[arrandale] Black screen on all outputs when external VGA is connected

There is a lot of noise in CurrentDmesg.txt because rtl8192 is spamming the kernel log, but this is the first abnormal drm message:

[ 4425.992922] [drm:i915_gem_do_execbuffer] *ERROR* Object ffff8801b0d60600 [handle 639, index 1] appears more than once in object list

Hi Jane and Matt,

It appears there is a fix for this in libdrm-2.4.24. I've packaged this in a PPA for testing convenience:
https://launchpad.net/~bryce/+archive/cherry

To verify the fix I think you could insert this step in the process:

  3.5. Install libdrm dev from https://launchpad.net/~bryce/+archive/cherry

Since this is an upstream cherrypick from a change in a recently released libdrm version, I feel pretty confident we can just roll this into the distro. However I see there's also a small number of other (possibly important) bugfixes in the new version, so I'm going to investigate if we should just update to the newer package.

  3.5. Install libdrm dev from https://launchpad.net/~bryce/+archive/cherry

^ that should be 'deb' not 'dev'.

The version in the PPA for this is 2.4.23-1ubuntu5, which includes the specific patch from 2.4.24 that fixes the problem.

On reviewing all the changes in libdrm 2.4.24 over what we have in natty, aside from this fix most of the remaining changes don't look important for natty. There are three nouveau patches for GeForce 400/500 chips that might be worth taking, however we there haven't been any LP bug reports filed for chips of that family (maybe folks are just installing nvidia).

So... I think for now I'll skip libdrm 2.4.24 and just do this one fix. We can backport other fixes as we need them. If anyone has strong feelings one way or the other on updating to the libdrm release let me know; we'd need to do a FFe, etc.

This bug was fixed in the package libdrm - 2.4.23-1ubuntu6

---------------
libdrm (2.4.23-1ubuntu6) natty; urgency=low

  * Add 100_intel_remember_named_bo.patch: Reuse already requested buffer
    objects if re-requested. Fixes error where dmesg shows "Object
    ... appears more than once in object list". Cherrypick from upstream
    libdrm 2.4.24 tree.
    (LP: #747205)
 -- Bryce Harrington <email address hidden> Fri, 01 Apr 2011 17:53:28 -0700

Re-opening this because the original problem still exists (fully updated Natty beta as of 6 April). If I boot with external monitor attached, the screen of laptop and monitor turn black as soon as I log in. If I boot without external monitor, I can log in and the desktop appears as expected. If I plug an external monitor or projector in after the desktop is running, then it either turns black or a blinking solid dark red colour.

Thanks Jane, I've notified Intel their fix didn't fix it.

We'd need a new set of logs to send upstream, so reproduce the issue and then collect the dmesg output and Xorg.0.log and attach here. Matt, can you give Jane a hand again? (running apport-collect would do, if it's possible to get to the machine while it's hung)

Architecture: amd64
CompizPlugins: [core,bailer,detection,composite,opengl,decor,mousepoll,vpswitch,regex,animation,snap,expo,move,compiztoolbox,place,grid,imgpng,gnomecompat,wall,ezoom,workarounds,staticswitcher,resize,fade,unitymtgrabhandles,scale,session,unityshell]
DRM.card0.DP.1:
 status: disconnected
 enabled: disabled
 dpms: Off
 modes:
 edid-base64:
DRM.card0.HDMI.A.1:
 status: disconnected
 enabled: disabled
 dpms: Off
 modes:
 edid-base64:
DRM.card0.LVDS.1:
 status: connected
 enabled: enabled
 dpms: Off
 modes: 1440x900
 edid-base64: AP///////wAwrhRAAAAAABQUAQOAGhB46uWVk1ZPkCgoUFQAAAABAQEBAQEBAQEBAQEBAQEB8BygoFCEGjAYEDYABaMQAAAY8BygoFCEGjAYEDYABaMQAAAYAAAADwCVCjKVCjIZAQAwZABVAAAA/gBMVEQxMjFFUTNCCiAgAKc=
DRM.card0.VGA.1:
 status: connected
 enabled: enabled
 dpms: Off
 modes: 1440x900 1440x900 1152x864 1024x768 1024x768 1024x768 832x624 800x600 800x600 800x600 640x480 640x480 640x480 640x480 720x400
 edid-base64: AP///////wAmzQFWAQEBAQMSAQNsNCF46lrVp1ZLmyQTUFS974BxT4GQgYCBjKlAswCVD5UAKDyAoHCwI0AwIDYAB0QhAAAaAAAA/wAwNTc0MjgwMzUwMjUxAAAA/QA4TB5TEQAKICAgICAgAAAA/ABQTEIyNDAzV1MKICAgAH8=
DistUpgraded: Log time: 2011-04-01 10:25:39.688526
DistroCodename: natty
DistroRelease: Ubuntu 11.04
DistroVariant: ubuntu
GraphicsCard:
 Intel Corporation Core Processor Integrated Graphics Controller [8086:0046] (rev 02) (prog-if 00 [VGA controller])
   Subsystem: Lenovo Device [17aa:215a]
InstallationMedia: Ubuntu 10.10 "Maverick Meerkat" - Release amd64 (20101007)
MachineType: LENOVO 5129CTO
Package: libdrm (not installed)
ProcEnviron:
 LANG=en_GB.UTF-8
 SHELL=/bin/bash
ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-2.6.38-8-generic root=UUID=52f3f8dd-ab0d-4f3b-8cef-9bc54492e659 ro quiet splash vt.handoff=7
ProcVersionSignature: Ubuntu 2.6.38-8.42-generic 2.6.38.2
Renderer: Unknown
Tags: unity-2d natty running-unity ubuntu
Uname: Linux 2.6.38-8-generic x86_64
UpgradeStatus: Upgraded to natty on 2011-04-11 (0 days ago)
UserGroups: adm admin cdrom dialout lpadmin plugdev sambashare
dmi.bios.date: 09/15/2010
dmi.bios.vendor: LENOVO
dmi.bios.version: 6QET53WW (1.23 )
dmi.board.name: 5129CTO
dmi.board.vendor: LENOVO
dmi.board.version: Not Available
dmi.chassis.asset.tag: No Asset Information
dmi.chassis.type: 10
dmi.chassis.vendor: LENOVO
dmi.chassis.version: Not Available
dmi.modalias: dmi:bvnLENOVO:bvr6QET53WW(1.23):bd09/15/2010:svnLENOVO:pn5129CTO:pvrThinkPadX201s:rvnLENOVO:rn5129CTO:rvrNotAvailable:cvnLENOVO:ct10:cvrNotAvailable:
dmi.product.name: 5129CTO
dmi.product.version: ThinkPad X201s
dmi.sys.vendor: LENOVO
version.compiz: compiz 1:0.9.4+bzr20110411-0ubuntu1
version.ia32-libs: ia32-libs 20090808ubuntu11
version.libdrm2: libdrm2 2.4.23-1ubuntu6
version.libgl1-mesa-dri: libgl1-mesa-dri 7.10.2-0ubuntu1
version.libgl1-mesa-dri-experimental: libgl1-mesa-dri-experimental N/A
version.libgl1-mesa-glx: libgl1-mesa-glx 7.10.2-0ubuntu1
version.xserver-xorg: xserver-xorg 1:7.6+4ubuntu3
version.xserver-xorg-video-ati: xserver-xorg-video-ati 1:6.14.0-0ubuntu4
version.xserver-xorg-video-intel: xserver-xorg-video-intel 2:2.14.0-4ubuntu6
version.xserver-xorg-video-nouveau: xserver-xorg-video-nouveau 1:0.0.16+git20110107+b795ca6e-0ubuntu7

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

Jane reproduced the problem, switched to a text console and ran apport-collect. Since the bug is now open on libdrm, only the libdrm hook was run, whereas the xserver-xorg-video-intel one was run originally. This means you don't have the same logs, but hopefully this is what you wanted. If not, please let us know.

I don't see anything of interest in the updated CurrentDmesg.txt or XorgLog.txt, but hopefully I'm wrong. :-)

hmm right.. they don't show much of interest :/

Is there a difference if the monitor is plugged in _after_ the desktop has loaded, will it again break once the system tries to use the screen?

I'm not sure if I understand your question, but I think the answer is "no". Specifically, I don't notice a difference in behaviour that is dependent on when the monitor is plugged in. I get the same black screen result if I
(a) plug the monitor in before logging in (i.e., it goes black after I press enter to actually log in), or
(b) if I am already logged in and plug the monitor in (i.e., it goes black).

In both cases, unplugging the monitor doesn't return to normal operations - I need to reboot/relogin. Let me know if there are any other logs or debugging steps you'd like me to take.

thanks,
Jane

On Tue, Apr 12, 2011 at 11:45:17AM -0000, Timo Aaltonen wrote:
> hmm right.. they don't show much of interest :/
>
> Is there a difference if the monitor is plugged in _after_ the desktop
> has loaded, will it again break once the system tries to use the screen?

What happens is:

1. Using the desktop normally
2. Plug in the external monitor
3. Black screens

We're not even getting the chance to launch the display properties dialog,
or anything. As soon as it's connected, the symptoms begin.

--
 - mdz

On Tue, Apr 12, 2011 at 12:23:40PM -0000, Jane Silber wrote:
> I'm not sure if I understand your question, but I think the answer is "no". Specifically, I don't notice a difference in behaviour that is dependent on when the monitor is plugged in. I get the same black screen result if I
> (a) plug the monitor in before logging in (i.e., it goes black after I press enter to actually log in), or
> (b) if I am already logged in and plug the monitor in (i.e., it goes black).
>
> In both cases, unplugging the monitor doesn't return to normal
> operations - I need to reboot/relogin. Let me know if there are any
> other logs or debugging steps you'd like me to take.

It might be worth trying this:

- reproduce the problem
- disconnect the VGA
- login on a text console via ctrl+alt+f1
- run "xrandr -display :0.0 > xrandr.txt"
- run "sleep 10;xrandr -display :0.0 --auto --verbose >> xrandr.txt 2>&1"
- within 10 seconds, switch back to X with ctrl+alt+f7
- wait 10 seconds

and see if that has any effect (what it *should* do is turn the built-in
screen back on).

Regardless, it would be useful to get the resulting xrandr.txt file and see
what it says.

--
 - mdz

A quick update of my experience -

When booting with an external monitor connected to my laptop, I get to the gdm login and successfully enter my password. Once the login event starts occuring, the laptop monitor goes black and the external monitor randomly and quickly flashes colors. I can switch to a terminal; switching back with ctrl+alt+f7 shows BOTH screens as black, but I will now seen a mouse cursor.

As a workaround I boot without the montor attached, go to System Properties > Multiple Screens, turn OFF the laptop monitor, plug in the external monitor.. and the display correctly comes up on the external. Enabling the laptop monitor again will reproduce the behavior described above.

I will be happy to do any diagnostic testing you need when I have access to my machine tonight around 5pm EST. Please advise as to any logs etc which may be of assistance. Thanks.

I tried mdz's suggestion from comment #31. After ctrl+alt+f7, the screen switched to what looked like a different text console and I saw a dozen or so status messages, the last two of which were "Stopping anac(h)ronistic cron" and "Starting CUPS printing spooler/server".

xrandr.txt attached.

Timo asked on IRC for:

intel_reg_dumper output before the problem occurs (sudo intel_reg_dumper > intel-regs-before.txt)
intel_reg_dumper output after the problem occurs (sudo intel_reg_dumper > intel-regs-after.txt)

"Since the bug is now open on libdrm, only the libdrm hook was run, whereas the xserver-xorg-video-intel one was run originally. This means you don't have the same logs, but hopefully this is what you wanted."

libdrm and -intel share the same apport hook (symlinks) so should be good, thanks.

Per Matt's request in comment #34 I have attached the before and after outputs from intel_reg_dumper.

The process I followed:
Boot system to login prompt
ctrl+alt+1 to prompt, login, execute 'sudo intel_reg_dumper > intel-regs-before.txt' and exit
ctrl+alt+7 to login prompt
login, laptop screen goes black, monitor begins flashing colors. Login sound is played.
ctrl+alt+1 to prompt, login, execute 'sudo intel_reg_dumper > intel-regs-after.txt'
reboot

Please advise if there are any additional diagnostic steps I can run. Thanks.

and here is the after file...

https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/759258 was marked as a duplicate, so it's seemingly the same bug. I didn't plug in an external monitor though, and I didn't get a a black screen, my got a while overlay for a few seconds, then it got back to normal and apport said that a system error occurred. I hope these logs can be of help.

Jeb: those are identical, but I sent one for upstream anyway.. thanks

https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/759258 was marked as a duplicate, so it's seemingly the same bug. I didn't plug in an external monitor though, and I didn't get a a black screen, my got a while overlay for a few seconds, then it got back to normal and apport said that a system error occurred. I hope these logs can be of help.

mine got a white overlay*

Oops. I seem to have submitted the same comment twice.

I'm going to bump this back from libdrm to -intel; given the libdrm patch seems to not solve it, there's not a particular reason to think the bug is in libdrm. If nothing else, the bug report will be easier to find if it's with the other -intel reports.

On Wed, Apr 13, 2011 at 08:05:47PM -0000, Timothy Kross wrote:
> https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-
> intel/+bug/759258 was marked as a duplicate, so it's seemingly the same
> bug. I didn't plug in an external monitor though, and I didn't get a a
> black screen, my got a while overlay for a few seconds, then it got back
> to normal and apport said that a system error occurred. I hope these
> logs can be of help.

Looking at the dupes, it appears apport is noticing the gpu lockup that
occurred with the original bug report and duping gpu lockups with the
same signature.

So, quite potentially there were two separate bugs going on:

  a) A gpu lockup (IPEHR: 0x7a005502)
  b) Black screen on VGA connection

The dupes to this bug report (so far) are all showing (a) but none have
much of a description so hard to say if they're seeing (b) as well.

However, separately there have been other reports of just issue (b):

750259 [arrandale] Dual screen not working/shows corruption (x86_64)
734756 [Arrandale] Blank screen when booting with external monitor (x86_64)
754799 [arrandale] external VGA output doesn't output anything (i686)
745112 [arrandale] desktop is messed up (goes black) when laptop is docked (x86_64)
756624 [pineviewgm] black screen attaching external monitor (x86_64)
758785 [arrandale] Display flickers or blank on login (x86_64)
761236 [Arrandale] unable to correctly enable monitors connected to VGA (x86_64)
762855 [sandybridge-m-gt2+] Display on external monitor (vga) is black on 13" MacBook Pro 8,1 (x86_64)

Possibly (a) and (b) are related, but for now let's treat them as
separate bugs.

I've reviewed the aforementioned bug reports. It is striking they have such similar symptoms and are mostly Arrandale/x86_64 specific, but there are subtle differences in symptoms so I'm unsure they're all duplicates.

There are a variety of Arrandale (aka Ironlake) kernel patches proposed upstream at http://lists.freedesktop.org/archives/intel-gfx/2011-April/010071.html.

Some of those patches are experimental bits, others are merely shuffling code around for refactoring, so it's difficult to spot the actual fixes we should care about.

In any case, I think the next action for this bug report is that we need people to test this intel-drm-next-proposed branch and tell us whether it fixes the issue; if so then the actual fixes can be git-bisected, or if not we should forward the bugs upstream.

On Mon, Apr 18, 2011 at 07:18:37PM -0000, Bryce Harrington wrote:
> Looking at the dupes, it appears apport is noticing the gpu lockup that
> occurred with the original bug report and duping gpu lockups with the
> same signature.
>
> So, quite potentially there were two separate bugs going on:
>
> a) A gpu lockup (IPEHR: 0x7a005502)
> b) Black screen on VGA connection
>
> The dupes to this bug report (so far) are all showing (a) but none have
> much of a description so hard to say if they're seeing (b) as well.

I fear I may have contributed to this confusion by filing the original
report by way of apport-gpu-error-intel.py rather than ubuntu-bug. I
(perhaps wrongly) inferred from the symptoms that there was a GPU lockup,
even though the error handler hadn't triggered, or else that the GPU dump
would be useful in diagnosing the problem regardless.

--
 - mdz

I have seen issues with Unity when the overall width of the combined screens exceeds the underlying hardware texture limit. Could you post the output of the (non-destructive) command below please (which would need running in a terminal window):

    glxinfo -l | grep GL_MAX_TEXTURE_SIZE

Thanks.

GL_MAX_TEXTURE_SIZE = 4096

The internal laptop display is 1440x900 and the external monitor is 1920x1200

--- intel_regs_before.txt 2011-04-19 16:35:21.000000000 +0100
+++ intel_regs_after.txt 2011-04-19 16:36:05.000000000 +0100
@@ -11,27 +11,27 @@
 DISPLAY_PORT_PLL_BIO: 0x00000000
 DISPLAY_PORT_PLL_BIO: 0x00000000
     FDI_PLL_FREQ_CTL: 0x00053687
- PIPEACONF: 0xc0000054 (enabled, active, 6bpc)
- HTOTAL_A: 0x063f059f (1440 active, 1600 total)
- HBLANK_A: 0x063f059f (1440 start, 1600 end)
- HSYNC_A: 0x05c705b7 (1464 start, 1480 end)
- VTOTAL_A: 0x039d0383 (900 active, 926 total)
- VBLANK_A: 0x039d0383 (900 start, 926 end)
- VSYNC_A: 0x038c0386 (903 start, 909 end)
+ PIPEACONF: 0xc0000000 (enabled, active, 8bpc)
+ HTOTAL_A: 0x081f077f (1920 active, 2080 total)
+ HBLANK_A: 0x081f077f (1920 start, 2080 end)
+ HSYNC_A: 0x07cf07af (1968 start, 2000 end)
+ VTOTAL_A: 0x04d204af (1200 active, 1235 total)
+ VBLANK_A: 0x04d204af (1200 start, 1235 end)
+ VSYNC_A: 0x04b804b2 (1203 start, 1209 end)
         VSYNCSHIFT_A: 0x00000000
- PIPEASRC: 0x059f0383 (1440, 900)
- PIPEA_DATA_M1: 0x7e1458c0 (TU 64, val 0x1458c0 1333440)
- PIPEA_DATA_N1: 0x0020f580 (val 0x20f580 2160000)
+ PIPEASRC: 0x077f04af (1920, 1200)
+ PIPEA_DATA_M1: 0x7e386580 (TU 64, val 0x386580 3696000)
+ PIPEA_DATA_N1: 0x0041eb00 (val 0x41eb00 4320000)
        PIPEA_DATA_M2: 0x00000000 (TU 1, val 0x0 0)
        PIPEA_DATA_N2: 0x00000000 (val 0x0 0)
- PIPEA_LINK_M1: 0x00012160 (val 0x12160 74080)
+ PIPEA_LINK_M1: 0x00025990 (val 0x25990 154000)
        PIPEA_LINK_N1: 0x00041eb0 (val 0x41eb0 270000)
        PIPEA_LINK_M2: 0x00000000 (val 0x0 0)
        PIPEA_LINK_N2: 0x00000000 (val 0x0 0)
             DSPACNTR: 0xd8004400 (enabled)
             DSPABASE: 0x00000000
- DSPASTRIDE: 0x00001800 (96)
- DSPASURF: 0x00539000
+ DSPASTRIDE: 0x00001e00 (120)
+ DSPASURF: 0x0c024000
          DSPATILEOFF: 0x00000000 (0, 0)
            PIPEBCONF: 0x00000000 (disabled, inactive, 8bpc)
             HTOTAL_B: 0x031f027f (640 active, 800 total)
@@ -100,11 +100,11 @@
    TRANSB_DP_LINK_N1: 0x00000000 (val 0x0 0)
    TRANSB_DP_LINK_M2: 0x00000000 (val 0x0 0)
    TRANSB_DP_LINK_N2: 0x00000000 (val 0x0 0)
- TRANSACONF: 0xc0000040 (enable, active)
+ TRANSACONF: 0xc0000000 (enable, active)
           TRANSBCONF: 0x00000000 (disable, inactive)
- FDI_TXA_CTL: 0xb0044000 (enable, train pattern not train, voltage swing 0.4V,pre-emphasis none, port width X1, enhanced framing enable, FDI PLL enable, scrambing enable, master mode disable)
+ FDI_TXA_CTL: 0xb00c4000 (enable, train pattern not train, voltage swing 0.4V,pre-emphasis none, port width X2, enhanced framing enable, FDI PLL enable, scrambing enable, master mode disable)
          FDI_TXB_CTL: 0x00044000 (disable, train pattern pattern_1, voltage swing 0.4V,pre-emphasis none, port width X1, enhanced framing enable, FDI PLL enable, scrambing enable, master mode disable)
- FDI_RXA_CTL: 0xb0022050 (enable, train pattern not train, port width X1, 6bpc,link_reverse_strap_overwrite no, ...

By running xrandr --auto when it's in the broken state, I can get the backlight to turn back on, and a mouse cursor appears, but no windows are drawn.

Something similar happens when I suspend and resume, but the mouse cursor appears over a text console background.

I'm still able to type into my fullscreen terminal window, so everything seems to be running OK, but I can't get the display to return to normal.

I'm able to get the internal display back to normal by restarting the X server from a text console.

At all points I've checked (before connecting external display, after connecting external display, after switching to text console, after restarting X) the i915_error_state sysfs file reads "no error state collected"

Here's what i915_gem_seqno is doing (there are two, in dri/0 and dri/64):

--- i915_gem_seqno.txt.old 2011-04-19 17:07:54.000000000 +0100
+++ i915_gem_seqno.txt 2011-04-19 17:08:10.000000000 +0100
@@ -1,12 +1,12 @@
-Current sequence (render ring): 187113
+Current sequence (render ring): 187195
 Waiter sequence (render ring): 0
-IRQ sequence (render ring): 187112
+IRQ sequence (render ring): 187194
 Current sequence (bsd ring): 0
 Waiter sequence (bsd ring): 0
 IRQ sequence (bsd ring): 0
-Current sequence (render ring): 187113
+Current sequence (render ring): 187195
 Waiter sequence (render ring): 0
-IRQ sequence (render ring): 187112
+IRQ sequence (render ring): 187194
 Current sequence (bsd ring): 0
 Waiter sequence (bsd ring): 0
 IRQ sequence (bsd ring): 0

As an experiment, we tried creating a new user and logging in with Unity, and the problem vanished! It drives the external monitor just fine.

If I then switch over to Jane's account, it exhibits the same problem.

But then I can switch back to my account, and it works again!

On Alberto's advice on IRC, I tried moving ~/.config/monitors.xml out of the way, and that has resolved the problem.

I'm attaching a copy of the old monitors.xml here

After moving monitors.xml out of the way, Jane went into the monitor preferences and put things back the way they were (with the laptop display switched off), the problem reoccurred.

Copying in the old monitors.xml reliably reproduces the problem.

mdz thinks this is the monitors.xml that corresponds to the test done in comment #59.

On Tue, Apr 19, 2011 at 09:45:42AM -0000, Andy Whitcroft wrote:
> I have seen issues with Unity when the overall width of the combined
> screens exceeds the underlying hardware texture limit. Could you post
> the output of the (non-destructive) command below please (which would
> need running in a terminal window):
>
> glxinfo -l | grep GL_MAX_TEXTURE_SIZE

Fwiw, yes I've seen this issue too, however for this particular bug
we're not running up against this (it's a single-external monitor
situation).

Here's a patched X driver with DPMS forcing code that might work around the kernel bug. YMMV. I'm having a few other folks on the similar arrandale bugs test this as well.

https://launchpad.net/~bryce/+archive/elderberry

You can also try to manually flip dpms on via this command:

 xset dpms force on

...however in testing I'm not finding this has much effect at all so YMM not V at all on that.

Here is another thing to test. This is a package of the current Intel drm experimental kernel tree:

  http://kernel.ubuntu.com/~kernel-ppa/mainline/drm-intel-next-proposed/

Please test the patched X driver in the previous comment first, then if that doesn't do it, revert back to stock -intel and then try this kernel.

Several other testers have found that drm-intel-next-proposed solves it.

There is a similar bug against sandybridge (#761065) which was isolated down to 4 kernel patches recently, provided in this test kernel: http://people.canonical.com/~acelan/bugs/lp753189/. Of those 4 patches, the fix is suspected to be the following kernel patch:

http://<email address hidden>/msg03038.html

Reviewing that patch, it looks to me this should similarly affect other >965 chips. So I have a hunch this may be the patch from drm-intel-next-proposed that we want. However, I'd want to see some additional testing to be certain.

Anyway, at the least I am 99% sure that a lot of these Arrandale/external-screen/blank issues are the same, so am going to dupe them together. 737891 has the best testing feedback so far, and is one of the oldest reported ones so I'll make that the master.

(Hmm, actually 737891 is a bit of a mess; I'll dupe to a more appropriate one.)