| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
| |
The whole point of fence_server_sync is that it can be used to
avoid waiting in the application thread.
Reviewed-by: Andres Rodriguez <[email protected]>
Reviewed-by: Marek Olšák <[email protected]>
|
|
|
|
|
|
|
|
|
| |
We need to account for SGPR locations in merged shaders.
This case is exercised by KHR-GL45.enhanced_layouts.vertex_attrib_locations
Fixes: 79c2e7388c7f ("radeonsi/gfx9: use SPI_SHADER_USER_DATA_COMMON")
Reviewed-by: Marek Olšák <[email protected]>
|
|
|
|
|
|
|
|
| |
The original spec I had didn't expose integer textures and suggested that
you use unfiltered floats. Now there are proper formats for them.
Fixes 16- and 32-bit texwrap integer tests in piglit, and
dEQP-GLES3.functional.fbo.completeness.renderable.renderbuffer.color0.rgb10_a2ui.
|
|
|
|
|
|
|
|
| |
When we tried to clear color while storing depth, it assertion failed
about basically not having enough information to decide which color RT to
clear. It turns out the STORE_GENERAL picks the buffer according to the
color buffer being stored, or all of them if NONE. If you're doing depth,
it doesn't know which to pick.
|
|
|
|
| |
Signed-off-by: Rob Clark <[email protected]>
|
|
|
|
|
|
|
|
| |
The OVERWRITE bit disables destination fetches, which is exactly what
we want when there is no valid color buffer bound.
Signed-off-by: Lucas Stach <[email protected]>
Reviewed-by: Wladimir J. van der Laan <[email protected]>
|
|
|
|
|
| |
This centralizes the calculation in the surface, instead of in each
load/store.
|
|
|
|
| |
Fixes dEQP-GLES3.functional.depth_stencil_clear.depth.*
|
|
|
|
| |
Fixes piglit masked-clear.
|
|
|
|
| |
Fixes part of piglit masked-clear.
|
| |
|
| |
|
|
|
|
|
|
| |
To silence compiler warnings.
Reviewed-by: Marek Olšák <[email protected]>
|
|
|
|
| |
Roland Scheidegger <[email protected]>
|
|
|
|
|
|
|
| |
To avoid problems with MSVC. And verify size with ASSERT_BITFIELD_SIZE().
Reviewed-by: Marek Olšák <[email protected]>
Reviewed-by: Charmaine Lee <[email protected]>
|
|
|
|
|
|
|
| |
Enable h.264 encode for vcn hardware (raven)
Signed-off-by: Boyuan Zhang <[email protected]>
Acked-by: Christian König <[email protected]>
|
|
|
|
|
|
|
| |
Enable vcn encode by creating radeon_encoder for vcn.
Signed-off-by: Boyuan Zhang <[email protected]>
Acked-by: Christian König <[email protected]>
|
|
|
|
|
|
|
| |
Add implementation for create_encoder interface for vcn encode.
Signed-off-by: Boyuan Zhang <[email protected]>
Acked-by: Christian König <[email protected]>
|
|
|
|
|
|
|
| |
Add implementation for get_feedback interface for vcn encode.
Signed-off-by: Boyuan Zhang <[email protected]>
Acked-by: Christian König <[email protected]>
|
|
|
|
|
|
|
| |
Add implementation for destroy interface for vcn encode.
Signed-off-by: Boyuan Zhang <[email protected]>
Acked-by: Christian König <[email protected]>
|
|
|
|
|
|
|
| |
Add implementation for end_frame interface for vcn encode.
Signed-off-by: Boyuan Zhang <[email protected]>
Acked-by: Christian König <[email protected]>
|
|
|
|
|
|
|
| |
Add implementation for encode_bitstream interface for vcn encode.
Signed-off-by: Boyuan Zhang <[email protected]>
Acked-by: Christian König <[email protected]>
|
|
|
|
|
|
|
| |
Add implementation for begin_frame interface for vcn encode.
Signed-off-by: Boyuan Zhang <[email protected]>
Acked-by: Christian König <[email protected]>
|
|
|
|
|
|
|
|
| |
Implement encoding of sps, pps, and silce headers using the newly added h.264
header coding descriptors functions based on h.264 specs.
Signed-off-by: Boyuan Zhang <[email protected]>
Acked-by: Christian König <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Since bitstream headers, e.g. sps, pps, slice, are encoded in driver side, we
need to add corresponding algorithms that required to generate those headers.
According to h.264 specs, signed/unsigned interger Exp-Golomb-coded syntax
element with left bit first (code_se and code_ue) and unsigned integer using
n bits (code_fixed_bits) descriptors function are needed. Therefore, adding
those algorithms and related variables and output algorithms here.
Signed-off-by: Boyuan Zhang <[email protected]>
Acked-by: Christian König <[email protected]>
|
|
|
|
|
|
|
| |
Implement required ibs and command buffer submission interfaces for vcn encode
Signed-off-by: Boyuan Zhang <[email protected]>
Acked-by: Christian König <[email protected]>
|
|
|
|
|
|
|
|
|
| |
Add a skeleton pipe video interface and encode ib interface for video encode
on vcn hardware. Add function defines and structures for vcn encode. Update
Makefile.sources and meson.build with newly added files.
Signed-off-by: Boyuan Zhang <[email protected]>
Acked-by: Christian König <[email protected]>
|
|
|
|
|
|
|
|
|
| |
pic_order_cnt_type is a required variable when encoding both sps and
slice header, therefore we need to get this value from st, e.g. vaapi
interface, and then pass it to radeon driver for encoding headers.
Signed-off-by: Boyuan Zhang <[email protected]>
Acked-by: Christian König <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
| |
Different from vce encoding, vcn encoding requires driver side to encode
bitstream header, such as pps, sps and slice header. pic_order_cnt_type
is a required variable when encoding both sps and slice header, therefore
we need to add this new variable here, and hold the value passed from st,
e.g. vaapi interface
Signed-off-by: Boyuan Zhang <[email protected]>
Acked-by: Christian König <[email protected]>
|
|
|
|
|
|
|
| |
New cs support is needed for vcn encode
Signed-off-by: Boyuan Zhang <[email protected]>
Acked-by: Christian König <[email protected]>
|
|
|
|
|
|
|
| |
New ring type is needed for vcn encode
Signed-off-by: Boyuan Zhang <[email protected]>
Acked-by: Christian König <[email protected]>
|
|
|
|
|
|
|
| |
Add a new header file for vcn encode interface
Signed-off-by: Boyuan Zhang <[email protected]>
Acked-by: Christian König <[email protected]>
|
|
|
|
|
|
|
|
| |
* Explicitely convert values to int in comparison.
* Remove one MAYBE_UNUSED that is actually not needed.
Signed-off-by: Gert Wollny <[email protected]>
Reviewed-by: Brian Paul <[email protected]>
|
|
|
|
|
|
|
| |
Decorate the according parameters with UNUSED.
Signed-off-by: Gert Wollny <[email protected]>
Reviewed-by: Brian Paul <[email protected]>
|
|
|
|
|
|
|
| |
Decorate the unused parameters with UNUSED.
Signed-off-by: Gert Wollny <[email protected]>
Reviewed-by: Brian Paul <[email protected]>
|
|
|
|
|
|
|
| |
Silence warnings by decoration the parameters with UNUSED.
Signed-off-by: Gert Wollny <[email protected]>
Reviewed-by: Brian Paul <[email protected]>
|
|
|
|
|
|
|
| |
function headers
Signed-off-by: Gert Wollny <[email protected]>
Reviewed-by: Brian Paul <[email protected]>
|
|
|
|
|
|
|
|
| |
Use size_t instread of unsigned for new_max. realloc later expects
size_t as parameter anyway.
Signed-off-by: Gert Wollny <[email protected]>
Reviewed-by: Brian Paul <[email protected]>
|
|
|
|
|
|
|
|
| |
With --disable-debug a parameter is not used. Silence this
warning by fake-using it.
Signed-off-by: Gert Wollny <[email protected]>
Reviewed-by: Brian Paul <[email protected]>
|
|
|
|
|
|
|
|
|
| |
Annotate the according parameters accordingly.
v2: move UNUSED decoration in front of parameter declaration
Signed-off-by: Gert Wollny <[email protected]>
Reviewed-by: Brian Paul <[email protected]> (v1)
|
|
|
|
|
|
|
|
|
| |
Annotate the parameters accordingly.
v2: move UNUSED decoration in front of parameter declaration
Signed-off-by: Gert Wollny <[email protected]>
v1: Reviewed-by: Brian Paul <[email protected]> (v1)
|
|
|
|
|
|
|
|
|
| |
Decorate the params with "UNUSED" accordingly.
v2: move UNUSED decoration in front of parameter declaration
Signed-off-by: Gert Wollny <[email protected]>
Reviewed-by: Brian Paul <[email protected]> (v1)
|
|
|
|
|
|
|
|
|
|
| |
Decorate the params accordingly with UNUSED or MAYBE_UNUSED (for params
that are used in debug mode).
v2: move *UNUSED decoration in front of parameter declaration
Signed-off-by: Gert Wollny <[email protected]>
Reviewed-by: Brian Paul <[email protected]> (v1)
|
|
|
|
|
|
|
| |
Explicitely convert one value to compare.
Signed-off-by: Gert Wollny <[email protected]>
Reviewed-by: Brian Paul <[email protected]>
|
|
|
|
|
|
|
| |
Silence the warning by making the conversion to int explicit.
Signed-off-by: Gert Wollny <[email protected]>
Reviewed-by: Brian Paul <[email protected]>
|
|
|
|
|
|
|
|
|
| |
Decorate the unused param accordingly with "UNUSED".
v2: move UNUSED decoration in front of parameter declaration
Signed-off-by: Gert Wollny <[email protected]>
Reviewed-by: Brian Paul <[email protected]> (v1)
|
|
|
|
|
|
|
|
|
| |
Decorate the params accordingly with "UNUSED".
v2: move UNUSED decoration in front of parameter declaration
Signed-off-by: Gert Wollny <[email protected]>
Reviewed-by: Brian Paul <[email protected]> (v1)
|
|
|
|
|
|
|
|
|
| |
Decorate the params accordingly with "UNUSED".
v2: move UNUSED decoration in front of parameter declaration
Signed-off-by: Gert Wollny <[email protected]>
Reviewed-by: Brian Paul <[email protected]> (v1)
|
|
|
|
|
|
|
|
|
| |
Decorate the unused params with "UNUSED".
v2: move UNUSED decoration in front of parameter declaration
Signed-off-by: Gert Wollny <[email protected]>
Reviewed-by: Brian Paul <[email protected]> (v1)
|
|
|
|
|
|
|
|
|
| |
Decorate the unused params with "UNUSED".
v2: move UNUSED decoration in front of parameter declaration
Signed-off-by: Gert Wollny <[email protected]>
Reviewed-by: Brian Paul <[email protected]> (v1)
|