meta: Use _mesa_unpack_format_to_base_format() to handle integer formats

Replace a call to mesa_base_tex_format() that handles only internal
formats with a call to the new _mesa_unpack_format_to_base_format()
function that handles allowed unpack formats and does not care for
internal formats at all.

Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>

mesa: Add a helper function _mesa_unpack_format_to_base_format()

Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>

mesa: Set green, blue channels to zero only for formats with these components

This is an optimization which avoids setting pixel transfer operations
when not required. _mesa_ReadPixels falls back to slower path if
transfer operations are set.

Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>

meta: Don't do fragment color clamping in _mesa_meta_pbo_GetTexSubImage

_mesa_meta_pbo_GetTexSubImage() uses _mesa_meta_BlitFrameBuffer(),
which will do fragment clamping if enabled. But fragment clamping
doesn't affect ReadPixels and GetTexImage.

Without this patch, piglit test arb_color_buffer_float-clear fails,
when forced to use the meta pbo path.

v2: Apply this fix to both glReadPixels and glGetTexImage.

Cc: <mesa-stable@lists.freedesktop.org>
Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>

meta: Abort meta pbo path if readpixels need signed-unsigned conversion

Meta pbo path for ReadPixels rely on BlitFramebuffer which doesn't support
signed to unsigned integer conversions and vice versa.

Without this patch, piglit test fbo_integer_readpixels_sint_uint fails, when
forced to use the meta pbo path.

v2: Make need_signed_unsigned_int_conversion() a static function. (Iago)
    Bump up the comment and the commit message. (Jason)

Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Cc: <mesa-stable@lists.freedesktop.org>
Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
Reviewed-by: Iago Toral <itoral@igalia.com>

meta: Fix transfer operations check in meta pbo path for readpixels

Currently used ctx->_ImageTransferState check is not sufficient
because it doesn't include the read color clamping enabled with
GL_CLAMP_READ_COLOR. So, use the helper function
_mesa_get_readpixels_transfer_ops().

Also, transfer operations don't affect glGetTexImage(). So, do
the check only for glReadPixles.

Without this patch, arb_color_buffer_float-readpixels test fails, when
forced to use meta pbo path.

V2: Add a comment and bump up the commit message.

Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Cc: <mesa-stable@lists.freedesktop.org>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>

mesa: Turn get_readpixels_transfer_ops() in to a global function

This utility function is utilized in a later patch.

Cc: <mesa-stable@lists.freedesktop.org>
Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>

i965: Use updated kernel interface for accurate TIMESTAMP reads

I was mistaken, I thought we already had fixed this in the kernel a
couple of years ago. We had not, and the broken read (the hardware
shifts the register output on 64bit kernels, but not on 32bit kernels) is
now enshrined into the ABI. I also had the buggy architecture reversed,
believing it to be 32bit that had the shifted results. On the basis of
those mistakes, I wrote

commit c8d3ebaffc0d7d915c1c19d54dba61fd1e57b338
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Wed Apr 29 13:32:38 2015 +0100

    i965: Query whether we have kernel support for the TIMESTAMP register once

Now that we do have an extended register read interface for always
reporting the full 36bit TIMESTAMP (irrespective of whether the hardware
is buggy or not), make use of it and in the process fix my reversed
detection of the buggy reads for unpatched kernels.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Martin Peres <martin.peres@linux.intel.com>
Cc: Kenneth Graunke <kenneth@whitecape.org>
Cc: Michał Winiarski <michal.winiarski@intel.com>
Cc: Daniel Vetter <daniel@ffwll.ch>
Tested-and-acked-by: Chris Forbes <chrisf@ijw.co.nz>
Reviewed-by: Daniel Vetter <daniel@ffwll.ch>

glsl/glcpp: fix SIGSEGV when checking error condition for macro redefinition

Commit a6e9cd14c does not take into account than node_{a,b}->next could be NULL
in some circumstances, such as in a shader containing this code:

  #define A 1 /* comment */
  #define A 1 /* comment */

This patch fixes the segmentation fault for cases like that.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=91290
Signed-off-by: Samuel Iglesias Gonsalvez <siglesias@igalia.com>
Reviewed-by: Anuj Phogat <anuj.phogat@gmail.com>
Cc: mesa-stable@lists.freedesktop.org

nvc0/ir: per-patch vars are in a separate address space

There's no need to attempt to avoid overlapping generic i/o with patch
i/o. By the same token, we can't merge patch and non-patch loads/stores.

This fixes at least the

  tes-both-input-array-*-index-rd

tessellation variable-indexing tests.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>

nvc0/ir: kepler can't do indirect shader input/output loads directly

There's a special AL2P instruction (called AFETCH in nv50 ir) which
computes a "physical" value to be used with indirect addressing with ALD.

Fixes

  tcs-input-array-*-index-rd
  tcs-output-array-*-index-wr

varying-indexing tessellation tests on Kepler.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>

radeon: Silence GCC unused-but-set-variable warnings.

radeon_fbo.c: In function 'radeon_map_renderbuffer_s8z24':
radeon_fbo.c:162:9: warning: variable 'ret' set but not used [-Wunused-but-set-variable]
     int ret;
         ^
radeon_fbo.c: In function 'radeon_map_renderbuffer_z16':
radeon_fbo.c:200:9: warning: variable 'ret' set but not used [-Wunused-but-set-variable]
     int ret;
         ^
radeon_fbo.c: In function 'radeon_map_renderbuffer':
radeon_fbo.c:242:8: warning: variable 'ret' set but not used [-Wunused-but-set-variable]
    int ret;
        ^
radeon_fbo.c: In function 'radeon_unmap_renderbuffer':
radeon_fbo.c:419:14: warning: variable 'ok' set but not used [-Wunused-but-set-variable]
    GLboolean ok;
              ^

Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>

doxygen: Link GLvector4f struct members properly, avoiding invalid XML/HTML warning

Signed-off-by: Rhys Kidd <rhyskidd@gmail.com>
Reviewed-by: Brian Paul <brianp@vmware.com>

doxygen: Correct grammatical typo in math/m_vector.h

Signed-off-by: Rhys Kidd <rhyskidd@gmail.com>
Reviewed-by: Brian Paul <brianp@vmware.com>

mesa: minor clean-ups in shaderapi.c

80-column wrapping.  Move break statements.  Indentation fixes.

mesa: fix _mesa_error() compiler warnings in shaderapi.c

Fix many instances of:
main/shaderapi.c: In function '_mesa_GetSubroutineUniformLocation':
main/shaderapi.c:2176:7: warning: format not a string literal and no format arguments [-Wformat-security]
       _mesa_error(ctx, GL_INVALID_OPERATION, api_name);
       ^

Ideally, many of these error messages should be improved to indicate
which argument is incorrect as we do in other parts of Mesa.

Reviewed-by: Kai Wasserbäch <kai@dev.carbon-project.org>
Tested-by: Kai Wasserbäch <kai@dev.carbon-project.org>

st/mesa: remove unused 'samp' function parameters

Reviewed-by: Marek Olšák <marek.olsak@amd.com>

st/mesa: add comments on a few sampler view functions

Trivial.

mesa: do more thorough target checking in compressed_subtexture_target_check()

When we're error-checking the target, we also need to check if the
corresponding extension is supported.

Reviewed-by: Roland Scheidegger <sroland@vmware.com>

mesa: another target fix in compressed_subtexture_target_check()

The previous fix added GL_TEXTURE_CUBE_MAP_ARRAY but we also need
to support GL_TEXTURE_CUBE_MAP (via DSA).

So in the end, GL_TEXTURE_3D is the only (legal) target for
glCompressedTex*SubImage3D() which needs additional compression
format checking.  GL_TEXTURE_2D_ARRAY, GL_TEXTURE_CUBE_MAP_ARRAY
and GL_TEXTURE_CUBE_MAP are basically 2D images which support all
compressed formats.

Reviewed-by: Roland Scheidegger <sroland@vmware.com>

mesa: simplify format check in compressed_subtexture_target_check()

Lose the invalidformat local variable.

Reviewed-by: Roland Scheidegger <sroland@vmware.com>

mesa: initialize variables to silence compiler warnings

Reviewed-by: Anuj Phogat <anuj.phogat@gmail.com>

apiexec: remove leading gl from shader subroutine interfaces

Remove the gl at the start, stared at this for a while
yesterday, totally missed it.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=91441
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Signed-off-by: Dave Airlie <airlied@redhat.com>

docs: remove expanded ARB_dsa notes

This doesn't provide much value since it's all done. The qbo interaction
is fairly trivial.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Dave Airlie <airlied@redhat.com>

nvc0/ir: tess factors are now sysvals, adapt codegen to expect that

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>

docs/GL3.txt: ARB_shader_precision

This extension is about setting expectation on GL4.1 implementations
rather than actually enforcing things. So once you support GLSL 410
then you support this in theory.

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

i965: add support for ARB_shader_subroutine

This just adds some missing pieces to nir/i965,
it is lightly tested on my Haswell.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>

mesa: rearrange texture error checking order

This moves the width/height/depth == 0 check to the front and avoids
doing any other checking when that is the case.

Also moves the dimensions check after the format/type checks so that we
don't bail out with success on a width/height/depth == 0 request when
the format/type don't match.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=91425
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Brian Paul <brianp@vmware.com>

mesa: adjust error message when there's a missing teximage

The current message makes it seem like the zoffset is invalid.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Brian Paul <brianp@vmware.com>

Revert "Match swrast modes more loosely."

This reverts commit f3728a16c9c6a02fc1f44b8069b0060e2358f22e.

It broke glxgears on radeonsi. The window was just black.

gallivm: Fix profile build.

gallium/util: Stop bundling our snprintf implementation.

Use MSVCRT functions instead.  Their semantics are slightly
different but they can be made to work as expected.

Also, use the same code paths for both MSVCRT and MinGW.

https://bugs.freedesktop.org/show_bug.cgi?id=91418

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

Match swrast modes more loosely.

https://bugs.freedesktop.org/show_bug.cgi?id=90817

Reviewed-by: Jose Fonseca <jfonseca@vmware.com>

mesa: Fix error in target validation of glCompressedTex(ture)SubImage3D() calls

Basically, two different target error checks are chained consecutively, and the
second one is executed regardless the result of the first one. This produces an
incorrect error if the first check fails but is overrided by the second.

This patch conditions the execution of the second check to a successful pass of
the first one.

Fixes 1 dEQP test:
* dEQP-GLES3.functional.negative_api.texture.compressedtexsubimage3d

Reviewed-by: Laura Ekstrand <laura@jlekstrand.net>

gallivm: Add ifdefs so raw_debug_stream is only defined when used

Its only use is to implement a custom version of LLVMDumpValue
on some Windows and embedded platforms.

Reviewed-by: Jose Fonseca <jfonseca@vmware.com>

gallivm: Don't use raw_debug_ostream for dissasembling

All LLVM API calls that require an ostream object have been removed from
the disassemble() function, so we don't need to use this class to wrap
_debug_printf() we can just call this function directly.

Reviewed-by: Jose Fonseca <jfonseca@vmware.com>

docs: mark off tess for nvc0

gk110/ir: fake BAR support

Makes things sorta work until we figure out the real way to do this.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>

nvc0/ir: cleanup private enums that have graduated to gallium

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>

nvc0/ir: allow tess eval output loads to be CSE'd

These only happen for gl_TessCoord which are constant.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>

nvc0/ir: add hazard for 2nd dim of vfetch/load indirect argument

Apparently a multi-word load can potentially overwrite the indirect
sources, so make sure that RA picks different registers for those.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>

nvc0/ir: patch vertex count is stored in the upper bits

nvc0/ir: add support for reading outputs in tess control shaders

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>

nvc0/ir: set perPatch flag on load/stores to per-patch varyings

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>

nvc0/ir: populate info structure based on new tess properties

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>

nvc0/ir: mark varyings as per-patch based on semantic name

Also add proper handling for PATCH semantics

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>

nvc0: TESSCOORD comes in as a sysval, not an input

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>

nvc0: add handling for set_tess_state callback

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>

nvc0: add support for setting patch vertices at draw time

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>

nvc0: support MAX_SHADER_PATCH_VARYINGS

nvc0: preliminary tess support

Uncomment the various functionality that was already there and add in
obvious missing bits that parallel vp/gp/fp functionality.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>

docs/GL3.txt: update ARB_shader_subroutine status.

Acked-by: Kenneth Graunke <kenneth@whitecape.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>

st/mesa: enable shader subroutine

since this touches drivers, only enable it on gallium
for now for drivers reporting GLSL 1.30 or above.

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

st/mesa: add subroutine bits (v1.1)

Just add support for the subroutine type to the
glsl->tgsi convertor.

v1.1: add subroutine to int support.

Acked-by: Kenneth Graunke <kenneth@whitecape.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>

mesa: fill out the ARB_shader_subroutine APIs

This fleshes out the APIs, using the program resource
APIs where they should match.

It also sets the default values to valid subroutines.

Acked-by: Kenneth Graunke <kenneth@whitecape.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>

program: add subroutine uniform support (v1.1)

Add support for the subroutine uniform type ir->mesa.cpp

v1.1: add subroutine to int to switch

Acked-by: Kenneth Graunke <kenneth@whitecape.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>

program_resource: add subroutine support (v3.1)

This fleshes out the ARB_program_query support for the
APIs that ARB_shader_subroutine introduces, leaving
some TODOs for later addition.

v2: reworked for lots of the ARB_program_interface_query
entry points and tests
v3: use common function to test for subroutine support
v3.1: add tess, fix missing breaks

Acked-by: Kenneth Graunke <kenneth@whitecape.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>

glsl: add uniform and program resource support (v2)

This adds linker support for subroutine uniforms, they
have some subtle differences from real uniforms, we also hide
them and they are given internal uniform names.

This also adds the subroutine locations and subroutine uniforms
to the program resource tracking for later use.

v1.1: drop is_subroutine_def

v2: handle explicit location properly, ARB_explicit_location
has a lot of language for subroutine shaders.
Calculate a link time the number of compatible subroutines
for a uniform, to make program resource easier later.

Acked-by: Kenneth Graunke <kenneth@whitecape.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>

mesa/mtypes: add gl_subroutine_function and uniform storage to shader (v2)

This adds the necessary storage for subroutine info to gl_shader.

v2: add comments, rename one member
Acked-by: Kenneth Graunke <kenneth@whitecape.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>

glsl/ir: add subroutine lowering pass (v2.3)

This lowers the enhanced ir_call using the lookaside table
of subroutines into an if ladder. This initially was done
at the AST level but it caused some ordering issues so a separate
pass was required.

v2: clone return value derefs.
v2.1: update for subroutine->int convert.
v2.2: add a clone for the array index

Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>
Signed-off-by: Dave Airlie <airlied@redhat.com>

glsl: add ast/parser support for subroutine parsing storage (v3.2)

This is the guts of the GLSL parser and AST support for
shader subroutines.

The code creates a subroutine type in the parser, and
uses that there to validate the identifiers. The parser
also distinguishes between subroutine types/function prototypes
/uniforms and subroutine defintions for functions.

Then in the AST conversion it recreates the types, and
stores the subroutine definition info or subroutine info
into the ir_function along with a side lookup table in
the parser state. It also converts subroutine calls into
the enhanced ir_call.

v2: move to handling method calls in
function handling not in field selection.
v3: merge Chris's previous parser patches in here, to
make it clearer what's changed in one place.
v3.1: add more documentation, drop unused include
v3.2: drop is_subroutine_def

Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>
Signed-off-by: Dave Airlie <airlied@redhat.com>

glsl/ir: allow ir_call to handle subroutine calling

This adds a ir_variable which contains the subroutine uniform
and an array rvalue for the deref of that uniform, these
are stored in the ir_call and lowered later.

Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>
Signed-off-by: Dave Airlie <airlied@redhat.com>

glsl/ir: add subroutine information storage to ir_function (v1.1)

We need to store two sets of info into the ir_function,
if this is a function definition with a subroutine list
(subroutine_def) or if it a subroutine prototype.

v1.1: add some more documentation.

Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>
Signed-off-by: Dave Airlie <airlied@redhat.com>

mesa: add function to check if shader subroutines are enabled.

This checks if core profile and shader subroutine extension
is enabled.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>

mesa: add inline conversion functions for ARB_shader_subroutine (v2)

This handles converting the shader stages to the internal
prefix along with the program resource interfaces.

v2: add tess support

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

glsl: don't eliminate subroutine types.

This stops dead code from removing subroutines types,
we need these for the queries to work properly.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>

glsl/types: add new subroutine type (v3.2)

This type will be used to store the name of subroutine types

as in subroutine void myfunc(void);
will store myfunc into a subroutine type.

This is required to the parser can identify a subroutine
type in a uniform decleration as a valid type, and also for
looking up the type later.

Also add contains_subroutine method.

v2: handle subroutine to int comparisons, needed
for lowering pass.
v3: do subroutine to int with it's own IR
operation to avoid hacking on asserts (Kayden)
v3.1: fix warnings in this patch, fix nir,
fix tgsi
v3.2: fixup tests

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>
Signed-off-by: Dave Airlie <airlied@redhat.com>
tests: fix warnings

glsl: Make `subroutine` a reserved keyword

Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Signed-off-by: Chris Forbes <chrisf@ijw.co.nz>
Signed-off-by: Dave Airlie <airlied@redhat.com>

dispatch_sanity: add shader subroutine to fix make check

Add the shader subroutine to the core only API list,
and fixup dispatch_sanity to suit.

Acked-by: Kenneth Graunke <kenneth@whitecape.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>

glsl: Add extension plumbing and define for ARB_shader_subroutine

Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Signed-off-by: Chris Forbes <chrisf@ijw.co.nz>
Signed-off-by: Dave Airlie <airlied@redhat.com>

mesa: Add glGet support for ARB_shader_subroutine implementation limits

Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Signed-off-by: Chris Forbes <chrisf@ijw.co.nz>
Signed-off-by: Dave Airlie <airlied@redhat.com>

mesa: Add extension tracking for arb_shader_subroutine (v2)

v2: [airlied]: merge version check update.

Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Signed-off-by: Chris Forbes <chrisf@ijw.co.nz>
Signed-off-by: Dave Airlie <airlied@redhat.com>

glapi: Add ARB_shader_subroutine functions and enums (v2)

v2: fix output="true" and LENGTH typo

Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Signed-off-by: Chris Forbes <chrisf@ijw.co.nz>
Signed-off-by: Dave Airlie <airlied@redhat.com>

mesa: Add stubs for ARB_shader_subroutine entrypoints

Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Signed-off-by: Chris Forbes <chrisf@ijw.co.nz>
Signed-off-by: Dave Airlie <airlied@redhat.com>

dispatch_sanity.cpp: remove commented out tess entries

These entries were put in the GL4.0 section, so removed the commented
out ones.

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

glsl/tests: fix varying_test since tess changes.

This fixes make check since the tess changes.

Tested-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

i965: fix warning since tess merge.

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

radeonsi: enable tessellation, update GL3.txt & release notes

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

radeonsi: add scratch buffer support for tessellation shaders

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

radeonsi: update invariant registers for tessellation

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

radeonsi: add assertions into draw_vbo and check tessellation

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

radeonsi: set the rasterization primitive type for tessellation

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

radeonsi: set VGT_LS_HS_CONFIG for tessellation

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

radeonsi: update IA_MULTI_VGT_PARAM for tessellation

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

radeonsi: add derived tessellation state

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

radeonsi: implement a fixed-function tessellation control shader and its state

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

radeonsi: set up a ring buffer for tessellation factors

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

radeonsi: add tessellation shader states

ls_rsrc# will be emitted as part of the derived tessellation state

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

radeonsi: update si_get_vs_info and si_get_vs_state for tessellation

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

radeonsi: add shader code generation for tessellation

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

radeonsi: implement TGSI_OPCODE_BARRIER

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

radeonsi: program VGT_SHADER_STAGES_EN for tessellation

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

radeonsi: add translation of PATCH primitives

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

radeonsi: add support for tessellation shader resources and samplers

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

radeonsi: add code for creating, binding and destroying tessellation shaders

This doesn't do anything yet.

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

radeonsi: add debug flags for dumping tessellation shaders

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

radeonsi: rework how shader pointers to descriptors are set

This is mainly needed for tessellation where a VS can be bound as VS, ES,
or LS, and TES (tess. evaluationshader) can be bound as VS or ES or neither.
Therefore we need the ability to move pointers to descriptors between
shaders arbitrarily.

The idea is that the context has a mapping from PIPE_SHADER_x to
SPI_SHADER_USER_DATA_x. After a shader is enabled or disabled,
si_shader_change_notify should be called to update this mapping accordingly.

There is a dirty flag for each shader pointer, but only one emit function
for all pointers in the whole context, whose code and logic is separated
from descriptors.

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

radeonsi: rename build_streamout_store -> build_tbuffer_store_dwords

It will be reused later.

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

radeonsi: separate primitive ID computation

Support for new shader stages will be added here.

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

radeonsi: move declaring streamout parameters to its own function

It will be reused later.

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