... except when running with software rendering as apps do not
know that they would need to request additional CPU_* usage
for GPU_* usage.
Bug: b/356845188
Test: cts -m CtsNativeHardwareTestCases
Change-Id: I3536d80469d2187550558e9d02795896de4f9827
... to ensure the lifetime of the underlying cros_gralloc_driver
matches those of the various Mapper* interfaces. Without this, the
the destruction of the static cros_gralloc_driver `s_instance`
could race with a potential CrosGralloc4Mapper user on another
thread which is still alive at the time of exit() (for example,
a buffer being cleaned up on a libhwbinder worker thread).
Bug: b/300616292
Bug: b/300194143
Bug: b/294661992
Test: launch Cuttlefish, run SurfaceFlinger_test, inspect
tracing around lifetime of cros_gralloc_driver
Change-Id: Id8e2dbff471c03f7b663952524905a06cc03fbc1
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/minigbm/+/4897926
Reviewed-by: Yiwei Zhang <zzyiwei@chromium.org>
Tested-by: Jason Macnak <natsu@google.com>
Commit-Queue: Jason Macnak <natsu@google.com>
Move mt8183_camera_quirk_ into struct cross_domain_private.
Add a helper `drv_get_os_option` in `drv_helper` to query os option.
Advertise DRM_FORMAT_MTISP_SXYZW10 as supported format in virtgpu_cross_domain.c.
Implement cross domain's own resolve_format_and_use_flags entry point.
BUG=b:269982880
TEST=tested on a kukui machine after applying the change. Camera and
screen recording works fine.
Change-Id: Ia0d6e2cc701477aa6b9542de84e24f8aeb5a29e0
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/minigbm/+/4304961
Commit-Queue: Dawn Han <dawnhan@google.com>
Reviewed-by: Dominik Behr <dbehr@chromium.org>
Tested-by: Dawn Han <dawnhan@google.com>
Add support for the MT8183 private camera format to gralloc. This is
done as a vendor specific workaround that bypasses the normal minigbm
format resolution, targeting kukui+ARCVM.
This is aimed at working around the fact that the virtgpu cross domain
backend needs to directly resolve formats itself. Proper support for
involving the host when resolving formats would be cleaner. However,
that would be a substantial amount of work for what is (hopefully) a
one-off situation.
BUG=b:158617858
TEST=testBasicOpaqueToJpegReprocessing on kukui-arc-r DUT
Change-Id: I8e1d74aac0a16babd7d7e2e630ff8affb3104e8b
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/minigbm/+/4258986
Tested-by: David Stevens <stevensd@chromium.org>
Reviewed-by: Yiwei Zhang <zzyiwei@chromium.org>
Commit-Queue: David Stevens <stevensd@chromium.org>
... so that devices that have API level 31 but do not have the
system dmabuf heap installed (as it is not required) do not spam
error logs before falling back to memfd.
BUG=b:227375257
TEST=build Cuttlefish with aosp/2051732
TEST=cvd start
TEST=vts -m VtsHalGraphicsMapperV4_0Target
Change-Id: I7ab79ce472b1e3d84ffa8bb06044ecc933847656
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/minigbm/+/3563512
Commit-Queue: Jason Macnak <natsu@google.com>
Reviewed-by: Yiwei Zhang <zzyiwei@chromium.org>
Commit-Queue: Yiwei Zhang <zzyiwei@chromium.org>
Tested-by: Yiwei Zhang <zzyiwei@chromium.org>
"Prefer to use a struct instead of a pair or a tuple
whenever the elements can have meaningful names."
BUG=b:207388558
TEST=cvd start
TEST=vts -m VtsHalGraphicsMapperV4_0Target
Change-Id: I688267a4c40f0260476029c0d70bc8bb19943e21
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/minigbm/+/3481955
Reviewed-by: Yiwei Zhang <zzyiwei@chromium.org>
Tested-by: Jason Macnak <natsu@google.com>
Commit-Queue: Jason Macnak <natsu@google.com>
... as interacting with a cloned handle should not
succeed until it is imported:
```
native_handle_t handle1;
gralloc->alloc(..., &handle1);
native_handle_t handle2 = native_handle_clone(handle1);
gralloc->lock(handle2) <--- should fail!
```
This is a partial revert of crrev.com/c/3433118
BUG=b:207388558
BUG=b:220537482
TEST=cvd start
TEST=vts -m VtsHalGraphicsMapperV4_0Target
Change-Id: Ia9a9f91fca14ea9f30017fb934c5c48064090d43
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/minigbm/+/3481795
Reviewed-by: Yiwei Zhang <zzyiwei@chromium.org>
Tested-by: Jason Macnak <natsu@google.com>
Commit-Queue: Jason Macnak <natsu@google.com>
... for reserved region / metadata as dmabuf heaps are the hot new
thing replacing ion.
Also, Android already has sepolicy in place for passing around
dmabufs (see `dmabuf_system_heap_device:chr_file` references in
system/sepolicy) while memfds require manual additions (see
aosp/1960673).
BUG=b:207388558
TEST=cvd start
TEST=vts -m VtsHalGraphicsMapperV4_0Target
Change-Id: I9acb30a10afb1f9f273ceb17dd44a9ac96da1a97
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/minigbm/+/3449613
Reviewed-by: Roman Stratiienko <r.stratiienko@gmail.com>
Reviewed-by: Yiwei Zhang <zzyiwei@chromium.org>
Tested-by: Jason Macnak <natsu@google.com>
Commit-Queue: Jason Macnak <natsu@google.com>
... to support upcoming change where CrosGralloc4Mapper's
get(<metadata>) will want to access metadata that may be
in either the buffer's handle or the buffer's reserved
region. The for-each-[handle|buffer] methods support
dumpBuffers() where we need to iterate over all of the
buffers while holding a lock on the cros_gralloc_driver's
mutex.
BUG=b:207388558
TEST=cvd start
TEST=vts -m VtsHalGraphicsMapperV4_0Target
TEST=cvd start --gpu_mode=gfxstream
TEST=vts -m VtsHalGraphicsMapperV4_0Target
Change-Id: I674cab487d7446af722f94d796f504cf4f97de27
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/minigbm/+/3433121
Reviewed-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Roman Stratiienko <r.stratiienko@gmail.com>
Reviewed-by: Dominik Behr <dbehr@chromium.org>
Tested-by: Jason Macnak <natsu@google.com>
Commit-Queue: Jason Macnak <natsu@google.com>
... to make it easier handle ownership (no more conditional
ownership if the buffer was allocated versus imported).
This also helps avoid a deadlock in CrosGralloc4Mapper's
dumpBuffers() in an upcoming change where CrosGralloc4Mapper
wants to interact with both metadata in the handle and
metadata in the cros_gralloc_buffer's reserved region. By
having the cros_gralloc_buffer always have a handle, the
CrosGralloc4Mapper can have access to all metadata when it
has a cros_gralloc_buffer parameter.
BUG=b:207388558
TEST=cvd start
TEST=vts -m VtsHalGraphicsMapperV4_0Target
TEST=cvd start --gpu_mode=gfxstream
TEST=vts -m VtsHalGraphicsMapperV4_0Target
Change-Id: I56fffbbfacf8436db165a961bec8ce546b4acc6c
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/minigbm/+/3433118
Reviewed-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Roman Stratiienko <r.stratiienko@gmail.com>
Reviewed-by: Dominik Behr <dbehr@chromium.org>
Tested-by: Jason Macnak <natsu@google.com>
Commit-Queue: Jason Macnak <natsu@google.com>
It makes more sense to resovle the format and use_flags together to
avoid duplicates.
Add the helper to vc4 and dumb_driver backends since that's missed
from prior frontend refactoring.
Split the resolve helper in virtgpu_virgl for the 3d and 2d paths to
make code logic cleaner.
BUG=b:199524294
TEST=CQ
Change-Id: I89b633ea484f0fc5bb9b4e0548cad017b9970cc6
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/minigbm/+/3200143
Tested-by: Yiwei Zhang <zzyiwei@chromium.org>
Tested-by: Jason Macnak <natsu@google.com>
Reviewed-by: Chia-I Wu <olv@google.com>
Reviewed-by: Jason Macnak <natsu@google.com>
Reviewed-by: Dominik Behr <dbehr@chromium.org>
Commit-Queue: Yiwei Zhang <zzyiwei@chromium.org>
This refactoring is to prepare for later unifying use_flags resolution
for all gralloc frontends, so that all fallbacks and special handlings
can be placed in the centralized private helper.
BUG=b:199524294
TEST=CtsNativeHardwareTestCases
Change-Id: I6e37822b7dbaa123c6956ff31bcef6a35a171f74
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/minigbm/+/3163215
Tested-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Jason Macnak <natsu@google.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>
Commit-Queue: Yiwei Zhang <zzyiwei@chromium.org>
Only a single cros_gralloc_driver instance is allowed to exist in a
process. However, currently when Gralloc3.0 or Gralloc4.0 is used, it's
possible for 2 independent driver instances to be created. This change
refactors cros_gralloc_driver to be a singleton and ensures the driver
initialization only happens once per process.
BUG=b:191895066
TEST=No regression of gralloc0 in ARCVM
Change-Id: I3cb8b7c3d13717cfe531b557b9d69b72efdbf3da
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/minigbm/+/3002266
Tested-by: Yiwei Zhang <zzyiwei@chromium.org>
Tested-by: Jason Macnak <natsu@google.com>
Tested-by: John Stultz <john.stultz@linaro.org>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>
Reviewed-by: Jason Macnak <natsu@google.com>
Commit-Queue: Yiwei Zhang <zzyiwei@chromium.org>
Vulkan on Android relis on VkImageDrmFormatModifierExplicitCreateInfoEXT
to import the gralloc allocated buffer, which requires buffer info of
the host storage.
BUG=b:186679287
TEST=venus wsi on arcvm and egl/gles
TEST=CtsNativeHardwareTestCases and related CtsGraphicsTestCases
Change-Id: I7ab0e80ab67a2daae9ea547a8d5abe9417af17d8
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/minigbm/+/2862161
Tested-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>
Commit-Queue: Yiwei Zhang <zzyiwei@chromium.org>
Implements the allocator 3.0 and mapper 3.0 interfaces
which:
- Implements HIDL interface directly (older versions
are wrapped in a passthrough HIDL interface)
- Adds isSupported() to allow checking for format and
usage combination support before allocating.
Adds emulated multi-planar buffer support to virtio
backend for non gbm enabled hosts.
Updates cros_gralloc_handle to use uint64_t instead of
two uint32_t for some members.
Updates cros_gralloc_handle to have a single format
modifier.
Replaces Android makefiles with Android bp files.
BUG=b:146515640
TEST=run Cuttlefish w/ gralloc3 and run CTS tests
Change-Id: I43ed9788a2413201bddce17ffb69b76006ef39fd
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/minigbm/+/2273554
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>
Tested-by: Jason Macnak <natsu@google.com>
Commit-Queue: Jason Macnak <natsu@google.com>
In ARC++, the wayland sevice and the video stack rely on
GRALLOC_DRM_GET_STRIDE and (*lock_ycbcr) with zero flags to return
the metadata associated with the buffer.
In the past, we've simply returned the metadata that was calculated
during allocation.
Since the current virtio-gpu API relies on shadow buffers, there's
actually two different sets of metadata:
1) The metadata of the shadow buffer --> useful for mapping
2) The metadata of the host resource --> useful for passing to Chrome
For the wayland_service and video stack, we want to return (2).
For the Android framework, we want to return (1).
BUG=b:132939420
TEST=compile
Change-Id: I1134d651396ba68e064eaf2e3cad3cb3225d7c5c
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/minigbm/+/1681383
Reviewed-by: David Stevens <stevensd@chromium.org>
Commit-Queue: Gurchetan Singh <gurchetansingh@chromium.org>
Tested-by: Gurchetan Singh <gurchetansingh@chromium.org>
This will allow drivers to tile or detile only the regions requested
by the user. Note that the gralloc spec states that:
"This address will represent the top-left corner of the entire buffer,
even if accessRegion does not begin at the top-left corner."
(see hardware/interfaces/graphics/mapper/2.0/IMapper.hal in AOSP)
Also, the gralloc API makes it difficult to maintain two mappings of
the same buffer. For example, say you have two access regions:
module->lock(mod, handle1, 0, 0, 5, 5, &addr);
module->lock(mod, handle1, 5, 5, 10, 10, &addr);
module->unlock(mod, handle1); // which access region should be unlocked?
In practice, this scenario never happens on Android.
It's not exactly clear what gbm should return. Let's just return the
top left of the access region because that's what we where doing before.
BUG=chromium:764871
TEST=gbmtest, mmap_test -g, the following CTS tests:
android.view.cts.SurfaceViewSyncTests
android.media.cts.EncodeDecodeTest
android.video.cts.VideoEncoderDecoderTest
Change-Id: I7ca0713871e03928b1d4402aa161588990c7e775
Reviewed-on: https://chromium-review.googlesource.com/758147
Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org>
Tested-by: Gurchetan Singh <gurchetansingh@chromium.org>
Reviewed-by: Stéphane Marchesin <marcheu@chromium.org>
It's helpful to differentiate map flags from normal buffer creation
flags. Note gralloc doesn't differentiate between map flags and buffer
creation flags. However, since flags are passed in with gralloc
(*lock)(), we can use a separate conversion function there.
BUG=chromium:764871
TEST=Boot Android and play games on Eve
Change-Id: Ic8aee84d9ac945abf93d9a9bda78fe3f77711cc3
Reviewed-on: https://chromium-review.googlesource.com/691424
Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org>
Tested-by: Gurchetan Singh <gurchetansingh@chromium.org>
Reviewed-by: Stéphane Marchesin <marcheu@chromium.org>
Let's support GRALLOC_MODULE_API_VERSION_0_3. This will more closely
mirror what gralloc1 and HIDL gralloc does.
We have the option to asychronously unlock with this version of the API,
but decided the added complexity wouldn't lead to any performance benefits.
We'll just set the release fence pointer to be -1, indicating the
the buffer is ready to use after the (*unlock_Async) call.
BUG=b:62069164
TEST=Android boots, play Youtube app, ./gralloctest all passes
Change-Id: Ia21a11b541796c4b36003c50cd8627d189b4bc56
Reviewed-on: https://chromium-review.googlesource.com/422661
Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org>
Tested-by: Gurchetan Singh <gurchetansingh@chromium.org>
Reviewed-by: Tomasz Figa <tfiga@chromium.org>
We want the ability to support gralloc0, gralloc1 and HIDL
gralloc in this repo. This commit moves gralloc0 specific code
to the gralloc0 subdirectory, and refactors the rest of the code
so it can be eventually used by whichever APIs we choose to
implement.
In addition, all of the data by the module is now handled by
managed pointers instead of process data, since we get a number
of crash reports when we try to access data in gralloc_close().
BUG=b:62069164, b:62221166
TEST=gralloctest all, Android boots, Youtube videos, 5-10
CTS tests
Change-Id: Ic29a777573936216d99498d0b814ea50015fd435
Reviewed-on: https://chromium-review.googlesource.com/521794
Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org>
Tested-by: Gurchetan Singh <gurchetansingh@chromium.org>
Reviewed-by: Stéphane Marchesin <marcheu@chromium.org>
