radeonsi: Fix LLVM context leak.

Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>

gallium/radeon: Don't assign virtual address space for BO that already has one.

We'd end up re-using the old one and throwing away the new one anyway, but only
after a roundtrip to the kernel.

Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>

gallium/radeon: Create hole for waste when allocating from va_offset.

Otherwise, the wasted area could never be used for an allocation again.

Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>

gallium/radeon: Fix potential address space loss in radeon_bomgr_force_va().

Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>

gallium/radeon: Delete uppermost virtual address space hole if it's at the top.

Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>

gallium/radeon: Fix losing holes when allocating virtual address space.

If a hole exactly matches the allocated size plus alignment, we would fail to
preserve the alignment as a hole. This would result in never being able to use
the alignment area for an allocation again.

Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>

gallium/radeon: Merge holes when freeing virtual address space.

Otherwise we'll likely end up with an ever increasing amount of ever smaller
holes.

Requires keeping the list ordered wrt offsets.

Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>

gallium/radeon: Make va_offset 64 bits wide.

Otherwise we'd wrap around after 32 bits. The kernel currently limits GPU
virtual address space to 4GB anyway, but that will probably change sooner or
later, and this would result in confusing error messages when running out of
virtual address space even now.

Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>

llvmpipe: Silence Coverity incorrect sizeof expression defect.

Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: José Fonseca <jfonseca@vmware.com>

scons: Add option to enable floating-point textures.

Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: José Fonseca <jfonseca@vmware.com>

glx/dri2: add dri2 prime support.

This adds support for having libGL pick a different driver for prime support.

DRI_PRIME env var is set to the value retrieved from the server randr
provider calls, by the calling process. (generally DRI_PRIME=1 will be
the right answer).

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

radeon/llvm: Enable if-cvt

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

radeon/llvm: Add callbacks needed by if-cvt

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

radeon/llvm: Lower branch/branch_cond into predicated jump

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

radeon/llvm: Add a predicated JUMP instruction

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

radeon/llvm: Support for predicate bit

Tom Stellard:
  - A few changes to predicate register defs

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

r600g: Glue to handle predicate aware output from llvm

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

r600g: Fix instruction group merge when there are predicated insts.

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

radeon/llvm: Do not use PV/PS if PRED_SEL does not match

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

r600g: Add support for predicates

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

radeonsi: move ps sampler state into PM4 stream

Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>

radeonsi: move ps sampler views into PM4 stream

Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>

radeonsi: move vertex state descriptors into PM4 stream

Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>

radeonsi: add shader data infrastructure

With this we can embed data for the shaders (like resource
descriptors) into the PM4 stream.

Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>

radeon/llvm: add support to fetch temps as vectors

Necessary for texture fetches with temp regs as source on SI.

Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>

radeon/llvm: Remove AMDGPUUtil.cpp

radeon/llvm: Cleanup AMDGPUUtil.cpp

radeon/llvm: Lower loads from USE_SGPR adddress space during DAG lowering

radeon/llvm: Add live-in registers during DAG lowering

Psuedo instructions emulating live-in registers have been removed
and their corresponding intrinsics are now being lowered during DAG
lowering.

radeon/llvm: Lower store_output intrinsic during DAG lowering

radeon/llvm: Force VTX_READ instructions to use same reg for src and dst

I was seeing some GPU hangs that seemed to be cause by ALU instructions
writing to the same register used as the source for VTX_READ.  Adding
this constraint to the VTX_READ instructions avoids this situation.

radeonsi: fix build breakage after u_blitter changes

gallium/u_blitter: document custom meta helpers

r600g: disable handling of DISCARD_RANGE

https://bugs.freedesktop.org/show_bug.cgi?id=53130

r600g: implement timestamp query and get_timestamp hook

Reviewed-by: Alex Deucher <alexander.deucher@amd.com>

r600g: enable MSAA on evergreen by default

v2: add the DRM version check

r600g: implement copying between MSAA textures

r600g: implement MSAA color resolve

r600g: implement MSAA depth-stencil decompression and resolve

and integer textures, which are resolved the same as depth, I think.

r600g: implement TXQ_LZ opcode

r600g: implement MSAA rendering and texturing for evergreen and cayman

r600g: implement set_sample_mask

r600g: implement alpha-to-coverage

r600g: implement alpha-to-one

r600g: remove support for 3-channel colorbuffers

We have no sampler support for them.

configure.ac: bump libdrm_radeon requirement to 2.6.38

winsys/radeon: print error if CS is overflowed

and don't submit the CS to the kernel.

gallium/u_blitter: implement X and Y texture flipping

gallium/u_blitter: implement blitting multisample resources

It can blit only one sample at a time (it should be called in a loop).

gallium: add TGSI support for multisample textures

The only allowed instructions are TXQ_LZ and TXF.

TXQ_LZ is like TXQ, but without the LOD parameter (which is always zero
with MSAA textures)

The 3rd or the 4th texcoord component in TXF should contain the sample index
for a 2D_MSAA or 2D_ARRAY_MSAA texture, respectively.

gallium/tgsi: fix TGSI text parser

The problem was that the string matching succeeded e.g. for "2D" when there
was actually "2D_MSAA" and then failed parsing "_MSAA".

To prevent similar failures in the future, let's fix this kind of error
everywhere.

gallium/u_blit: set dst format from pipe_resource, not pipe_surface

We use it to decide whether we can use resource_copy_region.

NOTE: This is a candidate for the 8.0 branch.

gallium: make pipe_box signed in order to represent flipped blits

This will be used by u_blitter.

st/mesa: don't clamp fragment color with integer colorbuffer

mesa: flush vertices in test_framebuffer_completeness

st/egl: Fix up for ClientVersion -> ClientMajorVersion rename.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=53513

Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>

i965: add ARB_texture_rgb10_a2ui support

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Brian Paul <brianp@vmware.com>

meta: allow CopyTexSubImage on integer formats

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Brian Paul <brianp@vmware.com>

mesa ReadPixels: handle signed/unsigned integer clamping

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Brian Paul <brianp@vmware.com>

mesa pack: handle packed integer formats with clamping

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Brian Paul <brianp@vmware.com>

mesa unpack: call _mesa_problem when unpack function is not available

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Brian Paul <brianp@vmware.com>

mesa texstore: handle signed/unsigned integer clamping

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Brian Paul <brianp@vmware.com>

mesa GetTexImage: handle signed/unsigned integer clamping

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Brian Paul <brianp@vmware.com>

mesa pack: handle uint and int clamping properly

Rename _mesa_pack_rgba_span_int to _mesa_pack_rgba_span_from_uints.
Add _mesa_pack_rgba_span_from_ints.

These separate routines allow the integer clamping to be handled
properly for signed versus unsigned integers.

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Brian Paul <brianp@vmware.com>

intel: Fix rendering to a multisample front buffer

We need to downsample before flushing BUFFER_FAKE_FRONT_LEFT to
BUFFER_FRONT_LEFT in intel_flush_front.

Reviewed-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>

intel: Clean up intel_flush_front

Stop repeating ourselves. Replace the 4 instances of
`driContext->driDrawablePriv` with `driDrawable`.

Reviewed-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>

intel: Refactor intel_downsample_for_dri2_flush

Move it from intel_screen.c to intel_context.c. Redeclare as non-static.
A future commit will use it in multiple files.

Reviewed-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>

docs: Add EGL extensions to release notes

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

egl: Allow OpenGL ES 3.0 as a version

In the DRI2 back-end this will get the same API as GLES 2.0.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

dri2: Note that __DRI_API_GLES2 is also used for OpenGL ES 3.0

Unlike 1.x to 2.0, OpenGL ES 3.0 is backwards compatible with 2.0.  Use the
same API flag for both.  Applications that specifically want 3.0 will specify
this using the major / minor version attributes.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

egl_dri2: Add support for EGL_KHR_create_context and EGL_EXT_create_context_robustness

Just like in GLX, EGL_KHR_create_context requires DRI2 version >= 3, and
EGL_EXT_create_context_robustness requires both DRI2 version >= 3 and the
__DRI2_ROBUSTNESS extension.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

egl: Implement front-end support for EGL_EXT_create_context_robustness

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

egl: Implement front-end support for EGL_KHR_create_context

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

egl_dri2: Silence warnings about missing initializers

egl_dri2.c: At top level:
egl_dri2.c:325:4: warning: missing initializer [-Wmissing-field-initializers]
egl_dri2.c:325:4: warning: (near initialization for 'swrast_driver_extensions[2].version') [-Wmissing-field-initializers]
egl_dri2.c:330:4: warning: missing initializer [-Wmissing-field-initializers]
egl_dri2.c:330:4: warning: (near initialization for 'swrast_core_extensions[1].version') [-Wmissing-field-initializers]

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

egl: Rename ClientVersion to ClientMajorVersion, add ClientMinorVersion

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

egl_dri2: Use createContextAttribs if DRI2 version >= 3

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

egl_dri2: Require DRI2 version 2

The extra block in dri2_create_context is to prevent extra white space noise
in the next patch.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

dri_util: Compare against the correct API enums

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

mesa: Enable GL_ARB_invalidate_subdata

v2: Add GL_ARB_invalidate_subdata to release notes at Brian's
suggestion.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Brian Paul <brianp@vmware.com>

mesa: Add skeleton implementations of glInvalidateTex{Sub,}Image

These are part of GL_ARB_invalidate_subdata (but not OpenGL ES 3.0).

v2: Add comment explaining why minimum dimensions are set to 1 for some
texture targets.  Add default case to switch statement to silence
compiler warnings and detect new texture targets.  Both changes
suggested by Brian.  Also use _mesa_is_desktop_gl as suggested by Ken.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

mesa: Add skeleton implementations of glInvalidateBuffer{Sub,}Data

These are part of GL_ARB_invalidate_subdata (but not OpenGL ES 3.0).

v2: Use _mesa_bufferobj_mapped instead of testing
gl_buffer_object::Pointer as suggested by Brian.  Also use
_mesa_is_desktop_gl as suggested by Ken.

v3: Add a comment by the map subrange / discard range overlap test and
fix an off-by-one error noticed by Ken.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

mesa/es: Pass context to _mesa_init_bufferobj_dispatch

With this change _mesa_init_bufferobj_dispatch won't set function
pointers that don't exist in OpenGL ES.

v2: Use _mesa_is_desktop_gl and _mesa_is_gles3 as suggested by Ken.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

mesa: Add skeleton implementations of glInvalidate{Sub,}Framebuffer

These are part of GL_ARB_invalidate_subdata and OpenGL ES 3.0.

v2: Reject aux buffers in core context, and use _mesa_is_desktop_gl and
_mesa_is_gles3.  Both suggested by Ken.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

glapi: Add GL_ARB_invalidate_subdata

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Brian Paul <brianp@vmware.com>

mesa/es3: Add _mesa_is_gles3 predicate

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

intel: Implement ARB_texture_storage

This is basically cut-and-paste from the swrast implementation, and it
could probably be (slightly) more optimal.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

mesa: update glext.h to version 83

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Brian Paul <brianp@vmware.com>

build: Use MKDIR_P in src/mesa/Makefile.am

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

build: Use AM_V_GEN in src/mesa/Makefile.am

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

build: Fix autogen.sh to allow out-of-tree builds

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

build: Fix out-of-tree generation of builtin_function.cpp

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

build: Fix gtest out-of-tree build

Introduced by 3d000e7dd.

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

build: Fix out-of-tree generation of api_exec_es{1,2}.c

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

build/sources.mak: Add src/glsl/glcpp to INCLUDE_DIRS

Fixes problem where libdricore's of-out-tree build couldn't find
glcpp.h.

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

build/sources.mak: Remove unused GLSL_LIBS

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

mesa: Kill GL_ARB_shadow_ambient with fire

No driver supports this extension, and it seems unlikely than any driver
ever will.  I think r300c may have supported it at one time, but that
driver has already been removed.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Marek Olšák <maraeo@gmail.com>

radeon/llvm: Inline immediate offset when lowering implicit parameters

radeon/llvm: Use correct opcocde for BREAK_LOGICALNZ_i32

scons: Populate top_srcdir and top_builddir variables when reading Makefiles.sources.

This is not entirely correct, as scons doesn't put binaries in a
"src" subdirectory, but doesn't seem to be a problem for now.

mesa: Use GLdouble for depthMax in final unpack conversions.

The final step of _mesa_unpack_depth_span is to take the temporary
GLfloat depth values and convert them to the desired format.  When
converting to GL_UNSIGNED_INTEGER with depthMax > 0xffffff, we use
double-precision math to avoid overflow and precision problems.

Or at least that's the idea.  Unfortunately

   GLdouble z = depthValues[i] * (GLfloat) depthMax;

actually causes single-precision multiplication, since both operands are
GLfloats.  Casting depthMax to GLdouble causes the scaling to be done
with double-precision math.

Fixes a regression in oglconform's depth-stencil basic.read.ds test
since c60ac7b17993d28af65b04f9bbbf3ee74c35358c, where the expected and
actual values differed slightly.  For example, 0xcfa7a6 vs. 0xcfa7a4.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=49772
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>