With that Unigine Heaven goes from around 56 FPS to 58
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
With that shaders that are used more frequently will remain at the
top of the list, and since the search in the shader program list is
linear, this should save some time.
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
This gives a 6% speed improvement with Unigine Heaven
v2: remove the unlikely for dual_src, the whole statement will
be changed later (Chia-I Wu)
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
if-chains can usually not be optimized very well, switch/case might be
implemented as a jump table.
v2: Add a break in bese before default (Chia-I Wu)
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
This makes the code a bit easier to read and may help the complier
to avoid needless dereferencing of pointers.
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
v2: add comment about that INDEX is not used (Chia-I Wu)
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
It gives clients access to virgl_renderer_context_create_with_flags.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Isaac Bosompem <mrisaacb@google.com>
Comparing to VCMD_GET_{CAPS,CAPS2}, it allows any capset id/version
to be queried.
This is similar to linux's DRM_IOCTL_VIRTGPU_GET_CAPS.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Isaac Bosompem <mrisaacb@google.com>
There is no parameter defined yet, but the command can be used to query
optional features. It also allows new features to be introduced without
bumping up the protocol version.
This is similar to linux's DRM_IOCTL_VIRTGPU_GETPARAM.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Isaac Bosompem <mrisaacb@google.com>
The goal is to allow clients to negoticate protocol versions, capsets,
and in the future, parameters and context types before
virgl_renderer_context_create is called.
Specifically, these commands should not trigger
virgl_renderer_context_create
VCMD_PING_PROTOCOL_VERSION
VCMD_PROTOCOL_VERSION
VCMD_GET_CAPS
VCMD_GET_CAPS2
VCMD_RESOURCE_BUSY_WAIT when res_id==0 (for legacy reasons)
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Isaac Bosompem <mrisaacb@google.com>
Move virgl_renderer_context_create call from vtest_create_context to
the new vtest_lazy_init_context. This defers context initialization
until the first command after VCMD_CREATE_RENDERER.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Isaac Bosompem <mrisaacb@google.com>
struct vtest_context is about to get more complex. Use
vtest_destroy_context so that we have a single point that does context
destruction.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Isaac Bosompem <mrisaacb@google.com>
Comparing to virgl_renderer_context_create, it allows flags to be
specified. flags can only be used to specify the capset id in this
commit, but that may change in the future.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Isaac Bosompem <mrisaacb@google.com>
Unlike out_fourcc, where the default value 0 is DRM_FORMAT_INVALID, the
default value 0 for out_modifier is DRM_FORMAT_MOD_LINEAR.
This changes nothing in practice as the only user of
virgl_renderer_execute is crosvm, and crosvm takes the
ENABLE_MINIGBM_ALLOCATION path instead.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Isaac Bosompem <mrisaacb@google.com>
Add virgl_resource::map_info and return it in
virgl_renderer_resource_get_map_info.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Isaac Bosompem <mrisaacb@google.com>
This allows the callers to modify the returned virgl_resource without
looking up.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Isaac Bosompem <mrisaacb@google.com>
When a virgl_resource wraps a pipe_resource, it is said to be typed
because pipe_resource contains the type information.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Isaac Bosompem <mrisaacb@google.com>
Most of the time we pass a string that can simply be forwarded to
the perfetto trace function.
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
This silences a warning about non-static variable not being previously
declared.
v2: make table const (Chia-I Wu)
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Lock crates to version specified by crosvm when building
to ensure that the build succeeds.
Signed-off-by: Rohan Garg <rohan.garg@collabora.com>
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Rename it to vrend_resource_alloc_texture to match
vrend_resource_alloc_buffer.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Add vrend_resource_fixup_emulated_bgra to fix up emulated bgra.
This is a refactor with no real change.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Add vrend_resource_alloc_buffer to allocate the storage for PIPE_BUFFER.
This is a refactor with no real change.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Add vrend_resource_create helper to allocate and initialize the
vrend_resource struct.
This is a refactor with no real change.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Most of the work of this MR was done by Tomeu.
Changes:
- update the Dockerfile to use the new virglrenderer tracing option
- use command line parsing to pass the parameters when running a trace
- move some files and rename scripts
v2: - unify gfx-pps build (Tomeu) and use to suggested branch (Fahien)
- use upstream tagged version for igt-gpu-tools (Tomeu)
- add parameters to set benchmark loop count (Louis-Francis)
- add parameter to set perfetto loop count or loopless run
v3: - compile virglrenderer against minigbm (Rohan)
- compile minigbm with driver i915 (Rohan)
v4: Don't build mesa-gbm (instead of deleting the library and
header later)
v5: add option to run script whether to wait after each frame
v6: correct gfx-pps version and add gdb
v7: fix tag for gfx-pps
v8: - when building the container use the user ID and group
id of the user who runs the container build
- rework passing command line parameters
- add a command line parameter to record per-frame
images
- Use debugoptimized build for gfx-pps (Tomeu)
v9: disable buffer storage, there seems to be a bug in
context handling triggered by this
v10: move to upstream kernel 5.9.12
v11: - remove writing txt version of host trace
- add debug flagto enable tracing EGL
- don't force-remove libgbm1 it is no longer a problem to
have it
v12: update kernel to 5.20.5
v13: refactor trace merge (Rohan)
v14: - Fix refactoring (Rohan)
- Add license infro to perfetto merge script
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-By: Rohan Garg <rohan.garg@collabora.com>
This is now already checked before we decode the command.
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
v2: drop decode state and use local variables to handle keeping
track of the decode buffer state (Chia-I Wo)
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
With that we don't need to resolve the pointer chain in each decode
function.
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Since each switch case only results is a function call we can also execute
the call directly by using a call table.
v2: Add VIRGL_CCMD_NOP and validate table when a context is
created (Chai-I Wu).
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
If GL_POINT_SPRITE mode is enabled and a texture has
GL_COORD_REPLACE enabled, the texture will apply coord_replace
to all primitive types, instead of only GL_POINTS as expected.
This change checks the prim_mode for every draw call and selects
the appropriate shader variant to enable coord_replace only when
rendering GL_POINTS primitives.
Closes: https://gitlab.freedesktop.org/virgl/virglrenderer/-/issues/188
Signed-off-by: Ryan Neph <ryanneph@google.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
We don't know what internal format the GL driver uses and we have to
glFinish and do a gbm transfer unless VREND_STORAGE_GL_IMMUTABLE is set.
This is motivated by Mesa's internal format change from GL_RGB to
GL_RGB8 for VIRGL_FORMAT_B8G8R8X8_UNORM in
5e4d69ec78.
That change caused glTexSubImage2D(..., GL_BGRA_EXT, GL_UNSIGNED_BYTE)
to be rejected by _mesa_gles_error_check_format_and_type.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>
The bo contents may be undefined when only GBM_BO_TRANSFER_WRITE is set
(e.g., with radeonsi). When transferring to a subregion, we need to set
GBM_BO_TRANSFER_READ as well.
gbm_bo_map2 does allow us to map a subregion, but not all
implementations work correctly.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>
We can now use gbm_bo_map(..) directly to match Mesa.
If minigbm doesn't define the needed use flags, we define it
internally.
This should allow us to build with minigbm allocation enabled and
use the Mesa gbm header.
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Fixes the order of params passed to emit_ios() calls
(swaps winsys_adjust_y_emitted and force_color_two_side).
Signed-off-by: Ryan Neph <ryanneph@google.com>
Reviewed-By: David Riley <davidriley@chromium.org>
EGL_ANDROID_native_fence_sync only works on gles profiles.
Signed-off-by: Lepton Wu <lepton@chromium.org>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>
Gert Wollny
Chia-I Wu
Rohan Garg
Eric Anholt
Ryan Neph
Gurchetan Singh
Lepton Wu