mesa.git
13 years agocell: comment-out unused fields, functions
Brian Paul [Tue, 27 Jul 2010 02:50:17 +0000 (20:50 -0600)]
cell: comment-out unused fields, functions

13 years agocell: make functions static
Brian Paul [Tue, 27 Jul 2010 02:49:54 +0000 (20:49 -0600)]
cell: make functions static

13 years agocell: fix segfault when freeing samplers
Brian Paul [Tue, 27 Jul 2010 02:49:35 +0000 (20:49 -0600)]
cell: fix segfault when freeing samplers

13 years agocell: comment-out unneeded padding field
Brian Paul [Tue, 27 Jul 2010 02:49:09 +0000 (20:49 -0600)]
cell: comment-out unneeded padding field

13 years agocell: added const qualifier
Brian Paul [Tue, 27 Jul 2010 02:48:46 +0000 (20:48 -0600)]
cell: added const qualifier

13 years agoutil: fix unused function warning on non-x86
Brian Paul [Tue, 27 Jul 2010 02:48:29 +0000 (20:48 -0600)]
util: fix unused function warning on non-x86

13 years agoglx: Remove function prototypes no longer necessary
Kristian Høgsberg [Mon, 26 Jul 2010 21:23:23 +0000 (17:23 -0400)]
glx: Remove function prototypes no longer necessary

13 years agoi965: Fix reversed naming of the operations in compute-to-mrf optimization.
Eric Anholt [Mon, 26 Jul 2010 20:01:45 +0000 (13:01 -0700)]
i965: Fix reversed naming of the operations in compute-to-mrf optimization.

Also fix up comments, so that the difference between the two passes is
clarified.

13 years agoi965: Clean up a few magic numbers to use brw_defines.h defs.
Eric Anholt [Mon, 26 Jul 2010 19:55:28 +0000 (12:55 -0700)]
i965: Clean up a few magic numbers to use brw_defines.h defs.

13 years agoi965: Use MIN2, MAX2 instead of rolling our own.
Eric Anholt [Mon, 26 Jul 2010 19:51:22 +0000 (12:51 -0700)]
i965: Use MIN2, MAX2 instead of rolling our own.

13 years agoi965: Fold the "is arithmetic" bit of 965 opcodes into the opcode list.
Eric Anholt [Mon, 26 Jul 2010 19:49:32 +0000 (12:49 -0700)]
i965: Fold the "is arithmetic" bit of 965 opcodes into the opcode list.

13 years agoi965: Remove some duped register size/count definitions
Eric Anholt [Mon, 26 Jul 2010 19:43:51 +0000 (12:43 -0700)]
i965: Remove some duped register size/count definitions

13 years agoi965: Move the GRF-to-MRF optimizations to brw_optimize.c.
Eric Anholt [Mon, 26 Jul 2010 19:41:39 +0000 (12:41 -0700)]
i965: Move the GRF-to-MRF optimizations to brw_optimize.c.

13 years agoi965: Improve (i.e. remove) some grf-to-mrf unnecessary moves
Benjamin Segovia [Mon, 26 Jul 2010 04:30:19 +0000 (21:30 -0700)]
i965: Improve (i.e. remove) some grf-to-mrf unnecessary moves

Several routines directly analyze the grf-to-mrf moves from the Gen
binary code. When it is possible, the mov is removed and the message
register is directly written in the arithmetic instruction

Also redundant mrf-to-grf moves are removed (frequently for example,
when sampling many textures with the same uv)

Code was tested with piglit, warsow and nexuiz on an Ironlake
machine. No regression was found there

Note that the optimizations are *deactivated* on Gen4 and Gen6 since I
did test them properly yet. No reason there are bugs but who knows

The optimizations are currently done in branch free programs *only*.
Considering branches is more complicated and there are actually two
paths: one for branch free programs and one for programs with branches

Also some other optimizations should be done during the emission
itself but considering that some code is shader between vertex shaders
(AOS) and pixel shaders (SOA) and that we may have branches or not, it
is pretty hard to both factorize the code and have one good set of
strategies

13 years agoglx: Drop debug fprintf that snug in with the previous commit
Kristian Høgsberg [Mon, 26 Jul 2010 20:06:05 +0000 (16:06 -0400)]
glx: Drop debug fprintf that snug in with the previous commit

13 years agoglx: Enable copy subbuffer patch when GLX_DIRECT_RENDERING is #defined
Kristian Høgsberg [Mon, 26 Jul 2010 19:50:02 +0000 (15:50 -0400)]
glx: Enable copy subbuffer patch when GLX_DIRECT_RENDERING is #defined

Depending on __DRI_COPY_SUB_BUFFER doesn't work when we no longer include
dri_interface.h.

https://bugs.freedesktop.org/show_bug.cgi?id=29264

13 years agost/egl: Fix debug line
nobled [Tue, 8 Jun 2010 13:00:17 +0000 (13:00 +0000)]
st/egl: Fix debug line

Acked-by: Jakob Bornecrantz <jakob@vmware.com>
13 years agoutil: fix CPU detection on OS X
nobled [Sat, 24 Jul 2010 12:59:40 +0000 (12:59 +0000)]
util: fix CPU detection on OS X

s/PIPE_OS_DARWIN/PIPE_OS_APPLE, since there is no PIPE_OS_DARWIN.

Acked-by: Vinson Lee <vlee@vmware.com>
13 years agost/xorg: fix use-after-free
nobled [Sat, 24 Jul 2010 12:05:30 +0000 (12:05 +0000)]
st/xorg: fix use-after-free

Acked-by: Jakob Bornecrantz <jakob@vmware.com>
13 years agoi965g: Enable llvm in dri driver if built
Jakob Bornecrantz [Fri, 23 Jul 2010 04:50:27 +0000 (21:50 -0700)]
i965g: Enable llvm in dri driver if built

13 years agoi915g: Fix llvm build
nobled [Sat, 24 Jul 2010 12:04:29 +0000 (12:04 +0000)]
i915g: Fix llvm build

Acked-by: Jakob Bornecrantz <jakob@vmware.com>
13 years agoglx: Drop duplicate psc field in dri context struct
Kristian Høgsberg [Mon, 26 Jul 2010 19:17:00 +0000 (15:17 -0400)]
glx: Drop duplicate psc field in dri context struct

Same problem as fixed for drisw in 4d58b5b482d06ab8d4c4b2db33d0b48b7c82d064.

13 years agoi965: Allow VS MOVs to use immediate constants.
Eric Anholt [Mon, 26 Jul 2010 11:20:33 +0000 (04:20 -0700)]
i965: Allow VS MOVs to use immediate constants.

Clarifies program assembly, and with a little tweak to always use
constant_map, we could cut down on constant buffer payload.

13 years agor600g: implememt the LIT instruction
Stephan Schmid [Mon, 26 Jul 2010 05:52:12 +0000 (07:52 +0200)]
r600g: implememt the LIT instruction

13 years agoutil: fix another mutex leak in mempool
Marek Olšák [Mon, 26 Jul 2010 12:56:48 +0000 (14:56 +0200)]
util: fix another mutex leak in mempool

By fixing one, I introduced another. Crap.

13 years agoutil: fix mutex leaks in mempool
Marek Olšák [Mon, 26 Jul 2010 12:53:06 +0000 (14:53 +0200)]
util: fix mutex leaks in mempool

13 years agor300g: fix macro substitution problem
Dave Airlie [Mon, 26 Jul 2010 01:56:12 +0000 (11:56 +1000)]
r300g: fix macro substitution problem

isn't a problem yet, but have issues in hiz branch.

Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agor300g: implement D24X8 texture sampling for r3xx-r4xx
Marek Olšák [Sun, 25 Jul 2010 21:40:51 +0000 (23:40 +0200)]
r300g: implement D24X8 texture sampling for r3xx-r4xx

Because the hw can't sample it, I reinterpret the format as G16R16 and
sample the G component. This gives 16 bits of precision, which should be
enough for depth texturing (surprisingly, the sampled values are exactly
the same as in D16 textures).

This also enables EXT_packed_depth_stencil on those old chipsets, finally.

13 years agor300g: make sure a texture is large enough for the CBZB clear
Marek Olšák [Sun, 25 Jul 2010 08:07:12 +0000 (10:07 +0200)]
r300g: make sure a texture is large enough for the CBZB clear

The number of macrotiles in the Y direction must be even, otherwise memory
corruption may happen (e.g. broken fonts). Basically, if we get a buffer
in resource_from_handle, we can determine from the buffer size whether it's
safe to use the CBZB clear or not.

13 years agor300g: do not use TXPITCH_EN if the width is POT and the height is NPOT
Marek Olšák [Sat, 24 Jul 2010 21:05:40 +0000 (23:05 +0200)]
r300g: do not use TXPITCH_EN if the width is POT and the height is NPOT

13 years agor300g: do not use TXPITCH_EN for power-of-two textures from the DDX
Marek Olšák [Sat, 24 Jul 2010 20:52:01 +0000 (22:52 +0200)]
r300g: do not use TXPITCH_EN for power-of-two textures from the DDX

We were using TXPITCH_EN for textures from the DDX since ever, for nothing.

13 years agor300g: cleanup texture creation code
Marek Olšák [Sat, 24 Jul 2010 19:32:53 +0000 (21:32 +0200)]
r300g: cleanup texture creation code

This decouples initializing a texture layout/miptree description
from an actual texture creation, it also partially unifies texture_create
and texture_from_handle.

r300_texture inherits r300_texture_desc, which inherits u_resource.

The CBZB clear criteria are moved to r300_texture_desc::cbzb_allowed[level].

And other minor cleanups.

13 years agor300g: reject resources from handles which are not large enough
Marek Olšák [Sat, 24 Jul 2010 01:34:18 +0000 (03:34 +0200)]
r300g: reject resources from handles which are not large enough

The driver gets a buffer and its size in resource_from_handle.
It computes the required minimum buffer size from given texture
properties, and compares the two sizes.

This is to early detect DDX bugs.

13 years agor300g: cleanup texture debug logging
Marek Olšák [Sat, 24 Jul 2010 01:17:32 +0000 (03:17 +0200)]
r300g: cleanup texture debug logging

13 years agor300g: do not align texture height to 2^n for 1D and 2D non-mipmapped textures
Marek Olšák [Sat, 24 Jul 2010 01:07:45 +0000 (03:07 +0200)]
r300g: do not align texture height to 2^n for 1D and 2D non-mipmapped textures

I don't remember why the alignment was there, but it seems to be
no longer needed. I guess it was a dirty fix for some other bug.

13 years agonvfx: Move declaration before code.
Vinson Lee [Sun, 25 Jul 2010 04:10:45 +0000 (21:10 -0700)]
nvfx: Move declaration before code.

13 years agoglx: Drop duplicate psc field in drisw context struct
Kristian Høgsberg [Sun, 25 Jul 2010 01:43:25 +0000 (21:43 -0400)]
glx: Drop duplicate psc field in drisw context struct

Causing a crash in drisw MakeCurrent.

13 years agoglx: Fix another case of confusing driContext and dri2_context *
Eric Anholt [Fri, 23 Jul 2010 22:02:14 +0000 (15:02 -0700)]
glx: Fix another case of confusing driContext and dri2_context *

13 years agoglx: Correctly look up the dri2 context pointer for SetTexBuffer.
Eric Anholt [Fri, 23 Jul 2010 21:51:06 +0000 (14:51 -0700)]
glx: Correctly look up the dri2 context pointer for SetTexBuffer.

gc->driContext points at the second member of the dri2 context.  The
dri2 context is just a subclass of the GLX context.

Fixes piglit tfp testcase.

13 years agoglx: Don't destroy context with XID 0
Kristian Høgsberg [Sat, 24 Jul 2010 02:05:21 +0000 (22:05 -0400)]
glx: Don't destroy context with XID 0

We use XID 0 to indicate the context has already been destroyed, but it's
currently bound.

13 years agoglx: Fix use after free case when destroying screens
Kristian Høgsberg [Sat, 24 Jul 2010 01:51:58 +0000 (21:51 -0400)]
glx: Fix use after free case when destroying screens

13 years agoglx: zero out drawable structs after allocation
Kristian Høgsberg [Sat, 24 Jul 2010 01:45:05 +0000 (21:45 -0400)]
glx: zero out drawable structs after allocation

13 years agoglx: Refactor and simplify context creation
Kristian Høgsberg [Fri, 23 Jul 2010 20:15:31 +0000 (16:15 -0400)]
glx: Refactor and simplify context creation

This lets us better separate context creation between the different
backends.

13 years agoglx: Fix indirect screen initialization
Kristian Høgsberg [Fri, 23 Jul 2010 12:32:25 +0000 (08:32 -0400)]
glx: Fix indirect screen initialization

https://bugs.freedesktop.org/show_bug.cgi?id=29225

13 years agoglx: Move context destroy to context vtable
Kristian Høgsberg [Fri, 23 Jul 2010 03:45:18 +0000 (23:45 -0400)]
glx: Move context destroy to context vtable

13 years agoutil: Add PIPE_OS_CYGWIN to u_network.
Vinson Lee [Sat, 24 Jul 2010 01:47:21 +0000 (18:47 -0700)]
util: Add PIPE_OS_CYGWIN to u_network.

13 years agor600g: first pass at texture support
Jerome Glisse [Sat, 24 Jul 2010 00:55:48 +0000 (20:55 -0400)]
r600g: first pass at texture support

This add texture support to the assembler, generated code is
wrong (tested against working dump).

Signed-off-by: Jerome Glisse <jglisse@redhat.com>
13 years agoscons: Add sunos5 to list of accepted platforms.
Vinson Lee [Sat, 24 Jul 2010 00:53:35 +0000 (17:53 -0700)]
scons: Add sunos5 to list of accepted platforms.

13 years agor600g: Fix SCons build.
Vinson Lee [Sat, 24 Jul 2010 00:50:35 +0000 (17:50 -0700)]
r600g: Fix SCons build.

13 years agoscons: Use '-Wmissing-field-initializers' on GCC 4.0 and greater only.
Vinson Lee [Sat, 24 Jul 2010 00:24:21 +0000 (17:24 -0700)]
scons: Use '-Wmissing-field-initializers' on GCC 4.0 and greater only.

13 years agomesa: Fix Cygwin build with llvm enabled.
Vinson Lee [Fri, 23 Jul 2010 23:24:35 +0000 (16:24 -0700)]
mesa: Fix Cygwin build with llvm enabled.

On Cygwin locale_t in not available but 'llvm-config --cppflags' adds
the compiler flag -D_GNU_SOURCE to the build.

13 years agogallium: Fix build with llvm installed in non-standard location
Christopher James Halse Rogers [Fri, 11 Jun 2010 10:33:44 +0000 (20:33 +1000)]
gallium: Fix build with llvm installed in non-standard location

The es1, es2 and gl state trackers include draw_pipe.h, which includes
the llvm headers if MESA_LLVM is true, so we also need to add the
llvm seachpaths.

Similarly, gallivm and other gallium drivers need LLVM_CFLAGS to build when enabled.

Also fix xorg drivers, they didn't include LDFLAGS.

13 years agor600g: fix dp2, dp3, dp4 tokens
Jerome Glisse [Fri, 23 Jul 2010 22:19:13 +0000 (18:19 -0400)]
r600g: fix dp2, dp3, dp4 tokens

We need to make sure dp are all mirror accross the alu unit.

Signed-off-by: Jerome Glisse <jglisse@redhat.com>
13 years agor600g: add RSQ token support
Jerome Glisse [Fri, 23 Jul 2010 21:49:26 +0000 (17:49 -0400)]
r600g: add RSQ token support

Could serve as an example on how to add more token
support.

Signed-off-by: Jerome Glisse <jglisse@redhat.com>
13 years agor600g: drop compiler stuff and switch over dumb tgsi assembler
Jerome Glisse [Fri, 23 Jul 2010 21:32:32 +0000 (17:32 -0400)]
r600g: drop compiler stuff and switch over dumb tgsi assembler

Writing a compiler is time consuming and error prone in
order to allow r600g to further progress in the meantime
i wrote a simple tgsi assembler, it does stupid thing but
i would rather keep the code simple than having people
trying to optimize code it does.

Signed-off-by: Jerome Glisse <jglisse@redhat.com>
13 years agoglx: Don't try to swap a front buffer if we don't have one.
Eric Anholt [Fri, 23 Jul 2010 20:41:13 +0000 (13:41 -0700)]
glx: Don't try to swap a front buffer if we don't have one.

Fixes glean glsl1 since 7b7845a076c933e096ac511b4184141ba194449a

13 years agost/mesa: get rid of unneeded ureg_writemask()
Brian Paul [Fri, 23 Jul 2010 19:50:09 +0000 (13:50 -0600)]
st/mesa: get rid of unneeded ureg_writemask()

13 years agost/mesa: fix bug in emit_adjusted_wpos()
Brian Paul [Fri, 23 Jul 2010 19:39:36 +0000 (13:39 -0600)]
st/mesa: fix bug in emit_adjusted_wpos()

If we bias x,y we still need to pass through z,w in case the shader
reads gl_FragCoord.z or .w.

Fixes fd.o bug 29183 (piglit glsl-bug-22603).

NOTE: This is a candidate for the 7.8 branch.

13 years agoi965: Cleanly fail programs with unsupported array access.
Eric Anholt [Fri, 23 Jul 2010 17:01:45 +0000 (10:01 -0700)]
i965: Cleanly fail programs with unsupported array access.

This should be more useful for developers and for bug triaging than
just generating wrong code.

13 years agoi965: Add support for VS relative addressing of temporary arrays.
Eric Anholt [Fri, 23 Jul 2010 17:34:29 +0000 (10:34 -0700)]
i965: Add support for VS relative addressing of temporary arrays.

Fixes glsl-vs-arrays.  Bug #27388.

13 years agodraw: add small ybias factor for drawing wide points
Brian Paul [Fri, 23 Jul 2010 16:32:25 +0000 (10:32 -0600)]
draw: add small ybias factor for drawing wide points

Fixes minor rasterization error detected by some tests.

13 years agodocs: document new extensions for Gallium
Brian Paul [Thu, 22 Jul 2010 22:55:41 +0000 (16:55 -0600)]
docs: document new extensions for Gallium

13 years agosoftpipe: Check for NULL pointer in sp_destroy_tile_cache().
Michal Krol [Thu, 22 Jul 2010 16:32:50 +0000 (18:32 +0200)]
softpipe: Check for NULL pointer in sp_destroy_tile_cache().

13 years agosoftpipe: Check for NULL pointer in sp_destroy_tex_tile_cache().
Michal Krol [Thu, 22 Jul 2010 16:32:31 +0000 (18:32 +0200)]
softpipe: Check for NULL pointer in sp_destroy_tex_tile_cache().

13 years agonv50: implement depth clamp
Christoph Bumiller [Fri, 23 Jul 2010 10:03:33 +0000 (12:03 +0200)]
nv50: implement depth clamp

13 years agoxorg/vmwgfx: Implement early mode pruning based on max fb size.
Thomas Hellstrom [Fri, 23 Jul 2010 02:44:26 +0000 (04:44 +0200)]
xorg/vmwgfx: Implement early mode pruning based on max fb size.

Also move some initialization from screen init to pre-init, now
that it is possible.

Also import a new vmwgfx drm (1.3) header.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
13 years agost/xorg: Init the Gallium3D / libkms resources in pre-init.
Thomas Hellstrom [Fri, 23 Jul 2010 02:36:43 +0000 (04:36 +0200)]
st/xorg: Init the Gallium3D / libkms resources in pre-init.

This makes it possible to prune modes already in pre-init.
We also keep these resources alive across server generations, and
they are implicitly closed on server exit.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
13 years agost/xorg: Kill a couple of compilation warnings
Thomas Hellstrom [Thu, 22 Jul 2010 01:41:10 +0000 (03:41 +0200)]
st/xorg: Kill a couple of compilation warnings

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
13 years agost/xorg vmwgfx/xorg: Add a pre-init customizer callback
Thomas Hellstrom [Thu, 22 Jul 2010 00:05:35 +0000 (02:05 +0200)]
st/xorg vmwgfx/xorg: Add a pre-init customizer callback

Add a customizer callback just before initial config setting, so that the
customizer code can initialize the mode validator using the drm
file-descriptor.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
13 years agost/xorg: Add a possibility to prune modes and limit fb allocation size based on max...
Thomas Hellstrom [Wed, 21 Jul 2010 23:49:46 +0000 (01:49 +0200)]
st/xorg: Add a possibility to prune modes and limit fb allocation size based on max fb size.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
13 years agollvmpipe: Partially fix resource texture from_handle
Jakob Bornecrantz [Fri, 23 Jul 2010 03:18:51 +0000 (20:18 -0700)]
llvmpipe: Partially fix resource texture from_handle

13 years agollvmpipe: Don't align values already aligned
Jakob Bornecrantz [Fri, 23 Jul 2010 03:17:35 +0000 (20:17 -0700)]
llvmpipe: Don't align values already aligned

13 years agoi915g: Rename winsys debug options
Jakob Bornecrantz [Fri, 23 Jul 2010 03:13:02 +0000 (20:13 -0700)]
i915g: Rename winsys debug options

13 years agoi915g: Allow wrapping with software pipes
Jakob Bornecrantz [Fri, 23 Jul 2010 03:11:20 +0000 (20:11 -0700)]
i915g: Allow wrapping with software pipes

13 years agoi915g: Set total_nblocksy in from_handle
Jakob Bornecrantz [Fri, 23 Jul 2010 03:03:29 +0000 (20:03 -0700)]
i915g: Set total_nblocksy in from_handle

13 years agoi915g: Add some debug prints in texture code
Jakob Bornecrantz [Fri, 23 Jul 2010 03:02:58 +0000 (20:02 -0700)]
i915g: Add some debug prints in texture code

13 years agoi915g: Ifdef out debug code on non-debug builds
Jakob Bornecrantz [Fri, 23 Jul 2010 03:00:28 +0000 (20:00 -0700)]
i915g: Ifdef out debug code on non-debug builds

13 years agotgsi: Fix error message on invalid swizzle parse
Jakob Bornecrantz [Fri, 16 Jul 2010 04:11:24 +0000 (21:11 -0700)]
tgsi: Fix error message on invalid swizzle parse

13 years agoglx: Drop screen argument to GetGLXDRIDrawable
Kristian Høgsberg [Fri, 23 Jul 2010 02:36:37 +0000 (22:36 -0400)]
glx: Drop screen argument to GetGLXDRIDrawable

We'll just get it from the returned drawable when we need it.

13 years agoglx: Move WaitGL, WaitX, UseXFont to context vtable functions
Kristian Høgsberg [Fri, 23 Jul 2010 02:24:00 +0000 (22:24 -0400)]
glx: Move WaitGL, WaitX, UseXFont to context vtable functions

13 years agoglx: Allocate the __GLXcontext in the DRI drivers
Kristian Høgsberg [Fri, 23 Jul 2010 01:24:14 +0000 (21:24 -0400)]
glx: Allocate the __GLXcontext in the DRI drivers

13 years agoglx: Use _Xglobal_lock for protecting extension display list
Kristian Høgsberg [Fri, 23 Jul 2010 01:11:02 +0000 (21:11 -0400)]
glx: Use _Xglobal_lock for protecting extension display list

Avoids double locking glXLock in the X wire to event handlers.

13 years agoglsl: remove invalid _mesa_problem() call
Brian Paul [Thu, 22 Jul 2010 22:35:43 +0000 (16:35 -0600)]
glsl: remove invalid _mesa_problem() call

Fixes fd.o bug 29206.

13 years agodraw: re-order optimization passes depending on LLVM version, 32/64-bit
Brian Paul [Thu, 22 Jul 2010 22:07:18 +0000 (16:07 -0600)]
draw: re-order optimization passes depending on LLVM version, 32/64-bit

This is a work-around for an apparent bug in LLVM seen with piglit's
glsl-vs-sqrt-zero test.

13 years agodraw: added new assertions to clipping code
Brian Paul [Thu, 22 Jul 2010 17:18:45 +0000 (11:18 -0600)]
draw: added new assertions to clipping code

13 years agoi965: Respect VS/VP point size result when enabled.
Eric Anholt [Thu, 22 Jul 2010 17:04:24 +0000 (10:04 -0700)]
i965: Respect VS/VP point size result when enabled.

Fixes glsl-vs-point-size.

13 years agoi965: Fix the disasm output for da16 src widths.
Eric Anholt [Thu, 22 Jul 2010 08:53:43 +0000 (01:53 -0700)]
i965: Fix the disasm output for da16 src widths.

This has confused me twice now.  It's a fixed width of 4 (usually a
region description of <4,4,1>), not 1.  If it was 1, we'd have been
skipping all over register space.

13 years agoi965: Avoid extra MOV in VS indirect register reads.
Eric Anholt [Thu, 22 Jul 2010 08:27:15 +0000 (01:27 -0700)]
i965: Avoid extra MOV in VS indirect register reads.

13 years agoi965: Fix up VS temporary array access for fixed index offset != 0.
Eric Anholt [Thu, 22 Jul 2010 07:06:10 +0000 (00:06 -0700)]
i965: Fix up VS temporary array access for fixed index offset != 0.

13 years agor600: Flip point sprite coordinates when rendering to an FBO.
Henri Verbeet [Thu, 22 Jul 2010 15:10:11 +0000 (11:10 -0400)]
r600: Flip point sprite coordinates when rendering to an FBO.

This supersedes http://lists.freedesktop.org/archives/mesa-dev/2010-July/001442.html.

13 years agoi965: In the VS, multiply the address reg by the appropriate register size.
Eric Anholt [Thu, 22 Jul 2010 04:45:34 +0000 (21:45 -0700)]
i965: In the VS, multiply the address reg by the appropriate register size.

The ARL value is increments of vec4 in the register file.  But
PROGRAM_TEMPORARY or PROGRAM_INPUT are stored as vec4s interleaved
between the two verts being executed (thus a vec8 each), compared to
PROGRAM_STATE_VAR being packed vec4s.

Fixes:
glsl-vs-arrays-2
glsl-vs-mov-after-deref
(without regressing glsl-vs-arrays-3)

13 years agoi965: Clean up brw_dp_READ_4_vs() now that it has fewer options to support.
Eric Anholt [Thu, 22 Jul 2010 02:49:39 +0000 (19:49 -0700)]
i965: Clean up brw_dp_READ_4_vs() now that it has fewer options to support.

13 years agoi965: Support relative addressed VS constant reads using the appropriate msg.
Eric Anholt [Thu, 22 Jul 2010 01:29:06 +0000 (18:29 -0700)]
i965: Support relative addressed VS constant reads using the appropriate msg.

The previous support was overly complicated by trying to use the same
1-OWORD message for both offsets.

13 years agoi965: Fix the DP read msg_control definitions other than plain OWORD.
Eric Anholt [Thu, 22 Jul 2010 01:41:06 +0000 (18:41 -0700)]
i965: Fix the DP read msg_control definitions other than plain OWORD.

13 years agoi965: Clean up dead code from the VS get_constant/get_reladdr_constant split.
Eric Anholt [Thu, 22 Jul 2010 00:35:21 +0000 (17:35 -0700)]
i965: Clean up dead code from the VS get_constant/get_reladdr_constant split.

13 years agoi956: Set the execution size correctly for scratch space writes.
Eric Anholt [Wed, 21 Jul 2010 22:54:53 +0000 (15:54 -0700)]
i956: Set the execution size correctly for scratch space writes.

Otherwise, the second half isn't written, and we end up reading back
black.

Fixes the remaining junk drawn in glsl-max-varyings, and will likely
help with a number of large real-world shaders.

13 years agoi965: Set the GEM domain flags for the scratch space.
Eric Anholt [Wed, 21 Jul 2010 20:52:37 +0000 (13:52 -0700)]
i965: Set the GEM domain flags for the scratch space.

They go into the render cache, so while we don't care about their
contents after execution, failing to note them could cause the writes
to be flushed over important buffer contents later.

13 years agoi965: Use the pretty define for 4-oword DP reads.
Eric Anholt [Wed, 21 Jul 2010 20:49:14 +0000 (13:49 -0700)]
i965: Use the pretty define for 4-oword DP reads.

13 years agoi965: Set the send commit bit on register spills as required pre-gen6.
Eric Anholt [Wed, 21 Jul 2010 20:07:12 +0000 (13:07 -0700)]
i965: Set the send commit bit on register spills as required pre-gen6.

Otherwise, the subsequent read may not get the written value.

13 years agoi965: Add disasm for dataport reads (register unspilling).
Eric Anholt [Wed, 21 Jul 2010 18:19:52 +0000 (11:19 -0700)]
i965: Add disasm for dataport reads (register unspilling).