Jason Ekstrand [Fri, 15 Apr 2016 20:52:34 +0000 (13:52 -0700)]
Merge branch 'vulkan'
Roland Scheidegger [Fri, 15 Apr 2016 19:11:52 +0000 (21:11 +0200)]
llvmpipe: (trivial) initialize src1_alpha var to NULL
The blend code would do a conditional assignment based on it, causing valgrind
to complain. Since that variable was actually unused in this case, this
doesn't fix anything but the warning.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94955
Cc: "11.1 11.2" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Brian Paul <brianp@vmware.com>
Jason Ekstrand [Fri, 15 Apr 2016 20:34:04 +0000 (13:34 -0700)]
Merge remote-tracking branch 'public/master' into vulkan
Jason Ekstrand [Thu, 14 Apr 2016 20:04:22 +0000 (13:04 -0700)]
configure: Add support for the Intel Vulkan driver
This adds a --with-vulkan-drivers option with one driver, "intel". In the
future, we may add more drivers to this list.
v2: Don't enable any drivers by default. This should prevent this patch
from breaking anyone's build.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Thu, 14 Apr 2016 17:45:24 +0000 (10:45 -0700)]
i965/surface_formats: Update some formats for more recent gens
The surface format table hasn't entirely been kept up-to-date. This commit
marks a couple more compressed formats as sampleable on gen8+ and adds the
A4B4G4R4 format as renderable on gen9.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Thu, 30 Jul 2015 18:28:22 +0000 (11:28 -0700)]
util/list: Add list splicing functions
This adds functions for splicing one list into another. These have
more-or-less the same API as the kernel list splicing functions. The
implementation, however, was stolen from the Wayland list implementation.
Reviewed-by: Mark Janes <mark.a.janes@intel.com>
Reviewed-by: Rob Clark <robclark@freedesktop.org>
Jason Ekstrand [Fri, 15 Apr 2016 20:17:08 +0000 (13:17 -0700)]
Remove the Intel Vulkan readme
Tim Rowley [Thu, 14 Apr 2016 23:23:33 +0000 (18:23 -0500)]
gallium/swr: confine c++11 flag to swr driver
On the philosophy that a driver shouldn't change the compile flags
for the entire tree, take the clove approach of moving the c++11 flag
to the swr driver directory.
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
Tim Rowley [Fri, 15 Apr 2016 17:38:25 +0000 (12:38 -0500)]
gallium/swr: allow swr use as a swrast dri driver
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
Tested-by: Ilia Mirkin <imirkin@alum.mit.edu>
Eric Anholt [Fri, 15 Apr 2016 18:43:00 +0000 (11:43 -0700)]
vc4: Fix subimage accesses to LT textures.
This code started out like the T case, iterating over utile offsets, but I
had partially switched it to iterating over pixel offsets. I hadn't
caught this before because it's unusual to do piecemeal uploads to small
textures.
Fixes bad text rendering in QT5 apps, which use a 256x16 glyph cache.
Also fixes 6 piglit tests related to glTexSubImage() and
glGetTexSubImage().
Cc: "11.1 11.2" <mesa-stable@lists.freedesktop.org>
Mark Janes [Tue, 12 Apr 2016 18:52:53 +0000 (11:52 -0700)]
util: Fix race condition on libgcrypt initialization
Fixes intermittent Vulkan CTS failures within the test groups:
dEQP-VK.api.object_management.multithreaded_per_thread_device
dEQP-VK.api.object_management.multithreaded_per_thread_resources
dEQP-VK.api.object_management.multithreaded_shared_resources
Signed-off-by: Mark Janes <mark.a.janes@intel.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94904
Reviewed-by: Edward O'Callaghan <eocallaghan@alterapraxis.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Jason Ekstrand [Fri, 15 Apr 2016 16:54:42 +0000 (09:54 -0700)]
i965: Default to scalar GS
Jason Ekstrand [Fri, 15 Apr 2016 16:48:56 +0000 (09:48 -0700)]
i965/surface_formats: Mark A4B4G4R4_UNORM as SKL+ only
This is what is indicated by the bspec.
Jason Ekstrand [Fri, 15 Apr 2016 16:45:55 +0000 (09:45 -0700)]
Revert "i965/fs: Properly write-mask spills"
This reverts commit
9c0109a1f6e0c9c8dd81fc215f8c0c5976e05790.
Jason Ekstrand [Fri, 15 Apr 2016 16:53:42 +0000 (09:53 -0700)]
Revert "i965/fs: Feel free to spill partial reads/writes"
This reverts commit
2434ceabf41e66f2a3627ea8591e5ca427a78cce.
Jason Ekstrand [Thu, 14 Apr 2016 20:02:48 +0000 (13:02 -0700)]
configure: Add support for detecting valgrind headers
We have several places where the Vulkan driver explicitly hooks into
valgrind when it's available. We need to be able to detect it.
Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
Eduardo Lima Mitev [Fri, 15 Apr 2016 08:00:05 +0000 (10:00 +0200)]
nir/print: Fix printing variable mode
nir_variable_mode is currently a bitflag enum, while
nir_print::print_var_decl() assumes is still a numbered list.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
John Sheu [Fri, 1 Apr 2016 23:52:22 +0000 (16:52 -0700)]
xlib: remove MESA_GLX_VISUAL_HACK
This removes a hack introduced in 1999 in the first version of
fakeglx.c, with the comment:
/* XXX revisit this after 3.0 is finished. */
Mesa 4.0 was released in 2001. It is now 2016, and Mesa 11.0 was
released last year.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
John Sheu [Fri, 1 Apr 2016 23:52:21 +0000 (16:52 -0700)]
xlib: fix leaks of returned values from XGetVisualInfo
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
John Sheu [Fri, 1 Apr 2016 23:52:20 +0000 (16:52 -0700)]
xlib: fix memory leak of and remove vishandle from XMesaVisualInfo
The vishandle member of XMesaVisualInfo is used to support the
comparison of XVisualInfo instances by pointer value, in
find_glx_visual(). The comparison however will always be false, as in
every case the comparison is made, the VisualInfo instance being
compared to is a new allocation passed in through a GLX API call.
In addition, the XVisualInfo instance pointed to by vishandle is itself
never freed, causing a memory leak. Since vishandle is essentially
useless, we just remove it and thereby also fix the leak.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
John Sheu [Wed, 13 Apr 2016 20:57:42 +0000 (13:57 -0700)]
xlib: do not cache return value of glXChooseVisual/glXGetVisualFromFBConfig
The returned XVisualInfo from glXChooseVisual/glXGetVisualFromFBConfig
is being cached in XMesaVisual.vishandle (and unconditionally
overwritten on subsequent calls). However, these entry points are
specified to return XVisualInfo instances to be owned by the caller and
freed with XFree(), so the return values should not be retained.
With this change, XMesaVisual.vishandle is essentially unused and will
be removed in a subsequent change.
v2: update commit message
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Jason Ekstrand [Fri, 15 Apr 2016 01:30:52 +0000 (18:30 -0700)]
Merge remote-tracking branch 'public/master' into vulkan
Jason Ekstrand [Fri, 15 Apr 2016 01:23:59 +0000 (18:23 -0700)]
main/mtypes: Remove the "set" parameter from gl_uniform_block
This is a left-over from the early days of the Vulkan driver
Jason Ekstrand [Fri, 15 Apr 2016 01:22:08 +0000 (18:22 -0700)]
Revert "i965/vec4: Add support for SHADER_OPCODE_MOV_INDIRECT"
This reverts commit
4115648a6be2e846660a35a0e260ae53b809b7e0. This commit
was half-baked and probably never should have been committed. We'll add
this back in properly later when we need it.
Jason Ekstrand [Thu, 14 Apr 2016 17:42:29 +0000 (10:42 -0700)]
i965: Expose the surface format table
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Thu, 14 Apr 2016 17:31:27 +0000 (10:31 -0700)]
nir/lower_io: Add UBOs and SSBOs to get_io_offset_src
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Fri, 25 Mar 2016 17:17:28 +0000 (10:17 -0700)]
nir/intrinsics: Add a vulkan_resource_index intrinsic
This is used to facilitate the Vulkan binding model where each resource is
described by a (descriptor set, binding, array index) tuple.
Reviewed-by: Rob Clark <robdclark@gmail.com>
Jason Ekstrand [Fri, 25 Mar 2016 17:16:23 +0000 (10:16 -0700)]
nir: Add a descriptor_set field to nir_variable
This is needed for supporting the Vulkan binding model
Reviewed-by: Rob Clark <robdclark@gmail.com>
Chad Versace [Mon, 11 Apr 2016 22:12:15 +0000 (15:12 -0700)]
dri: Fix robust context creation via EGL attribute
driCreateContextAttribs() emits an error if bit
__DRI_CTX_FLAG_ROBUST_BUFFER_ACCESS is set for an ES context. But,
EGL_EXT_create_context_robustness and EGL 1.5 both allow creation of
robust ES contexts. One requests a robust ES context by setting the
EGL_CONTEXT_OPENGL_ROBUST_ACCESS *attribute*, which Mesa's EGL layer
translates into the __DRI_CTX_FLAG_ROBUST_BUFFER_ACCESS *bit*.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Jason Ekstrand [Fri, 15 Apr 2016 00:14:28 +0000 (17:14 -0700)]
Merge remote-tracking branch 'public/master' into vulkan
Leo Liu [Tue, 12 Apr 2016 19:19:31 +0000 (15:19 -0400)]
radeon/uvd: fix tonga feedback buffer size
This only applies to tonga
Signed-off-by: Leo Liu <leo.liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Cc: "11.1 11.2" <mesa-stable@lists.freedesktop.org>
Jason Ekstrand [Tue, 5 Apr 2016 23:22:26 +0000 (16:22 -0700)]
i965: Push everything if pull_param == NULL
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Wed, 9 Dec 2015 01:34:38 +0000 (17:34 -0800)]
i965/fs: Push small uniform arrays
Unfortunately, this also means that we need to use a slightly different
algorithm for assign_constant_locations. The old algorithm worked based on
the assumption that each read of a uniform value read exactly one float.
If it encountered a MOV_INDIRECT, it would immediately bail and push the
whole thing. Since we can now read ranges using MOV_INDIRECT, we need to
be able to push a series of floats without breaking them up. To do this,
we use an algorithm similar to the on in split_virtual_grfs.
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Acked-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Wed, 9 Dec 2015 01:14:49 +0000 (17:14 -0800)]
i965/fs: Rename demote_pull_constants to lower_constant_loads
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Wed, 9 Dec 2015 01:02:16 +0000 (17:02 -0800)]
i965/vec4: Get rid of the uniform_size array
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Wed, 25 Nov 2015 17:36:34 +0000 (09:36 -0800)]
i965/vec4: Use MOV_INDIRECT instead of reladdr for indirect push constants
This commit moves us to an instruction based model rather than a
register-based model for indirects. This is more accurate anyway as we
have to emit instructions to resolve the reladdr. It's also a lot simpler
because it gets rid of the recursive reladdr problem by design.
One side-effect of this is that we need a whole new algorithm in
move_uniform_array_access_to_pull_constants. This new algorithm is much
more straightforward than the old one and is fairly similar to what we're
already doing in the FS backend.
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Acked-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Wed, 25 Nov 2015 01:02:01 +0000 (17:02 -0800)]
i965/fs: Get rid of the param_size array
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Tue, 24 Nov 2015 23:16:14 +0000 (15:16 -0800)]
i965/fs: Stop relying on param_size in assign_constant_locations
Now that we have MOV_INDIRECT opcodes, we have all of the size information
we need directly in the opcode. With a little restructuring of the
algorithm used in assign_constant_locations we don't need param_size
anymore. The big thing to watch out for now, however, is that you can have
two ranges overlap where neither contains the other. In order to deal with
this, we make the first pass just flag what needs pulling and handle
assigning pull constant locations until later.
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Acked-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Wed, 25 Nov 2015 00:54:41 +0000 (16:54 -0800)]
i965/fs: Get rid of reladdr
We aren't using it anymore.
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Tue, 24 Nov 2015 23:12:20 +0000 (15:12 -0800)]
i965/fs: Use MOV_INDIRECT for all indirect uniform loads
Instead of using reladdr, this commit changes the FS backend to emit a
MOV_INDIRECT whenever we need an indirect uniform load. We also have to
rework some of the other bits of the backend to handle this new form of
uniform load. The obvious change is that demote_pull_constants now acts
more like a lowering pass when it hits a MOV_INDIRECT.
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Acked-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Tue, 24 Nov 2015 21:52:49 +0000 (13:52 -0800)]
nir: Add another index to load_uniform to specify the range read
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Tue, 24 Nov 2015 17:01:11 +0000 (09:01 -0800)]
i965/fs: Add support for MOV_INDIRECT on pre-Broadwell hardware
While we're at it, we also add support for the possibility that the
indirect is, in fact, a constant. This shouldn't happen in the common case
(if it does, that means NIR failed to constant-fold something), but it's
possible so we should handle it.
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Tue, 24 Nov 2015 19:24:57 +0000 (11:24 -0800)]
i965/fs: Fix regs_read() for MOV_INDIRECT with a non-zero subnr
The subnr field is in bytes so we don't need to multiply by type_sz.
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Tue, 24 Nov 2015 17:03:29 +0000 (09:03 -0800)]
i965/fs: Don't force MASK_DISABLE on INDIRECT_MOV instructions
It should work fine without it and the visitor can set it if it wants.
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Tue, 24 Nov 2015 02:32:38 +0000 (18:32 -0800)]
i965/fs: Add support for doing MOV_INDIRECT on uniforms
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Wed, 30 Mar 2016 17:34:58 +0000 (10:34 -0700)]
anv: Install the installable ICD
Jason Ekstrand [Wed, 30 Mar 2016 17:30:42 +0000 (10:30 -0700)]
anv/intel_icd: Don't provide an absolute path
The driver will be installed to $(libdir)/libvulkan_intel.so and just
providing a driver name is enough for the loader. This also ensures that
multi-arch systems work ok.
Jason Ekstrand [Tue, 29 Mar 2016 19:00:24 +0000 (12:00 -0700)]
configure: Add initial support for enabling Vulkan drivers
Jason Ekstrand [Thu, 14 Apr 2016 22:12:41 +0000 (15:12 -0700)]
anv/pipeline: Use the right mask for lower_indirect_derefs
Ben Widawsky [Mon, 11 Apr 2016 16:49:41 +0000 (09:49 -0700)]
i965: Make intel_get_param return an int
This will fix the spurious error message: "Failed to query GPU properties."
that was unintentionally added in
cc01b63d730.
This patch changes the function to return an int so that the caller is able to
do stuff based on the return value.
The equivalent of this patch was in the original series that fixed up the
warning, but I dropped it at the last moment. It is required to make the desired
behavior of not warning when trying to query GPU properties from the kernel
unless there is something the user can do about it.
v2: Use strerror (Jason)
Make EINVAL check similar in all places (Ian)
NOTE: Broadwell appears to actually have some issue where the kernel returns
ENODEV when it shouldn't be. I will investigate this separately.
Reported-by: Chris Forbes <chrisf@ijw.co.nz>
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Brian Paul [Thu, 14 Apr 2016 21:07:53 +0000 (15:07 -0600)]
st/mesa: fix sampler view leak in st_DrawAtlasBitmaps()
I neglected to free the sampler view which was created earlier in the
function. So for each glCallLists() command that used the bitmap atlas
to draw text, we'd leak a sampler view object.
Also, check for st_create_texture_sampler_view() failure and record
GL_OUT_OF_MEMORY.
Cc: "11.1 11.2" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Nicolai Hähnle [Wed, 13 Apr 2016 15:55:29 +0000 (10:55 -0500)]
gallium/radeon: handle failure when mapping staging buffer
Cc: "11.1 11.2" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Nicolai Hähnle [Wed, 13 Apr 2016 17:17:28 +0000 (12:17 -0500)]
radeonsi: mark ssbo and images descriptor pointers dirty at beginning of CS
Without this, we were getting non-deterministic VM faults under high pressure.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Jason Ekstrand [Fri, 8 Apr 2016 17:49:02 +0000 (10:49 -0700)]
i965/vec4: Use UD rather than D for uniform indirects
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Wed, 25 Nov 2015 17:59:03 +0000 (09:59 -0800)]
i965/fs: Use UD type for offsets in VARYING_PULL_CONSTANT_LOAD
Reveiewed-by: Kristian Høgsberg <krh@bitplanet.net>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Samuel Pitoiset [Thu, 14 Apr 2016 19:47:15 +0000 (21:47 +0200)]
nvc0: do not break the universe on GK110+
I removed that return 0 by mistake. Ooops.
Fixes: 6e23fd4 ("nvc0: allow to use compute support on GM200")
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Samuel Pitoiset [Wed, 13 Apr 2016 22:46:50 +0000 (00:46 +0200)]
nvc0: allow to use compute support on GM200
This works like a charm but please not that NVF0_COMPUTE have to be set
because compute support is still not enabled by default on GK110+. This
will require more testing to make sure it won't break the 3D state.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Jason Ekstrand [Thu, 14 Apr 2016 17:39:54 +0000 (10:39 -0700)]
i965: remove pointless diff with the master branch
Jason Ekstrand [Thu, 14 Apr 2016 17:35:37 +0000 (10:35 -0700)]
nir/opt_algebraic: Remove the encoding line
This is an unneeded diff between the vulkan and master branches
Jason Ekstrand [Thu, 14 Apr 2016 17:28:45 +0000 (10:28 -0700)]
spirv: Move to compiler/
While it does rely on NIR, it's not really part of the NIR core. At the
moment, it still builds as part of libnir but that can be changed later if
desired.
Jason Ekstrand [Thu, 14 Apr 2016 17:24:33 +0000 (10:24 -0700)]
nir: Remove some pointless delta between vulkan and master
Jose Fonseca [Wed, 13 Apr 2016 17:19:52 +0000 (18:19 +0100)]
scons: Build NIR.
Emil Velikov:
- Attribute the src/{glsl,compiler}/nir move
- Flesh out to separate SConscript
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
Jose Fonseca [Wed, 13 Apr 2016 17:19:51 +0000 (18:19 +0100)]
nir: Use _snprintf on Windows.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
Jose Fonseca [Wed, 13 Apr 2016 17:19:50 +0000 (18:19 +0100)]
nir: Avoid structure initalization expressions.
Not supported by MSVC, and completely unnecessary -- inline functions
work just as well.
NIR_SRC_INIT/NIR_DEST_INIT could and probably should be replaced by the
inline functions.
Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
Jose Fonseca [Wed, 13 Apr 2016 17:19:49 +0000 (18:19 +0100)]
nir: Remove unistd.h include.
It doesn't seem needed, and is not available on MSVC.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
Jose Fonseca [Wed, 13 Apr 2016 17:19:48 +0000 (18:19 +0100)]
nir: Avoid empty {} struct initializer.
Not supported by MSVC and consistent through NIR.
[Emil Velikov: rebase]
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
Emil Velikov [Wed, 13 Apr 2016 14:38:48 +0000 (15:38 +0100)]
gallium/swr: fold the almost identical Makefiles
Rather than having two almost identical Makefiles, with various VPATH
hacks just fold them, using COMMON_* variables and actually getting
things buildable/shipable.
v2: whitespace fixes, remove Makefile.sources-arch
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Tim Rowley <timothy.o.rowley@intel.com>
Tim Rowley [Wed, 13 Apr 2016 20:06:45 +0000 (15:06 -0500)]
install-gallium-links.mk: handle multiple libraries
Need to prevent bash from interpreting whitespace between libraries
as a command line.
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
Marek Olšák [Wed, 13 Apr 2016 15:46:23 +0000 (17:46 +0200)]
radeonsi: don't overwrite the scratch offset in shader prologs
Prologs only look at num_input_sgprs.
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Marek Olšák [Wed, 13 Apr 2016 12:15:16 +0000 (14:15 +0200)]
radeonsi: fold num_user_sgprs where it is possible
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Marek Olšák [Wed, 13 Apr 2016 11:50:04 +0000 (13:50 +0200)]
radeonsi: fix SGPRS calculation once more
This fixes GS piglit failures after adding SI_PARAM_SHADER_BUFFERS,
which bumped NUM_USER_SGPRS and uncovered this bug on SI.
If this was fixed in LLVM, these workarounds wouldn't be needed.
LLVM would have to look at the calling convention to know how many SGPR
inputs are declared, and add VCC and the scratch wave offset (which is
enabled even if we spill SGPRs but not VGPRs, oh well).
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Marek Olšák [Wed, 13 Apr 2016 14:31:35 +0000 (16:31 +0200)]
radeonsi: disable hw ETC2 on Polaris
not supported by hw directly, but it's still fully supported by the driver
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Emil Velikov [Thu, 14 Apr 2016 08:47:50 +0000 (09:47 +0100)]
doxygen: remove git rebase fallouts
Should never have been (git) added in the first place.
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Jose Fonseca [Wed, 13 Apr 2016 12:58:57 +0000 (13:58 +0100)]
appveyor: Run unit tests.
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Jose Fonseca [Wed, 13 Apr 2016 12:31:04 +0000 (13:31 +0100)]
scons: Add a "check" target to run all unit tests.
Except:
- u_cache_test -- too long
- translate_test -- unreliable (it's probably testing corner cases that
translate module doesn't care about.)
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Jose Fonseca [Wed, 13 Apr 2016 12:29:12 +0000 (13:29 +0100)]
test/unit: Make translate_test invoke translate_create by default.
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Jose Fonseca [Wed, 13 Apr 2016 12:28:36 +0000 (13:28 +0100)]
test/unit: Make pipe_barrier_test actually check correct bahavior.
So it can run unattended.
Also make it silent by default.
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Jason Ekstrand [Thu, 14 Apr 2016 03:25:39 +0000 (20:25 -0700)]
Merge remote-tracking branch 'public/master' into vulkan
Michel Dänzer [Wed, 13 Apr 2016 06:57:50 +0000 (15:57 +0900)]
clover: Fix build against LLVM SVN >= r266163
createInternalizePass now takes a callback instead of a StringSet.
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Nanley Chery [Thu, 31 Mar 2016 00:13:01 +0000 (17:13 -0700)]
anv: Remove default scissor and viewport concepts
Users should never provide a scissor or viewport count of 0 because
they are required to set such state in a graphics pipeline. This
behavior was previously only used in Meta, which actually just
disables those hardware operations at pipeline creation time.
Kristian noticed that the current assignment of viewport count
reduces the number of viewport uploads, so it is not removed.
Signed-off-by: Nanley Chery <nanley.g.chery@intel.com>
Reviewed-by: Kristian Høgsberg Kristensen <kristian.h.kristensen@intel.com>
Nanley Chery [Wed, 13 Apr 2016 18:59:54 +0000 (11:59 -0700)]
anv: Replace ::disable_scissor with ::use_rectlists
Meta currently uses screenspace RECTLIST primitives that lie within
the framebuffer rectangle. Since this behavior shouldn't change in the
future, disable the scissor operation whenever rectlists are used.
Signed-off-by: Nanley Chery <nanley.g.chery@intel.com>
Reviewed-by: Kristian Høgsberg Kristensen <kristian.h.kristensen@intel.com>
Nanley Chery [Mon, 11 Apr 2016 16:57:21 +0000 (09:57 -0700)]
anv: Delete anv_graphics_pipeline_create_info::disable_viewport
There are no users of this field.
Signed-off-by: Nanley Chery <nanley.g.chery@intel.com>
Reviewed-by: Kristian Høgsberg Kristensen <kristian.h.kristensen@intel.com>
Nanley Chery [Fri, 1 Apr 2016 06:16:12 +0000 (23:16 -0700)]
gen{7,8}_pipeline: Always set ViewportXYClipTestEnable
For the following reasons, there is no behavioural change with this
commit: the ViewportXYClipTest function of the CLIP stage will continue
to be enabled outside of Meta (where disable_viewport is always false),
and the CLIP stage is turned off within Meta, so this function will
continue to be disabled in that case.
Signed-off-by: Nanley Chery <nanley.g.chery@intel.com>
Reviewed-by: Kristian Høgsberg Kristensen <kristian.h.kristensen@intel.com>
Nanley Chery [Thu, 31 Mar 2016 17:04:46 +0000 (10:04 -0700)]
gen{7,8}_pipeline: Apply 3DPRIM_RECTLIST restrictions
According to 3D Primitives Overview in the Bspec, when the RECTLIST
primitive is in use, the CLIP stage should be disabled or set to have
a different Clip Mode, and Viewport Mapping must be disabled:
Clipping: Must not require clipping or rely on the CLIP unit’s
ClipTest logic to determine if clipping is required. Either the CLIP
unit should be DISABLED, or the CLIP unit’s Clip Mode should be set
to a value other than CLIPMODE_NORMAL.
Viewport Mapping must be DISABLED (as is typical with the use of
screen-space coordinates).
We swap out ::disable_viewport for ::use_rectlist, because we currently
always use the RECTLIST primitive when we disable viewport mapping, and
we'll likely continue to use this primitive.
Signed-off-by: Nanley Chery <nanley.g.chery@intel.com>
Reviewed-by: Kristian Høgsberg Kristensen <kristian.h.kristensen@intel.com>
Nanley Chery [Thu, 7 Apr 2016 09:47:28 +0000 (02:47 -0700)]
anv_cmd_buffer: Don't make the initial state dirty
Avoid excessive state emission. Relevant state for an action command
will get set by the user:
From Chapter 5. Command Buffers,
When a command buffer begins recording, all state in that command
buffer is undefined.
[...]
Whenever the state of a command buffer is undefined, the application
must set all relevant state on the command buffer before any state
dependent commands such as draws and dispatches are recorded, otherwise
the behavior of executing that command buffer is undefined.
Signed-off-by: Nanley Chery <nanley.g.chery@intel.com>
Reviewed-by: Kristian Høgsberg Kristensen <kristian.h.kristensen@intel.com>
Nanley Chery [Tue, 29 Mar 2016 20:31:30 +0000 (13:31 -0700)]
anv/meta: Don't set the dynamic state for disabled operations
CmdSet* functions dirty the CommandBuffer's dynamic state. This causes
the new state to be emitted when CmdDraw is called. Since we don't need
the state that would be emitted, don't call the CmdSet* functions.
Signed-off-by: Nanley Chery <nanley.g.chery@intel.com>
Reviewed-by: Kristian Høgsberg Kristensen <kristian.h.kristensen@intel.com>
Nanley Chery [Wed, 6 Apr 2016 22:57:32 +0000 (15:57 -0700)]
anv/clear: Disable the scissor operation
Since the scissor rectangle always matches that of the framebuffer,
this operation isn't needed.
Signed-off-by: Nanley Chery <nanley.g.chery@intel.com>
Reviewed-by: Kristian Høgsberg Kristensen <kristian.h.kristensen@intel.com>
Jason Ekstrand [Fri, 25 Mar 2016 21:26:11 +0000 (14:26 -0700)]
nir/dead_variables: Configurably work with any variable mode
The old version of the pass only worked on globals and locals and always
left inputs, outputs, uniforms, etc. alone.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Kenneth Graunke [Tue, 12 Apr 2016 09:30:25 +0000 (02:30 -0700)]
i965: Switch to NIR for ldexp lowering.
The old GLSL IR based lowering doesn't quite work right in all cases,
and fails several dEQP-GLES31 and Vulkan CTS tests. Jason's new
approach in NIR passes all the tests. There's not likely to be a ton
of advantage to lowering early in GLSL IR anyway, so...switch.
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Jason Ekstrand [Fri, 25 Mar 2016 19:12:38 +0000 (12:12 -0700)]
nir/algebraic: Add lowering for ldexp
The algorithm used is different from both the naive suggestion from the
GLSL spec and the one used in GLSL IR today. Unfortunately, the GLSL IR
implementation that we have today doesn't handle denormals (for those that
care) or the case where the float source is +-inf.
Reviewed-by: Matt Turner <mattst88@gmail.com>
Jason Ekstrand [Fri, 25 Mar 2016 18:17:53 +0000 (11:17 -0700)]
i965: Implement the new imod and irem opcodes
Reviewed-by: Matt Turner <mattst88@gmail.com>
Jason Ekstrand [Fri, 25 Mar 2016 18:13:40 +0000 (11:13 -0700)]
nir: Add more modulus opcodes
These are all needed for SPIR-V
Reviewed-by: Rob Clark <robdclark@gmail.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Jason Ekstrand [Wed, 25 Nov 2015 17:12:37 +0000 (09:12 -0800)]
i965/vec4: Inline get_pull_constant_offset
It's not really doing enough anymore to justify a helper function.
Reviewed-by: Eduardo Lima Mitev <elima@igalia.com>
Reveiewed-by: Kristian Høgsberg <krh@bitplanet.net>
Jason Ekstrand [Mon, 11 Apr 2016 20:43:27 +0000 (13:43 -0700)]
nir/lower_io: Allow for a full bitmask of modes
Acked-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Rob Clark <robdclark@gmail.com>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Jason Ekstrand [Mon, 11 Apr 2016 20:38:02 +0000 (13:38 -0700)]
nir/lower_indirect: nir_variable_mode is now a bitfield
Acked-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Rob Clark <robdclark@gmail.com>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Jason Ekstrand [Mon, 11 Apr 2016 20:32:59 +0000 (13:32 -0700)]
nir: Convert nir_variable_mode to a bitfield
There are several passes where we need to specify some set of variable
modes that the pass needs top operate on. This lets us easily do that.
Acked-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Rob Clark <robdclark@gmail.com>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
George Kyriazis [Wed, 13 Apr 2016 15:48:55 +0000 (10:48 -0500)]
gallium/swr: Make flat shading tris work.
- Incorporate flatshade flag into the shader generation
- Use provoking vertex (vc) in shader when flat shading.
Reviewed-by: Bruce Cherniak <bruce.cherniak@intel.com>
Rob Clark [Tue, 12 Apr 2016 20:06:17 +0000 (16:06 -0400)]
Revert "freedreno/a4xx: better occlusion/sample counting"
This reverts commit
62fa868728c729152af0d7cecd1d3e47e831cb7d.
dEQP-GLES3.functional.occlusion_query.* was unhappy about that change.
Still not really sure *what* the other slots in the sample results
buffer are.
Signed-off-by: Rob Clark <robclark@freedesktop.org>
Rob Clark [Tue, 12 Apr 2016 18:51:33 +0000 (14:51 -0400)]
freedreno/a4xx: rasterizer_discard support
This one is slightly annoying, since trying to write RBRC from draw
would clobber values set in the tiling/gmem code. We could do command-
stream patching for RBRC, as is done on a3xx. Although since it seems
to be a rarely used feature, it is easier just to do RMW to set/clear
the bit.
Fixes dEQP-GLES3.functional.rasterizer_discard.basic.write_depth_triangles
and related tests.
a3xx still needs the same feature, although there it probably makes more
sense to take advantage of the existing cmdstream patching which is
required for RBRC for other reasons.
Signed-off-by: Rob Clark <robclark@freedesktop.org>
Rob Clark [Tue, 12 Apr 2016 17:56:41 +0000 (13:56 -0400)]
freedreno/ir3: fix array textures on a4xx
Seems like a4xx needs offset added to array index for all arrays,
whereas a3xx only for cubemap arrays. Fixes a whole swath of dEQP fails
(roughly *sampler2darray*).
Signed-off-by: Rob Clark <robclark@freedesktop.org>