Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Ivi diff v8 #109

Merged
merged 16 commits into from
Apr 16, 2024
Merged

Ivi diff v8 #109

merged 16 commits into from
Apr 16, 2024

Conversation

zhonghuis
Copy link

No description provided.

@sysopenci
Copy link

Improper Commit Message
Tracked on not found in commit message,
make sure Tracked-On: Jira-ticket is present.

@sysopenci
Copy link

Android CI has started Engineering Build for this issue ,Please check Android CI Web for more details.

@sysopenci
Copy link

SUCCESS: Android CI has completed Engineering Build for this issue, Please check Android CI Web for more details.

@sysopenci
Copy link

Android CI has started Engineering Build for this issue ,Please check the linked Tracked-On issue/Android CI Web for more details.

@sysopenci
Copy link

Android CI has started Engineering Build for this issue ,Please check the linked Tracked-On issue/Android CI Web for more details.

@sysopenci
Copy link

FAILURE: Android CI has completed Engineering Build for this issue.Please check the linked Tracked-On issue/Android CI Web for more details.

Copy link

@feijiang1 feijiang1 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

@sysopenci
Copy link

FAILURE: Android CI has completed Engineering Build for this issue.Please check the linked Tracked-On issue/Android CI Web for more details.

Remove GEN3, which is no longer supported.
Add more device id for GEN9, GEN12, ADL-P from UPSTREAM.
Make Y-Tile bo creation available for framebuffer.

As ADL-P tiled framebuffer allocation code in kernel depends
on DPT. The DPT kernel code in smr0 branch is not the latest, which
will cause tiled framebuffer alloc fail in gvt-d mode. Therefore,
just disable tiled fb for ADL-P in vm mode. This is not needed in
newer kernel version.

Tracked-On: OAM-117251
Signed-off-by: chenyanxzhu <[email protected]>
Alloc no-tile buffer when sriov mode
If has dGPU, do not use Tiling Y.

Tracked-On:
Signed-off-by: HeYue <[email protected]>
Signed-off-by: chenyanxzhu <[email protected]>
For sw usage, allocate linear buffer.

Tracked-On:
Signed-off-by: HeYue <[email protected]>
Use i915 for both render and display buffer alloc.
Kvm doesn't support tiled fb in adl-p, turn off tiled fb
for that. Disable tiling-y for all render format for dGPU
doesn't support it.

Tracked-On:
Signed-off-by: chenyanxzhu <[email protected]>
The hypervisor signatures of kernel 5.10 and
5.15 are different.

Tracked-On:
Signed-off-by: HeYue <[email protected]>
when we use format HAL_PIXEL_FORMAT_IMPLEMENTATION_DEFINED to
allocate the buffer, we get the buffer with format
DRM_FORMAT_XBGR8888. The validateBufferSize will return an error
because format mismatch. So we should ignore this type.

Tracked-On:
Signed-off-by: zhangyichix <[email protected]>
The function validateBufferSize fails when sw codec allocates buffer
with HAL_PIXEL_FORMAT_YCBCR_P010. The reason is that the pixel format
will be converted to drm format for allocating buffer, then
the format of the handle is assigned to HAL_PIXEL_ format type with
i915_private_invert_format, but the descriptor.format is still
the pixel format, they must be different when checked in
validateBufferSize. So we check the format type after uniformly
converting the pixel/HAL format to drm format.

Tracked-On:
Signed-off-by: zhangyichix <[email protected]>
Instead of printing all logs with ERROR level, some logs just
need INFO level.

Tracked-On:
Signed-off-by: chenyanxzhu <[email protected]>
Changes include
- Using tile 4 instead of tile y.
- Querying memory regions using ioctl
- Horizontal and veritical alignment based on gen version

Tracked-On:
Change-Id: Id3f799ce26f6e154e75a81f22c92e15799759ca6
Signed-off-by: Huang Rui <[email protected]>
Signed-off-by: Mao, Marc <[email protected]>
Signed-off-by: Jeevaka Prabu Badrappan <[email protected]>
Signed-off-by: Kanlii Hu <[email protected]>
Signed-off-by: Fei Jiang <[email protected]>
Signed-off-by: He Yue <[email protected]>
Kernel supports flex and arc series cards with prelim uapi.
Switch to use prelim uapi in default and fallback to legacy uapi if it
works on upstream kernels.

Tracked-On:
Change-Id: Id3f799ce26f6e154e75a81f22c92e15799759ca6
Signed-off-by: Huang Rui <[email protected]>
Signed-off-by: Mao, Marc <[email protected]>
Signed-off-by: Jeevaka Prabu Badrappan <[email protected]>
Signed-off-by: Kanlii Hu <[email protected]>
Signed-off-by: Fei Jiang <[email protected]>
Signed-off-by: He Yue <[email protected]>
Tests done: Tested with antutu benchmark, 'video editing' score
has increased.

Tracked-On:
Signed-off-by: ZhuChenyanX <[email protected]>
In VM if run vulkan on iGPU, i915_dev->has_fence_reg
is 0, so DRM_IOCTL_I915_GEM_SET_TILING and
DRM_IOCTL_I915_GEM_GET_TILING will never be invoked.
Meanwhile if on dg2, i915_dev->has_fence_reg also is 0.
So the flag 'i915_dev->has_fence_reg' can't be using to
check if invoke GET_TILING,SET_TILING.
Consider if run vulkan app not on dg2, should invoke
GET_TILING,SET_TILING.

Tracked-On:
Signed-off-by: Li, HaihongX <[email protected]>
Neusoft Demo app performance is low
It is caused by the Choreographer->doFrame->travesal->draw function.
This function will do the copyBlit operation.
The WB flag will use the cache to gain performance.
So we should not use the WC flag for the
case of BO_USE_SW_READ_OFTEN,due to the frequent CPU data reads.

Tracked-On:
Signed-off-by: Li,Gaoshun [email protected]
Cause kernel doesn't support this op for dgpu.

Tracked-On:
Signed-off-by: chenyanxzhu <[email protected]>
This is to help fix Coverity issues in minigbm project.

CIDs:222122 188451 203271 255440

Tracked-On:
Signed-off-by: manxiaoliang <[email protected]>
Tracked-On: OAM-117251
Signed-off-by: HeYue <[email protected]>
@sysopenci
Copy link

Android CI has started Engineering Build for this issue ,Please check the linked Tracked-On issue/Android CI Web for more details.

1 similar comment
@sysopenci
Copy link

Android CI has started Engineering Build for this issue ,Please check the linked Tracked-On issue/Android CI Web for more details.

@sysopenci
Copy link

FAILURE: Android CI has completed Engineering Build for this issue.Please check the linked Tracked-On issue/Android CI Web for more details.

1 similar comment
@sysopenci
Copy link

FAILURE: Android CI has completed Engineering Build for this issue.Please check the linked Tracked-On issue/Android CI Web for more details.

@sysopenci
Copy link

one of the dependent pr cannot be Rebase and Merged; please fix the issue and commit your changes once again to run EB.

@sysopenci
Copy link

Android CI has started Engineering Build for this issue ,Please check the linked Tracked-On issue/Android CI Web for more details.

@sysopenci
Copy link

SUCCESS: Android CI has completed Engineering Build for this issue.Please check the linked Tracked-On issue/Android CI Web for more details.

@sysopenci
Copy link

Android CI has started MERGE Build for this pr ,Please check the linked Tracked-On issue/Android CI Web for more details.

@sysopenci sysopenci merged commit b5053aa into projectceladon:upstream_main Apr 16, 2024
15 checks passed
@sysopenci
Copy link

Android CI has completed MERGE Build for this pr, build is SUCCESS. Please check the linked Tracked-On issue/Android CI Web for more details. For Binaries: /cactus-absp-or-local/celadon-merge/404

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants