mesa.git
6 years agoradv: make use of the output usage mask in GS copy shader
Samuel Pitoiset [Thu, 13 Sep 2018 13:57:58 +0000 (15:57 +0200)]
radv: make use of the output usage mask in GS copy shader

This is just for consistency because LLVM can detect and
remove unused loads.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
6 years agoradv: improve a comment in si_emit_set_predication_state()
Samuel Pitoiset [Wed, 12 Sep 2018 21:20:39 +0000 (23:20 +0200)]
radv: improve a comment in si_emit_set_predication_state()

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
6 years agoradv: fix VK_EXT_conditional_rendering visibility
Samuel Pitoiset [Wed, 12 Sep 2018 21:20:38 +0000 (23:20 +0200)]
radv: fix VK_EXT_conditional_rendering visibility

It's actually just the opposite.

This fixes the new Sascha conditionalrender demo.

CC: 18.2 <mesa-stable@lists.freedesktop.org>
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
6 years agoradv: make use of ac_unpack_param() instead of ac_build_bfe()
Samuel Pitoiset [Thu, 13 Sep 2018 14:36:45 +0000 (16:36 +0200)]
radv: make use of ac_unpack_param() instead of ac_build_bfe()

Same code is generated because LLVM ends up by using bfe, but
that seems cleaner to me.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
6 years agonir: add loop unroll support for complex wrapper loops
Timothy Arceri [Wed, 11 Jul 2018 00:53:41 +0000 (10:53 +1000)]
nir: add loop unroll support for complex wrapper loops

In GLSL IR we cheat with switch statements and simply convert them
into loops with a single iteration. This allowed us to make use of
the existing jump instruction handling provided by the loop handing
code, it also allows dead code to be cleaned up once we have
wrapped the code in a loop.

However using loops in this way created previously unrollable loops
which limits further optimisations. Here we provide a way to unroll
loops that end in a break and have multiple other exits.

All shader-db changes are from the dolphin uber shaders. There is a
small amount of HURT shaders but in general the improvements far
exceed the HURT.

shader-db results IVB:

total instructions in shared programs: 10018187 -> 10016468 (-0.02%)
instructions in affected programs: 104080 -> 102361 (-1.65%)
helped: 36
HURT: 15

total cycles in shared programs: 220065064 -> 154529655 (-29.78%)
cycles in affected programs: 126063017 -> 60527608 (-51.99%)
helped: 51
HURT: 0

total loops in shared programs: 2515 -> 2308 (-8.23%)
loops in affected programs: 903 -> 696 (-22.92%)
helped: 51
HURT: 0

total spills in shared programs: 4370 -> 4124 (-5.63%)
spills in affected programs: 1397 -> 1151 (-17.61%)
helped: 9
HURT: 12

total fills in shared programs: 4581 -> 4419 (-3.54%)
fills in affected programs: 2201 -> 2039 (-7.36%)
helped: 9
HURT: 15

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
6 years agonir: propagates if condition evaluation down some alu chains
Timothy Arceri [Mon, 23 Jul 2018 04:38:34 +0000 (14:38 +1000)]
nir: propagates if condition evaluation down some alu chains

v2:
 - only allow nir_op_inot or nir_op_b2i when alu input is 1.
 - use some helpers as suggested by Jason.

v3:
 - evaluate alu op for single input alu ops
 - add helper function to decide if to propagate through alu
 - make use of nir_before_src in another spot

shader-db IVB results:

total instructions in shared programs: 9993483 -> 9993472 (-0.00%)
instructions in affected programs: 1300 -> 1289 (-0.85%)
helped: 11
HURT: 0

total cycles in shared programs: 219476091 -> 219476059 (-0.00%)
cycles in affected programs: 7675 -> 7643 (-0.42%)
helped: 10
HURT: 1

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
6 years agonir: evaluate if condition uses inside the if branches
Timothy Arceri [Wed, 29 Aug 2018 15:24:43 +0000 (10:24 -0500)]
nir: evaluate if condition uses inside the if branches

Since we know what side of the branch we ended up on we can just
replace the use with a constant.

All the spill changes in shader-db are from Dolphin uber shaders,
despite some small regressions the change is clearly positive.

V2: insert new constant after any phis in the
    use->parent_instr->type == nir_instr_type_phi path.

v3:
 - use nir_after_block_before_jump() for inserting const
 - check dominance of phi uses correctly

v4:
 - create some helpers as suggested by Jason.

v5 (Jason Ekstrand):
 - Use LIST_ENTRY to get the phi src

shader-db results IVB:

total instructions in shared programs: 9999201 -> 9993483 (-0.06%)
instructions in affected programs: 163235 -> 157517 (-3.50%)
helped: 132
HURT: 2

total cycles in shared programs: 231670754 -> 219476091 (-5.26%)
cycles in affected programs: 143424120 -> 131229457 (-8.50%)
helped: 115
HURT: 24

total spills in shared programs: 4383 -> 4370 (-0.30%)
spills in affected programs: 1656 -> 1643 (-0.79%)
helped: 9
HURT: 18

total fills in shared programs: 4610 -> 4581 (-0.63%)
fills in affected programs: 374 -> 345 (-7.75%)
helped: 6
HURT: 0

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
6 years agovirgl: adjust strides when mapping temp-resources
Erik Faye-Lund [Wed, 12 Sep 2018 07:48:41 +0000 (09:48 +0200)]
virgl: adjust strides when mapping temp-resources

When we're mapping temp-resources, we clip the resource to the
transfer-box, which means the stride might not be correct any more.

So let's update the stride from the temp-resource, and recompute the
layer-stride.

This fixes crashes when running dEQP with --deqp-gl-config-name=rgba8888d24s8ms4

Signed-off-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Fixes: a8987b88ff1 "virgl: add driver for virtio-gpu 3D (v2)"
Reviewed-by: Dave Airlie <airlied@redhat.com>
6 years agonvir: Always split 64-bit IMAD/IMUL operations
Pierre Moreau [Mon, 4 Dec 2017 23:51:04 +0000 (00:51 +0100)]
nvir: Always split 64-bit IMAD/IMUL operations

Those operations do not map to actual hardware instructions, therefore
those should always be lowered to 32-bit instructions.

Fixes: 009c54aa7af "nv50/ir: Split 64-bit integer MAD/MUL operations"
Signed-off-by: Pierre Moreau <pierre.morrow@free.fr>
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Signed-off-by: Karol Herbst <kherbst@redhat.com>
6 years agost/vdpau: Use output buffer as back buffer with 24-bit color only
Leo Liu [Fri, 7 Sep 2018 13:26:08 +0000 (09:26 -0400)]
st/vdpau: Use output buffer as back buffer with 24-bit color only

Using output buffer with 8 bits video RGB as back buffer
certainly is not working for 30 bits color depth visual.

Signed-off-by: Leo Liu <leo.liu@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
6 years agovl/dri: add color depth to vl winsys
Leo Liu [Mon, 10 Sep 2018 20:02:29 +0000 (16:02 -0400)]
vl/dri: add color depth to vl winsys

For VDPAU use later

Signed-off-by: Leo Liu <leo.liu@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
6 years agovl/dri3: add support for 10 bits format
Leo Liu [Mon, 10 Sep 2018 19:33:28 +0000 (15:33 -0400)]
vl/dri3: add support for 10 bits format

Signed-off-by: Leo Liu <leo.liu@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
6 years agovl/dri: add 10 bits format supports
Leo Liu [Fri, 7 Sep 2018 19:17:31 +0000 (15:17 -0400)]
vl/dri: add 10 bits format supports

v2: Tell B10G10R10X2 and R10G10B10X2 formats for different HW.

Signed-off-by: Leo Liu <leo.liu@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
6 years agoegl/android: Declare droid_load_driver() static
Kristian H. Kristensen [Wed, 12 Sep 2018 18:38:22 +0000 (11:38 -0700)]
egl/android: Declare droid_load_driver() static

Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Signed-off-by: Kristian H. Kristensen <hoegsberg@chromium.org>
6 years agoradv: fix function names for VK_EXT_conditional_rendering
Samuel Pitoiset [Wed, 12 Sep 2018 20:54:27 +0000 (22:54 +0200)]
radv: fix function names for VK_EXT_conditional_rendering

Otherwise they are not exported.

CC: 18.2 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Reviewed-by: Dave Airlie <airlied@redhat.com
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
6 years agoanv: Silence a couple compiler warnings
Jason Ekstrand [Wed, 12 Sep 2018 22:02:57 +0000 (17:02 -0500)]
anv: Silence a couple compiler warnings

[63/93] Compiling C object 'src/intel/vulkan/...intel@vulkan@@anv_common@sta/anv_device.c.o'.
../src/intel/vulkan/anv_device.c:685:30: warning: passing 'const char *' to parameter of type 'void *' discards qualifiers [-Wincompatible-pointer-types-discards-qualifiers]
   vk_free(&instance->alloc, instance->app_info.app_name);
                             ^~~~~~~~~~~~~~~~~~~~~~~~~~~
../src/vulkan/util/vk_alloc.h:62:51: note: passing argument to parameter 'data' here
vk_free(const VkAllocationCallbacks *alloc, void *data)
                                                  ^
../src/intel/vulkan/anv_device.c:686:30: warning: passing 'const char *' to parameter of type 'void *' discards qualifiers [-Wincompatible-pointer-types-discards-qualifiers]
   vk_free(&instance->alloc, instance->app_info.engine_name);
                             ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../src/vulkan/util/vk_alloc.h:62:51: note: passing argument to parameter 'data' here
vk_free(const VkAllocationCallbacks *alloc, void *data)
                                                  ^
[65/93] Compiling C object 'src/intel/vulkan/...ommon@sta/anv_nir_apply_pipeline_layout.c.o'.
../src/intel/vulkan/anv_nir_apply_pipeline_layout.c:519:13: warning: unused variable 'image_uniform' [-Wunused-variable]
   unsigned image_uniform;

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
6 years agoloader/dri3: Don't wait for fence of old buffer when re-allocating it
Michel Dänzer [Tue, 4 Sep 2018 16:18:57 +0000 (18:18 +0200)]
loader/dri3: Don't wait for fence of old buffer when re-allocating it

We only need to wait for the fence before drawing to a buffer, not
before reading from it.

This might avoid hangs when re-allocating the fake front buffer, similar
to the previous change. But I haven't seen any evidence that this was
actually happening in practice.

Tested-by: Olivier Fourdan <ofourdan@redhat.com>
6 years agoloader/dri3: Only wait for back buffer fences in dri3_get_buffer
Michel Dänzer [Tue, 4 Sep 2018 10:18:19 +0000 (12:18 +0200)]
loader/dri3: Only wait for back buffer fences in dri3_get_buffer

We don't need to wait before drawing to the fake front buffer, as front
buffer rendering by definition is allowed to produce artifacts.

Fixes hangs in some cases when re-using the fake front buffer, due to it
still being busy (i.e. in use for presentation).

Cc: mesa-stable@lists.freedesktop.org
Bugzilla: https://bugs.freedesktop.org/106404
Bugzilla: https://bugs.freedesktop.org/107757
Tested-by: Olivier Fourdan <ofourdan@redhat.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
6 years agoglsl/linker: Check the invariance of built-in special variables
Vadym Shovkoplias [Fri, 7 Sep 2018 12:14:52 +0000 (15:14 +0300)]
glsl/linker: Check the invariance of built-in special variables

From Section 4.6.4 (Invariance and Linkage) of the GLSL ES 1.0 specification

    "The invariance of varyings that are declared in both the vertex and
     fragment shaders must match. For the built-in special variables,
     gl_FragCoord can only be declared invariant if and only if
     gl_Position is declared invariant. Similarly gl_PointCoord can only
     be declared invariant if and only if gl_PointSize is declared
     invariant. It is an error to declare gl_FrontFacing as invariant.
     The invariance of gl_FrontFacing is the same as the invariance of
     gl_Position."

Fixes:
    * glsl-pcoord-invariant.shader_test
    * glsl-fcoord-invariant.shader_test
    * glsl-fface-invariant.shader_test

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=107734
Signed-off-by: Vadym Shovkoplias <vadym.shovkoplias@globallogic.com>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
6 years agointel/tools: fix initial position of window in aubinator viewer
Tapani Pälli [Tue, 11 Sep 2018 07:32:32 +0000 (10:32 +0300)]
intel/tools: fix initial position of window in aubinator viewer

Currently position is set before widgets are sized by gtk and
calculation can get wrong results where window is positioned
offscreen. Patch fixes this by setting aubfile window position
as 0,0 only when size_allocate has been called to the widget.

Now window is always positioned to 0,0 if imgui.ini is missing.

Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
6 years agowinsys/virgl: avoid unintended behavior
Erik Faye-Lund [Thu, 6 Sep 2018 10:21:53 +0000 (11:21 +0100)]
winsys/virgl: avoid unintended behavior

If we end up never taking the loop that writes ret, we can end up with
an uninitialized value, and if we're *really* unlucky, that value can
be -1, causing us to go down an error-path instead of a success path.

This was obviously not intended, so let's just initialize this to zero.

Noticed by Valgrind:

Conditional jump or move depends on uninitialised value(s)
   at 0xBA640A0: virgl_drm_winsys_resource_cache_create (virgl_drm_winsys.c:348)
   by 0xBA62FCF: virgl_buffer_create (virgl_buffer.c:170)
   by 0xBA605AC: virgl_resource_create (virgl_resource.c:60)
   by 0xBCF816F: bufferobj_data (st_cb_bufferobjects.c:344)
   by 0xBCF816F: st_bufferobj_data (st_cb_bufferobjects.c:390)
   by 0xBB7E836: vbo_use_buffer_objects (vbo_exec_api.c:1136)
   by 0xBCFCC6E: st_create_context_priv (st_context.c:414)
   by 0xBCFD3CD: st_create_context (st_context.c:590)
   by 0xBBB30CA: st_api_create_context (st_manager.c:896)
   by 0xB981E76: dri_create_context (dri_context.c:155)
   by 0xB97BDCE: driCreateContextAttribs (dri_util.c:473)
   by 0x5288331: dri3_create_context_attribs (dri3_glx.c:309)
   by 0x5264D64: glXCreateContextAttribsARB (create_context.c:78)

Fixes: a8987b88ff1 ("virgl: add driver for virtio-gpu 3D (v2)")
Signed-off-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
6 years agotravis: use python3.5 for meson
Juan A. Suarez Romero [Mon, 10 Sep 2018 12:15:33 +0000 (14:15 +0200)]
travis: use python3.5 for meson

Newer Meson versions require python >=3.5. But in Trusty default python3
version is 3.4.x.

Install python3.5 and makes it the default version for Meson using
update-alternatives method.

CC: Jan Vesely <jano.vesely@gmail.com>
CC: Andres Gomez <agomez@igalia.com>
CC: Emil Velikov <emil.l.velikov@gmail.com>
CC: Jon Turney <jon.turney@dronecode.org.uk>
CC: Eric Engestrom <eric.engestrom@intel.com>
CC: Dylan Baker <dylan@pnwbakers.com>
Fixes: 3824c8e7cda97c3bf856 "meson: disable asserts by default on release builds"
Reviewed-by: Eric Engestrom <eric.engestrom@intel.com>
Reviewed-by: Andres Gomez <agomez@igalia.com>
6 years agoradv: adjust ESGS ring buffer size computation on VI+
Samuel Pitoiset [Mon, 10 Sep 2018 16:14:41 +0000 (18:14 +0200)]
radv: adjust ESGS ring buffer size computation on VI+

Noticed while working in this area. Ported from RadeonSI.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
6 years agomesa/texture: Also check for LA texture when querying intensity component size
Gert Wollny [Mon, 10 Sep 2018 10:39:44 +0000 (12:39 +0200)]
mesa/texture: Also check for LA texture when querying intensity component size

Gallium may pick L16A16_FLOAT to represent GL_INTENSITY16F if no intensity
format is provided by the driver. However, when calling

   glGetTexLevelParameteriv(..., GL_TEXTURE_INTENSITY_SIZE, ...)

mesa will return a zero size because the actually used format has no
intensity channel and as a fallback only the sizes of the red/green
channels are checked.

Also checking for LA sizes in the allocated texture resolves this problem.

v2: Only check alpha channel size and return it (Marek)
    L and A size are always the same in this case.

Fixes (on virgl):
  ext_framebuffer_multisample-fast-clear GL_ARB_texture_float *

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=107832

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
6 years agonv50,nvc0: warn on not-explicitly-handled caps
Ilia Mirkin [Fri, 7 Sep 2018 02:33:13 +0000 (22:33 -0400)]
nv50,nvc0: warn on not-explicitly-handled caps

Not handling caps explicitly means that we're likely getting incorrect
values -- these need to be reviewed and set appropriately.

While we're at it, add in some missing caps, and set all the subpixel
stuff to 8 as that seems to be what the blob reports.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
6 years agomesa: remove duplicate dispatch sanity tests
Timothy Arceri [Fri, 7 Sep 2018 22:21:21 +0000 (08:21 +1000)]
mesa: remove duplicate dispatch sanity tests

This removes duplicate tests from gl_core_functions_possible
that are already covered by common_desktop_functions_possible.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
6 years agomesa: tidy up init_matrix_stack()
Timothy Arceri [Mon, 10 Sep 2018 10:41:27 +0000 (20:41 +1000)]
mesa: tidy up init_matrix_stack()

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
6 years agoradeon: fix ColorMask
Christopher Egert [Sat, 8 Sep 2018 00:16:48 +0000 (02:16 +0200)]
radeon: fix ColorMask

Since commit af3685d14936844f79e6f372b4b258e29375f21b various OpenGL applications regressed
on the classic mesa radeon driver.

Signed-off-by: Christopher Egert <cme3000@gmail.com>
CC: 18.1 18.2 <mesa-stable@lists.freedesktop.org>
Signed-off-by: Marek Olšák <marek.olsak@amd.com>
6 years agogallium: Correctly handle no config context creation
Elie Tournier [Thu, 6 Sep 2018 09:46:32 +0000 (10:46 +0100)]
gallium: Correctly handle no config context creation

This patch fixes the following Piglit test:
spec@egl_mesa_configless_context@basic
It also fixes few test in a virgl guest.

v2: Evaluate the value of no_config (Ilia)

Suggested-by: Emil Velikov <emil.velikov@collabora.com>
Signed-off-by: Elie Tournier <elie.tournier@collabora.com>
Signed-off-by: Marek Olšák <marek.olsak@amd.com>
6 years agoradv: Support v3 of VK_EXT_vertex_attribute_divisor.
Bas Nieuwenhuizen [Mon, 10 Sep 2018 18:34:00 +0000 (20:34 +0200)]
radv: Support v3 of VK_EXT_vertex_attribute_divisor.

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
CC: 18.2 <mesa-stable@lists.freedesktop.org>
6 years agoradeonsi/nir: port some bindless and sampler code from TGSI
Marek Olšák [Wed, 29 Aug 2018 06:45:49 +0000 (02:45 -0400)]
radeonsi/nir: port some bindless and sampler code from TGSI

These might be all missing changes for bindless textures.

Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
6 years agoradeonsi: adjust and simplify max_alloc_size determination
Marek Olšák [Tue, 28 Aug 2018 23:13:18 +0000 (19:13 -0400)]
radeonsi: adjust and simplify max_alloc_size determination

Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
6 years agoradeonsi: split si_copy_buffer
Marek Olšák [Fri, 24 Aug 2018 04:28:16 +0000 (00:28 -0400)]
radeonsi: split si_copy_buffer

compute and SDMA will be added into it.

Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
6 years agoradeonsi: don't call VBO prefetch with size=0
Marek Olšák [Tue, 28 Aug 2018 21:45:00 +0000 (17:45 -0400)]
radeonsi: don't call VBO prefetch with size=0

for the next commit.

Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
6 years agoradeonsi: merge SI and CI dma_clear_buffer and remove the callback
Marek Olšák [Fri, 24 Aug 2018 04:04:11 +0000 (00:04 -0400)]
radeonsi: merge SI and CI dma_clear_buffer and remove the callback

also use assertions for the requirements that offset and size are a multiple
of 4.

Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
6 years agoradeonsi: fix GPU hangs with bindless textures and LLVM 7.0
Marek Olšák [Wed, 29 Aug 2018 05:34:46 +0000 (01:34 -0400)]
radeonsi: fix GPU hangs with bindless textures and LLVM 7.0

Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
6 years agoac: remove deprecated use of LLVMInt1Type()
Marek Olšák [Wed, 29 Aug 2018 04:17:26 +0000 (00:17 -0400)]
ac: remove deprecated use of LLVMInt1Type()

Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
6 years agoac: use iN_0/1 constants
Marek Olšák [Wed, 29 Aug 2018 04:15:16 +0000 (00:15 -0400)]
ac: use iN_0/1 constants

Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
6 years agoac: add radeon_info::num_good_cu_per_sh
Marek Olšák [Sat, 25 Aug 2018 02:57:17 +0000 (22:57 -0400)]
ac: add radeon_info::num_good_cu_per_sh

Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
6 years agoac: revert new LLVM 7.0 behavior for fdiv
Marek Olšák [Sat, 25 Aug 2018 03:00:37 +0000 (23:00 -0400)]
ac: revert new LLVM 7.0 behavior for fdiv

Cc: 18.2 <mesa-stable@lists.freedesktop.org>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
6 years agoradeonsi: fix printing a BO list into ddebug reports
Marek Olšák [Tue, 28 Aug 2018 18:59:25 +0000 (14:59 -0400)]
radeonsi: fix printing a BO list into ddebug reports

important for debugging

Cc: 18.1 18.2 <mesa-stable@lists.freedesktop.org>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
6 years agor600: fix HTILE for NPOT textures with mipmapping
Marek Olšák [Tue, 28 Aug 2018 19:49:31 +0000 (15:49 -0400)]
r600: fix HTILE for NPOT textures with mipmapping

Cc: 18.1 18.2 <mesa-stable@lists.freedesktop.org>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
6 years agowinsys/radeon: fix CMASK fast clear for NPOT textures with mipmapping on SI/CI
Marek Olšák [Tue, 28 Aug 2018 18:39:09 +0000 (14:39 -0400)]
winsys/radeon: fix CMASK fast clear for NPOT textures with mipmapping on SI/CI

Cc: 18.2 <mesa-stable@lists.freedesktop.org>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
6 years agoradeonsi: fix HTILE for NPOT textures with mipmapping on SI/CI
Marek Olšák [Tue, 28 Aug 2018 19:49:31 +0000 (15:49 -0400)]
radeonsi: fix HTILE for NPOT textures with mipmapping on SI/CI

VI uses addrlib so it's unaffected.

Cc: 18.1 18.2 <mesa-stable@lists.freedesktop.org>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
6 years agodocs: document new features/extensions in driver for WS 15 / Fusion 11
Brian Paul [Mon, 10 Sep 2018 17:29:12 +0000 (11:29 -0600)]
docs: document new features/extensions in driver for WS 15 / Fusion 11

Reviewed-by: Charmaine Lee <charmainel@vmware.com>
6 years agosvga: assorted fixes/changes in svga_pipe_blit.c
Brian Paul [Thu, 6 Sep 2018 21:06:22 +0000 (15:06 -0600)]
svga: assorted fixes/changes in svga_pipe_blit.c

To align the code with VMware's in-house copy.

Signed-off-by: Brian Paul <brianp@vmware.com>
6 years agosvga: set buffer bind_flags in svga_buffer_add_host_surface()
Brian Paul [Thu, 6 Sep 2018 20:56:47 +0000 (14:56 -0600)]
svga: set buffer bind_flags in svga_buffer_add_host_surface()

To match the in-house VMware code.

Signed-off-by: Brian Paul <brianp@vmware.com>
6 years agosvga: add format conversion for legacy formats
Charmaine Lee [Sun, 29 Apr 2018 22:13:05 +0000 (15:13 -0700)]
svga: add format conversion for legacy formats

This patch extends the format_conversion table to support
different view formats on texture buffer.
For legacy image formats such as INTENSITY, LUMINANCE, LUMINANCE_ALPHA,
special swizzle masks will be used on the red or RG channels.

This fixes piglit test arb_texture_buffer_object-formats fs|vs arb

Reviewed-by: Brian Paul <brianp@vmware.com>
6 years agosvga: remove obsolete code to reemit gs binding
Charmaine Lee [Thu, 15 Feb 2018 20:11:43 +0000 (12:11 -0800)]
svga: remove obsolete code to reemit gs binding

The svga_reemit_gs_bindings function is no longer needed. Remove it.

Reviewed-by: Neha Bhende <bhenden@vmware.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
6 years agosvga: move variant->fs_shadow_compare_units assignment
Brian Paul [Wed, 27 Dec 2017 18:36:03 +0000 (11:36 -0700)]
svga: move variant->fs_shadow_compare_units assignment

Fixes a crash since the variant object isn't allocated until later
in the function.  Not sure how this got through.

Reviewed-by: Charmaine Lee <charmainel@vmware.com>
6 years agosvga: fix resource checking in is_blending_enabled()
Charmaine Lee [Mon, 18 Dec 2017 17:41:54 +0000 (09:41 -0800)]
svga: fix resource checking in is_blending_enabled()

This patch makes sure a valid color buffer is bound before
checking its resource. This fixes Unigine Valley running in SM41 device.

Reviewed-by: Brian Paul <brianp@vmware.com>
6 years agosvga: Use texture_copy_region instead of texture_copy_handle for multisampling
Neha Bhende [Wed, 13 Dec 2017 23:02:13 +0000 (15:02 -0800)]
svga: Use texture_copy_region instead of texture_copy_handle for multisampling

This fixes some of tests cases in arb_copy_image-formats and also fixes
SurfaceCopy related errors in vmware.log when multi sampled surfaces are
used.

Tested with piglit, glretrace on windows and linux VM.

v2: As per Brian's comment

Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
6 years agosvga: add missing devcap check for texture array support
Charmaine Lee [Fri, 8 Dec 2017 05:29:49 +0000 (21:29 -0800)]
svga: add missing devcap check for texture array support

The patch checks DXFMT_ARRAY devcap for texture array support.

Tested with MTT-piglit. No regressions.

Reviewed-by: Brian Paul <brianp@vmware.com>
6 years agosvga: no need to check MULTISAMPLE devcap for view format
Charmaine Lee [Wed, 6 Dec 2017 23:33:45 +0000 (15:33 -0800)]
svga: no need to check MULTISAMPLE devcap for view format

According to the current SVGA contract, any view format can be
used on the underlying resource that is multisample. So there
is no need to check the MULTISAMPLE devcap for the view format.

Fixes black rendering issue with Tropics running with 4xMSAA.

Reviewed-by: Brian Paul <brianp@vmware.com>
6 years agosvga: sync devcap name changes in svga3d_devcaps.h
Charmaine Lee [Thu, 7 Dec 2017 17:52:47 +0000 (09:52 -0800)]
svga: sync devcap name changes in svga3d_devcaps.h

Reviewed-by: Brian Paul <brianp@vmware.com>
6 years agosvga: explicit set DXFMT_SHADER_SAMPLE for DS format for pre-SM41 device
Charmaine Lee [Thu, 7 Dec 2017 17:35:00 +0000 (09:35 -0800)]
svga: explicit set DXFMT_SHADER_SAMPLE for DS format for pre-SM41 device

Explicit set the DXFMT_SHADER_SAMPLE bit for depth stencil formats
for pre-SM41 device only. This bit is now set by the SM41 device.

Reviewed-by: Brian Paul <brianp@vmware.com>
6 years agosvga: remove unused variable
Charmaine Lee [Thu, 7 Dec 2017 20:04:17 +0000 (12:04 -0800)]
svga: remove unused variable

Trivial.

6 years agosvga: draw round points when msaa is enabled
Brian Paul [Fri, 8 Dec 2017 04:12:10 +0000 (21:12 -0700)]
svga: draw round points when msaa is enabled

See comments for details.  This allows the piglit
ext_framebuffer_multisample-point-smooth test to pass.

Also, test the pipe_rasterizer_state::point_quad_rasterization field
to see if sprite point rasterization is needed because it's possible
for no sprite_coord_enable bits to be set when drawing sprites.

Finally, remove old, stale comments.

Reviewed-by: Charmaine Lee <charmainel@vmware.com>
6 years agosvga: check number of samples before emitting MSAA decls/opcodes
Brian Paul [Wed, 6 Dec 2017 03:41:55 +0000 (20:41 -0700)]
svga: check number of samples before emitting MSAA decls/opcodes

If real MSAA is not available, we only support 1 sample/pixel.  In that
case, we must not declare MSAA resources or emit MSAA opcodes.  Do that
by checking the sample count.

Fixes several piglit MSAA tests, such as
arb_texture_multisample-sample-depth (when the hard-coded sample count
of 4 is fixed in that test).

Reviewed-by: Charmaine Lee <charmainel@vmware.com>
6 years agosvga: remove obsolete comment on format_cap_table[]
Brian Paul [Tue, 5 Dec 2017 00:09:24 +0000 (17:09 -0700)]
svga: remove obsolete comment on format_cap_table[]

We removed the special cases referred to in this comment in the commit
"svga: add a separate function to get dx format capabilities from
vgpu10 device".

Reviewed-by: Charmaine Lee <charmainel@vmware.com>
6 years agosvga: allow TGSI_TEXTURE_CUBE_ARRAY in emit_tg4()
Brian Paul [Fri, 17 Nov 2017 18:18:38 +0000 (11:18 -0700)]
svga: allow TGSI_TEXTURE_CUBE_ARRAY in emit_tg4()

Technically, SM4.1 doesn't support cube map arrays, but our backend
renderers actually do.  This allows the Piglit textureGather cube
map array tests to pass.

Tested with GLrenderer, DX11renderer and SWrenderer.

Reviewed-by: Neha Bhende <bhenden@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
6 years agosvga: no dma on multisample surface
Charmaine Lee [Mon, 4 Dec 2017 18:03:51 +0000 (10:03 -0800)]
svga: no dma on multisample surface

Force direct map on multisample surface.

Fixes SVGA Driver Errors running multisample piglit tests on Linux VM

v2: use texture for the check.

Reviewed-by: Neha Bhende <bhenden@vmware.com>
6 years agosvga: src surface for IntraSurfaceCopy cannot be multisample
Charmaine Lee [Sat, 2 Dec 2017 04:59:42 +0000 (20:59 -0800)]
svga: src surface for IntraSurfaceCopy cannot be multisample

Fixes SVGA Driver Errors with piglit test arb_copy_image-targets

Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Neha Bhende <bhenden@vmware.com>
6 years agosvga: fix missing format multisample devcap check
Charmaine Lee [Fri, 1 Dec 2017 20:45:30 +0000 (12:45 -0800)]
svga: fix missing format multisample devcap check

In commit e4048f6cd1, svga_is_dx_format_supported() is supposed to
also check the SVGA3D_DXFMT_MULTISAMPLE bit for multisample
support of a format. Somehow that code is not included in that commit.
This patch fixes it.

Fixes piglit test spec@ext_framebuffer_multisample@formats all_samples.

Reviewed-by: Brian Paul <brianp@vmware.com>
6 years agosvga: fix incorrect multisample support in VGPU9 device
Charmaine Lee [Thu, 30 Nov 2017 16:54:44 +0000 (08:54 -0800)]
svga: fix incorrect multisample support in VGPU9 device

Commit e4048f6cd1 unintentionally allows multisample support for VGPU9 device.
This patch fixes this regression.

Reviewed-by: Neha Bhende <bhenden@vmware.com>
6 years agosvga: fix the missing devcap for SVGA3D_BC3_UNORM_SRGB
Charmaine Lee [Wed, 29 Nov 2017 21:27:58 +0000 (13:27 -0800)]
svga: fix the missing devcap for SVGA3D_BC3_UNORM_SRGB

Set the devcap to SVGA3D_DEVCAP_DXFMT_BC3_UNORM_SRGB

Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Neha Bhende <bhenden@vmware.com>
6 years agosvga: add a separate function to get dx format capabilities from vgpu10 device
Charmaine Lee [Wed, 29 Nov 2017 21:25:30 +0000 (13:25 -0800)]
svga: add a separate function to get dx format capabilities from vgpu10 device

Currently we have one function to get format capabailities and
we convert DX10 devcaps back to DX9. This can be confusing.
Going forward we will have a separate function for dealing with dx formats.

This patch also fixes the depth stencil devcap. Instead of hardcoding
the capabilities for the depth stencil formats, we will inquire the
device for the capabilities. Note: we will still need to explicity set
the SVGA3D_DXFMT_SHADER_SAMPLE bit for SVGA3D_R32_FLOAT_X8X24 and
SVGA3D_R24_UNORM_X8 since this bit is not advertised but supported
by the device.

v2: reapply the patch after svga_is_format_supported is moved to svga_format.c

Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Neha Bhende <bhenden@vmware.com>
6 years agosvga: assign a separate function for is_format_supported() for vgpu10 device
Charmaine Lee [Wed, 29 Nov 2017 21:09:26 +0000 (13:09 -0800)]
svga: assign a separate function for is_format_supported() for vgpu10 device

This patch adds a new function svga_is_dx_format_supported() to check
for format support in a VGPU10 device.

v2: reapply the patch after svga_is_format_supported is moved to svga_format.c

Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Neha Bhende <bhenden@vmware.com>
6 years agosvga: add some devcap debugging code
Brian Paul [Wed, 29 Nov 2017 18:28:10 +0000 (11:28 -0700)]
svga: add some devcap debugging code

Reviewed-by: Charmaine Lee <charmainel@vmware.com>
6 years agosvga: fix depth and coverage mask output declaration
Charmaine Lee [Mon, 27 Nov 2017 21:24:26 +0000 (13:24 -0800)]
svga: fix depth and coverage mask output declaration

Set the component mask to zero for both registers.

Reviewed-by: Brian Paul <brianp@vmware.com>
6 years agosvga: add sample positions for 2 samples
Charmaine Lee [Fri, 17 Nov 2017 03:46:35 +0000 (19:46 -0800)]
svga: add sample positions for 2 samples

Fixes piglit tests spec@arb_sample_shading@builtin-gl-sample-position 2
                   spec@arb_texture_multisample@fb-completeness@2

Reviewed-by: Brian Paul <brianp@vmware.com>
6 years agosvga: check sample count devcaps
Charmaine Lee [Thu, 16 Nov 2017 06:21:22 +0000 (22:21 -0800)]
svga: check sample count devcaps

Check sample count devcaps from the svga device to determine the
supported sample counts.

Reviewed-by: Brian Paul <brianp@vmware.com>
6 years agosvga: fix 1-element cube map array issue
Brian Paul [Tue, 14 Nov 2017 16:36:23 +0000 (09:36 -0700)]
svga: fix 1-element cube map array issue

As with 1D and 2D array textures, if there's only one array element
(one cubemap in this case) we have to issue different shader code.

This fixes a number of Piglit cubemap array tests.

Reviewed-by: Charmaine Lee <charmainel@vmware.com>
6 years agosvga: simplify array test in svga_init_shader_key_common()
Brian Paul [Tue, 14 Nov 2017 16:34:29 +0000 (09:34 -0700)]
svga: simplify array test in svga_init_shader_key_common()

And squash commit a patch to silence a compiler warning (add
default case to the switch statement).

Reviewed-by: Charmaine Lee <charmainel@vmware.com>
6 years agowinsys/drm: check for CAPS2/SM41 support if VGPU10 is enabled
Charmaine Lee [Mon, 6 Nov 2017 22:13:47 +0000 (14:13 -0800)]
winsys/drm: check for CAPS2/SM41 support if VGPU10 is enabled

No need to check for HW_CAPS2 or SM4_1 support if VGPU10 is not
enabled or is explicitly disabled via the environment variable
SVGA_VGPU10.

Reviewed-by: Deepak Rawat <drawat@vmware.com>
6 years agowinsys/drm: Add support for quality level in surface ioctl
Deepak Rawat [Thu, 2 Nov 2017 00:40:50 +0000 (17:40 -0700)]
winsys/drm: Add support for quality level in surface ioctl

A new argument "quality level" is added in surface define v3 which
represets precision settings for surface. This commit add support
for quality level in DRM_VMW_GB_SURFACE_CREATE_EXT and
DRM_VMW_GB_SURFACE_REF_EXT.

Signed-off-by: Deepak Rawat <drawat@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
6 years agosvga: sync svga3d_types.h with upstream changes
Charmaine Lee [Mon, 30 Oct 2017 14:30:33 +0000 (07:30 -0700)]
svga: sync svga3d_types.h with upstream changes

Reviewed-by: Brian Paul <brianp@vmware.com>
6 years agowinsys/drm: enable intra_surface_copy if HW_CAP2 is supported
Charmaine Lee [Fri, 27 Oct 2017 22:41:37 +0000 (15:41 -0700)]
winsys/drm: enable intra_surface_copy if HW_CAP2 is supported

With drm version 2_15, we can inquire for support of HW_CAP2.
If it is supported, we can enable intra_surface_copy support.

Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Sinclair Yeh <syeh@vmware.com>
Reviewed-by: Deepak Rawat <drawat@vmware.com>
6 years agosvga: add git version logging at init time
Charmaine Lee [Fri, 27 Oct 2017 17:24:46 +0000 (10:24 -0700)]
svga: add git version logging at init time

Before we can log the git version in the host log,
we'll add the git version in the init debug message.

Reviewed-by: Brian Paul <brianp@vmware.com>
6 years agosvga: fix a typo in svga_texture_copy_region()
Charmaine Lee [Wed, 25 Oct 2017 05:10:20 +0000 (22:10 -0700)]
svga: fix a typo in svga_texture_copy_region()

Trivial.

6 years agosvga: use helper function to do copy region
Charmaine Lee [Tue, 24 Oct 2017 17:57:42 +0000 (10:57 -0700)]
svga: use helper function to do copy region

Use the common helper function svga_texture_copy_region
for copy region command.

Reviewed-by: Brian Paul <brianp@vmware.com>
6 years agosvga: fix cubemap array rendering with backed surface view
Charmaine Lee [Thu, 19 Oct 2017 23:03:44 +0000 (16:03 -0700)]
svga: fix cubemap array rendering with backed surface view

This patch fixes the layer index when rendering to a
backed surface view of a cubemap array.

Fixes piglit test fbo-generatemipmap-cubemap array.

Reviewed-by: Brian Paul <brianp@vmware.com>
6 years agosvga: add a helper function to send ResolveCopy command
Charmaine Lee [Mon, 16 Oct 2017 18:31:15 +0000 (11:31 -0700)]
svga: add a helper function to send ResolveCopy command

Reviewed-by: Brian Paul <brianp@vmware.com>
6 years agosvga: sync svga3d header files
Charmaine Lee [Fri, 13 Oct 2017 22:08:02 +0000 (15:08 -0700)]
svga: sync svga3d header files

This is a squash of what was orginally three commits.

Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Neha Bhende <bhenden@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
6 years agosvga: add SM4_1 enable debug print
Charmaine Lee [Wed, 11 Oct 2017 21:26:52 +0000 (14:26 -0700)]
svga: add SM4_1 enable debug print

Reviewed-by: Brian Paul <brianp@vmware.com>
6 years agosvga: fix swizzling for texture gather
Charmaine Lee [Tue, 10 Oct 2017 22:30:21 +0000 (15:30 -0700)]
svga: fix swizzling for texture gather

Texture swizzling for texture gather needs to be done to the selected texels
rather than to the returned vector. This patch has specical cases
for the different swizzles in emit_tg4().

Fixes a lot of piglit texture gather tests.

Reviewed-by: Neha Bhende <bhenden@vmware.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
6 years agosvga: fix starting index for system values
Charmaine Lee [Wed, 4 Oct 2017 23:51:35 +0000 (16:51 -0700)]
svga: fix starting index for system values

Currently, the starting index for system values is assigned to
the next index after the highest index of the tgsi declared input registers.
But the tgsi index might be different from the actual assigned index, hence
this might cause overlap of indices.
With this patch, the shader linker keeps track of the highest index of the
translated input registers, and the next index will be used for the
starting index for system values.

Fixes SHIM errors running arb_copy_image-formats on SM4_1 device.

Reviewed-by: Brian Paul <brianp@vmware.com>
6 years agowinsys/svga: Add support for new surface ioctl, multisample pattern
Deepak Rawat [Tue, 19 Sep 2017 22:56:26 +0000 (15:56 -0700)]
winsys/svga: Add support for new surface ioctl, multisample pattern

Kernel driver version 2.15 added new surface ioctl named:
DRM_VMW_GB_SURFACE_CREATE_EXT
DRM_VMW_GB_SURFACE_REF_EXT

The new ioctl has support for 64-bit svga3d_flags if
DRM_VMW_PARAM_SM4_1 is available.

Multisampling surface mob size calculation is added. Also synced the
relevant header update.

svga device modified the surface define command V3 with new parameter
multisampling pattern. Adding support for that in winsys.

Signed-off-by: Deepak Rawat <drawat@vmware.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Sinclair Yeh <syeh@vmware.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
6 years agosvga: enable MSAA for SM4_1 device
Brian Paul [Wed, 5 Sep 2018 15:48:35 +0000 (09:48 -0600)]
svga: enable MSAA for SM4_1 device

The SVGA device is deprecating the DX9 MSAA support.
This patch enables MSAA for SM4_1 device by explicitly
setting the SVGA3D_SURFACE_MULTISAMPLE bit.
For SM4_1 device, only 4 samples is supported.

Reviewed-by: Brian Paul <brianp@vmware.com>
6 years agosvga: add sample count to the surface_can_create interface
Charmaine Lee [Tue, 26 Sep 2017 23:23:58 +0000 (16:23 -0700)]
svga: add sample count to the surface_can_create interface

With this patch, sample count is also taken into account
when determining if a resource can be created.

Reviewed-by: Brian Paul <brianp@vmware.com>
6 years agosvga: implement support for GL_ARB_texture_query_lod
Brian Paul [Wed, 4 Oct 2017 17:21:18 +0000 (11:21 -0600)]
svga: implement support for GL_ARB_texture_query_lod

Just translate the TGSI LODQ intruction to VGPU10 LOD instruction.
All (4) Piglit GL_ARB_texture_query_lod tests pass.

Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Reviewed-by: Neha Bhende <bhenden@vmware.com>
6 years agosvga: Add support for arb_texture_gather
Neha Bhende [Fri, 25 Aug 2017 12:06:41 +0000 (05:06 -0700)]
svga: Add support for arb_texture_gather

With sm4_1, we can support single channel 2D or CubeMap textures.
This patch exercises this feature.

Tested with piglit

v2: As per Brian's comment

Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
6 years agosvga: add support for interpolation at sample position
Brian Paul [Tue, 1 Aug 2017 03:12:07 +0000 (21:12 -0600)]
svga: add support for interpolation at sample position

Vs. sampling at the centroid or the fragment center.

Note that this does not fix failures with the Piglit
arb_sample_shading-interpolate-at-sample-position or
arb_sample_shading-ignore-centroid-qualifier.exe tests at this time.

Reviewed-by: Charmaine Lee <charmainel@vmware.com>
6 years agosvga: clarify sys value -> input register mapping
Brian Paul [Tue, 1 Aug 2017 03:12:07 +0000 (21:12 -0600)]
svga: clarify sys value -> input register mapping

We translate TGSI system value registers to VGPU10 input registers.
Add a comment and set file = TGSI_FILE_INPUT.  That's not stricly
necessary since we map both TGSI_FILE_INPUT and TGSI_FILE_SYSTEM_VALUE
to VGPU10_OPERAND_TYPE_INPUT, but this makes the code a bit more
understandable.

Reviewed-by: Charmaine Lee <charmainel@vmware.com>
6 years agosvga: add support for FS sample mask output
Brian Paul [Tue, 1 Aug 2017 03:12:07 +0000 (21:12 -0600)]
svga: add support for FS sample mask output

This, with the previous work for sample position/id query, allows
us to enable per-sample shading for VGPU 10.1.

Note that quite a few Piglit arb_sample_shading tests still do not
pass, but many do.

Reviewed-by: Charmaine Lee <charmainel@vmware.com>
6 years agosvga: add support for sample id, sample position
Brian Paul [Tue, 1 Aug 2017 03:12:07 +0000 (21:12 -0600)]
svga: add support for sample id, sample position

Sample ID is just a system value.  Sample position must be implemented
with the VGPU10_OPCODE_SAMPLE_POS instruction.

Reviewed-by: Charmaine Lee <charmainel@vmware.com>
6 years agosvga: implement no-op svga_set_min_samples()
Brian Paul [Tue, 1 Aug 2017 03:12:07 +0000 (21:12 -0600)]
svga: implement no-op svga_set_min_samples()

This is part of the per-sample shading feature (PIPE_CAP_SAMPLE_SHADING).

Reviewed-by: Charmaine Lee <charmainel@vmware.com>
6 years agosvga: add support for independent blend function per render target
Charmaine Lee [Wed, 4 Oct 2017 17:05:58 +0000 (11:05 -0600)]
svga: add support for independent blend function per render target

This patch adds support for GL_ARB_draw_buffers_blend extension
for SM4_1 device.

Fixes piglit test fbo-draw-buffers-blend.

This patch is squashed with a subsequent patch which fixed a
regression.

Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Neha Bhende <bhenden@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
6 years agosvga: emit shader version as 4.0 or 4.1 depending on device support
Brian Paul [Thu, 25 May 2017 19:46:49 +0000 (13:46 -0600)]
svga: emit shader version as 4.0 or 4.1 depending on device support

Reviewed-by: Charmaine Lee <charmainel@vmware.com>
6 years agosvga: restructure nested if's in emit_src_register()
Brian Paul [Thu, 25 May 2017 18:45:01 +0000 (12:45 -0600)]
svga: restructure nested if's in emit_src_register()

To make it cleaner for subsequent changes.

Reviewed-by: Charmaine Lee <charmainel@vmware.com>