Marek Olšák [Sun, 18 Oct 2015 20:07:01 +0000 (22:07 +0200)]
radeonsi: add one more SWITCH_ON_EOI requirement for Hawaii and VI
The VI condition depends on geometry shaders and MAX_PRIMGRP_IN_WAVE.
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Marek Olšák [Sun, 18 Oct 2015 19:51:41 +0000 (21:51 +0200)]
radeonsi: only apply the instancing bug workaround to Bonaire
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Marek Olšák [Sun, 18 Oct 2015 19:43:30 +0000 (21:43 +0200)]
radeonsi: add SWITCH_ON_EOI requirement for 4 SE parts
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Marek Olšák [Sun, 18 Oct 2015 19:28:54 +0000 (21:28 +0200)]
radeonsi: remove unnecessary PARTIAL_VS_WAVE setting for streamout
hardware does this automatically
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Marek Olšák [Thu, 22 Oct 2015 19:32:23 +0000 (21:32 +0200)]
radeonsi: allow unbinding vertex shaders
Draw calls without a vertex shader are skipped.
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Marek Olšák [Thu, 22 Oct 2015 20:19:34 +0000 (22:19 +0200)]
radeonsi: allow unbinding pixel shaders and remove the dummy shader
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Marek Olšák [Thu, 22 Oct 2015 20:18:49 +0000 (22:18 +0200)]
radeonsi: add draw_vbo check for a NULL pixel shader
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Marek Olšák [Thu, 22 Oct 2015 20:17:28 +0000 (22:17 +0200)]
radeonsi: add checks for a NULL pixel shader
This will allow removing the dummy PS.
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Marek Olšák [Thu, 22 Oct 2015 20:14:53 +0000 (22:14 +0200)]
gallium/util: add a test for NULL fragment shaders
Just to validate that radeonsi doesn't crash.
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Marek Olšák [Thu, 22 Oct 2015 17:46:07 +0000 (19:46 +0200)]
st/mesa: don't load state parameters if there are none
Out of 7063 shaders from my shader-db:
- 6564 (93%) shaders don't have any state parameters.
- 347 (5%) shaders have 1 state parameter for WPOS lowering.
- The remaining 2% have more state parameters, usually matrices.
Reviewed-by: Brian Paul <brianp@vmware.com>
Samuel Li [Thu, 22 Oct 2015 16:06:43 +0000 (12:06 -0400)]
radeonsi: add Stoney pci ids
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Samuel Li <samuel.li@amd.com>
Cc: mesa-stable@lists.freedesktop.org
Samuel Li [Fri, 21 Aug 2015 19:35:46 +0000 (15:35 -0400)]
radeonsi: add support for Stoney asics (v3)
v2 (agd): rebase on mesa master, split pci ids to
separate commit
v3 (agd): use carrizo for llvm processor name for
llvm 3.7 and older
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Signed-off-by: Samuel Li <samuel.li@amd.com>
Cc: mesa-stable@lists.freedesktop.org
Ilia Mirkin [Fri, 23 Oct 2015 06:14:31 +0000 (02:14 -0400)]
nvc0: respect edgeflag attribute width
The edgeflag comes in as ubyte with glEdgeFlagPointer but as float with
plain immediate glEdgeFlag. Avoid reading bytes that weren't meant for
the edgeflag in the pointer case.
Fixes intermittent failures with gl-2.0-edgeflag piglit (and valgrind
complaints about reading uninitialized memory).
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Cc: mesa-stable@lists.freedesktop.org
Jose Fonseca [Fri, 23 Oct 2015 10:40:25 +0000 (11:40 +0100)]
gallivm: Explicitly disable unsupported CPU features.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=92214
CC: "10.6 11.0" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Eric Anholt [Wed, 19 Aug 2015 16:38:14 +0000 (09:38 -0700)]
vc4: Convert blending to being done in 4x8 unorm normally.
We can't do this all the time, because you want blending to be done in
linear space, and sRGB would lose too much precision being done in 4x8.
The win on instructions is pretty huge when you can, though.
total uniforms in shared programs: 32065 -> 32168 (0.32%)
uniforms in affected programs: 327 -> 430 (31.50%)
total instructions in shared programs: 92644 -> 89830 (-3.04%)
instructions in affected programs: 15580 -> 12766 (-18.06%)
Improves openarena performance at 1920x1080 from 10.7fps to 11.2fps.
Eric Anholt [Fri, 9 Jan 2015 02:07:15 +0000 (18:07 -0800)]
vc4: Add QIR/QPU support for the 8-bit vector instructions.
Eric Anholt [Fri, 23 Oct 2015 15:34:01 +0000 (16:34 +0100)]
vc4: Don't try to CSE non-SSA instructions.
This can happen when we're doing destination packing -- we don't know
what's in the rest of the register.
Signed-off-by: Eric Anholt <eric@anholt.net>
Eric Anholt [Wed, 19 Aug 2015 05:38:34 +0000 (22:38 -0700)]
nir: Add opcodes for saturated vector math.
This corresponds to instructions used on vc4 for its blending inside of
shaders. I've seen these opcodes on other architectures before, but I
think it's the first time these are needed in Mesa.
v2: Rename to 'u' instead of 'i', since they're all 'u'norm (from review
by jekstrand)
Eric Anholt [Fri, 23 Oct 2015 14:26:12 +0000 (15:26 +0100)]
vc4: Add dumping of VC4_PACKET_GL_INDEXED_PRIMITIVE.
Eric Anholt [Fri, 23 Oct 2015 13:43:41 +0000 (14:43 +0100)]
vc4: Add a workaround for HW-2116 (state counter wrap fails).
I haven't proven that this happens (I've got other GPU hangs in the
way), but the closed driver also does this and it's documented as an
errata.
Eric Anholt [Fri, 23 Oct 2015 13:41:47 +0000 (14:41 +0100)]
vc4: Fix missing \n in a perf_debug().
Kristian Høgsberg Kristensen [Wed, 21 Oct 2015 06:22:27 +0000 (23:22 -0700)]
i965/fs: Allow copy propagating into new surface access opcodes
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Kristian Høgsberg Kristensen <krh@bitplanet.net>
Kristian Høgsberg Kristensen [Thu, 22 Oct 2015 06:43:34 +0000 (23:43 -0700)]
i965/fs: Optimize ssbo stores
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Write groups of enabled components together.
Signed-off-by: Kristian Høgsberg Kristensen <krh@bitplanet.net>
Kristian Høgsberg Kristensen [Thu, 22 Oct 2015 05:49:14 +0000 (22:49 -0700)]
i965/fs: Drop offset_reg temporary in ssbo load
Now that we don't read each component one-by-one, we don't need the
temoprary vgrf for the offset. More importantly, this register was type
UD while the nir source was type D. This broke copy propagation and left
a redundant MOV in the generated code.
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Kristian Høgsberg Kristensen <krh@bitplanet.net>
Kristian Høgsberg Kristensen [Wed, 21 Oct 2015 06:31:49 +0000 (23:31 -0700)]
i965/fs: Avoid scalar destinations in emit_uniformize()
The scalar destination registers break copy propagation. Instead compute
the results to a regular register and then reference a component when we
later use the result as a source.
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Kristian Høgsberg Kristensen <krh@bitplanet.net>
Kristian Høgsberg Kristensen [Wed, 21 Oct 2015 06:16:51 +0000 (23:16 -0700)]
i965/fs: Don't uniformize surface index twice
The emit_untyped_read and emit_untyped_write helpers already uniformize
the surface index argument. No need to do it before calling them.
Reviewed-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Kristian Høgsberg Kristensen <krh@bitplanet.net>
Kristian Høgsberg Kristensen [Wed, 21 Oct 2015 06:07:56 +0000 (23:07 -0700)]
i965/fs: Use unsigned immediate 0 when eliminating SHADER_OPCODE_FIND_LIVE_CHANNEL
The destination for SHADER_OPCODE_FIND_LIVE_CHANNEL is always a UD
register. When we replace the opcode with a MOV, make sure we use a UD
immediate 0 so copy propagation doesn't bail because of non-matching
types.
Reviewed-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Kristian Høgsberg Kristensen <krh@bitplanet.net>
Kristian Høgsberg Kristensen [Sat, 17 Oct 2015 04:58:14 +0000 (21:58 -0700)]
i965/fs: Read all components of a SSBO field with one send
Instead of looping through single-component reads, read all components
in one go.
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Signed-off-by: Kristian Høgsberg Kristensen <krh@bitplanet.net>
Kristian Høgsberg Kristensen [Sat, 17 Oct 2015 04:38:05 +0000 (21:38 -0700)]
i965: Don't use message headers for untyped reads
We always set the mask to 0xffff, which is what it defaults to when no
header is present. Let's drop the header instead.
v2: Only remove header for untyped reads. Typed reads always need the
header.
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Signed-off-by: Kristian Høgsberg Kristensen <krh@bitplanet.net>
Alejandro Piñeiro [Fri, 23 Oct 2015 13:32:30 +0000 (15:32 +0200)]
i965/vec4: check opcode on vec4_instruction::reads_flag(channel)
Commit f17b78 added an alternative reads_flag(channel) that returned
if the instruction was reading a specific channel flag. By mistake it
only took into account the predicate, but when the opcode is
VS_OPCODE_UNPACK_FLAGS_SIMD4X2 there isn't any predicate, but the flag
are used.
That mistake caused some regressions on old hw. More information on
this bug:
https://bugs.freedesktop.org/show_bug.cgi?id=92621
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Eric Anholt [Wed, 21 Oct 2015 15:40:46 +0000 (16:40 +0100)]
vc4: Use Rob's NIR-based user clip lowering.
Eric Anholt [Mon, 10 Aug 2015 18:28:53 +0000 (11:28 -0700)]
vc4: Also dump the decimation mode for resolved stores.
Eric Anholt [Mon, 10 Aug 2015 18:27:21 +0000 (11:27 -0700)]
vc4: Use VC4_GET_FIELD and other defines in dumping VC4_RENDER_CONFIG.
Eric Anholt [Thu, 22 Oct 2015 10:31:56 +0000 (11:31 +0100)]
vc4: Add a sentinel after simulator buffers for buffer overflow detection.
This is a little bit like the mprotect-based fencing I've experimented
with, but it's simple and low overhead. The downside is that only catches
writes, not reads.
It didn't catch any bad writes on a current piglit run, but may be useful
in the future.
Samuel Iglesias Gonsalvez [Wed, 21 Oct 2015 07:46:48 +0000 (09:46 +0200)]
glsl: fix shader storage block member rules when adding program resources
Commit f24e5e did not take into account arrays of named shader
storage blocks.
Fixes 20 dEQP-GLES31.functional.ssbo.* tests:
dEQP-GLES31.functional.ssbo.layout.single_struct_array.per_block_buffer.shared_instance_array
dEQP-GLES31.functional.ssbo.layout.single_struct_array.per_block_buffer.packed_instance_array
dEQP-GLES31.functional.ssbo.layout.single_struct_array.per_block_buffer.std140_instance_array
dEQP-GLES31.functional.ssbo.layout.single_struct_array.per_block_buffer.std430_instance_array
dEQP-GLES31.functional.ssbo.layout.single_struct_array.single_buffer.shared_instance_array
dEQP-GLES31.functional.ssbo.layout.single_struct_array.single_buffer.packed_instance_array
dEQP-GLES31.functional.ssbo.layout.single_struct_array.single_buffer.std140_instance_array
dEQP-GLES31.functional.ssbo.layout.single_struct_array.single_buffer.std430_instance_array
dEQP-GLES31.functional.ssbo.layout.single_nested_struct_array.per_block_buffer.shared_instance_array
dEQP-GLES31.functional.ssbo.layout.single_nested_struct_array.per_block_buffer.packed_instance_array
dEQP-GLES31.functional.ssbo.layout.single_nested_struct_array.per_block_buffer.std140_instance_array
dEQP-GLES31.functional.ssbo.layout.single_nested_struct_array.per_block_buffer.std430_instance_array
dEQP-GLES31.functional.ssbo.layout.single_nested_struct_array.single_buffer.shared_instance_array
dEQP-GLES31.functional.ssbo.layout.single_nested_struct_array.single_buffer.packed_instance_array
dEQP-GLES31.functional.ssbo.layout.single_nested_struct_array.single_buffer.std140_instance_array
dEQP-GLES31.functional.ssbo.layout.single_nested_struct_array.single_buffer.std430_instance_array
dEQP-GLES31.functional.ssbo.layout.random.all_per_block_buffers.2
dEQP-GLES31.functional.ssbo.layout.random.all_per_block_buffers.29
dEQP-GLES31.functional.ssbo.layout.random.all_per_block_buffers.33
dEQP-GLES31.functional.ssbo.layout.random.all_shared_buffer.3
V2:
- Rename some variables (Timothy)
Signed-off-by: Samuel Iglesias Gonsalvez <siglesias@igalia.com>
Reviewed-by: Timothy Arceri <timothy.arceri@collabora.com>
Chia-I Wu [Thu, 22 Oct 2015 16:45:49 +0000 (00:45 +0800)]
ilo: add support for scratch spaces
When a kernel reports a non-zero per-thread scratch space size, make sure the
hardware state is correctly set up, and a scratch bo is allocated.
Chia-I Wu [Thu, 22 Oct 2015 16:24:26 +0000 (00:24 +0800)]
ilo: fix scratch space setup in core
Move scratch_size out of ilo_state_shader_kernel_info and
ilo_state_compute_interface_info. A scratch space is shared by all
kernels/interfaces. Update builder to emit relocs for scratch bos.
Timothy Arceri [Thu, 22 Oct 2015 05:18:44 +0000 (16:18 +1100)]
glsl: remove excess location qualifier validation
Location has never been able to be a negative value because it has
always been validated in the parser.
Also the linker doesn't check for negatives like the comment claims.
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Dave Airlie [Fri, 23 Oct 2015 02:11:43 +0000 (12:11 +1000)]
docs: update relnotes to mention virgl driver.
Dave Airlie [Fri, 13 Mar 2015 04:15:47 +0000 (14:15 +1000)]
virgl/vtest: add vtest driver
virgl/vtest is a swrast driver that allows the
virgl acceleration to be tested without having
a virtual machine.
The backend has a unix socket server that
this connects to.
This is run by setting
LIBGL_ALWAYS_SOFTWARE=y
GALLIUM_DRIVER=virpipe
In this mode all renderering is sent over
a socket to the remote renderer, and the
results are readback and copies to the screen
using drisw. This works well enough to develop
new features and to help debug.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Dave Airlie [Thu, 22 Jan 2015 05:11:47 +0000 (15:11 +1000)]
virgl: add driver for virtio-gpu 3D (v2)
virgl is the 3D acceleration backend for the
virtio-gpu shipping with qemu.
The 3D acceleration is designed around gallium
and TGSI as the virtualisation layer. The backend
renderer translates the virgl interface into
OpenGL currently.
This is the initial import of the driver to mesa.
The kernel driver portions are lined up for drm-next.
Currently this driver supports up to GL3.3 and some
misc extensions if the host driver exposes it. It is
planned to iterate the virgl API to new GL levels
as mesa host drivers gain features.
v2: fix resource tracking across flushes to avoid
->bind hack in mapping.
consolidate mapping and waiting code for transfers.
use u_range for dirt tracking.
handle larger shaders in protocol.
include virtgpu_drm.h in mesa for now.
add translation layer for gallium tgsi to virgl tgsi.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Dave Airlie [Thu, 22 Jan 2015 05:18:05 +0000 (15:18 +1000)]
tgsi: try and handle overflowing shaders. (v2)
This is used to detect error in virgl if we overflow the shader
dumping buffers.
v2: return a bool.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Dave Airlie [Mon, 16 Sep 2013 00:13:55 +0000 (10:13 +1000)]
tgsi: add option to dump floats as hex values
This adds support to the parser to accept hex values as floats,
and then adds support to the dumper to allow the user to select
to dump float as 32-bit hex numbers.
This is required to get accurate values for virgl use of TGSI.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Sinclair Yeh [Tue, 13 Oct 2015 19:58:26 +0000 (12:58 -0700)]
svga: Condition preemptive flush on draw emission
On ultra high resolution modes, the preemptive flush flag can be
set midway through command submission, a condition that cannot be
recovered from a flush-retry, causing rendering artifacts.
This patch prevents a preemtive_flush until a draw has been
emitted.
Signed-off-by: Sinclair Yeh <syeh@vmware.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Brian Paul [Fri, 16 Oct 2015 22:14:46 +0000 (16:14 -0600)]
svga: try to avoid index generation for some primitive types
The svga device doesn't directly support quads, quad strips or polygons
so we have to convert those types to indexed triangle lists. But we
can sometimes avoid that if we're drawing flat/constant-colored prims
and we don't have to worry about provoking vertex.
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Reviewed-by: José Fonseca <jfonseca@vmware.com>
Brian Paul [Fri, 16 Oct 2015 22:12:19 +0000 (16:12 -0600)]
svga: avoid provoking vertex conversion when possible
Provoking vertex comes into play when doing flat shading. But if we know
that all fragments in a primitive are the same color, the provoking vertex
doesn't matter. Check for that case and use whichever provoking vertex
convention is supported by the device.
This avoids generating an index buffer to do the PV conversion.
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Reviewed-by: José Fonseca <jfonseca@vmware.com>
Brian Paul [Thu, 22 Oct 2015 21:36:25 +0000 (15:36 -0600)]
svga: detect constant color writes in fragment shaders
Examine the fragment shader to try to detect TGSI shaders which use
"MOV OUT[0], CONST[i]" to write a constant value for the fragment color.
In this case, all fragments will have the same color (unless blending is
enabled).
This is a common case for OpenGL code such as: glColor(), glBegin(),
glVertex(), ..., glEnd() when lighting/fog/etc are disabled. In this
case, the Mesa/gallium state tracker actually generates a simple
"MOV OUT[0], CONST[i]" fragment shader.
This will be used by the next commit to avoid provoking vertex conversion
(creating/rewriting an index buffer) when drawing flat-shaded primitives.
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Reviewed-by: José Fonseca <jfonseca@vmware.com>
Brian Paul [Wed, 21 Oct 2015 19:58:04 +0000 (13:58 -0600)]
mesa: check for unchanged line width before error checking
Reviewed-by: Matt Turner <mattst88@gmail.com>
Brian Paul [Wed, 21 Oct 2015 19:42:37 +0000 (13:42 -0600)]
st/mesa: use _mesa_RasterPos() when possible
The st_RasterPos() function goes to great pains to implement the
rasterpos transformation. It basically uses gallium's draw module to
execute the vertex shader to draw a point, then capture that point's
attributes.
But glRasterPos isn't typically used with a vertex shader so we can
usually use the old/fixed-function implementation which is a lot simpler
and faster.
This can add up for legacy apps that make a lot of calls to glRasterPos.
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Brian Paul [Wed, 21 Oct 2015 19:42:21 +0000 (13:42 -0600)]
tnl: remove t_rasterpos.c
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Brian Paul [Wed, 21 Oct 2015 19:40:58 +0000 (13:40 -0600)]
drivers/common: use _mesa_RasterPos instead of _tnl_RasterPos
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Brian Paul [Wed, 21 Oct 2015 19:39:15 +0000 (13:39 -0600)]
mesa: copy rasterpos evaluation code into core Mesa
We'll remove it from the tnl module next. By lifting this code into core
Mesa we can use it from the gallium state tracker.
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Brian Paul [Wed, 21 Oct 2015 19:38:23 +0000 (13:38 -0600)]
vbo: optimize vertex copying when 'wrapping'
Instead of calling memcpy() 'n' times, we can do it all at once since
the source and dest regions are all contiguous.
Reviewed-by: Matt Turner <mattst88@gmail.com>
Alex Deucher [Thu, 22 Oct 2015 16:24:42 +0000 (12:24 -0400)]
radeon/uvd: don't expose HEVC on old UVD hw (v3)
The section for UVD 2 and older was not updated
when HEVC support was added. Reported by Kano
on irc.
v2: integrate the UVD2 and older checks into the
main switch statement.
v3: handle encode checking as well. Encode is
already checked in the top case statement, so
drop encode checks in the lower case statement.
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: mesa-stable@lists.freedesktop.org
Alejandro Piñeiro [Fri, 9 Oct 2015 16:39:42 +0000 (18:39 +0200)]
i965/vec4: print predicate control at brw_vec4 dump_instruction
v2: externalize pred_ctrl_align16 from brw_disasm.c instead of adding
a copy on brw_vec4.c, as suggested by Matt Turner
Reviewed-by: Matt Turner <mattst88@gmail.com>
Alejandro Piñeiro [Thu, 1 Oct 2015 14:41:30 +0000 (16:41 +0200)]
i965/vec4: use an envvar to decide to print the assembly on cmod_propagation tests
The complete way to do this would be parse INTEL_DEBUG and
print the output if DEBUG_VS (or a new one) is present
(see intel_debug.c).
But that seems like an overkill for the unit tests, that
after all, the most common use case is being run when
calling make check.
v2: use the same idea for the fs counterpart too, as suggested by
Matt Turner
Reviewed-by: Matt Turner <mattst88@gmail.com>
Alejandro Piñeiro [Wed, 30 Sep 2015 11:39:30 +0000 (13:39 +0200)]
i965/vec4: Add unit tests for cmod propagation pass
This include the same tests coming from test_fs_cmod_propagation, (non
vector glsl types included) plus some new with vec4 types, inspired on
the regressions found while the optimization was a work in progress.
Additionally, the check of number of instructions after the
optimization was changed from EXPECT_EQ to ASSERT_EQ. This was done to
avoid a crash on failing tests that expected no optimization, as after
checking the number of instructions, there were some checks related to
this last instruction opcode/conditional mod.
v2: update tests after Matt Turner's review of the optimization pass
v3: tweaks on the tests (mostly on the comments), after Matt Turner's
review
Reviewed-by: Matt Turner <mattst88@gmail.com>
Alejandro Piñeiro [Mon, 28 Sep 2015 15:00:19 +0000 (17:00 +0200)]
i965/vec4: adding vec4_cmod_propagation optimization
vec4 port of fs_cmod_propagation.
Shader-db results (no vec4 grepping):
total instructions in shared programs:
6240413 ->
6235841 (-0.07%)
instructions in affected programs: 401933 -> 397361 (-1.14%)
total loops in shared programs: 1979 -> 1979 (0.00%)
helped: 2265
HURT: 0
v2: remove extra space and combine two if blocks, as suggested by
Matt Turner
v3: add condition check to bail out if current inst and inst being
scanned has different writemask, as pointed by Matt Turner
v3: updated shader-db numbers
v4: remove block from foreach_inst_in_block_*_starting_from after
commit
801f151917fedb13c5c6e96281a18d833dd6901f
Reviewed-by: Matt Turner <mattst88@gmail.com>
Alejandro Piñeiro [Wed, 14 Oct 2015 18:26:43 +0000 (20:26 +0200)]
i965/vec4: track and use independently each flag channel
vec4_live_variables tracks now each flag channel independently, so
vec4_dead_code_eliminate can update the writemask of null registers,
based on which component are alive at the moment. This would allow
vec4_cmod_propagation to optimize out several movs involving null
registers.
v2: added support to track each flag channel independently at vec4
live_variables, as v1 assumed that it was already doing it, as
pointed by Francisco Jerez
v3: general cleaningn after Matt Turner's review
Reviewed-by: Matt Turner <mattst88@gmail.com>
Alejandro Piñeiro [Fri, 9 Oct 2015 14:59:20 +0000 (16:59 +0200)]
i965/vec4: nir_emit_if doesn't need to predicate based on all the channels
v2: changed comment, as suggested by Matt Turner
Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
Matt Turner [Thu, 22 Oct 2015 19:26:27 +0000 (12:26 -0700)]
i965/vec4/gs: Fix signed/unsigned comparison warning.
Matt Turner [Wed, 21 Oct 2015 01:31:02 +0000 (18:31 -0700)]
i965/fs: Emit a single ADD instruction for SET_SAMPLE_ID on Gen8+.
Gen8+ lifted the register region restriction that an instruction whose
destination spans two registers must have sources that also span two
registers.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Anuj Phogat <anuj.phogat@gmail.com>
Matt Turner [Wed, 21 Oct 2015 01:29:42 +0000 (18:29 -0700)]
i965/fs: Drop unnecessary write-enable-all from SET_SAMPLE_ID.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Anuj Phogat <anuj.phogat@gmail.com>
Matt Turner [Wed, 21 Oct 2015 01:23:50 +0000 (18:23 -0700)]
i965/fs: Trim unneeded channels in SampleID setup.
The AND and SHR produce a scalar value that we had been replicating
across $dispatch_width channels. The immediate MOV produces only four
useful channels of data.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Anuj Phogat <anuj.phogat@gmail.com>
Matt Turner [Wed, 21 Oct 2015 00:51:12 +0000 (17:51 -0700)]
i965/fs: Use type-W for immediate in SampleID setup.
Not a functional difference, but register is loaded with a signed
immediate (V) and added to a signed type (D) producing a signed result
(D).
Also change the type of g0 to allow for compaction.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Anuj Phogat <anuj.phogat@gmail.com>
Matt Turner [Mon, 19 Oct 2015 17:51:42 +0000 (10:51 -0700)]
i965/vec4: Initialize LOD to 0.0f for textureQueryLevels() and texture().
We implement textureQueryLevels (which takes no arguments, save the
sampler) using the resinfo message (which takes an argument of LOD).
Without initializing it, we'd generate a MOV from the null register to
load the LOD argument.
Essentially the same logic applies to texture. A vertex shader cannot
compute derivatives and so cannot produce an LOD, so TXL with an LOD of
0.0 is used.
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Matt Turner [Mon, 29 Jun 2015 21:04:14 +0000 (14:04 -0700)]
i965: Note that the UV immediate type is Gen6+.
Jose Fonseca [Wed, 21 Oct 2015 16:19:41 +0000 (17:19 +0100)]
gallivm: Translate all util_cpu_caps bits to LLVM attributes.
This should prevent disparity between features Mesa and LLVM
believe are supported by the CPU.
http://lists.freedesktop.org/archives/mesa-dev/2015-October/thread.html#96990
Tested on a i7-3720QM w/ LLVM 3.3 and 3.6.
v2: Increase SmallVector initial size as suggested by Gustaw Smolarczyk.
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
CC: "10.6 11.0" <mesa-stable@lists.freedesktop.org>
Jordan Justen [Tue, 20 Oct 2015 06:13:09 +0000 (23:13 -0700)]
i965/fs: Disable CSE optimization for untyped & typed surface reads
An untyped surface read is volatile because it might be affected by a
write.
In the ES31-CTS.compute_shader.resources-max test, two back to back
read/modify/writes of an SSBO variable looked something like this:
r1 = untyped_surface_read(ssbo_float)
r2 = r1 + 1
untyped_surface_write(ssbo_float, r2)
r3 = untyped_surface_read(ssbo_float)
r4 = r3 + 1
untyped_surface_write(ssbo_float, r4)
And after CSE, we had:
r1 = untyped_surface_read(ssbo_float)
r2 = r1 + 1
untyped_surface_write(ssbo_float, r2)
r4 = r1 + 1
untyped_surface_write(ssbo_float, r4)
Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Chia-I Wu [Thu, 22 Oct 2015 03:29:17 +0000 (11:29 +0800)]
ilo: make sure there is HiZ before resolving
We do not want to perform a depth resolve on an MCS enabled surface.
Chia-I Wu [Thu, 22 Oct 2015 03:19:05 +0000 (11:19 +0800)]
ilo: fix max thread count for HS on Gen8
It is in DW2 on Gen8.
Ben Widawsky [Tue, 20 Oct 2015 21:29:41 +0000 (14:29 -0700)]
i965: Advertise ARB_shader_stencil_export (gen9+)
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Ben Widawsky [Tue, 20 Oct 2015 21:29:39 +0000 (14:29 -0700)]
i965: Implement ARB_shader_stencil_export (gen9+)
v2: remove useless source_stencil_to_render_target (Ken)
Squash in the actual packing function, which also got to
v2:
Move the definition of the OPCODE outside of FB_WRITE opcodes (Matt)
Reorder the regioning to be in VWH order (Matt)
Don't retype src in the backend, just assert instead (Matt)
Rename the debug prints to something better (Matt)
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Ben Widawsky [Tue, 20 Oct 2015 21:29:37 +0000 (14:29 -0700)]
i965/fs: Enumerate logical fb writes arguments
Gen9 adds the ability to write out a stencil value, so we need to expand the
virtual payload by one. Abstracting this now makes that change easier to read.
I was admittedly confused early on about some of the hardcoding. If people
believe the resulting code is inferior, I am not super attached to the patch.
v2:
Remove explicit numbering from the enumeration (Matt).
Use a real naming scheme, and reference it in the opcode definition (Curro)
Add a missed hardcoded logical position in get_lowered_simd_width (Ben)
Add an assertion to make sure the component numbering is correct (Ben)
Cc: Matt Turner <mattst88@gmail.com>
Cc: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Brian Paul [Wed, 21 Oct 2015 22:07:33 +0000 (16:07 -0600)]
svga: fix clip plane regression after recent tgsi_scan change
Before the change "tgsi/scan: use properties for clip/cull distance
writemasks", the tgsi_shader_info::num_written_clipdistance field
was a multiple of four, now it's an accurate count. In the svga
driver, we need a minor change to the loop test.
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Kenneth Graunke [Fri, 10 Jul 2015 07:16:19 +0000 (00:16 -0700)]
i965: Implement gl_InvocationID.
It's stored in bits 31:27 of g1 (along with the URB handles).
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Kenneth Graunke [Thu, 24 Sep 2015 01:59:57 +0000 (18:59 -0700)]
i965: Implement nir_intrinsic_load_primitive.
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Kenneth Graunke [Tue, 30 Jun 2015 05:50:28 +0000 (22:50 -0700)]
i965: Add a fs_visitor constructor that takes a brw_gs_compile.
Unlike the vs/wm structs, brw_gs_compile is actually useful: it contains
the input VUE map and information about the control data headers.
Passing this in allows us to share that code in brw_gs.c, and calculate
them before deciding on vec4 vs. scalar mode, as it's independent of
that choice.
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Kenneth Graunke [Fri, 26 Jun 2015 23:20:21 +0000 (16:20 -0700)]
i965: Add a brw->scalar_gs flag controlled by INTEL_SCALAR_GS=1.
This patch introduces a brw->scalar_gs flag, similar to brw->scalar_vs,
which controls whether or not to use SIMD8 geometry shaders.
For now, we control it via a new environment variable, INTEL_SCALAR_GS.
This provides a convenient way to try it out.
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Kenneth Graunke [Wed, 6 May 2015 03:19:04 +0000 (20:19 -0700)]
i965: Make emit_urb_writes() reserve space for GS header information.
Geometry shaders have additional header data at the beginning of their
output URB entries. Shaders that use EndPrimitive() or multiple streams
have a control data header; shaders with a dynamic vertex count have an
additional vec4 slot to hold the 32-bit vertex count (and 96 bits of
padding).
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Kenneth Graunke [Thu, 12 Mar 2015 08:55:44 +0000 (01:55 -0700)]
i965: Make emit_urb_writes() only set EOT for the VS.
The GS will emit a bunch of vertices, and we don't want to do an EOT
prematurely. We'll emit GS_OPCODE_THREAD_END when we want to terminate
the thread.
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Kenneth Graunke [Thu, 12 Mar 2015 12:52:13 +0000 (05:52 -0700)]
i965: Make fs_visitor::emit_urb_writes reusable for scalar GS.
GS doesn't have ClampVertexColor, and we don't want to go through VS
structures.
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Kenneth Graunke [Tue, 29 Sep 2015 21:43:29 +0000 (14:43 -0700)]
i965: Introduce a brw_vue_prog_data::include_vue_handles flag.
Tessellation shaders and SIMD8 geometry shaders may need to resort to
the pull model for inputs at times. When set, the state upload code
will tell the hardware to provide URB handles for input data.
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Kenneth Graunke [Tue, 29 Sep 2015 21:32:02 +0000 (14:32 -0700)]
i965: Introduce a new SHADER_OPCODE_URB_READ_SIMD8 opcode.
In scalar mode, geometry shader inputs can easily take up hundreds of
registers. This makes pushing VUE entries impractical; we'll need to
resort to the pull model in some cases.
To support this, we introduce a new opcode corresponding to the "URB
Read SIMD8" message.
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Kenneth Graunke [Wed, 6 May 2015 07:04:10 +0000 (00:04 -0700)]
i965: Introduce new SHADER_OPCODE_URB_WRITE_SIMD8_MASKED/PER_SLOT opcodes.
In the vec4 backend, we have a vec4_instruction::urb_write_flags field.
There are many kinds of flags for SIMD4x2 messages.
However, there are really only two (per-slot offset, use channel masks)
for SIMD8 messages. Rather than adding a boolean flag for per-slot
offsets (polluting all instructions), I decided to just make three new
opcodes.
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Jason Ekstrand [Wed, 21 Oct 2015 19:03:21 +0000 (12:03 -0700)]
i965/gs: Do prog_data setup and other calculations in brw_compile_gs
This commit moves the large pile of setup calculations we have to do for
geometry shaders out of brw_gs_emit and into brw_compile_gs. This has a
couple of nice implications. First, it's less work that the caller of
brw_compile_gs has to do. Second, it's consistent with the vertex and
fragment stages. Finally, it allows us to put brw_gs_compile back behind
the API boundary where it belongs.
v2 (Jason Ekstrand):
- Pull the changes to use nir info into a separate patch
- Put brw_gs_compile into brw_shader.h rather than brw_vec4_gs_visitor.h
so that we can use it for scalar GS.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Wed, 21 Oct 2015 19:02:08 +0000 (12:02 -0700)]
i965/gs: Use NIR info for setting up prog_data
Previously, we were pulling bits from GL data structures in order to set up
the prog_data. However, in this brave new world of NIR, we want to be
pulling it out of the NIR shader whenever possible. This way, we can move
all this setup code into brw_compile_gs without depending on the old GL
stuff.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Wed, 21 Oct 2015 00:12:03 +0000 (17:12 -0700)]
i965/gs: Pull prog_data out of brw_gs_compile
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Tue, 20 Oct 2015 23:40:30 +0000 (16:40 -0700)]
i965/gs: Use NIR instead of the brw_geometry_program for GS metadata
With this, we can remove the geometry program from brw_gs_compile.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Tue, 20 Oct 2015 23:21:09 +0000 (16:21 -0700)]
i965/gs: Move the mem_ctx argument to brw_compile_gs
This makes it better match the other brw_compile_* functions.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Tue, 20 Oct 2015 23:46:50 +0000 (16:46 -0700)]
i965/gs: Set static_vertex_count unconditionally on GEN8+
We always have NIR, so there's no reason for the check.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Wed, 21 Oct 2015 00:40:19 +0000 (17:40 -0700)]
nir: Constify nir_gs_count_vertices
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Tue, 20 Oct 2015 23:35:44 +0000 (16:35 -0700)]
nir/info: Add more information about geometry shaders
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Ben Widawsky [Tue, 20 Oct 2015 21:29:38 +0000 (14:29 -0700)]
i965: (trivial) rename computes stencil to gen9
All the documentation I can find says that this bit (and functionality) only
exists on SKL+. Since the bit isn't yet used, there is no real impact here.
The original code was added by Ken here (a surprisingly long time ago):
commit
f3c6d6f1e151f6a44a76038dccebe4434038dcb1
Author: Kenneth Graunke <kenneth@whitecape.org>
Date: Thu Nov 29 21:00:27 2012 -0800
i965: Update 3DSTATE_PS, 3DSTATE_WM, and add 3DSTATE_PS_EXTRA.
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Ben Widawsky [Tue, 20 Oct 2015 21:29:36 +0000 (14:29 -0700)]
i965: Correct the comment about fb write payload
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Nanley Chery [Thu, 8 Oct 2015 23:44:30 +0000 (16:44 -0700)]
mesa/glformats: Undo code changes from _mesa_base_tex_format() move
The refactoring commit,
c6bf1cd, accidentally reverted
cd49b97
and
99b1f47. These changes caused more code to be added to the
function and removed the existing support for ASTC. This patch
reverts those modifications.
v2. Actually include ASTC support again.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=92221
Cc: "11.0" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Nanley Chery <nanley.g.chery@intel.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
Matt Turner [Tue, 20 Oct 2015 02:02:16 +0000 (19:02 -0700)]
i965: Mark compacted 3-src instructions as Gen8+.
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Matt Turner [Tue, 20 Oct 2015 01:59:53 +0000 (18:59 -0700)]
i965: Add const to brw_compact_inst_bits.
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Matt Turner [Mon, 19 Oct 2015 22:08:28 +0000 (15:08 -0700)]
i965: Add mask_control_ex field and handle it in compaction.
Documentation is sparse, but it appears to have existed on G45 and ILK
as a second bit extension of the mask_control field. Setting the pair of
bits to 0b11 enables "NoCMask".
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Matt Turner [Mon, 19 Oct 2015 21:47:17 +0000 (14:47 -0700)]
i965: Add devinfo->gen assertions for acc_wr_control.
... and for flag_subreg_nr since it's right near by.
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>