Mir

Merge lp://staging/~vanvugt/mir/scanout into lp://staging/~mir-team/mir/trunk

scanout
Merge into trunk

Proposed by Daniel van Vugt on 2013-09-11

Status:

Merged

Approved by:

Alexandros Frantzis on 2013-09-12

Approved revision:

no longer in the source branch.

Merged at revision:

1063

Proposed branch:

lp://staging/~vanvugt/mir/scanout

Merge into:

lp://staging/~mir-team/mir/trunk

Diff against target:

287 lines (+111/-3)

13 files modified

include/platform/mir/graphics/buffer_ipc_packer.h (+1/-0)
include/shared/mir_toolkit/mir_native_buffer.h (+10/-1)
include/test/mir_test_doubles/mock_buffer_packer.h (+1/-0)
src/client/mir_surface.cpp (+1/-0)
src/server/frontend/protobuf_buffer_packer.cpp (+6/-0)
src/server/frontend/protobuf_buffer_packer.h (+1/-0)
src/server/graphics/gbm/gbm_buffer.cpp (+1/-0)
src/server/graphics/gbm/gbm_platform.cpp (+1/-0)
src/shared/protobuf/mir_protobuf.proto (+2/-1)
tests/acceptance-tests/test_client_library.cpp (+71/-0)
tests/mir_test_framework/testing_server_options.cpp (+12/-1)
tests/unit-tests/frontend/test_protobuf_buffer_packer.cpp (+2/-0)
tests/unit-tests/graphics/gbm/test_gbm_platform.cpp (+2/-0)

To merge this branch:

bzr merge lp://staging/~vanvugt/mir/scanout

Related bugs:

Bug #1218735: [intel] Bypass mode makes the cursor break up into horizontal line artefacts	High	Invalid
Bug #1218815: [radeon] Graphic glitches and screen corruption (vertical lines) on XMir surfaces only	Critical	Invalid

Link a bug report

Reviewer	Review Type	Date Requested	Status
Robert Ancell			Approve on 2013-09-11
Kevin DuBois (community)			Approve on 2013-09-11
Alan Griffiths		2013-09-11	Approve on 2013-09-11
PS Jenkins bot (community)	continuous-integration		Approve on 2013-09-11
Review via email: mp+184961@code.staging.launchpad.net

Commit message

Add a "flags" field to MirBufferPackage so that clients can find out if the
buffer they've been given is scanout-capable. This is normally something a
client should never need to know. However there are two specialized cases
where it's required to fix bugs in the intel and radeon X drivers:
LP: #1218735, LP: #1218815
The intel fix (already landed) contains a hack which will be updated after
this branch lands.

Description of the change

The weird looking reshuffle in mir_native_buffer.h is necessary to avoid breaking the ABI for Mesa. It seems only Mesa is inflexible enough that we can't extend MirBufferPackage without breaking it. So we avoid growing the size of MirBufferPackage for now.

Revision history for this message

PS Jenkins bot (ps-jenkins) wrote on 2013-09-11:

PASSED: Continuous integration, rev:1067
http://jenkins.qa.ubuntu.com/job/mir-ci/1466/
Executed test runs:
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-android-saucy-i386-build/1981
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-clang-saucy-amd64-build/1866
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-saucy-amd64-ci/704
        deb: http://jenkins.qa.ubuntu.com/job/mir-saucy-amd64-ci/704/artifact/work/output/*zip*/output.zip

Click here to trigger a rebuild:
http://s-jenkins:8080/job/mir-ci/1466/rebuild

review: Approve (continuous-integration)

Revision history for this message

Alan Griffiths (alan-griffiths) wrote on 2013-09-11:

Looks messy but I guess it is justified in the description

review: Approve

Revision history for this message

Kevin DuBois (kdub) wrote on 2013-09-11:

the abi-preserving changes for the gbm buffer look good. Android has a flag on its 'usage' field in its native type that can indicate scanout surfaces

It would be nice to get rid of the #ifndef ANDROID around the test case... but looks okay otherwise.

review: Approve

Revision history for this message

Robert Ancell (robert-ancell) wrote on 2013-09-11:

The ABI preserving is a little bit unclear, it would be nice to add a big comment saying "we are stealing the last element of data[] and fd[] to use for flags - we can revert this once we bump ABI.

Otherwise makes sense to me

review: Approve