platform/upstream/mesa.git
9 years agoi965/vec4: Add a devinfo field to the generator and use it for gen checks
Jason Ekstrand [Thu, 16 Apr 2015 17:30:05 +0000 (10:30 -0700)]
i965/vec4: Add a devinfo field to the generator and use it for gen checks

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agoi965/fs: Add a devinfo field to the generator and use it for gen checks
Jason Ekstrand [Wed, 15 Apr 2015 00:45:40 +0000 (17:45 -0700)]
i965/fs: Add a devinfo field to the generator and use it for gen checks

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agoi965/device_info: Add a supports_simd16_3src flag
Jason Ekstrand [Fri, 17 Apr 2015 00:52:03 +0000 (17:52 -0700)]
i965/device_info: Add a supports_simd16_3src flag

This also involves moving revision checking to screen creation time and
passing that into brw_get_device_info so that we can get the right
device_info for early versions of SKL.  Since the only place we used
revision was to check for SIMD16 3-src instruction support, it's safe to
remove the revision field from brw_context.

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agoi965/device_info: Add a HSW_FEATURES macro
Jason Ekstrand [Fri, 17 Apr 2015 00:50:43 +0000 (17:50 -0700)]
i965/device_info: Add a HSW_FEATURES macro

It's basically just a copy of GEN7_FEATURES only with is_haswell set

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agoi965: Make the annotation code take a device_info instead of a context
Jason Ekstrand [Wed, 15 Apr 2015 22:01:25 +0000 (15:01 -0700)]
i965: Make the annotation code take a device_info instead of a context

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agoi965/fs: Remove the GL context from the generator
Jason Ekstrand [Wed, 15 Apr 2015 21:51:18 +0000 (14:51 -0700)]
i965/fs: Remove the GL context from the generator

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agoi965: Remove the context field from brw_compiler
Jason Ekstrand [Wed, 15 Apr 2015 21:13:58 +0000 (14:13 -0700)]
i965: Remove the context field from brw_compiler

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agoi965: Make the disassembler take a device_info instead of a context
Jason Ekstrand [Wed, 15 Apr 2015 20:46:21 +0000 (13:46 -0700)]
i965: Make the disassembler take a device_info instead of a context

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agoi965: Make instruction compaction take a device_info instead of a context
Jason Ekstrand [Wed, 15 Apr 2015 20:19:21 +0000 (13:19 -0700)]
i965: Make instruction compaction take a device_info instead of a context

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agoi965: Make the brw_inst helpers take a device_info instead of a context
Jason Ekstrand [Wed, 15 Apr 2015 01:00:06 +0000 (18:00 -0700)]
i965: Make the brw_inst helpers take a device_info instead of a context

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agoi965/eu: Add a devinfo parameter to brw_compile
Jason Ekstrand [Tue, 14 Apr 2015 23:57:24 +0000 (16:57 -0700)]
i965/eu: Add a devinfo parameter to brw_compile

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agoi965: Do better fake context setup in unit tests
Jason Ekstrand [Fri, 17 Apr 2015 22:23:18 +0000 (15:23 -0700)]
i965: Do better fake context setup in unit tests

In future tests, we will start relying on devinfo and not just brw in the
compiler.  Changing this now keeps these tests from failing in the future.

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agoi965: Remove the context parameter from brw_texture_offset
Jason Ekstrand [Tue, 14 Apr 2015 21:23:40 +0000 (14:23 -0700)]
i965: Remove the context parameter from brw_texture_offset

It wasn't really being used anyway.  We used it to assert that gpu_shader5
is supported in the back-end but that should be caught by the front-end.

Signed-off-by: Jason Ekstrand <jason.ekstrand@intel.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agosoftpipe: fix stencil write to use an integer value
Dave Airlie [Wed, 22 Apr 2015 07:13:06 +0000 (17:13 +1000)]
softpipe: fix stencil write to use an integer value

This fixes a number of regressions since
61393bdcdc3b63624bf6e9730444f5e9deeedfc8
u_tile: fix stencil texturing tests under softpipe

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=89960
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
9 years agomesa: Fix typo in a comment
Anuj Phogat [Wed, 22 Apr 2015 18:23:49 +0000 (11:23 -0700)]
mesa: Fix typo in a comment

Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
9 years agofreedreno: misc minor cleanups
Rob Clark [Wed, 22 Apr 2015 17:11:33 +0000 (13:11 -0400)]
freedreno: misc minor cleanups

Signed-off-by: Rob Clark <robclark@freedesktop.org>
9 years agofreedreno/a4xx: (partial) gl_FragCoord.zw
Rob Clark [Wed, 22 Apr 2015 17:09:28 +0000 (13:09 -0400)]
freedreno/a4xx: (partial) gl_FragCoord.zw

The bit to enable .z is still commented out, as it is triggering gpu
hangs in 0ad.  But at least gl_FragCoord.w works now, and we know what
bits we are *supposed* to set for .z (with that uncommented all piglit
fragcoord tests are passing).

Signed-off-by: Rob Clark <robclark@freedesktop.org>
9 years agofreedreno/a4xx: primitive-restart
Rob Clark [Wed, 22 Apr 2015 17:08:44 +0000 (13:08 -0400)]
freedreno/a4xx: primitive-restart

This was the missing bit to get dolphin-emu working on a4xx.

Signed-off-by: Rob Clark <robclark@freedesktop.org>
9 years agofreedreno/nir: sysval fixes
Rob Clark [Wed, 22 Apr 2015 17:07:33 +0000 (13:07 -0400)]
freedreno/nir: sysval fixes

Signed-off-by: Rob Clark <robclark@freedesktop.org>
9 years agofreedreno/a4xx: wire up integer texture sampling
Rob Clark [Wed, 22 Apr 2015 17:05:53 +0000 (13:05 -0400)]
freedreno/a4xx: wire up integer texture sampling

Similar to a3xx, the compiler needs to know the return type of the sam,
etc, instructions.

Signed-off-by: Rob Clark <robclark@freedesktop.org>
9 years agofreedreno/a4xx: formats updates/fixes
Rob Clark [Wed, 22 Apr 2015 17:00:03 +0000 (13:00 -0400)]
freedreno/a4xx: formats updates/fixes

Update formats table with new formats that Ilia has figured out, and fix
sampling from srgb texture and integer vbo's.

Signed-off-by: Rob Clark <robclark@freedesktop.org>
9 years agofreedreno: update generated headers
Rob Clark [Sat, 18 Apr 2015 16:08:53 +0000 (12:08 -0400)]
freedreno: update generated headers

Signed-off-by: Rob Clark <robclark@freedesktop.org>
9 years agogallium/targets/d3dadapter9: drop the libdrm prefix for drm.h
Emil Velikov [Fri, 20 Mar 2015 21:20:10 +0000 (21:20 +0000)]
gallium/targets/d3dadapter9: drop the libdrm prefix for drm.h

The path is provided by libdrm.pc and already used appropriately by
the build system.

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
9 years agocso: minor comment fix
Brian Paul [Tue, 21 Apr 2015 14:10:22 +0000 (08:10 -0600)]
cso: minor comment fix

9 years agoglsl: rewrite glsl_type::record_key_hash() to avoid buffer overflow
Brian Paul [Thu, 16 Apr 2015 21:29:18 +0000 (15:29 -0600)]
glsl: rewrite glsl_type::record_key_hash() to avoid buffer overflow

This should be more efficient than the previous snprintf() solution.
But more importantly, it avoids a buffer overflow bug that could result
in crashes or unpredictable results when processing very large interface
blocks.

For the app in question, key->length = 103 for some interfaces.  The check
if size >= sizeof(hash_key) was insufficient to prevent overflows of the
hash_key[128] array because it didn't account for the terminating zero.
In this case, this caused the call to hash_table_string_hash() to return
different results for identical inputs, and then shader linking failed.

This new solution also takes all structure fields into account instead
of just the first 15 when sizeof(pointer)==8.

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
9 years agomesa: add check for NV_texture_barrier in _mesa_TextureBarrierNV()
Brian Paul [Thu, 16 Apr 2015 21:27:41 +0000 (15:27 -0600)]
mesa: add check for NV_texture_barrier in _mesa_TextureBarrierNV()

If an app called glTextureBarrierNV() without checking if the
extension was available, we'd crash with some gallium drivers
in st_TextureBarrier() because the pipe_context::texture_barrier()
pointer was NULL.

Generate GL_INVALID_OPERATION instead.

Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Anuj Phogat <anuj.phogat@gmail.com>
9 years agomain: silence missing return value warning in array_index_of_resource()
Brian Paul [Thu, 16 Apr 2015 21:27:20 +0000 (15:27 -0600)]
main: silence missing return value warning in array_index_of_resource()

v2: return -1 instead of 0, per Emil Velikov.

Reviewed-by: Anuj Phogat <anuj.phogat@gmail.com>
9 years agoandroid: re-build all mesa binaries properly
Chih-Wei Huang [Fri, 3 Apr 2015 21:01:26 +0000 (05:01 +0800)]
android: re-build all mesa binaries properly

The clean steps ensure both 32-bit and 64-bit objects are cleaned.

Signed-off-by: Chih-Wei Huang <cwhuang@linux.org.tw>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
9 years agoandroid: xmlpool: cleanup the generation rules
Emil Velikov [Wed, 22 Apr 2015 14:38:57 +0000 (15:38 +0100)]
android: xmlpool: cleanup the generation rules

 - Do not attempt to create the save folder twice - both dir $@ and
PRIVATE_LOCALEDIR point to the same place.
 - Use @ and $(hide), for mkdir and python, to avoid spamming the
output.

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
9 years agoandroid: xmlpool: Get rid of the last use of intermediates-dir-for
Chih-Wei Huang [Fri, 3 Apr 2015 21:01:24 +0000 (05:01 +0800)]
android: xmlpool: Get rid of the last use of intermediates-dir-for

v2 [Emil Velikov]
 - Keep the PRIVATE_LOCALEDIR variable.
 - Do not use $(@D) but the more widespead $(dir $@)

Signed-off-by: Chih-Wei Huang <cwhuang@linux.org.tw>
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
9 years agoandroid: export the path of the generated headers
Chih-Wei Huang [Fri, 3 Apr 2015 21:01:23 +0000 (05:01 +0800)]
android: export the path of the generated headers

The modules need the headers can get the path automatically.

Signed-off-by: Chih-Wei Huang <cwhuang@linux.org.tw>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
9 years agoandroid: fix the building rules for Android 5.0
Chih-Wei Huang [Fri, 3 Apr 2015 21:01:22 +0000 (05:01 +0800)]
android: fix the building rules for Android 5.0

Android 5.0 allows modules to generate source into $OUT/gen, which will
then be copied into $OUT/obj and $OUT/obj_$(TARGET_2ND_ARCH) as necessary.
Modules will need to change calls to local-intermediates-dir into
local-generated-sources-dir.

The patch changes local-intermediates-dir into local-generated-sources-dir.
If the Android version is less than 5.0, fallback to local-intermediates-dir.

The patch also fixes the 64-bit building issue of Android 5.0.

v2 [Emil Velikov]
 - Keep the LOCAL_UNSTRIPPED_PATH variable.

Signed-off-by: Chih-Wei Huang <cwhuang@linux.org.tw>
9 years agoandroid: fix building issues of host binaries
Chih-Wei Huang [Fri, 3 Apr 2015 21:01:21 +0000 (05:01 +0800)]
android: fix building issues of host binaries

Define _GNU_SOURCE to enable features (__USE_XOPEN2K and __USE_UNIX98)
required to build the host binaries.

Signed-off-by: Chih-Wei Huang <cwhuang@linux.org.tw>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
9 years agoandroid: fix a building error of libmesa_program
Chih-Wei Huang [Thu, 2 Apr 2015 06:16:33 +0000 (14:16 +0800)]
android: fix a building error of libmesa_program

Add libmesa_glsl to LOCAL_STATIC_LIBRARIES to get
its exported include path (for nir_opcodes.h).

Signed-off-by: Chih-Wei Huang <cwhuang@linux.org.tw>
9 years agoandroid: mesa: fold the ARCH_X86_HAVE_SSE4_1 conditionals
Emil Velikov [Wed, 22 Apr 2015 14:53:11 +0000 (15:53 +0100)]
android: mesa: fold the ARCH_X86_HAVE_SSE4_1 conditionals

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
9 years agoandroid: mesa: fix the path of the SSE4_1 optimisations
Emil Velikov [Sat, 28 Mar 2015 18:23:01 +0000 (18:23 +0000)]
android: mesa: fix the path of the SSE4_1 optimisations

Commit dd6f641303c(mesa: Build with subdir-objects.) removed the SRCDIR
variable, but forgot to update all references of it.

v2: Fix path - must be relative to LOCAL_PATH. (Chih-Wei)

Cc: "10.5" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Chih-Wei Huang <cwhuang@linux.org.tw>
9 years agoandroid: build the Mesa IR -> NIR translator
Emil Velikov [Sat, 28 Mar 2015 18:12:23 +0000 (18:12 +0000)]
android: build the Mesa IR -> NIR translator

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
9 years agoandroid: nir: add build rules for nir_builder_opcodes.h
Emil Velikov [Wed, 22 Apr 2015 13:51:31 +0000 (14:51 +0100)]
android: nir: add build rules for nir_builder_opcodes.h

Missed out with commit 2a135c470e3(nir: Add an ALU op builder kind of
like ir_builder.h)

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
9 years agoandroid: add inital NIR build
Mauro Rossi [Fri, 27 Mar 2015 22:25:45 +0000 (22:25 +0000)]
android: add inital NIR build

Required by the i965 driver.

v2:
 - Split out the nir_builder_opcodes.h rules.
 - Do not unconditionally hide the python command - use $(hide)
 - Use LOCAL_EXPORT_C_INCLUDE_DIRS to manage includes for the generated
sources.

Cc: "10.5" <mesa-stable@lists.freedesktop.org>
[Emil Velikov: Split from a larger commit, v2]
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Chih-Wei Huang <cwhuang@linux.org.tw>
9 years agoandroid: dri: link against libmesa_util
Emil Velikov [Fri, 27 Mar 2015 21:39:15 +0000 (21:39 +0000)]
android: dri: link against libmesa_util

The dri modules depend on symbols provided by it.

Cc: "10.5" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Chih-Wei Huang <cwhuang@linux.org.tw>
9 years agoandroid: add $(mesa_top)/src/mesa/main to the includes list
Emil Velikov [Sat, 28 Mar 2015 01:11:54 +0000 (01:11 +0000)]
android: add $(mesa_top)/src/mesa/main to the includes list

Required by the format_{un,}pack rework. Otherwise the build will fail
to locate the respective headers - format_{un,}pack.h

Cc: "10.5" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Chih-Wei Huang <cwhuang@linux.org.tw>
9 years agoandroid: add HAVE__BUILTIN_* and HAVE_FUNC_ATTRIBUTE_* defines
Emil Velikov [Fri, 27 Mar 2015 20:10:35 +0000 (20:10 +0000)]
android: add HAVE__BUILTIN_* and HAVE_FUNC_ATTRIBUTE_* defines

All of those are available on gcc 4.5 and later with the current android
build using gcc 4.7.

Cc: "10.4 10.5" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Chih-Wei Huang <cwhuang@linux.org.tw>
9 years agoandroid: add gallium dirs to more places in the tree
Emil Velikov [Sat, 28 Mar 2015 01:10:52 +0000 (01:10 +0000)]
android: add gallium dirs to more places in the tree

Similar to e8c5cbfd921(mesa: Add gallium include dirs to more parts of
the tree.)

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Chih-Wei Huang <cwhuang@linux.org.tw>
9 years agoandroid: dri/common: conditionally include drm_cflags/set __NOT_HAVE_DRM_H
Emil Velikov [Fri, 27 Mar 2015 18:57:46 +0000 (18:57 +0000)]
android: dri/common: conditionally include drm_cflags/set __NOT_HAVE_DRM_H

Otherwise we'll fail to find the drm.h header.

Cc: "10.4 10.5" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
9 years agoandroid: egl: add libsync_cflags to the build
Emil Velikov [Fri, 27 Mar 2015 18:36:10 +0000 (18:36 +0000)]
android: egl: add libsync_cflags to the build

... via local_shared_libraries. Otherwise the sync/sync.h header won't
be found.

Note: 10.5 and earlier will need similar change in st/egl.

v2: Append the library to the local_shared_libraries list. (Chih-Wei)

Cc: "10.4 10.5" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Chih-Wei Huang <cwhuang@linux.org.tw>
9 years agoandroid: mesa: generate the format_{un,}pack.[ch] sources
Mauro Rossi [Fri, 27 Mar 2015 18:20:53 +0000 (18:20 +0000)]
android: mesa: generate the format_{un,}pack.[ch] sources

Missed out with commit e1fdcddafe9(mesa: Autogenerate format_unpack.c)

v2: Conditionaly print the python commands - s/@/$(hide) / (Chih-Wei)

Cc: "10.5" <mesa-stable@lists.freedesktop.org>
[Emil Velikov: Split our from a larger commit.]
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
9 years agoandroid: add $(mesa_top)/src include to the whole of mesa
Emil Velikov [Fri, 27 Mar 2015 16:13:50 +0000 (16:13 +0000)]
android: add $(mesa_top)/src include to the whole of mesa

Many parts of mesa already have the include with others depending on it
but it's missing. Add it once at the top makefile and be done with it.

Cc: "10.4 10.5" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Chih-Wei Huang <cwhuang@linux.org.tw>
9 years agoandroid: use := operator for assigning MESA_VERSION
Emil Velikov [Fri, 27 Mar 2015 14:18:24 +0000 (14:18 +0000)]
android: use := operator for assigning MESA_VERSION

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
9 years agoutil: android: optimize the rules to generate format_srgb.c
Chih-Wei Huang [Wed, 21 Jan 2015 17:21:00 +0000 (01:21 +0800)]
util: android: optimize the rules to generate format_srgb.c

Signed-off-by: Chih-Wei Huang <cwhuang@linux.org.tw>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
9 years agoandroid: simplify the subdirs including rules
Chih-Wei Huang [Wed, 21 Jan 2015 17:17:42 +0000 (01:17 +0800)]
android: simplify the subdirs including rules

Use the macro defined in the Android build system.

Signed-off-by: Chih-Wei Huang <cwhuang@linux.org.tw>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
9 years agoandroid: use LOCAL_SHARED_LIBRARIES over TARGET_OUT_HEADERS
Emil Velikov [Fri, 27 Mar 2015 11:40:42 +0000 (11:40 +0000)]
android: use LOCAL_SHARED_LIBRARIES over TARGET_OUT_HEADERS

... to manage the LIBDRM*_CFLAGS. The former is the recommended approach
by the Android build system developers while the latter has been
depreciated for quite some time.

Cc: "10.4 10.5" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
9 years agoilo: remove unused include from Android.mk
Emil Velikov [Fri, 27 Mar 2015 11:13:21 +0000 (11:13 +0000)]
ilo: remove unused include from Android.mk

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Chih-Wei Huang <cwhuang@linux.org.tw>
9 years agodrirc: Add "Second Life" quirk (allow_glsl_extension_directive_midshader).
Kenneth Graunke [Fri, 10 Apr 2015 17:24:33 +0000 (10:24 -0700)]
drirc: Add "Second Life" quirk (allow_glsl_extension_directive_midshader).

Appears to fix shader compilation.  Tested by starting the client,
dragging the "quality and speed" slider back and forth, and watching the
console output - instead of piles of "shader failed to compile", the CPU
seems to be busy compiling shaders.  I haven't actually tried to play.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=69226
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=71591
Cc: mesa-stable@lists.freedesktop.org
9 years agonir: Fix per-component negation in prog_to_nir's SWZ handling.
Kenneth Graunke [Fri, 17 Apr 2015 21:08:14 +0000 (14:08 -0700)]
nir: Fix per-component negation in prog_to_nir's SWZ handling.

I missed the fact that the ARB_fragment_program SWZ instruction allows
per-component negation.  To fix this, move Abs/Negate handling into both
the simple case and the SWZ case's per-component loop.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=90000
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agoglsl: correct indentation of comment, Trivial.
Tapani Pälli [Tue, 21 Apr 2015 17:11:43 +0000 (20:11 +0300)]
glsl: correct indentation of comment, Trivial.

Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
9 years agoi965/fs: Combine pixel center calculation into one inst.
Matt Turner [Tue, 14 Apr 2015 20:17:38 +0000 (13:17 -0700)]
i965/fs: Combine pixel center calculation into one inst.

The X and Y values come interleaved in g1 (.4-.11 inclusive), so we can
calculate them together with a single add(32) instruction on some
platforms like Broadwell and newer or in SIMD8 elsewhere.

Note that I also moved the PIXEL_X/PIXEL_Y virtual opcodes from before
LINTERP to after it. That's because the writes_accumulator_implicitly()
function in backend_instruction tests for <= LINTERP for determining
whether the instruction indeed writes the accumulator implicitly. The
old FS_OPCODE_PIXEL_X/Y emitted ADD instructions, which did, but the new
opcodes just emit MOVs, which don't. It doesn't matter, since we don't
use these opcodes on Gen4/5 anymore, but in the case that we do...

On Broadwell:
total instructions in shared programs: 7192355 -> 7186224 (-0.09%)
instructions in affected programs:     1190700 -> 1184569 (-0.51%)
helped:                                6131

On Haswell:
total instructions in shared programs: 6155979 -> 6152800 (-0.05%)
instructions in affected programs:     652362 -> 649183 (-0.49%)
helped:                                3179

Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
9 years agoi965/fs: Calculate delta_x and delta_y together.
Matt Turner [Tue, 7 Apr 2015 00:44:40 +0000 (17:44 -0700)]
i965/fs: Calculate delta_x and delta_y together.

This lets SIMD16 programs on G45 and Gen5 use the PLN instruction.

On Ironlake:

total instructions in shared programs: 5634757 -> 5518055 (-2.07%)
instructions in affected programs:     1745837 -> 1629135 (-6.68%)
helped:                                11439
HURT:                                  4

Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
9 years agoi965/fs: Emit ADDs for gl_FragCoord, not virtual opcodes.
Matt Turner [Thu, 2 Apr 2015 23:15:53 +0000 (16:15 -0700)]
i965/fs: Emit ADDs for gl_FragCoord, not virtual opcodes.

These were used only on Gen4 and 5. emit_interpolation_setup_gen6() emits
ADDs directly. The virtual opcodes weren't providing anything useful.

I'm going to repurpose these opcodes, so deleting and readding them makes
it simpler to see what's going on.

Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
9 years agoi965/fs: Manually set source regioning on PLN instructions.
Matt Turner [Tue, 7 Apr 2015 04:46:54 +0000 (21:46 -0700)]
i965/fs: Manually set source regioning on PLN instructions.

Like LINE (commit 92346db0), src0 must have a scalar region. Setting
src1's region to <8,8,1> lets us pass a properly sized combined delta_xy
argument in a few commits without getting a bogus <16,16,1> region.

Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
9 years agoi965/fs: Add LINTERP's src0 to fs_inst::regs_read().
Matt Turner [Sat, 11 Apr 2015 22:00:19 +0000 (15:00 -0700)]
i965/fs: Add LINTERP's src0 to fs_inst::regs_read().

LINTERP's src0 is PLN's src1, and PLN's src1 reads exec_size / 4
registers.

Having that information lets us drop the delta_x/y special case code in
split_virtual_grfs().

Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
9 years agoi965/fs: Set compression only if writing two registers.
Matt Turner [Sat, 11 Apr 2015 21:51:13 +0000 (14:51 -0700)]
i965/fs: Set compression only if writing two registers.

We don't want to set compression control on a SIMD16 instruction
operating on words or smaller.

Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
9 years agoi965/fs: Allow an execution size of 32.
Matt Turner [Sat, 4 Apr 2015 00:44:06 +0000 (17:44 -0700)]
i965/fs: Allow an execution size of 32.

In a few commits, we'll start emitting an add(32) instruction on some
platforms.

Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
9 years agoi965: Make type_sz() return unsigned.
Matt Turner [Sat, 11 Apr 2015 21:49:50 +0000 (14:49 -0700)]
i965: Make type_sz() return unsigned.

Avoids annoying warnings when comparing with sizeof(...).

Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
9 years agoi965: Replace guess_execution_size with something simpler.
Matt Turner [Tue, 14 Apr 2015 19:40:34 +0000 (12:40 -0700)]
i965: Replace guess_execution_size with something simpler.

guess_execution_size() does two things:

   1. Cope with small destination registers.
   2. Cope with SIMD8 vs SIMD16 mode.

This patch replaces the first with a simple if block in brw_set_dest: if
the destination register width is less than 8, you probably want the
execution size to match.  (I didn't put this in the 3src block because
it doesn't seem to matter.)

Since only the FS compiler cares about SIMD16 mode, it's easy to just
set the default execution size there.

This pattern was already been proven in the Gen8+ generator, but we
didn't port it back to the existing generator when we combined the two.

This is based on a patch from Ken from about a year ago. I've rebased it
and and fixed a few bugs.

Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
9 years agoi965/fs: Ensure delta_x/y are even-aligned registers on Gen6.
Matt Turner [Thu, 2 Apr 2015 23:57:10 +0000 (16:57 -0700)]
i965/fs: Ensure delta_x/y are even-aligned registers on Gen6.

The BSpec says this applies to Gen6 as well.

Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
9 years agomain: remove __FUNCTION__ defined because it is obsolete
Marius Predut [Thu, 16 Apr 2015 10:36:07 +0000 (13:36 +0300)]
main: remove __FUNCTION__ defined because it is obsolete

Consistently just use C99's __func__ everywhere.
No functional changes.

Signed-off-by: Marius Predut <marius.predut@intel.com>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
9 years agoradeon: replace __FUNCTION__ with __func__
Marius Predut [Thu, 16 Apr 2015 10:35:50 +0000 (13:35 +0300)]
radeon: replace __FUNCTION__ with __func__

Consistently just use C99's __func__ everywhere.
No functional changes.

Signed-off-by: Marius Predut <marius.predut@intel.com>
Acked-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
9 years agomesa: add missing break in switch statement
Tapani Pälli [Tue, 21 Apr 2015 05:26:03 +0000 (08:26 +0300)]
mesa: add missing break in switch statement

Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-By: Martin Peres <martin.peres@linux.intel.com>
9 years agoglsl: add fallthrough comment on switch
Tapani Pälli [Tue, 21 Apr 2015 05:25:16 +0000 (08:25 +0300)]
glsl: add fallthrough comment on switch

Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-By: Martin Peres <martin.peres@linux.intel.com>
9 years agomesa: fix UBO queries for active uniforms
Tapani Pälli [Mon, 20 Apr 2015 12:41:06 +0000 (15:41 +0300)]
mesa: fix UBO queries for active uniforms

Commit 34df5eb introduced regression to GetActiveUniformBlockiv
when querying one of the following properties:

   GL_UNIFORM_BLOCK_ACTIVE_UNIFORMS
   GL_UNIFORM_BLOCK_ACTIVE_UNIFORM_INDICES

Implementation counted all uniforms in ubo directly while query should
check first if the uniform in question is _active_.

Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=90109
Reviewed-By: Martin Peres <martin.peres@linux.intel.com>
9 years agoi965/skl: Fix the qpitch value
Neil Roberts [Wed, 18 Feb 2015 18:41:58 +0000 (18:41 +0000)]
i965/skl: Fix the qpitch value

On Skylake the qpitch value is uploaded as part of the surface state
so we don't need to add the extra rows that are done for other
generations. However for 3D textures it needs to be aligned to the
tile height and for depth/stencil textures it needs to be a multiple
of 8. Unlike previous generations the qpitch is measured as a multiple
of the block size for compressed surfaces. When the horizontal mipmap
layout is used for 1D textures then the qpitch is measured in pixels
instead of rows.

v2: Align the depth/stencil textures to a multiple of 8
v3: Add an assert that ALL_SLICES_AT_EACH_LOD is not used. Ignore the
    vertical alignment when picking the qpitch for 1D_ARRAY textures.

Reviewed-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Anuj Phogat <anuj.phogat@gmail.com>
9 years agoi965/skl: Don't use ALL_SLICES_AT_EACH_LOD
Neil Roberts [Fri, 20 Feb 2015 19:11:46 +0000 (19:11 +0000)]
i965/skl: Don't use ALL_SLICES_AT_EACH_LOD

The render surface state command for Skylake doesn't have the surface
array spacing bit so it's not possible to select this layout. I think
it was only used in order to make it pick a tightly-packed qpitch
value that doesn't include space for the mipmaps. However this won't
be necessary after the next patch because it will automatically pick a
packed qpitch value whenever first_level==last_level. It is better to
remove this layout entirely on Gen8+ because although it can
effectively be implemented with a small qpitch value when there are no
mipmaps it isn't possible to support the case where there are mipmaps
because in that case the layout is very different.

It could be good to make a similar change for Gen8 if we also change
the layouting code to pick the qpitch value in a similar way.

v2: Make the commit message and comments more convincing

Reviewed-by: Ben Widawsky <ben@bwidawsk.net>
Tested-by: Ben Widawsky <ben@bwidawsk.net>
9 years agoclover: remove pre llvm 3.5.0 compatibility code
EdB [Sun, 19 Apr 2015 08:50:19 +0000 (10:50 +0200)]
clover: remove pre llvm 3.5.0 compatibility code

Acked-by: Francisco Jerez <currojerez@riseup.net>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
9 years agoclover: make llvm >= 3.5.0 and c++11 mandatory
EdB [Mon, 20 Apr 2015 18:28:09 +0000 (20:28 +0200)]
clover: make llvm >= 3.5.0 and c++11 mandatory

Clover not longer compile with llvm <= 3.5.0 since e1d363b3.
e1d363b3 implies c++11 and llvm 3.5.0 CXXFLAGS provided it.
No one seems to have noticed it, it's now official.

Acked-by: Francisco Jerez <currojerez@riseup.net>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
9 years agodocs/GL3.txt: update ARB_shader_subroutine status
Dave Airlie [Mon, 20 Apr 2015 08:41:52 +0000 (18:41 +1000)]
docs/GL3.txt: update ARB_shader_subroutine status

Admit to having started working on this, I don't admit to ever finishing it

Signed-off-by: Dave Airlie <airlied@redhat.com>
9 years agogallivm: Fix build against LLVM 3.7 SVN r235265
Nick Sarnie [Mon, 20 Apr 2015 03:51:26 +0000 (23:51 -0400)]
gallivm: Fix build against LLVM 3.7 SVN r235265

LLVM removed JITEmitDebugInfo from TargetOptions since they weren't used

v2: Be consistent with the LLVM version check (Aaron Watry)

Signed-off-by: Nick Sarnie <commendsarnex@gmail.com>
Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com>
9 years agodoc: Add GL_ARB_shader_image_size dependency for OpenGL ES 3.1
Ian Romanick [Thu, 16 Apr 2015 19:17:19 +0000 (12:17 -0700)]
doc: Add GL_ARB_shader_image_size dependency for OpenGL ES 3.1

imageSize() is in the GLSL ES 3.1 spec.  Trivial.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
9 years agoindices: fix provoking vertex for quads/quadstrips
Ilia Mirkin [Tue, 7 Apr 2015 05:42:05 +0000 (01:42 -0400)]
indices: fix provoking vertex for quads/quadstrips

This allows drivers to provide consistent flat shading for quads.
Otherwise a driver that only supported tris would have to force last
provoking vertex when drawing quads (and would have to say that quads
don't follow the provoking vertex convention).

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Rob Clark <robclark@freedesktop.org>
9 years agoprimconvert: select pv convention only from flatshade_first
Ilia Mirkin [Tue, 7 Apr 2015 15:43:52 +0000 (11:43 -0400)]
primconvert: select pv convention only from flatshade_first

This should match to how drivers program hardware. flatshade relates to
whether color inputs are interpolated, not the provoking vertex
convention.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Rob Clark <robclark@freedesktop.org>
9 years agofreedreno/a3xx: enable polymode setting with non-fill modes
Ilia Mirkin [Tue, 7 Apr 2015 04:57:34 +0000 (00:57 -0400)]
freedreno/a3xx: enable polymode setting with non-fill modes

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
9 years agofreedreno/a3xx: fix integer and 32-bit float border colors
Ilia Mirkin [Mon, 6 Apr 2015 05:15:09 +0000 (01:15 -0400)]
freedreno/a3xx: fix integer and 32-bit float border colors

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
9 years agofreedreno/a3xx: add support for float R/RG render targets
Ilia Mirkin [Mon, 6 Apr 2015 01:38:57 +0000 (21:38 -0400)]
freedreno/a3xx: add support for float R/RG render targets

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
9 years agomesa: add .mesa-install-links files to gitignore
Connor Abbott [Fri, 17 Apr 2015 16:53:53 +0000 (12:53 -0400)]
mesa: add .mesa-install-links files to gitignore

Signed-off-by: Connor Abbott <cwabbott0@gmail.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
9 years agomesa/main: add autogenerated format-info.c to gitignore
Connor Abbott [Fri, 17 Apr 2015 16:50:30 +0000 (12:50 -0400)]
mesa/main: add autogenerated format-info.c to gitignore

v2: move to right after format-info.h

Signed-off-by: Connor Abbott <cwabbott0@gmail.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
9 years agoi965: Issue perf_debug messages for unsynchronized maps on !LLC systems.
Kenneth Graunke [Wed, 25 Feb 2015 05:34:30 +0000 (21:34 -0800)]
i965: Issue perf_debug messages for unsynchronized maps on !LLC systems.

We haven't implemented proper unsynchronized map support on !LLC systems
(pre-SNB, Atom).  MapBufferRange with GL_MAP_UNSYNCHRONIZE_BIT will
actually do a synchronized map, probably killing performance.

Also warn on BufferSubData, when we should be doing an unsynchronized
upload, but instead have to do a synchronous map.

v2: Only complain if the buffer is actually busy - we use unsynchronized
    maps internally for vertex upload and such, but expect those to not
    be busy.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ben Widawsky <ben@bwidawsk.net>
Tested-by: Ben Widawsky <ben@bwidawsk.net>
9 years agoi965: Make shader_time store names/ids instead of referencing shaders.
Kenneth Graunke [Wed, 15 Apr 2015 09:16:47 +0000 (02:16 -0700)]
i965: Make shader_time store names/ids instead of referencing shaders.

Jason noticed that shader_time was bumping the reference count on the
gl_shader_program and gl_program structures, in code called during
compilation.

Not only were these never unreferenced, but it meant fragment shaders
might be referenced twice (SIMD8 and SIMD16)...or only once.

We don't actually need the programs.  We just need their numeric ID and
their language (GLSL/ARB/FF) or KHR_debug label.  If there's a label, we
have to strdup it since the underlying program could be deleted.

To be fair, we're not exactly cleaning that up either, but we at least
ralloc it out of the shader_time arrays, so if we ever bother cleaning
those up, they'll go away properly.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
Acked-by: Anuj Phogat <anuj.phogat@gmail.com>
9 years agoi965: Delete some unnecessary code in brw_report_shader_time().
Kenneth Graunke [Wed, 15 Apr 2015 09:04:17 +0000 (02:04 -0700)]
i965: Delete some unnecessary code in brw_report_shader_time().

It is true that a gl_shader_program with ID 0 will be a fixed-function
fragment program; a gl_program with ID 0 but NULL gl_shader_program
means that it's a fixed-function vertex shader.

But that's not terribly interesting or relevant to what we're doing.
We just need to know that ID 0 means "fixed function".

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
Reviewed-by: Anuj Phogat <anuj.phogat@gmail.com>
9 years agoi965: Make shader_time use 0 instead of -1 for "no meaningful ID".
Kenneth Graunke [Wed, 15 Apr 2015 08:57:52 +0000 (01:57 -0700)]
i965: Make shader_time use 0 instead of -1 for "no meaningful ID".

0 is not a valid GLSL shader or ARB program ID.  For some reason,
shader_time used -1 instead...so we had code to detect 0, then override
it to -1.

We can just delete that.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
Reviewed-by: Anuj Phogat <anuj.phogat@gmail.com>
9 years agoadjust a couple of ifdefs to handle NetBSD correctly
Tobias Nygren [Fri, 17 Apr 2015 18:27:55 +0000 (20:27 +0200)]
adjust a couple of ifdefs to handle NetBSD correctly

Acked-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Tobias Nygren <tnn@NetBSD.org>
9 years agoconfigure.ac: fix bashism
Tobias Nygren [Fri, 17 Apr 2015 18:18:48 +0000 (20:18 +0200)]
configure.ac: fix bashism

Reviewed-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Tobias Nygren <tnn@NetBSD.org>
9 years agoi965: Render R16G16B16X16 as R16G16B16A16
Anuj Phogat [Fri, 10 Apr 2015 11:35:24 +0000 (04:35 -0700)]
i965: Render R16G16B16X16 as R16G16B16A16

This enables using _mesa_meta_pbo_TexSubImage() to upload data
to R16G16B16X16 texture. Earlier it fell back to slower paths.

Jenkins run shows no piglit regressions.

Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
9 years agoi965: Update the comment about platforms supporting blorp
Anuj Phogat [Mon, 13 Apr 2015 17:20:29 +0000 (10:20 -0700)]
i965: Update the comment about platforms supporting blorp

Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Chad Versace <chad.versace@intel.com>
9 years agonir: Allow abs/neg in select peephole pass.
Matt Turner [Thu, 2 Apr 2015 17:21:16 +0000 (10:21 -0700)]
nir: Allow abs/neg in select peephole pass.

total instructions in shared programs: 4314531 -> 4308949 (-0.13%)
instructions in affected programs:     429085 -> 423503 (-1.30%)
helped:                                1680
HURT:                                  0
GAINED:                                0
LOST:                                  111

Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
9 years agoi965/fs: Change SEL and MOV types as needed to propagate source modifiers
Jason Ekstrand [Fri, 3 Apr 2015 18:07:47 +0000 (11:07 -0700)]
i965/fs: Change SEL and MOV types as needed to propagate source modifiers

SEL and MOV instructions, as long as they don't have source modifiers, are
just copying bits around.  This commit adds support to copy propagation to
switch the type of a SEL or MOV instruction as needed so that it can
propagate source modifiers.  This is needed because NIR generates integer
SEL and MOV instructions whenver it doesn't know what else to generate.

shader-db results with NIR:
total FS instructions in shared programs: 4360910 -> 4360186 (-0.02%)
FS instructions in affected programs:     59094 -> 58370 (-1.23%)
helped:                                   341
HURT:                                     0
GAINED:                                   2
LOST:                                     0

Signed-off-by: Jason Ekstrand <jason.ekstrand@intel.com>
Reviewed-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agoi965/fs: Use the source type when looking for UD negations in copy prop
Jason Ekstrand [Fri, 3 Apr 2015 19:15:48 +0000 (12:15 -0700)]
i965/fs: Use the source type when looking for UD negations in copy prop

There can be problems with floats and conditional modifiers when
copy-propagating a negated UD source.  The problem arises when a source
modifier is applied to a UD value.  In this case, a 33-bit representation
is internally used.  If you do the following:

   1: mov foo:UD 7U
   2: mov bar:UD -foo:UD
   3: mov out:F bar:UD

the out register will have the value (float)(unt32_t)-7 which is some very
large floating-point number.  However, if we allow copy-propagation of the
second mov, we get

   1: mov foo:UD 7U
   3: mov out:f -bar:UD

and, since the negation is computed in 33-bits, we get a value of -7.0f
which is clearly not the same.  This is a similar problem if the
instruction has a conditional modifier where the 33-bit value is used in
the comparison and not the 32-bit version.

Previously, we checked the source to be copied for the negate and then
checked the source being propagated to for the type.  This isn't quite what
we want because we are really just looking for negated UD sources.  A check
later in the file ensures that both ends of the propagate have the right
type so it works.  However, if we relax the restriction that both ends of
the propagation have the same type, it ends up causing us to bail early in
cases we don't want.

Signed-off-by: Jason Ekstrand <jason.ekstrand@intel.com>
Reviewed-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agofreedreno/ir3/nir: few little fixes
Rob Clark [Thu, 16 Apr 2015 19:35:50 +0000 (15:35 -0400)]
freedreno/ir3/nir: few little fixes

isaml needs to scale up coords based on LoD.  Also fix bogus bary.f
varying # when there are non-bary frag shader inputs.  And use sub.s of
a positive immediate rather than add.s of negative (since CP is better
about figuring out that those can be collapsed into the cat2 instr).

Signed-off-by: Rob Clark <robclark@freedesktop.org>
9 years agofreedreno/ir3/nir: lower if/else
Rob Clark [Wed, 15 Apr 2015 14:16:31 +0000 (10:16 -0400)]
freedreno/ir3/nir: lower if/else

For now, completely flatten if/else blocks.  That will almost certainly
change once we have flow control.

Signed-off-by: Rob Clark <robclark@freedesktop.org>
9 years agofreedreno/a4xx: support for large shaders
Rob Clark [Sun, 12 Apr 2015 16:58:52 +0000 (12:58 -0400)]
freedreno/a4xx: support for large shaders

Signed-off-by: Rob Clark <robclark@freedesktop.org>
9 years agofreedreno: update generated headers
Rob Clark [Sun, 12 Apr 2015 18:25:29 +0000 (14:25 -0400)]
freedreno: update generated headers

Signed-off-by: Rob Clark <robclark@freedesktop.org>
9 years agofreedreno/ir3/nir: UBO support
Rob Clark [Sun, 12 Apr 2015 13:47:40 +0000 (09:47 -0400)]
freedreno/ir3/nir: UBO support

Signed-off-by: Rob Clark <robclark@freedesktop.org>