mesa.git
9 years agowinsys/amdgpu: add a new winsys for the new kernel driver
Marek Olšák [Thu, 16 Apr 2015 20:43:23 +0000 (22:43 +0200)]
winsys/amdgpu: add a new winsys for the new kernel driver

v2: - lots of changes according to Emil Velikov's comments
    - implemented radeon_winsys::read_registers

v3: - a lot of new work, many of them adapt to libdrm interface changes
Squashed patches:
winsys/amdgpu: implement radeon_winsys context support
winsys/amdgpu: add reference counting for contexts
winsys/amdgpu: add userptr support
winsys/amdgpu: allocate IBs like normal buffers
winsys/amdgpu: add IBs to the buffer list, adapt to interface changes
winsys/amdgpu: don't use KMS handles as reloc hash keys
winsys/amdgpu: sync buffer accesses to different rings
winsys/amdgpu: use dependencies instead of waiting for last fence v2
gallium/radeon: unify buffer_wait and buffer_is_busy in the winsys interface (amdgpu part)
winsys/amdgpu: track fences per ring and be thread-safe
winsys/amdgpu: simplify waiting on a variable in amdgpu_fence_wait
gallium/radeon: allow the winsys to choose the IB size (amdgpu part)
winsys/amdgpu: switch to new amdgpu_cs_query_fence_status interface
winsys/amdgpu: handle fence and dependencies merge
winsys/amdgpu follow libdrm change to move user fence into UMD
winsys/amdgpu: use amdgpu_bo_va_op for va map/unmap v2
winsys/amdgpu: use the new tiling flags
winsys/amdgpu: switch to new GTT_USWC definition
winsys/amdgpu: expose amdgpu_cs_query_reset_state to drivers
winsys/amdgpu: fix valgrind warnings
winsys/amdgpu: don't use VRAM with APUs that don't have much of it
winsys/amdgpu: require LLVM 3.6.1 for VI because of bug fixes there
winsys/amdgpu: remove amdgpu_winsys::num_cpus
winsys/amdgpu: align BO size to page size
winsys/amdgpu: reduce BO cache timeout
winsys/amdgpu: remove useless flushing and waiting in amdgpu_bo_set_tiling
winsys/amdgpu: use amdgpu_device_handle as a unique device ID instead of fd
winsys/amdgpu: use safer access to amdgpu_fence_wait::signalled
winsys/amdgpu: allow maximum IB size of 4 MB
winsys/amdgpu: add ip_instance into amdgpu_fence
gallium/radeon: add RING_COMPUTE instead of RADEON_FLUSH_COMPUTE
winsys/amdgpu: set the ring type at CS initilization
winsys/amdgpu: query the GART page size from the kernel
winsys/amdgpu: correctly wait for shared buffers to become idle
winsys/amdgpu: set the amdgpu_cs_fence structure only once at fence creation
winsys/amdgpu: add a specific error message for cs_submit -> -ENOMEM
winsys/amdgpu: check num_active_ioctls before calling amdgpu_bo_wait_for_idle
winsys/amdgpu: clear user fence BO after allocating it
winsys/amdgpu: fix user fences
winsys/amdgpu: make amdgpu_winsys_create public
winsys/amdgpu: remove thread offloading
winsys/amdgpu: flatten the amdgpu_cs_context structure and simplify more

v4: require libdrm 2.4.63

9 years agost/vdpau: add HEVC support v2
Christian König [Wed, 29 Apr 2015 13:35:02 +0000 (15:35 +0200)]
st/vdpau: add HEVC support v2

v2: fix return code

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Leo Liu <leo.liu@amd.com>
9 years agost/omx/enc: stack frame tasks for the gathering
Leo Liu [Thu, 25 Jun 2015 17:19:56 +0000 (13:19 -0400)]
st/omx/enc: stack frame tasks for the gathering

Put tasks to the FIFO queue for results

Signed-off-by: Leo Liu <leo.liu@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
9 years agost/omx/enc: flush after eos handling v2
Leo Liu [Fri, 29 May 2015 18:50:44 +0000 (14:50 -0400)]
st/omx/enc: flush after eos handling v2

v2 (chk): reorder the flush

Signed-off-by: Leo Liu <leo.liu@amd.com>
Signed-off-by: Christian König <christian.koenig@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
9 years agovl: add HEVC profiles and defines
Christian König [Tue, 28 Apr 2015 13:31:37 +0000 (15:31 +0200)]
vl: add HEVC profiles and defines

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Leo Liu <leo.liu@amd.com>
9 years agovl: add cap for stacking frames
Leo Liu [Thu, 25 Jun 2015 16:09:11 +0000 (12:09 -0400)]
vl: add cap for stacking frames

Signed-off-by: Leo Liu <leo.liu@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
9 years agoswrast: fix EXT_depth_bounds_test
Marek Olšák [Wed, 12 Aug 2015 23:51:37 +0000 (01:51 +0200)]
swrast: fix EXT_depth_bounds_test

zMin and zMax can't use _DepthMaxF, because the test is done in Z32_UNORM.

Probably a useless patch given how popular swrast is nowadays, but it helped
create and validate the piglit test.

v2: add an explicit cast to GLuint

Reviewed-by: Brian Paul <brianp@vmware.com>
9 years agoradeonsi: add support for EXT_depth_bounds_test
Marek Olšák [Mon, 10 Aug 2015 00:23:21 +0000 (02:23 +0200)]
radeonsi: add support for EXT_depth_bounds_test

Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
9 years agost/mesa: add EXT_depth_bounds_test
Marek Olšák [Mon, 10 Aug 2015 00:18:43 +0000 (02:18 +0200)]
st/mesa: add EXT_depth_bounds_test

Reviewed-by: Brian Paul <brianp@vmware.com>
9 years agogallium: add an interface for EXT_depth_bounds_test
Marek Olšák [Mon, 10 Aug 2015 00:11:48 +0000 (02:11 +0200)]
gallium: add an interface for EXT_depth_bounds_test

Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
9 years agost/mesa: small cleanup in st_extensions.c
Marek Olšák [Mon, 10 Aug 2015 17:53:22 +0000 (19:53 +0200)]
st/mesa: small cleanup in st_extensions.c

Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Edward O'Callaghan <eocallaghan at alterapraxis.com>
9 years agogallium: add support for GLES texture float extensions (v3)
Marek Olšák [Mon, 10 Aug 2015 17:37:01 +0000 (19:37 +0200)]
gallium: add support for GLES texture float extensions (v3)

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=74329

v2: add a CAP for half floats
    drivers should not expose the CAPs if they don't support the formats

v3: update relnotes

Reviewed-by: Edward O'Callaghan <eocallaghan@alterapraxis.com>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
9 years agor600,compute: setup compute sampler states and views
Zoltan Gilian [Tue, 7 Jul 2015 21:38:27 +0000 (23:38 +0200)]
r600,compute: setup compute sampler states and views

v2: Add compute mode flag to sampler state setup (Marek).
    Drop branches which avoid reference counting (Marek).
    Simplify unset branch condition (Marek).

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
9 years agost/clover: Fix build against LLVM 3.8 SVN r244928
Michel Dänzer [Fri, 14 Aug 2015 06:16:12 +0000 (15:16 +0900)]
st/clover: Fix build against LLVM 3.8 SVN r244928

raw_svector_ostream::flush() is now unnecessary and forbidden:

  CXX      llvm/libclllvm_la-invocation.lo
../../../../../src/gallium/state_trackers/clover/llvm/invocation.cpp: In function 'clover::module {anonymous}::build_module_llvm(llvm::Module*, unsigned int (&)[7])':
../../../../../src/gallium/state_trackers/clover/llvm/invocation.cpp:574:29: error: use of deleted function 'void llvm::raw_svector_ostream::flush()'
       bitcode_ostream.flush();
                             ^
In file included from /home/daenzer/src/llvm-git/llvm/include/clang/Basic/VirtualFileSystem.h:22:0,
                 from /home/daenzer/src/llvm-git/llvm/include/clang/Basic/FileManager.h:20,
                 from /home/daenzer/src/llvm-git/llvm/include/clang/Basic/SourceManager.h:38,
                 from /home/daenzer/src/llvm-git/llvm/include/clang/Frontend/CompilerInstance.h:16,
                 from ../../../../../src/gallium/state_trackers/clover/llvm/invocation.cpp:25:
/home/daenzer/src/llvm-git/llvm/include/llvm/Support/raw_ostream.h:512:8: note: declared here
   void flush() = delete;
        ^
Makefile:862: recipe for target 'llvm/libclllvm_la-invocation.lo' failed

Reviewed-by: Francisco Jerez <currojerez@riseup.net>
9 years agomesa: set correct error for non-renderable multisample textures
Tapani Pälli [Thu, 13 Aug 2015 06:30:35 +0000 (09:30 +0300)]
mesa: set correct error for non-renderable multisample textures

v2: same common error on gles31 and desktop OpenGL
    (spotted by Erik Faye-Lund)

Signed-off-by: Marta Lofstedt <marta.lofstedt@linux.intel.com>
Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Timothy Arceri <t_arceri@yahoo.com.au>
9 years agoi965/skl: Remove early platform support
Ben Widawsky [Fri, 7 Aug 2015 20:58:37 +0000 (13:58 -0700)]
i965/skl: Remove early platform support

We do not want bug reports from this early stepping of SKL. Few if any were ever
shipped outside of Intel to early enabling partners, and none will be sold.

There is a functional change here. If you're using new mesa on an old
kernel/libdrm, the revid will be -1, and we'll use new SKL values instead of
early ones (a hopefully irrelevant improvement IMO).

v2: Remove hunk which warned before dying. Instead, default to normal SKL
support (Ken)

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Neil Roberts <neil@linux.intel.com>
9 years agoegl: improve attribute checking for eglCreateContext
Frank Binns [Wed, 12 Aug 2015 15:36:00 +0000 (16:36 +0100)]
egl: improve attribute checking for eglCreateContext

The EGL 1.4 spec states for eglCreateContext:

"attribute EGL_CONTEXT_CLIENT_VERSION is only valid when the current
 rendering API is EGL_OPENGL_ES_API"

Additionally, if the EGL_KHR_create_context EGL extension is supported
(this is mandatory in EGL 1.5) then the EGL_CONTEXT_MAJOR_VERSION_KHR,
which is an alias for EGL_CONTEXT_CLIENT_VERSION, and
EGL_CONTEXT_MINOR_VERSION_KHR attributes are also accepted by
eglCreateContext with the extension spec stating:

"The values for attributes EGL_CONTEXT_MAJOR_VERSION_KHR and
 EGL_CONTEXT_MINOR_VERSION_KHR specify the requested client API
 version. They are only meaningful for OpenGL and OpenGL ES
 contexts, and specifying them for other types of contexts will
 generate an error."

Add the necessary checks against the extension and rendering APIs when
validating these attributes as part of eglCreateContext.

Signed-off-by: Frank Binns <frank.binns@imgtec.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
[Emil Velikov: Add newline before the spec quote (Matt)]
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
9 years agoegl: don't allow eglGetConfigs to set num_configs param to a negative value
Frank Binns [Wed, 12 Aug 2015 15:35:59 +0000 (16:35 +0100)]
egl: don't allow eglGetConfigs to set num_configs param to a negative value

When a buffer is provided to eglGetConfigs it's supposed to set the value
of the num_config parameter to the total number of configs that have been
copied into this buffer. For some reason the EGL spec doesn't consider it
to be an error to pass this function a buffer while specifying its size to
be less than 0. Given this, one would expect this combination to result in
the num_config parameter being set to 0 but this wasn't the case. This was
due to the buffer size being copied straight into num_configs without being
clamped to 0.

This was causing the following dEQP EGL test to fail:
dEQP-EGL.functional.query_config.get_configs.get_configs_bounds

Signed-off-by: Frank Binns <frank.binns@imgtec.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
9 years agoegl/x11: don't abort when creating a DRI2 drawable fails
Frank Binns [Tue, 4 Aug 2015 13:32:45 +0000 (14:32 +0100)]
egl/x11: don't abort when creating a DRI2 drawable fails

When calling either eglCreateWindowSurface or eglCreatePixmapSurface it
was possible for an application to be aborted as a result of it failing
to create a DRI2 drawable on the server. This could happen due to an
application passing in an invalid native drawable handle, for example.

v2: Handle the case where an error has been set on the connection

Cc: <mesa-stable@lists.freedesktop.org>
Signed-off-by: Frank Binns <frank.binns@imgtec.com>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
9 years agoegl/x11: set EGL_BAD_NATIVE_(PIXMAP|WINDOW) for invalid pixmaps/windows
Frank Binns [Tue, 4 Aug 2015 13:32:44 +0000 (14:32 +0100)]
egl/x11: set EGL_BAD_NATIVE_(PIXMAP|WINDOW) for invalid pixmaps/windows

Both eglCreatePixmapSurface and eglCreateWindowSurface were incorrectly
setting the EGL error to be EGL_BAD_ALLOC when an invalid native drawable
handle was being passed in. The EGL spec states the following for
eglCreatePixmapSurface:

"If pixmap is not a valid native pixmap handle, then an EGL_BAD_-
 NATIVE_PIXMAP error should be generated."

(eglCreateWindowSurface has similar text)

Correctly set the EGL error value based on xcb_get_geometry_reply returning
an error structure containing something other than BadAlloc.

v2: Check for BadAlloc error and update commit message to reflect this

Signed-off-by: Frank Binns <frank.binns@imgtec.com>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
9 years agoegl/x11: fix use of EGL_BAD_NATIVE_WINDOW
Frank Binns [Tue, 4 Aug 2015 13:32:43 +0000 (14:32 +0100)]
egl/x11: fix use of EGL_BAD_NATIVE_WINDOW

Commit 4ed23fd590 introduced some calls to _eglError inappropriately
passing it EGL_BAD_NATIVE_WINDOW. This was actually harmless in two of the
cases as _eglError gets called later on with a more appropriate error code
but (just to be safe) switch these to _eglLog calls instead.

The final case is a little trickier as it actually needs to set an error
of which the following are available (according to the EGL spec):
EGL_BAD_MATCH, EGL_BAD_CONFIG, EGL_BAD_NATIVE_(PIXMAP|WINDOW) and
EGL_BAD_ALLOC.

Of these, EGL_BAD_ALLOC seems to be the most appropriate given that
failure can occur either as a result of xcb_get_setup failing due to an
earlier error on the connection (where the most commonly occurring error
code is XCB_CONN_CLOSED_MEM_INSUFFICIENT) or as a result of the
xcb_screen_iterator_t 'rem' field being 0.

In addition to this, commit af2aea40d2 unconditionally set the error to
EGL_BAD_NATIVE_WINDOW when creating a window or pixmap surface with a NULL
native handle. Change this to correctly set the error based on surface
type.

v2: Updated patch description (Emil Velikov)
    Return EGL_BAD_NATIVE_PIXMAP when eglCreatePixmapSurface is called
    with a NULL native pixmap handle

Signed-off-by: Frank Binns <frank.binns@imgtec.com>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
9 years agomesa: remove extern from texture function
Timothy Arceri [Thu, 13 Aug 2015 08:42:54 +0000 (18:42 +1000)]
mesa: remove extern from texture function

Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
9 years agoglsl: make linker error message more informative
Timothy Arceri [Wed, 12 Aug 2015 07:01:52 +0000 (17:01 +1000)]
glsl: make linker error message more informative

Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
9 years agoi965: Stop aux data compare preventing program binary re-use
Topi Pohjolainen [Thu, 25 Jun 2015 11:00:41 +0000 (14:00 +0300)]
i965: Stop aux data compare preventing program binary re-use

Items in the program cache consist of three things: key, the data
representing the instructions and auxiliary data representing
uniform storage. The data consisting of instructions is stored into
a drm buffer object while the key and the auxiliary data reside in
malloced section. Now the cache uploading is equipped with a check
that iterates over existing items and seeks to find a another item
using identical instruction data than the one being just uploaded.
If such is found there is no need to add another section into the
drm buffer object holding identical copy of the existing one. The
item just being uploaded should instead simply point to the same
offset in the underlying drm buffer object.

Unfortunately the check for the matching instruction data is
coupled with a check for matching auxiliary data also. This
effectively prevents the cache from ever containing two items
that could share a section in the drm buffer object.

The constraint for the instruction data and auxiliary data to
match is, fortunately, unnecessary strong. When items are stored
into the cache they will anyway contain their own copy of the
auxiliary data (even if they matched - which they in real world
never will). The only thing the items would be sharing is the
instruction data and hence we should only check for that to match
and nothing else.

No piglit regression in jenkins.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
9 years agoi965: Only write program to cache when it doesn't exist yet
Topi Pohjolainen [Thu, 25 Jun 2015 11:35:26 +0000 (14:35 +0300)]
i965: Only write program to cache when it doesn't exist yet

Current logic re-writes the same data when existing data is found.
Not that this actually matters at the moment in practice, the
contraint for finding matching data is too severe to ever allow
data to be shared between two items in the cache.

Reviewed-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
9 years agoi965: Rename brw_upload_item_data to brw_alloc_item_data
Topi Pohjolainen [Thu, 25 Jun 2015 11:31:03 +0000 (14:31 +0300)]
i965: Rename brw_upload_item_data to brw_alloc_item_data

and simplify the interface to take directly the size and to return
the offset. The routine does nothing more than allocate, it doesn't
upload anything.

Reviewed-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
9 years agomesa: update MaxShaderStorageBlockSize to 2^27
Tapani Pälli [Wed, 12 Aug 2015 08:13:40 +0000 (11:13 +0300)]
mesa: update MaxShaderStorageBlockSize to 2^27

Extension spec originally required 2^24 but 2^27 is the minimum value
required by OpenGL 4.5 and OpenGL ES 3.1 specifications.

Fixes:
   ES31-CTS.shader_storage_buffer_object.basic-max

Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Timothy Arceri <t_arceri@yahoo.com.au>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
9 years agomesa: fix name returned for XFB varyings
Tapani Pälli [Mon, 3 Aug 2015 07:46:33 +0000 (10:46 +0300)]
mesa: fix name returned for XFB varyings

_mesa_get_program_resource_name has logic to append '[0]' in name
if variable is an array, this should be skipped for XFB varyings
that have array index already appended.

v2: fix comment, change also GL_NAME_LENGTH query to match
    the behaviour

Fixes:
   ES31-CTS.program_interface_query.transform-feedback-types

Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Martin Peres <martin.peres@linux.intel.com>
9 years agomesa: Fix printf format specifier warn of the ptrdiff_t
Edward O'Callaghan [Wed, 29 Jul 2015 11:49:45 +0000 (21:49 +1000)]
mesa: Fix printf format specifier warn of the ptrdiff_t

See §7.19.6.1, paragraph 7 of the ISO C specification.

Signed-off-by: Edward O'Callaghan <eocallaghan@alterapraxis.com>
Signed-off-by: Marek Olšák <marek.olsak@amd.com>
9 years agor600g: allow setting geometry shader sampler states
Marek Olšák [Tue, 11 Aug 2015 19:37:59 +0000 (21:37 +0200)]
r600g: allow setting geometry shader sampler states

We were ignoring them. This is both hilarious and sad.

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Edward O'Callaghan <eocallaghan at alterapraxis.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
9 years agor600g: fix polygon offset scale
Marek Olšák [Tue, 11 Aug 2015 20:36:51 +0000 (22:36 +0200)]
r600g: fix polygon offset scale

The value was copied from r300g, which uses 1/12 subpixels, but this hw
uses 1/16 subpixels.

Should fix piglit: gl-1.4-polygon-offset (formerly a glean test)
(untested, ported from radeonsi)

Reviewed-by: Edward O'Callaghan <eocallaghan at alterapraxis.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Cc: mesa-stable@lists.freedesktop.org
9 years agoradeonsi: fix polygon offset scale
Marek Olšák [Tue, 11 Aug 2015 20:36:51 +0000 (22:36 +0200)]
radeonsi: fix polygon offset scale

The value was copied from r300g, which uses 1/12 subpixels, but this hw
uses 1/16 subpixels.

Fixes piglit: gl-1.4-polygon-offset (formerly a glean test)

Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Cc: mesa-stable@lists.freedesktop.org
9 years agoradeonsi: enable VS_OUT_MISC_SIDE_BUS_ENA
Marek Olšák [Mon, 10 Aug 2015 00:28:01 +0000 (02:28 +0200)]
radeonsi: enable VS_OUT_MISC_SIDE_BUS_ENA

This is recommended for better performance.
Diag tests always enable this.

Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
9 years agoradeonsi: add support for gl_PrimitiveID in the fragment shader
Marek Olšák [Sun, 9 Aug 2015 23:50:11 +0000 (01:50 +0200)]
radeonsi: add support for gl_PrimitiveID in the fragment shader

It must be obtained from the VS.

The GS scenario A must be enabled for PrimID to be generated for the VS.

+ 4 piglits

Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
9 years agoradeonsi: move VGT_GS_MODE to the VS state
Marek Olšák [Sun, 9 Aug 2015 22:52:21 +0000 (00:52 +0200)]
radeonsi: move VGT_GS_MODE to the VS state

The VS will want to select GS scenario A here (VS with PrimitiveID).

Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
9 years agofreedreno/a4xx: format updates
Rob Clark [Wed, 12 Aug 2015 15:39:24 +0000 (11:39 -0400)]
freedreno/a4xx: format updates

Signed-off-by: Rob Clark <robclark@freedesktop.org>
9 years agofreedreno/a3xx+a4xx: add texture buffer object support
Rob Clark [Tue, 11 Aug 2015 20:47:16 +0000 (16:47 -0400)]
freedreno/a3xx+a4xx: add texture buffer object support

Basic texture buffer support.  Should be straightforward to add first/
last_element support.  And with a bit of work in ir3 emulate larger
texture buffer sizes.  But this seems to be enough for stk gl31 render
paths.

Signed-off-by: Rob Clark <robclark@freedesktop.org>
9 years agottn: add buffer texture type
Rob Clark [Tue, 11 Aug 2015 20:33:14 +0000 (16:33 -0400)]
ttn: add buffer texture type

Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Signed-off-by: Rob Clark <robclark@freedesktop.org>
9 years agofreedreno/ir3: 'keeps' need neighbors found too
Rob Clark [Tue, 11 Aug 2015 20:11:04 +0000 (16:11 -0400)]
freedreno/ir3: 'keeps' need neighbors found too

This shows up with a glamor shader, which does a TXF and uses the result
for conditional kill.  Before we wouldn't group the fanin (collect)
neighbors which need to be allocated adjacently at RA, resulting in
badness.

Signed-off-by: Rob Clark <robclark@freedesktop.org>
9 years agofreedreno/ir3/print: print left/right neighbors too
Rob Clark [Tue, 11 Aug 2015 20:09:48 +0000 (16:09 -0400)]
freedreno/ir3/print: print left/right neighbors too

When debugging compiler, this is useful to see.

Signed-off-by: Rob Clark <robclark@freedesktop.org>
9 years agofreedreno/ir3: use nir pass to lower const to scalar
Rob Clark [Tue, 11 Aug 2015 15:47:46 +0000 (11:47 -0400)]
freedreno/ir3: use nir pass to lower const to scalar

Signed-off-by: Rob Clark <robclark@freedesktop.org>
9 years agofreedreno/a4xx: point-size and spritelist fixes
Rob Clark [Tue, 11 Aug 2015 12:48:34 +0000 (08:48 -0400)]
freedreno/a4xx: point-size and spritelist fixes

a4xx needs similar treatment as 995f55a6

Also fixup a few point-size and vpsrepl issues and drop fix_blit_fp()
hack previously needed for mem2gmem.

Signed-off-by: Rob Clark <robclark@freedesktop.org>
9 years agofreedreno: cap cleanups
Rob Clark [Tue, 11 Aug 2015 00:41:45 +0000 (20:41 -0400)]
freedreno: cap cleanups

Move a few things around to group stuff that is common to a3xx/a4xx
together.  Also, introduce is_ir3() for things that are more specific to
the compiler / shader-ISA than to the gpu generation.

Signed-off-by: Rob Clark <robclark@freedesktop.org>
9 years agomesa: add NV_read_{depth,stencil,depth_stencil} extensions
Rob Clark [Mon, 10 Aug 2015 10:58:37 +0000 (06:58 -0400)]
mesa: add NV_read_{depth,stencil,depth_stencil} extensions

These extensions allow reading depth/stencil for GLES contexts, which is
useful for tools like apitrace.

Signed-off-by: Rob Clark <robclark@freedesktop.org>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
9 years agoi965/shader: Don't use OptimizeForAOS for NIR vec4 vertex shaders
Jason Ekstrand [Thu, 6 Aug 2015 21:26:47 +0000 (14:26 -0700)]
i965/shader: Don't use OptimizeForAOS for NIR vec4 vertex shaders

Shader-db results for vec4 programs using NIR on HSW:

   total instructions in shared programs: 1838157 -> 1828469 (-0.53%)
   instructions in affected programs:     275978 -> 266290 (-3.51%)
   helped:                                2827
   HURT:                                  244
   GAINED:                                0
   LOST:                                  0

Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Eduardo Lima Mitev <elima@igalia.com>
9 years agomesa/teximage: report the correct function which triggered the error
Nanley Chery [Fri, 7 Aug 2015 23:37:47 +0000 (16:37 -0700)]
mesa/teximage: report the correct function which triggered the error

This function would always report that a dimension or size error occurred
in glTexImage even when it was called from glCompressedTexImage. Replace
the static string with the dynamically determined caller name.

Reviewed-by: Tapani Palli <tapani.palli@intel.com>
Signed-off-by: Nanley Chery <nanley.g.chery@intel.com>
9 years agomesa/formats: don't byteswap when building array formats
Oded Gabbay [Wed, 12 Aug 2015 15:22:53 +0000 (18:22 +0300)]
mesa/formats: don't byteswap when building array formats

Because we build here an array format, we don't need to swap the
bytes for big endian.
If it isn't an array format, the bytes will be swapped in
_mesa_format_convert.

v2: remove temp variable

Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
Cc: "10.5 10.6" <mesa-stable@lists.freedesktop.org>
9 years agomesa/formats: Don't flip channels of null array formats
Jason Ekstrand [Mon, 10 Aug 2015 08:32:23 +0000 (01:32 -0700)]
mesa/formats: Don't flip channels of null array formats

Before, if we encountered an array format of 0 on a BE system, we would
flip all the channels even though it's an invalid format.  This would
result in a mostly invalid format with a swizzle of yyyy or wwww.  Instead,
we should just return 0 if the array format stashed in the format info is
invalid.

Cc: "10.6 10.5" <mesa-stable@lists.freedesktop.org>
9 years agomesa/formats: Fix swizzle flipping for big-endian targets
Jason Ekstrand [Mon, 10 Aug 2015 06:45:44 +0000 (23:45 -0700)]
mesa/formats: Fix swizzle flipping for big-endian targets

The swizzle defines where in the format you should look for any given
channel.  When we flip the format around for BE targets, we need to change
the destinations of the swizzles, not the sources.  For example, say the
format is an RGBX format with a swizzle of xyz1 on LE.  Then it should be
wzy1 on BE;  however, the code as it was before, would have made it 1zyx on
BE which is clearly wrong.

Reviewed-by: Iago Toral <itoral@igalia.com>
Reviewed-by: Oded Gabbay <oded.gabbay@gmail.com>
Cc: "10.6 10.5" <mesa-stable@lists.freedesktop.org>
9 years agomesa/formats: Only do byteswapping for packed formats
Jason Ekstrand [Sat, 8 Aug 2015 16:00:21 +0000 (09:00 -0700)]
mesa/formats: Only do byteswapping for packed formats

Reviewed-by: Iago Toral <itoral@igalia.com>
Cc: "10.6 10.5" <mesa-stable@lists.freedesktop.org>
9 years agoconfigure.ac: Always define __STDC_LIMIT_MACROS.
Matt Turner [Tue, 11 Aug 2015 22:21:03 +0000 (15:21 -0700)]
configure.ac: Always define __STDC_LIMIT_MACROS.

... which ensures that we get defines like LONG_MAX in C++.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=91591
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
9 years agoi965: Optimize brw_inst_set_bits() and brw_compact_inst_set_bits().
Matt Turner [Tue, 11 Aug 2015 01:50:48 +0000 (18:50 -0700)]
i965: Optimize brw_inst_set_bits() and brw_compact_inst_set_bits().

Cuts about 2k of .text.

   text     data      bss      dec      hex  filename
5017141   197160    27672  5241973   4ffc75  i965_dri.so before
5014981   197160    27672  5239813   4ff405  i965_dri.so after

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
9 years agoi965: Optimize brw_inst_bits() and brw_compact_inst_bits().
Matt Turner [Mon, 10 Aug 2015 23:57:58 +0000 (16:57 -0700)]
i965: Optimize brw_inst_bits() and brw_compact_inst_bits().

Cuts about 1k of .text.

   text     data      bss      dec      hex  filename
5018165   197160    27672  5242997   500075  i965_dri.so before
5017141   197160    27672  5241973   4ffc75  i965_dri.so after

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
9 years agodocs: add news item and link release notes for 10.6.4
Emil Velikov [Tue, 11 Aug 2015 18:00:03 +0000 (19:00 +0100)]
docs: add news item and link release notes for 10.6.4

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
9 years agodocs: add sha256 checksums for 10.6.4
Emil Velikov [Tue, 11 Aug 2015 17:54:18 +0000 (18:54 +0100)]
docs: add sha256 checksums for 10.6.4

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
(cherry picked from commit 99793e2541510fe208d29e69fedf97a6fff006f8)

9 years agodocs: add release notes for 10.6.4
Emil Velikov [Tue, 11 Aug 2015 15:39:10 +0000 (16:39 +0100)]
docs: add release notes for 10.6.4

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
(cherry picked from commit 6b2fcee64edadbd4db2293f5f4fc1a70e80c7251)

9 years agogallium/radeon: fix r600g build if LLVM is disabled
Marek Olšák [Sat, 8 Aug 2015 12:03:54 +0000 (14:03 +0200)]
gallium/radeon: fix r600g build if LLVM is disabled

MESA_LLVM_VERSION_PATCH is undefined.

Reviewed-by: Edward O'Callaghan <eocallaghan at alterapraxis.com>
Tested-by: Benjamin Bellec <b.bellec@gmail.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
9 years agor600g: use a bitfield to track dirty atoms
Grazvydas Ignotas [Sun, 9 Aug 2015 21:42:35 +0000 (00:42 +0300)]
r600g: use a bitfield to track dirty atoms

r600 currently has 73 atoms and looping through their dirty flags has
become costly because checking each flag requires a pointer
dereference before the read. To avoid having to do that add additional
bitfield which can be checked really quickly thanks to tzcnt instruction.

id field was added to struct r600_atom but that doesn't affect memory
usage for both 32 and 64 bit CPUs because it was stuffed into padding.

The performance improvement is ~2% for benchmarks that can have FPS in
the thousands but is hardly measurable in "real" programs.

Signed-off-by: Marek Olšák <marek.olsak@amd.com>
9 years agor600g: don't mark unused atom dirty
Grazvydas Ignotas [Sun, 9 Aug 2015 21:42:34 +0000 (00:42 +0300)]
r600g: don't mark unused atom dirty

On evergreen config_state is not used, so don't mark it dirty.

Signed-off-by: Marek Olšák <marek.olsak@amd.com>
9 years agor600g: use a helper to add an initialized atom
Grazvydas Ignotas [Sun, 9 Aug 2015 21:42:33 +0000 (00:42 +0300)]
r600g: use a helper to add an initialized atom

Instead of writing to rctx->atoms directly use a helper to take
advantage of assert checks.

Signed-off-by: Marek Olšák <marek.olsak@amd.com>
9 years agogallium/radeon: use helper functions to mark atoms dirty
Grazvydas Ignotas [Sun, 9 Aug 2015 21:42:32 +0000 (00:42 +0300)]
gallium/radeon: use helper functions to mark atoms dirty

This is analogous to r300_mark_atom_dirty() used by r300, and will
be used by later patches. For common radeon code, appropriate helper
is called through a function pointer.

No functional changes.

Signed-off-by: Marek Olšák <marek.olsak@amd.com>
9 years agodocs: Mark ARB_shader_image_load_store as done on i965.
Francisco Jerez [Fri, 1 May 2015 14:00:02 +0000 (17:00 +0300)]
docs: Mark ARB_shader_image_load_store as done on i965.

9 years agoi965: Expose ARB_shader_image_load_store.
Francisco Jerez [Thu, 7 May 2015 15:56:01 +0000 (18:56 +0300)]
i965: Expose ARB_shader_image_load_store.

Reviewed-by: Paul Berry <stereotype441@gmail.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
9 years agoi965/fs: Clamp image array indices to the array bounds on IVB.
Francisco Jerez [Mon, 27 Jul 2015 11:33:06 +0000 (14:33 +0300)]
i965/fs: Clamp image array indices to the array bounds on IVB.

This fixes the spec@arb_shader_image_load_store@invalid index bounds
piglit tests on IVB, which were causing a GPU hang and then a crash
due to the invalid binding table index result of the array index
calculation.  Other generations seem to behave sensibly when an
invalid surface is provided so it doesn't look like we need to care.

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>
9 years agoi965/fs: Translate image load, store and atomic NIR intrinsics.
Francisco Jerez [Mon, 27 Jul 2015 13:26:52 +0000 (16:26 +0300)]
i965/fs: Translate image load, store and atomic NIR intrinsics.

v2: Move array coordinate workaround into the surface builder.

Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
9 years agoi965/fs: Handle image uniforms in NIR programs.
Francisco Jerez [Sun, 28 Jun 2015 18:16:31 +0000 (21:16 +0300)]
i965/fs: Handle image uniforms in NIR programs.

v2: Move the image_params array back to brw_stage_prog_data.

Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
9 years agoi965: Implement logic to set up and upload an image uniform.
Francisco Jerez [Tue, 5 May 2015 18:07:15 +0000 (21:07 +0300)]
i965: Implement logic to set up and upload an image uniform.

v2: Move the image_params array back to brw_stage_prog_data.

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
9 years agoi965: Teach type_size() about the size of an image uniform.
Francisco Jerez [Tue, 5 May 2015 18:05:45 +0000 (21:05 +0300)]
i965: Teach type_size() about the size of an image uniform.

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
9 years agoi965/fs: Implement image load, store and atomic.
Francisco Jerez [Thu, 30 Jul 2015 12:46:40 +0000 (15:46 +0300)]
i965/fs: Implement image load, store and atomic.

v2: Drop VEC4 suport.
v3: Rebase.
v4: Move array coordinate workaround into the surface builder.

Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
9 years agoi965/fs: Import image format conversion primitives.
Francisco Jerez [Thu, 30 Jul 2015 12:51:58 +0000 (15:51 +0300)]
i965/fs: Import image format conversion primitives.

Define bitfield packing, unpacking and type conversion operations in
terms of which the image format conversion code will be implemented.
These don't directly know about image formats: The packing and
unpacking functions take a 4-tuple of bit shifts and a 4-tuple of bit
widths as arguments, determining the bitfield position of each
component.  Most of the remaining functions perform integer, fixed
point normalized, and floating point type conversions, mapping between
a target type with per-component bit widths given by a parameter and a
matching native representation of the same type.

v2: Drop VEC4 suport.
v3: Rebase.
v4: Fix clamping of negative floats in the unsigned case of
    emit_convert_to_scaled().

Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
9 years agoi965/fs: Import image format metadata queries.
Francisco Jerez [Wed, 22 Apr 2015 13:45:28 +0000 (16:45 +0300)]
i965/fs: Import image format metadata queries.

Define some utility functions to query the bitfield layout of a given
image format and whether it satisfies a number of more or less
hardware-specific properties.

v2: Drop VEC4 suport.
v3: Add SKL support.

Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
9 years agoi965/fs: Import code to transform image coordinates into surface coordinates.
Francisco Jerez [Thu, 23 Jul 2015 16:32:08 +0000 (19:32 +0300)]
i965/fs: Import code to transform image coordinates into surface coordinates.

Accounting for the padding required for 1D arrays in certain cases.

Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
9 years agoi965/fs: Import image memory offset calculation code.
Francisco Jerez [Wed, 22 Apr 2015 13:44:18 +0000 (16:44 +0300)]
i965/fs: Import image memory offset calculation code.

Define a function to calculate the memory address of the image
location given by a vector of coordinates.  This is required in cases
where we need to fall back to untyped surface access, which take a raw
memory offset and know nothing about surface coordinates, type
conversion or memory tiling and swizzling.  They are still useful
because typed surface reads don't support any 64 or 128-bit formats on
IVB, and they don't support any 128-bit formats on HSW and BDW.

The tiling algorithm is implemented based on a number of parameters
which are passed in as uniforms and determine whether the surface
layout is X-tiled, Y-tiled or untiled.  This allows binding surfaces
of different tiling layouts to the pipeline without recompiling the
program.

v2: Drop VEC4 suport.
v3: Rebase.
v4: Add plenty of comments (Jason).

Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
9 years agoi965/fs: Import image access validity checks.
Francisco Jerez [Wed, 22 Apr 2015 13:43:51 +0000 (16:43 +0300)]
i965/fs: Import image access validity checks.

These utility functions check whether an image access is valid.
According to the spec an invalid image access should have no effect on
the image and yield well-defined results.  Typically the hardware
implements correct bounds and surface checking by itself, but in some
cases (typed atomics on IVB and untyped messages elsewhere) we need to
implement it in software to work around lacking hardware support.

v2: Drop VEC4 suport.
v3: Rebase.

Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
9 years agoi965: Define implementation constants for ARB_shader_image_load_store.
Francisco Jerez [Sat, 23 Nov 2013 00:00:33 +0000 (16:00 -0800)]
i965: Define implementation constants for ARB_shader_image_load_store.

Reviewed-by: Paul Berry <stereotype441@gmail.com>
v2: Drop VS support pre-Gen8, drop GS support.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
9 years agoi965/gen7-8: Set up early depth/stencil control appropriately for image load/store.
Francisco Jerez [Mon, 9 Feb 2015 19:04:53 +0000 (21:04 +0200)]
i965/gen7-8: Set up early depth/stencil control appropriately for image load/store.

v2: Store early fragment test mode in brw_wm_prog_data instead of
    getting it from core mesa data structures (Ken).

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
9 years agoi965/gen7-8: Poke the 3DSTATE UAV access enable bits.
Francisco Jerez [Mon, 13 Jul 2015 11:21:07 +0000 (14:21 +0300)]
i965/gen7-8: Poke the 3DSTATE UAV access enable bits.

v2: Set the PS UAV-only bit on HSW (Ken).

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
9 years agoi965/gen7: Enable fragment shader dispatch if the program has image uniforms.
Francisco Jerez [Tue, 3 Feb 2015 15:14:10 +0000 (17:14 +0200)]
i965/gen7: Enable fragment shader dispatch if the program has image uniforms.

Shaders with image uniforms may have side effects.  Make sure that
fragment shader threads are dispatched if the shader has any image
uniforms.

v2: Use brw_stage_prog_data::nr_image_params to find out if the shader
    has image uniforms instead of checking core mesa data structures
    (Ken).

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
9 years agoi965: Hook up image state upload.
Francisco Jerez [Mon, 20 Jul 2015 14:13:17 +0000 (17:13 +0300)]
i965: Hook up image state upload.

v2: Add CS support.  Move the image_params array back to
    brw_stage_prog_data.

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Acked-by: Jason Ekstrand <jason@jlekstrand.net>
9 years agoi965: Reserve enough parameter entries for all image uniforms used in the program.
Francisco Jerez [Mon, 13 Jul 2015 14:19:29 +0000 (17:19 +0300)]
i965: Reserve enough parameter entries for all image uniforms used in the program.

v2: Add CS support.

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
9 years agoi965: Define and initialize image parameter structure.
Francisco Jerez [Wed, 21 Jan 2015 15:34:49 +0000 (17:34 +0200)]
i965: Define and initialize image parameter structure.

This will be used to pass image meta-data to the shader when we cannot
use typed surface reads and writes.  All entries except surface_idx
and size are otherwise unused and will get eliminated by the uniform
packing pass.  size will be used for bounds checking with some image
formats and will be useful for ARB_shader_image_size too.  surface_idx
is always used.

v2: Add CS support.  Move the image_params array back to
    brw_stage_prog_data.
v3: Improve documentation.

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
9 years agoi965: Implement surface state set-up for shader images.
Francisco Jerez [Sat, 2 May 2015 13:58:24 +0000 (16:58 +0300)]
i965: Implement surface state set-up for shader images.

v2: Add SKL support.

Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
9 years agoi965: Fix brw_memory_barrier() for SKL.
Francisco Jerez [Tue, 12 May 2015 12:56:54 +0000 (15:56 +0300)]
i965: Fix brw_memory_barrier() for SKL.

This works as-is on SKL, only the assertion needs to be relaxed.

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
9 years agoi965: Add SKL support to brw_miptree_get_horizontal_slice_pitch().
Francisco Jerez [Tue, 12 May 2015 13:10:07 +0000 (16:10 +0300)]
i965: Add SKL support to brw_miptree_get_horizontal_slice_pitch().

Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>
9 years agoglsl: Add missing spec quote about atomic counter in structs
Timothy Arceri [Sun, 9 Aug 2015 04:44:30 +0000 (14:44 +1000)]
glsl: Add missing spec quote about atomic counter in structs

Reviewed-by: Thomas Helland <thomashelland90@gmail.com>
9 years agoradeonsi: add new OLAND pci id
Alex Deucher [Mon, 10 Aug 2015 19:35:21 +0000 (15:35 -0400)]
radeonsi: add new OLAND pci id

Reviewed-by: Edward O'Callaghan <eocallaghan@alterapraxis.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: mesa-stable@lists.freedesktop.org
9 years agonouveau: no need to do tnl wakeup, state updates are always hooked up
Ilia Mirkin [Mon, 10 Aug 2015 21:41:36 +0000 (17:41 -0400)]
nouveau: no need to do tnl wakeup, state updates are always hooked up

A TNL state update now requires a DrawBuffer to be set, which it isn't
early on in context creation. Since we init swtnl from context init,
this caused crashes.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=91570
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Cc: "10.6" <mesa-stable@lists.freedesktop.org>
9 years agoi965/fs: Make resolve_source_modifiers consistent with the vec4 version
Jason Ekstrand [Mon, 10 Aug 2015 18:52:50 +0000 (11:52 -0700)]
i965/fs: Make resolve_source_modifiers consistent with the vec4 version

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agoi965/vec4_visitor: Make some function arguments const references
Jason Ekstrand [Mon, 10 Aug 2015 18:48:14 +0000 (11:48 -0700)]
i965/vec4_visitor: Make some function arguments const references

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agoi965/fs: Don't do redundant RA setup on IVB+
Jason Ekstrand [Fri, 31 Jul 2015 15:36:35 +0000 (08:36 -0700)]
i965/fs: Don't do redundant RA setup on IVB+

Acked-by: Matt Turner <mattst88@gmail.com>
9 years agoi965/fs: Use dispatch_width instead of reg_width in alloc_reg_sets
Jason Ekstrand [Fri, 31 Jul 2015 15:35:57 +0000 (08:35 -0700)]
i965/fs: Use dispatch_width instead of reg_width in alloc_reg_sets

reg_width is kind of an outdated concept.

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agora: Delete the conflict lists in ra_set_finalize
Jason Ekstrand [Fri, 31 Jul 2015 03:53:04 +0000 (20:53 -0700)]
ra: Delete the conflict lists in ra_set_finalize

They are never used after the set is finalized so there's no reason to keep
them around.

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agora: Refactor ra_set_finalize
Jason Ekstrand [Fri, 31 Jul 2015 03:49:22 +0000 (20:49 -0700)]
ra: Refactor ra_set_finalize

All this commit does is change an early return to an if with an else
clause.

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agoi965/vec4_nir: Properly handle integer multiplies on BDW+
Jason Ekstrand [Mon, 3 Aug 2015 22:21:59 +0000 (15:21 -0700)]
i965/vec4_nir: Properly handle integer multiplies on BDW+

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agoi965/vec4_nir: Do boolean source modifier resolves on BDW+
Jason Ekstrand [Mon, 3 Aug 2015 21:37:41 +0000 (14:37 -0700)]
i965/vec4_nir: Do boolean source modifier resolves on BDW+

On BDW+, the negation source modifier on NOT, AND, OR, and XOR, is actually
a boolean negate and not an integer negate.  However, NIR's soruce
modifiers are the integer version.  We have to resolve it with a MOV prior
to emitting the actual instruction.  This is basically the same thing we do
in the FS backend.

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agoi965/vec4-nir: Handle boolean resolvese on ILK-
Jason Ekstrand [Mon, 3 Aug 2015 17:00:38 +0000 (10:00 -0700)]
i965/vec4-nir: Handle boolean resolvese on ILK-

The analysis code was already there and running, we just weren't doing
anything with the result of it yet.

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agoi965/nir: Don't mark bany or ball instructions for resolve
Jason Ekstrand [Mon, 3 Aug 2015 23:25:18 +0000 (16:25 -0700)]
i965/nir: Don't mark bany or ball instructions for resolve

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agoi965/nir: Use nir_op_info.output_type for determining when to resolve
Jason Ekstrand [Mon, 3 Aug 2015 21:12:35 +0000 (14:12 -0700)]
i965/nir: Use nir_op_info.output_type for determining when to resolve

Previously, we were explicitly listing every instruction that needs a
resolve.  However, those instructions were precicely the ones that returned
booleans so there's no reason why we shouldn't just have that check.  Also,
all of the reduction opcodes such as bany and ball were missing so it
didn't properly flag stuff on vec4.  If an opcode gets added in the future
that returns a bool but doesn't need a resolve, we can special-case that.

Reviewed-by: Matt Turner <mattst88@gmail.com>
9 years agomesa/format_utils: Add src_bits == dst_bits cases to unorm_to_unorm
Jason Ekstrand [Mon, 10 Aug 2015 05:03:00 +0000 (22:03 -0700)]
mesa/format_utils: Add src_bits == dst_bits cases to unorm_to_unorm

This better ensures that the src_bits == dst_bits case gets optimized away.

Reviewed-by: Neil Roberts <neil@linux.intel.com>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
9 years agogallium/radeon: add a debug flag not to use write combining (v2)
Marek Olšák [Sun, 2 Aug 2015 14:22:43 +0000 (16:22 +0200)]
gallium/radeon: add a debug flag not to use write combining (v2)

v2: just clear the flag before the allocation

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