softpipe: add depth clamping support. (v2)

This passes the piglit depth clamp tests.

this is required for GL 3.2.

v2: move min/max up one level, could go further, thanks
to Roland for suggestion.

v1: Reviewed-by: Brian Paul <brianp@vmware.com>

Signed-off-by: Dave Airlie <airlied@redhat.com>

tgsi/gs: bound max output vertices in shader

This limits the number of emitted vertices to the shaders max output
vertices, and avoids us writing things into memory that isn't big
enough for it.

Reviewed-by: Zack Rusin <zackr@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

i965: Add GPU BLIT of texture image to PBO in Intel driver

Add Intel driver hook for glGetTexImage to accelerate the case of reading
texture image into a PBO.  This case gets huge performance gains by using
GPU BLIT directly to PBO rather than GPU BLIT to temporary texture followed
by memcpy.

No regressions on Piglit tests  with Intel driver.
Performance gain (1280 x 800 FBO, Ivybridge):
glGetTexImage + glMapBufferRange  with patch 1.45 msec
glGetTexImage + glMapBufferRange without patch 4.68 msec

v3: (by Kenneth Graunke)
 - Fix compile after Eric's change to drop the tiling argument
   to intel_miptree_create_for_bo.
 - Add GL_TEXTURE_3D to blacklisted texture targets to prevent Piglit
   regressions.
 - Squash in several whitespace and coding style fixes.

i965: Invalidate live intervals when inserting Gen4 SEND workarounds.

We need to invalidate the live intervals when inserting new
instructions.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Cc: mesa-stable@lists.freedesktop.org

i965: Don't use the head sentinel as an fs_inst in Gen4 workaround code.

When walking backwards, we want to stop at the head sentinel, which is
where scan_inst->prev->prev == NULL, not scan_inst->prev == NULL.

Fixes random crashes, as well as valgrind errors.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Cc: mesa-stable@lists.freedesktop.org

meta: Label the meta GLSL clear program.

Giving the meta clear program a meaningful name makes it easier to find
in output such as INTEL_DEBUG=fs or INTEL_DEBUG=shader_time.  We already
did so for integer programs, but neglected to label the primary program.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Matt Turner <mattst88@gmail.com>

i965/fs: Combine generate_math[12]_gen6 methods.

These used to call different math emitters (brw_math vs. brw_math2).
Now that they both call gen6_math, they're virtually identical.

When unrolling SIMD16 to multiple SIMD8 operations, we should take care
not to apply sechalf to brw_null_reg for src1.  Otherwise, we'd end up
with BRW_ARF_NULL + 1 as the register number, and I'm not sure if that's
valid.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>

i965/fs: Drop the generate_math[12]_gen7 methods.

These functions are basically identical, so we should combine them.
However, they're so trivial, we may as well just fold them into their
only call sites.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>

i965/vec4: Combine generate_math[12]_gen6 methods.

These are trivial to combine: we should just avoid checking the second
operand if it's brw_null_reg.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>

i965/vec4: Drop the generate_math2_gen7() method.

It's now a single line of code, so we may as well fold it into the
caller.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>

i965: Rename brw_math to gen4_math.

Usually, I try to use "brw" for functions that apply to all generations,
and "gen4" for dead end/legacy code that is only used on Gen4-5.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>

i965: Split Gen4-5 and Gen6+ MATH instruction emitters.

Our existing functions, brw_math and brw_math2, had unclear roles:

Gen4-5 used brw_math for both unary and binary math functions; it never
used brw_math2.  Since operands are already in message registers, this
is reasonable.

Gen6+ used brw_math for unary math functions, and brw_math2 for binary
math functions, duplicating a lot of code.  The only real difference was
that brw_math used brw_null_reg() for src1.

This patch improves brw_math2's assertions to allow both unary and
binary operations, renames it to gen6_math(), and drops the Gen6+ code
out of brw_math().

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>

i965: Make src_reg::equals() take a constant reference, not a pointer.

This is more typical C++ style.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>

i965: Don't set the "switch" flag on control flow instructions on Gen6+.

Thread switching on control flow instructions is a documented workaround
for Gen4-5 errata.  As far as I can tell, it hasn't been needed since
Sandybridge.  Thread switching is not free, so in theory this may help
performance slightly.

Flow control instructions with the "switch" flag cannot be compacted, so
removing it will make these instructions compactable.  (Of course, we
still have to implement compaction for flow control instructions...)

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Matt Turner <mattst88@gmail.com>

i965/fs: Allow CSE on math opcodes on Gen6+.

total instructions in shared programs: 2081469 -> 2081248 (-0.01%)
instructions in affected programs:     22606 -> 22385 (-0.98%)
No programs were hurt by this patch.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>

glsl: Remove unused include in expr.flatt.

Found with IWYU. Compile-tested on my Ivy-bridge system.

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Thomas Helland <thomashelland90@gmail.com>

glsl: Remove unused include in ir.cpp

Found with IWYU. Compile-tested on my Ivy-bridge system.

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Thomas Helland <thomashelland90@gmail.com>

glsl: Remove unused include from ir_constant_expression.cpp

Found with IWYU. Compile-tested on my Ivy-bridge system.

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Thomas Helland <thomashelland90@gmail.com>

glsl: Remove unused include from ir_basic_block.cpp

Found with IWYU. Compile-tested on my Ivy-bridge system.

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Thomas Helland <thomashelland90@gmail.com>

glsl: Remove unused include from hir_field_selection.cpp

Found with IWYU. Compile-tested on my Ivy-bridge system

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Thomas Helland <thomashelland90@gmail.com>

glsl: Remove unused include from glsl_symbol_table.h

Only function-defs use glsl_type so forward declare instead.
Compile-tested on my Ivy-bridge system.

IWYU also suggests removing #include <new>, and this compiles fine.
I'm not familiar enough with memory management in C/C++ that I feel
comfortable removing this. Insights would be appreciated.

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Thomas Helland <thomashelland90@gmail.com>

glsl: Remove unused include from glsl_types.cpp

Found with IWYU. Compile-tested on my Ivy-bridge system.
Added comment about core.h being used for MAX2.

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Thomas Helland <thomashelland90@gmail.com>

glsl: Remove unused include from builtin_variables.cpp

Found with IWYU. Compile-tested on my Ivy-bridge system.

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Thomas Helland <thomashelland90@gmail.com>

glsl: Remove unused include in ast_to_hir.cpp

Found with IWYU. Comment says it's for struct gl_extensions.
Grepping for gl_extensions shows no uses.
Tested by compiling on my Ivy-bridge system.

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Thomas Helland <thomashelland90@gmail.com>

glsl: Remove unused includes in link_uniform_block_active_visitor.h

Found with IWYU, compile-tested on my Ivy-bridge system.
This is not used in the header, and is included in the source.

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Thomas Helland <thomashelland90@gmail.com>

glsl: Remove unused includes in link_uniform_init.

Found with IWYU, confirmed with grepping for "hash" and "symbol".
No negative effects on compilation.

IWYU also reported core.h and linker.h could be removed,
but I'm unsure if those are false positives.

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Thomas Helland <thomashelland90@gmail.com>

i965: Replace open-coded linked list with exec_list.

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>

glsl: Add an exec_node_init() function, usable from C.

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>

glsl: Make foreach macros usable from C by adding struct keyword.

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>

glsl: Make exec_list members just wrap the C API.

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>

glsl: Make exec_node members just wrap the C API.

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>

glsl: Add C API for exec_list.

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>

glsl: Add C API for exec_node.

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>

glsl: Move definition of exec_list member functions out of the struct.

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>

glsl: Move definition of exec_node member functions out of the struct.

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>

r600g/compute: Use %u as the unsigned format

This fixes an issue when running cl-program-bitcoin-phatk
piglit test where some of the inputs have negative values

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>

r600g/compute: align items correctly

Now, items whose size is a multiple of 1024 dw won't leave
1024 dw between itself and the following item

The rest of the cases is left as it was

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>

r600g/compute: Cleanup of compute_memory_pool.h

Removed compute_memory_defrag declaration because it seems
to be unimplemented.

I think that this function would have been the one that solves
the problem with fragmentation that compute_memory_finalize_pending has.

Also removed comments that are already at compute_memory_pool.c

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>

r600g/compute: Tidy a bit compute_memory_finalize_pending

Explanation of the changes, as requested by Tom Stellard:

Let's take need after is calculated as
item->size_in_dw+2048 - (pool->size_in_dw - allocated)

BEFORE:
If need is positive or 0:
    we calculate need += 1024 - (need % 1024), which is like
        cealing to the nearest multiple of 1024, for example
        0 goes to 1024, 512 goes to 1024 as well, 1025 goes
        to 2048 and so on. So now need is always possitive,
        we do compute_memory_grow_pool, check its output
        and continue.

If need is negative:
    we calculate need += 1024 - (need % 1024), in this case
        we will have negative numbers, and if need is
        [-1024:-1] 0, so now we take the else, recalculate
        need as need = pool->size_in_dw / 10 and
        need += 1024 - (need % 1024), we do
        compute_memory_grow_pool, check its output and continue.

AFTER:
If need is positive or 0:
    we jump the if, calculate need += 1024 - (need % 1024)
        compute_memory_grow_pool, check its output and continue.

If need is negative:
    we enter the if, and need is now pool->size_in_dw / 10.
        Now we calculate need += 1024 - (need % 1024)
        compute_memory_grow_pool, check its output and continue.

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>

r600g/compute: Add more NULL checks

In this case, NULL checks are added to compute_memory_grow_pool,
so it returns -1 when it fails. This makes necesary
to handle such cases in compute_memory_finalize_pending
when it is needed to grow the pool

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>

r600g/compute: Adding checks for NULL after CALLOC

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>

r600g/compute: Fixing a typo and some indentation

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>

mesa: Fix substitution of large shaders

Signed-off-by: Cody Northrop <cody@lunarg.com>
Reviewed-by: Brian Paul <brianp@vmware.com>

configure: Only check for OpenCL without LLVM when the latter is certain

LLVM is enabled by default for some architectures, but the test was failing
before that.

Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
Cc: "10.1 10.2" <mesa-stable@lists.freedesktop.org>

r600g,radeonsi: implement PIPE_QUERY_TIMESTAMP_DISJOINT

v2 Marek: set the query result correctly

Signed-off-by: David Heidelberger <david.heidelberger@ixit.cz>
Signed-off-by: Marek Olšák <marek.olsak@amd.com>

configure: Always default to --enable-driglx-direct

Always default to --enable-driglx-direct, now that will build driswrast, but
won't try to use dri[123] on platforms which don't have that.

Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>

glx: Fix build in GLX_DIRECT_RENDERING !GLX_USE_APPLEGL !GLX_USE_DRM case

Some untangling to fix building in the dri_platform=none, --enable-driglx-direct
case, where only driswast can be used.

Turn the test for including the glXGetScreenDriver()/glXGetScreenDriver()
interface used by xdriinfo from !GLX_USE_APPLEGL into a positive form, as it is
only useful when dri_platform=drm

Add additional GLX_USE_DRM tests so DRI[123] renderers are only used when
dri_platform=drm

Note that swrast and indirect must still be disabled in the APPLEGL case at the
moment, which makes things more complex than they need to be.  More untangling
is needed to allow that

Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>

i965: Make gen7_pi field of brw_instruction use unsigned instead of GLuint

Nothing else uses GL-types here.

Signed-off-by: Kristian Høgsberg <krh@bitplanet.net>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>

i965: Don't include mtypes.h in brw_disasm.c

It's not used.

Signed-off-by: Kristian Høgsberg <krh@bitplanet.net>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>

i965/fs: initialize src as reg_undef for texture opcodes on Gen4.

Untested.

i965/fs: initialize src as reg_undef for texture opcodes on Gen5/6.

Commit 07af0ab changed fs_inst to have 0 sources for texture opcodes
in emit_texture_gen5 (Ironlake, Sandybrige) while fs_generator still
uses a single source from brw_reg struct. Patch sets src as reg_undef
which matches the behavior before the constructor got changed.

Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=79534

egl/dri2: do not leak dri2_dpy->driver_name

Originally all hardware drivers duplicate the driver_name string
from an external source, while for the software rasterizer we set
it to "swrast". Follow the example set by hw drivers this way
we can free the string at dri2_terminate().

v2: Use strdup over strndup. Suggested by Ilia Mirkin.
v3: Handle platform_drm in a similar manner. Cleanup swrast
driver_name in error path.

Cc: Chia-I Wu <olv@lunarg.com>
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>

egl/dri2/x11: use standard strndup function

Using a custom version of the function brings no benefit.

Cc: Chad Versace <chad.versace@linux.intel.com>
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>

android, dricore: undefined reference to _mesa_streaming_load_memcpy

_mesa_streaming_load_memcpy is defined in main/streaming-load-memcpy.c
I'm adding it to the dricore lib

Cc: "10.1 10.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Adrian Negreanu <adrian.m.negreanu@intel.com>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>

android, mesa_gen_matypes: pull in timespec POSIX definition

This fixes:
  include/c11/threads_posix.h: In function 'cnd_timedwait':
  include/c11/threads_posix.h:140:21: error: storage size of 'abs_time' isn't known

Cc: "10.1 10.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Adrian Negreanu <adrian.m.negreanu@intel.com>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>

android, egl: typo dri2_fallback_pixmap_surface -> dri2_fallback_create_pixmap_surface

I used commit bc8b07a6 as reference, and only the droid_display_vtbl had this issue.

This fixes:
src/egl/drivers/dri2/platform_android.c:641:29:
  error: 'dri2_fallback_pixmap_surface' undeclared here (not in a function)

Cc: "10.1 10.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Adrian Negreanu <adrian.m.negreanu@intel.com>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>

android, egl: add correct drm include for libmesa_egl_dri2

Fixes:
  src/egl/drivers/dri2/platform_android.c:38:
  include/GL/internal/dri_interface.h:51:17:
    fatal error: drm.h: No such file or directory

Cc: "10.1 10.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Adrian Negreanu <adrian.m.negreanu@intel.com>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>

android: add src/gallium/auxiliary as include path for libmesa_dricore

This fixes:
In file included from
/home/adrian/workspace/mesa/mesa-master.git/src/mesa/vbo/vbo_exec_api.c:445:0:
/home/adrian/workspace/mesa/mesa-master.git/src/mesa/vbo/vbo_attrib_tmp.h:28:38:
fatal error: util/u_format_r11g11b10f.h: No such file or directory

Cc: "10.1 10.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Adrian Negreanu <adrian.m.negreanu@intel.com>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>

android: add libloader to libGLES_mesa and libmesa_egl_dri2

This fixes
  src/egl/drivers/dri2/platform_android.c:664: error: undefined reference to 'loader_set_logger'
  src/egl/drivers/dri2/platform_android.c:678: error: undefined reference to 'loader_get_driver_for_fd'

Cc: "10.1 10.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Adrian Negreanu <adrian.m.negreanu@intel.com>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>

android: adapt to the megadriver mechanism

Fixes linker error:
  ld:
  .../libmesa_dri_common_intermediates/libmesa_dri_common.a(dri_util.o):
    in function globalDriverAPI:dri_util.c(.data.rel+0x0): error:
    undefined reference to 'driDriverAPI'

As an example, you can see that mesa_dri_drivers
also uses common/libmegadriver_stub (src/mesa/drivers/dri/Makefile.am)

The _stub part might be confusing, but
it actually provides the dri-driver shared lib constructor,
megadriver_stub_init, which will later on load the real
platform dependent part and call
l __driDriverGetExtensions_<platform>

Cc: "10.1 10.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Adrian Negreanu <adrian.m.negreanu@intel.com>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>

add megadriver_stub_FILES

So that android part can also use $(megadriver_stub_FILES)

Cc: "10.1 10.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Adrian Negreanu <adrian.m.negreanu@intel.com>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>

scons: remove dri-i915 build target

Unmaintained and broken.

Cc: Jakob Bornecrantz <jakob@vmware.com>
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Acked-by: Jakob Bornecrantz <jakob@vmware.com>

configure: error out when building opencl without LLVM

Cc: Tom Stellard <thomas.stellard@amd.com>
Cc: "10.1 10.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>

i965/disasm: Properly debug negate source modifier for logical instructions

Reviewed-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Abdiel Janulgue <abdiel.janulgue@linux.intel.com>

i965/vec4: skip copy-propate for logical instructions with negated src entries

The negation source modifier on src registers has changed meaning in Broadwell when
used with logical operations. Don't copy propagate when negate src modifier is set
and when the destination instruction is a logical op.

Reviewed-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Abdiel Janulgue <abdiel.janulgue@linux.intel.com>

i965/fs: skip copy-propate for logical instructions with negated src entries

The negation source modifier on src registers has changed meaning in Broadwell when
used with logical operations. Don't copy propagate when negate src modifier is set
and when the destination instruction is a logical op.

Reviewed-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Abdiel Janulgue <abdiel.janulgue@linux.intel.com>

i965/fs: Refactor check for potential copy propagated instructions.

Reviewed-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Abdiel Janulgue <abdiel.janulgue@linux.intel.com>

docs: add link to 10.1.5 on news page

docs: fix version number in 10.2.1 release notes

docs: import the 10.1.5 release notes

glsl: Validate aux storage qualifier combination with other qualifiers.

We've been allowing `centroid` and `sample` in all kinds of weird places
where they're not valid.

Insist that `sample` is combined with `in` or `out`;
and that `centroid` is combined with `in`, `out`, or the deprecated
`varying`.

V2: Validate this in a more sensible place. This does require an extra
case for uniform blocks members and struct members, though, since they
don't go through the normal path.

V3: Improve error message wording; eliminate redundant error generation
for inputs in VS or outputs in FS.

Signed-off-by: Chris Forbes <chrisf@ijw.co.nz>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

i965: Ensure that we end instruction streams properly.

Threads must terminate with a SEND message to a particular shared function,
such as a URB write or FB write, so the instruction stream really shouldn't
ever end in an IF/ELSE/ENDIF or similar block structure.

However, if the instruction stream (incorrectly) ends in a block structure
the last block's end pointer will not be set, leading to a crash later on in
fs_live_variables::setup_def_use(). It is better to detect this earlier, so
assert on that.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

i965/fs: Add Gen < 6 runtime checks for line antialiasing.

In Gen < 6 the hardware generates a runtime bit that indicates whether AA data
has to be sent as part of the framebuffer write SEND message. This affects the
specific case where we have setup antialiased line rendering and we render
polygons which have one face setup in GL_LINE mode (line antialiasing
will be used) and the other one in GL_FILL mode (no line antialiasing needed).

Currently we are not doing this runtime test and instead we always send AA
data, which produces incorrect rendering of the GL_FILL face of the polygon in
in the aforementioned scenario (verified in ironlake and gm45).

In Gen4 this is, likely, a regression introduced with commit 098acf6c843. In
Gen5 this has never worked properly. Gen > 5 are not affected by this.

The patch fixes the problem by adding the appropriate runtime check and
adjusting the framebuffer write message accordingly in the conflictive
scenario.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=78679

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

i965/fs: Let the gen < 8 generator know about runtime_check_aads_emit

In gen < 6 we need to produce conditional code based on this flag when doing
framebuffer writes.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

docs: Mark off ARB_compressed_texture_pixel_storage

.. and add to release notes for 10.3

Signed-off-by: Chris Forbes <chrisf@ijw.co.nz>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>

mesa: Add extension enable for ARB_compressed_texture_pixel_storage

Signed-off-by: Chris Forbes <chrisf@ijw.co.nz>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>

mesa: Add pixel storage support for GetCompressedTexImage

Signed-off-by: Chris Forbes <chrisf@ijw.co.nz>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>

mesa: Compute proper strides for compressed texture pixel storage.

Signed-off-by: Chris Forbes <chrisf@ijw.co.nz>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>

mesa: Extract computation of compressed pixel store params

This logic is reusable across CompressedTex*Image* and
GetCompressedTexImage; the strides calculated will also be needed
in the PBO validation functions to ensure that the referenced range of
bytes is valid.

Signed-off-by: Chris Forbes <chrisf@ijw.co.nz>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>

mesa: Emit errors for inconsistent compressed pixel store state

V2: Use bool rather than GLboolean for internal function

Signed-off-by: Chris Forbes <chrisf@ijw.co.nz>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>

mesa: Add new pixel pack/unpack state for
ARB_compressed_texture_pixel_storage

Signed-off-by: Chris Forbes <chrisf@ijw.co.nz>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>

tests: Add new enum strings for ARB_compressed_texture_pixel_storage

Signed-off-by: Chris Forbes <chrisf@ijw.co.nz>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>

glapi: Add XML infrastructure for ARB_compressed_texture_pixel_storage

Signed-off-by: Chris Forbes <chrisf@ijw.co.nz>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>

mesa: Make CompressedTexSubImage errors more consistent

Signed-off-by: Chris Forbes <chrisf@ijw.co.nz>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>

mesa: Trim down PixelStorei implementation

Move _mesa_error call for INVALID_VALUE to one place.
Remove checks for previous value matching -- this was important when we
were flushing vertices before the update, but that hasn't happened for a
long time now.

Signed-off-by: Chris Forbes <chrisf@ijw.co.nz>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>

mesa/main: Prevent sefgault on glGetIntegerv(GL_ATOMIC_COUNTER_BUFFER_BINDING).

A recent ApiTrace change, that tries to dump more buffer state
causes Mesa from my distro (10.1.4) to segfaults here.

I haven't actually confirm this fixes it (I can't repro on master),
but it seems a good idea to be defensive here anyway.

Cc: "10.1 10.2" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

Revert "i965: Move brw_land_fwd_jump() to compilation unit of its use."

This reverts commit f3cb2e6ed7059b22752a6b7d7a98c07ba6b5552e.

brw_land_fwd_jump() is convenient wherever we produce JMPI instructions
and we will use JMPI to implement framebuffer writes that involve line
antialiasing in gen < 6.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

i965: Fix else and brace placement in brw_eu_emit.c.

I'm making a lot of changes to this area, and I figured I may as well
not conflate these trivial changes.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>

i965: Drop the remaining default predication whacking.

With my earlier cleaning in place (see git log brw_eu_emit.c), nothing
relies on the instruction emitters for IF/WHILE/JMPI disabling
predication.  Drop it in favor of making callers do the right thing
explicitly.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>

i965/sf: Use brw_set_default_predicate_control().

This is a bit tidier than poking at p->current directly.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>

gk110/ir: fix bfind emission

There is a short-immediate version as well, but it should never end up
getting used since it would have gotten folded earlier.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Cc: "10.2" <mesa-stable@lists.freedesktop.org>

docs: Add MD5 checksum, etc. for 10.2.1 release

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
(cherry picked from commit 70ce1031e7947c28133cb90211ecd6866c079d8b)

docs: Add initial 10.2.1 release notes

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
(cherry picked from commit 8c4845d29b21ec902c59d84e51d95639f0709085)

configure.ac: Do not use Pthreads with MinGW.

Match the behavior of the SCons MinGW build.

This patch also fixes these build errors.

  CC       glapi_entrypoint.lo
glapi_entrypoint.c: In function 'init_glapi_relocs_once':
glapi_entrypoint.c:341:4: error: unknown type name 'pthread_once_t'
    static pthread_once_t once_control = PTHREAD_ONCE_INIT;
    ^
glapi_entrypoint.c:341:41: error: 'PTHREAD_ONCE_INIT' undeclared (first use in this function)
    static pthread_once_t once_control = PTHREAD_ONCE_INIT;
                                         ^
glapi_entrypoint.c:341:41: note: each undeclared identifier is reported only once for each function it appears in
glapi_entrypoint.c:342:4: error: implicit declaration of function 'pthread_once' [-Werror=implicit-function-declaration]
    pthread_once( & once_control, init_glapi_relocs );
    ^

Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Matt Turner <mattst88@gmail.com>

gk110/ir: fix emitting constbuf file index

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Cc: "10.2" <mesa-stable@lists.freedesktop.org>

docs: Add MD5 checksum, etc. for 10.1 release

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
(cherry picked from commit 28d41e409d74aabe35f1031e759179e82cb760b4)

gk110/ir: emit saturate flag on fadd when needed

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Cc: "10.2" <mesa-stable@lists.freedesktop.org>

gk110/ir: fix slct emission

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Cc: "10.2" <mesa-stable@lists.freedesktop.org>

st/mesa: remove extra calculation of sampler count

This code was originally introduced to fix https://bugs.freedesktop.org/show_bug.cgi?id=53617. The comment says you need to pass NULL in order to unref old views however cso_set_sampler_views() already takes care of old views with the second for loop. Also as of 2355a6441435b8e66a032c44f0794066338e30a3 cso_set_sampler_views() passes the max of the old and new views to the driver for all state trackers making this code obsolete.

Signed-off-by: Timothy Arceri <t_arceri@yahoo.com.au>
Reviewed-by: Brian Paul <brianp@vmware.com>

gk110/ir: fix interp mode emission

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Cc: "10.2" <mesa-stable@lists.freedesktop.org>