kernel apparently has broken EDID
parsing/handling, that's where problem lies. Here's build which uses
older kernel and doesn't have such issues (and few others too):
https://android-build.linaro.org/builds/~linaro-android/leb-panda/#build=98

In other words, we have regressions going to newer kernel versions and
would rather track them.

--
Best Regards,
Paul

[Alexander Sack (asac)]

paul, you see this in latest build i figure?

[Paul Sokolovsky (pfalcon)]

On Thu, 30 Jun 2011 08:54:53 -0000
Alexander Sack <email address hidden> wrote:

> paul, you see this in latest build i figure?

I tested against releases at
http://releases.linaro.org/platform/linaro-n/android/11.06/

--
Best Regards,
Paul

[Zach Pfeffer (pfefferz)]

Per monitor bugs are more useful than a blanket bug in this case. People should file a per monitor bug with all the specs of the monitor including serial and model number.