Re: [Mesa-dev] [PATCH] r600g: Replace needless flush in texture upload.

2011-10-27 Thread Mathias Fröhlich
Hi, On Saturday, October 22, 2011 13:36:37 Mathias Fröhlich wrote: The attached patch reduces the amount of pipe flushes for r600g. I am not exactly sure if we could skip this flush entirely because of the internal r600g winsys flush logic. But what we can do is the attached patch

Re: [Mesa-dev] [PATCH] r600g: Replace needless flush in texture upload.

2011-10-29 Thread Mathias Fröhlich
Hi, On Thursday, October 27, 2011 18:17:41 Marek Olšák wrote: Looks good to me. Ideally even texture_barrier should not be needed. set_framebuffer_state() and flush() should flush caches automatically. If they don't, there is a bug. I have pushed this now as an improovement above the former

Re: [Mesa-dev] Performance glxSwapBuffers 32 bit vs. 64 bit

2011-11-10 Thread Mathias Fröhlich
Michael, On Thursday, November 10, 2011 18:42:13 Michel Dänzer wrote: On Don, 2011-11-10 at 11:01 +0100, Theiss, Ingo wrote: The function calls of mesa/state_tracker/st_cb_readpixels.c:382 - st_readpixels and mesa/main/pack.c:552 - _mesa_pack_rgba_span_float clearly stands out when

[Mesa-dev] [PATCH 01/10] mesa: Introduce more symbolic VERT_{ATTRIB, BIT}* defines.

2011-11-11 Thread Mathias Fröhlich
Introduce a set of defines for VERT_ATTRIB_* and VERT_BIT_* that will be used in the followup patches. --- src/mesa/main/context.c | 29 ++ src/mesa/main/mtypes.h | 60 -- 2 files changed, 71 insertions(+), 18 deletions(-)

[Mesa-dev] [PATCH 03/10] vbo: Use The VERT_{ATTRIB,BIT} defines.

2011-11-11 Thread Mathias Fröhlich
--- src/mesa/vbo/vbo_context.c| 26 +++--- src/mesa/vbo/vbo_exec.h |4 ++-- src/mesa/vbo/vbo_exec_api.c | 24 src/mesa/vbo/vbo_exec_array.c | 41 + src/mesa/vbo/vbo_exec_draw.c | 22

[Mesa-dev] [PATCH 02/10] mesa: Replace _NEW_ARRAY_* bits with VERT_BIT_*

2011-11-11 Thread Mathias Fröhlich
Consolidate the two distinct set of flags to use VERT_BIT_*. --- src/mesa/main/arrayobj.c |2 +- src/mesa/main/enable.c | 20 ++-- src/mesa/main/mtypes.h | 37 ++--- src/mesa/main/varray.c | 36 ++--

[Mesa-dev] [PATCH 04/10] mesa: Make gl_array_object::_Enabled 64 bits.

2011-11-11 Thread Mathias Fröhlich
--- src/mesa/main/context.c |2 +- src/mesa/main/enable.c |2 +- src/mesa/main/mtypes.h | 54 +++--- src/mesa/main/varray.c |2 +- 4 files changed, 30 insertions(+), 30 deletions(-) diff --git a/src/mesa/main/context.c

[Mesa-dev] [PATCH 05/10] mesa: Make gl_program::InputsRead 64 bits.

2011-11-11 Thread Mathias Fröhlich
--- src/mesa/main/context.c|2 +- src/mesa/main/ff_fragment_shader.cpp |2 +- src/mesa/main/ffvertex_prog.c | 18 +- src/mesa/main/mtypes.h |4 ++-- src/mesa/main/state.c

[Mesa-dev] [PATCH 00/10] Make gl_array_object use an array of gl_client_attribs

2011-11-11 Thread Mathias Fröhlich
Hi all, following a patch series to make gl_array_object use a VERT_ATTRIB_* indexed array of gl_client_array structs. Since we have already 33 client arrays in an array object the VERT_BIT_* and for vertex shader inputs bitmaps are extended to 64 bits. Drivers and shader code is updated to use

[Mesa-dev] [PATCH 07/10] radeon: Make gl_program::InputsRead 64 bits.

2011-11-11 Thread Mathias Fröhlich
--- src/mesa/drivers/dri/r200/r200_vertprog.c | 31 ++-- 1 files changed, 20 insertions(+), 11 deletions(-) diff --git a/src/mesa/drivers/dri/r200/r200_vertprog.c b/src/mesa/drivers/dri/r200/r200_vertprog.c index 697073e..bf87d80 100644 ---

[Mesa-dev] [PATCH 06/10] intel: Make gl_program::InputsRead 64 bits.

2011-11-11 Thread Mathias Fröhlich
--- src/mesa/drivers/dri/i915/i915_fragprog.c |4 ++-- src/mesa/drivers/dri/i965/brw_context.h |2 +- src/mesa/drivers/dri/i965/brw_draw_upload.c |6 +++--- src/mesa/drivers/dri/i965/brw_vs.c |4 ++-- src/mesa/drivers/dri/i965/brw_vs_constval.c |2 +-

[Mesa-dev] [PATCH 09/10] mesa: Make NV generic attributes alias the legacy arrays.

2011-11-11 Thread Mathias Fröhlich
The NV_vertex_program generic attributes should alias the legacy attributes. Simplify aliasing by using the same gl_client_arrays for legacy and NV_vertex_program attributes. The same is already done for the current state values. --- src/mesa/main/enable.c| 10 ++-- src/mesa/main/get.c

[Mesa-dev] [PATCH 10/10] vbo: Clean up recalculate_input_bindings.

2011-11-11 Thread Mathias Fröhlich
Now the gl_array_object's layout matches the one used in recalculate_input_bindings. Make use of this and remove the bind_array_obj function. --- src/mesa/vbo/vbo_exec.h |6 src/mesa/vbo/vbo_exec_array.c | 53 +++- 2 files changed, 15

Re: [Mesa-dev] [PATCH 00/10] Make gl_array_object use an array of gl_client_attribs

2011-11-16 Thread Mathias Fröhlich
Hi, On Thursday, November 17, 2011 00:47:22 Brian Paul wrote: I did a quick read of these patches and they look pretty good. I'd like to go over them again though. Another set of eyes would be good too. I'll try to finish reviewing by the end of the week. Thanks, I have a busy week too,

Re: [Mesa-dev] [PATCH 00/10] Make gl_array_object use an array of gl_client_attribs

2011-11-22 Thread Mathias Fröhlich
Hi, On Thursday, November 17, 2011 00:47:22 Brian Paul wrote: On 11/11/2011 10:07 AM, Mathias Fröhlich wrote: Hi all, following a patch series to make gl_array_object use a VERT_ATTRIB_* indexed array of gl_client_array structs. Since we have already 33 client arrays in an array

[Mesa-dev] [PATCH 0/6] Make gl_array_object use an array of gl_client_attribs V2

2011-11-27 Thread Mathias Fröhlich
. The patchset survives a r600 piglit run on my favourite rv670 development machine with no regressions. Please review. Thanks Mathias Mathias Fröhlich (6): mesa: Introduce more symbolic VERT_{ATTRIB,BIT}* defines. mesa: Replace _NEW_ARRAY_* bits with VERT_BIT_* vbo: Use The VERT_{ATTRIB

[Mesa-dev] [PATCH 1/6] mesa: Introduce more symbolic VERT_{ATTRIB, BIT}* defines.

2011-11-27 Thread Mathias Fröhlich
Introduce a set of defines for VERT_ATTRIB_* and VERT_BIT_* that will be used in the followup patches. Signed-off-by: Mathias Froehlich mathias.froehl...@web.de --- src/mesa/main/mtypes.h | 63 ++- 1 files changed, 45 insertions(+), 18 deletions(-)

[Mesa-dev] [PATCH 2/6] mesa: Replace _NEW_ARRAY_* bits with VERT_BIT_*

2011-11-27 Thread Mathias Fröhlich
Consolidate the two distinct set of flags to use VERT_BIT_*. Signed-off-by: Mathias Froehlich mathias.froehl...@web.de Reviewed-by: Brian Paul bri...@vmware.com --- src/mesa/main/arrayobj.c |2 +- src/mesa/main/enable.c | 20 ++-- src/mesa/main/mtypes.h | 37

[Mesa-dev] [PATCH 3/6] vbo: Use The VERT_{ATTRIB,BIT} defines.

2011-11-27 Thread Mathias Fröhlich
Signed-off-by: Mathias Froehlich mathias.froehl...@web.de Reviewed-by: Brian Paul bri...@vmware.com --- src/mesa/vbo/vbo_context.c| 26 +++--- src/mesa/vbo/vbo_exec.h |4 ++-- src/mesa/vbo/vbo_exec_api.c | 24

[Mesa-dev] [PATCH 4/6] mesa: Make gl_array_object::_Enabled 64 bits.

2011-11-27 Thread Mathias Fröhlich
Signed-off-by: Mathias Froehlich mathias.froehl...@web.de Reviewed-by: Brian Paul bri...@vmware.com --- src/mesa/main/enable.c |2 +- src/mesa/main/mtypes.h | 54 src/mesa/main/varray.c |2 +- 3 files changed, 29 insertions(+), 29

[Mesa-dev] [PATCH 5/6] mesa: Make gl_program::InputsRead 64 bits.

2011-11-27 Thread Mathias Fröhlich
Make gl_program::InputsRead a 64 bits bitfield. Adapt the intel and radeon driver to handle a 64 bits InputsRead value. Signed-off-by: Mathias Froehlich mathias.froehl...@web.de --- src/mesa/drivers/dri/i915/i915_fragprog.c |4 +- src/mesa/drivers/dri/i965/brw_context.h|2 +-

Re: [Mesa-dev] [PATCH 6/6] mesa: Use VERT_ATTRIB_* indexed array in gl_array_object.

2011-11-28 Thread Mathias Fröhlich
Hi, On Sunday, November 27, 2011 22:52:29 Eric Anholt wrote: Not an issue for this patch, but I noticed that there's this _mesa_update_array_object_max_element that does something similar, but doesn't follow this rule and would trash the _MaxElement value. Some day, the body of this

Re: [Mesa-dev] [PATCH] mesa: distinct gl_client_array arrays are gone

2011-11-29 Thread Mathias Fröhlich
Hi, Apologies for braking this! Thanks, looks great. Mathias On Tuesday, November 29, 2011 08:22:58 Chia-I Wu wrote: From: Chia-I Wu o...@lunarg.com Fix build errors since 762c9766c93697af8d7fbaa729aed118789dbe8e. --- src/mesa/main/ffvertex_prog.c |2 +- src/mesa/main/get.c

Re: [Mesa-dev] [PATCH] mesa: distinct gl_client_array arrays are gone

2011-11-30 Thread Mathias Fröhlich
Hi, On Wednesday, November 30, 2011 07:58:08 Jose Fonseca wrote: It looks like there is a regression, https://bugs.freedesktop.org/show_bug.cgi?id=43353 . Can you investigate this? Sure Mathias ___ mesa-dev mailing list

[Mesa-dev] [PATCH 0/2] update tnl module to the recent VERT_ATTRIB* changes.

2011-12-01 Thread Mathias Fröhlich
These two patches should fix: https://bugs.freedesktop.org/show_bug.cgi?id=3D43353 Please review. Mathias Fröhlich (2): mesa: Extend BITSET64_*_RANGE to work on arbitrary ranges. mesa: Renumber the tnl attributes to match VERT_ATTRIB*. src/mesa/main/bitset.h | 21

[Mesa-dev] [PATCH 1/2] mesa: Extend BITSET64_*_RANGE to work on arbitrary ranges.

2011-12-01 Thread Mathias Fröhlich
The BITSET64_{TEST,SET,CLEAR}_RANGE macros only work on ranges wither in the lower 32 or in the upper 32 bits of the bitset. This change extends these macros to work on arbitrary ranges possibly crossing the bitset word boundary. Signed-off-by: Mathias Froehlich mathias.froehl...@web.de ---

[Mesa-dev] [PATCH 2/2] mesa: Renumber the tnl attributes to match VERT_ATTRIB*.

2011-12-01 Thread Mathias Fröhlich
Also renumber the tnl modules vertex attributes to match the renumbered VERT_ATTRIB_* values. This should fix https://bugs.freedesktop.org/show_bug.cgi?id=43353 Signed-off-by: Mathias Froehlich mathias.froehl...@web.de --- src/mesa/tnl/t_context.h | 58

[Mesa-dev] [PATCH] swrast: Fix signed/unsigned problems with negative strides.

2011-12-01 Thread Mathias Fröhlich
Hi, The attached patch fixes a core dump in glReadPixels with traditional swrast. The problem looks pretty much the same like the one the VTK guy Kevin H. Hobbs reports. At least his backtrace is mostly the same. Please review. Thanks Mathias From 92eaecb0f471d6705121ca346b43e3709989847f Mon

[Mesa-dev] [PATCH] mesa: Use VERT_{ATTRIB, BIT}* for ARB input validation.

2011-12-01 Thread Mathias Fröhlich
Hi, Attached is a change to the ARB program parser that should fix https://bugs.freedesktop.org/show_bug.cgi?id=43407 Please review. Thanks Mathias From 4a3d9f06651c5cd51edcb394fa598153e5fea2f3 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Mathias=20Fr=C3=B6hlich?= mathias.froehl...@web.de Date:

Re: [Mesa-dev] [PATCH 1/2] r600g: Make tiling alignment restrictions match the DDX

2011-12-05 Thread Mathias Fröhlich
Hi, On Thursday, December 01, 2011 17:53:45 Simon Farnsworth wrote: It's unclear exactly what the alignment requirements are for Radeon tiled surfaces, but they differ between the DDX and Mesa. Make Mesa match the DDX. Signed-off-by: Simon Farnsworth simon.farnswo...@onelan.co.uk ---

Re: [Mesa-dev] vertex array regression

2011-12-15 Thread Mathias Fröhlich
Brian, On Thursday, December 15, 2011 17:08:57 Brian Paul wrote: There's a regression in vertex array drawing with this commit: commit ed42c2580717527b2005580940fc766d95bb6b0b Author: Mathias Fröhlich mathias.froehl...@web.de Date: Mon Oct 31 16:23:40 2011 +0100 vbo: Use

Re: [Mesa-dev] vertex array regression

2011-12-16 Thread Mathias Fröhlich
Brian, On Thursday, December 15, 2011 22:36:24 you wrote: I found the problem. It's this chunk in vbo_context.c: [...] For fixed function, the point is to simply place the per-vertex material attributes in the generic attribute arrays. There are 12 such material attributes. So there's

[Mesa-dev] [PATCH] mesa: Set _NEW_ARRAY flag on rebind in any case.

2011-12-16 Thread Mathias Fröhlich
Hi, attached is a change that fixes a problem with the isosurf mesa demo. Please review. Mathias From ce63927e24df800ab6c89554ae9b2096cd79122a Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Mathias=20Fr=C3=B6hlich?= mathias.froehl...@web.de Date: Fri, 16 Dec 2011 20:16:52 +0100 Subject: [PATCH]

Re: [Mesa-dev] vertex array regression

2011-12-19 Thread Mathias Fröhlich
Brian, On Monday, December 19, 2011 17:01:07 you wrote: I still don't get what the purpose of the special mapping of the last four/five elements is all about. Can you explain that? As far as I can see the last ones should not map to anything. The map is used only in fixed function mode. And

Re: [Mesa-dev] vertex array regression

2011-12-19 Thread Mathias Fröhlich
Marek, On Tuesday, December 20, 2011 01:40:46 Marek Olšák wrote: The problem back in March and earlier was that the vertex array state was completely recomputed for each draw call even if the user did not change any state at all! There was the validation in the vbo module and then in

Re: [Mesa-dev] [PATCH] vbo: signal _NEW_ARRAY when transitioning between glBegin/End, glDrawArrays

2011-12-21 Thread Mathias Fröhlich
Hi, On Wednesday, December 21, 2011 01:58:08 Brian Paul wrote: This fixes a regression seen with the isosurf demo when switching between glBegin/End and glDrawArrays (do it several times). The problem was the driver wasn't getting _NEW_ARRAY when the arrays were subtly changed: (vertex3f,

Re: [Mesa-dev] [PATCH] vbo: signal _NEW_ARRAY when transitioning between glBegin/End, glDrawArrays

2011-12-22 Thread Mathias Fröhlich
Hi, On Thursday, December 22, 2011 18:30:44 Brian Paul wrote: I'm not sure if playback_vertex_list is more like DRAW_BEGIN_END or DRAW_ARRAYS. Maybe add a DRAW_DISPLAY_LIST enum value? It's more like begin/end I think. The begin/end code just sets the array state below the state tracking of

Re: [Mesa-dev] [PATCH] vbo: signal _NEW_ARRAY when transitioning between glBegin/End, glDrawArrays

2011-12-25 Thread Mathias Fröhlich
Hi Brian, On Saturday, December 24, 2011 16:38:09 Brian Paul wrote: From: Brian Paul bri...@vmware.com This fixes a regression seen with the isosurf demo when switching between glBegin/End and glDrawArrays (do it several times). The problem was the driver wasn't getting _NEW_ARRAY when

[Mesa-dev] [PATCH] vbo: Clean up recalculate_input_bindings.

2011-12-25 Thread Mathias Fröhlich
Hi, I have posted a very similar patch already with the array object change series. But I have omitted applying this one because of a general issue that Brian pointed me to in a patch prior to this one in the series. But it turns out that this one should be still safe. Please review Thanks

Re: [Mesa-dev] vertex array regression

2011-12-25 Thread Mathias Fröhlich
Hi, On Sunday, December 25, 2011 10:03:34 Marek Olšák wrote: I did not fix Torcs. I think optimizing apps is useless if the same optimization can be done in Mesa too. Ok, so any torcs version should show a performance difference to check against. Do you have any test program that could be

[Mesa-dev] [PATCH 0/5] Replace RENDERINPUTS with direct GLbitfield64 usage.

2011-12-26 Thread Mathias Fröhlich
reviews and tests are very welcome. Please review. Thanks Mathias Mathias Fröhlich (5): mesa: Convert RENDERINPUTS* macros to GLbitfield64. i915: Convert to use GLbitfield64 directly. nouveau: Convert to use GLbitfield64 directly. radeon: Convert to use GLbitfield64 directly. mesa

[Mesa-dev] [PATCH 2/5] i915: Convert to use GLbitfield64 directly.

2011-12-26 Thread Mathias Fröhlich
Signed-off-by: Mathias Froehlich mathias.froehl...@web.de --- src/mesa/drivers/dri/i915/i830_context.h |2 +- src/mesa/drivers/dri/i915/i830_vtbl.c| 24 +++- 2 files changed, 12 insertions(+), 14 deletions(-) diff --git a/src/mesa/drivers/dri/i915/i830_context.h

[Mesa-dev] [PATCH 1/5] mesa: Convert RENDERINPUTS* macros to GLbitfield64.

2011-12-26 Thread Mathias Fröhlich
Signed-off-by: Mathias Froehlich mathias.froehl...@web.de --- src/mesa/main/mtypes.h |6 ++ src/mesa/tnl/t_context.h | 25 + 2 files changed, 19 insertions(+), 12 deletions(-) diff --git a/src/mesa/main/mtypes.h b/src/mesa/main/mtypes.h index 02452aa..28cf6d7

[Mesa-dev] [PATCH 3/5] nouveau: Convert to use GLbitfield64 directly.

2011-12-26 Thread Mathias Fröhlich
Signed-off-by: Mathias Froehlich mathias.froehl...@web.de --- src/mesa/drivers/dri/nouveau/nouveau_swtnl_t.c |2 +- src/mesa/drivers/dri/nouveau/nv04_render.c |2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/src/mesa/drivers/dri/nouveau/nouveau_swtnl_t.c

[Mesa-dev] [PATCH 5/5] mesa: Convert to use GLbitfield64 directly.

2011-12-26 Thread Mathias Fröhlich
Signed-off-by: Mathias Froehlich mathias.froehl...@web.de --- src/mesa/swrast_setup/ss_context.c | 30 +- src/mesa/swrast_setup/ss_context.h |2 +- src/mesa/tnl/t_context.c | 20 +--- src/mesa/tnl/t_context.h | 14

Re: [Mesa-dev] [PATCH] vbo: Clean up recalculate_input_bindings.

2011-12-26 Thread Mathias Fröhlich
Hi, On Tuesday, December 27, 2011 00:17:09 Brian Paul wrote: LGTM. I presume you've done a piglit run and tested with a few other things? I have run piglit quick with classic swrast an piglit r600g on gallium rv670 with this change. Also I am pretty sure that I have started osgviewer and

Re: [Mesa-dev] [PATCH 0/5] Replace RENDERINPUTS with direct GLbitfield64 usage.

2011-12-27 Thread Mathias Fröhlich
Hi, On Tuesday, December 27, 2011 00:37:13 Brian Paul wrote: For the series: Reviewed-by: Brian Paul bri...@vmare.com Thanks. What is your preference to do with the remaining bitset.h header? * Remove this as ununsed? * Remove the 64 bit bitset implementation as this could be better done with

Re: [Mesa-dev] [PATCH 0/5] Replace RENDERINPUTS with direct GLbitfield64 usage.

2011-12-28 Thread Mathias Fröhlich
Hi, On Tuesday, December 27, 2011 09:47:41 Brian Paul wrote: If bitset.h is no longer used let's just remove it. pushed. Thanks Mathias ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev

Re: [Mesa-dev] [PATCH 0/5] Replace RENDERINPUTS with direct GLbitfield64 usage.

2011-12-28 Thread Mathias Fröhlich
Hi, On Wednesday, December 28, 2011 03:22:31 Jose Fonseca wrote: It looks like bitset.h is still used by classic nouveau driver -- see `git grep '\BITSET_'`-- and the state stored in it is too big to fit in 64bit integers (it requires approximately 87 bits). There doesn't seem to be an

Re: [Mesa-dev] [PATCH] r600g: Manage fences per screen rather than per context.

2011-12-29 Thread Mathias Fröhlich
Hi, On Thursday, December 29, 2011 13:35:19 Michel Dänzer wrote: From: Michel Dänzer michel.daen...@amd.com Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=44151 Probably fixes: https://bugs.freedesktop.org/show_bug.cgi?id=44007

Re: [Mesa-dev] [PATCH] r600g: Manage fences per screen rather than per context.

2011-12-29 Thread Mathias Fröhlich
: Add pipe_mutex_destroy() call for newly added mutex. Reviewed-by: Mathias Fröhlich mathias.froehl...@web.de Greetings Mathias ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev

[Mesa-dev] [PATCH 0/3] Array object cleanup.

2011-12-29 Thread Mathias Fröhlich
a piglit quick run for r600g/rv770 and swrast without regressions. Please review. Thanks Mathias Mathias Fröhlich (3): mesa: Use BITFIELD64_RANGE for VERT_BIT_*_ALL. mesa: Introduce enabled bitfield helper functions. mesa: Clean up gl_array_object::_MaxElement computation. src/mesa/main

[Mesa-dev] [PATCH 1/3] mesa: Use BITFIELD64_RANGE for VERT_BIT_*_ALL.

2011-12-29 Thread Mathias Fröhlich
Signed-off-by: Mathias Fröhlich mathias.froehl...@web.de --- src/mesa/main/mtypes.h | 10 +- 1 files changed, 5 insertions(+), 5 deletions(-) diff --git a/src/mesa/main/mtypes.h b/src/mesa/main/mtypes.h index 107371e..451d442 100644 --- a/src/mesa/main/mtypes.h +++ b/src/mesa/main

[Mesa-dev] [PATCH 3/3] mesa: Clean up gl_array_object::_MaxElement computation.

2011-12-29 Thread Mathias Fröhlich
Use a bitmask approach to compute gl_array_object::_MaxElement. To make this work reliably depending on the shader type actually used, make use of the newly introduced typed bitmaks getters. With this change I gain about 5% draw time on some osgviewer examples. Signed-off-by: Mathias Fröhlich

Re: [Mesa-dev] [PATCH 3/3] mesa: Clean up gl_array_object::_MaxElement computation.

2012-01-11 Thread Mathias Fröhlich
Hi Eric, Thanks for the review. I will send shortly an updated patch. ... except: On Friday, December 30, 2011 12:13:36 Eric Anholt wrote: Also, I think that helper would be easier to read as; + GLbitfield64 enabled = arrayObj-_Enabled; + + if (enabled VERT_ATTRIB_GENERIC0) +

[Mesa-dev] [PATCH 0/3] Array object Improvement v2

2012-01-11 Thread Mathias Fröhlich
Hi, Thanks for the first review. Following a short followup series to the gl_array_object change. The most important change is the use ot an ffs based loop to compute gl_array_object::_MaxElement. This change provides a noticable performance improovement for my average workloads. To implement

[Mesa-dev] [PATCH 1/3] mesa: Use BITFIELD64_RANGE for VERT_BIT_*_ALL.

2012-01-11 Thread Mathias Fröhlich
Signed-off-by: Mathias Fröhlich mathias.froehl...@web.de --- src/mesa/main/mtypes.h | 10 +- 1 files changed, 5 insertions(+), 5 deletions(-) diff --git a/src/mesa/main/mtypes.h b/src/mesa/main/mtypes.h index 9fdabf9..b381ad2 100644 --- a/src/mesa/main/mtypes.h +++ b/src/mesa/main

[Mesa-dev] [PATCH 3/3] mesa: Fix and speedup gl_array_object::_MaxElement computation.

2012-01-11 Thread Mathias Fröhlich
Use a bitmask approach to compute gl_array_object::_MaxElement. To make this work correctly depending on the shader type actually used, make use of the newly introduced typed bitmask getters. With this change I gain about 5% draw time on some osgviewer examples. Signed-off-by: Mathias Fröhlich

[Mesa-dev] [PATCH] st/wgl: Release stw_framebuffer::mutex after pbuffer creation

2011-10-13 Thread Mathias Fröhlich
Hi, We have experienced a problem with the win32 wgl pbuffer implementation using gallium together with threading. This attached change releases the stw_framebuffer::mutex past creation of the pbuffer stw_framebuffer. Without this change the pbuffers lock is never released. Since on win32

[Mesa-dev] [PATCH 1/2] r600g: Use the bitfield define matching the register it is used for.

2011-10-13 Thread Mathias Fröhlich
Hi, The attached change just makes the bitfield macro match the register number it is used for. Both actually expand into the same values. Please apply Thanks! Mathias From eaf61856ba5e40700a4b562fc38750f66cf7cba1 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Mathias=20Fr=C3=B6hlich?=

[Mesa-dev] [PATCH 2/2] r600g: restore the old r600g winsys memory characteristics.

2011-10-13 Thread Mathias Fröhlich
Hi, This patch restores the previous memory characteristic of the recently removed r600g winsys. Static and immutable buffers were placed in any case in vram. This really improoves rendering performance for some OpenSceneGraph based applications I often use for testing here. Please review and

Re: [Mesa-dev] [PATCH] st/wgl: Release stw_framebuffer::mutex after pbuffer creation

2011-10-13 Thread Mathias Fröhlich
Hi Jose, On Thursday, October 13, 2011 20:18:25 Jose Fonseca wrote: We should also acquire/release the mutex on every entry point on that file for consistency, but it would make no difference, given that pbuffers can't be resized like normal windows are. Do you have commit access or should

Re: [Mesa-dev] [PATCH] st/wgl: Release stw_framebuffer::mutex after pbuffer creation

2011-10-14 Thread Mathias Fröhlich
Hi, On Thursday, October 13, 2011 21:35:05 Alex Deucher wrote: Follow the instructions here: http://www.freedesktop.org/wiki/AccountRequests Will do, thanks! Mathias ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org

[Mesa-dev] [PATCH] mesa: Fix multithreaded buffer object refcounting.

2011-10-17 Thread Mathias Fröhlich
the buffer objects mutex while modifying reference counts for push/pop of client array attributes. Also ensure that each taken reference on push is released on pop. Signed-off-by: Mathias Fröhlich mathias.froehl...@web.de --- src/mesa/main/attrib.c | 108

Re: [Mesa-dev] [PATCH] mesa: Fix multithreaded buffer object refcounting.

2011-10-19 Thread Mathias Fröhlich
Hi Brian, On Tuesday, October 18, 2011 17:05:53 Brian Paul wrote: This was brought up a few weeks ago by someone else. I had posted the start of a patch to fix it. I'm attaching an updated version that I think is pretty clean. But there's a few lines missing and it's totally untested.

Re: [Mesa-dev] [PATCH] mesa: Fix multithreaded buffer object refcounting.

2011-10-22 Thread Mathias Fröhlich
Hi, On Wednesday, October 19, 2011 16:32:57 Brian Paul wrote: As a first cut, I want something that works correctly and is easy to understand and maintain. We can optimize later if needed. I think most seasoned OpenGL programmers know that glPush/PopAttrib isn't exactly a high-performance

[Mesa-dev] [PATCH 1/2] mesa: Avoid ABA problem on buffer object bind.

2011-10-22 Thread Mathias Fröhlich
name that was just deleted and never rebound in between. The explicit rebinding to the debault object in the current context prevents the above in the current context, but an other context sharing the same objects might suffer from this problem. Signed-off-by: Mathias Fröhlich mathias.froehl

[Mesa-dev] [PATCH 2/2] mesa: Fix multithreaded buffer object refcounting. v2

2011-10-22 Thread Mathias Fröhlich
the array attrib push/pop implementation to be thread safe. Make use of more library functions for this purpose. Signed-off-by: Mathias Fröhlich mathias.froehl...@web.de --- src/mesa/main/attrib.c | 223 1 files changed, 151 insertions(+), 72

[Mesa-dev] [PATCH] r600g: Replace needless flush in texture upload.

2011-10-22 Thread Mathias Fröhlich
Hi, The attached patch reduces the amount of pipe flushes for r600g. I am not exactly sure if we could skip this flush entirely because of the internal r600g winsys flush logic. But what we can do is the attached patch: Replace pipe-flush() with pipe-texture_barrier() in the texture upload

[Mesa-dev] [PATCH] dri-r600: Hook up a drm_descriptor configuration function

2011-10-22 Thread Mathias Fröhlich
Hi, the attached patch enables opengl client side throtteling for r600g that recently slipped in as an infrastructure but is just enabled for vmgfx so far. The benefits of the client side throtteling are great, since a single render intensive application is no longer able to just saturate the

Re: [Mesa-dev] [PATCH 1/2] mesa: Avoid ABA problem on buffer object bind.

2011-10-23 Thread Mathias Fröhlich
Hi, On Sunday, October 23, 2011 17:53:53 Brian Paul wrote: Looks good. Just minor things: In other places where we have a field like this it's called DeletePending. The comment you added in _mesa_DeleteBuffersARB() is a bit hard to understand. If you need help, I could rewrite it for

Re: [Mesa-dev] [PATCH 2/2] mesa: Fix multithreaded buffer object refcounting. v2

2011-10-23 Thread Mathias Fröhlich
Hi, On Sunday, October 23, 2011 17:55:52 Brian Paul wrote: Reviewed-by: Brian Paul bri...@vmware.com I think it would be really good to have even a simple, single-threaded piglit test for this (glPush, glDeleteBufferObj, glPop). Could you write one? Pushing this on top of the todo list

Re: [Mesa-dev] [Bug 59877] Build fail since r600g: Don't build llvm_wrapper.cpp when we aren't using LLVM

2013-01-26 Thread Mathias Fröhlich
On Saturday, January 26, 2013 08:42:41 bugzilla-dae...@freedesktop.org wrote: https://bugs.freedesktop.org/show_bug.cgi?id=59877 Tom Stellard tstel...@gmail.com changed: What|Removed |Added

Re: [Mesa-dev] r600g: status of my work on the shader optimization

2013-02-15 Thread Mathias Fröhlich
Hi, On Friday, February 15, 2013 15:00:24 Vadim Girlin wrote: LLVM backend is the future is a pretty abstract argument. I prefer to operate with real facts. After a year of LLVM backend development what are the real benefits for the users? What are the real use cases where the users might

Re: [Mesa-dev] [PATCH] radeon/llvm: Use LLVM C API for compiling LLVM IR to ISA.

2013-04-17 Thread Mathias Fröhlich
Tom, -class LLVMEnsureMultithreaded { -public: - LLVMEnsureMultithreaded() - { - llvm_start_multithreaded(); - } -}; - -static LLVMEnsureMultithreaded lLVMEnsureMultithreaded; Removing this leads to crashes in llvm with applications that concurrently work on different gl

Re: [Mesa-dev] [PATCH] radeon/llvm: Use LLVM C API for compiling LLVM IR to ISA.

2013-04-20 Thread Mathias Fröhlich
it works stable if we ensure llvm being multithreaded. So, I tried to distill a piglit testcase out of this somehow huger setup with flightgear, OpenSceneGraph, multiple gpu's and what not. On Friday, April 19, 2013 20:08:54 Tom Stellard wrote: On Wed, Apr 17, 2013 at 07:54:32AM +0200, Mathias

Re: [Mesa-dev] [PATCH] radeon/llvm: Use LLVM C API for compiling LLVM IR to ISA.

2013-04-22 Thread Mathias Fröhlich
Hi all, On Monday, April 22, 2013 00:39:57 Tom Stellard wrote: [...] The only pro for further investigating the dlopen flags is that I fear the distribution builders who invented dynamic linking in the drivers. That change destroyed symbol isolation in the drivers at that point. They will

Re: [Mesa-dev] [PATCH] radeon/llvm: Use LLVM C API for compiling LLVM IR to ISA.

2013-04-24 Thread Mathias Fröhlich
/tests/general/driver-isolation-library.c @@ -0,0 +1,36 @@ +/* + * Copyright (c) 2013 Mathias Fröhlich. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the Software), + * to deal in the Software without

Re: [Mesa-dev] [PATCH] radeon/llvm: Use LLVM C API for compiling LLVM IR to ISA.

2013-04-25 Thread Mathias Fröhlich
Hi, On Wednesday, April 24, 2013 14:15:06 Tom Stellard wrote: I've thought about this some more, and I think that the best solution might be to move all LLVM API calls into gallivm and build it as a shared object with it's own private copy of LLVM statically linked. This way we would still

Re: [Mesa-dev] [PATCH] radeon/llvm: Use LLVM C API for compiling LLVM IR to ISA.

2013-04-25 Thread Mathias Fröhlich
Hi, On Wednesday, April 24, 2013 21:54:02 Jose Fonseca wrote: I don't see how this would work -- llvmpipe/draw has LLVMBuildXxxx calls too. So to prevent symbol collision with apps that use them, we'd need to expose all LLVM calls we need under nome unique prefix. Also note that gallivm

Re: [Mesa-dev] [PATCH] radeon/llvm: Use LLVM C API for compiling LLVM IR to ISA.

2013-04-25 Thread Mathias Fröhlich
Jose, On Thursday, April 25, 2013 01:38:46 Jose Fonseca wrote: What I'm suggesting doesn't require huge effort. In detail, I'm suggesting: (1) have a custom build of LLVM libraries with -fvisibility=hidden (2) have a src/mesallvm/mesallvm.c containing wrappers #include

Re: [Mesa-dev] [PATCH] radeon/llvm: Use LLVM C API for compiling LLVM IR to ISA.

2013-04-27 Thread Mathias Fröhlich
Jose, On Thursday, April 25, 2013 03:52:44 Jose Fonseca wrote: There is no paradox. To be clear: Ok, thanks! Mathias ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev

Re: [Mesa-dev] [PATCH] radeon/llvm: Use LLVM C API for compiling LLVM IR to ISA.

2013-04-27 Thread Mathias Fröhlich
Hi, On Thursday, April 25, 2013 10:29:27 Jose Fonseca wrote: - There are a bunch of options that need to be set via globals, (see lp_set_target_options), so app/drivers could tamper with each other options. - llvm::cl::ParseCommandLineOptions will complain if called multiple times -- I

Re: [Mesa-dev] [PATCH] radeon/llvm: Use LLVM C API for compiling LLVM IR to ISA.

2013-05-01 Thread Mathias Fröhlich
Tom, Jose, On Tuesday, April 30, 2013 16:56:56 Tom Stellard wrote: I took the linker script from your email and took at shot at creating libMesaLLVM.so within Mesa. I've pushed my initial code here: http://cgit.freedesktop.org/~tstellar/mesa/log/?h=libmesallvm Thank you very much and sorry

Re: [Mesa-dev] [PATCH] radeon/llvm: Use LLVM C API for compiling LLVM IR to ISA.

2013-05-05 Thread Mathias Fröhlich
Hi, Sorry for these late response. I am in the middle of moving to a new appartements coordinating work in the new one and so on. It might even be that I do have no good network access for a few days ... On Wednesday, May 01, 2013 17:56:33 Tom Stellard wrote: Thanks for pointing this out,

Re: [Mesa-dev] [PATCH 1/3] gallivm: Move LLVMStartMultithreaded() static initializer into gallivm

2013-05-05 Thread Mathias Fröhlich
Tom, The series Reviewed-by: mathias.froehl...@web.de Greetings Mathias ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev

Re: [Mesa-dev] No luck with tstellar/llvm on HD4890

2012-11-06 Thread Mathias Fröhlich
Hi, On Monday, November 05, 2012 20:38:24 Tom Stellard wrote: I think you will have better luck narrowing down the issue if you bisect Mesa, rather than LLVM. I see something that looks for the first cut at least quiet similar. Reverting commit 84b437213294ff4e1a3bcae2f9cbb36a9b4955c4

Re: [Mesa-dev] No luck with tstellar/llvm on HD4890

2012-11-06 Thread Mathias Fröhlich
Hi Vincent, On Tuesday, November 06, 2012 12:14:52 Vincent Lejeune wrote: Do you use the master branch ? Yes, e3213f01f7764af573ed641a7bc98dde5824e321 is my current head in llvm. This commit generates some llvm.R600.store.pixel.color/stencil/depth intrinsics, which were introduced by this

Re: [Mesa-dev] No luck with tstellar/llvm on HD4890

2012-11-07 Thread Mathias Fröhlich
Vincent, On Wednesday, November 07, 2012 08:32:03 Vincent Lejeune wrote: git seems to have trouble sending my patch to ML atm, can you manually apply it ? It fixes lock up here diff --git a/src/gallium/drivers/r600/r600_asm.c b/src/gallium/drivers/r600/r600_asm.c index 5f2548e..f06af44

Re: [Mesa-dev] [PATCH 0/2] First attempt at asynchronous SwapBuffers

2012-11-15 Thread Mathias Fröhlich
Marek, Regarding swap handling, there was a lengthy discussion about this topic at XDC in Nürnberg. May be Keith can comment on that as he also presented these ideas. But short, the overall aim of this discusstion was to make the dri2 buffer swap more asynchronous. This does not exactly solve

Re: [Mesa-dev] [PATCH 0/2] First attempt at asynchronous SwapBuffers

2012-11-17 Thread Mathias Fröhlich
Hi, On Thursday, November 15, 2012 19:04:51 Marek Olšák wrote: Yes, that's what I am looking for, though I'd like to have a solution that works with current DRI2 as well. The triple flush is unrelated to DRI2, it's actually an internal issue with libGL and can be fixed easily. So may be

Re: [Mesa-dev] thread safe llvmpipe

2013-01-19 Thread Mathias Fröhlich
Jose, On Saturday, January 12, 2013 17:22:03 Jose Fonseca wrote: That work has been merged (w/ LLVM 3.1 and MC-JIT support), however it's quite likely that's there are still race conditions, as nobody has been pursuing llvmpipe thread-safety so far. Given the findings that lead to commit

[Mesa-dev] [PATCH] Add EXT_timer_query to the mesa state tracker and softpipe

2010-05-16 Thread Mathias Fröhlich
Hi, The attached patch adds EXT_timer_query support to the mesa state tracker. It does so by adding a new pipe capability for the timer queries and a new type for the query objects. Also a softpipe driver implementation for this capability is included in this patch. Please review and may be

Re: [Mesa-dev] [PATCH] Add EXT_timer_query to the mesa state tracker and softpipe

2010-05-17 Thread Mathias Fröhlich
Hi, On Monday 17 May 2010 20:51:09 Corbin Simpson wrote: I'm going to be proactive here, and pull in both this patch and a docs update. Great! Thanks! anyway a small followup on the time units is attached. Please apply. Mathias diff --git a/src/gallium/docs/source/context.rst

Re: [Mesa-dev] [PATCH] Add EXT_timer_query to the mesa state tracker and softpipe

2010-05-18 Thread Mathias Fröhlich
Hi, On Monday 17 May 2010 21:02:00 Brian Paul wrote: It would be nice to have a piglit test to exercise EXT_timer_query too. We should really have tests for all extensions. Not yet there. I was using osgviewer to test. I will build something that shows that it runs ... and how long. May be

Re: [Mesa-dev] [PATCH] Add EXT_timer_query to the mesa state tracker and softpipe

2010-05-22 Thread Mathias Fröhlich
) add_executable (vbo-map-remap vbo-map-remap.c) /* * Copyright © 2009 Intel Corporation * Copyright © 2010 Mathias Fröhlich * * Permission is hereby granted, free of charge, to any person obtaining a * copy of this software and associated documentation files (the Software), * to deal

Re: [Mesa-dev] Enable HW GenerateMipmap for radeon

2010-05-23 Thread Mathias Fröhlich
Hi, On Tuesday 18 May 2010 21:09:56 Will Dyson wrote: Wire up _mesa_meta_GenerateMipmap() for radeon. This greatly improves the playability of some games, such as FlightGear and the recently GPL'd game Lugaru. Great thanks! So, turning the view most probably no longer hangs due to generating

Re: [Mesa-dev] [PATCH] Add EXT_timer_query to the mesa state tracker and softpipe

2010-05-23 Thread Mathias Fröhlich
Hi, On Monday 17 May 2010 20:51:09 Corbin Simpson wrote: I'm going to be proactive here, and pull in both this patch and a docs update. Ok. Now that the infrastructure is there. My initial aim was to have something to profile the r300g driver. It already runs nicer then the classic one for

Re: [Mesa-dev] [PATCH] Add EXT_timer_query to the mesa state tracker and softpipe

2010-05-24 Thread Mathias Fröhlich
Hi, On Monday 24 May 2010 23:08:41 Brian Paul wrote: I hacked on the code a bit: 1. remove dead/unused code 2. Test the 64-bit query functions 3. fix the GL_QUERY_COUNTER_BITS check 4. fix indentation Updated version attached. I'll commit in a day or so if there's no feedback. Looks

Re: [Mesa-dev] [PATCH] Add EXT_timer_query to the mesa state tracker and softpipe

2010-05-24 Thread Mathias Fröhlich
Hi, On Sunday 23 May 2010 18:15:09 you wrote: You might try latest xf86-video-ati from git, there's ColorTiling enabled by default, which should give you nice performance increase in all 3D apps. As for timer queries, I think there's quite a bit of work to do in the kernel too. However, we

  1   2   3   4   >