genxml: Define GENXML_XML_FILES in Makefile.sources

The future header genX_bits.h will depend on GENXML_XML_FILES.

Reviewed-by: Emil Velikov <emil.velikov@collabora.com>

clover: use pipe_resource references

v2: buffers are created with one reference.
v3: add pipe_resource reference to mapping object
v4: rename to pres and drop inline initializers

CC: "17.0 13.0" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
Reviewed-by: Francisco Jerez <currojerez@riseup.net>

i965: Fix symbolic size of next_offset[] array.

It's indexed by buffer, not stream.  BRW_MAX_SOL_BUFFERS and
MAX_VERTEX_STREAMS happen to both be 4, so there's no actual bug.

Reviewed-by: Matt Turner <mattst88@gmail.com>

i965: Remove pointless NULL check from Gen6 primitive counting code.

We create the BO when creating a transform feedback object, and only
destroy it when deleting that object.  So it won't be NULL.

CID: 1401410

Reviewed-by: Matt Turner <mattst88@gmail.com>

radeonsi: don't crash on compute shader compile failure

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>

radeonsi: don't hang on shader compile failure

Cc: 17.0 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>

radeonsi: fix dvec[34] attributes sourced from current attribute state

The state tracker no longer uploads those attributes for us,
so we must conservatively upload the size of the largest
attribute, which is a dvec4.

Fixes a regression of GL45-CTS.gpu_shader_fp64.varyings and
GL45-CTS.vertex_attrib_64bit.limits_test.

Fixes: 9b91e0b54cc2 ("radeonsi: allow unaligned vertex buffer offsets and strides on CIK-VI")
Reviewed-by: Marek Olšák <marek.olsak@amd.com>

anv: automake: ensure that the destination directory is created

Earlier commit unintentionally dropped the mkdir, as it was rebased.

Some versions of autotools will not create the output directory for
generated sources. Thus the issue went unnoticed by the original author.

Cc: Dylan Baker <dylan@pnwbakers.com>
Cc: Steven Newbury <steve@snewbury.org.uk>
Reported-by: Steven Newbury <steve@snewbury.org.uk> Fixes:
Fixes: 1610b3dede1 ("anv: don't pass xmlfile via stdin anv_entrypoints_gen.py")
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>

glsl_to_tgsi: don't rely on glsl types when visiting tex instructions

Instead add is_cube_shadow like is_cube_array.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>

anv/query: handle out of host memory without crashing in compute_query_result()

We don't need to make the caller (CmdCopyQueryPoolResults) aware of the
problem since compute_query_result() only emits state. The caller is also
expected to hit OOM in this scenario right after calling this function, but
it is already handling it safely.

Fixes:
dEQP-VK.api.out_of_host_memory.cmd_copy_query_pool_results

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>

anv/pipeline: make FragCoord include sample positions when sample shading

We need to know if sample shading has been requested during shader
compilation since that affects the way fragment coordinates are
computed.

Notice that the semantics of fragment coordinates only depend on
whether sample shading has been requested, not on whether more
than one sample will actually be produced (that is,
minSampleShading and rasterizationSamples do not affect this
behavior).

Because this setting affects the code we generate for the shader, we also
need to include it in the WM prog key. Notice we don't need to alter the
OpenGL code because it doesn't ever use this behavior, so they key's
value is always false (the default).

Fixes:
dEQP-VK.glsl.builtin_var.fragcoord_msaa.*

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

nir/lower_wpos_center: support adding sample position to fragment coordinate

According to section 14.6 of the Vulkan specification:

   "When sample shading is enabled, the x and y components of FragCoord
    reflect the location of the sample corresponding to the shader
    invocation."

So add a boolean parameter to the lowering pass to select this behavior
when we need it.

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

anv: return VK_ERROR_DEVICE_LOST immeditely when device is known to be lost

If we know the device has been lost we should return this error code for
any command that can report it before we attempt to do anything with the
device.

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

anv/device: keep track of 'device lost' state

The Vulkan specs say:

   "A logical device may become lost because of hardware errors, execution
    timeouts, power management events and/or platform-specific events. This
    may cause pending and future command execution to fail and cause hardware
    resources to be corrupted. When this happens, certain commands will
    return VK_ERROR_DEVICE_LOST (see Error Codes for a list of such commands).
    After any such event, the logical device is considered lost. It is not
    possible to reset the logical device to a non-lost state, however the lost
    state is specific to a logical device (VkDevice), and the corresponding
    physical device (VkPhysicalDevice) may be otherwise unaffected. In some
    cases, the physical device may also be lost, and attempting to create a
    new logical device will fail, returning VK_ERROR_DEVICE_LOST."

This means that we need to track if a logical device has been lost so we can
have the commands referenced by the spec return VK_ERROR_DEVICE_LOST
immediately.

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

anv/device: return VK_ERROR_DEVICE_LOST for errors during queue submissions

So that we don't have to do things like rolling back address relocations in
case that we ran into OOM after computing them, etc

Also, make sure that if the queue submission comes with a fence, we set it up
correctly so it behaves according to the spec after returning
VK_ERROR_DEVICE_LOST.

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

mesa/marshal: add custom BufferData/BufferSubData marshalling

GL_AMD_pinned_memory requires memory to be aligned correctly, so
we skip marshalling in this case. Also copying the data defeats
the purpose of EXTERNAL_VIRTUAL_MEMORY_BUFFER_AMD.

Fixes GL_AMD_pinned_memory piglit tests when glthread is enabled.

Acked-by: Edward O'Callaghan <funfunctor@folklore1984.net>

util/disk_cache: write cache entry keys to file header

This can be used to deal with key hash collisions from different
versions (should we find that to actually happen) and to find
which mesa version produced the cache entry.

V2: use blob created at cache creation.

v3: remove left over var from v1.

Reviewed-by: Grazvydas Ignotas <notasas@gmail.com>

util/disk_cache: hash pointer size and gpu name into cache keys

This allows to get rid of the arch and gpu name directories.

v2: (Timothy Arceri) don't use an opaque data type to store
    pointer size and gpu name.

v3: (Timothy Arceri) use blob to store driver keys just make sure
    to store null terminator for strings, and make sure blob is
    defined by disk_cache and not it's users.

v4: (Timothy Arceri) fix typo, and make ptr_size a uint8_t.

Signed-off-by: Grazvydas Ignotas <notasas@gmail.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>

util/disk_cache: hash timestamps into the cache keys

Instead of using a directory, hash the timestamps into the cache keys
themselves. Since there is no more timestamp directory, there is no more
need for deleting the cache of other mesa versions and we rely on
eviction to clean up the old cache entries. This solves the problem of
using several incarnations of disk_cache at the same time, where one
deletes a directory belonging to the other, like when both OpenGL and
gallium nine are used simultaneously (or several different mesa
installations).

v2: using additional blob instead of trying to clone sha1 state

v3: (Timothy Arceri) don't use an opaque data type to store
    timestamp.

V4: (Timothy Arceri) use blob to store driver keys just make sure
    to store null terminator for strings, and make sure blob is
    defined by disk_cache and not it's users.

Signed-off-by: Grazvydas Ignotas <notasas@gmail.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=100091

mesa: set thread name for glthread

Signed-off-by: Miklós Máté <mtmkls@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>

i965: Replace OPT_V() with OPT().

We want to be able to check the progress of each pass and dump the NIR
for debugging purposes if it changed.

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

i965/fs: Return progress from demote_sample_qualifiers().

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

i965/fs: Return progress from move_interpolation_to_top().

And mark as static at the same time.

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

i965: Return progress from brw_nir_lower_uniforms().

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

nir: Return progress from nir_convert_from_ssa().

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

nir: Return progress from nir_lower_io().

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

nir: Return progress from nir_lower_regs_to_ssa().

And from nir_lower_regs_to_ssa_impl() as well.

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

nir: Return progress from nir_lower_samplers().

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

nir: Return progress from nir_lower_atomics().

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

nir: Return progress from nir_lower_clamp_color_outputs().

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

nir: Return progress from nir_lower_clip_fs().

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

nir: Return progress from nir_lower_clip_vs().

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

nir: Return progress from nir_move_vec_src_uses_to_dest().

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

nir: Return progress from nir_lower_to_source_mods().

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

nir: Return progress from nir_lower_clip_cull_distance_arrays().

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

nir: Return progress from nir_lower_var_copies().

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

nir: Return progress from nir_lower_load_const_to_scalar().

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

nir: Return progress from nir_lower_64bit_pack().

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

nir: Return progress from nir_lower_doubles().

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

nir: Return progress from nir_lower_vars_to_ssa().

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

nir: Fix syntax.

et is not an abbreviation.

Reviewed-by: Dylan Baker <dylan@pnwbakers.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

nir: Fix misspellings.

Reviewed-by: Dylan Baker <dylan@pnwbakers.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

nir: Stop using apostrophes to pluralize.

Reviewed-by: Dylan Baker <dylan@pnwbakers.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

st/omx/enc: use PIPE_USAGE_STAGING for output buffer

Workaround an unknown bug with inside the transfer_map for certain
ASIC, also tested with un-affected ASICs, the performance actually
improved slightly.

Signed-off-by: Leo Liu <leo.liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>

gbm: Use unsigned for BO offset getter

The actual offset returned is uint32_t, however int64_t was used as the
return type from gbm_bo_get_offset to allow negative returns to signal
errors to the caller.

In case of an error getting the offset, the user will also be unable to
get the handle/FD, and thus have nothing to offset into. This means that
returning 0 as an error value is harmless, allowing us to change the
return type to uint32_t in order to avoid signed/unsigned confusion in
callers.

Signed-off-by: Daniel Stone <daniels@collabora.com>
Cc: Ben Widawsky <ben@bwidawsk.net>
Cc: Jason Ekstrand <jason@jlekstrand.net>

REVIEWERS: add autogen.sh to the autoconf group

Signed-off-by: Eric Engestrom <eric.engestrom@imgtec.com>
Acked-by: Emil Velikov <emil.l.velikov@gmail.com>

docs/submittingpatches: add mention about legal disclaimers

Signed-off-by: Eric Engestrom <eric.engestrom@imgtec.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>

r600_shader.c: fix indentation

Introduced by ad13bd2e51a5dc01b0f8a0eb927022f0deac0a0c

Signed-off-by: Julien Isorce <jisorce@oblong.com>
Reviewed-by: Nayan Deshmukh <nayan26deshmukh@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>

glx: Prefer library path given by pkgconfig over the system

Recent change to use drmGetDevices2() made me realize that
build configured using

PKG_CONFIG_PATH=my_drm_lib_path/pkgconfig ./autogen.sh

considers the libdrm path gotten from pkgconfig only during
make. When invoking "make install" the relink command puts
system library ahead of the path gotten from pkgconfig
(and starts to fail as system libdrm isn't new enough).

This change forces the relink command to respect pkgconfig
settings.

It looks to me that in

https://bugs.freedesktop.org/show_bug.cgi?id=100259

with Emil et al considering it a libtool bug.

Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
[Emil Velikov: add inline comment]
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>

intel: move gen_decoder.* to DECODER_FILES

patch adds DECODER_FILES for libintel_common, this is so that platforms
such as Android not currently using this functionality can opt out.

Fixes: 7d84bb3 ("intel: Move tools/decoder.[ch] to common/gen_decoder.[ch].")
Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>

android: fix vulkan build issues with anv_entrypoints

Patch fixes entrypoint generation for libmesa_anv_entrypoints that
still used old style of calling generator script.

Also small fixes to libmesa_vulkan_common where there was a typo
in target name (vulknan) and files were generated to wrong folder.

Fixes: 8211e3e6 ("anv: Generate anv_entrypoints header and code in one command")
Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Acked-by: Emil Velikov <emil.velikov@collabora.com>

android: i965: generate code for OA counter queries

Automake generation rules are replicated for android.
$* macro was expected to return "hsw" but instead gives "hsw.{h,c}"
so $(basename $*) is used as a workaround
to set the correct --chipset option for brw_oa.py script.

Build tested with nougat-x86

Fixes: e565505 "i965: Add script to gen code for OA counter queries"
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Acked-by: Robert Bragg <robert@sixbynine.org>
Acked-by: Emil Velikov <emil.velikov@collabora.com>

android: rename Intel Vulkan library to match desktop one

Original naming was following Vulkan HAL naming scheme for no good
purpose and we need same binary name for build-id code.

Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Acked-by: Emil Velikov <emil.velikov@collabora.com>

nouveau: enable glsl/tgsi on-disk cache

v2: Fix argument to nouveau_screen_get_name()

Signed-off-by: Boyan Ding <boyan.j.ding@gmail.com>
Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>

REVIEWERS: add myself as a reviewer for EGL and docs

Signed-off-by: Eric Engestrom <eric.engestrom@imgtec.com>

anv: Remove dead prototype from entrypoints

Spotted by Emil.

v2: - Add this patch

Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>

anv: use cElementTree in anv_entrypoints_gen.py

It's written in C rather than pure python and is strictly faster, the
only reason not to use it that it's classes cannot be subclassed.

Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>

anv: don't use Element.get in anv_entrypoints_gen.py

This has the potential to mask errors, since Element.get works like
dict.get, returning None if the element isn't found. I think the reason
that Element.get was used is that vulkan has one extension that isn't
really an extension, and thus is missing the 'protect' field.

This patch changes the behavior slightly by replacing get with explicit
lookup in the Element.attrib dictionary, and using xpath to only iterate
over extensions with a "protect" attribute.

Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>

anv: use dict.get in anv_entrypoints_gen.py

Instead of using an if and a check, use dict.get, which does the same
thing, but more succinctly.

Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>

anv: anv_entrypoints_gen.py: use reduce function.

Reduce is it's own reward.

Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>

anv: anv-entrypoints_gen.py: rename hash to cal_hash.

hash is reserved name in python, it's the interface to access an
object's hash protocol.

Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>

anv: Generate anv_entrypoints header and code in one command

This produces the header and the code in one command, saving the need to
call the same script twice, which parses the same XML file.

Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>

anv: anv_entrypoints_gen.py: directly write files instead of piping

This changes the output to be written as a file rather than being piped.
This had one critical advantage, it encapsulates the encoding. This
prevents bugs where a symbol (generally unicode like © [copyright]) is
printed and the system being built on doesn't have a unicode locale.

v2: - Update Android.mk
v3: - Don't generate both files at once
    - Fix Android.mk
    - drop --outdir, since the filename is passed in as an argument

Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>

anv: convert C generation to template in anv_entrypoints_gen.py

This produces a file that is identical except for whitespace, there is a
table that has 8 columns in the original and is easy to do with prints,
but is ugly using mako, so it doesn't have columns; the data is not
inherently tabular.

Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>

anv: convert header generation in anv_entrypoints_gen.py to mako

This produces an identical file except for whitespace.

Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>

anv: Update "do not edit" comments with proper filename

This does two things, first it updates both the .h and the .c file to
have the same do not edit string. Second, it uses __file__ to ensure
that even if the file is moved or renamed that the name will be correct.

One thing to note is the use of '{{' and '}}' in the C template. This is
to instruct python to print a literal '{' and '}' respectively, rather
than treating the contents as a formatter specifier.

v3: - add this patch

Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>

anv: split main into two functions in anv_entrypoints_gen.py

This is groundwork for the next patches, it will allows porting the
header and the code to mako separately, and will also allow both to be
run simultaneously.

Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>

anv: don't pass xmlfile via stdin anv_entrypoints_gen.py

It's slow, and has the potential for encoding issues.

v2: - pass xml file location via argument
    - update Android.mk

Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>

anv: make constants capitals in anv_entrypoints_gen.py

Again, it's standard python style.

Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>

anv: Use python style in anv_entrypoints_gen.py

These are all fairly small cleanups/tweaks that don't really deserve
their own patch.

- Prefer comprehensions to map() and filter(), since they're faster
- replace unused variables with _
- Use 4 spaces of indent
- drop semicolons from the end of lines
- Don't use parens around if conditions
- don't put spaces around brackets
- don't import modules as caps (ET -> et)
- Use docstrings instead of comments

v2: - Replace comprehensions with multiplication

Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>

anv: anv_entrypoints_gen.py: use a main function

This is just good practice.

Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>

radv: Invalidate L2 for TRANSFER_WRITE barriers

CP DMA and PKT3_WRITE_DATA (in CmdUpdateBuffer) don't (currently) write
through L2. Therefore, to make these writes visible to later accesses
we must invalidate L2 rather than just writing it back, to avoid the
possibility that stale data is read through L2.

Cc: "17.0" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Alex Smith <asmith@feralinteractive.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Signed-off-by: Dave Airlie <airlied@redhat.com>

glsl: Link glsl_compiler with CLOCK_LIB.

Fix linking error on CentOS 6.

  CXXLD  glsl_compiler
glsl/.libs/libstandalone.a(lt16-libmesautil_la-u_queue.o): In function `u_thread_get_time_nano':
src/util/../../src/util/u_thread.h:84: undefined reference to `clock_gettime'

Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>

util/disk_cache: use rand_xorshift128plus() to get our random int

Otherwise for apps that don't seed the regular rand() we will always
remove old cache entries from the same dirs.

V2: assume bits returned by rand are independent uniformly distributed
    bits and grab our hex value without taking the modulus of the whole
    value, this also fixes a bug where 'f' was always missing.

Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>

util/rand_xor: add function to seed rand

V2: pass the seed to the seed function so that we can isolate
    its uses. Stop leaking fd when urandom couldn't be read.

Reviewed-by: Grazvydas Ignotas <notasas@gmail.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>

util: move rand_xorshift128plus() to utils

V2: pass the seed to rand_xorshift128plus() so that we can isolate
    its uses.

Reviewed-by: Grazvydas Ignotas <notasas@gmail.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>

drirc: add force_glsl_abs_sqrt() for "Spec Ops: The Line"

Game ported from D3D9 which expects sqrt() to compute the absolute
value as explained in the spec.

This gets rid of the NaN values as well as the black squares
with RadeonSI.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=97338
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>

st/glsl_to_tgsi: enable lower_sqrt() conditionally

It relies on the force_glsl_abs_sqrt driconf option.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>

glsl: lower sqrt(abs()) and inversesqrt(abs()) if requested

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

driconf: add force_glsl_abs_sqrt option

This will allow to force computing the absolute value for sqrt()
and inversesqrt() in order to follow D3D9 behaviour for buggy
apps that rely on it.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>

swr: [rasterizer jitter] fix llvm >= 5.0 build break

Function::getArgumentList() doesn't exist anymore, switch to using
arg_begin() (existed back to at least llvm-3.6.0).

Reviewed-by: Vedran Miletić <vedran@miletic.net>
CC: <mesa-stable@lists.freedesktop.org>

Android: drop Android 4.4 (KitKat) support

Any users of KitKat are likely using an older version of Mesa and
KitKat support adds complexity to the make files. Dropping support
allows removing the MESA_LOLLIPOP_BUILD make variable in various make
files.

Signed-off-by: Rob Herring <robh@kernel.org>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>

Android: kill off {MESA_}ANDROID_VERSION defines aka Android 4.1 and older

The Android version defines are only needed for versions less than 4.2
which aren't really supported or tested.

Signed-off-by: Rob Herring <robh@kernel.org>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>

Android: fix libz dependency for host targets

Commit 6facb0c08fa1 ("android: fix libz dynamic library dependencies")
added libz as a dependency, but this breaks host targets as the host
dependency is libz-host. As no host lib needs libz, just remove the
dependency for them.

Fixes: 6facb0c08fa1 "android: fix libz dynamic library dependencies"
Signed-off-by: Rob Herring <robh@kernel.org>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>

Android: remove host libmesa_util

The host libmesa_util is never used for Android builds, so remove it.

Signed-off-by: Rob Herring <robh@kernel.org>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>

Android: clean-up trailing '\' in make variables

Fixed with the following command:

perl -pe 'BEGIN{undef $/;} s/ \\\n\n/\n\n/smg' $(find . -name 'Android.*')

Signed-off-by: Rob Herring <robh@kernel.org>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>

mesa/main: remove unused strndup.h include

No longer needed as of commit ac257f1070a ("glsl: calculate
TOP_LEVEL_ARRAY_SIZE and STRIDE when adding resources")

Reported-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>
Reviewed-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>

util: automake: beautify sources list

Remove trailing tabs and sort alphabetically.

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Acked-by: Vedran Miletić <vedran@miletic.net>
Acked-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>

util/strndup: move header inclusion as applicable

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Acked-by: Vedran Miletić <vedran@miletic.net>
Acked-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>

util: inline strndup implementation in the header

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Acked-by: Vedran Miletić <vedran@miletic.net>
Acked-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>

util: consistently use ifndef guards over pragma once

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Acked-by: Vedran Miletić <vedran@miletic.net>
Acked-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>

mesa/program: consistently use ifndef guards over pragma once

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Acked-by: Vedran Miletić <vedran@miletic.net>
Acked-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>

mesa/main: consistently use ifndef guards over pragma once

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Acked-by: Vedran Miletić <vedran@miletic.net>
Acked-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>

intel/compiler: consistently use ifndef guards over pragma once

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Acked-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Acked-by: Vedran Miletić <vedran@miletic.net>
Acked-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>

intel/common: consistently use ifndef guards over pragma once

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Acked-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Acked-by: Vedran Miletić <vedran@miletic.net>
Acked-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>

i965: consistently use ifndef guards over pragma once

The only remaining case is the brw_oa.py generator which pipes the
generated file to stdout. That will be resolved with later commits.

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Acked-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Acked-by: Vedran Miletić <vedran@miletic.net>
Acked-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>

st/wgl: consistently use ifndef guards over pragma once

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Acked-by: Vedran Miletić <vedran@miletic.net>
Acked-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>

egl/dri2: consistently use ifndef guards over pragma once

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Acked-by: Vedran Miletić <vedran@miletic.net>
Acked-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>

spirv: consistently use ifndef guards over pragma once

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Acked-by: Vedran Miletić <vedran@miletic.net>
Acked-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>

nir: consistently use ifndef guards over pragma once

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Acked-by: Vedran Miletić <vedran@miletic.net>
Acked-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>