Intel SNB/IVB platform supports rendering YUYV buffer to overlay plane, however
YUYV is missing support from mesa/dri-image yet.
so I go ahead to add it; basing on it, libva can send YUYV buffer directly to
wayland/weston, then weston output it to overlay plane.
v2 update:
- set
---
include/GL/internal/dri_interface.h |1 +
1 files changed, 1 insertions(+), 0 deletions(-)
mode change 100644 = 100755 include/GL/internal/dri_interface.h
diff --git a/include/GL/internal/dri_interface.h
b/include/GL/internal/dri_interface.h
old mode 100644
new mode 100755
index
---
.../state_trackers/egl/wayland/native_drm.c|3 +++
.../state_trackers/egl/wayland/native_wayland.h|3 ++-
2 files changed, 5 insertions(+), 1 deletions(-)
mode change 100644 = 100755
src/gallium/state_trackers/egl/wayland/native_drm.c
mode change 100644 = 100755
YUYV image works for overlay now
YUYV image works for texture will come soon
---
src/mesa/drivers/dri/intel/intel_screen.c|3 +++
src/mesa/drivers/dri/intel/intel_tex_image.c |6 ++
2 files changed, 9 insertions(+), 0 deletions(-)
mode change 100644 = 100755
---
src/gbm/backends/dri/gbm_dri.c |9 +
src/gbm/main/gbm.h |5 -
2 files changed, 13 insertions(+), 1 deletions(-)
mode change 100644 = 100755 src/gbm/backends/dri/gbm_dri.c
mode change 100644 = 100755 src/gbm/main/gbm.h
diff --git
---
src/egl/drivers/dri2/platform_wayland.c |7 ++-
1 files changed, 6 insertions(+), 1 deletions(-)
mode change 100644 = 100755 src/egl/drivers/dri2/platform_wayland.c
diff --git a/src/egl/drivers/dri2/platform_wayland.c
b/src/egl/drivers/dri2/platform_wayland.c
old mode 100644
new
it can be shared among processed
---
src/gbm/backends/dri/gbm_dri.c | 13 +
src/gbm/main/gbm.c | 18 ++
src/gbm/main/gbm.h |3 +++
src/gbm/main/gbmint.h |2 ++
4 files changed, 36 insertions(+), 0 deletions(-)
mode change
---
src/egl/wayland/wayland-drm/wayland-drm.c |3 +++
1 files changed, 3 insertions(+), 0 deletions(-)
mode change 100644 = 100755 src/egl/wayland/wayland-drm/wayland-drm.c
diff --git a/src/egl/wayland/wayland-drm/wayland-drm.c
b/src/egl/wayland/wayland-drm/wayland-drm.c
old mode 100644
add tiling support for intel_image_write()
all drm image can be written from app
---
src/mesa/drivers/dri/intel/intel_screen.c | 17 +++--
1 files changed, 15 insertions(+), 2 deletions(-)
diff --git a/src/mesa/drivers/dri/intel/intel_screen.c
---
include/EGL/eglmesaext.h|3 +++
src/egl/drivers/dri2/egl_dri2.c |9 -
2 files changed, 11 insertions(+), 1 deletions(-)
mode change 100644 = 100755 include/EGL/eglext.h
mode change 100644 = 100755 include/EGL/eglmesaext.h
diff --git a/include/EGL/eglext.h
---
tests/egl/CMakeLists.gl.txt |2 +
tests/egl/egl-create-drm-image.c | 145 ++
2 files changed, 147 insertions(+), 0 deletions(-)
create mode 100755 tests/egl/egl-create-drm-image.c
diff --git a/tests/egl/CMakeLists.gl.txt
According to GLSL ES spec, section 4.5.2:
If the precision cannot be determined by this method e.g. if an entire
expression is composed only of
operands with no precision qualifier, and the result is not assigned or passed
as an argument, then it
is evaluated at the default precision of the
On 13/06/2012 18:30, Eric Anholt wrote:
Here's a bunch more automake. We're at the point that an enterprising
person could potentially convert libmesa.a if they figure out how to
work that into the gallium build. I also haven't converted osmesa,
because I didn't think I knew of any tests for
From: Laurent Carlier lordhea...@gmail.com
This also currently fix the installation of libOSmesa.
v2: Remove old Makefile, libOSmesa is now versioned, fix typos
v3: Keep config substitution alphabetized
v4: Update .gitignore
Reviewed-by: Jon TURNEY jon.tur...@dronecode.org.uk
---
configure.ac
Signed-off-by: Jon TURNEY jon.tur...@dronecode.org.uk
---
configure.ac|2 +
src/glsl/.gitignore |1 +
src/glsl/Makefile | 144 ---
src/glsl/Makefile.am| 111
Signed-off-by: Jon TURNEY jon.tur...@dronecode.org.uk
---
src/glsl/tests/Makefile.am |2 +-
src/mesa/Makefile.am| 45 +--
src/mesa/Makefile.old | 57 +--
src/mesa/drivers/osmesa/Makefile.am |
---
configs/current.in |1 -
configure.ac |3 ---
2 files changed, 0 insertions(+), 4 deletions(-)
diff --git a/configs/current.in b/configs/current.in
index f4858bd..ca817c4 100644
--- a/configs/current.in
+++ b/configs/current.in
@@ -147,7 +147,6 @@ VG_LIB_DEPS =
Now mesa/drivers/dri is converted to automake, we want to update DRI_LIB_DEPS
so that we link with the libmesa or libdricore libtool library, as appropriate.
However, this is complicated by the fact that gallium/targets is not (yet)
converted, so we can't share the DRI_LIB_DEPS autoconf variable
Automake can't handle having both clip.S and clip.c, even though they have
different paths
src/mesa/Makefile.am: object `clip.lo' created by `$(SRCDIR)/sparc/clip.S' and
`$(SRCDIR)/main/clip.c'
Signed-off-by: Jon TURNEY jon.tur...@dronecode.org.uk
---
src/mesa/SConscript |2 +-
* configure substitutions are not allowed in _SOURCES variables in automake,
so remove the AC_SUBST'ed GLAPI_ASM_SOURCES and instead use some AM_CONDITIONALS
to choose which asm sources are used
* Change GLAPI_LIB to point to the .la file in other Makefile.am files, and
make a link
to the .a
We must either:
(i) Not link libmesa with libglapi, and link anything that uses libmesa with
libglapi as well, or
(ii) Link libmesa with libglapi, and not link anything that uses libmesa with
libglapi
I choose (ii) just because it's least change, but I'm not sure it's right.
Note that
automake doesn't allow hyphens in variable names
Signed-off-by: Jon TURNEY jon.tur...@dronecode.org.uk
---
configure.ac |2 +-
src/mapi/glapi/sources.mak |2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/configure.ac b/configure.ac
index 216ec3b..c7a42cb
Am Sonntag, 8. Juli 2012, 19:41:33 schrieb Kenneth Graunke:
On 07/08/2012 03:12 PM, Johannes Obermayr wrote:
---
configure.ac |1 +
src/mesa/Makefile.am |9 ++---
2 files changed, 7 insertions(+), 3 deletions(-)
It's never been clear to me whether one can build
Jon TURNEY jon.tur...@dronecode.org.uk writes:
We must either:
(i) Not link libmesa with libglapi, and link anything that uses libmesa with
libglapi as well, or
(ii) Link libmesa with libglapi, and not link anything that uses libmesa with
libglapi
I choose (ii) just because it's least
Jon TURNEY jon.tur...@dronecode.org.uk writes:
On 13/06/2012 18:30, Eric Anholt wrote:
Here's a bunch more automake. We're at the point that an enterprising
person could potentially convert libmesa.a if they figure out how to
work that into the gallium build. I also haven't converted
Marek Olšák mar...@gmail.com writes:
This is adds a new driver function to retrieve the timestamp.
Reviewed-by: Eric Anholt e...@anholt.net
pgpiAIdFyMQAr.pgp
Description: PGP signature
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
osmesa didn't install gl.pc in the past as far as I can see, and I don't
think it should -- it's not a normal GL implementation, so you want to
be specifically looking for osmesa in your pkgconfig checks.
pgp9AaCd4EExB.pgp
Description: PGP signature
Jon TURNEY jon.tur...@dronecode.org.uk writes:
diff --git a/src/mesa/Makefile.am b/src/mesa/Makefile.am
index bada760..08beaa6 100644
--- a/src/mesa/Makefile.am
+++ b/src/mesa/Makefile.am
+libmesa_la_SOURCES = \
+ $(MESA_FILES) \
+ $(MESA_CXX_FILES) \
+
Jan Engelhardt jeng...@inai.de writes:
On Monday 2012-07-09 13:41, Jon TURNEY wrote:
Automake can't handle having both clip.S and clip.c, even though they
have different paths
src/mesa/Makefile.am: object `clip.lo' created by `$(SRCDIR)/sparc/clip.S'
and `$(SRCDIR)/main/clip.c'
It should -
This series enables it for all hardware. The implementation simply
decodes the ETC1 data into RGB at time of glTexImage, so no hardware support
is needed. This series lives on my etc1-v8 branch.
Tested against pending Piglit test OES_compressed_ETC1_RGB8_texture/miptree,
which lives on my Piglit
Remove 'extern' from the functions declared in texcompress_etc.h.
Signed-off-by: Chad Versace chad.vers...@linux.intel.com
---
src/mesa/main/texcompress_etc.h | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/src/mesa/main/texcompress_etc.h b/src/mesa/main/texcompress_etc.h
- Add doxygen to etc1_texel_fetch().
- Rename ambiguously named `dst` variables to `bgr888` in order to be
better self-documenting.
Signed-off-by: Chad Versace chad.vers...@linux.intel.com
---
src/mesa/main/texcompress_etc.c | 22 +-
1 file changed, 13 insertions(+), 9
Add function _mesa_etc1_image_decode_to_abgr(). It is intended to be
used by glCompressedTexSubImage2D to translate ETC1 textures into ABGR.
Signed-off-by: Chad Versace chad.vers...@linux.intel.com
---
src/mesa/main/texcompress_etc.c | 82 +
Enable it for all hardware.
No current hardware supports ETC1, so this patch implements it by
translating the ETC1 data to RGBX data during the call to
glCompressedTexImage2D(). For details, see the doxygen for
intel_mipmap_tree::wraps_etc1.
Tested on Sandybridge with Piglit test
Am Montag, 9. Juli 2012, 09:45:56 schrieb Eric Anholt:
osmesa didn't install gl.pc in the past as far as I can see, and I don't
think it should -- it's not a normal GL implementation, so you want to
be specifically looking for osmesa in your pkgconfig checks.
Mesa-demos require gl.pc (line 28)
It didn't change performance on Lightsmark or Nexuiz, which both used
DYNAMIC_DRAW buffers, but it was killing performance (40% CPU wasted pwriting
buffers) on a closed-source app we're looking at.
---
src/mesa/drivers/dri/intel/intel_buffer_objects.c | 13 +
1 file changed, 5
In one workload, 45% of CPU time was spent in pwrite(), primarily
uploading large amounts of VBO data. This was in fact the hottest path
in the application, according to sysprof. Removing the system memory
buffer and just storing VBOs in actual BOs cut the application's CPU
usage by about half,
Previously we only did this pre-Gen6, and used pwrite on Gen6+.
In one workload, this cuts significant amount of overhead.
Signed-off-by: Kenneth Graunke kenn...@whitecape.org
---
src/mesa/drivers/dri/intel/intel_buffer_objects.c | 26 +
1 file changed, 11 insertions(+), 15
On 07/09/2012 11:18 AM, Kenneth Graunke wrote:
In one workload, 45% of CPU time was spent in pwrite(), primarily
uploading large amounts of VBO data. This was in fact the hottest path
in the application, according to sysprof. Removing the system memory
buffer and just storing VBOs in actual
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 07/04/2012 10:52 AM, Eric Anholt wrote:
We've been talking about doing this for a long time -- ignore the piece of
the GLX spec about clients sharing auxiliary buffers across address spaces,
and just allocate our own buffers. Ever since HiZ we
On 07/04/2012 10:52 AM, Eric Anholt wrote:
This means that GLX buffer sharing of these no longer works. On the
other hand, just *look* at this code reduction.
I enjoy seeing all this code die ;) I wish we had decided earlier to break GLX
sharing before the ugly handshake code was written.
Eric rightly pointed out that this code is still needed by i915. So,
NAK on my patch 1/2, and
Reviewed-by: Kenneth Graunke kenn...@whitecape.org
on Eric's patch.
Is it possible to salvage patch 1/2 so that it only applies to post-i915?
___
Tapani,
Is this patch still applicable? It seems that the autoconf work has calmed down,
and now is a good time to fix the Android build.
-Chad
On 06/07/2012 04:11 AM, Tapani Pälli wrote:
Some more of the files are now autogenerated, this caused build breakage,
patch adds generation of these
Hi everyone,
this series has two parts.
Patches 1-12 are Mesa/Gallium improvements. Most importantly, this adds
acceleration of stencil copying (including glBlitFramebuffer) using the shader
stencil export functionality. I am afraid r600g and radeonsi are the only
drivers which can make use
NOTE: This is a candidate for the 8.0 branch.
---
src/mesa/main/formats.c |6 ++
1 file changed, 2 insertions(+), 4 deletions(-)
diff --git a/src/mesa/main/formats.c b/src/mesa/main/formats.c
index ccc0b17..c65c9c2 100644
--- a/src/mesa/main/formats.c
+++ b/src/mesa/main/formats.c
@@
---
src/gallium/auxiliary/util/u_blitter.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/gallium/auxiliary/util/u_blitter.c
b/src/gallium/auxiliary/util/u_blitter.c
index 47d0227..99bde54 100644
--- a/src/gallium/auxiliary/util/u_blitter.c
+++
used for stencil sampler views.
---
src/gallium/auxiliary/util/u_format.h | 29 +
src/mesa/state_tracker/st_cb_drawpixels.c | 23 ++-
2 files changed, 31 insertions(+), 21 deletions(-)
diff --git a/src/gallium/auxiliary/util/u_format.h
This fixes stencil buffer write transfers on r600g.
---
src/gallium/auxiliary/util/u_blitter.c| 95 ++---
src/gallium/auxiliary/util/u_blitter.h|4 +-
src/gallium/auxiliary/util/u_simple_shaders.c | 54 ++
This doesn't seem to be used by anything yet, but better safe than sorry.
---
src/gallium/auxiliary/util/u_blitter.c| 28 ++-
src/gallium/auxiliary/util/u_simple_shaders.c | 46 +
src/gallium/auxiliary/util/u_simple_shaders.h |6
3 files
---
src/gallium/auxiliary/util/u_blitter.c | 34
src/gallium/auxiliary/util/u_inlines.h | 26
2 files changed, 30 insertions(+), 30 deletions(-)
diff --git a/src/gallium/auxiliary/util/u_blitter.c
Because u_blit couldn't sample a 1D, 3D, CUBE and ARRAY texture, we created
a 2D texture holding a copy of one slice of the source texture (even for 1D).
Let's just do it right.
---
src/gallium/auxiliary/util/u_blit.c | 72 +++
1 file changed, 39 insertions(+),
the structure is calloc'd.
---
src/gallium/auxiliary/util/u_blit.c |6 --
1 file changed, 6 deletions(-)
diff --git a/src/gallium/auxiliary/util/u_blit.c
b/src/gallium/auxiliary/util/u_blit.c
index 469f249..6cdfbca 100644
--- a/src/gallium/auxiliary/util/u_blit.c
+++
---
src/gallium/auxiliary/util/u_blit.c | 12
src/mesa/state_tracker/st_cb_blit.c |4 ++--
2 files changed, 10 insertions(+), 6 deletions(-)
diff --git a/src/gallium/auxiliary/util/u_blit.c
b/src/gallium/auxiliary/util/u_blit.c
index 6cdfbca..30f4a93 100644
---
---
src/gallium/auxiliary/postprocess/pp_mlaa.c |2 +-
src/gallium/auxiliary/postprocess/pp_run.c |2 +-
src/gallium/auxiliary/util/u_blit.c | 206 ++-
src/gallium/auxiliary/util/u_blit.h |4 +-
src/mesa/state_tracker/st_cb_blit.c |
to remove some overhead from draw_vbo. This is a derived state.
BTW, I've got no idea how compute interacts with 3D here, but it should
use cb_misc_state, so that 3D and compute don't conflict.
---
src/gallium/drivers/r600/evergreen_compute.c |3 +--
---
src/gallium/drivers/r600/r600_hw_context.c |1 -
src/gallium/drivers/r600/r600_pipe.h |3 ++-
src/gallium/drivers/r600/r600_state.c|4
src/gallium/drivers/r600/r600_state_common.c | 33 +++---
4 files changed, 20 insertions(+), 21
---
src/gallium/drivers/r600/evergreen_hw_context.c |2 --
src/gallium/drivers/r600/evergreen_state.c | 13 ---
src/gallium/drivers/r600/r600_hw_context.c |1 -
src/gallium/drivers/r600/r600_pipe.h|6 ++---
src/gallium/drivers/r600/r600_state.c
this will be useful for in-place DB decompression, otherwise should be harmless
---
src/gallium/drivers/r600/evergreen_state.c | 10 +++---
src/gallium/drivers/r600/evergreend.h |2 ++
src/gallium/drivers/r600/r600_state.c |6 ++
src/gallium/drivers/r600/r600d.h
we can just update the state when decompressing, there's no need to add
additional info into the DSA state
---
src/gallium/drivers/r600/r600_blit.c | 13 +
src/gallium/drivers/r600/r600_pipe.h |3 +--
src/gallium/drivers/r600/r600_state.c|9 ++---
- maintain a mask of which mipmap levels are dirty (instead of one big flag)
- only flush what was requested at a given point and not the whole resource
(most often only one level and one layer has to be flushed)
---
src/gallium/drivers/r600/r600_blit.c | 65 +-
This was missing/broken. There are also minor code cleanups.
---
src/gallium/drivers/r600/r600_blit.c | 29 +-
src/gallium/drivers/r600/r600_pipe.h |2 +-
src/gallium/drivers/r600/r600_state_common.c |2 +-
3 files changed, 17 insertions(+), 16
A flush depth texture is never set as a depth buffer and never flushed.
---
src/gallium/drivers/r600/r600_texture.c |2 --
1 file changed, 2 deletions(-)
diff --git a/src/gallium/drivers/r600/r600_texture.c
b/src/gallium/drivers/r600/r600_texture.c
index c34df2d..a6a83ca 100644
---
This fixes piglit/depth-level-clamp.
---
src/gallium/drivers/r600/r600_texture.c |5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/src/gallium/drivers/r600/r600_texture.c
b/src/gallium/drivers/r600/r600_texture.c
index a6a83ca..d16c252 100644
---
The only case a depth buffer can be set as a color buffer is when flushing.
That wasn't always the case, but now this code isn't required anymore.
---
src/gallium/drivers/r600/r600_blit.c | 18 --
1 file changed, 18 deletions(-)
diff --git
On 07/09/2012 11:44 AM, Chad Versace wrote:
Eric rightly pointed out that this code is still needed by i915. So,
NAK on my patch 1/2, and
Reviewed-by: Kenneth Graunke kenn...@whitecape.org
on Eric's patch.
Is it possible to salvage patch 1/2 so that it only applies to post-i915?
Sure.
Johannes Obermayr johannesoberm...@gmx.de writes:
Am Montag, 9. Juli 2012, 09:45:56 schrieb Eric Anholt:
osmesa didn't install gl.pc in the past as far as I can see, and I don't
think it should -- it's not a normal GL implementation, so you want to
be specifically looking for osmesa in your
Kenneth Graunke kenn...@whitecape.org writes:
Previously we only did this pre-Gen6, and used pwrite on Gen6+.
In one workload, this cuts significant amount of overhead.
Signed-off-by: Kenneth Graunke kenn...@whitecape.org
---
src/mesa/drivers/dri/intel/intel_buffer_objects.c | 26
@@ -192,6 +192,12 @@ struct brw_blorp_blit_prog_key
*/
unsigned tex_samples;
+ /* If tex_samples 0, whether or not the GPU pipeline will be configured
+* to read from it as though it were an interleaved MSAA layout. False if
+* tex_samples == 0.
+*/
+ bool
On 07/06/2012 03:29 PM, Paul Berry wrote:
Previously the DWORD used to control the CMS MSAA layout was just a
pad value, because we didn't use it.
---
src/mesa/drivers/dri/i965/brw_structs.h | 22 --
1 files changed, 20 insertions(+), 2 deletions(-)
The documentation
On 9 July 2012 13:35, Chad Versace chad.vers...@linux.intel.com wrote:
@@ -192,6 +192,12 @@ struct brw_blorp_blit_prog_key
*/
unsigned tex_samples;
+ /* If tex_samples 0, whether or not the GPU pipeline will be
configured
+* to read from it as though it were an
On 9 July 2012 13:36, Chad Versace chad.vers...@linux.intel.com wrote:
On 07/06/2012 03:29 PM, Paul Berry wrote:
To implement Gen7's CMS MSAA layout, we need an extra buffer, the MCS
(Multisample Control Surface) buffer. This patch introduces code for
allocating and deallocating the
On 06/11/2012 04:37 PM, Kenneth Graunke wrote:
Oh, sorry! Lost track of this. It looks like Eric's pushed it now.
Yes, thanks Eric!
However, since then another commit broke the feature. The build
always produces GL no matter the --with-gl-lib-name given.
The GLU and OSMesa options still
Here's the patch series with Erics suggestion and comments incorporated,
and I've also added support for the packed YUYV format in this version.
The changes since last series boils down to:
- Use __DRI_IMAGE_FORMAT_NONE when we create a __DRIimage we don't intend
to texture from, and don't
The additions in version 5 enables creating EGLImages for different planes
of a YUV buffer. createImageFromName is still used to create the containing
__DRIimage, and createSubImage can then be used no that __DRIimage to create
__DRIimages that correspond to the y, u, and v planes
This lets us specify an offset into the bo where the miptree starts,
which will let us set up a texture for a single plane in a planar buffer.
Signed-off-by: Kristian Høgsberg k...@bitplanet.net
---
src/mesa/drivers/dri/i915/i915_texstate.c |2 +-
We use the new miptree offset to pick out the sub-image when we bind
the EGLImage to a texture.
Signed-off-by: Kristian Høgsberg k...@bitplanet.net
---
src/mesa/drivers/dri/intel/intel_regions.h |1 +
src/mesa/drivers/dri/intel/intel_screen.c| 50 --
We're going to extend this to support multi-plane buffers, so pass this
to the driver so it can access the details.
---
src/egl/drivers/dri2/egl_dri2.c| 33 +---
src/egl/wayland/wayland-drm/wayland-drm.c | 19 +++
---
src/egl/drivers/dri2/egl_dri2.c| 11 ++-
.../wayland/wayland-drm/protocol/wayland-drm.xml | 16
src/egl/wayland/wayland-drm/wayland-drm.c | 79 +++-
src/egl/wayland/wayland-drm/wayland-drm.h |5 +-
This lets us specify the plane to create the image for for multiplanar
wl_buffers.
Signed-off-by: Kristian Høgsberg k...@bitplanet.net
---
docs/WL_bind_wayland_display.spec | 19 +-
include/EGL/eglmesaext.h |1 +
src/egl/drivers/dri2/egl_dri2.c | 117
This lets the compositor know about the structure of the wl_buffer, namely
which planes and components are available and in which order.
---
src/egl/drivers/dri2/egl_dri2.c |5 +
1 file changed, 5 insertions(+)
diff --git a/src/egl/drivers/dri2/egl_dri2.c
On 07/09/2012 02:00 PM, Paul Berry wrote:
On 9 July 2012 13:36, Chad Versace chad.vers...@linux.intel.com
mailto:chad.vers...@linux.intel.com wrote:
On 07/06/2012 03:29 PM, Paul Berry wrote:
To implement Gen7's CMS MSAA layout, we need an extra buffer, the MCS
(Multisample
On Mon, Jul 9, 2012 at 11:41 AM, Jon TURNEY jon.tur...@dronecode.org.uk wrote:
From: Laurent Carlier lordhea...@gmail.com
This also currently fix the installation of libOSmesa.
v2: Remove old Makefile, libOSmesa is now versioned, fix typos
v3: Keep config substitution alphabetized
v4:
On Sat, Jul 7, 2012 at 1:40 AM, Kristian Høgsberg hoegsb...@gmail.com wrote:
On Fri, Jul 06, 2012 at 02:10:57PM -0700, Eric Anholt wrote:
Kristian Høgsberg k...@bitplanet.net writes:
The additions in version 5 enables creating EGLImages for different planes
of a YUV buffer.
In these places, we care about any sort of send that hits more than one reg,
not just textures. We don't yet have anything else returning more than one
reg, so there's no change.
---
src/mesa/drivers/dri/i965/brw_fs.cpp | 16
1 file changed, 8 insertions(+), 8 deletions(-)
This was accidentally copy-and-pasted inside.
---
src/mesa/drivers/dri/i965/brw_fs_live_variables.cpp |9 +
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/src/mesa/drivers/dri/i965/brw_fs_live_variables.cpp
b/src/mesa/drivers/dri/i965/brw_fs_live_variables.cpp
index
Previously, if we were spilling the result of a texture call, we would store
all 4 regs, then for each use of one of those regs as the source of an
instruction, we would unspill all 4 regs even though only one was needed.
In an app we're looking at, one shader goes from 2817 instructions to 2179,
---
src/mesa/drivers/dri/i965/brw_fs.cpp | 25 ++--
src/mesa/drivers/dri/i965/brw_fs.h |1 +
.../drivers/dri/i965/brw_fs_copy_propagation.cpp | 16 ++---
src/mesa/drivers/dri/i965/brw_fs_cse.cpp |8 +--
4 files
count is a more useful name, since most of the time we're using it for
looping over the variables.
---
src/mesa/drivers/dri/i965/brw_fs.cpp |9
src/mesa/drivers/dri/i965/brw_fs.h |2 +-
.../drivers/dri/i965/brw_fs_live_variables.cpp |4 ++--
On Sat, 2012-07-07 at 01:48 +0200, Marek Olšák wrote:
On Wed, Jun 27, 2012 at 1:34 AM, Vadim Girlin vadimgir...@gmail.com wrote:
Use r600_resource_texture::flished_depth_texture for GPU access, and
allocate it in the VRAM. For transfers we'll allocate untiled texture in the
GTT and store it
On 07/09/2012 03:40 PM, Eric Anholt wrote:
[snip]
@@ -1729,10 +1731,8 @@ fs_visitor::compute_to_mrf()
* into a compute-to-MRF.
*/
+/* SENDs can only write to GRFs, so no compute-to-MRF. */
if (scan_inst-is_tex()) {
-/* texturing
90 matches
Mail list logo
