Chris Forbes [Wed, 29 Jun 2016 03:12:29 +0000 (15:12 +1200)]
layers: Require subpass attachments' sample counts to be consistent
Signed-off-by: Chris Forbes <chrisforbes@google.com>
Chris Forbes [Tue, 28 Jun 2016 23:50:31 +0000 (11:50 +1200)]
layers: Add validation for subpass multisample resolve rules
Signed-off-by: Chris Forbes <chrisforbes@google.com>
Chris Forbes [Thu, 30 Jun 2016 04:56:15 +0000 (16:56 +1200)]
layers: Make TransitionSubpassLayouts work with VK_ATTACHMENT_UNUSED
Signed-off-by: Chris Forbes <chrisforbes@google.com>
Chris Forbes [Thu, 30 Jun 2016 23:45:49 +0000 (11:45 +1200)]
layers: Fix crash in ValidateDependencies with VK_ATTACHMENT_UNUSED
Signed-off-by: Chris Forbes <chrisforbes@google.com>
Chris Forbes [Thu, 30 Jun 2016 23:28:00 +0000 (11:28 +1200)]
layers: Fix crash in ValidateLayouts with VK_ATTACHMENT_UNUSED
Signed-off-by: Chris Forbes <chrisforbes@google.com>
Chris Forbes [Thu, 30 Jun 2016 03:42:41 +0000 (15:42 +1200)]
layers: Apply subpass 0 layout transitions at CmdBeginRenderPass
Signed-off-by: Chris Forbes <chrisforbes@google.com>
Chris Forbes [Thu, 30 Jun 2016 00:43:36 +0000 (12:43 +1200)]
layers: Fix render pass begin handling of initial layout
Signed-off-by: Chris Forbes <chrisforbes@google.com>
Karl Schultz [Thu, 30 Jun 2016 22:34:29 +0000 (16:34 -0600)]
tests: Add missing TODO tests to layer details doc
Change-Id: I55fb267772e94f9228e0a8b19aae5049cfccfe02
Mark Young [Thu, 30 Jun 2016 19:02:42 +0000 (13:02 -0600)]
loader: GH370 - re-enable allocator usage
This reverts commit
5876cc2b7563f909c8ac8cb49f51f7ef715877f8.
This adds the previous fixes as well as additional fixes if Implicit
layers aren't present. This is not a case that gets tested on Windows
as most SDK installs have at least 1 Implicit layer (RenderDoc).
This change passes an individual run on Jenkins Linux system.
Change-Id: I466cc54b72946e2a9f6477bd4c3d4e4e72eb9579
Jon Ashburn [Thu, 30 Jun 2016 16:21:55 +0000 (10:21 -0600)]
layers: Fix compilation errors for KHR_display ext
For Android remove KHR_display entry points from intercept generation and
from header files.
Change-Id: Ic1386a94a2229c45a1ffaac3651326324105351e
Jon Ashburn [Thu, 30 Jun 2016 15:01:27 +0000 (09:01 -0600)]
layers: Add some comments to swapchain for more KHR_display validation
Change-Id: I9d5cf4f6f8f0f9c5177155deaa18fe521c7620a9
Jon Ashburn [Wed, 29 Jun 2016 22:12:51 +0000 (16:12 -0600)]
layers: Fix unique_objects creation of VkDisplayKHR
Make an explict_GetPhysicalDeviceDisplayProperties.
Change-Id: I39411d0dd97ca03dbf42d7e79b02243ef8988078
Jon Ashburn [Tue, 28 Jun 2016 22:59:36 +0000 (16:59 -0600)]
layers: FIx object_tracker displayKHR and displayModeKHR creation
For displayModeKHR there are two separate Vulkan commands that can
create these objects.
For displayKHR there is one command that creates the object:
GetPhysicalDeviceDisplayPropertiesKHR. GetDisplayPlaneSupportedDisplaysKHR
doesn't create new objects but returns existing objects. The existing code
had the creation in GetDisplayPlaneSupportedDisplaysKHR rather than in
GetPhysicalDeviceDisplayPropertiesKHR.
Change-Id: I63df7c41a0d17acd8b06ebb1a5742ac032159a09
Jon Ashburn [Tue, 28 Jun 2016 20:46:12 +0000 (14:46 -0600)]
layers: Fix GetDisplayModePropertiesKHR unique_objects code
Add the available DisplayModes to the map for use by CreateDisplayPlaneSurfaceKHR
Change-Id: I94754259069a844bb516d3cf9d4e130473dbc2e3
Jon Ashburn [Wed, 15 Jun 2016 14:19:07 +0000 (08:19 -0600)]
layers: Finish add support to validation for VK_KHR_display extension
Swapchain, object_tracker and unique_objects now intercept and perform
validation checks for this extension.
Change-Id: I2e78d1dc516103de165ef83b99665eaad05b73d9
Petros Bantolas [Fri, 15 Apr 2016 10:02:59 +0000 (11:02 +0100)]
Added WSI KHR_display support to validation
Add support for this extension to object_tracker and swapchain layers.
Change-Id: Ifdf6095958a9d9d01720fde92e0bbd0a8a9a3216
Mark Young [Thu, 30 Jun 2016 17:03:59 +0000 (11:03 -0600)]
Revert "loader: GH370 - re-enable allocator usage"
This reverts commit
0860f9dbac51b241a9690d23b53db14f307b0181.
Mark Young [Thu, 30 Jun 2016 00:33:53 +0000 (18:33 -0600)]
loader: GH370 - re-enable allocator usage
Re-enable reverted changes done in commit
d6f491b88eaf11f6953c02638d079f6a76806658. But also include
fixes for the Linux and Windows release runs.
Change-Id: I7644bb305faab068b3229eb9c1d8a67b052af165
Chris Forbes [Wed, 29 Jun 2016 23:56:45 +0000 (11:56 +1200)]
Revert "loader: GH370 - re-enable allocator usage"
This reverts commit
aa32dbbf147d86257a0c9e091258e0ea2be5a013.
This is broken in both debug and release 64bit linux.
Mun Gwan-gyeong [Sun, 26 Jun 2016 20:34:44 +0000 (05:34 +0900)]
demos: GH692 Added a supporting of wayland wsi
Added a supporting of wayland wsi to cube, tri and smoke.
Change-Id: Ifa291d334cb1ed54018573c970078051c6867d96
Signed-off-by: Mun Gwan-gyeong <elongbug@gmail.com>
Chris Forbes [Tue, 28 Jun 2016 22:50:41 +0000 (10:50 +1200)]
layers: Don't produce spurious validation errors for sample count
Previously we would emit a spurious validation error if the only
color or depth attachment references were VK_ATTACHMENT_UNUSED.
Signed-off-by: Chris Forbes <chrisforbes@google.com>
Mark Young [Fri, 3 Jun 2016 21:11:02 +0000 (15:11 -0600)]
loader: GH370 - re-enable allocator usage
Re-enable the allocator usage in the loader. Also, fix several
memory leaks. The leaks were especially noticeable during the Vulkan
conformance object_management.alloc_callback_fail.instance and
object_management.alloc_callback_fail.device tests because the
tests were designed to intentionally fail the Alloc calls.
This change now fully passes the Vulkan CTS
object_management.alloc_callback_fail tests when the loader
uses application-provided Allocators.
Change-Id: I03f2a09bc33259442e02c917a34d78f8937808d8
Karl Schultz [Wed, 29 Jun 2016 22:24:03 +0000 (16:24 -0600)]
misc: GH691 Remove duplicated line.
Change-Id: I5af06e880a7a71a42503e1ce6acdd1557e445169
Mark Lobodzinski [Tue, 28 Jun 2016 20:51:06 +0000 (14:51 -0600)]
layers: Update layer details doc
Updated new CV and PV checks
Change-Id: I3fb2a1ebe2319b07e146804b9d8714a58e4ad67a
Mark Lobodzinski [Wed, 29 Jun 2016 16:53:51 +0000 (10:53 -0600)]
layers: Take care of DevLimits cleanup items
Added some simplifying make_unique() calls, changed param_validation
Log_msg calls to use LayerName var, added variable init to
constructor, and fixed some c-style casts.
Change-Id: I718e0d51b62a4b91a4c9b3645ff796d114a4432f
Mark Lobodzinski [Tue, 28 Jun 2016 19:58:54 +0000 (13:58 -0600)]
layers: Move QueueFamilyProp Query checks into CV
Migrate CreateDevice-time QFP checks from device limits into
core_validation, fix up CV error enums.
Change-Id: I36458fd9f7666324faf69373dc581655a580a7f3
Mark Lobodzinski [Tue, 28 Jun 2016 17:51:16 +0000 (11:51 -0600)]
layers: Moved devlimits UpdateDescrSets to PV
DevLimits GetDeviceQueue validation was redundant, moved
updateDescriptorSets devlimits validation into parameter
validation.
Change-Id: Ic7bf2450c835cb5a35b687be5872cbbd0d70ac95
Mark Lobodzinski [Tue, 28 Jun 2016 17:23:08 +0000 (11:23 -0600)]
layers: Move devlimits validation into PV
Moved CreateRenderpass and BeginCommandBuffer validation into
parameter validation and hooked up enumeratePhysicalDevices.
Change-Id: Ie074b055c1d6252e503bce744cb7d718e6eb528f
Mark Lobodzinski [Tue, 28 Jun 2016 16:04:01 +0000 (10:04 -0600)]
layers: Add GetPhysDevQueueFamilyProps to CV
Added API, devlimits checks for this API, and did some renaming.
Change-Id: Ie27e8f5bd61f711b5c2121533ef92dd833678ee3
Mark Lobodzinski [Mon, 27 Jun 2016 22:47:24 +0000 (16:47 -0600)]
layers: Add EnumeratePhysicalDevices to CV
Change-Id: Ib7a635c4f0e42560198cdb208ce310b09648aca4
Mark Lobodzinski [Mon, 27 Jun 2016 21:24:48 +0000 (15:24 -0600)]
layers: Add DevLimits error codes to CV.h
Change-Id: I1c5a6eade6c1beeb9e9320357a000c3253e7d77c
Mark Lobodzinski [Mon, 27 Jun 2016 21:14:12 +0000 (15:14 -0600)]
layers: Move phys dev call state structs to CV
Change-Id: I92faf07ea2e47d2013bb206e04de26081049f79f
Mark Lobodzinski [Mon, 27 Jun 2016 21:13:00 +0000 (15:13 -0600)]
layers: Add phyDevProps call to Param Validation
Change-Id: I3a8249c741d1695675842f1be09d3d63a92f900d
Mark Lobodzinski [Fri, 24 Jun 2016 15:57:32 +0000 (09:57 -0600)]
layers: GH618, Correct WSI exension flag usage
Object_tracker had no notion of individual WSI extension flags,
and unique_objects was not checking them.
Change-Id: I93c1aa0c324aa602717f36e2975120dba8bc364e
Tobin Ehlis [Fri, 24 Jun 2016 23:22:16 +0000 (17:22 -0600)]
layers: Unify invalid command buffer handling
Created a single vector to track any cmd buffer dependencies that are destroyed
and thereby cause the cmd buffer to become invalid. Currently just tracking
the previous object types of descriptor sets and framebuffers, but will update
the tracking to include all cmd buffer dependencies in future commits.
Tobin Ehlis [Fri, 24 Jun 2016 20:41:20 +0000 (14:41 -0600)]
layers: Update cb_bindings to hold CB_NODE ptr
Store GLOBAL_CB_NODE ptr in cb_bindings to avoid additional look-ups.
Tobin Ehlis [Fri, 24 Jun 2016 20:28:59 +0000 (14:28 -0600)]
layers: Refactor framebuffer look-up
Make better use of getFramebuffer() and add framebuffer member to FRAMEBUFFER_NODE.
Tobin Ehlis [Fri, 24 Jun 2016 20:18:09 +0000 (14:18 -0600)]
layers: Make FB node inherit from BASE_NODE
Update FB_NODE to use cb_binding from BASE_NODE. Making this the common
way to track command buffers that objects are bound to.
Tobin Ehlis [Fri, 24 Jun 2016 18:31:29 +0000 (12:31 -0600)]
layers: Refactor command buffer look-up
Increase use of getCBNode() helper function.
Tobin Ehlis [Fri, 24 Jun 2016 17:16:36 +0000 (11:16 -0600)]
layers: Refactor image/buffer memory fetch
Kill get_mem_binding_from_object() and replace with getImageMemory() and getBufferMemory().
Those functions then both call a shared function to retrieve memory bindings from appropriate
underlying vulkan object. This is still early in the refactor and I hope to simplify these
further.
Tobin Ehlis [Thu, 23 Jun 2016 23:03:56 +0000 (17:03 -0600)]
layers: BUFFER_NODE clean-up
Initial work to prepare for larger overhaul of BUFFER_NODE as well
as unifying INVALID_CB treatment.
This adds cb_bindings set to BASE_NODE which will be used to track
command buffers that objects are bound to in order to flag them as
INVALID if objects are deleted while CB are still valid.
Initially making use of this with BUFFER_NODE, but it will move to
all objects that can be bound to command buffers.
Also, use getBufferNode() utility in DestroyBuffer() and add buffer
to BUFFER_NODE so it can be used for erasing node from map.
Tobin Ehlis [Tue, 21 Jun 2016 18:34:33 +0000 (12:34 -0600)]
layers: GH616 Improve descriptor set invalid binding handling
Prior to calls to get an global index from a binding, make sure that
HasBinding() is called to verify that the binding exists.
Added HasBinding() calls to a couple of validation functions for
early exist.
Updated code comment to indicate that HasBinding() call should
precede GetGlobal*IndexFromBinding() calls.
Added assert() to GetGlobal*IndexFromBinding() calls in the error case
to make error more obvious in debug builds.
Note that Perform*Update() functions already indicate that they should
be preceded by their matching Validate*Update() functions, which include
the appropriate checks for HasBinding().
case the Validate*Update() funciont
Cody Northrop [Fri, 24 Jun 2016 17:22:29 +0000 (11:22 -0600)]
toolchain: Simplify shaderc setup for Windows too, per
dc0ac6bc
Chris Forbes [Wed, 22 Jun 2016 22:36:09 +0000 (10:36 +1200)]
layers: Inline away markCommandBuffersInFlight
Signed-off-by: Chris Forbes <chrisforbes@google.com>
Chris Forbes [Wed, 22 Jun 2016 05:28:23 +0000 (17:28 +1200)]
layers: More use of getFenceNode / getQueueNode helpers
Signed-off-by: Chris Forbes <chrisforbes@google.com>
Chris Forbes [Tue, 21 Jun 2016 23:57:17 +0000 (11:57 +1200)]
layers: Make CB_SUBMISSION 1:1 with VkSubmitInfo
This gets rid of a bunch of weirdness where submission with no command
buffers (sema/fence only) would get mistracked, and cuts down on
resource counting churn.
Signed-off-by: Chris Forbes <chrisforbes@google.com>
Chris Forbes [Tue, 21 Jun 2016 06:59:28 +0000 (18:59 +1200)]
layers: Clean up around semaphore flagging
Signed-off-by: Chris Forbes <chrisforbes@google.com>
Chris Forbes [Tue, 21 Jun 2016 06:30:46 +0000 (18:30 +1200)]
layers: Slightly simplify logic in CreateRenderPass
Signed-off-by: Chris Forbes <chrisforbes@google.com>
Chris Forbes [Tue, 21 Jun 2016 06:27:03 +0000 (18:27 +1200)]
layers: misc minor tidying
Signed-off-by: Chris Forbes <chrisforbes@google.com>
Chris Forbes [Tue, 21 Jun 2016 06:17:41 +0000 (18:17 +1200)]
layers: Fix locking in ResetCommandPool
This got broken when some code moved around
Signed-off-by: Chris Forbes <chrisforbes@google.com>
Chris Forbes [Tue, 21 Jun 2016 06:10:47 +0000 (18:10 +1200)]
layers: Fix ResetFences validation vs update
Signed-off-by: Chris Forbes <chrisforbes@google.com>
Chris Forbes [Tue, 21 Jun 2016 02:32:00 +0000 (14:32 +1200)]
layers: Fix remaining direct uses of commandPoolMap
These would get us in trouble if the pool wasn't in the map.
Signed-off-by: Chris Forbes <chrisforbes@google.com>
Chris Forbes [Tue, 21 Jun 2016 01:24:04 +0000 (13:24 +1200)]
layers: Untangle check & update in FreeCommandBuffers
Fixes crash later in new test case
Signed-off-by: Chris Forbes <chrisforbes@google.com>
Chris Forbes [Tue, 21 Jun 2016 01:18:44 +0000 (13:18 +1200)]
layers: Untangle check from update around pool reset/destroy.
Signed-off-by: Chris Forbes <chrisforbes@google.com>
Chris Forbes [Tue, 21 Jun 2016 01:04:18 +0000 (13:04 +1200)]
layers: Move CMD_POOL_INFO to header, rename, add helper
Signed-off-by: Chris Forbes <chrisforbes@google.com>
Chris Forbes [Tue, 21 Jun 2016 00:35:16 +0000 (12:35 +1200)]
layers: Split check from clear of CB global inflight state
Tiny step toward untangling crashes after FreeCommandBuffers.
Signed-off-by: Chris Forbes <chrisforbes@google.com>
Tobin Ehlis [Thu, 23 Jun 2016 20:25:22 +0000 (14:25 -0600)]
layers: Clean up frameBufferMap unique_ptr use
Create local unique_ptr to FRAMEBUFFER_NODE and update that prior
to map insertion. Then move it directly into map in order to avoid
two map accesses.
Tobin Ehlis [Thu, 23 Jun 2016 14:23:25 +0000 (08:23 -0600)]
layers: Add Framebuffer dimension limits check
Make sure that requested framebuffer dimensions do not exceed the
underlying device limits.
Tobin Ehlis [Thu, 23 Jun 2016 13:49:12 +0000 (07:49 -0600)]
layers: Add FB attachment swizzle check
When creating a framebuffer, verify that all attachments were
created with the identity swizzle. This is a valid usage
restriction for vkCreateFramebuffer()
Tobin Ehlis [Wed, 22 Jun 2016 22:38:29 +0000 (16:38 -0600)]
layers: Add Framebuffer dimensions check
When creating a framebuffer, make sure that the dimensions for the
attachments are at least as big as the framebuffer dimensions.
Tobin Ehlis [Wed, 22 Jun 2016 22:12:58 +0000 (16:12 -0600)]
layers: Add INVALID_FRAMEBUFFER_CREATE_INFO error
Added new error enum DRAWSTATE_INVALID_FRAMEBUFFER_CREATE_INFO
and added initial check with this enum type that flags and error
if framebuffer attachment has non-1 mip levelCount.
Tobin Ehlis [Wed, 22 Jun 2016 20:16:06 +0000 (14:16 -0600)]
layers: Add sample count check at CreateFramebuffer time
When creating a framebuffer make sure that samples for its attachments
match samples for corresponding renderpass attachments. This is a
valid usage requirement for CreateFramebuffer.
Also need to store swapchain images with default samples value of
VK_SAMPLE_COUNT_1_BIT to avoid false errors.
Tobin Ehlis [Wed, 22 Jun 2016 19:32:23 +0000 (13:32 -0600)]
layers: Add format check between renderpass and framebuffer
Verify that corresponding attachment formats match between a framebuffer
and the renderpass it is being created with.
This is a valid usage requirement for vkCreateFramebuffer()
Tobin Ehlis [Wed, 22 Jun 2016 16:42:19 +0000 (10:42 -0600)]
tests: Added FramebufferCreateErrors test
Migrated AttachmentUsageMismatch test to FramebufferCreateErrors
and added a testcase where attachmentCount mis-matches between
the framebuffer and the renderpass.
Also updated documentation
Tobin Ehlis [Wed, 22 Jun 2016 16:02:02 +0000 (10:02 -0600)]
layers: Add framebuffer attachmentCount check
Verify that attachmentCount for a framebuffer matches the
attachmentCount for the renderPass that framebuffer is being
created with.
Tobin Ehlis [Wed, 22 Jun 2016 15:20:42 +0000 (09:20 -0600)]
layers: Refactor CreateFramebuffer
Migrate post-call CreateFramebuffer code into PostCallRecord* function.
Tobin Ehlis [Wed, 22 Jun 2016 15:06:48 +0000 (09:06 -0600)]
layers: Refactor CreateFramebuffer
Migrate validation code to PreCall* function.
Tobin Ehlis [Wed, 22 Jun 2016 14:29:24 +0000 (08:29 -0600)]
layers: Clean up frameBuffer map
Update frameBufferMap to store unique ptrs and kill a new/delete.
Add safe_ structs to FRAMEBUFFER_NODE for frameBuffer and renderPass
create info structs in order to correctly shadow the whole struct.
Mark Young [Thu, 23 Jun 2016 17:25:03 +0000 (11:25 -0600)]
docs: Fix broken link in LoaderAndLayerInterface
The Manifest file example link was broken.
Change-Id: I454392f14f526ef6fdf5b30d971bd06739acb994
Ian Elliott [Tue, 21 Jun 2016 17:08:43 +0000 (11:08 -0600)]
Layers: GH674: Improve locking in swapchain layer.
This fixes cases where the swapchain layer wasn't always unlocking before
returning from a function.
Mark Young [Wed, 22 Jun 2016 22:55:31 +0000 (16:55 -0600)]
winsdk: Fix multiple RT uninstall issue
Change originally by David Pinedo. If the same Vulkan Run-time
installer is executed multiple times, it will bump the install count
for the RT installation. There was a problem during uninstall
which failed to look at the install count and would always remove
some important files. This fixes that issue.
Change-Id: I4dda50b0dbdd9ec901aa25713be7d7348f3d67ec
Rene Lindsay [Wed, 22 Jun 2016 22:46:48 +0000 (15:46 -0700)]
demos: Check global extensions before creating surface
Tony Barbour [Wed, 22 Jun 2016 21:24:24 +0000 (15:24 -0600)]
layers: Update json files - VK_EXT_debug_report v3
Change-Id: Ifdabdd288b965b17d1af31c443909da7469df63a
Mark Young [Wed, 22 Jun 2016 21:25:00 +0000 (15:25 -0600)]
loader: Fix version info on manifest string.
Debug_report manifest string version was incorrect because of
the new version parsing code I had put in place.
Change-Id: I48d40b131bc3efd573d08576573ee9a4060a7332
David Pinedo [Wed, 22 Jun 2016 16:37:55 +0000 (10:37 -0600)]
windowsbuild: chmod -x update_external_sources.bat
Rene Lindsay [Tue, 21 Jun 2016 23:37:26 +0000 (17:37 -0600)]
demos: Add Format properties heading to vulkaninfo
Rene Lindsay [Tue, 21 Jun 2016 22:42:19 +0000 (16:42 -0600)]
demos: Put back the XSync, removed by previous commit
Change-Id: I8f4bbac6d124ef18f4fb3edd0da5acf308d6f7a1
Mark Lobodzinski [Tue, 21 Jun 2016 21:54:57 +0000 (15:54 -0600)]
Revert "layers: Add debug report header"
This reverts commit
2f50b5f4afa2198e13dfa8a80a460c15a6acc56c.
Change-Id: If67e7039a339ae39008a2c73f97b6a1eaedf1c3d
Mark Lobodzinski [Tue, 21 Jun 2016 21:53:37 +0000 (15:53 -0600)]
Revert "layers: Make vklayer_utils a shared library"
This reverts commit
5fa68457c16e57d90274d911adc49179024442b3.
Change-Id: I5bd6c6be23859f271c84aa44ab62cab6c518db5a
Mark Lobodzinski [Tue, 21 Jun 2016 21:51:12 +0000 (15:51 -0600)]
Revert "layers: Update docs for shared utils DLL"
This reverts commit
94dd6081959103c1c9b0bcfda173d4149d5555b9.
Change-Id: Ie5b437d822c5efe6baf0baf2830591f50bf3ee4d
Rene Lindsay [Tue, 21 Jun 2016 22:02:10 +0000 (15:02 -0700)]
demos: LX535 Remove deprecated device layers.
Rene Lindsay [Tue, 21 Jun 2016 21:58:57 +0000 (14:58 -0700)]
demos: Use XSync to fix intermittent xlib crash.
Chris Forbes [Tue, 21 Jun 2016 08:54:34 +0000 (20:54 +1200)]
layers: Fix CmdBeginRenderPass after non-inline renderpass in same CB
Fixes #667.
Signed-off-by: Chris Forbes <chrisforbes@google.com>
Mark Lobodzinski [Tue, 21 Jun 2016 16:28:30 +0000 (10:28 -0600)]
layers: GH669, Fix clearValueCount errors
Validation was assuming LOAD_OP_CLEAR for some attachments resulting
in bad validation messages.
Change-Id: I8ff75ca63f591e3002b1ee2b61a6844d0cf1c07d
Rene Lindsay [Tue, 21 Jun 2016 19:16:44 +0000 (12:16 -0700)]
demos: Fix intermittent crash on xlib
Tony Barbour [Tue, 21 Jun 2016 16:47:18 +0000 (10:47 -0600)]
smoke: Remove references to device layers
Change-Id: I2399cdefb76a086525fea29e71a7c321b5cfecdd
David Pinedo [Tue, 21 Jun 2016 15:51:48 +0000 (09:51 -0600)]
windowsbuild: +x of update_external_sources.bat
Mark Lobodzinski [Tue, 21 Jun 2016 00:49:25 +0000 (18:49 -0600)]
layers: Fix Windows crash in Dota2
In core_validation attachment_references_compatible(), having array
elements set to VK_ATTACHMENT_UNUSED was not handled resulting in
invalid memory references.
Change-Id: If78aea1467d191db905272bece16cc9f56643136
Karl Schultz [Tue, 21 Jun 2016 01:08:43 +0000 (19:08 -0600)]
demos: remove deprecated device layer code
Change-Id: I916f30da7e1b2d43826f3d2080fed9721b1796d6
Jeremy Hayes [Mon, 20 Jun 2016 22:30:21 +0000 (16:30 -0600)]
loader: update so version
Chris Forbes [Thu, 16 Jun 2016 05:16:00 +0000 (17:16 +1200)]
layers: Get rid of fence->swapchain link
Nothing used this after rationalizing the state tracking.
Change-Id: I67cc267336b5d64bea9645a8e0a072885db58c26
Signed-off-by: Chris Forbes <chrisforbes@google.com>
Chris Forbes [Thu, 16 Jun 2016 05:14:34 +0000 (17:14 +1200)]
layers: Untangle validate+update in QueuePresentKHR
Change-Id: I8f719c6e57cbab836b90f30414e6494a0643580e
Signed-off-by: Chris Forbes <chrisforbes@google.com>
Chris Forbes [Thu, 16 Jun 2016 00:20:32 +0000 (12:20 +1200)]
layers: Fix AcquireNextImageKHR handling of sync objects
Change-Id: I3e5a5d08763c0bf26c6c69aecf4bcef17d055a48
Signed-off-by: Chris Forbes <chrisforbes@google.com>
Chris Forbes [Thu, 16 Jun 2016 00:17:09 +0000 (12:17 +1200)]
layers: Add getSemaphoreNode helper
Change-Id: Iceda3b28474e289189ef1f97b271f9e9bb7b2083
Signed-off-by: Chris Forbes <chrisforbes@google.com>
Chris Forbes [Wed, 15 Jun 2016 23:47:24 +0000 (11:47 +1200)]
layers: Rework fence state tracking
Instead of having a combination of firstTimeFlag, in_use, and
needsSignaled, let's just have an enum with the three states a fence can
actually be in.
Change-Id: I6a40cd709717c75740c9977b149c4f71a02a27ab
Signed-off-by: Chris Forbes <chrisforbes@google.com>
Chris Forbes [Wed, 15 Jun 2016 03:48:52 +0000 (15:48 +1200)]
layers: Tidy decrementResources a bit
Change-Id: Ib8f9a917115323d8d43b971593b213ff04138a35
Signed-off-by: Chris Forbes <chrisforbes@google.com>
Chris Forbes [Wed, 15 Jun 2016 02:06:23 +0000 (14:06 +1200)]
layers: use getFenceNode in decrementResources
Change-Id: I8351405299f2f109489fdddfc111412164aded77
Signed-off-by: Chris Forbes <chrisforbes@google.com>
Chris Forbes [Wed, 15 Jun 2016 01:10:28 +0000 (13:10 +1200)]
layers: Pass layer_data to verifyWaitFenceState
Change-Id: Id3e0f09254197119b821710f0e9ce27d3458afc9
Signed-off-by: Chris Forbes <chrisforbes@google.com>
Chris Forbes [Wed, 15 Jun 2016 01:03:31 +0000 (13:03 +1200)]
layers: use getFenceNode in verifyWaitFenceState
Change-Id: I46f4ef4c225e199c0415507bd19dc4ffc98a77bc
Signed-off-by: Chris Forbes <chrisforbes@google.com>
Chris Forbes [Mon, 13 Jun 2016 23:05:39 +0000 (11:05 +1200)]
layers: Remove some MT leftovers
Change-Id: I2b62475951315c6b231342be6f7c74865e3ea2bb
Signed-off-by: Chris Forbes <chrisforbes@google.com>