Rob Clark [Fri, 1 Jul 2016 02:52:23 +0000 (22:52 -0400)]
freedreno: update valid_buffer_range for SO buffers
Signed-off-by: Rob Clark <robdclark@gmail.com>
Rob Clark [Wed, 22 Jun 2016 18:45:25 +0000 (14:45 -0400)]
freedreno/ir3: support non-user_buffer consts
Signed-off-by: Rob Clark <robdclark@gmail.com>
Rob Clark [Fri, 20 May 2016 20:00:26 +0000 (16:00 -0400)]
freedreno/a2xx: move setup/restore cmds into binning pass
Rather than doing a separate submit at context create, move these cmds
to before first tile, as is done on a3xx/a4xx. Otherwise state can
be overwritten by other contexts.
Signed-off-by: Rob Clark <robdclark@gmail.com>
Rob Clark [Fri, 20 May 2016 17:41:44 +0000 (13:41 -0400)]
freedreno: pass index buffer as a pipe_resource
This will be useful in a following patch.
Signed-off-by: Rob Clark <robdclark@gmail.com>
Rob Clark [Fri, 20 May 2016 16:54:29 +0000 (12:54 -0400)]
freedreno: switch emit_const_bo() to take prsc's
We can push the unwrap of pipe_resource down.
Signed-off-by: Rob Clark <robdclark@gmail.com>
Hans de Goede [Wed, 29 Jun 2016 12:32:19 +0000 (14:32 +0200)]
nv30: Fix "array subscript is below array bounds" compiler warning
gcc6 does not like the trick where we point to one entry before the
array start and then start a while with a pre-increment.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Hans de Goede [Wed, 29 Jun 2016 12:23:23 +0000 (14:23 +0200)]
nouveau: Fix a couple of "foo may be used uninitialized' compiler warnings
These are all new false positives with gcc6.
In nouveau_compiler.c: gcc6 no longer assumes that passing a pointer
to a variable into a function initialises that variable.
In nv50_ir_from_tgsi.cpp op and mode are not set if there are 0
enabled dst channels, this never happens, but gcc cannot know this.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ilia Mirkin <imirkin@alum.mit.edu>
Hans de Goede [Wed, 29 Jun 2016 11:09:11 +0000 (13:09 +0200)]
nouveau: Fix gcc6 / c++11 auto_ptr deprecation compiler warnings
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Hans de Goede [Mon, 25 Apr 2016 09:42:12 +0000 (11:42 +0200)]
nouveau: Add support for SV_WORK_DIM
Add support for SV_WORK_DIM for nvc0 and nve4.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Hans de Goede [Wed, 29 Jun 2016 09:23:45 +0000 (11:23 +0200)]
nvc0: Make NVC0_CB_AUX_GRID_INFO take an index argument
This brings it inline with the other macros like NVC0_CB_AUX_UBO_INFO
and NVC0_CB_AUX_TEX_INFO.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Hans de Goede [Fri, 22 Apr 2016 12:47:05 +0000 (14:47 +0200)]
clover: Pass work_dim parameter of clEnqueueNDRangeKernel() to driver
In order to implement get_work_dim() the driver may need to know the
clEnqueueNDRangeKernel() work_dim parameter, so pass it to the driver.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Hans de Goede [Fri, 22 Apr 2016 12:26:50 +0000 (14:26 +0200)]
tgsi: Add WORK_DIM System Value
Add a new WORK_DIM SV type, this is will return the grid dimensions
(1-4) for compute (opencl) kernels.
This is necessary to implement the opencl get_work_dim() function.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Alejandro Piñeiro [Thu, 30 Jun 2016 12:41:01 +0000 (14:41 +0200)]
mesa/main: fix error checking logic on CopyImageSubData
For the case (both src or dst) where we had a texobject, but the
texobject target was not the same that the method target, this spec
paragraph was appplied:
/* Section 18.3.2 (Copying Between Images) of the OpenGL 4.5 Core
* Profile spec says:
*
* "An INVALID_VALUE error is generated if either name does not
* correspond to a valid renderbuffer or texture object according
* to the corresponding target parameter."
*/
But for that case, the correct spec paragraph should be:
/* Section 18.3.2 (Copying Between Images) of the OpenGL 4.5 Core
* Profile spec says:
*
* "An INVALID_ENUM error is generated if either target is
* not RENDERBUFFER or a valid non-proxy texture target;
* is TEXTURE_BUFFER or one of the cubemap face selectors
* described in table 8.18; or if the target does not
* match the type of the object."
*/
specifically the last sentence: "or if the target does not match the
type of the object".
This patch fixes the error returned (s/INVALID/ENUM) for that case,
and moves up the INVALID_VALUE spec paragraph, as that case (invalid
texture object) was handled before.
Fixes:
GL44-CTS.copy_image.target_miss_match
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Dave Airlie [Mon, 6 Jun 2016 23:39:55 +0000 (09:39 +1000)]
st/glsl_to_tgsi: don't increase immediate index by 1.
Immediates are stored into a separate table, and are
consolidated, so if we get an immediate we don't need
to offset it as the index it has is correct.
Cc: "11.2 12.0" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Ilia Mirkin [Wed, 29 Jun 2016 04:11:57 +0000 (00:11 -0400)]
st/mesa: get max supported number of image samples from driver
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Ilia Mirkin [Wed, 29 Jun 2016 04:08:01 +0000 (00:08 -0400)]
nvc0: fix up image support for allowing multiple samples
Basically we just have to scale up the coordinates and then add the
relevant sample offset. The code to handle this was already largely
present from Christoph's earlier attempts to pipe images through back in
the dark ages, this just hooks it all up.
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Nicolai Hähnle [Fri, 24 Jun 2016 15:54:51 +0000 (17:54 +0200)]
st/mesa: check the texture image level in st_texture_match_image
Otherwise, 1x1 images of arbitrarily high level are accepted.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=96639#add_comment
Cc: 11.2 12.0 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Nicolai Hähnle [Fri, 24 Jun 2016 14:35:36 +0000 (16:35 +0200)]
st/mesa: an incomplete texture may have a zero-size first image
Fixes a regression introduced by commit
42624ea83 which triggered
an assertion in
dEQP-GLES2.functional.texture.completeness.cube.not_positive_level_0
While stImage must have a non-zero size as verified by the caller, we also
look at the size of the base image in an attempt to make a better guess at
the level0 size (this is important when the base image size is odd). However,
the base image may have a zero size even when it exists.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=96629
Cc: 12.0 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Nayan Deshmukh [Wed, 29 Jun 2016 13:24:01 +0000 (18:54 +0530)]
st/vdpau: use bicubic filter for scaling(v6.1)
use bicubic filtering as high quality scaling L1.
v2: fix a typo and add a newline to code
v3: -render the unscaled image on a temporary surface (Christian)
-apply noise reduction and sharpness filter on
unscaled surface
-render the final scaled surface using bicubic
interpolation
v4: support high quality scaling
v5: set dst_area and dst_clip in bicubic filter
v6: set buffer layer before setting dst_area
v6.1: add PIPE_BIND_LINEAR when creating resource
Signed-off-by: Nayan Deshmukh <nayan26deshmukh@gmail.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Nayan Deshmukh [Wed, 29 Jun 2016 13:24:00 +0000 (18:54 +0530)]
vl: add a bicubic interpolation filter(v5)
This is a shader based bicubic interpolater which uses cubic
Hermite spline algorithm.
v2: set dst_area and dst_clip during scaling (Christian)
v3: clear the render target before rendering
v4: intialize offsets while initializing shaders
use a constant buffer to send dst_size to frag shader
small changes to reduce calculation in shader
v5: send half pixel offset instead of sending dst_size
Signed-off-by: Nayan Deshmukh <nayan26deshmukh@gmail.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Vinson Lee [Thu, 30 Jun 2016 03:15:03 +0000 (20:15 -0700)]
mesa/st: Use 'struct nir_shader' instead of 'nir_shader'.
Fix this build error with GCC 4.4.
CC state_tracker/st_nir_lower_builtin.lo
In file included from state_tracker/st_nir_lower_builtin.c:61:
state_tracker/st_nir.h:34: error: redefinition of typedef ‘nir_shader’
../../src/compiler/nir/nir.h:1830: note: previous declaration of ‘nir_shader’ was here
Suggested-by: Rob Clark <robdclark@gmail.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=96235
Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Rob Clark <robdclark@gmail.com>
Alejandro Piñeiro [Mon, 27 Jun 2016 08:00:58 +0000 (10:00 +0200)]
docs: update MESA_DEBUG envvar documentation.
silent, flush, incomplete_tex and incomplete_fbo flags were not
documented (see src/mesa/main.debug.c for more info).
FP is not checked anymore.
v2 (Brian Paul):
* MESA_DEBUG accepts a comma-separated list of parameters.
* Clarify how MESA_DEBUG behaves with mesa debug and release builds.
* Updated wording.
v3: Better wording for one paragraph (Brian Paul)
Reviewed-by: Brian Paul <brianp@vmware.com>
Alejandro Piñeiro [Tue, 28 Jun 2016 11:16:33 +0000 (13:16 +0200)]
i965: intel_texture_barrier reimplemented
Fixes:
GL44-CTS.texture_barrier_ARB.same-texel-rw-multipass
On Haswell, Broadwell and Skylake (note that in order to execute that
test, it is needed to override GL and GLSL versions).
On gen6 this test was already working without this change. It keeps
working after it.
This commit replaces the call to brw_emit_mi_flush for gen6+ with two
calls to brw_emit_pipe_control_flush:
* The first one with RENDER_TARGET_FLUSH and CS_STALL set to initiate
a render cache flush after any concurrent rendering completes and
cause the CS to stop parsing commands until the render cache
becomes coherent with memory.
* The second one have TEXTURE_CACHE_INVALIDATE set (and no CS stall)
to clean up any stale data from the sampler caches before rendering
continues.
Didn't touch gen4-5, basically because I don't have a way to test
them.
More info on commits:
0aa4f99f562a05880a779707cbcd46be459863bf
72473658c51d5e074ce219c1e6385a4cce29f467
Thanks to Curro to help to tracking this down, as the root case was a
hw race condition.
v2: use two calls to pipe_control_flush instead of a combination of
gen7_emit_cs_stall_flush and brw_emit_mi_flush calls (Curro)
v3: no need to const cache invalidation (Curro)
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Ilia Mirkin [Fri, 1 Jul 2016 04:59:58 +0000 (00:59 -0400)]
nv30: go back to not using viewport validate function for swtnl
The output of draw requires a null viewport transform, which the regular
code is ill-equiped to do. Reinstate the original settings in the render
path, and add setting of the viewport clip polygon based on fb
width/height (as that is all taken care of by draw).
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Ilia Mirkin [Wed, 22 Jun 2016 02:51:36 +0000 (22:51 -0400)]
nv30: fix viewport clipping settings to be based on viewport, not rt
This fixes a ton of "*clip*" dEQP GLES2 tests, as well as
triangle-guardband-viewport in piglit.
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Brian Paul [Fri, 1 Jul 2016 00:15:19 +0000 (18:15 -0600)]
gallium/util: check for window cliprects in util_can_blit_via_copy_region()
We can't blit with resource_copy_region() if there are window clip rects.
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Chuck Atkins [Wed, 29 Jun 2016 02:32:02 +0000 (22:32 -0400)]
gallium: Force blend color to 16-byte alignment
This aligns the 4-element color float array to 16 byte boundaries. This
should allow compiler vectorizers to generate better optimizations.
Also fixes broken vectorization generated by Intel compiler.
v2: Fixed indentation and added a lengthy comment explaining the
reason for the alignment.
Cc: <mesa-stable@lists.freedesktop.org>
Reported-by: Tim Rowley <timothy.o.rowley@intel.com>
Tested-by: Tim Rowley <timothy.o.rowley@intel.com>
Signed-off-by: Chuck Atkins <chuck.atkins@kitware.com>
Acked-by: Roland Scheidegger <sroland@vmware.com>
Chuck Atkins [Tue, 28 Jun 2016 19:50:47 +0000 (15:50 -0400)]
swr: Refactor checks for compiler feature flags
Encapsulate the test for which flags are needed to get a compiler to
support certain features. Along with this, give various options to try
for AVX and AVX2 support. Ideally we want to use specific instruction
set feature flags, like -mavx2 for instance instead of -march=haswell,
but the flags required for certain compilers are different. This
allows, for AVX2 for instance, GCC to use -mavx2 -mfma -mbmi2 -mf16c
while the Intel compiler which doesn't support those flags can fall
back to using -march=core-avx2.
This addresses a bug where the Intel compiler will silently ignore the
AVX2 instruction feature flags and then potentially fail to build.
v2: Pass preprocessor-check argument as true-state instead of
false-state for clarity.
v3: Reduce AVX2 define test to just __AVX2__. Additional defines suchas
__FMA__, __BMI2__, and __F16C__ appear to be inconsistently defined
w.r.t thier availability.
v4: Fix C++11 flags being added globally and add more logic to
swr_require_cxx_feature_flags
Cc: <mesa-stable@lists.freedesktop.org>
Reviewed-by: Tim Rowley <timothy.o.rowley@intel.com>
Tested-by: Tim Rowley <timothy.o.rowley@Intel.com>
Signed-off-by: Chuck Atkins <chuck.atkins@kitware.com>
Brian Paul [Wed, 11 May 2016 15:20:04 +0000 (09:20 -0600)]
st/wgl: make own_mutex() non-static
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
Andres Gomez [Thu, 30 Jun 2016 13:37:25 +0000 (16:37 +0300)]
glsl: atomic counters are different than their uniforms
The linker deals with atomic counters in terms of uniforms but the
data structure are called after the atomic counters.
Renamed the data structures used in the linker for disambiguation.
Reviewed-by: Timothy Arceri <timothy.arceri@collabora.com>
Signed-off-by: Andres Gomez <agomez@igalia.com>
Andres Gomez [Thu, 30 Jun 2016 13:37:11 +0000 (16:37 +0300)]
glsl: count atomic counters correctly
Currently the linker uses the uniform count for the total number of
atomic counters. However uniforms don't include the innermost array
dimension in their count, but atomic counters are expected to include
them.
Although the spec doesn't directly state this, it's clear how offsets
will be assigned for arrays.
From OpenGL 4.2 (Core Profile), page 98:
" * Arrays of type atomic_uint are stored in memory by element
order, with array element member zero at the lowest offset. The
difference in offsets between each pair of elements in the
array in basic machine units is referred to as the array
stride, and is constant across the entire array. The stride can
be queried by calling GetIntegerv with a pname of
ATOMIC_COUNTER_- ARRAY_STRIDE after a program is linked."
From that it is clear how arrays of atomic counters will interact with
GL_MAX_ATOMIC_COUNTER_BUFFER_SIZE.
For other kinds of uniforms it's also clear that each entry in an
array counts against the relevant limits.
Hence, although inferred, this is the expected behavior.
Fixes GL44-CTS.arrays_of_arrays_gl.AtomicDeclaration
Reviewed-by: Timothy Arceri <timothy.arceri@collabora.com>
Signed-off-by: Andres Gomez <agomez@igalia.com>
Brian Paul [Tue, 28 Jun 2016 23:15:57 +0000 (17:15 -0600)]
svga: use SVGA3D_vgpu10_BufferCopy() for buffer copies
So that we do copies host-side rather than in the guest with map/memcpy.
Tested with piglit arb_copy_buffer-subdata-sync test and new
arb_copy_buffer-intra-buffer-copy test.
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Acked-by: Roland Scheidegger <sroland@vmware.com>
Brian Paul [Thu, 23 Jun 2016 02:38:06 +0000 (20:38 -0600)]
svga: add SVGA3D_vgpu10_BufferCopy()
Acked-by: Roland Scheidegger <sroland@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Brian Paul [Thu, 30 Jun 2016 19:27:57 +0000 (13:27 -0600)]
svga: flush buffers when mapping for reading
With host-side buffer copies (via SVGA3D_vgpu10_BufferCopy()) we have
to make sure any pending map-write operations are completed before reading
if the buffer is dirty. Otherwise the ReadbackSubResource operation could
get stale data from the host buffer.
This allows the piglit arb_copy_buffer-subdata-sync test to pass when
we start using the SVGA3D_vgpu10_BufferCopy command.
v2: check the sbuf->dirty flag in the outer conditional, per Charmaine.
Acked-by: Roland Scheidegger <sroland@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Neha Bhende [Thu, 23 Jun 2016 17:21:31 +0000 (11:21 -0600)]
svga: enable ARB_copy_image extension in the driver
Reviewed-by: Brian Paul <brianp@vmware.com>
Acked-by: Roland Scheidegger <sroland@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Brian Paul [Tue, 28 Jun 2016 23:13:57 +0000 (17:13 -0600)]
svga: try blitting with copy region in more cases
We previously could do blits with util_resource_copy_region() when doing
'loose' format checking. Also do blits with util_resource_copy_region()
when the blit src/dst formats (not the underlying resources) exactly
match. Needed for GL_ARB_copy_image.
Acked-by: Roland Scheidegger <sroland@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Brian Paul [Thu, 23 Jun 2016 17:57:08 +0000 (11:57 -0600)]
svga: use copy_region_vgpu10() for region copies when possible
v2: remove extra svga_define_texture_level() call, per Charmaine.
Acked-by: Roland Scheidegger <sroland@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Neha Bhende [Tue, 28 Jun 2016 23:20:43 +0000 (17:20 -0600)]
svga: use vgpu10 CopyRegion command when possible
Do texture->texture copies host-side with this command when possible.
Use the previous software fallback otherwise.
Reviewed-by: Brian Paul <brianp@vmware.com>
Acked-by: Roland Scheidegger <sroland@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Brian Paul [Mon, 27 Jun 2016 17:18:10 +0000 (11:18 -0600)]
svga: set render target flag for snorm surfaces
We don't normally support rendering to SNORM surfaces, but with
GL_ARB_copy_image we can copy to them if we treat them as typeless
and use a UNORM surface view.
Acked-by: Roland Scheidegger <sroland@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Brian Paul [Mon, 27 Jun 2016 17:17:45 +0000 (11:17 -0600)]
svga: add new svga_format_is_uncompressed_snorm() helper
Acked-by: Roland Scheidegger <sroland@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Brian Paul [Mon, 27 Jun 2016 17:16:03 +0000 (11:16 -0600)]
svga: adjust sampler view format for RGBX
We previously handled the case of a RGBX sampler view of a RGBA surface.
Add the reverse case too. For GL_ARB_copy_image.
Acked-by: Roland Scheidegger <sroland@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Brian Paul [Mon, 27 Jun 2016 17:15:07 +0000 (11:15 -0600)]
svga: adjust render target view format for RGBX
For GL_ARB_copy_image we may be asked to create an RGBA view of
a RGBX surface. Use an RGBX view format for that case.
Acked-by: Roland Scheidegger <sroland@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Neha Bhende [Thu, 2 Jun 2016 21:35:20 +0000 (14:35 -0700)]
svga: don't advertise support for R32G32B32_UINT/SINT surface formats
We want to be able to copy between different 32-bit, 3-channel surface
formats for GL_ARB_copy_image but since we don't support R32G32B32_FLOAT
for textures (it's not blendable and wouldn't work for render to texture)
we can't support 32-bit, 3-channel integer formats.
The state tracker will choose 4-channel formats instead.
Fixes the piglit arb_copy_image-format test for several cases.
Note: This change may need to be revisited if/when the texture_view exension
is enabled in driver.
Reviewed-by: Brian Paul <brianp@vmware.com>
Acked-by: Roland Scheidegger <sroland@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Brian Paul [Thu, 23 Jun 2016 02:42:37 +0000 (20:42 -0600)]
svga: use untyped surface formats in most cases
This allows us to do copies between different, but compatible, surface
formats such as RGBA8_UNORM, RGBA8_SINT, RGBA8_UINT, etc. for
GL_ARB_copy_image.
Acked-by: Roland Scheidegger <sroland@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Brian Paul [Thu, 23 Jun 2016 14:11:25 +0000 (08:11 -0600)]
gallium/util: add tight_format_check param to util_can_blit_via_copy_region()
The VMware driver will use this for implementing GL_ARB_copy_image.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Brian Paul [Wed, 8 Jun 2016 20:41:48 +0000 (14:41 -0600)]
gallium/util: simplify a few things in util_can_blit_via_copy_region()
Since only the src box can have negative dims for flipping, just
comparing the src/dst box sizes is enough to detect flips.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Brian Paul [Wed, 8 Jun 2016 20:36:08 +0000 (14:36 -0600)]
gallium/util: new util_try_blit_via_copy_region() function
Pulled out of the util_try_blit_via_copy_region() function. Subsequent
changes build on this.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Neha Bhende [Tue, 28 Jun 2016 19:59:19 +0000 (12:59 -0700)]
svga: Fix failures caused in fedora 24
SVGA_3D_CMD_DX_GENRATE_MIPMAP & SVGA_3D_CMD_DX_SET_PREDICATION commands
are not presents in fedora 24 kernel module. Because of this
reason application like supertuxkart are not running.
v2: Add few comments and code modifications suggested by Brian P.
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Brian Paul [Fri, 20 May 2016 20:24:32 +0000 (14:24 -0600)]
st/wgl: remove unneeded inline qualifiers
No effect on size of the .o files (optimized build).
Reviewed-by: José Fonseca <jfonseca@vmware.com>
Brian Paul [Fri, 20 May 2016 20:24:59 +0000 (14:24 -0600)]
st/wgl: add a stw_device::initialized field
Set when the stw_dev object's initialization is completed. We test
for this in the window callback function to avoid potential crashes
on start-up in multi-threaded applications.
Reviewed-by: José Fonseca <jfonseca@vmware.com>
Brian Paul [Fri, 20 May 2016 20:16:18 +0000 (14:16 -0600)]
st/wgl: refactor framebuffer locking code
Split the old stw_framebuffer_reference() function into two new
functions: stw_framebuffer_reference_locked() which increments
the refcount and stw_framebuffer_release_locked() which decrements
the refcount and destroys the buffer when the count hits zero.
Original patch by Jose. Modified by Brian (clean-ups, lock assertion
checks, etc).
Reviewed-by: José Fonseca <jfonseca@vmware.com>
José Fonseca [Fri, 20 May 2016 18:11:56 +0000 (12:11 -0600)]
st/wgl: rename curctx to old_ctx in stw_make_current()
Reviewed-by: Brian Paul <brianp@vmware.com>
Brian Paul [Thu, 12 May 2016 22:33:30 +0000 (16:33 -0600)]
st/wgl: release the pbuffer DC at the end of wglBindTexImageARB()
Otherwise we were leaking DC GDI objects and if wglBindTexImageARB()
was called enough we'd eventually hit the GDI limit of 10,000 objects.
Things started failing at that point.
v2: also release DC if we return early, per Charmaine.
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Reviewed-by: José Fonseca <jfonseca@vmware.com>
Matt Turner [Mon, 27 Jun 2016 23:31:09 +0000 (16:31 -0700)]
mesa: Close fp on error path.
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Matt Turner [Thu, 26 May 2016 19:09:33 +0000 (12:09 -0700)]
i965: Simplify foreach_inst_in_block_safe() macro.
We know what the end looks like without examining .tail: it's NULL. It's
always NULL.
Andres Gomez [Wed, 29 Jun 2016 13:02:27 +0000 (16:02 +0300)]
Revert "i965: get PrimitiveMode from the program rather than the shader struct"
This reverts commit
644e015f0b9236e955d679cac4bcc7a1523fc475.
PrimitiveMode from the program doesn't always hold a valid value that
is neither of GL_TRIANGLES, GL_QUADS nor GL_ISOLINES when reaching
this code. This caused regressions in the following CTS tests:
GL44-CTS.stencil_texturing.functional
GL44-CTS.shading_language_420pack.binding_images
GL44-CTS.shading_language_420pack.binding_samplers
GL44-CTS.shading_language_420pack.binding_uniform_single_block
GL44-CTS.shading_language_420pack.implicit_conversions
GL44-CTS.shading_language_420pack.initializer_list
GL44-CTS.shading_language_420pack.length_of_vector_and_matrix
GL44-CTS.shading_language_420pack.line_continuation
Hence, we rather take it from the linked shader.
Reviewed-by: Timothy Arceri <timothy.arceri@collabora.com>
Signed-off-by: Andres Gomez <agomez@igalia.com>
Timothy Arceri [Thu, 30 Jun 2016 04:44:59 +0000 (14:44 +1000)]
glsl/mesa: move duplicate shader fields into new struct gl_shader_info
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Timothy Arceri [Mon, 27 Jun 2016 23:41:59 +0000 (09:41 +1000)]
glsl/main: remove unused params and make function static
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Timothy Arceri [Mon, 27 Jun 2016 21:52:46 +0000 (07:52 +1000)]
glsl: simplify link_uniform_blocks()
There is only ever one shader so simplify the input params.
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Timothy Arceri [Thu, 30 Jun 2016 04:55:40 +0000 (14:55 +1000)]
glsl/mesa: split gl_shader in two
There are two distinctly different uses of this struct. The first
is to store GL shader objects. The second is to store information
about a shader stage thats been linked.
The two uses actually share few fields and there is clearly confusion
about their use. For example the linked shaders map one to one with
a program so can simply be destroyed along with the program. However
previously we were calling reference counting on the linked shaders.
We were also creating linked shaders with a name even though it
is always 0 and called the driver version of the _mesa_new_shader()
function unnecessarily for GL shader objects.
Acked-by: Iago Toral Quiroga <itoral@igalia.com>
Timothy Arceri [Thu, 30 Jun 2016 04:54:22 +0000 (14:54 +1000)]
mesa: don't print name in _mesa_append_uniforms_to_file()
This is only used to print linked shaders which always have a name of 0
so this was pointless.
Acked-by: Iago Toral Quiroga <itoral@igalia.com>
Timothy Arceri [Thu, 30 Jun 2016 04:52:21 +0000 (14:52 +1000)]
mesa: remove unreachable code from _mesa_write_shader_to_file()
_mesa_write_shader_to_file() is only used to print gl shader objects
so Program should never be set as it only gets set for linked shaders.
Acked-by: Iago Toral Quiroga <itoral@igalia.com>
Timothy Arceri [Mon, 27 Jun 2016 06:50:00 +0000 (16:50 +1000)]
glsl: pass symbols to find_matching_signature() rather than shader
This will allow us to later split gl_shader into two structs.
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Timothy Arceri [Mon, 27 Jun 2016 06:25:00 +0000 (16:25 +1000)]
glsl: pass symbols rather than shader to _mesa_get_main_function_signature()
This will allow us to split gl_shader into two different structs, one for
shader objects and one for linked shaders.
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Timothy Arceri [Mon, 27 Jun 2016 06:10:35 +0000 (16:10 +1000)]
mesa: don't use drivers NewShader function when creating shader objects
The drivers function only needs to be used when creating a struct for
linked shaders.
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Timothy Arceri [Mon, 27 Jun 2016 05:38:51 +0000 (15:38 +1000)]
glsl: make cross_validate_globals() more generic
Rather than passing in gl_shader we now pass in the IR. This will
allow us to later split gl_shader into two structs. One for use
as a linked per stage shader struct and one for use as a GL shader
object.
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Ian Romanick [Fri, 17 Jun 2016 02:51:15 +0000 (19:51 -0700)]
mapi: Export all GLES 3.1 functions in libGLESv2.so
Khronos recommends that the GLES 3.1 library also be called libGLESv2.
It also requires that functions be statically linkable from that
library.
NOTE: Mesa has supported the EGL_KHR_get_all_proc_addresses extension
since at least Mesa 10.5, so applications targeting Linux should use
eglGetProcAddress to avoid problems running binaries on systems with
older, non-GLES 3.1 libGLESv2 libraries.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Cc: "11.2 12.0" <mesa-stable@lists.freedesktop.org>
Cc: Mike Gorchak <mike.gorchak.qnx@gmail.com>
Reported-by: Mike Gorchak <mike.gorchak.qnx@gmail.com>
Acked-by: Chad Versace <chad.versace@intel.com>
Chad Versace [Mon, 27 Jun 2016 18:33:36 +0000 (11:33 -0700)]
i965: Use drmIoctl for DRM_I915_GETPARAM (v2)
Stop using drmCommandWriteRead for such a simple ioctl.
v2: Handle errno correctly. [ickle]
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
sonjiang [Tue, 28 Jun 2016 15:23:41 +0000 (11:23 -0400)]
radeon/uvd: fix a h265 context size bug
Signed-off-by: sonjiang <sonny.jiang@amd.com>
Cc: "12.0" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
sonjiang [Mon, 27 Jun 2016 21:19:01 +0000 (17:19 -0400)]
radeon/uvd: seperate uvd context buffer from DPB
Signed-off-by: sonjiang <sonny.jiang@amd.com>
Cc: "12.0" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
sonjiang [Wed, 29 Jun 2016 15:24:36 +0000 (11:24 -0400)]
radeon uvd add uvd fw version for amdgpu
Signed-off-by: sonjiang <sonny.jiang@amd.com>
Cc: "12.0" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Samuel Pitoiset [Wed, 29 Jun 2016 13:35:44 +0000 (15:35 +0200)]
nv50/ir: print EMIT subops in debug mode
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Samuel Pitoiset [Wed, 29 Jun 2016 13:34:35 +0000 (15:34 +0200)]
nv50/ir: print RSQ/RCP subops in debug mode
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Samuel Pitoiset [Wed, 29 Jun 2016 13:25:16 +0000 (15:25 +0200)]
nv50/ir: print PIXLD subops in debug mode
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Samuel Pitoiset [Wed, 29 Jun 2016 13:16:35 +0000 (15:16 +0200)]
nv50/ir: print SHFL subops in debug mode
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Rodrigo Vivi [Thu, 23 Jun 2016 21:38:18 +0000 (14:38 -0700)]
i965: Removing PCI IDs that are no longer listed as Kabylake.
This is unusual. Usually IDs listed on early stages of platform
definition are kept there as reserved for later use.
However these IDs here are not listed anymore in any of steppings
and devices IDs tables for Kabylake on configurations overview
section of BSpec.
So it is better removing them before they become used in any
other future platform.
Reviewed-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Acked-by: Kenneth Graunke <kenneth@whitecape.org>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Rodrigo Vivi [Thu, 23 Jun 2016 21:35:09 +0000 (14:35 -0700)]
i956: Add more Kabylake PCI IDs.
The spec has been updated adding new PCI IDs.
Reviewed-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Acked-by: Kenneth Graunke <kenneth@whitecape.org>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Marek Olšák [Mon, 27 Jun 2016 17:47:27 +0000 (19:47 +0200)]
gallium/radeon: remove zombie textures kept alive by DCC stat gathering
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Mon, 27 Jun 2016 17:46:39 +0000 (19:46 +0200)]
gallium/radeon: don't re-create queries for DCC stat gathering
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Mon, 27 Jun 2016 17:45:30 +0000 (19:45 +0200)]
gallium/radeon: assume X11 DRI3 can use at most 5 back buffers
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Tue, 28 Jun 2016 19:02:40 +0000 (21:02 +0200)]
gallium/radeon: separate DCC starts as disabled (ps_draw_ratio = 0)
DRI3:
- Only slows clears can enable it for the first frame.
- A good PS/draw ratio can enable it for other frames.
DRI2:
- Only slows clears can enable it for a frame.
- Page-flipped color buffers are unref'd at the end of each frame,
so it can't be enabled in any other way.
- Relying on slow clears is sufficient for our synthetic benchmarks.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Mon, 27 Jun 2016 17:44:45 +0000 (19:44 +0200)]
gallium/radeon: R600_DEBUG=nodccfb disables separate DCC
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Tue, 21 Jun 2016 19:17:44 +0000 (21:17 +0200)]
gallium/radeon: add and use r600_texture_reference
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Vedran Miletić <vedran@miletic.net>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Tue, 21 Jun 2016 17:36:14 +0000 (19:36 +0200)]
gallium/radeon: add a HUD query for PS draw ratio stats from separate DCC
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Tue, 21 Jun 2016 16:18:46 +0000 (18:18 +0200)]
gallium/radeon: add a heuristic enabling DCC for scanout surfaces (v2)
DCC for displayable surfaces is allocated in a separate buffer and is
enabled or disabled based on PS invocations from 2 frames ago (to let
queries go idle) and the number of slow clears from the current frame.
At least an equivalent of 5 fullscreen draws or slow clears must be done
to enable DCC. (PS invocations / (width * height) + num_slow_clears >= 5)
Pipeline statistic queries are always active if a color buffer that can
have separate DCC is bound, even if separate DCC is disabled. That means
the window color buffer is always monitored and DCC is enabled only when
the situation is right.
The tracking of per-texture queries in r600_common_context is quite ugly,
but I don't see a better way.
The first fast clear always enables DCC. DCC decompression can disable it.
A later fast clear can enable it again. Enable/disable typically happens
only once per frame.
The impact is expected to be negligible because games usually don't have
a high level of overdraw. DCC usually activates when too much blending
is happening (smoke rendering) or when testing glClear performance and
CMASK isn't supported (Stoney).
v2: rename stuff, add assertions
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Tue, 21 Jun 2016 14:16:15 +0000 (16:16 +0200)]
gallium/radeon: add state setup for a separate DCC buffer
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Tue, 21 Jun 2016 14:09:33 +0000 (16:09 +0200)]
radeonsi: always calculate DCC info even if it's not used immediately
for a later use
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Tue, 21 Jun 2016 13:52:03 +0000 (15:52 +0200)]
radeonsi: unreference framebuffer state with set_framebuffer_state
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Tue, 21 Jun 2016 13:49:25 +0000 (15:49 +0200)]
gallium/radeon: add flag R600_QUERY_HW_FLAG_BEGIN_RESUMES
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Chad Versace [Mon, 27 Jun 2016 18:50:17 +0000 (11:50 -0700)]
i965: Use intel_get_param() more often
Replace some open-coded ioctls with intel_get_param().
This is just a cleanup. No change in behavior.
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Chad Versace [Mon, 27 Jun 2016 18:29:27 +0000 (11:29 -0700)]
i965: Refactor intel_get_param()
Replace the function's __DRIscreen parameter with struct intel_screen.
The callsites feel more natural that way.
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Marek Olšák [Wed, 29 Jun 2016 09:19:58 +0000 (11:19 +0200)]
radeonsi: don't advertise multisample shader images
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Tue, 28 Jun 2016 12:19:04 +0000 (14:19 +0200)]
radeonsi: enable distributed tess on multi-SE parts only
ported from Vulkan
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Tue, 28 Jun 2016 12:11:12 +0000 (14:11 +0200)]
radeonsi: set optimal VGT_HS_OFFCHIP_PARAM
ported from Vulkan
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Tue, 28 Jun 2016 11:15:45 +0000 (13:15 +0200)]
radeonsi: enable CU0 in each SE for LS-HS execution
Offchip-only tessellation allows this.
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Tue, 28 Jun 2016 11:04:07 +0000 (13:04 +0200)]
radeonsi: use conformant line rasterization
AA lines are not completely correct (see TODO), but everything else
should be.
+ 3 linestipple piglits
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Rob Herring [Mon, 13 Jun 2016 18:45:53 +0000 (13:45 -0500)]
Android: add missing u_math.h include path for libmesa_isl
Commit
87d062a94080 ("i965: Fix shared local memory size for Gen9+.")
added u_math.h include which broke the Android build:
In file included from external/mesa3d/src/intel/isl/isl_storage_image.c:25:
In file included from external/mesa3d/src/mesa/drivers/dri/i965/brw_compiler.h:29:
external/mesa3d/src/mesa/main/macros.h:35:10: fatal error: 'util/u_math.h' file not found
^
Add the missing include paths for libmesa_isl.
Signed-off-by: Rob Herring <robh@kernel.org>
Reviewed-by: Kenneth Garunke <kenneth@whitecape.org>
Charmaine Lee [Thu, 23 Jun 2016 16:21:11 +0000 (09:21 -0700)]
svga: force direct map for transfering multiple slices
With commit
fb9fe35, we start using transfer_inline_write
for memcpy of TexSubImage. But SurfaceDMA command does not work
well with texture array. This patch forces direct map when
transfering multiple slices of a texture array.
Fixes piglit regression "texelFetch fs sampler1DArray"
Tested with MTT piglit, glretrace, conform.
Reviewed-by: Sinclair Yeh <syeh@vmware.com>
Brian Paul [Mon, 27 Jun 2016 17:40:00 +0000 (11:40 -0600)]
svga: whitespace, line wrapping fixes in svga_surface.c
Samuel Pitoiset [Mon, 27 Jun 2016 22:59:46 +0000 (00:59 +0200)]
gm107/ir: make sure that flagsDef is set when emitting setcond
Rely on the existence of a second destination when emitting a setcond
flag is dangerous, because this doesn't mean that the flag has been
correctly set. Instead rely on flagsDef like what emitX() does
for flagsSrc.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Cc: <mesa-stable@lists.freedesktop.org>