[Git][xorg-team/vulkan/vulkan-loader][debian-unstable] 73 commits: test: Fix framework to work on MacOS

["Timo Aaltonen (@tjaalton)" <tjaalton@debian.org>]

Title: GitLab

Timo Aaltonen pushed to branch debian-unstable at X Strike Force / vulkan / vulkan-loader

Commits:

• f4054d56
  by Charles Giessen at 2021-06-22T12:03:05-06:00

  test: Fix framework to work on MacOS
  
  Main issue was getting function redireciton to work, dyld-interposer accomplishes that.
  It required making the shim-library a dylib, but otherwise only required minor changes.
  
  Change-Id: I7b7e2c86cc0c0b082f58241a7e146b74af72e792
  

• f16c6ab1
  by Charles Giessen at 2021-06-22T12:03:05-06:00

  ci: Enable regression tests on MacOS
  
  Change-Id: I85b08bfd2914e03840a027240470e2596415f4dc
  

• 008212c3
  by Charles Giessen at 2021-06-23T15:15:56-06:00

  test: Add revamped allocation tests
  
  Creates a memory tracker class and rewrites the existing allocation tests.
  
  Change-Id: I09ca2dc5b4d83f775ae0e3fd362a3fa4f7255d27
  

• 6e352029
  by Charles Giessen at 2021-06-23T15:15:56-06:00

  test: Update googlestest to 1.10.0
  

• e3fd8bde
  by Charles Giessen at 2021-06-23T15:15:56-06:00

  test: Delete old temp folders when starting tests/framework/test_util
  
  Make sure the work space is clean before starting a test, this is mainly for developer builds, as CI
  runs will generally wipe the entire build tree instead.
  

• 840f63c4
  by xantares at 2021-06-24T10:36:34-06:00

  Loader shared lib: back to default cmake prefix
  
  but it changed the import lib name (libvulkan.dll.a instead of libvulkan-1.dll.a) which in turns breaks cmake detection in FindVulkan
  now we go back to using the default cmake prefix which is equivalent to what #595 wants, (prior to #523), but without the broken behavior
  
  cc @Biswa96
  

• cb11f9b6
  by Bob Ellison at 2021-06-29T17:27:16-06:00

  copyrights: updated copyrights on MIT files
  
  During SDK testing we identified that the runtime copyrights
  were out-of-date.  The loader runtime includes two separate
  MIT copyrights, each attributed to a pair of source files,
  one initiated by Dave Gamble and one by joseph werle.
  
  I examined these files to discover that the last LunarG change to
  three out of the four was actually in 2017, not 2016.  This
  date should be attributed to Khronos, Valve, and LunarG.
  
  By updating the copyright notice in these files, we ensure that
  the license registry remains up-to-date.
  
  Note that no real content changed (which is why the copyrights are not
  being updated to 2021).
  

• 873040b7
  by Charles Giessen at 2021-06-29T18:36:52-06:00

  docs: Add info on DXGI adapters in Windows ICD Discovery
  

• b70cf11a
  by Mike Schuchardt at 2021-06-30T15:37:42-06:00

  build: Update to header 1.2.183
  
  - Update known-good
  - Generate source (no change)
  
  Change-Id: I059e6c2d8430228b4f20cadb83929b220efa5d0d
  

• 0a1a346b
  by Charles Giessen at 2021-07-06T10:03:29-06:00

  test: Remove Allocation tests from old test suite
  
  loader_alloc_callback_tests.cpp now contain all of the removed tests.
  
  Change-Id: Ia99df3a8f86d40f0d5e89a62b7b0a5f9bebafb7c
  

• 5097a3d4
  by Charles Giessen at 2021-07-06T10:03:29-06:00

  test: Rename regressions test cases
  
  >From RegressionTests.CreateInstanceExampleTestCase
  To CreateInstance.ExampleTestCase
  
  Change-Id: Ibd74541f93edbcaccb3daa9016dcacf8ecccfa78
  

• 8074954f
  by Charles Giessen at 2021-07-06T10:03:29-06:00

  test: Remove some redundant tests in old test suite
  
  Change-Id: Icc84b25bb092e15e054433651112e86b30726e52
  

• 436222fd
  by Charles Giessen at 2021-07-06T10:03:29-06:00

  test: Drop fp_ from function pointer names
  
  Change-Id: Ibd047e99f442d4a9755ac186e16b3620d3544e37
  

• 77f74505
  by Charles Giessen at 2021-07-06T10:03:29-06:00

  test: Use a new `string_eq` function instead of raw strcmp
  
  Change-Id: I703763133c42eb78ae3c6e194450411e3f784e35
  

• 0a8af1ba
  by Charles Giessen at 2021-07-06T10:03:29-06:00

  test: Update Framework Readme.md
  
  Change-Id: Iaf3740b046b3f6457c337988c573a73f068ff1a3
  

• ba70a411
  by Charles Giessen at 2021-07-06T10:03:29-06:00

  test: Use MACRO=1 in TestICD for definition style
  
  This replaces instances of `#ifdef MACRO_NAME` with `#if MACRO_NAME` as now
  MACRO_NAME will have a value of 0 or 1, instead of it being defined or not
  defined.
  
  Change-Id: I27a4330511a0f0d779f41851fe816fc79e868de2
  

• 9cd092f7
  by Charles Giessen at 2021-07-06T10:03:29-06:00

  test: Flesh out TestLayer to include relevant exports
  
  Functions needed in the various layer interface versions are now
  exported as well as guarded behind macros configurable from CMake.
  The export_definitions folder was populated as well.
  
  Change-Id: Iebfdd8ae02062eb81b8aa9f705885c9384c95be9
  

• 3adb0474
  by Charles Giessen at 2021-07-06T10:03:29-06:00

  test: Add Enumeration Function stubs to TestLayer
  
  Change-Id: I68b3597877fed803f4284e7e67f19ce4e3e2fe53
  

• 17bc7b4c
  by Charles Giessen at 2021-07-06T10:03:29-06:00

  test: Fix TestLayer exports on windows
  
  * `get_instance_func` needed to have __stdcall (VKAPI_ATTR) added
  * Several .def files were incorrect
  * vk_layer.h forward declares vkNegotiateLoaderLayerInterfaceVersion. This is
  problomatic because it doesn't have __declspec(dllexport), a token added by
  FRAMEWORK_EXPORT on windows. We have to special case the export macro here to
  prevent that mismatched definition.
  

• 7ae227e8
  by Mike Schuchardt at 2021-07-06T14:00:46-06:00

  build: Update to header 1.2.184
  
  - Update known-good
  - Generate source
  
  Change-Id: Ib2cfed88095541a127c23cf7418e9a66d81bd35e
  

• 8daad818
  by Mark Young at 2021-07-08T08:48:19-06:00

  Report a warning if the number of GPUs doesn't match
  
  We've encountered situations where a layer removes physical devices
  from the list which can cause issues on some systems if it's done
  incorrectly.  For now, simply throw a warning if the loader determines
  that a layer has removed devices.
  
  NOTE: In the long-run, I think we should define a policy which states
  that layers can re-order devices, but not add or remove.  If a layer
  wants to add, it should become an ICD.
  

• 550b23a5
  by Charles Giessen at 2021-07-12T16:23:15-06:00

  test: Move more old tests to new test framework
  
  Remove them from the calling scripts where appropriate. Notably the 'get count
  then run tests and pass in count' are now using the framework where that is
  completely redundant.
  
  Change-Id: I00d561a85047742a849df48818398b10c62d3b74
  

• 8c45e953
  by Mark Young at 2021-07-13T08:12:54-06:00

  Update contact information
  

• 7033871e
  by Charles Giessen at 2021-07-13T10:36:25-06:00

  ci: Re-add code and commit formatting verification
  
  Change-Id: I240f912daa03d001f2dbd2527d8374bbf394ac4a
  

• 15e1bbe9
  by Bob Ellison at 2021-07-13T16:31:46-06:00

  windows: update the loader copyright date
  
  The date, specified in loader/loader.rc, was incorrectly set to 2020.
  

• 99c0b143
  by Charles Giessen at 2021-07-14T12:55:05-06:00

  loader: Dont return OOM on function load failure
  
  Win7 does not support D3DKMTEnumAdapters2, as such when the loader goes to
  query it, it gets NULL back. Except, instead of just erroring out, it would
  return VK_ERROR_OUT_OF_HOST_MEMORY. Problem was that change
  73296a04678a07fc4fb1eef50cf4fe2778f059f8
  starts checking for OOM, and finding one when running in Win7. This change
  alters that behavior so that it doesn't report an OOM when one hasn't occured.
  

• 9d05cd96
  by Jamie Madill at 2021-07-20T11:42:47-06:00

  build: Disable -Wimplicit-fallthrough in GN.
  
  This flag was recently enabled by default in Chrome and ANGLE
  and other projects inherit these build settings. Explicitly
  dsiable this warning until it is enabled in the non-GN build.
  
  Change-Id: I8f353f8abc3f7857481ff9cc774aff3618eaf35a
  

• a0c69f0d
  by Mike Schuchardt at 2021-07-20T15:29:56-06:00

  build: Update to header 1.2.185
  
  - Update known-good
  - Re-enable code generation for VK_HUAWEI_subpass_shading
  - Generate source
  
  Change-Id: I6607e1de097d1b8a2f631cd1f11b088204b8bb3a
  

• 99a74402
  by Mike Schuchardt at 2021-07-27T13:27:49-07:00

  build: Update to header 1.2.186
  
  - Update known-good
  - Generate source (no change)
  
  Change-Id: I7c0699f8613ecbfefe79981f34faa5e276703324
  

• 3aeed853
  by Charles Giessen at 2021-07-30T10:33:15-06:00

  test: Add basic threading test
  
  Requires adding ability for device to return functions with
  vkGetDeviceProcAddr that aren't necessarily useable functions.
  
  Also added checks to make sure that the handle passed in by the loader
  to the TestICD is the 'right handle' where necessary, as some functions
  do not need to know which device is being called on, just that it is correct.
  

• e07667a9
  by Mark Young at 2021-08-02T16:19:55-06:00

  Fix loader debug documentation
  
  The loader debug environment variable (VK_LOADER_DEBUG) is not inclusive
  of higher importance messages, it is exclusive to the types specifically
  enabled.
  

• fa67fe73
  by Mike Schuchardt at 2021-08-03T15:16:04-07:00

  build: Update to header 1.2.187
  
  - Update known-good
  - Generate source (no change)
  

• 56722cb3
  by Charles Giessen at 2021-08-04T09:57:16-06:00

  test: Implemente layer create instance/device
  
  Added the necessary bits to make calling create instance and device function,
  though it may not work in all circumstances
  

• 1550eeb1
  by Charles Giessen at 2021-08-04T09:57:16-06:00

  test: add CreateInstance.LayerNotPresent test
  
  Remove the corresponding test in loader_validation_tests.cpp
  

• 9d4e0bdc
  by Charles Giessen at 2021-08-04T09:57:16-06:00

  test: Add noexcept in test_environment
  

• 63a9f8dc
  by Charles Giessen at 2021-08-04T09:57:16-06:00

  test: Update framework_config to include paths to layer binaries
  

• 3d635ace
  by Charles Giessen at 2021-08-04T09:57:16-06:00

  test: Add CreateInstance.LayerPresent test
  

• ab977626
  by Charles Giessen at 2021-08-04T09:57:16-06:00

  test: Make all tests call vkDestroyInstance
  
  This caused spurious failures in other tests due to the dynamic library
  not being closed.
  

• 78c3e744
  by Charles Giessen at 2021-08-04T09:57:16-06:00

  test: Update tests/README.md
  
  Include note to always destroy the instance at the end of a test.
  

• f7aab801
  by Charles Giessen at 2021-08-04T09:57:16-06:00

  test: Clear TestICDs before each test
  
  Calling `dlclose` doesn't necessarily mean that the static variables will
  be reset nor the libraries destructor gets called. This caused cascading
  errors in tests on linux with GCC. By always manually resetting the ICD
  or Layer before each test, we prevent any funny business from occuring.
  

• 4c901a73
  by Samiullah Khawaja at 2021-08-04T14:09:17-06:00

  Deallocate the extension lists when deleting an item from layer list
  
  Vulkan Loader loads the meta layer list and filters the ones that are
  not valid. During filter it deallocates the attributes of the loaded
  meta layer but does not deallocate the associated extension list.
  

• 6259c575
  by Charles Giessen at 2021-08-06T16:00:06-06:00

  test: Add simple invalid-meta layer test
  
  Creates a meta layer whose component layers do not exist
  

• 363f736a
  by Charles Giessen at 2021-08-06T16:00:06-06:00

  test: Make invalid meta layer test more comprehensive
  
  This required fixing a few bugsin the Manifest file printing logic
  

• 1f522f55
  by Charles Giessen at 2021-08-06T16:00:06-06:00

  test: Make MetaLayer test add extensions
  
  Needed to fixup more of the manifest writer in the process
  

• a2edd936
  by Charles Giessen at 2021-08-06T16:00:06-06:00

  test: Prevent macro collision
  

• 9e438397
  by Mark Young at 2021-08-10T12:10:47-06:00

  Fix loader_log messages to use LOADER_ log levels
  
  The loader_log messages were using the DEBUG_REPORT log levels
  which just happened to coincide with the existing LOADER_ log levels.
  Going forward, to add layer logging in its own level, this would not
  have worked.
  

• 1941128d
  by Mark Young at 2021-08-10T12:10:47-06:00

  Add layer and implementation-specific logging
  
  Add logging that will allow end-users to specifically just look at
  messages for layers and implementations/ICDs.
  Setting VK_LOADER_DEBUG to include one of the following:
    - layer        : enables layer-specific logging
    - implem       : enables implementation/ICD-specific logging
  

• b0841ad1
  by Mark Young at 2021-08-10T12:10:47-06:00

  Allow "icd" as well as "implem" for VK_LOADER_DEBUG
  

• 1d30345f
  by Mike Schuchardt at 2021-08-10T12:40:45-07:00

  scripts: Use altlen instead of parsing latexmath
  
  KhronosGroup/Vulkan-ValidationLayers switched to this method a while ago
  and this syncs the KhronosGroup/Vulkan-Loader version of
  helper_file_generator.py to match.
  

• fbb56aac
  by Mike Schuchardt at 2021-08-10T12:40:45-07:00

  build: Update to header 1.2.188
  
  - Update known-good
  - Generate source (no change)
  

• baf0ffd7
  by Mark Young at 2021-08-11T10:20:01-06:00

  Fix environment test failure.
  
  It was looking for a specific output for the manifest file search string.
  Re-add that string for now.
  

• 773f1d20
  by Charles Giessen at 2021-08-11T15:16:33-06:00

  test: Revamp which linux paths are redirected
  
  Adds FALLBACK_DATA_DIRS, FALLBACK_CONFIG_DIRS, SYSCONFDIR, and EXTRASYSCONFDIR
  to the list of redirected paths. They were previously 'covered' by manual entries.
  This commit makes them responsive to changes in the build system.
  
  This commit also makes sure that the path used for redirection is one searched by
  the loader by using the SYSCONFIG variable. Previously it would use /usr/local/etc/
  which was only correct on systems where CMAKE_INSTALL_PREFIX was set to /usr/local.
  

• 52859e76
  by Charles Giessen at 2021-08-11T15:16:56-06:00

  loader: Reformated loader source code
  
  Previously, clang-format was required for all new commits but the codebase
  itself wasn't conformant. This commit formats the source files in the loader
  folder.
  

• de99f8f6
  by Charles Giessen at 2021-08-11T15:17:20-06:00

  loader: Make use of a generated header version
  
  This changes the loader to always set the version to the one which was used when
  generating the source files that are checked into the repo, instead of using the
  version gotten from the Vulkan-Headers find cmake file.
  
  Change-Id: Id0955ddfd10e35e0f358f5a77799d8baa4992b04
  

• 85765687
  by Charles Giessen at 2021-08-16T11:38:17-06:00

  test: Add ELF test binaries for 32 and 64 bit
  
  These binaries are necessary for testing whether the loader correctly
  ignores binaries of the wrong architecture. Since they are data files
  more than executables, they are being checked in.
  
  Currently only Unix binaries are added but in the future windows along
  with any other platform binaries will be added.
  

• 64665f7d
  by Charles Giessen at 2021-08-16T11:38:17-06:00

  test: Add linux x64 and x86 for wrong binaries
  
  Test only looks for ICD binaries.
  

• f06143bf
  by Charles Giessen at 2021-08-16T11:38:17-06:00

  test: Add windows 32/64 bit binary tests
  
  Added both layer and ICD tests. Included lengthy description of why the test
  successfully creates an instance even though it really shouldn't.
  

• f1accaba
  by Charles Giessen at 2021-08-16T11:38:17-06:00

  test: Simplify wrong arch tests by adding a macro
  
  Added the `CURRENT_PLATFORM_DUMMY_BINARY` macro which thanks to platform
  defines contains the path of the binary which is 'wrong' for the current
  architecture. EX: compiling with Win64, it points to the Win32 binary, and vice
  versa when compiling for Win32. Same logic applies for linux 64/32 bit.
  

• 024535bd
  by WenqingLiAMD at 2021-08-16T11:42:09-06:00

  Fix Vulkan CTS testcase bug: "create_instance_device_intentional_alloc_fail"
  
  In function loader_scanned_icd_add, if open_library fails, VK_SUCCESS
  is still returned, which would cause the bad number of good ICDs.
  

• de37ef3a
  by Charles Giessen at 2021-08-16T13:40:34-06:00

  loader: Fix accidental error propagation
  
  When a dll/so of the wrong architecture is found, the loader would
  return ERROR_INCOMPATIBLE_DRIVER from `loader_scanned_icd_add`.
  Previously it would return VK_SUCCESS, but that caused the logic to
  not skip over the ICD. This commit fixes the issue where the error
  code would be propagated out of the function, which shouldn't happen
  since there may be other drivers which do successfully load.
  

• 55e60732
  by Charles Giessen at 2021-08-16T15:18:43-06:00

  test: Add test for PR #639
  
  This test tries to recreate an out of memory condition which causes
  a binary of the wrong type to be confused with a valid binary, which
  causes VK_ERROR_INCOMPATIBLE_DRIVER to be returned instead of VK_SUCCESS
  (Or OOM if the case may be)
  

• 413673b2
  by Charles Giessen at 2021-08-16T15:18:43-06:00

  test: Use proper allocation count in tests
  
  The test used to only keep track of the maximum number of allocations and fail
  if it was exceeded, as calling free would decrease this number. Now tests use
  the total number of times allocate was called, and realloc was added to this
  metric. This should offer better OOM coverage due to failing at each and
  every possible OOM place. The code does make sure not to increase the count
  if realloc was called to 'downsize' the allocation, which shouldn't cause
  OOM to occur ever.
  

• b926e416
  by Charles Giessen at 2021-08-16T18:06:03-06:00

  test: Update Device Layers Match test
  
  Move the old EnumerateDeviceLayers.LayersMatch to the new framework,
  implementing the necessary components in the TestLayer code path.
  

• 8a3a1676
  by Charles Giessen at 2021-08-16T18:06:03-06:00

  test: Set VK_LAYER_PATH properly
  

• 90fd66f6
  by Mike Schuchardt at 2021-08-17T12:55:15-07:00

  build: Update to header 1.2.189
  
  - Update known-good
  - Generate source
  

• 56612520
  by Charles Giessen at 2021-08-26T10:51:15-06:00

  loader: Downgrade wrong arch errors on windows
  
  On linux, when the loader fails to load a .so if it was due to being on the
  wrong architecture, the loader downgraded the error to INFO level. This
  commit does the same thing for windows. Error code 193 is winapi's way to
  signify that the dll failed to load cause of the architecture.
  

• 669c41d6
  by Charles Giessen at 2021-08-26T10:51:15-06:00

  test: Add DebugUtilsWrapper
  
  This wrapper allows easy reading of the log output through the debug utils
  messenger infrastructure. Since getting the stdout/stderr is not well
  supported by googletest, this allows tests to be written which check for
  specific log messages to be emitted, allowing another way to verify the
  loader is behaving as expected.
  

• d7c8139c
  by Charles Giessen at 2021-08-26T10:51:15-06:00

  test: Update 32/64 tests to check the log
  
  Use the DebugUtilsLogger to verify that the loader is downgrading the error
  level of dll/so's from ERROR to INFO
  

• 39dd126e
  by Charles Giessen at 2021-08-26T11:38:55-06:00

  build: Silence MSVC warning of replacing /GR
  

• a628bea1
  by Timo Aaltonen at 2021-09-07T13:36:36+03:00

  Merge branch 'upstream-unstable' into debian-unstable
  

• b9ca2dc3
  by Timo Aaltonen at 2021-09-07T13:37:09+03:00

  bump the version
  

• ce1cdaec
  by Timo Aaltonen at 2021-09-07T13:38:05+03:00

  Refresh vulkan-headers to 1.2.189.0
  

• e9f29857
  by Timo Aaltonen at 2021-09-07T13:40:46+03:00

  release to sid
  

30 changed files:

• .gitattributes
• .github/workflows/build.yml
• BUILD.gn
• BUILD.md
• CMakeLists.txt
• GOVERNANCE.md
• README.md
• debian/changelog
• loader/CMakeLists.txt
• loader/LoaderAndLayerInterface.md
• loader/adapters.h
• loader/asm_offset.c
• loader/cJSON.c
• loader/cJSON.h
• loader/debug_utils.c
• loader/extension_manual.c
• loader/extension_manual.h
• + loader/generated/loader_generated_header_version.cmake
• loader/generated/vk_dispatch_table_helper.h
• loader/generated/vk_layer_dispatch_table.h
• loader/generated/vk_loader_extensions.c
• loader/loader.c
• loader/loader.h
• loader/loader.rc
• loader/murmurhash.c
• loader/trampoline.c
• loader/unknown_ext_chain.c
• loader/unknown_ext_chain_masm.asm
• loader/vk_loader_layer.h
• loader/vk_loader_platform.h

The diff was not included because it is too large.

—
View it on GitLab.
You're receiving this email because of your account on salsa.debian.org. If you'd like to receive fewer emails, you can adjust your notification settings.