Add latest maintenance extensions required by Android Vulkan
requirements, except VK_KHR_maintenance5 which enables dynamic rendering
on ANGLE and causes issues with some cuttlefish targets.
Test: CI
Reviewed-by: Aaron Ruby <aruby@qnx.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/41210>
This is a follow-up to:
- commit e27e41a8 ("vulkan,spirv: update headers")
- commit c4cecd9d ("gfxstream: cereal: fix 'None' in gfxstream codegen")
The second commit fixed compile, but still produced a
crash on the host side:
$0 __memcpy_evex_unaligned_erms () at ../sysdeps/x86_64/multiarch/memmove-vec-unaligned-erms.S:833
$1 0x00007f116f8ee285 in memcpy (__dest=0x7f0d258cd340, __dest@entry=0x7f0ed55e9458, __src=0x7f0e78072ff7, __src@entry=0x72795f61, __len=1920556897)
at prebuilts/gcc/linux-x86/host/x86_64-linux-glibc2.17-4.8/sysroot/usr/include/x86_64-linux-gnu/bits/string3.h:51
$2 gfxstream::host::vk::VulkanStream::loadStringInPlaceWithStreamPtr (this=this@entry=0x7f0e78022b50, forOutput=forOutput@entry=0x7f0d97fff018,
streamPtr=streamPtr@entry=0x7f0ed55e9458) at hardware/google/gfxstream/host/vulkan/vulkan_stream.cpp:100
$3 0x00007f116f8ee3de in gfxstream::host::vk::VulkanStream::loadStringArrayInPlaceWithStreamPtr (this=0x7f0e78022b50, forOutput=<optimized out>,
streamPtr=0x7f0ed55e9458) at hardware/google/gfxstream/host/vulkan/vulkan_stream.cpp:122
$4 0x00007f116f9f2fcb in gfxstream::host::vk::reservedunmarshal_VkDeviceCreateInfo (vkStream=vkStream@entry=0x7f0e78022b50,
Fix up lenAccess string marshaling to prevent this.
TEST=launch_cvd --gpu_mode=gfxstream_guest_angle_host_swiftshader
Reviewed-by: David Gilhooley <djgilhooley@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/41059>
Goldfish needs a non-stub QEMU pipe stream and access
to the goldfish_sync device.
All of this logic may be deleted once qemu over
virtio-gpu is fully enabled.
TEST=run meson2hermetic + m vulkan.ranchu
Reviewed-by: Bo Hu <bohu@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40851>
... to ensure that "mSyncHelper->close()" still happens on
non goldfish devices.
Android equivalent ag/39315505 for b/500332164
Tested with CtsGraphicsTestCases after the build file changes.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40836>
Commit e27e41a842 ("vulkan,spirv: update headers") exposed a
flaw in the cerealgenerator.
It modified -- among other things -- the VkDeviceCreateInfo
struct in vk.xml.
In the update, the len="enabledLayerCount,null-terminated"
attribute was removed from the ppEnabledLayerNames member.
The gfxstream code generator processes ppEnabledLayerNames
(which is a const char* const*), it identifies it as an "array of
strings". However, because the len attribute is now missing,
vulkanType.getLengthExpression() returns None.
This leads to errors like:
gfxstream_guest_vk_autogen_impl/gen/goldfish_vk_counting_guest.cpp:642:30:
error: use of undeclared identifier 'None'
642 | for (uint32_t i = 0; i < None; ++i)
| ^~~~
1 error generated.
This patch adds various length access checks to prevent this from
happening.
TEST=m vulkan.ranchu
Reviewed-by: David Gilhooley <djgilhooley@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40785>
... as this can lead to a deadlock with the following sequence:
Time1: guest-thread-1: vkDestroyImageView() called
Time2: VkEncoder grabs seqno 1
Time3: guest-thread-2: vkQueueSubmit() called
Time4: ResourceTracker::on_vkQueueSubmitTemplate() locks
mLock for using `info_VkFence`
Time5: ResourceTracker::on_vkQueueSubmitTemplate() calls
enc->vkQueueWaitIdle()
Time6: VkEncoder grabs seqno 2
Time7: VkEncoder sends the vkQueueWaitIdle with seqno
2 via ASG to host
Time8: VkEncoder waits for the `VkResult` from the
host via `stream->read()`
Time9: guest-thread-1: VkEncoder calls sResourceTracker->destroyMapping()
->mapHandles_VkImageView((VkBuffer*)&buffer);
which calls
ResourceTracker::unregister_VkImageView()
ResourceTracker::unregister_VkImageView() tries to
lock mLock to erase the info struct
!!! DEADLOCKED HERE !!!
guest-thread-1 is stuck waiting on mLock (currently locked by
guest-thread-2) before it would `stream->flush();` to finishing
sending the vkDestroyImageView() command to the host and potentially
ping its corresponding host-render-thread-1.
guest-thread-2 is stuck waiting on the result from host-render-thread-2
but host-render-thread-2 won't progress until host-render-thread-1
finishes seqno 1 which needs guest-thread-1 to finish sending/pinging.
Android equivalent change ag/39258728 for b/498964194
Test: cvd create --gpu_mode=gfxstream_guest_angle_host_swiftshader
open maps
pan/zoom/etc for a couple minutes
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40767>
Mesa's vk_common_SetDebugUtilsObjectNameEXT assumes every Vulkan object
handle is a pointer to a vk_object_base struct. In gfxstream, only a
subset of objects (instance, device, queue, command buffer, command pool,
buffer, fence, semaphore) carry a Mesa wrapper. All other non-dispatchable
handles (shader modules, pipelines, render passes, etc.) are opaque host
handles that are not valid pointers.
Passing such an unwrapped handle to the common path causes it to be cast
to a vk_object_base pointer and dereferenced, resulting in a SIGSEGV
(null-pointer dereference at offset 0x40).
Override the function in the gfxstream driver to store debug names on
vk_object_base for wrapped objects and return VK_SUCCESS for unwrapped
objects.
Fixes: 7b50e62179 ("gfxstream: mega-change to support guest Linux WSI with gfxstream")
Test: Verified with hellovk (with validation layers) on Android Emulator - no crashes.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40712>
The prior version of gfxstream-fuchsia featured an Android compat
layer, and only worked with Goldfish. The compat layer was
deleted a while ago.
The Fuchsia Goldfish integration was deleted here:
https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33050
since the Fuchsia emulator itself uses a frozen version of the
library. This goes further in that direction: deleting leftovers
of the Fuchsia Android compat layer, which used provided VK
dispatch.
This allows Fuchsia + gfxstream to build via meson2hermetic, with
the a few additional common core patches applied.
Reviewed-by: David Gilhooley <djgilhooley@gmail.com>
Reviewed-by: Aaron Ruby <aruby@qnx.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40442>
It is an output. Useful for meson-to-bazel conversion, where
Bazel happened to spot this.
Reviewed-by: David Gilhooley <djgilhooley@gmail.com>
Reviewed-by: Aaron Ruby <aruby@qnx.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40442>
Both inc_guest_iostream and inc_platform_virtgpu were used
twice in the same include_directories array.
Reviewed-by: David Gilhooley <djgilhooley@gmail.com>
Reviewed-by: Aaron Ruby <aruby@qnx.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40442>
Sometimes, it's used with fds (int32_t), sometimes with
gem_handle (uint32_t). Be consistent and explicitly cast
when needed.
Reviewed-by: David Gilhooley <djgilhooley@gmail.com>
Reviewed-by: Aaron Ruby <aruby@qnx.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40442>
This is a backport of f134cc5a1e:
("Update <type category="funcpointer"> schema to simplify")
in vulkan-docs, essentially. It changed things about how vk.xml
is parsed.
Fixes: b30f780c ("vulkan: update spec to 1.4.340")
Reviewed-by: Aaron Ruby <aruby@qnx.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39502>
- the vkSetDebugMetadataAsyncGOOGLE command should
not have an entry in the function table: it
leads to missing prototype errors
- Make gfxstream respect cpp_msvc_compat_args, since
it is a C++ project. -Wmissing-prototypes will be
made a cpp error *eventually*.
Reviewed-by: Aaron Ruby <aruby@qnx.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39418>
C++23 removed certain transitive includes so you need to add C++ headers
directly to utilize their functions and classes. We add cstdlib header
for abort.
Errors:
external/mesa3d/src/gfxstream/guest/vulkan_enc/goldfish_vk_counting_guest.cpp:9505:13:
error: use of undeclared identifier 'abort'
9505 | abort();
Test: m vulkan.ranchu
Reviewed-by: Aaron Ruby <aruby@qnx.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39360>
Previously, there were no meson.build rules for the Android
portion of gfxstream_vk, since nobody builds gfxstream_vk + Android
via the Android NDK.
Using Soong (and not the Android NDK) is an absolute requirement,
since Android Virtual Devices (AVDs) were amongst to fully
transition to Soong over Android makefiles, leading to
breath-taking and exhilarating build speeds.
The meson build rules that do exist works are designed for
gfxstream_vk + Linux. That leads to higher maintainence
costs: maintaining hand-written Android.bp files for Android,
and meson.build for Linux.
Enter meson2hermetic. With this tool, meson.build becomes the
source of truth, and Android.bp are generating automatically,
reducing maintainence costs.
But for this to work, the portion of gfxstream_vk that didn't
have meson build rules needs them. This patch does this
and there are two things to note:
1) gfxstream_vk + Android needs dependencies that don't have
any pkg-config files, and exist only with the AOSP tree.
These include things like:
- libqemupipe.ranchu
- libOpenglCodecCommon
- libgralloc_cb.ranchu
- renderControlEncoder
Most of these dependencies support the Goldfish AVD, and will
be deleted over time as the that emulator transitions to
virtio-gpu. There are more generic Android deps too, like
"android_base".
Generic Android dependencies are given the prefix "android-",
while Goldfish (a.k.a Android Emulator) is given the prefix
"android-aemu-".
The ability to use dependencies that don't have a direct
Linux-distro style analogue is the one of motivators
meson2hermetic.
2) There's a special "gfxstream_emulated_android" case. This
is for the GfxstreamEnd2EndTests target, which uses the code
defined here:
- src/gfxstream/guest/android/ANativeWindowEmulated.cpp
- src/gfxstream/guest/android/GrallocEmulated.cpp
This is used by Gfxstream Github for CI/CD.
Acked-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39279>
Otherwise, the following crash is observed on the host:
"Unhandled Vulkan structure type Unhandled VkStructureType [1000010002], aborting"
which corresponds to PHYSICAL_DEVICE_PRESENTATION_PROPERTIES_ANDROID.
We shouldn't be sending those structs down to the host. Don't
post-process vkGetPhysicalDeviceProperties2, pre-process it to
filter the guest-only structs.
Reviewed-by: David Gilhooley <djgilhooley.gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39205>
This helps Meson track when dependencies are modified. If they
are modified, running ninja -C actually re-generates the code.
Beforehand, this was not the case and contrary to the user
expectation.
Reviewed-by: David Gilhooley <djgilhooley.gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39205>
... and send data on connection creation.
Test: cvd create --gpu_mode=gfxstream_guest_angle
and go/cuttlefish-tracing to grab perfetto trace
Reviewed-by: Aaron Ruby <aruby@qnx.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39161>
Full command lines include full path to the output file, which triggers
reproducibility warnings (e.g. in Yocto builds). Drop the args and print
only a basename of the script used to generate the file.
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/38875>
Fixes things like:
GfxStreamVulkanMapper.cpp:45:10:
error: no previous prototype for function 'chooseGfxQueueFamily'[-Werror,-Wmissing-prototypes]
Reviewed-by: David Gilhooley <djgilhooley.gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/38751>
vulkan_gfxstream.h contains custom protocols not found
in vk.xml (vk_gfxstream.xml).
gfxstream_vk_entrypoints.h is codegen by Mesa common code,
and it does not accept the custom XML.
So avoid generating implementations for them:
guest/vulkan_enc/gfxstream_guest_vk_autogen_impl/gen/func_table.cpp:5321:1:
note: declare 'static' if the function is not intended to be
used outside of this translation unit
5321 | void gfxstream_vk_CollectDescriptorPoolIdsGOOGLE(
| ^
| static
fatal error: too many errors emitted, stopping now [-ferror-limit=]
Reviewed-by: David Gilhooley <djgilhooley.gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/38632>
Fixes errors like:
ResourceTracker.cpp:62:6: error: no previous prototype for function 'zx_handle_close'
[-Werror,-Wmissing-prototypes]
Reviewed-by: David Gilhooley <djgilhooley.gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/38632>
