[vivid-overlay] Servers can't start - input-stub.so fails to load on i386

Bug #1458689 reported by Ricardo Salveti
20
This bug affects 4 people
Affects Status Importance Assigned to Milestone
Canonical System Image
Fix Released
Critical
kevin gunn
Mir
Fix Released
Critical
Alan Griffiths
mir (Ubuntu)
Fix Released
Critical
Unassigned

Bug Description

phablet@ubuntu-phablet:~$ system-image-cli -i
current build number: 140
device name: generic_x86
channel: ubuntu-touch/rc-proposed/ubuntu
last update: 2015-05-25 12:02:10
version version: 140
version ubuntu: 20150525
version device: 20150210
version custom: 20150525

When starting unity8:

[1432591771.427440] Server: Starting
[1432591771.427679] Loader: Loading modules from: /usr/lib/i386-linux-gnu/mir/server-platform
[1432591771.427824] Loader: Loading module: /usr/lib/i386-linux-gnu/mir/server-platform/input-stub.so
[1432591771.428143] <WARNING> Loader: Failed to load module: /usr/lib/i386-linux-gnu/mir/server-platform/input-stub.so (error was:/usr/lib/i386-linux-gnu/mir/server-platform/input-stub.so: undefined symbol: _ZN3mir6events10make_eventExx17MirKeyboardActionjij)
[1432591771.428520] Loader: Loading module: /usr/lib/i386-linux-gnu/mir/server-platform/graphics-dummy.so
[1432591771.428875] Loader: Loading module: /usr/lib/i386-linux-gnu/mir/server-platform/graphics-android.so.2
[1432591771.429008] mirplatform: Found graphics driver: dummy
[1432591771.429180] mirplatform: Found graphics driver: android
[1432591771.429324] Platform Loader: Selected driver: android (version 0.13.1)

Related branches

Changed in mir (Ubuntu):
assignee: nobody → Andreas Pokorny (andreas-pokorny)
Changed in mir:
milestone: none → 0.14.0
assignee: nobody → Andreas Pokorny (andreas-pokorny)
Changed in mir (Ubuntu):
status: New → Confirmed
Changed in mir:
status: New → Confirmed
Revision history for this message
Andreas Pokorny (andreas-pokorny) wrote :

Has already been fixed in lp:mir.

Changed in mir:
status: Confirmed → Fix Committed
Changed in mir (Ubuntu):
assignee: Andreas Pokorny (andreas-pokorny) → nobody
Changed in mir:
assignee: Andreas Pokorny (andreas-pokorny) → nobody
Changed in mir:
importance: Undecided → Medium
assignee: nobody → Andreas Pokorny (andreas-pokorny)
Revision history for this message
Daniel van Vugt (vanvugt) wrote :

Wait a minute, fix committed where? What revision?

tags: added: lt-blocker lt-category-visible
summary: - [vivid-overlay] input-stub.so fails to load on i386
+ [vivid-overlay] Servers can't start - input-stub.so fails to load on
+ i386
Changed in mir:
importance: Medium → Critical
Changed in mir (Ubuntu):
importance: Undecided → Critical
Changed in canonical-devices-system-image:
status: New → Confirmed
tags: added: regression
Changed in canonical-devices-system-image:
assignee: nobody → kevin gunn (kgunn72)
importance: Undecided → Critical
milestone: none → ww34-2015
Revision history for this message
Andreas Pokorny (andreas-pokorny) wrote :

The error message in the log above means that the server hit input-stub.so while looking for graphics modules. It fails to load input-stub.so since it does not explicitly link to libmirclient.so which contains the given missing symbol. Additionally because qtmir is loaded as a plugin the input-stub.so cannot resolve that symbol through the already loaded libmirclient.so.

This warning message was fixed by adding the runtime dependency to stub-input.so. That fix will land with mir-0.14.0

Still this warning should not be related to any startup failure, since even due to the reported failure to probe stub-input the system picks a different graphics module. So something else is causing the failure.

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

From what I can find, the fix (if any) actually came in r2583 from:
lp:~alan-griffiths/mir/remove-ensure_loaded_with_rtld_global-hack-mostly

Changed in mir:
assignee: Andreas Pokorny (andreas-pokorny) → Alan Griffiths (alan-griffiths)
Revision history for this message
kevin gunn (kgunn72) wrote :

This should still be in mir0.14 then, which is stuck in wily proposed atm, will be in vivid+ next

Changed in canonical-devices-system-image:
status: Confirmed → In Progress
Revision history for this message
Launchpad Janitor (janitor) wrote :

This bug was fixed in the package mir - 0.14.0+15.10.20150722-0ubuntu1

---------------
mir (0.14.0+15.10.20150722-0ubuntu1) wily; urgency=medium

  [ Andreas Pokorny ]
  * Fix missing ABI renaming in Mirplatform
  * Bump Mirserver platform graphics to 3
  * Fix mirprotobuf ABI break

  [ CI Train Bot ]
  * New rebuild forced.

 -- CI Train Bot <email address hidden> Wed, 22 Jul 2015 18:01:49 +0000

Changed in mir (Ubuntu):
status: Confirmed → Fix Released
Changed in mir:
status: Fix Committed → Fix Released
Revision history for this message
kevin gunn (kgunn72) wrote :

waiting to get into image, this is in silo 11 for vivid+o

kevin gunn (kgunn72)
Changed in canonical-devices-system-image:
status: In Progress → Fix Committed
kevin gunn (kgunn72)
Changed in canonical-devices-system-image:
status: Fix Committed → Fix Released
tags: added: regression-release
removed: regression
Revision history for this message
Daniel van Vugt (vanvugt) wrote :

Please keep that tag. I use it for periodic quality analysis.

tags: added: regression
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.