mesa.git
9 years agomesa: trim down #includes in compiler.h
Brian Paul [Wed, 25 Feb 2015 00:48:31 +0000 (17:48 -0700)]
mesa: trim down #includes in compiler.h

Don't include stuff we don't need.  Fix a few #includes elsewhere to
keep thing building.

Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
9 years agor300g: remove dependency on compiler.h
Brian Paul [Wed, 25 Feb 2015 00:47:00 +0000 (17:47 -0700)]
r300g: remove dependency on compiler.h

It only needs typical stdio.h and stdlib.h functions.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
9 years agomesa: don't include limits.h in compiler.h
Brian Paul [Wed, 25 Feb 2015 00:05:48 +0000 (17:05 -0700)]
mesa: don't include limits.h in compiler.h

Not needed.

Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
9 years agomesa: don't include float.h in compiler.h
Brian Paul [Wed, 25 Feb 2015 00:03:47 +0000 (17:03 -0700)]
mesa: don't include float.h in compiler.h

Not needed.

Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
9 years agomesa: only include ctype.h where it's used
Brian Paul [Tue, 24 Feb 2015 23:56:54 +0000 (16:56 -0700)]
mesa: only include ctype.h where it's used

Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
9 years agomesa: include stdarg.h only where it's used
Brian Paul [Tue, 24 Feb 2015 23:55:53 +0000 (16:55 -0700)]
mesa: include stdarg.h only where it's used

Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
9 years agomesa: remove M_PI, M_E, M_LOG2E macro definitions
Brian Paul [Tue, 24 Feb 2015 23:44:53 +0000 (16:44 -0700)]
mesa: remove M_PI, M_E, M_LOG2E macro definitions

Should be defined in math.h.  If not, we can add them to c99_math.h

Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
9 years agoglsl: #include c99_math.h instead of core.h
Brian Paul [Tue, 24 Feb 2015 23:42:43 +0000 (16:42 -0700)]
glsl: #include c99_math.h instead of core.h

We only need the M_LOG2E definition.

Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
9 years agogallium: whitespace, comment formatting fixes in p_defines.h
Brian Paul [Thu, 26 Feb 2015 00:04:05 +0000 (17:04 -0700)]
gallium: whitespace, comment formatting fixes in p_defines.h

Just to keep things consistent.

Reviewed-by: José Fonseca <jfonseca@vmware.com>
9 years agoutil: add debug_print_bind_flags() debug helper
Brian Paul [Wed, 25 Feb 2015 23:59:54 +0000 (16:59 -0700)]
util: add debug_print_bind_flags() debug helper

Reviewed-by: José Fonseca <jfonseca@vmware.com>
9 years agogallium: renumber PIPE_BIND_ flags
Brian Paul [Wed, 25 Feb 2015 23:58:43 +0000 (16:58 -0700)]
gallium: renumber PIPE_BIND_ flags

Note that PIPE_BIND_COMMAND_ARGS_BUFFER and PIPE_BIND_LINEAR were both
bit 21 before.

Reviewed-by: José Fonseca <jfonseca@vmware.com>
9 years agometa: In pbo_{Get,}TexSubImage don't repeatedly rebind the source tex
Neil Roberts [Wed, 25 Feb 2015 15:10:01 +0000 (15:10 +0000)]
meta: In pbo_{Get,}TexSubImage don't repeatedly rebind the source tex

A layered PBO image is now interpreted as a single tall 2D image so
the z argument in _mesa_meta_bind_fbo_image is ignored. Therefore this
was just redundantly rebinding the same image repeatedly.

Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
9 years agomesa: use fi_type in vertex attribute code
Marius Predut [Wed, 25 Feb 2015 16:49:45 +0000 (09:49 -0700)]
mesa: use fi_type in vertex attribute code

For 32-bit builds, floating point operations use x86 FPU registers,
not SSE registers.  If we're actually storing an integer in a float
variable, the value might get modified when written to memory.  This
patch changes the VBO code to use the fi_type (float/int union) to
store/copy vertex attributes.

Also, this can improve performance on x86 because moving floats with
integer registers instead of FP registers is faster.

Neil Roberts review:
- include changes on all places that are storing attribute values.
- check with and without -O3 compiler flag.
Brian Paul review:
- use fi_type type instead gl_constant_value type
- fix a bunch of nit-picks.
- fix compiler warnings

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=82668
Signed-off-by: Marius Predut <marius.predut@intel.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
9 years agoi965/gen8: Use HALIGN_16 if MCS is enabled for non-MSRT
Anuj Phogat [Fri, 13 Feb 2015 19:21:21 +0000 (11:21 -0800)]
i965/gen8: Use HALIGN_16 if MCS is enabled for non-MSRT

Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Neil Roberts <neil@linux.intel.com>
9 years agoi965: Pass pointer to miptree as function parameter in intel_horizontal_texture_align...
Anuj Phogat [Fri, 20 Feb 2015 20:26:36 +0000 (12:26 -0800)]
i965: Pass pointer to miptree as function parameter in intel_horizontal_texture_alignment_unit

This will be used by next patch in the series.

Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Neil Roberts <neil@linux.intel.com>
9 years agoi965: Allocate texture buffer in intelTexImage
Anuj Phogat [Thu, 5 Feb 2015 20:24:57 +0000 (12:24 -0800)]
i965: Allocate texture buffer in intelTexImage

before calling _mesa_meta_pbo_TexSubImage(). This will be used in
later patches and will be required in Skylake to get the tile
resource mode of miptree before calling _mesa_meta_pbo_TexSubImage().

Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Neil Roberts <neil@linux.intel.com>
9 years agoi965: Make a function to check the conditions to use the blitter
Anuj Phogat [Fri, 13 Feb 2015 02:49:13 +0000 (18:49 -0800)]
i965: Make a function to check the conditions to use the blitter

No functional changes in the patch. Just makes the code look cleaner.

Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Neil Roberts <neil@linux.intel.com>
9 years agoi965: Move the comment to the right place
Anuj Phogat [Fri, 13 Feb 2015 00:59:12 +0000 (16:59 -0800)]
i965: Move the comment to the right place

Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Neil Roberts <neil@linux.intel.com>
9 years agoi965: Fix condition to use Y tiling in blitter in intel_miptree_create()
Anuj Phogat [Tue, 17 Feb 2015 18:40:58 +0000 (10:40 -0800)]
i965: Fix condition to use Y tiling in blitter in intel_miptree_create()

Y tiling is supported in blitter on SNB+.

Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Neil Roberts <neil@linux.intel.com>
9 years agometa: Pass null pointer for the pixel data to avoid unnecessary data upload
Anuj Phogat [Mon, 23 Feb 2015 18:56:14 +0000 (10:56 -0800)]
meta: Pass null pointer for the pixel data to avoid unnecessary data upload

to a temporary pbo created in _mesa_meta_pbo_GetTexSubImage().

Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Neil Roberts <neil@linux.intel.com>
9 years agometa: Fix buffer object assignment to account for both pack and unpack bo's
Anuj Phogat [Wed, 4 Feb 2015 00:56:04 +0000 (16:56 -0800)]
meta: Fix buffer object assignment to account for both pack and unpack bo's

create_texture_for_pbo() is shared by _mesa_meta_pbo_GetTexSubImage()
and _mesa_meta_pbo_TexSubImage() functions. So, we need to account
for both pack and unpack buffer objects.

Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Neil Roberts <neil@linux.intel.com>
9 years agometa: Use GL_STREAM_READ for pbo created with GL_PIXEL_PACK_BUFFER
Anuj Phogat [Wed, 4 Feb 2015 00:50:57 +0000 (16:50 -0800)]
meta: Use GL_STREAM_READ for pbo created with GL_PIXEL_PACK_BUFFER

create_texture_for_pbo() is used by both _mesa_meta_pbo_GetTexSubImage()
and _mesa_meta_pbo_TexSubImage() functions with different PBO targets.
Use GL_STREAM_READ with GL_PIXEL_PACK_BUFFER and GL_STREAM_DRAW with
GL_PIXEL_UNPACK_BUFFER.

Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Neil Roberts <neil@linux.intel.com>
9 years agometa: Add assertion check for ctx->Meta->SaveStackDepth
Anuj Phogat [Wed, 4 Feb 2015 01:04:45 +0000 (17:04 -0800)]
meta: Add assertion check for ctx->Meta->SaveStackDepth

before using it for derefrencing.

Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Neil Roberts <neil@linux.intel.com>
9 years agometa: Do power of two samples check only for samples > 0
Anuj Phogat [Mon, 1 Jan 2001 00:05:42 +0000 (16:05 -0800)]
meta: Do power of two samples check only for samples > 0

otherwise samples=0 passes the check, which is invalid.

Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Neil Roberts <neil@linux.intel.com>
9 years agoglsl: Rewrite and fix min/max to saturate optimization.
Matt Turner [Tue, 24 Feb 2015 18:41:52 +0000 (10:41 -0800)]
glsl: Rewrite and fix min/max to saturate optimization.

There were some bugs, and the code was really difficult to follow. We
would optimize

   min(max(x, b), 1.0) into max(sat(x), b)

but not pay attention to the order of min/max and also do

   max(min(x, b), 1.0) into max(sat(x), b)

Corrects four shaders from Champions of Regnum that do

   min(max(x, 1), 10)

and corrects rendering of Mass Effect under VMware Workstation.

Cc: "10.4 10.5" <mesa-stable@lists.freedesktop.org>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=89180
Reviewed-by: Abdiel Janulgue <abdiel.janulgue@linux.intel.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
9 years agofreedreno: drop ARRAY_SIZE macro
Rob Clark [Wed, 25 Feb 2015 13:35:32 +0000 (08:35 -0500)]
freedreno: drop ARRAY_SIZE macro

Since now ARRAY_SIZE has been added to util/macros.h.  Fixes a bunch of:

  freedreno_util.h:79:0: warning: "ARRAY_SIZE" redefined
   #define ARRAY_SIZE(arr) (sizeof(arr) / sizeof((arr)[0]))
   ^
  In file included from ../../../../src/gallium/include/pipe/p_compiler.h:36:0,
                   from ../../../../src/gallium/include/pipe/p_context.h:31,
                   from freedreno_context.h:32,
                   from freedreno_context.c:29:
  ../../../../src/util/macros.h:29:0: note: this is the location of the previous definition
   #  define ARRAY_SIZE(x) (sizeof(x) / sizeof(*(x)))
   ^

Signed-off-by: Rob Clark <robclark@freedesktop.org>
9 years agoi965: Don't force x-tiling for 16-bpp formats on Gen>7
Neil Roberts [Thu, 19 Feb 2015 16:09:41 +0000 (16:09 +0000)]
i965: Don't force x-tiling for 16-bpp formats on Gen>7

Sandybridge doesn't support y-tiling for surface formats with 16 or
more bpp. There was previously an override to explicitly allow this
for Gen7. However, this restriction is also removed in Gen8+ so we
should use y-tiling there too.

This is important to do for Skylake which doesn't support x-tiling for
3D surfaces.

Reviewed-by: Ben Widawsky <ben@bwidawsk.net>
9 years agoglx: Fix returned values of GLX_RENDERER_PREFERRED_PROFILE_MESA
Andreas Boll [Tue, 24 Feb 2015 19:01:30 +0000 (20:01 +0100)]
glx: Fix returned values of GLX_RENDERER_PREFERRED_PROFILE_MESA

If the renderer supports the core profile the query returned incorrectly
0x8 as value, because it was using (1U << __DRI_API_OPENGL_CORE) for the
returned value.

The same happened with the compatibility profile. It returned 0x1
(1U << __DRI_API_OPENGL) instead of 0x2.

Internal DRI defines:
   dri_interface.h: #define __DRI_API_OPENGL       0
   dri_interface.h: #define __DRI_API_OPENGL_CORE  3

Those two bits are supposed for internal usage only and should be
translated to GLX_CONTEXT_CORE_PROFILE_BIT_ARB (0x1) for a preferred
core context profile and GLX_CONTEXT_COMPATIBILITY_PROFILE_BIT_ARB (0x2)
for a preferred compatibility context profile.

This patch implements the above translation in the glx module.

v2: Fix the incorrect behavior in the glx module

Cc: "10.3 10.4 10.5" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Andreas Boll <andreas.boll.dev@gmail.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
9 years agodri/common: Update comment about driQueryRendererIntegerCommon
Andreas Boll [Mon, 9 Feb 2015 10:18:04 +0000 (11:18 +0100)]
dri/common: Update comment about driQueryRendererIntegerCommon

Since 87d3ae0b45b6b6bb50b583dafa55eb109449a005
driQueryRendererIntegerCommon handles __DRI2_RENDERER_PREFFERED_PROFILE
too.

Signed-off-by: Andreas Boll <andreas.boll.dev@gmail.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
9 years agoglsl: add double support for packing varyings
Ilia Mirkin [Fri, 20 Feb 2015 09:08:38 +0000 (04:08 -0500)]
glsl: add double support for packing varyings

Doubles are always packed, but a single double will never cross a slot
boundary -- single slots can still be wasted in some situations.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
9 years agocommon: Fix PBOs for 1D_ARRAY.
Laura Ekstrand [Thu, 19 Feb 2015 01:17:38 +0000 (17:17 -0800)]
common: Fix PBOs for 1D_ARRAY.

Corrects the way that _mesa_meta_pbo_TexSubImage and
_mesa_meta_pbo_GetTexSubImage handle 1D_ARRAY textures.  Fixes a failure in
the Piglit arb_direct_state_access/gettextureimage-targets test.

Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
Tested-by: Laura Ekstrand <laura@jlekstrand.net>
Cc: "10.4, 10.5" <mesa-stable@lists.freedesktop.org>
9 years agocommon: Correct PBO 2D_ARRAY handling.
Laura Ekstrand [Tue, 24 Feb 2015 22:51:29 +0000 (14:51 -0800)]
common: Correct PBO 2D_ARRAY handling.

Changes PBO uploads and downloads to use a tall (height * depth) 2D texture
for blitting. This fixes the bug where 2D_ARRAY, 3D, and CUBE_MAP_ARRAY
textures are not properly uploaded and downloaded.

Removes the option to use a 2D ARRAY texture for the PBO during upload and
download.  This option didn't work because the miptree couldn't be set up
reliably.

v2: Review from Jason Ekstrand and Neil Roberts:
   -Delete the depth parameter from create_texture_for_pbo
   -Abandon the option to create a 2D ARRAY texture in create_texture_for_pbo

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
Cc: "10.4, 10.5" <mesa-stable@lists.freedesktop.org>
9 years agocommon: Correct texture init for meta pbo uploads and downloads.
Laura Ekstrand [Tue, 24 Feb 2015 21:29:11 +0000 (13:29 -0800)]
common: Correct texture init for meta pbo uploads and downloads.

This moves the line setting immutability for the texture to after
_mesa_initialize_texture_object so that the initializer function will not
cancel it out. Moreover, because of the ARB_texture_view extension, immutable
textures must have NumLayers > 0, or depth will equal (0-1)=0xFFFFFFFF during
SURFACE_STATE setup, which triggers assertions.

v2: Review from Kenneth Graunke:
   - Include more explanation in the commit message.
   - Make texture setup bug fixes into a separate patch.

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
Cc: "10.4, 10.5" <mesa-stable@lists.freedesktop.org>
9 years agomesa: remove DEG2RAD macro
Brian Paul [Tue, 24 Feb 2015 22:26:34 +0000 (15:26 -0700)]
mesa: remove DEG2RAD macro

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agomesa: remove MAX_GLUSHORT, move MAX_GLUINT
Brian Paul [Tue, 24 Feb 2015 22:23:22 +0000 (15:23 -0700)]
mesa: remove MAX_GLUSHORT, move MAX_GLUINT

The later is only used in one place in swrast.

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agomesa: move signbit() macro to c99_math.h
Brian Paul [Tue, 24 Feb 2015 22:11:45 +0000 (15:11 -0700)]
mesa: move signbit() macro to c99_math.h

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agomesa: remove unused isblank() function
Brian Paul [Tue, 24 Feb 2015 22:05:47 +0000 (15:05 -0700)]
mesa: remove unused isblank() function

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agoglcpp: remove unneeded #include of core.h
Brian Paul [Tue, 24 Feb 2015 22:05:05 +0000 (15:05 -0700)]
glcpp: remove unneeded #include of core.h

isblank() is not used in the code.

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agomesa: remove sqrtf macro
Brian Paul [Tue, 24 Feb 2015 22:03:44 +0000 (15:03 -0700)]
mesa: remove sqrtf macro

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agoi965: Remove redundant discard jumps.
Kenneth Graunke [Sun, 6 Jul 2014 05:10:41 +0000 (22:10 -0700)]
i965: Remove redundant discard jumps.

With the previous optimization in place, some shaders wind up with
multiple discard jumps in a row, or jumps directly to the next
instruction.  We can remove those.

Without NIR on Haswell:
total instructions in shared programs: 5777258 -> 5775872 (-0.02%)
instructions in affected programs:     20312 -> 18926 (-6.82%)
helped:                                716

With NIR on Haswell:
total instructions in shared programs: 5773163 -> 5771785 (-0.02%)
instructions in affected programs:     21040 -> 19662 (-6.55%)
helped:                                717

v2: Use the CFG rather than the old instructions list.  Presumably
    the placeholder halt will be in the last basic block.

v3: Make sure placeholder_halt->prev isn't the head sentinel (caught
    twice by Eric Anholt).

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
9 years agoglsl: Optimize "if (cond) discard;" to a conditional discard.
Kenneth Graunke [Thu, 23 Oct 2014 03:48:21 +0000 (20:48 -0700)]
glsl: Optimize "if (cond) discard;" to a conditional discard.

st_glsl_to_tgsi and ir_to_mesa have handled conditional discards for a
long time; the previous patch added that capability to i965.

i965 (Haswell) shader-db stats:

Without NIR:
total instructions in shared programs: 5792133 -> 5776360 (-0.27%)
instructions in affected programs:     737585 -> 721812 (-2.14%)
helped:                                6300
HURT:                                  68
GAINED:                                2

With NIR:
total instructions in shared programs: 5787538 -> 5769569 (-0.31%)
instructions in affected programs:     767843 -> 749874 (-2.34%)
helped:                                6522
HURT:                                  35
GAINED:                                6

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Connor Abbott <cwabbott0@gmail.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
9 years agoi965/fs: Handle conditional discards.
Kenneth Graunke [Tue, 19 Aug 2014 22:22:43 +0000 (15:22 -0700)]
i965/fs: Handle conditional discards.

The discard condition tells us which channels we want killed.  We want
to invert that condition to get the channels that should survive (remain
live) in f0.1.  Emit a CMP to negate it.

Nothing generates these today, but that will change shortly.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Connor Abbott <cwabbott0@gmail.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
9 years agonir: Introduce nir_intrinsic_discard_if.
Kenneth Graunke [Tue, 24 Feb 2015 07:47:00 +0000 (23:47 -0800)]
nir: Introduce nir_intrinsic_discard_if.

This is a conditional discard, which takes a boolean source.

Note that we don't generate ir_discard::condition today, so this
shouldn't break drivers (since none implement this intrinsic yet).

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Connor Abbott <cwabbott0@gmail.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
9 years agoglsl: Delete dead discard conditions in constant folding.
Kenneth Graunke [Tue, 24 Feb 2015 09:00:22 +0000 (01:00 -0800)]
glsl: Delete dead discard conditions in constant folding.

opt_constant_folding() already detects conditional assignments where the
condition is constant, and either deletes the assignment or the
condition.

Make it handle discards in the same fashion.

Spotted happening in the wild in Tropico 5 shaders.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Connor Abbott <cwabbott0@gmail.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
9 years agoglsl: Handle conditional discards in lower_discard_flow().
Kenneth Graunke [Tue, 24 Feb 2015 07:12:39 +0000 (23:12 -0800)]
glsl: Handle conditional discards in lower_discard_flow().

This pass wasn't prepared to handle conditional discards.

Instead of initializing the "discarded" temporary to "true", set it to
the condition.  Then, refer to the variable for the condition, to avoid
duplicating the expression tree.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Connor Abbott <cwabbott0@gmail.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
9 years agoglsl: Make ir_rvalue_visitor visit ir_discard::condition.
Kenneth Graunke [Tue, 24 Feb 2015 06:00:34 +0000 (22:00 -0800)]
glsl: Make ir_rvalue_visitor visit ir_discard::condition.

This was forgotten.

I omitted the NULL check since we don't check ir_assignment::condition
either.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Connor Abbott <cwabbott0@gmail.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
9 years agoglsl: Make ir_validate check the type of ir_discard::condition.
Kenneth Graunke [Tue, 24 Feb 2015 06:56:47 +0000 (22:56 -0800)]
glsl: Make ir_validate check the type of ir_discard::condition.

Copy and pasted from the ir_if::condition handling, plus a NULL check.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Connor Abbott <cwabbott0@gmail.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
9 years agoRevert "i965/fs: Remove force_writemask_all assertion for execsize < 8."
Matt Turner [Tue, 24 Feb 2015 20:11:21 +0000 (12:11 -0800)]
Revert "i965/fs: Remove force_writemask_all assertion for execsize < 8."

This reverts commit 0d8f27eab7b7e8b7a16e76aabd3f6a0ab4880497.

"This doesn't seem to be necessary." <- I was wrong!

Tested-by: Markus Wick <markus@selfnet.de>
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
9 years agoi965/fs: Emit MOV(1) instructions with force_writemask_all.
Matt Turner [Tue, 24 Feb 2015 20:03:37 +0000 (12:03 -0800)]
i965/fs: Emit MOV(1) instructions with force_writemask_all.

Fixes rendering with Dolphin.

Tested-by: Markus Wick <markus@selfnet.de>
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
9 years agoi965/fs: Optimize (gl_FrontFacing ? x : y) where x and y are ±1.0.
Matt Turner [Mon, 22 Dec 2014 03:50:03 +0000 (19:50 -0800)]
i965/fs: Optimize (gl_FrontFacing ? x : y) where x and y are ±1.0.

total instructions in shared programs: 5695356 -> 5689775 (-0.10%)
instructions in affected programs:     486231 -> 480650 (-1.15%)
helped:                                2604
LOST:                                  1

9 years agoi965/fs/nir: Optimize integer multiply by a 16-bit constant.
Matt Turner [Sun, 15 Feb 2015 22:03:09 +0000 (14:03 -0800)]
i965/fs/nir: Optimize integer multiply by a 16-bit constant.

Gen8+ support was just broken, since MUL now consumes 32-bits from both
sources. Fixes 986 piglit tests on my BDW.

total instructions in shared programs: 7753873 -> 7753522 (-0.00%)
instructions in affected programs:     28164 -> 27813 (-1.25%)
helped:                                77
GAINED:                                47

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
9 years agoi965/fs/nir: Optimize (gl_FrontFacing ? x : y) where x and y are ±1.0.
Matt Turner [Sun, 15 Feb 2015 21:45:04 +0000 (13:45 -0800)]
i965/fs/nir: Optimize (gl_FrontFacing ? x : y) where x and y are ±1.0.

total instructions in shared programs: 7756214 -> 7753873 (-0.03%)
instructions in affected programs:     455452 -> 453111 (-0.51%)
helped:                                2333

Reviewed-by: Eric Anholt <eric@anholt.net>
9 years agonir/register: Add a parent_instr field
Jason Ekstrand [Fri, 30 Jan 2015 05:45:53 +0000 (21:45 -0800)]
nir/register: Add a parent_instr field

This adds a parent_instr field similar to the one for ssa_def.  The
difference here is that the parent_instr field on a nir_register can be
NULL if the register does not have a unique definition or if that
definition does not dominate all its uses.  We set this field in the
out-of-SSA pass so that backends can get SSA-like information even after
they have gone out of SSA.

Reviewed-by: Connor Abbott <cwabbott0@gmail.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
9 years agost/mesa: remove unused/broken function st_print_shaders
Marek Olšák [Sat, 21 Feb 2015 22:20:05 +0000 (23:20 +0100)]
st/mesa: remove unused/broken function st_print_shaders

Reviewed-by: Dave Airlie <airlied@redhat.com>
9 years agost/mesa: remove struct qualifier from st_src_reg parameter
Brian Paul [Tue, 24 Feb 2015 16:59:30 +0000 (09:59 -0700)]
st/mesa: remove struct qualifier from st_src_reg parameter

It's a class.  Silences MSVC warning.

9 years agomesa: remove INV_SQRTF() macro
Brian Paul [Tue, 24 Feb 2015 16:39:51 +0000 (09:39 -0700)]
mesa: remove INV_SQRTF() macro

Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agomesa: remove ceilf, floorf macros
Brian Paul [Tue, 24 Feb 2015 16:33:41 +0000 (09:33 -0700)]
mesa: remove ceilf, floorf macros

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agomesa: remove expf macro
Brian Paul [Tue, 24 Feb 2015 16:32:27 +0000 (09:32 -0700)]
mesa: remove expf macro

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agomesa: remove logf macro
Brian Paul [Tue, 24 Feb 2015 16:31:34 +0000 (09:31 -0700)]
mesa: remove logf macro

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agomesa: remove powf macro
Brian Paul [Tue, 24 Feb 2015 16:26:41 +0000 (09:26 -0700)]
mesa: remove powf macro

Use the wrapper in c99_math.h if needed.

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agomesa: remove unused exp2f, log2f, truncf wrappers
Brian Paul [Tue, 24 Feb 2015 16:24:27 +0000 (09:24 -0700)]
mesa: remove unused exp2f, log2f, truncf wrappers

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agomesa: remove unused acosf, asinf, atan2f, etc. macros
Brian Paul [Tue, 24 Feb 2015 16:20:12 +0000 (09:20 -0700)]
mesa: remove unused acosf, asinf, atan2f, etc. macros

Not used anywhere.  If any of these are needed, they should be added
to c99_math.h

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agomesa: replace FABSF with fabsf
Brian Paul [Tue, 24 Feb 2015 16:08:50 +0000 (09:08 -0700)]
mesa: replace FABSF with fabsf

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agomesa: replace FLOORF with floorf
Brian Paul [Tue, 24 Feb 2015 16:04:54 +0000 (09:04 -0700)]
mesa: replace FLOORF with floorf

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agomesa: remove unused CEILF macro
Brian Paul [Tue, 24 Feb 2015 16:03:25 +0000 (09:03 -0700)]
mesa: remove unused CEILF macro

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agomesa: replace LOGF, EXPF with logf, expf
Brian Paul [Tue, 24 Feb 2015 16:01:51 +0000 (09:01 -0700)]
mesa: replace LOGF, EXPF with logf, expf

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agomesa: replace FREXPF, LDEXPF with frexpf, ldexpf
Brian Paul [Tue, 24 Feb 2015 15:57:17 +0000 (08:57 -0700)]
mesa: replace FREXPF, LDEXPF with frexpf, ldexpf

Start getting rid of some imports.h macros.  Use the c99 functions instead.

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agotargets/libgl-xlib: add src/ include dir to fix build
Brian Paul [Tue, 24 Feb 2015 15:47:57 +0000 (08:47 -0700)]
targets/libgl-xlib: add src/ include dir to fix build

9 years agoswrast: fix a few release build warnings
Brian Paul [Tue, 24 Feb 2015 00:48:00 +0000 (17:48 -0700)]
swrast: fix a few release build warnings

9 years agor600g,radeonsi: fix streamout after pipeline stats have been used
Marek Olšák [Mon, 23 Feb 2015 17:03:02 +0000 (18:03 +0100)]
r600g,radeonsi: fix streamout after pipeline stats have been used

EVENT_TYPE_PIPELINESTAT_STOP disables streamout queries too.

Luckily, pipeline stats are enabled by default, so we don't even have to
emit EVENT_TYPE_PIPELINESTAT_START.

Tested on Hawaii, Bonaire, Redwood, RV730.

Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
9 years agoradeonsi: small cleanup around current_rast_prim
Marek Olšák [Sun, 22 Feb 2015 16:42:20 +0000 (17:42 +0100)]
radeonsi: small cleanup around current_rast_prim

- remove the last parameter of si_emit_rasterizer_prim_state
- remove the last unused parameter of si_emit_draw_registers
- use current_rast_prim in si_emit_draw_registers

Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
9 years agoradeonsi: set current_rast_prim in the right place
Marek Olšák [Sun, 22 Feb 2015 18:14:42 +0000 (19:14 +0100)]
radeonsi: set current_rast_prim in the right place

Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
9 years agoradeonsi: simplify obtaining a shader property in si_emit_clip_regs
Marek Olšák [Sun, 22 Feb 2015 14:36:20 +0000 (15:36 +0100)]
radeonsi: simplify obtaining a shader property in si_emit_clip_regs

Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
9 years agoradeonsi: only preload VertexID for the GS copy shader
Marek Olšák [Sun, 22 Feb 2015 14:11:49 +0000 (15:11 +0100)]
radeonsi: only preload VertexID for the GS copy shader

The copy shader doesn't use any other preloaded VGPRs.

Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
9 years agoradeonsi: dump the shader key when dumping shaders
Marek Olšák [Sun, 22 Feb 2015 13:58:30 +0000 (14:58 +0100)]
radeonsi: dump the shader key when dumping shaders

Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
9 years agor600g,radeonsi: cleanup of hex literals
Marek Olšák [Fri, 20 Feb 2015 11:22:00 +0000 (12:22 +0100)]
r600g,radeonsi: cleanup of hex literals

0x3F800000 -> fui(1.0)
0x00000000 -> 0

Reviewed-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
9 years agoradeonsi: set PA_SU_HARDWARE_SCREEN_OFFSET to 0
Marek Olšák [Fri, 20 Feb 2015 11:14:37 +0000 (12:14 +0100)]
radeonsi: set PA_SU_HARDWARE_SCREEN_OFFSET to 0

It was probably 0 already, but it doesn't hurt to set it.

Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
9 years agost/mesa: cleanup st_translate_geometry_program
Marek Olšák [Sun, 22 Feb 2015 12:22:07 +0000 (13:22 +0100)]
st/mesa: cleanup st_translate_geometry_program

Mostly dead code or code that didn't do anything.

Computing gs_num_outputs at the end was also useless. It's already set
correctly.

Reviewed-by: Dave Airlie <airlied@redhat.com>
9 years agost/mesa: inline st_free_tokens
Marek Olšák [Sat, 21 Feb 2015 21:41:39 +0000 (22:41 +0100)]
st/mesa: inline st_free_tokens

Reviewed-by: Dave Airlie <airlied@redhat.com>
9 years agost/mesa: cleanup st_geometry_program structure
Marek Olšák [Sat, 21 Feb 2015 21:37:47 +0000 (22:37 +0100)]
st/mesa: cleanup st_geometry_program structure

It's full of unused variables and variables only used
in st_translate_geometry_program.

Reviewed-by: Dave Airlie <airlied@redhat.com>
9 years agomesa: add a missing GS support check in GetActiveUniformBlockiv
Marek Olšák [Sat, 21 Feb 2015 23:33:03 +0000 (00:33 +0100)]
mesa: add a missing GS support check in GetActiveUniformBlockiv

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
9 years agor600g: Implement GL_ARB_draw_indirect for EG/CM
Glenn Kennard [Tue, 24 Feb 2015 14:59:16 +0000 (15:59 +0100)]
r600g: Implement GL_ARB_draw_indirect for EG/CM

Requires Evergreen/Cayman and radeon kernel module
2.41.0 or newer.

Expected piglit fails due to hardware limitations:
* arb_draw_indirect-draw-arrays-prim-restart
  Restarts not applied for DrawArrays commands
* arb_draw_indirect-vertexid
  Base vertex offset is not included in vertex id

Marek: bump vgt_state num_dw by 3 (= space needed for one register write)

Signed-off-by: Glenn Kennard <glenn.kennard@gmail.com>
Signed-off-by: Marek Olšák <marek.olsak@amd.com>
9 years agofreedreno/a4xx: aniso filtering
Rob Clark [Tue, 24 Feb 2015 16:55:28 +0000 (11:55 -0500)]
freedreno/a4xx: aniso filtering

Signed-off-by: Rob Clark <robclark@freedesktop.org>
9 years agofreedreno: update generated headers
Rob Clark [Tue, 24 Feb 2015 16:48:01 +0000 (11:48 -0500)]
freedreno: update generated headers

Signed-off-by: Rob Clark <robclark@freedesktop.org>
9 years agofreedreno/a4xx: add ARB_instanced_arrays support
Rob Clark [Sun, 21 Dec 2014 16:52:44 +0000 (11:52 -0500)]
freedreno/a4xx: add ARB_instanced_arrays support

Signed-off-by: Rob Clark <robclark@freedesktop.org>
9 years agofreedreno/a4xx: handle index_bias (i.e. base_vertex)
Rob Clark [Tue, 24 Feb 2015 17:29:06 +0000 (12:29 -0500)]
freedreno/a4xx: handle index_bias (i.e. base_vertex)

Signed-off-by: Rob Clark <robclark@freedesktop.org>
9 years agofreedreno/a4xx: add support for vertexid and instanceid sysvals
Rob Clark [Sun, 21 Dec 2014 16:38:34 +0000 (11:38 -0500)]
freedreno/a4xx: add support for vertexid and instanceid sysvals

ir3 bits of it already in place from a3xx patch..

Signed-off-by: Rob Clark <robclark@freedesktop.org>
9 years agofreedreno/a4xx: pass number of instances to draw
Rob Clark [Sun, 21 Dec 2014 16:14:24 +0000 (11:14 -0500)]
freedreno/a4xx: pass number of instances to draw

a4xx has it's own draw packet, so needs equivalent update to what a3xx
already got.

Signed-off-by: Rob Clark <robclark@freedesktop.org>
9 years agodocs: add news item and link release notes for mesa 10.4.5
Emil Velikov [Tue, 24 Feb 2015 16:09:12 +0000 (16:09 +0000)]
docs: add news item and link release notes for mesa 10.4.5

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
9 years agodocs: Add sha256 sums for the 10.4.5 release
Emil Velikov [Sat, 21 Feb 2015 12:31:51 +0000 (12:31 +0000)]
docs: Add sha256 sums for the 10.4.5 release

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
(cherry picked from commit 41bdeda10210b8f1d358c04b25af2fbfe3335531)

9 years agoAdd release notes for the 10.4.5 release
Emil Velikov [Sat, 21 Feb 2015 12:22:08 +0000 (12:22 +0000)]
Add release notes for the 10.4.5 release

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
(cherry picked from commit a5c608e951c7d41be99a1f84450a1d5aaf6ccadd)

9 years agost/omx/dec/h264: fix picture out-of-order with poc type 0 v2
Leo Liu [Mon, 23 Feb 2015 18:50:06 +0000 (13:50 -0500)]
st/omx/dec/h264: fix picture out-of-order with poc type 0 v2

poc counter should be reset with IDR frame,
otherwise there would be a re-order issue with
frames before and after IDR

v2: add commit message

Signed-off-by: Leo Liu <leo.liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Cc: "10.4 10.5" <mesa-stable@lists.freedesktop.org>
9 years agoinstall-lib-links: remove the .install-lib-links file
Emil Velikov [Tue, 24 Feb 2015 14:05:15 +0000 (14:05 +0000)]
install-lib-links: remove the .install-lib-links file

With earlier commit (install-lib-links: don't depend on .libs directory)
we moved the location of the file from .libs/ to the current dir.
Although we did not attribute that in the former case autotools was
doing us a favour and removing the file. Explicitly remove the file at
clean-local time, otherwise we'll end up with dangling files.

Cc: "10.3 10.4 10.5" <mesa-stable@lists.freedesktop.org>
Cc: Matt Turner <mattst88@gmail.com>
Cc: Lucas Stach <l.stach@pengutronix.de>
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
9 years agoclover: Set appropriate flag defaults on memory object creation.
Francisco Jerez [Sun, 22 Feb 2015 13:05:19 +0000 (15:05 +0200)]
clover: Set appropriate flag defaults on memory object creation.

According to the spec when no device access mode is specified
clCreateBuffer and clCreateImage* should default to read/write, and
clCreateSubBuffer should default to the parent's device access flags.

clCreateSubBuffer is also required to inherit the host access and
host pointer flags from the parent.

Reviewed-and-tested-by: EdB <edb+mesa@sigluy.net>
9 years agoclover: Add CL_MEM_HOST_* flag checks.
EdB [Sun, 22 Feb 2015 13:00:06 +0000 (15:00 +0200)]
clover: Add CL_MEM_HOST_* flag checks.

Those flags have been introduced in OpenCL 1.2.

[ Francisco Jerez: Rebase.  Throw CL_INVALID_VALUE from
  clCreateSubBuffer if the subbuffer drops access flags from its
  parent.  Use single function taking the set of allowed host access
  flags to validate memory transfer operands. ]

Reviewed-by: Francisco Jerez <currojerez@riseup.net>
9 years agoclover: Factor out memory object flags validation to a helper function.
Francisco Jerez [Sat, 21 Feb 2015 23:02:41 +0000 (01:02 +0200)]
clover: Factor out memory object flags validation to a helper function.

And define constants for commonly used subsets of flags to save some
typing.

Reviewed-and-tested-by: EdB <edb+mesa@sigluy.net>
9 years agovc4: Update to current kernel sources.
Eric Anholt [Mon, 23 Feb 2015 06:48:34 +0000 (06:48 +0000)]
vc4: Update to current kernel sources.

New BO create and mmap ioctls are added.  The submit ABI gains a flags
argument, and the pointers are fixed at 64-bit.  Shaders are now fixed at
the start of their BOs.

9 years agor600: Fix build after 984f3069370cd4a347cb38269d430b428385affd
Eric Anholt [Tue, 24 Feb 2015 13:47:58 +0000 (13:47 +0000)]
r600: Fix build after 984f3069370cd4a347cb38269d430b428385affd

Same as for the CLAMP macro, undef it before including a header file that
tries to make fields with that name.

9 years agost/nine: Mark end of non-void function unreachable
Tobias Klausmann [Thu, 12 Feb 2015 17:31:41 +0000 (18:31 +0100)]
st/nine: Mark end of non-void function unreachable

Signed-off-by: Tobias Klausmann <tobias.johannes.klausmann@mni.thm.de>
Reviewed-by: Eric Anholt <eric@anholt.net>
9 years agogallium: include util/macros.h
Tobias Klausmann [Thu, 12 Feb 2015 17:31:40 +0000 (18:31 +0100)]
gallium: include util/macros.h

The most common macros are defined there, no use to duplicate these
Clean up the already redefinded macros

Signed-off-by: Tobias Klausmann <tobias.johannes.klausmann@mni.thm.de>
Reviewed-by: Eric Anholt <eric@anholt.net>