Closes #141
Closes #142

v2 : drop colon in error description (Emil)

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>

- break line of resource creation function declaration

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>

While we are at it:
  - free memory if texture allocation fails

Closes #144
Closes #145
Closes #146

v2: Move the error string creation to extra patch (Emil)
v3: Fix whitespace errors (Emil) and one logic error

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>

Related #139 #141 #142 #143 #144 #146

v2: - fix autotools link order for the fuzzer tests
    - use a local name for egl

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Acked-by: Emil Velikov <emil.velikov@collabora.com>

The function is used from vrend_renderer.c, so it also should be
part of the libvrend.la conveniance library. This requires that the
CONTEXT enums and the variable egl are also part of this library.

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>

v2: Fix another missing return value warning
    Fix a warning about braces around an empty statement

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>

This is needed to check the resource creation parameters.

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>

For this introduce a NDEBUG_UNUSED macro

v2: Replace NDEBUG_UNUSED with MAYBE_UNUSED (Emil)
    Silence unused warnings only if NDEBUG is defined

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>

Lepton Wu authored 5 years ago and Gert Wollny committed 5 years ago

These values are unsigned, but the temporaries are always assumed to be in float
representation, and the temps values will be re-interpreted as integers when
needed. Hence above values must not be value-converted to float but the
bit-representation must be maintained when they are stored in a temp register.

Fixes #105

Signed-off-by: Lepton Wu <lepton@chromium.org>
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>

Another null pointer dereference.  Fixes #67.

Reviewed-by Gert Wollny <gert.wollny@collabora.com>

Fixes #65

Reviewed-by Gert Wollny <gert.wollny@collabora.com>

- Host plane offset queried from the gbm_bo should not be accumulated
  across plane transfer loop iterations.
- Host plane stride cannot be calculated from host plane0 stride in a
  generic way, as alignment adjustments cannot be subsampled. Instead of
  trying of trying to calculate the stride, just query the gbm_bo for
  it.

Signed-off-by: David Stevens <stevensd@chromium.org>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>

Gert Wollny authored 5 years ago and Gert Wollny committed 5 years ago

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Tested-by: Gurchetan Singh <gurchetansingh@chromium.org>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>

Gert Wollny authored 5 years ago and Gert Wollny committed 5 years ago

This patch adds the meson build system as alternative to the
autotools build system.

v2: Thanks to Alexandros for his comments that lead to the following changes:
    - Fix logic for platforms so that the ones selected by 'auto' don't
      fail if the dependencies are missing, but fail when explicitely
      requested
    - declare project version differently
    - set a minimun meson version
    - clean up some braces
    - reformat some code
    - squesh remaining separate patches

v3: require meson 0.46 (Gurchetan)

v4: set minimum gbm version to 18.0.0

v5: - fix gbm version requirement (Gurchetan)
    - don't duplicate glx files (Gurchetan)

Fixes: #129

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Tested-by: Gurchetan Singh <gurchetansingh@chromium.org>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>

Gert Wollny authored 5 years ago and Gert Wollny committed 5 years ago

In systems with more then one graphics card the rendernode is currently
picked based on the order in which they are listed in /dev/dri. Add an
command line option and a environment variable to make it possible to
specify the render node explicitely.

Fixes #135

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>

Gert Wollny authored 5 years ago and Gert Wollny committed 5 years ago

Closes #134

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>

Reviewed-by: David Riley <davidriley@chromium.org>
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>

Reviewed-by: David Riley <davidriley@chromium.org>
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>

Without this, attempting to map the dma-buf after an EGL image was created
from it will fail.

Reviewed-by: David Riley <davidriley@chromium.org>

Just a refactoring.

v2: modify struct order, don't cast pointer (@davidriley)

Reviewed-by: David Riley <davidriley@chromium.org>

David Stevens authored 5 years ago and David Stevens committed 5 years ago

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>

See:  https://github.com/docker/for-linux/issues/711



Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Alexandros Frantzis <alexandros.frantzis@collabora.com>

I can only guess why this is needed, it seems that dQEP is now capable to
ignore the mesa error messages that are printed out in debug mode and captures
the right result instead of indicating "Incomplete"

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>

Clean up some unused code too.

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>

With that we can trigger runs on master to make it easier to update to a
new good version.

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>

The build log is too long to see errors at the end, only log
stderr.

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>

v2: Update results as well

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>

Gert Wollny authored 5 years ago and Gert Wollny committed 5 years ago

If there are regressions dump the number of changes and the head
of the diff file so one doesn't have to search in the artifacts
for a quick look at the problem.

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
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>

The comment claims that this is wrong and only used for debugging,
so replace the code with a debug message.

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>

Gert Wollny authored 5 years ago and Gert Wollny committed 5 years ago

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>

Don't use the shader key when forcing it.

Changing a shader key within vrend_create_shader will always force
virglrenderer to recreate the shader the next time it is looked up, so
memory usage will grow and performance will go down.

Also it is actually not needed if with each emitted back color also has a
corresponding front color is emitted. So restrict forcing two-sided coloring
to the cases where a back color is emitted without a front color, and don't
change the shader key.

Fixes #130

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>

When transferring to/from a planar buffer, multiply by bytes-per-pixel
when computing each plane's stride from stride0. In particular, this
is necessary when transferring to/from biplanar yuv buffers.

Signed-off-by: David Stevens <stevensd@chromium.org>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>

With old DRM implementations, drmPrimeHandleToFD always returned a fd
that could be mapped read/write. However, new kernels require the
DRM_RDWR flag to be specified to create read/write mappings.

Try passing the DRM_RDWR flag but fall back to the old way if that
fails.

Signed-off-by: David Stevens <stevensd@chromium.org>
Acked-By: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>

Running "make check" is quite slow if valgrind is enabled, and most of the time
failures is the result of a leak in the GL driver. Since when building packages
testing is enabled, this will make the package build fail when valgrind is
installed. So instead of automagically enabling valgrind when it is installed,
make it a configure option that also may fail if valgrind is not installed, but
requested.

v2: Use some lowercase variable to report enabled valgrind (Gurchetan)

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>

Gert Wollny authored 5 years ago and Gert Wollny committed 5 years ago

Mesa has seen some fix for ARB_framebuffer_no_attachements that helps
softpipe, so update mesa and the results.

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>

Gert Wollny authored 5 years ago and Gert Wollny committed 5 years ago

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>

The interpolation identifier must come after the precise and invariant
specifier. Also add a newline after these specifiers to make it easier
parsable.

Fixes  #128

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>