In the capture-time tracking of resource usage, this change adds information on
the offset and size of accesses to Vulkan VkDeviceMemory. This information is
curretnly discarded, but will be used in a later change to implement
finer-grained usage tracking for VkDeviceMemory.
This change simplifies the `FrameRefType` enum by removing the separate states and
transitions, and replacing them with a smaller set of reference types, together
with functions that compose those reference types.
This change also introduces a `Clear` reference type, which is used to represent
an access to the resource that completely overwrites the previous contents; this
is in contrast with a `Write` access that may leave some of the previous
contents untouched.
* We need to locate it before setting the toolchain file, so checking when
building renderdoccmd is too late. This also lets us print where the SDK/NDK
is found, and add error messages if we don't locate it properly.
Intel published a performance query extension support defined back in
2013. This is available on Windows & Linux (in the Mesa driver). This
should provide the same types of counters as the MDAPI backend for
DX11.
Frameretrace [1] (a fork/branch of Apitrace) uses the same extension.
v2: Fix build without OpenGL
Simplify logic to enable counters
Warn about non enabled counters on Linux/Mesa
Generate counter Uuid
v3: Turn asserts into errors
Don't load perf entry points manually
v4: More clang-format
v5: Fix some Windows conversion warnings
v6: Fix errors on Windows where the driver reports an error on
glGetPerfQueryInfoINTEL as a mean to say that the queryId cannot
be used through the extension
v7: clang-format
v8: Initialize variable passed by pointers to GL entry points
v9: Only try to use the INTEL_performance_query on Mesa, experience
shows the Intel Windows driver doesn't report anything useful.
[1]: https://github.com/janesma/apitrace/wiki/screen-shots
* Nomachine is one program that requests these, but has a fallback path for if
they're not available. Denying the request makes things more compatible.
* When multiple buffers are aliased over the same heap memory, they'll have the
same GPU base address. Our mapping needs to ensure it removes the right entry
when a resource is released.
Benson Joeris