Commit Graph

  • b951c37fcf Wrap the fake backbuffer image view baldurk 2015-09-19 21:26:34 +02:00
  • 879dd422b5 Change descriptor set order back baldurk 2015-09-19 20:16:23 +02:00
  • 9bef542fc6 Fix typo - force initial states for real device memory baldurk 2015-09-19 17:58:09 +02:00
  • b4c00d9a7c Destroy instance through correct dispatch table baldurk 2015-09-19 17:48:02 +02:00
  • 98bfdd82db Fix in-application text rendering baldurk 2015-09-19 17:41:41 +02:00
  • 6a904ba032 Add a rather gross hack to get replay queues in advance and wrap later baldurk 2015-09-19 17:33:23 +02:00
  • 4db95cfe0c Add some missing wrap/unwraps baldurk 2015-09-19 17:32:45 +02:00
  • 949489b6b1 Don't dereference null pointer, use local device object baldurk 2015-09-19 17:23:48 +02:00
  • 802a5e4b0a Remove our own handling of dbg msg callbacks - come back to this later baldurk 2015-09-19 17:23:32 +02:00
  • 720a88f2f8 Serialise_vkCreateDevice needs to bootstrap with dummyDeviceTable baldurk 2015-09-19 17:20:03 +02:00
  • 58e0dfcdef Use VkInstance not ResourceId (problem with templated functions :() baldurk 2015-09-19 17:18:06 +02:00
  • d081efe65c GetLiveHandle / GetCurrentHandle should return wrapped object baldurk 2015-09-19 17:12:02 +02:00
  • 2559907c17 Resource tidy up isn't practical while writing as device is already dead baldurk 2015-09-19 17:11:47 +02:00
  • 2f8530b10b Tidy up destruction so all resources and records are deleted by shutdown baldurk 2015-09-19 17:02:37 +02:00
  • 1abe5c2c0d Handle releasing a NULL resource baldurk 2015-09-19 16:44:22 +02:00
  • 9260bfd0be Don't double-delete now that ReleaseWrappedResource does it all baldurk 2015-09-19 16:44:12 +02:00
  • fe9f89cb47 Only mark resources as frame referenced when actually frame capturing baldurk 2015-09-19 16:43:53 +02:00
  • 9aacd2ed5b Release baked commands when destroying command buffer baldurk 2015-09-19 16:43:39 +02:00
  • c1eecb28f0 Don't serialise initial states for debug-only objects baldurk 2015-09-19 16:42:50 +02:00
  • 4af3d1c00c Unwrap swapchain present info structure baldurk 2015-09-19 16:23:52 +02:00
  • ae0aef85c2 Fix small bugs baldurk 2015-09-19 16:14:55 +02:00
  • 7ba885fddd Use parent object when wrapping to fetch dispatch table baldurk 2015-09-19 16:13:33 +02:00
  • f3da5478fc Copy and unwrap any structures that contain object handles baldurk 2015-09-19 16:00:04 +02:00
  • 4254f54d92 Fix some more wrapping/unwrapping issues baldurk 2015-09-19 14:50:39 +02:00
  • 4816bcfa7b use layer specific dispatch table maps GregF 2015-09-18 15:22:24 -06:00
  • bc2dca08ba Fix linux compile issues with some temporaries baldurk 2015-09-18 19:07:35 +02:00
  • a00688a86a Add a little utility function so that code can get IDs from real objects baldurk 2015-09-17 19:29:34 +02:00
  • a6626ab31b Debug and replay properly wrapping and unwrapping everything baldurk 2015-09-17 19:12:32 +02:00
  • 3522ce25d5 Put 'real' handle into tiny struct similar to one in vulkan.h baldurk 2015-09-17 18:34:31 +02:00
  • 9dfcf7826b Fix 32-bit build baldurk 2015-09-17 18:32:25 +02:00
  • 725847e636 Move WrapResource in-place util function to manager baldurk 2015-09-17 17:56:49 +02:00
  • bad01d25fb Partial progress on wrapping/unwrapping objects correctly baldurk 2015-09-17 17:42:44 +02:00
  • 250b59e07c Move compile asserts out, as g++ can't use AllocByteSize in class def baldurk 2015-09-17 17:28:52 +02:00
  • 5bb681033d Add a handful of unwrapping up to instance/physical device point baldurk 2015-09-17 11:11:09 +02:00
  • 7afd99aff5 Use instance tables fetch instead of ObjDisp in vkCreateDevice/Instance baldurk 2015-09-17 11:10:53 +02:00
  • a6787430e8 Set up dispatch tables on creating wrapped resource, use ObjDisp baldurk 2015-09-17 09:36:17 +02:00
  • 93ad35e98a Make WrappedVulkan compile with wrapped resources baldurk 2015-09-16 20:12:22 +02:00
  • baba2a898b Make GetID(WrappedVkRes *) public since it's quite useful baldurk 2015-09-16 20:11:49 +02:00
  • 3a85a5f75b Add helper function to quickly get real from right place and cast baldurk 2015-09-16 20:11:34 +02:00
  • ba2473f27a Return new id from WrapResource baldurk 2015-09-16 20:10:57 +02:00
  • 302060bea6 ToStr function for VkResourceType baldurk 2015-09-16 20:10:50 +02:00
  • 4f744645e8 Add handy ObjDisp() function for dispatching via cached table baldurk 2015-09-16 19:10:49 +02:00
  • bd3f1c3002 Split dispatchable and non-dispatchable types up, to preserve tables baldurk 2015-09-16 18:56:11 +02:00
  • 9ab07a0ebb Add utility function for wrapping new resources baldurk 2015-09-16 18:11:51 +02:00
  • 55073df153 Refactor context record/id baldurk 2015-09-16 18:11:36 +02:00
  • 9b8431066a Add a couple of helpful utility functions for resources baldurk 2015-09-16 18:05:11 +02:00
  • 52c07cb789 Remove VkResource, switch to WrappedVkRes*, update all but vk_core.cpp baldurk 2015-09-16 16:06:18 +02:00
  • 6fa9007c2f Adjust pool counts/sizes baldurk 2015-09-16 15:35:55 +02:00
  • 2bbfa66d72 Add pool identification utility function baldurk 2015-09-16 15:16:36 +02:00
  • 585eba7c84 Add utility functions to unwrap, access ID or get record baldurk 2015-09-16 14:51:04 +02:00
  • e9342a32e7 Add VkWrappedRes struct with wrapping allocated pools baldurk 2015-09-16 13:30:06 +02:00
  • c5b3cf5849 Recompile text-rendering shaders to be LunarG sample driver friendly baldurk 2015-09-17 17:06:28 +02:00
  • eaf31123f7 Remove old code baldurk 2015-09-16 16:01:40 +02:00
  • 442386b84a Add text-rendering support for in-application renderdoc overlay baldurk 2015-12-31 22:29:38 +01:00
  • ad185a4042 A couple of runtime fixes, and create renderpass & framebuffers in-app baldurk 2015-09-16 12:14:17 +02:00
  • 332a281291 spin off 'debug data' into separate class so it can be used in-app baldurk 2015-09-16 12:10:05 +02:00
  • 2a043d6125 Implement getting backbuffer screenshot on capture baldurk 2015-09-16 11:54:14 +02:00
  • 088c3e1c4c Default to first swap image on replay output windows baldurk 2015-09-16 11:48:30 +02:00
  • 46d1d9c462 First pass at cleaning up resources. Remaining leak in mem initial state baldurk 2015-09-16 11:47:51 +02:00
  • 75a7a496f4 Release debug data (and wrap it up in its own struct) baldurk 2015-09-16 11:44:05 +02:00
  • 25cd48401b Change replay ID resource ID offset to decimal instead of hex baldurk 2015-09-16 11:36:56 +02:00
  • 429a93b4a0 Demote a VKTODO from HIGH to MED baldurk 2015-09-16 11:36:47 +02:00
  • 5a0e900a61 Fetch or add a queue family index that matches what we need baldurk 2015-09-16 11:36:31 +02:00
  • 3addb43f06 Warn about different API version, but proceed and use RenderDoc's baldurk 2015-09-16 11:28:37 +02:00
  • b14b149394 Dynamically allocate physical device array baldurk 2015-09-16 11:27:26 +02:00
  • 6d4cadf697 Remove VKTODOs about pNext/sType, it's handled in struct serialises baldurk 2015-09-16 11:23:01 +02:00
  • 1c0c09f4b9 Make sure to call vkResetCommandBuffer before re-begin'ing baldurk 2015-09-16 11:21:59 +02:00
  • e5e8fa29f5 Fix problem that might go away with resource record shutdown order baldurk 2015-09-16 11:17:17 +02:00
  • 3b3053369c Add frame references for current descriptor set bindings on queue submit baldurk 2015-09-16 11:17:06 +02:00
  • a675735499 Correctly query for backbuffer format and present mode from driver baldurk 2015-09-16 11:16:29 +02:00
  • 96e19cc4e9 List out descriptor set bindings and contents in vulkan pipeline state baldurk 2015-09-16 11:14:43 +02:00
  • 426aeb5a5e Keep track of current descriptor set bindings on replay baldurk 2015-09-16 11:14:14 +02:00
  • 26f773684a Track enough while READING to set drawcall topology/index width baldurk 2015-09-16 11:13:51 +02:00
  • 2eb87b804f Giant hack to set up 'fake' device/instance dispatch tables, for replay baldurk 2015-09-18 12:56:42 +02:00
  • 9b0b6ef3b6 Get compiling on win32 (hacks, not final code). baldurk 2015-09-18 11:40:11 +02:00
  • 5c8df903ef Revert "Use layer specific dispatch table maps" GregF 2015-09-18 16:52:07 -06:00
  • a8dfd36525 Use layer specific dispatch table maps GregF 2015-09-17 18:08:27 -06:00
  • e3f4a122fc Fixes to enable capture of cube with layer mechanism. GregF 2015-09-15 16:59:00 -06:00
  • 872a312af0 Use Vulkan layer to generate trace file GregF 2015-09-10 10:13:46 -06:00
  • 45d7b488d4 Typo fix - set handle to handle baldurk 2015-09-10 19:10:04 +02:00
  • 0bc36096f8 Track descriptor set updates and serialise out as initial states baldurk 2015-09-10 12:54:33 +02:00
  • ad687ae955 Pre-declare serialise of VkDescriptorInfo so it can be used elsewhere baldurk 2015-09-10 12:51:56 +02:00
  • 55ceffcdd2 In SerialiseObject explicitly handle VK_NULL_HANDLE baldurk 2015-09-10 12:51:44 +02:00
  • 8861110934 debug markers should be captured always when writing baldurk 2015-09-10 11:38:34 +02:00
  • c2623a55be Track descriptor set bindings as they are updated, don't record chunks baldurk 2015-09-10 11:38:24 +02:00
  • ab0bbf85a5 Store image type in image info baldurk 2015-09-10 10:11:57 +02:00
  • 0554c2779f Move platform-specific codepaths out into separate files baldurk 2015-09-10 09:50:54 +02:00
  • 53244eab7b Destroy output windows and all associated objects properly baldurk 2015-09-10 09:44:34 +02:00
  • e7a396b69c we need #includes to organise uniforms baldurk 2015-09-10 09:39:00 +02:00
  • 33eda64c8a Multiple maps don't happen baldurk 2015-09-09 10:32:30 +02:00
  • 8cec8cdd6a Consistently use VkResult vkr (res as a name is ambiguous) baldurk 2015-09-08 22:26:46 +02:00
  • 14e913ddeb Enable blending where needed for texture display baldurk 2015-09-08 20:09:41 +02:00
  • 1bb1168dae Change debug backbuffer clear from magenta to black baldurk 2015-09-08 20:03:44 +02:00
  • 0382a1bdce Move command buffer start drawcall to renderpass begin baldurk 2015-09-08 20:02:26 +02:00
  • 5db6e13505 Render output windows to off-screen backbuffer always, then copy in flip baldurk 2015-09-08 19:51:27 +02:00
  • c02d01c0b2 Check against special format, not is-special bool baldurk 2015-09-08 18:20:55 +02:00
  • 03a50d805e Use fake backbuffer format for image view baldurk 2015-09-08 18:20:40 +02:00
  • 85defbeeeb Linux compile fixes baldurk 2015-09-08 18:20:32 +02:00
  • ee6caa25c9 Implement window clearing (or checkerboard) showing up properly baldurk 2015-09-08 18:05:40 +02:00
  • 9853c73600 Try to properly calculate needed memory type index baldurk 2015-09-08 18:01:11 +02:00