This change makes a set of changes to increase the range of gbm-based
resources that can be allocated:
- Add VIRGL_BIND_LINEAR to support linear gbm allocations.
- Relax the bind flag argument check if VIRGL_BIND_SHARED or
VIRGL_BIND_LINEAR is set.
- For resources allocated from gbm, only try to create an image if one
of the render target or sampler view bind flags is set.
- Don't try to calculate the internal image format for external images.
This change also fixes a use-after-free that could occur if external
image creation failed.
Signed-off-by: David Stevens <stevensd@chromium.org>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>
The value should never haven been part of the shader key.
Fixes#132
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>
BGRA formats don't support texture storage on GLES, but GLES supports
creating multisample textures only by using glTexStorage*Multisample,
so switch to a swizzled RGB* format.
v2: Simplify handling of whether we should emulate (Gurchetan)
v3: Fix spaces and use VREND_DEBUG (Gurchetan)
Related #126
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>
When doing a blit in blit_int we use the same context that we use for drawing.
Now with qemu or crostini blits may be issued that are not issued by the guest
program so that it expects a GL_FRAMEBUFFER_SRGB state that might actually have
been clobbered.
Keep track of how GL_FRAMEBUFFER_SRGB is set when the framebuffer state is issued,
and restore its value after doing a blit via glBlitFramebuffer.
Thanks to Lepton Wu for tracking this down.
Fixes#126
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Lepton Wu <lepton@chromium.org>
This looks like a typo, we should use view->target instead of
view->texture->target since we are handling the view.
This fixes these 12 tests:
dEQP-EGL.functional.image.create.gles2_cubemap_positive_x_rgb_texture
dEQP-EGL.functional.image.create.gles2_cubemap_positive_y_rgb_texture
dEQP-EGL.functional.image.create.gles2_cubemap_positive_z_rgb_texture
dEQP-EGL.functional.image.create.gles2_cubemap_negative_x_rgb_texture
dEQP-EGL.functional.image.create.gles2_cubemap_negative_y_rgb_texture
dEQP-EGL.functional.image.create.gles2_cubemap_negative_z_rgb_texture
dEQP-EGL.functional.image.render_multiple_contexts.gles2_cubemap_positive_x_rgb8_texture
dEQP-EGL.functional.image.render_multiple_contexts.gles2_cubemap_positive_y_rgb8_texture
dEQP-EGL.functional.image.render_multiple_contexts.gles2_cubemap_positive_z_rgb8_texture
dEQP-EGL.functional.image.render_multiple_contexts.gles2_cubemap_negative_x_rgb8_texture
dEQP-EGL.functional.image.render_multiple_contexts.gles2_cubemap_negative_y_rgb8_texture
dEQP-EGL.functional.image.render_multiple_contexts.gles2_cubemap_negative_z_rgb8_texture
Signed-off-by: Lepton Wu <lepton@chromium.org>
Reviewed-By: Gert Wollny <gert.wollny@collabora.com>
Maybe it's a bug in mesa, maybe I didn't search long enough, but when I
use the source texture directly in the blit then occationally we hit #112,
and no unbinding of the textures from the texture unit or the framebuffers
involved seems to help.
Using a texture view, however, seem to solve the problem.
Fixes#112
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>
This should fix the following fuzzer failures.
SCARINESS: 10 (null-deref)
v2: Merge various checks (@davidriley)
Reviewed-by: David Riley <davidriley@chromium.org>
Signed-off-by: Gurchetan Singh <gurchetansingh@chromium.org>
This optimization is for some specific situation, but actually we
do see some android apps create 1x1 windows. So remove this optimization
now.
Signed-off-by: Lepton Wu <lepton@chromium.org>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>
"out_supported_structure_mask" is a bitmask representing the structures that
virglrenderer knows how to handle for a given "in_stype_version".
Reviewed-By: Gert Wollny <gert.wollny@collabora.com>
Signed-off-by: Gurchetan Singh <gurchetansingh@chromium.org>
On OpenGL with GLSL < 4.30 the invariant input specifiers must match the
invariant output specifiers of the previous stage.
Fixes#75
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>
In this case the paramters 'format' is not used in 'vrend_format_can_scanout'
and 'gr' are not used in 'vrend_allocate_using_gbm'.
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>
Gallium sends the shaders in the order FS-[GS]-[TES]-[TCS]-VS. If an old
shader program is still bound when a new shader is send, then it would use the
old bounds shader of the previous stage to evaluate the shader key and code
creation might create an invalid shader. Hence, if the shader to be translated
is not yet bound ignore the previous stage when evaluating the shader key
(Gurchtan)
Fixes#114
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>
It's desirable to query or modify virgl_renderer state for various values.
We've done it for resources (virgl_renderer_get_fd_for_texture /
virgl_renderer_get_fd_for_texture2 / virgl_renderer_resource_get_info) a
few times, but more parameters are often desired.
Let's define a protocol instead. That way, more queries/operations (is "ctx
{num} lost?", "is x host feature supported?") can be formulated.
It's also possible to put the protocol in virgl_hw.h, so in theory
the guest can recieve a response directly from virglrenderer. Any
opinions on this?
v2: virgl_renderer_query --> virgl_renderer_execute
Reviewed-by: David Riley <davidriley@chromium.org>
Signed-off-by: Gurchetan Singh <gurchetansingh@chromium.org>
It's good to tell the guest about these formats.
v2: move this check with the rest of the v2 caps
remove host version check bump (@kusma)
Reviewed-by: David Riley <davidriley@chromium.org>
Signed-off-by: Gurchetan Singh <gurchetansingh@chromium.org>
Certain 2D allocations will take the GBM path if the option
is set.
Reviewed-by: David Riley <davidriley@chromium.org>
Signed-off-by: Gurchetan Singh <gurchetansingh@chromium.org>
virgl_renderer_transfer_write_iov doesn't work for YUV buffers, since
it's based on glTexSubImage.
We assume the YUV image is not disjoint, since that's currently what
Mesa gbm supports.
For RGBA buffers, this will also eliminate a copy in the cases
where the gbm_bo_map(..) implementation doesn't use a shadow buffer
[i915, mediatek, rockchip].
Reviewed-by: David Riley <davidriley@chromium.org>
Signed-off-by: Gurchetan Singh <gurchetansingh@chromium.org>
vrend_clear executes the Gallium clear command which is only called
when the whole viewport is cleared. So far mesa was doing excessive checks
on the scissors, thereby updating the scissors to framebuffer size when they
were disabled, and the according state changes were transmitted to the host.
With mesa/2037478 this was optimized away, so that not disabling the scissors
in the clear command manifested itself as a regression in a number of tests.
Keeping track of the scissor state in the hardware and disabling the scissors
before the clear is executes, and re-enabling them according to the last state
fixes this.
Closes#116
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>
virglrenderer doesn't know with which parameters the external
image was created, so it doesn't make sense to emulate it.
This fixes Portal, HL2 etc. not going past the loading screen
in Crostini.
Reviewed-by: David Riley <davidriley@chromium.org>
Signed-off-by: Gurchetan Singh <gurchetansingh@chromium.org>
Instead of setting the depth range values directly just mark the viewport
as dirty and update the state later together with the other viewport
properties. This fixes some issues with the depth clamp emulation and
geometry shaders.
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>
With thie new version we can enable the depth clamp emulation in the
guest.
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>
The value is only supported for GL >= 3.2 or GLES >= 3.0, on earlier versions
assume the minimum required number.
Closes#113
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>
There are more instances where the sampler view and the texture are not
in the same context. So do more un-binding of textures after they were
used.
Related: #112
Signed-off-by: Gert Wollny's avatarGert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>
If nr_samples is 1 we allocate textures as MULTISAMPLE, so we also need to handle
the textures in later use as a multisample texture. In addition, on GLES blits to
a multisample texture are not allowed as target, and we must use the GL blit
fallback for all texture types.
v2: Correct more checks including the one in check_resource_valid (Gurchetan)
v3: Also update the tests to use only 0 for no samples and >= 1 for samples
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>
For guests that support this interface this:
Fixes#91Fixes#94
v2: (whole series) Switch from string interface to integer values
for tweaks (Gurchetan).
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>
By default emulate GL_SAMPLES_PASSED by using GL_ANY_SAMPLES_PASSED
and return the value 1024 if any sample has passed. Also add a tweak so
one can set this number to a different value.
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>
v2: fix use of bindings vs. flags
v3: correct typo (Gurchetan) fb55a453
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>
With vtest this swizzling is necessary, but with qemu it is not, so make
it a tweak.
v2: Use original blit format to check the format type in blitter
v3: Correct typo (Gurchetan)
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>
On GLES B8G8R8A8_SRGB is not available and B8G8R8A8_UNORM can not be
created as immutable. In order to have versions of these formats that
can be created immutable on GLES provide the formats R8G8B8A8_SRGB and
_UNORM with swizzling to BGRA that can then be used instead of the BGRA
formats. Since this may break things these swizzled versions will later
be only enabled when the guest requests them.
v2: Check the base swizzled BGRA format only on GL
v3: Correct type in method name (Gurchetan)
v4: Add PREFER_EMULATED_BRGA to according format binding flags (Gurchetan)
v5: Add comment about formats nou to be used in the guest (Gurchetan)
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>
When there is no change to UBOs of a shader stage, we still need to
increment next_ubo_id to match bind_ubo_locs. This was regressed by
commit 2192c92 (renderer: only update dirty constbufs).
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Do not use in/out parameters for sampler_id and ubo_id. They are
easier to get wrong.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
ubo_used_mask is similar to samplers_used_mask but for UBOs.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
As pointed out by Erik, this is a rogue-client behaviour which we
generally just ignore.
Suggested-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-By: Gert Wollny <gert.wollny@collabora.com>
GL 4.2 supports depth layouts that make it possible to optimize by
enabling early depth tests and still being able to write a new
z value in the fragment shader under specific circumstances.
Fixes#106
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>
The fuzzer series added a few warnings that need to be silenced.
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: David Riley <davidriley@chromium.org>
Destroy resources even if there is a pipe_reference.
Signed-off-by: David Riley <davidriley@chromium.org>
Reviewed-By: Gert Wollny <gert.wollny@collabora.com>
Transfer strides are either set internally or coming from
virgl_renderer_transfer_{write,read}_iov. As far as I can tell,
Mesa always passes sane values. We also expect sane values in
places like vrend_renderer_transfer_write_iov,
vrend_transfer_send_readpixels, and vrend_transfer_send_getteximage
already. Let's reject bad strides.
This fixes, for example, transfers to 1D array (thus box->height is
1) with non-default stride.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Alexandros Frantzis <alexandros.frantzis@collabora.com>
The origin change is already handled within Gallium, so it is only
necessary to deal with the Z range property.
v2: don't mark viewport as dirty when halfz is changed, this is already
handled by Gallium (Gurchetan)
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>
Just like all the other non-static vrend_shader functions.
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Ensure we can't create a texture resource with zero width, to
guarantee that we get a non-zero stride, since either:
1. A non-zero stride was specified in the transfer request, or
2. We will calculate a stride based on texture width which we guarantee
is not zero
Signed-off-by: Alexandros Frantzis <alexandros.frantzis@collabora.com>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Gurchetan Singh
David Stevens
Gert Wollny
Lepton Wu
Chia-I Wu
Emil Velikov
David Riley
Alexandros Frantzis