mesa.git
7 years agointel/genxml: Update genx_bits for gen10+
Anuj Phogat [Fri, 14 Apr 2017 22:09:42 +0000 (15:09 -0700)]
intel/genxml: Update genx_bits for gen10+

This commit adds a gen10 case to the switch statement and
drops some unneeded code for handling gen numbers which
doesn't work on gen10 and above.

V2: Drop "z = float(z)" and the "z *= 10" lines

Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
7 years agoi965/cnl: Add gen10 specific function declarations
Anuj Phogat [Thu, 11 May 2017 21:58:13 +0000 (14:58 -0700)]
i965/cnl: Add gen10 specific function declarations

These declarations will help the code start compiling
once we wire up the makefiles for gen10. Later patches
will start using these functions for gen10.

Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
7 years agoi965/cnl: Include gen10_pack.h
Anuj Phogat [Thu, 11 May 2017 21:35:10 +0000 (14:35 -0700)]
i965/cnl: Include gen10_pack.h

Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Nanley Chery <nanley.g.chery@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
7 years agoi965/cnl: Define genX(x) and GENX(x) for gen10
Anuj Phogat [Mon, 17 Apr 2017 20:48:55 +0000 (13:48 -0700)]
i965/cnl: Define genX(x) and GENX(x) for gen10

Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Nanley Chery <nanley.g.chery@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
7 years agoi965/genxml: Add gen10.xml
Jason Ekstrand [Thu, 11 May 2017 21:16:44 +0000 (14:16 -0700)]
i965/genxml: Add gen10.xml

V2(Anuj):
Add default value for length of 3DPRIMITIVE command
Add values for 'Attribute Active Component Format'
Rename few fields to match gen9.xml

V3 (Ander Conselvan de Oliveira)
Add gen10 alias for MOCS
Make 3DSTATE_CONSTANT_BODY on Gen10 use arrays

Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Tested-by: Anuj Phogat <anuj.phogat@gmail.com>
7 years agoi965: Make feature macros gen8 based
Ben Widawsky [Tue, 29 Dec 2015 17:11:02 +0000 (09:11 -0800)]
i965: Make feature macros gen8 based

All the "features" of the hardware are similar starting with GEN8, so remove as
much of the GEN9 uniqueness as possible. This makes implementing future gen
platforms a bit easier.

Signed-off-by: Ben Widawsky <benjamin.widawsky@intel.com>
Reviewed-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
7 years agoradv: set fmask state to all 0s when no fmask. (v2)
Dave Airlie [Fri, 9 Jun 2017 01:11:29 +0000 (02:11 +0100)]
radv: set fmask state to all 0s when no fmask. (v2)

The shader reads the descriptor to decide if it should take the
fmask value, however we weren't initing it always, which meant
random crap, esp with MSAA depth textures.

Fixes random hangs with:
dEQP-VK.glsl.builtin_var.fragdepth.*

v2: check fmask_state is not NULL

Fixes: f4e499ec791 "radv: add initial non-conformant radv vulkan driver"
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Signed-off-by: Dave Airlie <airlied@redhat.com>
7 years agoi965: Temporarily disable async mappings on non-LLC
Matt Turner [Thu, 8 Jun 2017 22:13:42 +0000 (15:13 -0700)]
i965: Temporarily disable async mappings on non-LLC

Fixes regressions from commits e0a9b261e593 and a16355d67d92 by
neutering async mappings on non-LLC to be synchronous, like they were
before those two commits. :(

The failing tests include

piglit-test piglit.spec.nv_primitive_restart.primitive-restart-vbo_index_only
piglit-test piglit.spec.nv_primitive_restart.primitive-restart-vbo_combined_vertex_and_index
piglit-test piglit.spec.nv_primitive_restart.primitive-restart-vbo_separate_vertex_and_index
piglit-test piglit.spec.nv_primitive_restart.primitive-restart-vbo_vertex_only
piglit-test piglit.spec.arb_pixel_buffer_object.texsubimage-unpack pbo

7 years agomesa/main/debug: Check if we successfully reopened the ppm file.
Rafael Antognolli [Mon, 5 Jun 2017 18:04:26 +0000 (11:04 -0700)]
mesa/main/debug: Check if we successfully reopened the ppm file.

Since we created the file, we should be able to reopen it for appending, but
some weird filesystem error could cause that to be false. So simply check
whether we could reopen it or not.

CID: 1177144
Signed-off-by: Rafael Antognolli <rafael.antognolli@intel.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
7 years agotgsi: clarify TGSI_SEMANTIC_SAMPLEMASK documentation
Brian Paul [Fri, 9 Jun 2017 13:45:02 +0000 (07:45 -0600)]
tgsi: clarify TGSI_SEMANTIC_SAMPLEMASK documentation

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
7 years agogallium/wgl: Allow context creation even if SetPixelFormat() wasn't called
Frank Richter [Wed, 7 Jun 2017 07:43:43 +0000 (09:43 +0200)]
gallium/wgl: Allow context creation even if SetPixelFormat() wasn't called

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101326
Reviewed-by: Brian Paul <brianp@vmware.com>
7 years agost/dri: support format modifier queries
Varad Gautam [Tue, 30 May 2017 11:53:47 +0000 (17:23 +0530)]
st/dri: support format modifier queries

ask the driver for supported modifiers for a given format.

v2: move to __DRIimageExtension v16.
v3: fail if the supplied format is not supported by driver.
v4: purge PIPE_CAP_QUERY_DMABUF_ATTRIBS.
v5:
- move to __DRIimageExtension v15, pass external_only to the driver.

Signed-off-by: Varad Gautam <varad.gautam@collabora.com>
Reviewed-by: Lucas Stach <l.stach@pengutronix.de> (v4)
Cc: Lucas Stach <l.stach@pengutronix.de>
7 years agogallium: introduce format modifier querying
Varad Gautam [Tue, 30 May 2017 11:53:46 +0000 (17:23 +0530)]
gallium: introduce format modifier querying

format modifiers tokens are driver specific, and hence, need to come
in from the driver. this allows drivers to be queried for supported
format modifiers for EGL_EXT_image_dma_buf_import_modifiers.

v2: rebase to master.
v3: drivers must return false on query failure.
v4: use pscreen->is_format_supported instead of adding a separate
    format query handle, remove PIPE_CAP_QUERY_DMABUF_ATTRIBS.
    (Lucas Stach)
v5: add external_only parameter.

Signed-off-by: Varad Gautam <varad.gautam@collabora.com>
Cc: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Lucas Stach <l.stach@pengutronix.de>
7 years agost/dri: support format queries
Varad Gautam [Tue, 30 May 2017 11:53:45 +0000 (17:23 +0530)]
st/dri: support format queries

ask the driver for supported dmabuf formats

v2: rebase to master.
v3: return false on failure.
v4: use pscreen->is_format_supported instead of adding a new query.
    (Lucas Stach)
v5: stylefix to conform to formatting rules (Brian Paul). add fourcc list
    here instead of using struct image_format from v4.

Signed-off-by: Varad Gautam <varad.gautam@collabora.com>
Reviewed-by: Lucas Stach <l.stach@pengutronix.de> (v4)
Cc: Lucas Stach <l.stach@pengutronix.de>
7 years agost/dri: implement DRIimage creation from dmabufs with modifiers
Varad Gautam [Tue, 30 May 2017 11:53:44 +0000 (17:23 +0530)]
st/dri: implement DRIimage creation from dmabufs with modifiers

support importing dmabufs into DRIimage while taking format modifiers
in account, as per DRIimage extension version 15.

v2: initialize winsys modifier to DRM_FORMAT_MOD_INVALID (Daniel Stone)
v3: do not bump DRIimageExtension version. split out winsys changes.

Signed-off-by: Varad Gautam <varad.gautam@collabora.com>
Reviewed-by: Lucas Stach <l.stach@pengutronix.de>
7 years agost/dri: implement createImageWithModifiers in DRIimage
Varad Gautam [Tue, 30 May 2017 11:53:43 +0000 (17:23 +0530)]
st/dri: implement createImageWithModifiers in DRIimage

adds a pscreen->resource_create_with_modifiers() to create textures
with modifier.

v2:
- stylefixes (Emil Velikov)
- don't return selected modifier from resource_create_with_modifiers. we can
  use the winsys_handle to get this.

Signed-off-by: Varad Gautam <varad.gautam@collabora.com>
Reviewed-by: Lucas Stach <l.stach@pengutronix.de> (v1)
Cc: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Lucas Stach <l.stach@pengutronix.de>
7 years agost/dri: enable DRIimage modifier queries
Varad Gautam [Tue, 30 May 2017 11:53:42 +0000 (17:23 +0530)]
st/dri: enable DRIimage modifier queries

return the modifier selected by the driver when creating this image.

v2: since we can use winsys_handle->modifier to serve these, remove
    DRIimage->modifier from v1.
    use DRM_API_HANDLE_TYPE_KMS instead of DRM_API_HANDLE_TYPE_FD to avoid
    ownership transfer. (Lucas)

Suggested-by: Daniel Stone <daniels@collabora.com>
Signed-off-by: Varad Gautam <varad.gautam@collabora.com>
Cc: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Lucas Stach <l.stach@pengutronix.de>
7 years agogallium/winsys/drm: introduce modifier field to winsys_handle
Varad Gautam [Tue, 30 May 2017 11:53:41 +0000 (17:23 +0530)]
gallium/winsys/drm: introduce modifier field to winsys_handle

we use this to import resources with format modifiers, and to support
per-resource modifier queries.

Signed-off-by: Varad Gautam <varad.gautam@collabora.com>
Cc: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Lucas Stach <l.stach@pengutronix.de>
7 years agomesa: make use of NewScissorTest driver flags
Samuel Pitoiset [Tue, 6 Jun 2017 13:23:58 +0000 (15:23 +0200)]
mesa: make use of NewScissorTest driver flags

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
7 years agomesa: make use of NewScissorRect driver flags
Samuel Pitoiset [Tue, 6 Jun 2017 13:21:43 +0000 (15:21 +0200)]
mesa: make use of NewScissorRect driver flags

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
7 years agomesa: add gl_driver_flags::NewScissor{Rect,Test}
Samuel Pitoiset [Tue, 6 Jun 2017 13:03:50 +0000 (15:03 +0200)]
mesa: add gl_driver_flags::NewScissor{Rect,Test}

_NEW_SCISSOR mesa flag is set when a scissor test is enabled/disabled
or when a new rectangle is defined. However, it triggers too much
changes in the state tracker.

Actually, ST_NEW_RASTERIZER should only be called when a scissor
test is enabled/disabled, while ST_NEW_SCISSOR should be called
in both situations.

In other words, this will avoid to update the rasterizer every
time a new rectangle is defined using glScissor*().

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
7 years agoegl: fix _eglQuerySurface in EGL_BUFFER_AGE_EXT case
Tapani Pälli [Thu, 8 Jun 2017 09:24:24 +0000 (12:24 +0300)]
egl: fix _eglQuerySurface in EGL_BUFFER_AGE_EXT case

Specification states that in case of error, value should not be
written, patch changes buffer age queries to return -1 in case of
error so that we can skip changing the value.

In addition, small change to droid_query_buffer_age to return 0
in case buffer does not have a back buffer available.

Fixes:
   dEQP-EGL.functional.negative_partial_update.not_postable_surface

Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Eric Engestrom <eric.engestrom@imgtec.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Chad Versace <chadversary@chromium.org>
Cc: mesa-stable@lists.freedesktop.org
7 years agoradv: introduce perf test env var and allow to enable chaining
Dave Airlie [Tue, 9 May 2017 03:17:30 +0000 (04:17 +0100)]
radv: introduce perf test env var and allow to enable chaining

We have some features that seem to slow things down or cause other
possible undesireable side effects, but it would be nice to test
games etc with them easily.

I forsee multisample DCC and maybe some shader opt changes using this.

For now use it for batch chaining.

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Signed-off-by: Dave Airlie <airlied@redhat.com>
7 years agomesa: add KHR_no_error support to glDrawRangeElements*()
Timothy Arceri [Thu, 8 Jun 2017 06:15:58 +0000 (16:15 +1000)]
mesa: add KHR_no_error support to glDrawRangeElements*()

Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
7 years agomesa: rework _ae_invalidate_state() so that it just sets a dirty flag
Timothy Arceri [Wed, 7 Jun 2017 04:02:03 +0000 (14:02 +1000)]
mesa: rework _ae_invalidate_state() so that it just sets a dirty flag

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
7 years agomesa: remove redundant _ae_invalidate_state() call
Timothy Arceri [Wed, 7 Jun 2017 03:58:34 +0000 (13:58 +1000)]
mesa: remove redundant _ae_invalidate_state() call

The FLUSH_VERTICES(ctx, _NEW_ARRAY) above this will already cause
this to be called.

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
7 years agomesa: inline vbo_exec_invalidate_state() and call from mesa core
Timothy Arceri [Wed, 7 Jun 2017 03:26:16 +0000 (13:26 +1000)]
mesa: inline vbo_exec_invalidate_state() and call from mesa core

Rather than calling it indirectly in each driver.

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
7 years agomesa: rework vbo_exec_init()
Timothy Arceri [Wed, 7 Jun 2017 02:43:27 +0000 (12:43 +1000)]
mesa: rework vbo_exec_init()

Here we make some assumptions about the AEcontext and set the
recalculate bools directly.

Some formating fixes are also made while we are here.

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
7 years agomesa: stop passing state bitfield to UpdateState()
Timothy Arceri [Wed, 7 Jun 2017 02:19:13 +0000 (12:19 +1000)]
mesa: stop passing state bitfield to UpdateState()

The code comment which seems to have been added in cab974cf6c2db
(from year 2000) says:

   "Set ctx->NewState to zero to avoid recursion if
   Driver.UpdateState() has to call FLUSH_VERTICES().  (fixed?)"

As far as I can tell nothing in any of the UpdateState() calls
should cause it to be called recursively.

V2: add a wrapper around the osmesa update function so it can still
    be used internally.

Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
7 years agost/mesa: add st_invalidate_buffers() helper
Timothy Arceri [Wed, 7 Jun 2017 02:14:54 +0000 (12:14 +1000)]
st/mesa: add st_invalidate_buffers() helper

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
7 years agor200/radeon: stop calling _ae_invalidate_state() directly
Timothy Arceri [Wed, 7 Jun 2017 01:28:23 +0000 (11:28 +1000)]
r200/radeon: stop calling _ae_invalidate_state() directly

It is already called via _vbo_InvalidateState().

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Tested-by: Ian Romanick <ian.d.romanick@intel.com>
7 years agoswr: relax c++ requirement from c++14 to c++11
Tim Rowley [Thu, 8 Jun 2017 15:38:52 +0000 (10:38 -0500)]
swr: relax c++ requirement from c++14 to c++11

Remove c++14 generic lambda to keep compiler requirement at c++11.

No regressions on piglit or vtk test suites.

Tested-by: Chuck Atkins <chuck.atkins@kitware.com>
Reviewed-by: Bruce Cherniak <bruce.cherniak@intel.com>
CC: mesa-stable@lists.freedesktop.org
7 years agoradeonsi: call LLVMAddEarlyCSEMemSSAPass only for LLVM >= 4.0
Juan A. Suarez Romero [Thu, 8 Jun 2017 19:22:08 +0000 (21:22 +0200)]
radeonsi: call LLVMAddEarlyCSEMemSSAPass only for LLVM >= 4.0

LLVMAddEarlyCSEMemSSAPass() is defined in LLVM 4.0.

Fixes: 257b538 ("radeonsi: do EarlyCSEMemSSA LLVM pass)
Signed-off-by: Marek Olšák <marek.olsak@amd.com>
7 years agogallium/radeon: don't allocate HTILE in a separate buffer
Marek Olšák [Tue, 6 Jun 2017 21:54:23 +0000 (23:54 +0200)]
gallium/radeon: don't allocate HTILE in a separate buffer

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
7 years agoradeonsi: rename depth decompress functions
Marek Olšák [Tue, 6 Jun 2017 22:58:46 +0000 (00:58 +0200)]
radeonsi: rename depth decompress functions

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
7 years agoradeonsi: rename shader resource decompress masks to their true meaning
Marek Olšák [Tue, 6 Jun 2017 22:51:23 +0000 (00:51 +0200)]
radeonsi: rename shader resource decompress masks to their true meaning

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
7 years agoradeonsi: rename is_compressed_colortex -> color_needs_decompression
Marek Olšák [Tue, 6 Jun 2017 22:27:14 +0000 (00:27 +0200)]
radeonsi: rename is_compressed_colortex -> color_needs_decompression

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
7 years agoradeonsi: disable the patch ID workaround on SI when the patch ID isn't used (v2)
Marek Olšák [Tue, 6 Jun 2017 13:23:42 +0000 (15:23 +0200)]
radeonsi: disable the patch ID workaround on SI when the patch ID isn't used (v2)

The workaround causes a massive performance decrease on 1-SE parts.
(Cape Verde, Hainan, Oland)

The performance regression is already part of 17.0 and 17.1.

v2: check tess_uses_prim_id

Cc: 17.0 17.1 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
7 years agoradeonsi: don't update dependent states if it has no effect (v2)
Marek Olšák [Mon, 5 Jun 2017 00:00:52 +0000 (02:00 +0200)]
radeonsi: don't update dependent states if it has no effect (v2)

This and the previous clip_regs commit decrease IB sizes and the number of
si_update_shaders invocations as follows:

                 IB size   si_update_shaders calls
Borderlands 2      -10%            -27%
Deus Ex: MD         -5%            -11%
Talos Principle     -8%            -30%

v2: always dirty cb_render_state in set_framebuffer_state

Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
7 years agoi965: Add format/modifier advertising
Varad Gautam [Tue, 30 May 2017 11:54:09 +0000 (17:24 +0530)]
i965: Add format/modifier advertising

v2: Rebase and reuse tiling/modifier map. (Daniel Stone)
v3: bump DRIimageExtension to version 15, fill external_only array.
v4: Y-tiling works since gen 6

Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
7 years agoi965: Support dmabuf import with modifiers
Varad Gautam [Tue, 30 May 2017 11:54:08 +0000 (17:24 +0530)]
i965: Support dmabuf import with modifiers

Add support for createImageFromDmaBufs2, adding a modifier to the
original, and allow importing CCS resources with auxiliary data from
dmabufs.

v2: avoid DRIimageExtension version bump, pass single modifier to
    createImageFromDmaBufs2.

Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
7 years agoi965: Improve same-buffer restriction for imports
Daniel Stone [Tue, 30 May 2017 11:53:48 +0000 (17:23 +0530)]
i965: Improve same-buffer restriction for imports

Intel hardware requires that all planes of an image come from the same
buffer, which is currently implemented by testing that all FDs are
numerically the same.

However, when going through a winsys (e.g.) or anything which transits
FDs individually, the FDs may be different even if the underlying buffer
is the same.

Instead of checking the FDs for equality, we must check if they actually
point to the same buffer (Jason).

Reviewed-by: Varad Gautam <varad.gautam@collabora.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
7 years agoi965: Allocate tile aligned height
Ben Widawsky [Tue, 30 May 2017 11:53:55 +0000 (17:23 +0530)]
i965: Allocate tile aligned height

This patch shouldn't actually do anything because the libdrm function
should already do this alignment. However, it preps us for a future
patch where we add in the CCS AUX size, and in the process it serves as
a good place to find bisectable issues if libdrm or kernel does
something incorrectly.

v2: Do proper alignment for X tiling, and make sure non-tiled case is
handled (Jason)
v3: Rebase (Daniel)

Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
7 years agoi965: Move fallback size assignment out of bufmgr
Daniel Stone [Tue, 6 Jun 2017 10:53:55 +0000 (11:53 +0100)]
i965: Move fallback size assignment out of bufmgr

The bufmgr took a mandatory size argument, which would only be used if
the kernel size query failed, i.e. an older kernel. It didn't actually
check that the BO size was sufficient for use.

Pull the check out of the bufmgr, and actually check that the BO is
sufficiently-sized for our import one level up. This also resolves a
chicken/egg we have when importing bufers without explicit modifiers,
namely that we need the tiling mode to calculate the size, but we need
the BO imported to query the tiling mode.

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
7 years agoi965: Invert image modifier/tiling inference
Daniel Stone [Tue, 30 May 2017 11:53:49 +0000 (17:23 +0530)]
i965: Invert image modifier/tiling inference

When allocating images, we record a tiling mode and then work backwards
to infer the modifier. Unfortunately this is the wrong way around, since
it is a one:many mapping (e.g. TILING_Y can be plain Y-tiling, or
Y-tiling with CCS).

Invert the mapping, so we record a modifier first and then map this to a
tiling mode.

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
7 years agoegl/dri2: Avoid sign extension when building modifier
Daniel Stone [Mon, 5 Jun 2017 13:30:02 +0000 (14:30 +0100)]
egl/dri2: Avoid sign extension when building modifier

Since the EGL attributes are signed integers, a straight OR would
also perform sign extension,

Fixes: 6f10e7c37a ("egl/dri2: Create EGLImages with dmabuf modifiers")
Reviewed-by: Eric Engestrom <eric.engestrom@imgtec.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
7 years agoi915g: Add blitter_context argument.
Vinson Lee [Thu, 8 Jun 2017 07:15:59 +0000 (00:15 -0700)]
i915g: Add blitter_context argument.

Fix build error.

  CC       i915_surface.lo
i915_surface.c:108:63: error: too few arguments to function call, expected 4, have 3
   util_blitter_default_src_texture(&src_templ, src, src_level);
   ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~                           ^
../../../../src/gallium/auxiliary/util/u_blitter.h:271:1: note: 'util_blitter_default_src_texture' declared here
void util_blitter_default_src_texture(struct blitter_context *blitter,
^

Fixes: a893c9169733 ("gallium/u_blitter: use 2D_ARRAY for cubemap blits if possible")
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101340
Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Reviewed-by: Juan A. Suarez Romero <jasuarez@igalia.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
7 years agoetnaviv: flush resource when binding as sampler view
Lucas Stach [Mon, 15 May 2017 15:06:41 +0000 (17:06 +0200)]
etnaviv: flush resource when binding as sampler view

As TS is also allowed on sampler resources, we need to make sure to resolve
to self when binding the resource as a texture, to avoid stale content
being sampled.

Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
7 years agoetnaviv: don't flush resource to self without TS
Lucas Stach [Sun, 4 Jun 2017 04:24:20 +0000 (06:24 +0200)]
etnaviv: don't flush resource to self without TS

A resolve to self is only necessary if the resource is fast cleared, so
there is never a need to do so if there is no TS allocated.

Signed-off-by: Lucas Stach <dev@lynxeye.de>
Reviewed-by: Wladimir J. van der Laan <laanwj@gmail.com>
7 years agoetnaviv: upgrade DISCARD_RANGE to DISCARD_WHOLE_RESOURCE if possible
Lucas Stach [Thu, 18 May 2017 15:05:02 +0000 (17:05 +0200)]
etnaviv: upgrade DISCARD_RANGE to DISCARD_WHOLE_RESOURCE if possible

Stolen from VC4. As we don't do any fancy reallocation tricks yet, it's
possible to upgrade also coherent mappings and shared resources.

Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Wladimir J. van der Laan <laanwj@gmail.com>
7 years agoetnaviv: simplify transfer tiling handling
Lucas Stach [Thu, 18 May 2017 14:44:18 +0000 (16:44 +0200)]
etnaviv: simplify transfer tiling handling

There is no need to special case compressed resources, as they are already
marked as linear on allocation. With that out of the way, there is room to
cut down on the number of if clauses used.

Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Wladimir J. van der Laan <laanwj@gmail.com>
7 years agoetnaviv: don't read back resource if transfer discards contents
Lucas Stach [Thu, 18 May 2017 14:30:02 +0000 (16:30 +0200)]
etnaviv: don't read back resource if transfer discards contents

Reduces bandwidth usage of transfers which discard the buffer contents,
as well as skipping unnecessary command stream flushes and CPU/GPU
synchronization.

Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Wladimir J. van der Laan <laanwj@gmail.com>
7 years agoetnaviv: honor PIPE_TRANSFER_UNSYNCHRONIZED flag
Lucas Stach [Thu, 18 May 2017 16:20:12 +0000 (18:20 +0200)]
etnaviv: honor PIPE_TRANSFER_UNSYNCHRONIZED flag

This gets rid of quite a bit of CPU/GPU sync on frequent vertex buffer
uploads and I haven't seen any of the issues mentioned in the comment,
so this one seems stale.

Ignore the flag if there exists a temporary resource, as those ones are
never busy.

Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Wladimir J. van der Laan <laanwj@gmail.com>
7 years agoetnaviv: slim down resource waiting
Lucas Stach [Thu, 18 May 2017 13:37:02 +0000 (15:37 +0200)]
etnaviv: slim down resource waiting

cpu_prep() already does all the required waiting, so the only thing that
needs to be done is flushing the commandstream, if a GPU write is pending.

Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Wladimir J. van der Laan <laanwj@gmail.com>
7 years agoglsl: Fix gl_shader_stage enum unsigned comparison
Rob Herring [Thu, 1 Jun 2017 00:56:56 +0000 (19:56 -0500)]
glsl: Fix gl_shader_stage enum unsigned comparison

Replace -1 with MESA_SHADER_NONE enum value to fix sign related warning:

external/mesa3d/src/compiler/glsl/link_varyings.cpp:1415:25: warning: comparison of constant -1 with expression of type 'gl_shader_stage' is always true [-Wtautological-constant-out-of-range-compare]
        (consumer_stage != -1 && consumer_stage != MESA_SHADER_FRAGMENT))) {
         ~~~~~~~~~~~~~~ ^  ~~

Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Signed-off-by: Rob Herring <robh@kernel.org>
7 years agoAndroid: vulkan: fix build error due to extra )
Rob Herring [Thu, 8 Jun 2017 01:56:07 +0000 (20:56 -0500)]
Android: vulkan: fix build error due to extra )

Commit 621b3410f5f8 ("util/vulkan: Move Vulkan utilities to
src/vulkan/util") broke the Android build with the following error:

build/core/binary.mk:1427: error: external/mesa3d/src/vulkan/Android.mk: libmesa_vulkan_util: Unused source files: util/vk_util.h).

Fixes: 621b3410f5f8 ("util/vulkan: Move Vulkan utilities to src/vulkan/util")
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Cc: Alex Smith <asmith@feralinteractive.com>
Signed-off-by: Rob Herring <robh@kernel.org>
7 years agoFix glcpp test expectations
Iago Toral Quiroga [Thu, 8 Jun 2017 06:38:29 +0000 (08:38 +0200)]
Fix glcpp test expectations

With commit f7741985be0234 we have changed some preprocessor
error messages and warnings. Adapt related glcpp tests
expectations accordingly.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101336
Tested-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
7 years agoutil: make set's deleted_key_value declaration consistent with hash table one
Vlad Golovkin [Mon, 29 May 2017 23:51:32 +0000 (02:51 +0300)]
util: make set's deleted_key_value declaration consistent with hash table one

This also silences following clang warnings:
no previous extern declaration for non-static variable 'deleted_key' [-Werror,-Wmissing-variable-declarations]
const void *deleted_key = &deleted_key_value;
            ^
no previous extern declaration for non-static variable 'deleted_key_value'
      [-Werror,-Wmissing-variable-declarations]
uint32_t deleted_key_value;
         ^

Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
7 years agoi965: Delete intel_resolve_map
Jason Ekstrand [Fri, 26 May 2017 19:18:49 +0000 (12:18 -0700)]
i965: Delete intel_resolve_map

Now that we've moved over to the new array mechanism, it's no longer
needed.

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Acked-by: Chad Versace <chadversary@chromium.org>
7 years agoi965: Use the new tracking mechanism for HiZ
Jason Ekstrand [Fri, 26 May 2017 19:12:06 +0000 (12:12 -0700)]
i965: Use the new tracking mechanism for HiZ

This is similar to the previous commit only for HiZ.  For HiZ, apart
from everything looking different, there is really only one functional
change:  We now track the ISL_AUX_STATE_COMPRESSED_NO_CLEAR state.
Previously, if you rendered to a resolved slice of the miptree and then
did a fast-clear with a different clear color, that slice would get
resolved even though it hadn't been fast-cleared.  Now that we can track
COMPRESSED_NO_CLEAR, we know that it doesn't have any blocks in the
"clear" state so we can skip the resolve.

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Acked-by: Chad Versace <chadversary@chromium.org>
7 years agoi965/miptree: Make level_has_hiz take a const miptree
Jason Ekstrand [Wed, 31 May 2017 18:50:24 +0000 (11:50 -0700)]
i965/miptree: Make level_has_hiz take a const miptree

Acked-by: Chad Versace <chadversary@chromium.org>
7 years agoi965: Wholesale replace the color resolve tracking code
Jason Ekstrand [Fri, 26 May 2017 00:18:30 +0000 (17:18 -0700)]
i965: Wholesale replace the color resolve tracking code

This commit reworks the resolve tracking for CCS and MCS to use the new
isl_aux_state enum.  This should provide much more accurate and easy to
reason about tracking.  In order to understand, for instance, the
intel_miptree_prepare_ccs_access function, one only has to go look at
the giant comment for the isl_aux_state enum and follow the arrows.
Unfortunately, there's no good way to split this up without making a
real mess so there are a bunch of changes in here:

 1) We now do partial resolves.  I really have no idea how this ever
    worked before.  So far as I can tell, the only time the old code
    ever did a partial resolve was when it was using CCS_D where a
    partial resolve and a full resolve are the same thing.

 2) We are now tracking 4 states instead of 3 for CCS_E.  In particular,
    we distinguish between compressed with clear and compressed without
    clear.  The end result is that you will never get two partial
    resolves in a row.

 3) The texture view rules are now more correct.  Previously, we would
    only bail if compression was not supported by the destination
    format.  However, this is not actually correct.  Not all format
    pairs are supported for texture views with CCS even if both support
    CCS individually.  Fortunately, ISL has a helper for this.

 4) We are no longer using intel_resolve_map for tracking aux state but
    are instead using a simple array of enum isl_aux_state indexed by
    level and layer.  This is because, now that we're tracking 4
    different states, it's no longer clear which should be the "default"
    and array lookups are faster than linked list searches.

 5) The new code is very assert-happy.  Incorrect transitions will now
    get caught by assertions rather than by rendering corruption.

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Acked-by: Chad Versace <chadversary@chromium.org>
7 years agoi965: Delete most of the old resolve interface
Jason Ekstrand [Thu, 25 May 2017 22:02:23 +0000 (15:02 -0700)]
i965: Delete most of the old resolve interface

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Acked-by: Chad Versace <chadversary@chromium.org>
7 years agoi965: Use the new get/set_aux_state functions for color clears
Jason Ekstrand [Thu, 25 May 2017 23:09:04 +0000 (16:09 -0700)]
i965: Use the new get/set_aux_state functions for color clears

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Acked-by: Chad Versace <chadversary@chromium.org>
7 years agoi965: Move blorp to the new resolve functions
Jason Ekstrand [Thu, 25 May 2017 21:50:26 +0000 (14:50 -0700)]
i965: Move blorp to the new resolve functions

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Acked-by: Chad Versace <chadversary@chromium.org>
7 years agoi965: Move depth to the new resolve functions
Jason Ekstrand [Thu, 25 May 2017 21:15:44 +0000 (14:15 -0700)]
i965: Move depth to the new resolve functions

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Chad Versace <chadversary@chromium.org>
7 years agoi965: Move images to the new resolve functions
Jason Ekstrand [Thu, 25 May 2017 19:30:50 +0000 (12:30 -0700)]
i965: Move images to the new resolve functions

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Chad Versace <chadversary@chromium.org>
7 years agoi965: Move framebuffer fetch to the new resolve functions
Jason Ekstrand [Thu, 25 May 2017 19:26:00 +0000 (12:26 -0700)]
i965: Move framebuffer fetch to the new resolve functions

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Chad Versace <chadversary@chromium.org>
7 years agoi965: Remove an unneeded render_cache_set_check_flush
Jason Ekstrand [Thu, 25 May 2017 19:14:52 +0000 (12:14 -0700)]
i965: Remove an unneeded render_cache_set_check_flush

This is only needed to fix rendering corruptions caused by not flushing
after doing a resolve operation.  The resolve now does all the needed
flushing so this is unnecessary.

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Chad Versace <chadversary@chromium.org>
7 years agoi965: Move color rendering to the new resolve functions
Jason Ekstrand [Thu, 25 May 2017 18:58:40 +0000 (11:58 -0700)]
i965: Move color rendering to the new resolve functions

This also removes an unneeded brw_render_cache_set_check_flush() call.
We were calling it in the case where the surface got resolved to satisfy
the flushing requirements around resolves.  However, blorp now does this
itself, so the extra is just redundant.

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Chad Versace <chadversary@chromium.org>
7 years agoi965: Move texturing to the new resolve functions
Jason Ekstrand [Thu, 25 May 2017 18:29:17 +0000 (11:29 -0700)]
i965: Move texturing to the new resolve functions

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Chad Versace <chadversary@chromium.org>
7 years agoi965: Use the new resolve function for several simple cases
Jason Ekstrand [Thu, 25 May 2017 05:09:51 +0000 (22:09 -0700)]
i965: Use the new resolve function for several simple cases

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Chad Versace <chadversary@chromium.org>
7 years agoi965/miptree: Add new entrypoints for resolve management
Jason Ekstrand [Thu, 25 May 2017 05:09:30 +0000 (22:09 -0700)]
i965/miptree: Add new entrypoints for resolve management

This commit adds a new unified interface for doing resolves.  The basic
format is that, prior to any surface access such as texturing or
rendering, you call intel_miptree_prepare_access.  If the surface was
written, you call intel_miptree_finish_write.  These two functions take
parameters which tell them whether or not auxiliary compression and fast
clears are supported on the surface.  Later commits will add wrappers
around these two functions for texturing, rendering, etc.

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
7 years agointel/isl: Add an enum for describing auxiliary compression state
Jason Ekstrand [Thu, 25 May 2017 03:01:12 +0000 (20:01 -0700)]
intel/isl: Add an enum for describing auxiliary compression state

This enum describes all of the states that a auxiliary compressed
surface can have.  All of the states as well as normative language for
referring to each of the compression operations is provided in the
truly colossal comment for the new isl_aux_state enum.  There is also
a diagram showing how surfaces move between the different states.

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Nanley Chery <nanley.g.chery@intel.com>
Reviewed-by: Chad Versace <chadversary@chromium.org>
7 years agoi965: Combine render target resolve code
Jason Ekstrand [Thu, 25 May 2017 18:47:13 +0000 (11:47 -0700)]
i965: Combine render target resolve code

We have two different bits of resolve code for render targets: one in
brw_draw where it's always been and one in brw_context to deal with sRGB
on gen9.  Let's pull them together.

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Chad Versace <chadversary@chromium.org>
7 years agoi965: Be a bit more conservative about certain resolves
Jason Ekstrand [Wed, 31 May 2017 17:06:31 +0000 (10:06 -0700)]
i965: Be a bit more conservative about certain resolves

There are several places where we were resolving the entire miptree
when we really only needed to resolve a single slice.  Let's avoid the
unneeded resolving.

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Chad Versace <chadversary@chromium.org>
7 years agoi965/blorp: Move MCS allocation earlier for clears
Jason Ekstrand [Fri, 26 May 2017 02:13:47 +0000 (19:13 -0700)]
i965/blorp: Move MCS allocation earlier for clears

This way it happens before we call get_aux_state.

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Chad Versace <chadversary@chromium.org>
7 years agoi965/blorp: Refactor do_single_blorp_clear
Jason Ekstrand [Tue, 23 May 2017 04:27:50 +0000 (21:27 -0700)]
i965/blorp: Refactor do_single_blorp_clear

Previously, we had two checks for can_fast_clear and a tiny bit of
shared code in between.  This commit pulls all of the fast clear code
together and duplicates the tiny bit that declares some surface structs
and calls blorp_surf_for_miptree.  The duplication is no real loss and
we're about to change the two in slightly different ways.

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Chad Versace <chadversary@chromium.org>
7 years agoi965/blorp: Take an explicit fast clear op in resolve_color
Jason Ekstrand [Thu, 25 May 2017 23:59:12 +0000 (16:59 -0700)]
i965/blorp: Take an explicit fast clear op in resolve_color

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Chad Versace <chadversary@chromium.org>
7 years agoi965/miptree: Move color resolve on map to intel_miptree_map
Jason Ekstrand [Thu, 25 May 2017 05:06:29 +0000 (22:06 -0700)]
i965/miptree: Move color resolve on map to intel_miptree_map

None of the other methods such as blit work with CCS either so we need
to do the resolve for all maps.  This change also makes us only resolve
the one slice we're mapping and not the entire image.

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
7 years agoi965: Inline renderbuffer_att_set_needs_depth_resolve
Jason Ekstrand [Thu, 25 May 2017 18:04:38 +0000 (11:04 -0700)]
i965: Inline renderbuffer_att_set_needs_depth_resolve

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
7 years agoi965: Get rid of intel_renderbuffer_resolve_*
Jason Ekstrand [Thu, 25 May 2017 04:55:59 +0000 (21:55 -0700)]
i965: Get rid of intel_renderbuffer_resolve_*

There is exactly one caller so it's a bit pointless to have all of this
plumbing.  Just inline it at the one place it's used.

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Chad Versace <chadversary@chromium.org>
7 years agoi965/miptree: Refactor intel_miptree_resolve_color
Jason Ekstrand [Mon, 22 May 2017 16:23:45 +0000 (09:23 -0700)]
i965/miptree: Refactor intel_miptree_resolve_color

The new version now takes a range of levels as well as a range of
layers.  It should also be a tiny bit faster because it only walks the
resolve_map list once instead of once per layer.

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Chad Versace <chadversary@chromium.org>
7 years agoi965/miptree: Clean up the depth resolve helpers a little
Jason Ekstrand [Mon, 22 May 2017 15:59:49 +0000 (08:59 -0700)]
i965/miptree: Clean up the depth resolve helpers a little

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Chad Versace <chadversary@chromium.org>
7 years agoi965/surface_state: Images can't handle CCS at all
Jason Ekstrand [Fri, 26 May 2017 16:03:42 +0000 (09:03 -0700)]
i965/surface_state: Images can't handle CCS at all

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Chad Versace <chadversary@chromium.org>
7 years agoi965: Mark depth surfaces as needing a HiZ resolve after blitting
Jason Ekstrand [Fri, 26 May 2017 16:33:55 +0000 (09:33 -0700)]
i965: Mark depth surfaces as needing a HiZ resolve after blitting

Cc: "17.0 17.1" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Chad Versace <chadversary@chromium.org>
7 years agost_glsl_to_tgsi: cleanup variable storage search.
Dave Airlie [Thu, 8 Jun 2017 03:27:46 +0000 (13:27 +1000)]
st_glsl_to_tgsi: cleanup variable storage search.

I forgot to put the cleanup in earlier.

Signed-off-by: Dave Airlie <airlied@redhat.com>
7 years agomesa/main: fix gl_buffer_index enum comparison
Rob Herring [Tue, 30 May 2017 20:30:38 +0000 (15:30 -0500)]
mesa/main: fix gl_buffer_index enum comparison

For clang, enums are unsigned by default and gives the following warning:

external/mesa3d/src/mesa/main/buffers.c:764:21: warning: comparison of constant -1 with expression of type 'gl_buffer_index' is always false [-Wtautological-constant-out-of-range-compare]
      if (srcBuffer == -1) {
          ~~~~~~~~~ ^  ~~

Replace -1 with an enum value to fix this.

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Signed-off-by: Rob Herring <robh@kernel.org>
7 years agoglsl: fix bounds check in blob_overwrite_bytes
Rob Herring [Thu, 1 Jun 2017 00:07:40 +0000 (19:07 -0500)]
glsl: fix bounds check in blob_overwrite_bytes

clang gives a warning in blob_overwrite_bytes because offset type is
size_t which is unsigned:

src/compiler/glsl/blob.c:110:15: warning: comparison of unsigned expression < 0 is always false [-Wtautological-compare]
   if (offset < 0 || blob->size - offset < to_write)
       ~~~~~~ ^ ~

Remove the less than 0 check to fix this.

Additionally, if offset is greater than blob->size, the 2nd check would
be false due to unsigned math. Rewrite the check to avoid subtraction.

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Signed-off-by: Rob Herring <robh@kernel.org>
7 years agost_glsl_to_tgsi: replace variables tracking list with a hash table
Dave Airlie [Tue, 30 May 2017 05:52:14 +0000 (15:52 +1000)]
st_glsl_to_tgsi: replace variables tracking list with a hash table

This removes the linear search which is fail when number of variables
goes up to 30000 or so.

Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
7 years agost_glsl_to_tgsi: rewrite rename registers to use array fully.
Dave Airlie [Tue, 30 May 2017 05:52:13 +0000 (15:52 +1000)]
st_glsl_to_tgsi: rewrite rename registers to use array fully.

Instead of having to search the whole array, just use the whole
thing and store a valid bit in there with the rename.

Removes this from the profile on some of the fp64 tests

Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
7 years agost_glsl_to_tgsi: bump index back up to 32-bit
Dave Airlie [Tue, 30 May 2017 05:52:11 +0000 (15:52 +1000)]
st_glsl_to_tgsi: bump index back up to 32-bit

with some of the fp64 emulation, we are seeing shaders coming in with
> 32K temps, they go out with 40 or so used, but while doing register
renumber we need to store a lot of them.

So bump this fields back up to 32-bit.

Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
7 years agoutil/u_queue: fix a use-before-initialization race for queue->threads
Marek Olšák [Wed, 7 Jun 2017 21:13:35 +0000 (23:13 +0200)]
util/u_queue: fix a use-before-initialization race for queue->threads

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
7 years agoac/nir: remove another unused variable
Grazvydas Ignotas [Wed, 7 Jun 2017 21:00:02 +0000 (00:00 +0300)]
ac/nir: remove another unused variable

Declared by each loop already.
Trivial.

Signed-off-by: Grazvydas Ignotas <notasas@gmail.com>
7 years agoradv/meta: remove an unused variable
Grazvydas Ignotas [Tue, 6 Jun 2017 23:00:36 +0000 (02:00 +0300)]
radv/meta: remove an unused variable

Trivial.

Signed-off-by: Grazvydas Ignotas <notasas@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
7 years agoac/nir: convert several ifs to a switch
Grazvydas Ignotas [Tue, 6 Jun 2017 22:55:26 +0000 (01:55 +0300)]
ac/nir: convert several ifs to a switch

Also solve "outinfo may be used uninitialized" warning by putting in an
unreachable().

Signed-off-by: Grazvydas Ignotas <notasas@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
7 years agoac/nir: mark some arguments const
Grazvydas Ignotas [Tue, 6 Jun 2017 22:52:30 +0000 (01:52 +0300)]
ac/nir: mark some arguments const

Most functions are only inspecting nir, so nir related arguments can be
marked const. Some more can be done if/when some nir changes are
accepted.

Signed-off-by: Grazvydas Ignotas <notasas@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
7 years agoradeonsi: Use libdrm to get chipset name
Samuel Li [Tue, 6 Jun 2017 22:21:08 +0000 (18:21 -0400)]
radeonsi: Use libdrm to get chipset name

v2: Add a func pointer to radeon_winsys to support radeon later.

Change-Id: I614ea71424f9e5c97e4ae68654315d28c89eaa5f
Signed-off-by: Samuel Li <Samuel.Li@amd.com>
Signed-off-by: Marek Olšák <marek.olsak@amd.com>
7 years agoutil: Add extern c to u_dynarray.h
Thomas Helland [Sat, 3 Jun 2017 17:59:07 +0000 (19:59 +0200)]
util: Add extern c to u_dynarray.h

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Eric Engestrom <eric.engestrom@imgtec.com>
7 years agonir: Delete nir_array.h
Thomas Helland [Wed, 7 Jun 2017 18:46:05 +0000 (20:46 +0200)]
nir: Delete nir_array.h

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Eric Engestrom <eric.engestrom@imgtec.com>