Rework crashes on window resize

[beezir]

Not sure where the issue is, but the code currently crashes for me on swapchain recreate. It was not doing this with the previous versions of the code. The following crash occurs using the exact code in the tutorial as of right now.

thread 'main' panicked at 'called Result::unwrap()on anErr value: AccessError { error: SwapchainImageAcquireOnly, command_name: "vkCmdBeginRenderPass", command_param: "attachment 0", command_offset: 0 }', libcore\result.rs:945:5 stack backtrace: 0: std::sys::windows::backtrace::unwind_backtrace at C:\projects\rust\src\libstd\sys\windows\backtrace\mod.rs:65 1: std::sys_common::backtrace::_print at C:\projects\rust\src\libstd\sys_common\backtrace.rs:71 2: std::sys_common::backtrace::print at C:\projects\rust\src\libstd\sys_common\backtrace.rs:59 3: std::panicking::default_hook::{{closure}} at C:\projects\rust\src\libstd\panicking.rs:211 4: std::panicking::default_hook at C:\projects\rust\src\libstd\panicking.rs:227 5: std::panicking::rust_panic_with_hook at C:\projects\rust\src\libstd\panicking.rs:511 6: std::panicking::continue_panic_fmt at C:\projects\rust\src\libstd\panicking.rs:426 7: std::panicking::rust_begin_panic at C:\projects\rust\src\libstd\panicking.rs:337 8: core::panicking::panic_fmt at C:\projects\rust\src\libcore\panicking.rs:92 9: core::result::unwrap_failed<vulkano::command_buffer::traits::CommandBufferExecError> at C:\projects\rust\src\libcore\macros.rs:26 10: core::result::Result<vulkano::command_buffer::traits::CommandBufferExecFuture<vulkano::sync::future::join::JoinFuture<alloc::boxed::Box<GpuFuture>, vulkano::swapchain::swapchain::SwapchainAcquireFuture<winit::Window>>, alloc::arc::Arc<vulkano::command_buffer::auto::AutoCommandBuffer<vulkano::command_buffer::pool::standard::StandardCommandPoolAlloc>>>, vulkano::command_buffer::traits::CommandBufferExecError>::unwrap<vulkano::command_buffer::traits::CommandBufferExecFuture<vulkano::sync::future::join::JoinFuture<alloc::boxed::Box<GpuFuture>, vulkano::swapchain::swapchain::SwapchainAcquireFuture<winit::Window>>, alloc::arc::Arc<vulkano::command_buffer::auto::AutoCommandBuffer<vulkano::command_buffer::pool::standard::StandardCommandPoolAlloc>>>,vulkano::command_buffer::traits::CommandBufferExecError> at C:\projects\rust\src\libcore\result.rs:782 11: triangletest::HelloTriangleApplication::draw_frame at .\src\main.rs:607 12: triangletest::HelloTriangleApplication::main_loop at .\src\main.rs:573 13: triangletest::main at .\src\main.rs:649 14: std::rt::lang_start::{{closure}}<()> at C:\projects\rust\src\libstd\rt.rs:74 15: std::rt::lang_start_internal::{{closure}} at C:\projects\rust\src\libstd\rt.rs:59 16: std::panicking::try::do_call<closure,i32> at C:\projects\rust\src\libstd\panicking.rs:310 17: panic_unwind::__rust_maybe_catch_panic at C:\projects\rust\src\libpanic_unwind\lib.rs:105 18: std::panicking::try at C:\projects\rust\src\libstd\panicking.rs:289 19: std::panic::catch_unwind at C:\projects\rust\src\libstd\panic.rs:392 20: std::rt::lang_start_internal at C:\projects\rust\src\libstd\rt.rs:58 21: std::rt::lang_start<()> at C:\projects\rust\src\libstd\rt.rs:74 22: main 23: invoke_main at f:\dd\vctools\crt\vcstartup\src\startup\exe_common.inl:64 24: __scrt_common_main_seh at f:\dd\vctools\crt\vcstartup\src\startup\exe_common.inl:259 25: BaseThreadInitThunk 26: RtlUserThreadStart

[bwasty]

Interesting, will try to look into it later today. Haven't tested on Windows yet, actually.

[beezir]

This may be card-specific. I reran the same built binary on my laptop and it works just fine. Laptop has Intel GPU, my other computer is nvidia.

[bwasty]

Ok, I can reproduce it on macOS/Intel and confirm it worked before the rework - but I don't know what's different yet. And the error seems weird - the comment for it says Trying to use a swapchain image without depending on a corresponding acquire image future., but it's joined just before:

vulkan-tutorial-rs/src/main.rs

Lines 607 to 610 in e58aac8

	let future = self.previous_frame_end.take().unwrap()
	.join(acquire_future)
	.then_execute(self.graphics_queue.clone(), command_buffer)
	.unwrap()

[bwasty]

Found the problem - see linked commit above.

[beezir]

Very nice find, I completely missed that.

On an unrelated side note found while tinkering with this issue - I may be entirely wrong, but I don't think you need the unsafe device wait line in the swapchain recreation. Vulkano appears to do all the sync needed with the previous frame cleanup and the passing of the old swapchain during recreation. I am unfortunately too new in both Rust and Vulkan to know for sure, but it works in my limited testing just fine without any validation errors or crash, and the documentation indicates that you shouldn't generally need to directly call that function.

[bwasty]

I went through the Vulkano code and I'm not entirely sure it's safe to omit it, but I removed it for now because the docs say that and none of the examples use it either.