Pekka Paalanen p...@iki.fi writes:
On Mon, 18 Jul 2011 08:09:17 -0600
Brian Paul bri...@vmware.com wrote:
On 07/15/2011 02:59 PM, Pekka Paalanen wrote:
On Fri, 15 Jul 2011 12:22:41 -0600
Brian Paulbri...@vmware.com wrote:
On 07/15/2011 10:07 AM, Dave Airlie wrote:
On Fri, Jul 15,
Signed-off-by: Francisco Jerez curroje...@riseup.net
---
[...]
Coincidentally nouveau's texture handling has always been structured
around two teximage map/unmap hooks (e.g. compare nouveau_teximage() and
the look of _mesa_store_teximage2d() after the API change), and AFAICT
the changes would
to _mesa_enable_imaging_extensions (via
driInitExtensions) is removed, EXT_blend_color, EXT_blend_logic_op,
EXT_blend_minmax, and EXT_blend_subtract are explicitly added to the
list.
Cc: Ben Skeggs bske...@redhat.com
Cc: Francisco Jerez curroje...@riseup.net
Cc: Viktor Novotný novik...@seznam.cz
Reviewed
Ian Romanick i...@freedesktop.org writes:
From: Ian Romanick ian.d.roman...@intel.com
The following extensions are always enabled, and drivers do not have
to option to disable them:
GL_ARB_copy_buffer
GL_ARB_draw_buffers / GL_ATI_draw_buffers
GL_ARB_multisample
not convinced that this makes sense. Vertex programs aren't
functional on nv2x at this moment, and on nv1x they probably never will
because there's no hardware support for it.
Cc: Ben Skeggs bske...@redhat.com
Cc: Francisco Jerez curroje...@riseup.net
Cc: Viktor Novotný novik...@seznam.cz
---
src/mesa
Ian Romanick i...@freedesktop.org writes:
On 08/30/2011 03:42 AM, Francisco Jerez wrote:
Ian Romanick i...@freedesktop.org writes:
From: Ian Romanick ian.d.roman...@intel.com
The following extensions are always enabled, and drivers do not have
to option to disable them
Ian Romanick i...@freedesktop.org writes:
On 08/30/2011 03:39 AM, Francisco Jerez wrote:
Ian Romanick i...@freedesktop.org writes:
From: Ian Romanick ian.d.roman...@intel.com
Specifically, the following extensions are enabled:
GL_ARB_vertex_program
GL_EXT_gpu_program_paramaters
From: Nicholas Miell nmi...@gmail.com
Add support for NV_fog_distance to the NV10 and NV20 drivers.
[ Francisco Jerez: Fix fog coord. signedness for
GL_EYE_RADIAL_NV/GL_EYE_PLANE on nv20 ]
---
Nicholas Miell nmi...@gmail.com writes:
This has not been tested on real hardware. I have no idea
Tom Stellard t...@stellard.net writes:
From: Tom Stellard thomas.stell...@amd.com
Acked-by: Francisco Jerez curroje...@riseup.net
---
.../state_trackers/clover/llvm/invocation.cpp | 47 ---
1 files changed, 39 insertions(+), 8 deletions(-)
diff --git a/src/gallium
to pass a
more a more detailed description of the hardware to compiler frontends.
Looks OK to me.
Reviewed-by: Francisco Jerez curroje...@riseup.net
v2:
- Adapt to libclc changes
---
src/gallium/docs/source/screen.rst |8 +-
src/gallium/drivers/r600/r600_llvm.c
Dmitry Cherkassov dcherkas...@gmail.com writes:
Signed-off-by: Dmitry Cherkassov dcherkas...@gmail.com
Thanks, pushed.
---
src/gallium/state_trackers/clover/core/kernel.cpp | 16
1 file changed, 16 insertions(+)
diff --git
Dmitry Cherkassov dcherkas...@gmail.com writes:
image2d is dynamic_cast-ed down from buffer, but is not it's
direct subclass. Make it inhernit buffer.
Hm, nope, you're not supposed to be able to dynamic_cast a buffer into
an image2d. If you found this problem in clEnqueueCopyBufferToImage()
Tom Stellard t...@stellard.net writes:
From: Tom Stellard thomas.stell...@amd.com
Hi Tom. This should probably leave out any devices that don't match the
type argument. To avoid duplicating the type-matching logic I think the
best would be to keep it implemented in terms of clGetDeviceIDs,
Tom Stellard t...@stellard.net writes:
Hi,
This series improves clover's kernel argument handling by allowing drivers to
specify an alignment for the input buffer and also by sign/zero extending the
arguments when required. In order to test these patches on r600g and
radeonsi,
you will
Tom Stellard t...@stellard.net writes:
On Sun, Jul 21, 2013 at 07:46:36PM +0200, Francisco Jerez wrote:
Tom Stellard t...@stellard.net writes:
Hi,
This series improves clover's kernel argument handling by allowing drivers
to
specify an alignment for the input buffer and also
Tom Stellard t...@stellard.net writes:
On Mon, Jul 22, 2013 at 09:24:12AM -0400, Jonathan Charest wrote:
To have non-static buffers in local memory, it is necessary to pass them
as arguments to the kernel. This was almost supported but the address
space mapping was missing to perform the
instead of a c-style cast, to make clear you're only interested in
safe integer-to-enum casts.
In any case this seems like a good thing to do, already in this state,
Reviewed-by: Francisco Jerez curroje...@riseup.net
For what it's worth, this patch is
Reviewed-by: Chad Versace chad.vers
Ian Romanick i...@freedesktop.org writes:
[...]
Disadvantage (b) can be made painless with the macro I discuss below.
IMHO it would be nicer to define generic templates implementing
overloads for all bitwise operators. They would have to reference the
bitmask_enumeration_traits structure
Chí-Thanh Christopher Nguyễn chith...@gentoo.org writes:
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=68504
Bugzilla: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58241
CC: mesa-sta...@lists.freedesktop.org
---
configure.ac | 11 +++
Chí-Thanh Christopher Nguyễn chith...@gentoo.org writes:
Francisco Jerez schrieb:
[...]
How about disabling AltiVec code generation by adding -mno-altivec to
CXXFLAGS? AFAICT what's going on is that the -maltivec option (which
is implied by your -mcpu=7450) enables a number of language
Chí-Thanh Christopher Nguyễn chith...@gentoo.org writes:
Chí-Thanh Christopher Nguyễn schrieb:
How about disabling AltiVec code generation by adding -mno-altivec to
CXXFLAGS? AFAICT what's going on is that the -maltivec option (which
is implied by your -mcpu=7450) enables a number of
Ian Romanick i...@freedesktop.org writes:
On 08/26/2013 01:10 PM, Francisco Jerez wrote:
[...]
The thing is that defining bitwise operators separately for each enum
type, as this patch and the macro solution do, doesn't stop the compiler
From using the corresponding built-in integer
Tom Stellard t...@stellard.net writes:
From: Tom Stellard thomas.stell...@amd.com
This way users can see compiler warnings.
---
src/gallium/state_trackers/clover/llvm/invocation.cpp | 3 +++
1 file changed, 3 insertions(+)
diff --git
-by: Francisco Jerez curroje...@riseup.net
src/gallium/state_trackers/clover/core/resource.cpp | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/gallium/state_trackers/clover/core/resource.cpp
b/src/gallium/state_trackers/clover/core/resource.cpp
index 6d348e6..3b003e9
comments. Does anyone else have anything to add before
I merge the gallium-compute branch to master?
On Tue, May 01, 2012 at 05:27:39PM +0200, Francisco Jerez wrote:
Define an interface that exposes the minimal functionality required to
implement some of the popular compute APIs. This commit adds
Francisco Jerez curroje...@riseup.net writes:
Tom Stellard thomas.stell...@amd.com writes:
Hi,
I've been testing these updated compute patches all week and they look
good to me. I don't think there are any outstanding complaints, so I'll
give my ACK for merging these into master.
Very
Christoph Bumiller e0425...@student.tuwien.ac.at writes:
This is not necessarily the product of MAX_BLOCK_SIZE[i].
Thanks, looks good. I've pushed it along with a fix for
clEnqueueNDRangeKernel() to check that the work-group dimensions
specified by the user are within the limit set by this
Tom Stellard tstel...@gmail.com writes:
---
.../state_trackers/clover/llvm/invocation.cpp | 174
+++-
1 files changed, 165 insertions(+), 9 deletions(-)
diff --git a/src/gallium/state_trackers/clover/llvm/invocation.cpp
Tom Stellard tstel...@gmail.com writes:
On Sun, May 13, 2012 at 12:40:43AM +0200, Francisco Jerez wrote:
[...]
I can think of two different ways this could work (your solution would
be somewhere in between):
- The r600g LLVM back-end could support some well-defined output object
format
Tom Stellard tstel...@gmail.com writes:
On Sun, May 13, 2012 at 04:10:50PM +0200, Francisco Jerez wrote:
Tom Stellard tstel...@gmail.com writes:
On Sun, May 13, 2012 at 12:40:43AM +0200, Francisco Jerez wrote:
[...]
I can think of two different ways this could work (your solution would
Tom Stellard thomas.stell...@amd.com writes:
On Mon, May 14, 2012 at 05:19:10PM +0200, Francisco Jerez wrote:
Tom Stellard tstel...@gmail.com writes:
On Sun, May 13, 2012 at 04:10:50PM +0200, Francisco Jerez wrote:
Tom Stellard tstel...@gmail.com writes:
[...]
I guess this might
Paul Berry stereotype...@gmail.com writes:
This patch adds .gitignore files to ignore the makefiles generated by
the gallium pipe loader and the clover OpenCL state tracker.
Reviewed-by: Francisco Jerez curroje...@riseup.net
---
src/gallium/auxiliary/pipe-loader/.gitignore |1 +
src
Hi Tom, thanks for the patches, just a few suggestions inline.
Tom Stellard tstel...@gmail.com writes:
This is for the llvm code that can't use extended initializers.
---
src/gallium/state_trackers/clover/core/module.hpp | 13 +
1 files changed, 13 insertions(+), 0
Tom Stellard tstel...@gmail.com writes:
---
configs/autoconf.in |1 +
configure.ac|3 +++
2 files changed, 4 insertions(+), 0 deletions(-)
diff --git a/configs/autoconf.in b/configs/autoconf.in
index 3c8f4c1..28f3cb4 100644
--- a/configs/autoconf.in
+++
), size(size), data(data) { }
+ section() : id(0), type(type), size(0), data() { }
+
type(type) doesn't initialize it to any definite value. Use
type(text), for example.
Other than that, Reviewed-by: Francisco Jerez curroje...@riseup.net
resource_id id;
type type
)
AM_CONDITIONAL(HAVE_DRM_LOADER_GALLIUM, test x$enable_gallium_drm_loader =
xyes)
+AM_CONDITIONAL(HAVE_GALLIUM_COMPUTE, test x$enable_opencl = xyes)
AC_SUBST([GALLIUM_MAKE_DIRS])
dnl prepend CORE_DIRS to SRC_DIRS
Reviewed-by: Francisco Jerez curroje...@riseup.net
pgpepkZUjnJyl.pgp
Description: PGP
Tom Stellard tstel...@gmail.com writes:
v2:
-Separate IR type and LLVM triple
-Do the OpenCL C-LLVM IR and linking steps for all PIPE_SHADER_IR
types.
---
There really isn't much functional change in this patch from v1, the only
really difference is that the decision to emit LLVM
Tom Stellard thomas.stell...@amd.com writes:
On Wed, May 23, 2012 at 02:04:20PM +0200, Francisco Jerez wrote:
[...]
Do we need this compatibility code? Don't you have to build a patched
version of LLVM from sources for this to work anyway?
You only need a patch if you want to use the r600
Tom Stellard tstel...@gmail.com writes:
v2:
-Separate IR type and LLVM triple
-Do the OpenCL C-LLVM IR and linking steps for all PIPE_SHADER_IR
types.
v3:
- Coding style fixes
- Removed compatibility code for LLVM 3.1
- Split build_module_llvm() into three functions:
-by: Francisco Jerez curroje...@riseup.net
pgpz5WbEJ49Ih.pgp
Description: PGP signature
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Tom Stellard tstel...@gmail.com writes:
From: Francisco Jerez curroje...@riseup.net
---
src/gallium/docs/source/screen.rst |5 +
src/gallium/include/pipe/p_defines.h |1 +
2 files changed, 6 insertions(+), 0 deletions(-)
diff --git a/src/gallium/docs/source/screen.rst
b
Tom Stellard tstel...@gmail.com writes:
A device now has two function for getting information about the IR
it needs to return.
ir_format() = returns the preferred IR
ir_target() = returns the triple for the target that is understood by
clang/llvm.
v2:
- renamed
Tom Stellard tstel...@gmail.com writes:
From: Francisco Jerez curroje...@riseup.net
v2: Tom Stellard
- Update CAP description
---
src/gallium/docs/source/screen.rst |5 +
src/gallium/include/pipe/p_defines.h |1 +
2 files changed, 6 insertions(+), 0 deletions(-)
diff
@@ public:
std::vectorsize_t max_block_size() const;
std::string device_name() const;
std::string vendor_name() const;
+ enum pipe_shader_ir ir_format() const;
std::string ir_target() const;
friend struct _cl_command_queue;
Reviewed-by: Francisco Jerez curroje
target) {
+clover::compile_program_tgsi(const compat::string source) {
const char *body = source.find(COMP\n);
module m;
Reviewed-by: Francisco Jerez curroje...@riseup.net
pgpRc90oOwVDN.pgp
Description: PGP signature
___
mesa-dev mailing list
tstel...@gmail.com writes:
From: Tom Stellard thomas.stell...@amd.com
This structure is used to pass non-TGSI programs to the drivers. clover
will use this for LLVM bytecode programs.
---
src/gallium/include/pipe/p_state.h |9 +
1 files changed, 9 insertions(+), 0
tstel...@gmail.com writes:
From: Tom Stellard thomas.stell...@amd.com
v2:
-Separate IR type and LLVM triple
-Do the OpenCL C-LLVM IR and linking steps for all PIPE_SHADER_IR
types.
v3:
- Coding style fixes
- Removed compatibility code for LLVM 3.1
- Split
Tom Stellard tstel...@gmail.com writes:
On Sun, May 27, 2012 at 02:25:05PM +0200, Francisco Jerez wrote:
Tom Stellard tstel...@gmail.com writes:
[...]
+* ``PIPE_COMPUTE_CAP_IR_TARGET``: A description of the target as a target
+ triple specification of the form ``processor-manufacturer-os
Tom Stellard tstel...@gmail.com writes:
v2:
-Separate IR type and LLVM triple
-Do the OpenCL C-LLVM IR and linking steps for all PIPE_SHADER_IR
types.
v3:
- Coding style fixes
- Removed compatibility code for LLVM 3.1
- Split build_module_llvm() into three functions:
Tom Stellard tstel...@gmail.com writes:
From: Francisco Jerez curroje...@riseup.net
v2: Tom Stellard
- Update CAP description
v3: Tom Stellard
- TGSI targets should pass an empty string for this CAP.
---
I think we can let TGSI drivers leave this as an empty string for now
Tom Stellard thomas.stell...@amd.com writes:
On Mon, May 28, 2012 at 10:03:27PM +0200, Francisco Jerez wrote:
Tom Stellard tstel...@gmail.com writes:
v2:
-Separate IR type and LLVM triple
-Do the OpenCL C-LLVM IR and linking steps for all PIPE_SHADER_IR
types.
v3
Blaž Tomažič blaz.toma...@gmail.com writes:
Hi,
I am applying for XOrg EVoC and have written a proposal for working on a
OpenCL testing framework for piglit. I am including project details I've
written in my proposal. I am looking for comments and suggestions.
Name:
OpenCL testing
Tom Stellard tstel...@gmail.com writes:
From: Francisco Jerez curroje...@riseup.net
v2: Tom Stellard
- Update CAP description
v3: Tom Stellard
- TGSI targets should pass an empty string for this CAP.
v4: Tom Stellard
- TGSI targets can ignore this CAP.
---
src/gallium/docs
. */
+};
+
struct pipe_compute_state
{
const void *prog; /** Compute program to be executed. */
Reviewed-by: Francisco Jerez curroje...@riseup.net
pgpmcvIa4S4Q4.pgp
Description: PGP signature
___
mesa-dev mailing list
mesa-dev
();
+ default:
+ return build_module_llvm(mod);
+ }
}
Reviewed-by: Francisco Jerez curroje...@riseup.net
pgp05W4zCqLWe.pgp
Description: PGP signature
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman
Pauli Nieminen pauli.niemi...@linux.intel.com writes:
ARB_sampler_object is very simple software only extension to support.
I want to make it mandator extension for Mesa drivers to allow meta
module to use it.
This patch add support for the extension to nouveau. It is completely
untested
Tom Stellard tstel...@gmail.com writes:
We need to increment the reference count for objects, like cl_event,
that the user is responsible for destroying when they are returned from
the API. Otherwise, the object will be destroyed when clover is done with
it, even though the user will still
Tom Stellard thomas.stell...@amd.com writes:
I'll commit this tomorrow unless there are objections.
Looks OK to me, aside from the std::vector's being passed around by
value instead of by const reference.
-Tom
On Thu, Jun 21, 2012 at 02:05:24PM -0400, Tom Stellard wrote:
The function
Tom Stellard thomas.stell...@amd.com writes:
On Thu, Jun 28, 2012 at 01:09:20AM +0200, Francisco Jerez wrote:
Tom Stellard tstel...@gmail.com writes:
We need to increment the reference count for objects, like cl_event,
that the user is responsible for destroying when they are returned
Tom Stellard tstel...@gmail.com writes:
If devs is NULL, then the kernel should be compiled for all devices
associated with the program's context.
---
src/gallium/state_trackers/clover/api/program.cpp | 16 ++--
1 files changed, 10 insertions(+), 6 deletions(-)
diff --git
- Close the X connection when done
- Split auth code into separate function
Thanks, Reviewed-by: Francisco Jerez curroje...@riseup.net
---
configure.ac |6 ++
.../auxiliary/pipe-loader/pipe_loader_drm.c| 61
2 files
and so
any device not being build should not have this information erased.
---
src/gallium/state_trackers/clover/core/program.cpp | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
Thanks. For both patches:
Reviewed-by: Francisco Jerez curroje...@riseup.net
diff --git a/src/gallium
Kristian Høgsberg k...@bitplanet.net writes:
2010/5/16 Xavier Chantry chantry.xav...@gmail.com:
Patrice reported this a few days ago in nouveau channel and wiki,
using xorg 1.7 and probably nvfx driver.
I can reproduce with xorg 1.7 and nv50 with current git master.
Window resizing (for
randrianas...@gmail.com writes:
Very hackish patch series for nv0x hardware, please take look and comment on
changes, blame me for errors etc.
Francisco, might be nouveau list is better for this stuff?
It's fine either way, do it as you prefer.
From
Ian Romanick i...@freedesktop.org writes:
[...]
At some point it may be worth adding a dri-conf option to enable fake
ARB_texture_env_combine support. There are a few apps that check for
that but not for the EXT. I'd swear that NV drivers used to support the
ARB extension on TNT2
son_of_the_osi...@interia.pl writes:
Hi. I have running linux on my very old machine with nv17 card and I
am impressed that kde4 with nouveau driver work so fast. I am also
curious about few things:
1. In nvfx was implemented sot so long time ago a new 2d engine. It
wil be ported to
This function is executed outside _mesa_meta_begin/end(), that means
that e.g. _mesa_meta_Bitmap() clobbers the texturing state because it
changes the currently active texture object.
There's no need to bind the new texture when it's created, it's done
again later anyway (from
---
src/mesa/drivers/common/meta.c |3 ++-
1 files changed, 2 insertions(+), 1 deletions(-)
diff --git a/src/mesa/drivers/common/meta.c b/src/mesa/drivers/common/meta.c
index c487097..9615b52 100644
--- a/src/mesa/drivers/common/meta.c
+++ b/src/mesa/drivers/common/meta.c
@@ -474,7 +474,8 @@
Sep 17 00:00:00 2001
From: Francisco Jerez curroje...@riseup.net
Date: Fri, 29 Oct 2010 21:20:34 +0200
Subject: [PATCH] meta: Handle bitmaps with alpha test enabled.
---
src/mesa/drivers/common/meta.c | 56 +--
1 files changed, 47 insertions(+), 9 deletions
Brian Paul bri...@vmware.com writes:
[...]
- bitmap8 = (GLubyte *) calloc(1, width * height);
+ bitmap8 = (GLubyte *) malloc(width * height);
if (bitmap8) {
+ memset(bitmap8, ~alpha, width * height);
_mesa_expand_bitmap(width, height, unpackSave, bitmap1,
-
---
src/mesa/drivers/common/meta.c |5 +
1 files changed, 5 insertions(+), 0 deletions(-)
diff --git a/src/mesa/drivers/common/meta.c b/src/mesa/drivers/common/meta.c
index 9615b52..7ad1e72 100644
--- a/src/mesa/drivers/common/meta.c
+++ b/src/mesa/drivers/common/meta.c
@@ -104,6 +104,8
---
src/mesa/drivers/common/meta.c | 45 ++-
1 files changed, 39 insertions(+), 6 deletions(-)
diff --git a/src/mesa/drivers/common/meta.c b/src/mesa/drivers/common/meta.c
index 7ad1e72..f96e931 100644
--- a/src/mesa/drivers/common/meta.c
+++
Brian Paul bri...@vmware.com writes:
On 11/01/2010 08:02 PM, Francisco Jerez wrote:
Brian Paulbri...@vmware.com writes:
[...]
- bitmap8 = (GLubyte *) calloc(1, width * height);
+ bitmap8 = (GLubyte *) malloc(width * height);
if (bitmap8) {
+ memset(bitmap8, ~alpha, width
Brian Paul bri...@vmware.com writes:
On 11/02/2010 12:25 PM, Francisco Jerez wrote:
We have several options.
- Fall back to swrast in the floating point renderbuffer case. In other
cases the conversion to fixed point guarantees that EQUAL will behave
as expected. Simple and good
Brian Paul bri...@vmware.com writes:
+static GLboolean
+alpha_test_bitmap(struct gl_context *ctx)
+{
+ struct gl_colorbuffer_attrib *color = ctx-Color;
+ GLenum func = (color-AlphaEnabled ? color-AlphaFunc : GL_ALWAYS);
+ GLubyte ref = FLOAT_TO_UBYTE(color-AlphaRef);
+ GLubyte
Brian Paul bri...@vmware.com writes:
On 11/03/2010 09:12 AM, Francisco Jerez wrote:
Brian Paulbri...@vmware.com writes:
On 11/02/2010 12:25 PM, Francisco Jerez wrote:
We have several options.
- Fall back to swrast in the floating point renderbuffer case. In other
cases
Francisco Jerez curroje...@riseup.net writes:
Brian Paul bri...@vmware.com writes:
+static GLboolean
+alpha_test_bitmap(struct gl_context *ctx)
+{
+ struct gl_colorbuffer_attrib *color = ctx-Color;
+ GLenum func = (color-AlphaEnabled ? color-AlphaFunc : GL_ALWAYS);
+ GLubyte ref
reason.
OK, here's a patch doing the same with floats.
-Brian
From 4158e267659eeff15ef5a6cbc29c6c23cb682ed3 Mon Sep 17 00:00:00 2001
From: Francisco Jerez curroje...@riseup.net
Date: Tue, 2 Nov 2010 02:49:32 +0100
Subject: [PATCH] meta: Handle bitmaps with alpha test enabled.
---
src/mesa
Brian Paul bri...@vmware.com writes:
On 11/10/2010 05:41 AM, Francisco Jerez wrote:
Brian Paulbri...@vmware.com writes:
Of course, we _could_ do the comparison with ints but we'd have to
query the number of framebuffer alpha bits, convert the floats to ints
with the right number of bits
Jerome Glisse j.gli...@gmail.com writes:
Hi,
[...]
optimized further). Also it seems that we are suffering from call
overhead (likely TLS or others similar optimization in our GL
dispatching stuff), nvidia is a lot better at facing millions of call.
Yeah, the nvidia binary blob is really
Bob Gleitsmann rjgle...@bellsouth.net writes:
Hello everyone,
When trying the demo program copytex for the first time recently, I noticed
pathological behavior: after running for a long time it asserted out and
locked up X. Investigation showed this to be due to the glFinish function
Michel Dänzer mic...@daenzer.net writes:
On Mon, 2011-01-03 at 22:50 +0100, Francisco Jerez wrote:
Bob Gleitsmann rjgle...@bellsouth.net writes:
When trying the demo program copytex for the first time recently, I
noticed
pathological behavior: after running for a long time
nobled nob...@dreamwidth.org writes:
It's a GNU extension that isn't supported by clang right now:
http://gcc.gnu.org/onlinedocs/gcc-4.6.3/gcc/Nested-Functions.html
http://clang.llvm.org/docs/UsersManual.html#c_unimpl_gcc
With this, clang now compiles the nouveau classic driver.
Bugzilla:
programs, patch 2 is a refactoring required by
patch 10.
Comments, questions and r-b's are much appreciated.
[1] http://cgit.freedesktop.org/~tstellar/mesa/
[2] https://github.com/curro/mesa
Francisco Jerez (11):
gallium: Basic compute interface.
gallium/tgsi: Move interpolation info from
Define an interface that exposes the minimal functionality required to
implement some of the popular compute APIs. This commit adds entry
points to set the grid layout and other state required to keep track
of the usual address spaces employed in compute APIs, to bind a
compute program, and
Move Interpolate, Centroid and CylindricalWrap from tgsi_declaration
to a separate token -- they only make sense for FS inputs and we need
room for other flags in the top-level declaration token.
---
src/gallium/auxiliary/draw/draw_pipe_aaline.c |3 +-
---
src/gallium/auxiliary/tgsi/tgsi_scan.c |3 ++-
src/gallium/auxiliary/tgsi/tgsi_strings.c |5 +++--
src/gallium/auxiliary/tgsi/tgsi_strings.h |2 +-
src/gallium/auxiliary/tgsi/tgsi_text.c |2 ++
src/gallium/include/pipe/p_shader_tokens.h |1 +
5 files changed, 9
This texture type was already referred to by the documentation but it
was never defined. Define it as 0 to match the pipe_texture_target
enumeration values.
---
src/gallium/auxiliary/tgsi/tgsi_build.c|4 ++--
src/gallium/auxiliary/tgsi/tgsi_strings.c |2 +-
Normal resource access (e.g. the LOAD TGSI opcode) is supposed to
perform a series of conversions to turn the texture data as it's found
in memory into the data format that was specified in the resource
declaration.
In compute programs it's often the case that we only want to access
the raw bits
Define a new STORE opcode with a role dual to the LOAD opcode, and add
flags to specify that a resource, sampler view, or shader resource
binding point is intended for writing.
---
src/gallium/auxiliary/tgsi/tgsi_build.c|4 +++
src/gallium/auxiliary/tgsi/tgsi_dump.c |2 ++
---
src/gallium/auxiliary/tgsi/tgsi_strings.c |6 +-
src/gallium/include/pipe/p_shader_tokens.h |6 +-
2 files changed, 10 insertions(+), 2 deletions(-)
diff --git a/src/gallium/auxiliary/tgsi/tgsi_strings.c
b/src/gallium/auxiliary/tgsi/tgsi_strings.c
index 9e9c0dc..cad46e8
---
src/gallium/auxiliary/tgsi/tgsi_info.c |4 +++
src/gallium/docs/source/tgsi.rst | 49
src/gallium/include/pipe/p_shader_tokens.h |7 +++-
3 files changed, 59 insertions(+), 1 deletion(-)
diff --git a/src/gallium/auxiliary/tgsi/tgsi_info.c
This change will be useful to implement function parameter passing on
top of TGSI. As we don't have a proper stack, a register-based
calling convention will be used instead, which isn't necessarily a bad
thing given that GPUs often have plenty of registers to spare.
Using the same register space
This is mainly to accomodate AMD's LLVM compiler back-end by letting
it bypass the TGSI representation. Other drivers will keep using the
common TGSI instruction set for compute shaders.
---
src/gallium/docs/source/screen.rst |5 +
src/gallium/include/pipe/p_defines.h |1 +
in patches 10 and 11 was to make the linking process of winsys/loader
a bit less painful.
Patches 2, 8, 15 aren't strictly required by anything else, they're
mainly clean-ups that are made possible by other commits of the
series.
Francisco Jerez (14):
gallium/util: Define util_strchrnul
---
src/gallium/auxiliary/util/u_string.h | 24 ++--
1 file changed, 18 insertions(+), 6 deletions(-)
diff --git a/src/gallium/auxiliary/util/u_string.h
b/src/gallium/auxiliary/util/u_string.h
index cc7992d..ed15981 100644
--- a/src/gallium/auxiliary/util/u_string.h
+++
---
src/gallium/auxiliary/tgsi/tgsi_ureg.c | 43
1 file changed, 16 insertions(+), 27 deletions(-)
diff --git a/src/gallium/auxiliary/tgsi/tgsi_ureg.c
b/src/gallium/auxiliary/tgsi/tgsi_ureg.c
index 75be6cf..082fec3 100644
---
---
src/gallium/auxiliary/tgsi/tgsi_ureg.c | 54
src/gallium/auxiliary/tgsi/tgsi_ureg.h |3 ++
2 files changed, 51 insertions(+), 6 deletions(-)
diff --git a/src/gallium/auxiliary/tgsi/tgsi_ureg.c
b/src/gallium/auxiliary/tgsi/tgsi_ureg.c
index
Local makes more sense in most places because non-inline function
calls are unimplemented anyway.
---
src/mesa/state_tracker/st_glsl_to_tgsi.cpp |8
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
The 'is_branch' TGSI opcode info flag was being interpreted
incorrectly by tgsi_text.c as if it implied that the opcode requires a
label token -- that's not the case on e.g. IF, BGNLOOP or ENDLOOP.
Add a new opcode info field that marks an opcode as requiring a TGSI
label token, and replace the
1 - 100 of 2990 matches
Mail list logo