mesa/src/gallium/drivers/radeonsi/r600_blit.c

630 lines
21 KiB
C
Raw Normal View History

radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
/*
* Copyright 2010 Jerome Glisse <glisse@freedesktop.org>
*
* 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 restriction, including without limitation
* on the rights to use, copy, modify, merge, publish, distribute, sub
* license, and/or sell copies of the Software, and to permit persons to whom
* the Software is furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice (including the next
* paragraph) shall be included in all copies or substantial portions of the
* Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
* THE AUTHOR(S) AND/OR THEIR SUPPLIERS BE LIABLE FOR ANY CLAIM,
* DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
* OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
* USE OR OTHER DEALINGS IN THE SOFTWARE.
*/
#include "util/u_surface.h"
#include "util/u_blitter.h"
#include "util/u_format.h"
#include "radeonsi_pipe.h"
#include "si_state.h"
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
enum r600_blitter_op /* bitmask */
{
R600_SAVE_TEXTURES = 1,
R600_SAVE_FRAMEBUFFER = 2,
R600_DISABLE_RENDER_COND = 4,
R600_CLEAR = 0,
R600_CLEAR_SURFACE = R600_SAVE_FRAMEBUFFER,
R600_COPY = R600_SAVE_FRAMEBUFFER | R600_SAVE_TEXTURES |
R600_DISABLE_RENDER_COND,
R600_BLIT = R600_SAVE_FRAMEBUFFER | R600_SAVE_TEXTURES |
R600_DISABLE_RENDER_COND,
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
R600_DECOMPRESS = R600_SAVE_FRAMEBUFFER | R600_DISABLE_RENDER_COND,
R600_COLOR_RESOLVE = R600_SAVE_FRAMEBUFFER | R600_DISABLE_RENDER_COND
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
};
static void r600_blitter_begin(struct pipe_context *ctx, enum r600_blitter_op op)
{
struct r600_context *rctx = (struct r600_context *)ctx;
r600_context_queries_suspend(rctx);
util_blitter_save_blend(rctx->blitter, rctx->queued.named.blend);
util_blitter_save_depth_stencil_alpha(rctx->blitter, rctx->queued.named.dsa);
util_blitter_save_stencil_ref(rctx->blitter, &rctx->stencil_ref);
util_blitter_save_rasterizer(rctx->blitter, rctx->queued.named.rasterizer);
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
util_blitter_save_fragment_shader(rctx->blitter, rctx->ps_shader);
util_blitter_save_vertex_shader(rctx->blitter, rctx->vs_shader);
util_blitter_save_vertex_elements(rctx->blitter, rctx->vertex_elements);
if (rctx->queued.named.viewport) {
util_blitter_save_viewport(rctx->blitter, &rctx->queued.named.viewport->viewport);
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
}
util_blitter_save_vertex_buffer_slot(rctx->blitter, rctx->vertex_buffer);
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
util_blitter_save_so_targets(rctx->blitter, rctx->num_so_targets,
(struct pipe_stream_output_target**)rctx->so_targets);
if (op & R600_SAVE_FRAMEBUFFER)
util_blitter_save_framebuffer(rctx->blitter, &rctx->framebuffer);
if (op & R600_SAVE_TEXTURES) {
util_blitter_save_fragment_sampler_states(
rctx->blitter, rctx->samplers[PIPE_SHADER_FRAGMENT].n_samplers,
(void**)rctx->samplers[PIPE_SHADER_FRAGMENT].samplers);
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
util_blitter_save_fragment_sampler_views(rctx->blitter,
util_last_bit(rctx->samplers[PIPE_SHADER_FRAGMENT].views.desc.enabled_mask),
rctx->samplers[PIPE_SHADER_FRAGMENT].views.views);
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
}
if ((op & R600_DISABLE_RENDER_COND) && rctx->current_render_cond) {
rctx->saved_render_cond = rctx->current_render_cond;
rctx->saved_render_cond_cond = rctx->current_render_cond_cond;
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
rctx->saved_render_cond_mode = rctx->current_render_cond_mode;
rctx->context.render_condition(&rctx->context, NULL, FALSE, 0);
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
}
}
static void r600_blitter_end(struct pipe_context *ctx)
{
struct r600_context *rctx = (struct r600_context *)ctx;
if (rctx->saved_render_cond) {
rctx->context.render_condition(&rctx->context,
rctx->saved_render_cond,
rctx->saved_render_cond_cond,
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
rctx->saved_render_cond_mode);
rctx->saved_render_cond = NULL;
}
r600_context_queries_resume(rctx);
}
void si_blit_uncompress_depth(struct pipe_context *ctx,
struct r600_texture *texture,
struct r600_texture *staging,
radeonsi: More assorted depth/stencil changes ported from r600g. [ Squashed port of the following r600g commits: - Michel Dänzer ] commit 428e37c2da420f7dc14a2ea265f2387270f9bee1 Author: Marek Olšák <maraeo@gmail.com> Date: Tue Oct 2 22:02:54 2012 +0200 r600g: add in-place DB decompression and texturing with DB tiling The decompression is done in-place and only the compressed tiles are decompressed. Note: R6xx-R7xx can do that only with Z16 and Z32F. The texture unit is programmed to use non-displayable tiling and depth ordering of samples, so that it can fetch the texture in the native DB format. The latest version of the libdrm surface allocator is required for stencil texturing to work. The old one didn't create the mipmap tree correctly. We need a separate mipmap tree for stencil, because the stencil mipmap offsets are not really depth offsets/4. There are still some known bugs, but this should save some memory and it also improves performance a little bit in Lightsmark (especially with low resolutions; tested with Radeon HD 5000). The DB->CB copy is still used for transfers. commit e2f623f1d6da9bc987582ff68d0471061ae44030 Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 28 13:55:59 2012 +0200 r600g: don't decompress depth or stencil if there isn't any commit 43e226b6efb77db2247741cc2057d9625a2cfa05 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:32:50 2012 +0200 r600g: optimize uploading depth textures Make it only copy the portion of a depth texture being uploaded and not the whole 2D layer. There is also a little code cleanup. commit b242adbe5cfa165b252064a1ea36f802d8251ef1 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:17:46 2012 +0200 r600g: remove needless wrapper r600_texture_depth_flush commit 611dd529425281d73f1f0ad2000362d4a5525a25 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:05:14 2012 +0200 r600g: init_flushed_depth_texture should be able to report errors commit 80755ff56317446a8c89e611edc1fdf320d6779b Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 14 17:06:27 2012 +0200 r600g: properly track which textures are depth This fixes the issue with have_depth_texture never being set to false. commit fe1fd675565231b49d3ac53d0b4bec39d8bc6781 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 03:10:37 2012 +0200 r600g: don't flush depth textures set as colorbuffers 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. commit 5a17d8318ec2c20bf86275044dc8f715105a88e7 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 02:14:18 2012 +0200 r600g: flush depth textures bound to vertex shaders This was missing/broken. There are also minor code cleanups. commit dee58f94af833906863b0ff2955b20f3ab407e63 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 01:54:24 2012 +0200 r600g: do fine-grained depth texture flushing - 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) Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 18:36:41 +00:00
unsigned first_level, unsigned last_level,
unsigned first_layer, unsigned last_layer)
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
{
struct r600_context *rctx = (struct r600_context *)ctx;
radeonsi: More assorted depth/stencil changes ported from r600g. [ Squashed port of the following r600g commits: - Michel Dänzer ] commit 428e37c2da420f7dc14a2ea265f2387270f9bee1 Author: Marek Olšák <maraeo@gmail.com> Date: Tue Oct 2 22:02:54 2012 +0200 r600g: add in-place DB decompression and texturing with DB tiling The decompression is done in-place and only the compressed tiles are decompressed. Note: R6xx-R7xx can do that only with Z16 and Z32F. The texture unit is programmed to use non-displayable tiling and depth ordering of samples, so that it can fetch the texture in the native DB format. The latest version of the libdrm surface allocator is required for stencil texturing to work. The old one didn't create the mipmap tree correctly. We need a separate mipmap tree for stencil, because the stencil mipmap offsets are not really depth offsets/4. There are still some known bugs, but this should save some memory and it also improves performance a little bit in Lightsmark (especially with low resolutions; tested with Radeon HD 5000). The DB->CB copy is still used for transfers. commit e2f623f1d6da9bc987582ff68d0471061ae44030 Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 28 13:55:59 2012 +0200 r600g: don't decompress depth or stencil if there isn't any commit 43e226b6efb77db2247741cc2057d9625a2cfa05 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:32:50 2012 +0200 r600g: optimize uploading depth textures Make it only copy the portion of a depth texture being uploaded and not the whole 2D layer. There is also a little code cleanup. commit b242adbe5cfa165b252064a1ea36f802d8251ef1 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:17:46 2012 +0200 r600g: remove needless wrapper r600_texture_depth_flush commit 611dd529425281d73f1f0ad2000362d4a5525a25 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:05:14 2012 +0200 r600g: init_flushed_depth_texture should be able to report errors commit 80755ff56317446a8c89e611edc1fdf320d6779b Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 14 17:06:27 2012 +0200 r600g: properly track which textures are depth This fixes the issue with have_depth_texture never being set to false. commit fe1fd675565231b49d3ac53d0b4bec39d8bc6781 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 03:10:37 2012 +0200 r600g: don't flush depth textures set as colorbuffers 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. commit 5a17d8318ec2c20bf86275044dc8f715105a88e7 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 02:14:18 2012 +0200 r600g: flush depth textures bound to vertex shaders This was missing/broken. There are also minor code cleanups. commit dee58f94af833906863b0ff2955b20f3ab407e63 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 01:54:24 2012 +0200 r600g: do fine-grained depth texture flushing - 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) Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 18:36:41 +00:00
unsigned layer, level, checked_last_layer, max_layer;
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
float depth = 1.0f;
radeonsi: More assorted depth/stencil changes ported from r600g. [ Squashed port of the following r600g commits: - Michel Dänzer ] commit 428e37c2da420f7dc14a2ea265f2387270f9bee1 Author: Marek Olšák <maraeo@gmail.com> Date: Tue Oct 2 22:02:54 2012 +0200 r600g: add in-place DB decompression and texturing with DB tiling The decompression is done in-place and only the compressed tiles are decompressed. Note: R6xx-R7xx can do that only with Z16 and Z32F. The texture unit is programmed to use non-displayable tiling and depth ordering of samples, so that it can fetch the texture in the native DB format. The latest version of the libdrm surface allocator is required for stencil texturing to work. The old one didn't create the mipmap tree correctly. We need a separate mipmap tree for stencil, because the stencil mipmap offsets are not really depth offsets/4. There are still some known bugs, but this should save some memory and it also improves performance a little bit in Lightsmark (especially with low resolutions; tested with Radeon HD 5000). The DB->CB copy is still used for transfers. commit e2f623f1d6da9bc987582ff68d0471061ae44030 Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 28 13:55:59 2012 +0200 r600g: don't decompress depth or stencil if there isn't any commit 43e226b6efb77db2247741cc2057d9625a2cfa05 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:32:50 2012 +0200 r600g: optimize uploading depth textures Make it only copy the portion of a depth texture being uploaded and not the whole 2D layer. There is also a little code cleanup. commit b242adbe5cfa165b252064a1ea36f802d8251ef1 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:17:46 2012 +0200 r600g: remove needless wrapper r600_texture_depth_flush commit 611dd529425281d73f1f0ad2000362d4a5525a25 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:05:14 2012 +0200 r600g: init_flushed_depth_texture should be able to report errors commit 80755ff56317446a8c89e611edc1fdf320d6779b Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 14 17:06:27 2012 +0200 r600g: properly track which textures are depth This fixes the issue with have_depth_texture never being set to false. commit fe1fd675565231b49d3ac53d0b4bec39d8bc6781 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 03:10:37 2012 +0200 r600g: don't flush depth textures set as colorbuffers 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. commit 5a17d8318ec2c20bf86275044dc8f715105a88e7 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 02:14:18 2012 +0200 r600g: flush depth textures bound to vertex shaders This was missing/broken. There are also minor code cleanups. commit dee58f94af833906863b0ff2955b20f3ab407e63 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 01:54:24 2012 +0200 r600g: do fine-grained depth texture flushing - 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) Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 18:36:41 +00:00
const struct util_format_description *desc;
void *custom_dsa;
struct r600_texture *flushed_depth_texture = staging ?
staging : texture->flushed_depth_texture;
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
if (!staging && !texture->dirty_level_mask)
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
return;
radeonsi: More assorted depth/stencil changes ported from r600g. [ Squashed port of the following r600g commits: - Michel Dänzer ] commit 428e37c2da420f7dc14a2ea265f2387270f9bee1 Author: Marek Olšák <maraeo@gmail.com> Date: Tue Oct 2 22:02:54 2012 +0200 r600g: add in-place DB decompression and texturing with DB tiling The decompression is done in-place and only the compressed tiles are decompressed. Note: R6xx-R7xx can do that only with Z16 and Z32F. The texture unit is programmed to use non-displayable tiling and depth ordering of samples, so that it can fetch the texture in the native DB format. The latest version of the libdrm surface allocator is required for stencil texturing to work. The old one didn't create the mipmap tree correctly. We need a separate mipmap tree for stencil, because the stencil mipmap offsets are not really depth offsets/4. There are still some known bugs, but this should save some memory and it also improves performance a little bit in Lightsmark (especially with low resolutions; tested with Radeon HD 5000). The DB->CB copy is still used for transfers. commit e2f623f1d6da9bc987582ff68d0471061ae44030 Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 28 13:55:59 2012 +0200 r600g: don't decompress depth or stencil if there isn't any commit 43e226b6efb77db2247741cc2057d9625a2cfa05 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:32:50 2012 +0200 r600g: optimize uploading depth textures Make it only copy the portion of a depth texture being uploaded and not the whole 2D layer. There is also a little code cleanup. commit b242adbe5cfa165b252064a1ea36f802d8251ef1 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:17:46 2012 +0200 r600g: remove needless wrapper r600_texture_depth_flush commit 611dd529425281d73f1f0ad2000362d4a5525a25 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:05:14 2012 +0200 r600g: init_flushed_depth_texture should be able to report errors commit 80755ff56317446a8c89e611edc1fdf320d6779b Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 14 17:06:27 2012 +0200 r600g: properly track which textures are depth This fixes the issue with have_depth_texture never being set to false. commit fe1fd675565231b49d3ac53d0b4bec39d8bc6781 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 03:10:37 2012 +0200 r600g: don't flush depth textures set as colorbuffers 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. commit 5a17d8318ec2c20bf86275044dc8f715105a88e7 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 02:14:18 2012 +0200 r600g: flush depth textures bound to vertex shaders This was missing/broken. There are also minor code cleanups. commit dee58f94af833906863b0ff2955b20f3ab407e63 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 01:54:24 2012 +0200 r600g: do fine-grained depth texture flushing - 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) Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 18:36:41 +00:00
desc = util_format_description(flushed_depth_texture->resource.b.b.format);
switch (util_format_has_depth(desc) | util_format_has_stencil(desc) << 1) {
default:
assert(!"No depth or stencil to uncompress");
case 3:
custom_dsa = rctx->custom_dsa_flush_depth_stencil;
break;
case 2:
custom_dsa = rctx->custom_dsa_flush_stencil;
break;
case 1:
custom_dsa = rctx->custom_dsa_flush_depth;
break;
}
for (level = first_level; level <= last_level; level++) {
if (!staging && !(texture->dirty_level_mask & (1 << level)))
radeonsi: More assorted depth/stencil changes ported from r600g. [ Squashed port of the following r600g commits: - Michel Dänzer ] commit 428e37c2da420f7dc14a2ea265f2387270f9bee1 Author: Marek Olšák <maraeo@gmail.com> Date: Tue Oct 2 22:02:54 2012 +0200 r600g: add in-place DB decompression and texturing with DB tiling The decompression is done in-place and only the compressed tiles are decompressed. Note: R6xx-R7xx can do that only with Z16 and Z32F. The texture unit is programmed to use non-displayable tiling and depth ordering of samples, so that it can fetch the texture in the native DB format. The latest version of the libdrm surface allocator is required for stencil texturing to work. The old one didn't create the mipmap tree correctly. We need a separate mipmap tree for stencil, because the stencil mipmap offsets are not really depth offsets/4. There are still some known bugs, but this should save some memory and it also improves performance a little bit in Lightsmark (especially with low resolutions; tested with Radeon HD 5000). The DB->CB copy is still used for transfers. commit e2f623f1d6da9bc987582ff68d0471061ae44030 Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 28 13:55:59 2012 +0200 r600g: don't decompress depth or stencil if there isn't any commit 43e226b6efb77db2247741cc2057d9625a2cfa05 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:32:50 2012 +0200 r600g: optimize uploading depth textures Make it only copy the portion of a depth texture being uploaded and not the whole 2D layer. There is also a little code cleanup. commit b242adbe5cfa165b252064a1ea36f802d8251ef1 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:17:46 2012 +0200 r600g: remove needless wrapper r600_texture_depth_flush commit 611dd529425281d73f1f0ad2000362d4a5525a25 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:05:14 2012 +0200 r600g: init_flushed_depth_texture should be able to report errors commit 80755ff56317446a8c89e611edc1fdf320d6779b Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 14 17:06:27 2012 +0200 r600g: properly track which textures are depth This fixes the issue with have_depth_texture never being set to false. commit fe1fd675565231b49d3ac53d0b4bec39d8bc6781 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 03:10:37 2012 +0200 r600g: don't flush depth textures set as colorbuffers 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. commit 5a17d8318ec2c20bf86275044dc8f715105a88e7 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 02:14:18 2012 +0200 r600g: flush depth textures bound to vertex shaders This was missing/broken. There are also minor code cleanups. commit dee58f94af833906863b0ff2955b20f3ab407e63 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 01:54:24 2012 +0200 r600g: do fine-grained depth texture flushing - 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) Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 18:36:41 +00:00
continue;
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
radeonsi: More assorted depth/stencil changes ported from r600g. [ Squashed port of the following r600g commits: - Michel Dänzer ] commit 428e37c2da420f7dc14a2ea265f2387270f9bee1 Author: Marek Olšák <maraeo@gmail.com> Date: Tue Oct 2 22:02:54 2012 +0200 r600g: add in-place DB decompression and texturing with DB tiling The decompression is done in-place and only the compressed tiles are decompressed. Note: R6xx-R7xx can do that only with Z16 and Z32F. The texture unit is programmed to use non-displayable tiling and depth ordering of samples, so that it can fetch the texture in the native DB format. The latest version of the libdrm surface allocator is required for stencil texturing to work. The old one didn't create the mipmap tree correctly. We need a separate mipmap tree for stencil, because the stencil mipmap offsets are not really depth offsets/4. There are still some known bugs, but this should save some memory and it also improves performance a little bit in Lightsmark (especially with low resolutions; tested with Radeon HD 5000). The DB->CB copy is still used for transfers. commit e2f623f1d6da9bc987582ff68d0471061ae44030 Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 28 13:55:59 2012 +0200 r600g: don't decompress depth or stencil if there isn't any commit 43e226b6efb77db2247741cc2057d9625a2cfa05 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:32:50 2012 +0200 r600g: optimize uploading depth textures Make it only copy the portion of a depth texture being uploaded and not the whole 2D layer. There is also a little code cleanup. commit b242adbe5cfa165b252064a1ea36f802d8251ef1 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:17:46 2012 +0200 r600g: remove needless wrapper r600_texture_depth_flush commit 611dd529425281d73f1f0ad2000362d4a5525a25 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:05:14 2012 +0200 r600g: init_flushed_depth_texture should be able to report errors commit 80755ff56317446a8c89e611edc1fdf320d6779b Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 14 17:06:27 2012 +0200 r600g: properly track which textures are depth This fixes the issue with have_depth_texture never being set to false. commit fe1fd675565231b49d3ac53d0b4bec39d8bc6781 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 03:10:37 2012 +0200 r600g: don't flush depth textures set as colorbuffers 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. commit 5a17d8318ec2c20bf86275044dc8f715105a88e7 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 02:14:18 2012 +0200 r600g: flush depth textures bound to vertex shaders This was missing/broken. There are also minor code cleanups. commit dee58f94af833906863b0ff2955b20f3ab407e63 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 01:54:24 2012 +0200 r600g: do fine-grained depth texture flushing - 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) Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 18:36:41 +00:00
/* The smaller the mipmap level, the less layers there are
* as far as 3D textures are concerned. */
max_layer = util_max_layer(&texture->resource.b.b, level);
radeonsi: More assorted depth/stencil changes ported from r600g. [ Squashed port of the following r600g commits: - Michel Dänzer ] commit 428e37c2da420f7dc14a2ea265f2387270f9bee1 Author: Marek Olšák <maraeo@gmail.com> Date: Tue Oct 2 22:02:54 2012 +0200 r600g: add in-place DB decompression and texturing with DB tiling The decompression is done in-place and only the compressed tiles are decompressed. Note: R6xx-R7xx can do that only with Z16 and Z32F. The texture unit is programmed to use non-displayable tiling and depth ordering of samples, so that it can fetch the texture in the native DB format. The latest version of the libdrm surface allocator is required for stencil texturing to work. The old one didn't create the mipmap tree correctly. We need a separate mipmap tree for stencil, because the stencil mipmap offsets are not really depth offsets/4. There are still some known bugs, but this should save some memory and it also improves performance a little bit in Lightsmark (especially with low resolutions; tested with Radeon HD 5000). The DB->CB copy is still used for transfers. commit e2f623f1d6da9bc987582ff68d0471061ae44030 Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 28 13:55:59 2012 +0200 r600g: don't decompress depth or stencil if there isn't any commit 43e226b6efb77db2247741cc2057d9625a2cfa05 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:32:50 2012 +0200 r600g: optimize uploading depth textures Make it only copy the portion of a depth texture being uploaded and not the whole 2D layer. There is also a little code cleanup. commit b242adbe5cfa165b252064a1ea36f802d8251ef1 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:17:46 2012 +0200 r600g: remove needless wrapper r600_texture_depth_flush commit 611dd529425281d73f1f0ad2000362d4a5525a25 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:05:14 2012 +0200 r600g: init_flushed_depth_texture should be able to report errors commit 80755ff56317446a8c89e611edc1fdf320d6779b Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 14 17:06:27 2012 +0200 r600g: properly track which textures are depth This fixes the issue with have_depth_texture never being set to false. commit fe1fd675565231b49d3ac53d0b4bec39d8bc6781 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 03:10:37 2012 +0200 r600g: don't flush depth textures set as colorbuffers 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. commit 5a17d8318ec2c20bf86275044dc8f715105a88e7 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 02:14:18 2012 +0200 r600g: flush depth textures bound to vertex shaders This was missing/broken. There are also minor code cleanups. commit dee58f94af833906863b0ff2955b20f3ab407e63 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 01:54:24 2012 +0200 r600g: do fine-grained depth texture flushing - 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) Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 18:36:41 +00:00
checked_last_layer = last_layer < max_layer ? last_layer : max_layer;
for (layer = first_layer; layer <= checked_last_layer; layer++) {
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
struct pipe_surface *zsurf, *cbsurf, surf_tmpl;
surf_tmpl.format = texture->real_format;
surf_tmpl.u.tex.level = level;
surf_tmpl.u.tex.first_layer = layer;
surf_tmpl.u.tex.last_layer = layer;
zsurf = ctx->create_surface(ctx, &texture->resource.b.b, &surf_tmpl);
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
surf_tmpl.format = flushed_depth_texture->real_format;
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
cbsurf = ctx->create_surface(ctx,
(struct pipe_resource*)flushed_depth_texture, &surf_tmpl);
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
r600_blitter_begin(ctx, R600_DECOMPRESS);
radeonsi: More assorted depth/stencil changes ported from r600g. [ Squashed port of the following r600g commits: - Michel Dänzer ] commit 428e37c2da420f7dc14a2ea265f2387270f9bee1 Author: Marek Olšák <maraeo@gmail.com> Date: Tue Oct 2 22:02:54 2012 +0200 r600g: add in-place DB decompression and texturing with DB tiling The decompression is done in-place and only the compressed tiles are decompressed. Note: R6xx-R7xx can do that only with Z16 and Z32F. The texture unit is programmed to use non-displayable tiling and depth ordering of samples, so that it can fetch the texture in the native DB format. The latest version of the libdrm surface allocator is required for stencil texturing to work. The old one didn't create the mipmap tree correctly. We need a separate mipmap tree for stencil, because the stencil mipmap offsets are not really depth offsets/4. There are still some known bugs, but this should save some memory and it also improves performance a little bit in Lightsmark (especially with low resolutions; tested with Radeon HD 5000). The DB->CB copy is still used for transfers. commit e2f623f1d6da9bc987582ff68d0471061ae44030 Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 28 13:55:59 2012 +0200 r600g: don't decompress depth or stencil if there isn't any commit 43e226b6efb77db2247741cc2057d9625a2cfa05 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:32:50 2012 +0200 r600g: optimize uploading depth textures Make it only copy the portion of a depth texture being uploaded and not the whole 2D layer. There is also a little code cleanup. commit b242adbe5cfa165b252064a1ea36f802d8251ef1 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:17:46 2012 +0200 r600g: remove needless wrapper r600_texture_depth_flush commit 611dd529425281d73f1f0ad2000362d4a5525a25 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:05:14 2012 +0200 r600g: init_flushed_depth_texture should be able to report errors commit 80755ff56317446a8c89e611edc1fdf320d6779b Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 14 17:06:27 2012 +0200 r600g: properly track which textures are depth This fixes the issue with have_depth_texture never being set to false. commit fe1fd675565231b49d3ac53d0b4bec39d8bc6781 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 03:10:37 2012 +0200 r600g: don't flush depth textures set as colorbuffers 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. commit 5a17d8318ec2c20bf86275044dc8f715105a88e7 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 02:14:18 2012 +0200 r600g: flush depth textures bound to vertex shaders This was missing/broken. There are also minor code cleanups. commit dee58f94af833906863b0ff2955b20f3ab407e63 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 01:54:24 2012 +0200 r600g: do fine-grained depth texture flushing - 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) Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 18:36:41 +00:00
util_blitter_custom_depth_stencil(rctx->blitter, zsurf, cbsurf, ~0, custom_dsa, depth);
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
r600_blitter_end(ctx);
pipe_surface_reference(&zsurf, NULL);
pipe_surface_reference(&cbsurf, NULL);
}
radeonsi: More assorted depth/stencil changes ported from r600g. [ Squashed port of the following r600g commits: - Michel Dänzer ] commit 428e37c2da420f7dc14a2ea265f2387270f9bee1 Author: Marek Olšák <maraeo@gmail.com> Date: Tue Oct 2 22:02:54 2012 +0200 r600g: add in-place DB decompression and texturing with DB tiling The decompression is done in-place and only the compressed tiles are decompressed. Note: R6xx-R7xx can do that only with Z16 and Z32F. The texture unit is programmed to use non-displayable tiling and depth ordering of samples, so that it can fetch the texture in the native DB format. The latest version of the libdrm surface allocator is required for stencil texturing to work. The old one didn't create the mipmap tree correctly. We need a separate mipmap tree for stencil, because the stencil mipmap offsets are not really depth offsets/4. There are still some known bugs, but this should save some memory and it also improves performance a little bit in Lightsmark (especially with low resolutions; tested with Radeon HD 5000). The DB->CB copy is still used for transfers. commit e2f623f1d6da9bc987582ff68d0471061ae44030 Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 28 13:55:59 2012 +0200 r600g: don't decompress depth or stencil if there isn't any commit 43e226b6efb77db2247741cc2057d9625a2cfa05 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:32:50 2012 +0200 r600g: optimize uploading depth textures Make it only copy the portion of a depth texture being uploaded and not the whole 2D layer. There is also a little code cleanup. commit b242adbe5cfa165b252064a1ea36f802d8251ef1 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:17:46 2012 +0200 r600g: remove needless wrapper r600_texture_depth_flush commit 611dd529425281d73f1f0ad2000362d4a5525a25 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:05:14 2012 +0200 r600g: init_flushed_depth_texture should be able to report errors commit 80755ff56317446a8c89e611edc1fdf320d6779b Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 14 17:06:27 2012 +0200 r600g: properly track which textures are depth This fixes the issue with have_depth_texture never being set to false. commit fe1fd675565231b49d3ac53d0b4bec39d8bc6781 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 03:10:37 2012 +0200 r600g: don't flush depth textures set as colorbuffers 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. commit 5a17d8318ec2c20bf86275044dc8f715105a88e7 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 02:14:18 2012 +0200 r600g: flush depth textures bound to vertex shaders This was missing/broken. There are also minor code cleanups. commit dee58f94af833906863b0ff2955b20f3ab407e63 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 01:54:24 2012 +0200 r600g: do fine-grained depth texture flushing - 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) Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 18:36:41 +00:00
/* The texture will always be dirty if some layers aren't flushed.
* I don't think this case can occur though. */
if (!staging && first_layer == 0 && last_layer == max_layer) {
texture->dirty_level_mask &= ~(1 << level);
radeonsi: More assorted depth/stencil changes ported from r600g. [ Squashed port of the following r600g commits: - Michel Dänzer ] commit 428e37c2da420f7dc14a2ea265f2387270f9bee1 Author: Marek Olšák <maraeo@gmail.com> Date: Tue Oct 2 22:02:54 2012 +0200 r600g: add in-place DB decompression and texturing with DB tiling The decompression is done in-place and only the compressed tiles are decompressed. Note: R6xx-R7xx can do that only with Z16 and Z32F. The texture unit is programmed to use non-displayable tiling and depth ordering of samples, so that it can fetch the texture in the native DB format. The latest version of the libdrm surface allocator is required for stencil texturing to work. The old one didn't create the mipmap tree correctly. We need a separate mipmap tree for stencil, because the stencil mipmap offsets are not really depth offsets/4. There are still some known bugs, but this should save some memory and it also improves performance a little bit in Lightsmark (especially with low resolutions; tested with Radeon HD 5000). The DB->CB copy is still used for transfers. commit e2f623f1d6da9bc987582ff68d0471061ae44030 Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 28 13:55:59 2012 +0200 r600g: don't decompress depth or stencil if there isn't any commit 43e226b6efb77db2247741cc2057d9625a2cfa05 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:32:50 2012 +0200 r600g: optimize uploading depth textures Make it only copy the portion of a depth texture being uploaded and not the whole 2D layer. There is also a little code cleanup. commit b242adbe5cfa165b252064a1ea36f802d8251ef1 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:17:46 2012 +0200 r600g: remove needless wrapper r600_texture_depth_flush commit 611dd529425281d73f1f0ad2000362d4a5525a25 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:05:14 2012 +0200 r600g: init_flushed_depth_texture should be able to report errors commit 80755ff56317446a8c89e611edc1fdf320d6779b Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 14 17:06:27 2012 +0200 r600g: properly track which textures are depth This fixes the issue with have_depth_texture never being set to false. commit fe1fd675565231b49d3ac53d0b4bec39d8bc6781 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 03:10:37 2012 +0200 r600g: don't flush depth textures set as colorbuffers 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. commit 5a17d8318ec2c20bf86275044dc8f715105a88e7 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 02:14:18 2012 +0200 r600g: flush depth textures bound to vertex shaders This was missing/broken. There are also minor code cleanups. commit dee58f94af833906863b0ff2955b20f3ab407e63 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 01:54:24 2012 +0200 r600g: do fine-grained depth texture flushing - 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) Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 18:36:41 +00:00
}
}
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
}
radeonsi: More assorted depth/stencil changes ported from r600g. [ Squashed port of the following r600g commits: - Michel Dänzer ] commit 428e37c2da420f7dc14a2ea265f2387270f9bee1 Author: Marek Olšák <maraeo@gmail.com> Date: Tue Oct 2 22:02:54 2012 +0200 r600g: add in-place DB decompression and texturing with DB tiling The decompression is done in-place and only the compressed tiles are decompressed. Note: R6xx-R7xx can do that only with Z16 and Z32F. The texture unit is programmed to use non-displayable tiling and depth ordering of samples, so that it can fetch the texture in the native DB format. The latest version of the libdrm surface allocator is required for stencil texturing to work. The old one didn't create the mipmap tree correctly. We need a separate mipmap tree for stencil, because the stencil mipmap offsets are not really depth offsets/4. There are still some known bugs, but this should save some memory and it also improves performance a little bit in Lightsmark (especially with low resolutions; tested with Radeon HD 5000). The DB->CB copy is still used for transfers. commit e2f623f1d6da9bc987582ff68d0471061ae44030 Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 28 13:55:59 2012 +0200 r600g: don't decompress depth or stencil if there isn't any commit 43e226b6efb77db2247741cc2057d9625a2cfa05 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:32:50 2012 +0200 r600g: optimize uploading depth textures Make it only copy the portion of a depth texture being uploaded and not the whole 2D layer. There is also a little code cleanup. commit b242adbe5cfa165b252064a1ea36f802d8251ef1 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:17:46 2012 +0200 r600g: remove needless wrapper r600_texture_depth_flush commit 611dd529425281d73f1f0ad2000362d4a5525a25 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:05:14 2012 +0200 r600g: init_flushed_depth_texture should be able to report errors commit 80755ff56317446a8c89e611edc1fdf320d6779b Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 14 17:06:27 2012 +0200 r600g: properly track which textures are depth This fixes the issue with have_depth_texture never being set to false. commit fe1fd675565231b49d3ac53d0b4bec39d8bc6781 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 03:10:37 2012 +0200 r600g: don't flush depth textures set as colorbuffers 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. commit 5a17d8318ec2c20bf86275044dc8f715105a88e7 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 02:14:18 2012 +0200 r600g: flush depth textures bound to vertex shaders This was missing/broken. There are also minor code cleanups. commit dee58f94af833906863b0ff2955b20f3ab407e63 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 01:54:24 2012 +0200 r600g: do fine-grained depth texture flushing - 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) Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 18:36:41 +00:00
static void si_blit_decompress_depth_in_place(struct r600_context *rctx,
struct r600_texture *texture,
radeonsi: More assorted depth/stencil changes ported from r600g. [ Squashed port of the following r600g commits: - Michel Dänzer ] commit 428e37c2da420f7dc14a2ea265f2387270f9bee1 Author: Marek Olšák <maraeo@gmail.com> Date: Tue Oct 2 22:02:54 2012 +0200 r600g: add in-place DB decompression and texturing with DB tiling The decompression is done in-place and only the compressed tiles are decompressed. Note: R6xx-R7xx can do that only with Z16 and Z32F. The texture unit is programmed to use non-displayable tiling and depth ordering of samples, so that it can fetch the texture in the native DB format. The latest version of the libdrm surface allocator is required for stencil texturing to work. The old one didn't create the mipmap tree correctly. We need a separate mipmap tree for stencil, because the stencil mipmap offsets are not really depth offsets/4. There are still some known bugs, but this should save some memory and it also improves performance a little bit in Lightsmark (especially with low resolutions; tested with Radeon HD 5000). The DB->CB copy is still used for transfers. commit e2f623f1d6da9bc987582ff68d0471061ae44030 Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 28 13:55:59 2012 +0200 r600g: don't decompress depth or stencil if there isn't any commit 43e226b6efb77db2247741cc2057d9625a2cfa05 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:32:50 2012 +0200 r600g: optimize uploading depth textures Make it only copy the portion of a depth texture being uploaded and not the whole 2D layer. There is also a little code cleanup. commit b242adbe5cfa165b252064a1ea36f802d8251ef1 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:17:46 2012 +0200 r600g: remove needless wrapper r600_texture_depth_flush commit 611dd529425281d73f1f0ad2000362d4a5525a25 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:05:14 2012 +0200 r600g: init_flushed_depth_texture should be able to report errors commit 80755ff56317446a8c89e611edc1fdf320d6779b Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 14 17:06:27 2012 +0200 r600g: properly track which textures are depth This fixes the issue with have_depth_texture never being set to false. commit fe1fd675565231b49d3ac53d0b4bec39d8bc6781 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 03:10:37 2012 +0200 r600g: don't flush depth textures set as colorbuffers 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. commit 5a17d8318ec2c20bf86275044dc8f715105a88e7 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 02:14:18 2012 +0200 r600g: flush depth textures bound to vertex shaders This was missing/broken. There are also minor code cleanups. commit dee58f94af833906863b0ff2955b20f3ab407e63 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 01:54:24 2012 +0200 r600g: do fine-grained depth texture flushing - 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) Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 18:36:41 +00:00
unsigned first_level, unsigned last_level,
unsigned first_layer, unsigned last_layer)
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
{
radeonsi: More assorted depth/stencil changes ported from r600g. [ Squashed port of the following r600g commits: - Michel Dänzer ] commit 428e37c2da420f7dc14a2ea265f2387270f9bee1 Author: Marek Olšák <maraeo@gmail.com> Date: Tue Oct 2 22:02:54 2012 +0200 r600g: add in-place DB decompression and texturing with DB tiling The decompression is done in-place and only the compressed tiles are decompressed. Note: R6xx-R7xx can do that only with Z16 and Z32F. The texture unit is programmed to use non-displayable tiling and depth ordering of samples, so that it can fetch the texture in the native DB format. The latest version of the libdrm surface allocator is required for stencil texturing to work. The old one didn't create the mipmap tree correctly. We need a separate mipmap tree for stencil, because the stencil mipmap offsets are not really depth offsets/4. There are still some known bugs, but this should save some memory and it also improves performance a little bit in Lightsmark (especially with low resolutions; tested with Radeon HD 5000). The DB->CB copy is still used for transfers. commit e2f623f1d6da9bc987582ff68d0471061ae44030 Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 28 13:55:59 2012 +0200 r600g: don't decompress depth or stencil if there isn't any commit 43e226b6efb77db2247741cc2057d9625a2cfa05 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:32:50 2012 +0200 r600g: optimize uploading depth textures Make it only copy the portion of a depth texture being uploaded and not the whole 2D layer. There is also a little code cleanup. commit b242adbe5cfa165b252064a1ea36f802d8251ef1 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:17:46 2012 +0200 r600g: remove needless wrapper r600_texture_depth_flush commit 611dd529425281d73f1f0ad2000362d4a5525a25 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:05:14 2012 +0200 r600g: init_flushed_depth_texture should be able to report errors commit 80755ff56317446a8c89e611edc1fdf320d6779b Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 14 17:06:27 2012 +0200 r600g: properly track which textures are depth This fixes the issue with have_depth_texture never being set to false. commit fe1fd675565231b49d3ac53d0b4bec39d8bc6781 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 03:10:37 2012 +0200 r600g: don't flush depth textures set as colorbuffers 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. commit 5a17d8318ec2c20bf86275044dc8f715105a88e7 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 02:14:18 2012 +0200 r600g: flush depth textures bound to vertex shaders This was missing/broken. There are also minor code cleanups. commit dee58f94af833906863b0ff2955b20f3ab407e63 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 01:54:24 2012 +0200 r600g: do fine-grained depth texture flushing - 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) Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 18:36:41 +00:00
struct pipe_surface *zsurf, surf_tmpl = {{0}};
unsigned layer, max_layer, checked_last_layer, level;
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
radeonsi: More assorted depth/stencil changes ported from r600g. [ Squashed port of the following r600g commits: - Michel Dänzer ] commit 428e37c2da420f7dc14a2ea265f2387270f9bee1 Author: Marek Olšák <maraeo@gmail.com> Date: Tue Oct 2 22:02:54 2012 +0200 r600g: add in-place DB decompression and texturing with DB tiling The decompression is done in-place and only the compressed tiles are decompressed. Note: R6xx-R7xx can do that only with Z16 and Z32F. The texture unit is programmed to use non-displayable tiling and depth ordering of samples, so that it can fetch the texture in the native DB format. The latest version of the libdrm surface allocator is required for stencil texturing to work. The old one didn't create the mipmap tree correctly. We need a separate mipmap tree for stencil, because the stencil mipmap offsets are not really depth offsets/4. There are still some known bugs, but this should save some memory and it also improves performance a little bit in Lightsmark (especially with low resolutions; tested with Radeon HD 5000). The DB->CB copy is still used for transfers. commit e2f623f1d6da9bc987582ff68d0471061ae44030 Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 28 13:55:59 2012 +0200 r600g: don't decompress depth or stencil if there isn't any commit 43e226b6efb77db2247741cc2057d9625a2cfa05 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:32:50 2012 +0200 r600g: optimize uploading depth textures Make it only copy the portion of a depth texture being uploaded and not the whole 2D layer. There is also a little code cleanup. commit b242adbe5cfa165b252064a1ea36f802d8251ef1 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:17:46 2012 +0200 r600g: remove needless wrapper r600_texture_depth_flush commit 611dd529425281d73f1f0ad2000362d4a5525a25 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:05:14 2012 +0200 r600g: init_flushed_depth_texture should be able to report errors commit 80755ff56317446a8c89e611edc1fdf320d6779b Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 14 17:06:27 2012 +0200 r600g: properly track which textures are depth This fixes the issue with have_depth_texture never being set to false. commit fe1fd675565231b49d3ac53d0b4bec39d8bc6781 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 03:10:37 2012 +0200 r600g: don't flush depth textures set as colorbuffers 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. commit 5a17d8318ec2c20bf86275044dc8f715105a88e7 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 02:14:18 2012 +0200 r600g: flush depth textures bound to vertex shaders This was missing/broken. There are also minor code cleanups. commit dee58f94af833906863b0ff2955b20f3ab407e63 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 01:54:24 2012 +0200 r600g: do fine-grained depth texture flushing - 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) Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 18:36:41 +00:00
surf_tmpl.format = texture->resource.b.b.format;
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
radeonsi: More assorted depth/stencil changes ported from r600g. [ Squashed port of the following r600g commits: - Michel Dänzer ] commit 428e37c2da420f7dc14a2ea265f2387270f9bee1 Author: Marek Olšák <maraeo@gmail.com> Date: Tue Oct 2 22:02:54 2012 +0200 r600g: add in-place DB decompression and texturing with DB tiling The decompression is done in-place and only the compressed tiles are decompressed. Note: R6xx-R7xx can do that only with Z16 and Z32F. The texture unit is programmed to use non-displayable tiling and depth ordering of samples, so that it can fetch the texture in the native DB format. The latest version of the libdrm surface allocator is required for stencil texturing to work. The old one didn't create the mipmap tree correctly. We need a separate mipmap tree for stencil, because the stencil mipmap offsets are not really depth offsets/4. There are still some known bugs, but this should save some memory and it also improves performance a little bit in Lightsmark (especially with low resolutions; tested with Radeon HD 5000). The DB->CB copy is still used for transfers. commit e2f623f1d6da9bc987582ff68d0471061ae44030 Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 28 13:55:59 2012 +0200 r600g: don't decompress depth or stencil if there isn't any commit 43e226b6efb77db2247741cc2057d9625a2cfa05 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:32:50 2012 +0200 r600g: optimize uploading depth textures Make it only copy the portion of a depth texture being uploaded and not the whole 2D layer. There is also a little code cleanup. commit b242adbe5cfa165b252064a1ea36f802d8251ef1 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:17:46 2012 +0200 r600g: remove needless wrapper r600_texture_depth_flush commit 611dd529425281d73f1f0ad2000362d4a5525a25 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:05:14 2012 +0200 r600g: init_flushed_depth_texture should be able to report errors commit 80755ff56317446a8c89e611edc1fdf320d6779b Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 14 17:06:27 2012 +0200 r600g: properly track which textures are depth This fixes the issue with have_depth_texture never being set to false. commit fe1fd675565231b49d3ac53d0b4bec39d8bc6781 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 03:10:37 2012 +0200 r600g: don't flush depth textures set as colorbuffers 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. commit 5a17d8318ec2c20bf86275044dc8f715105a88e7 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 02:14:18 2012 +0200 r600g: flush depth textures bound to vertex shaders This was missing/broken. There are also minor code cleanups. commit dee58f94af833906863b0ff2955b20f3ab407e63 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 01:54:24 2012 +0200 r600g: do fine-grained depth texture flushing - 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) Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 18:36:41 +00:00
for (level = first_level; level <= last_level; level++) {
if (!(texture->dirty_level_mask & (1 << level)))
radeonsi: More assorted depth/stencil changes ported from r600g. [ Squashed port of the following r600g commits: - Michel Dänzer ] commit 428e37c2da420f7dc14a2ea265f2387270f9bee1 Author: Marek Olšák <maraeo@gmail.com> Date: Tue Oct 2 22:02:54 2012 +0200 r600g: add in-place DB decompression and texturing with DB tiling The decompression is done in-place and only the compressed tiles are decompressed. Note: R6xx-R7xx can do that only with Z16 and Z32F. The texture unit is programmed to use non-displayable tiling and depth ordering of samples, so that it can fetch the texture in the native DB format. The latest version of the libdrm surface allocator is required for stencil texturing to work. The old one didn't create the mipmap tree correctly. We need a separate mipmap tree for stencil, because the stencil mipmap offsets are not really depth offsets/4. There are still some known bugs, but this should save some memory and it also improves performance a little bit in Lightsmark (especially with low resolutions; tested with Radeon HD 5000). The DB->CB copy is still used for transfers. commit e2f623f1d6da9bc987582ff68d0471061ae44030 Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 28 13:55:59 2012 +0200 r600g: don't decompress depth or stencil if there isn't any commit 43e226b6efb77db2247741cc2057d9625a2cfa05 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:32:50 2012 +0200 r600g: optimize uploading depth textures Make it only copy the portion of a depth texture being uploaded and not the whole 2D layer. There is also a little code cleanup. commit b242adbe5cfa165b252064a1ea36f802d8251ef1 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:17:46 2012 +0200 r600g: remove needless wrapper r600_texture_depth_flush commit 611dd529425281d73f1f0ad2000362d4a5525a25 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:05:14 2012 +0200 r600g: init_flushed_depth_texture should be able to report errors commit 80755ff56317446a8c89e611edc1fdf320d6779b Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 14 17:06:27 2012 +0200 r600g: properly track which textures are depth This fixes the issue with have_depth_texture never being set to false. commit fe1fd675565231b49d3ac53d0b4bec39d8bc6781 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 03:10:37 2012 +0200 r600g: don't flush depth textures set as colorbuffers 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. commit 5a17d8318ec2c20bf86275044dc8f715105a88e7 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 02:14:18 2012 +0200 r600g: flush depth textures bound to vertex shaders This was missing/broken. There are also minor code cleanups. commit dee58f94af833906863b0ff2955b20f3ab407e63 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 01:54:24 2012 +0200 r600g: do fine-grained depth texture flushing - 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) Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 18:36:41 +00:00
continue;
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
radeonsi: More assorted depth/stencil changes ported from r600g. [ Squashed port of the following r600g commits: - Michel Dänzer ] commit 428e37c2da420f7dc14a2ea265f2387270f9bee1 Author: Marek Olšák <maraeo@gmail.com> Date: Tue Oct 2 22:02:54 2012 +0200 r600g: add in-place DB decompression and texturing with DB tiling The decompression is done in-place and only the compressed tiles are decompressed. Note: R6xx-R7xx can do that only with Z16 and Z32F. The texture unit is programmed to use non-displayable tiling and depth ordering of samples, so that it can fetch the texture in the native DB format. The latest version of the libdrm surface allocator is required for stencil texturing to work. The old one didn't create the mipmap tree correctly. We need a separate mipmap tree for stencil, because the stencil mipmap offsets are not really depth offsets/4. There are still some known bugs, but this should save some memory and it also improves performance a little bit in Lightsmark (especially with low resolutions; tested with Radeon HD 5000). The DB->CB copy is still used for transfers. commit e2f623f1d6da9bc987582ff68d0471061ae44030 Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 28 13:55:59 2012 +0200 r600g: don't decompress depth or stencil if there isn't any commit 43e226b6efb77db2247741cc2057d9625a2cfa05 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:32:50 2012 +0200 r600g: optimize uploading depth textures Make it only copy the portion of a depth texture being uploaded and not the whole 2D layer. There is also a little code cleanup. commit b242adbe5cfa165b252064a1ea36f802d8251ef1 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:17:46 2012 +0200 r600g: remove needless wrapper r600_texture_depth_flush commit 611dd529425281d73f1f0ad2000362d4a5525a25 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:05:14 2012 +0200 r600g: init_flushed_depth_texture should be able to report errors commit 80755ff56317446a8c89e611edc1fdf320d6779b Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 14 17:06:27 2012 +0200 r600g: properly track which textures are depth This fixes the issue with have_depth_texture never being set to false. commit fe1fd675565231b49d3ac53d0b4bec39d8bc6781 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 03:10:37 2012 +0200 r600g: don't flush depth textures set as colorbuffers 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. commit 5a17d8318ec2c20bf86275044dc8f715105a88e7 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 02:14:18 2012 +0200 r600g: flush depth textures bound to vertex shaders This was missing/broken. There are also minor code cleanups. commit dee58f94af833906863b0ff2955b20f3ab407e63 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 01:54:24 2012 +0200 r600g: do fine-grained depth texture flushing - 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) Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 18:36:41 +00:00
surf_tmpl.u.tex.level = level;
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
radeonsi: More assorted depth/stencil changes ported from r600g. [ Squashed port of the following r600g commits: - Michel Dänzer ] commit 428e37c2da420f7dc14a2ea265f2387270f9bee1 Author: Marek Olšák <maraeo@gmail.com> Date: Tue Oct 2 22:02:54 2012 +0200 r600g: add in-place DB decompression and texturing with DB tiling The decompression is done in-place and only the compressed tiles are decompressed. Note: R6xx-R7xx can do that only with Z16 and Z32F. The texture unit is programmed to use non-displayable tiling and depth ordering of samples, so that it can fetch the texture in the native DB format. The latest version of the libdrm surface allocator is required for stencil texturing to work. The old one didn't create the mipmap tree correctly. We need a separate mipmap tree for stencil, because the stencil mipmap offsets are not really depth offsets/4. There are still some known bugs, but this should save some memory and it also improves performance a little bit in Lightsmark (especially with low resolutions; tested with Radeon HD 5000). The DB->CB copy is still used for transfers. commit e2f623f1d6da9bc987582ff68d0471061ae44030 Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 28 13:55:59 2012 +0200 r600g: don't decompress depth or stencil if there isn't any commit 43e226b6efb77db2247741cc2057d9625a2cfa05 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:32:50 2012 +0200 r600g: optimize uploading depth textures Make it only copy the portion of a depth texture being uploaded and not the whole 2D layer. There is also a little code cleanup. commit b242adbe5cfa165b252064a1ea36f802d8251ef1 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:17:46 2012 +0200 r600g: remove needless wrapper r600_texture_depth_flush commit 611dd529425281d73f1f0ad2000362d4a5525a25 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:05:14 2012 +0200 r600g: init_flushed_depth_texture should be able to report errors commit 80755ff56317446a8c89e611edc1fdf320d6779b Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 14 17:06:27 2012 +0200 r600g: properly track which textures are depth This fixes the issue with have_depth_texture never being set to false. commit fe1fd675565231b49d3ac53d0b4bec39d8bc6781 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 03:10:37 2012 +0200 r600g: don't flush depth textures set as colorbuffers 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. commit 5a17d8318ec2c20bf86275044dc8f715105a88e7 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 02:14:18 2012 +0200 r600g: flush depth textures bound to vertex shaders This was missing/broken. There are also minor code cleanups. commit dee58f94af833906863b0ff2955b20f3ab407e63 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 01:54:24 2012 +0200 r600g: do fine-grained depth texture flushing - 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) Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 18:36:41 +00:00
/* The smaller the mipmap level, the less layers there are
* as far as 3D textures are concerned. */
max_layer = util_max_layer(&texture->resource.b.b, level);
radeonsi: More assorted depth/stencil changes ported from r600g. [ Squashed port of the following r600g commits: - Michel Dänzer ] commit 428e37c2da420f7dc14a2ea265f2387270f9bee1 Author: Marek Olšák <maraeo@gmail.com> Date: Tue Oct 2 22:02:54 2012 +0200 r600g: add in-place DB decompression and texturing with DB tiling The decompression is done in-place and only the compressed tiles are decompressed. Note: R6xx-R7xx can do that only with Z16 and Z32F. The texture unit is programmed to use non-displayable tiling and depth ordering of samples, so that it can fetch the texture in the native DB format. The latest version of the libdrm surface allocator is required for stencil texturing to work. The old one didn't create the mipmap tree correctly. We need a separate mipmap tree for stencil, because the stencil mipmap offsets are not really depth offsets/4. There are still some known bugs, but this should save some memory and it also improves performance a little bit in Lightsmark (especially with low resolutions; tested with Radeon HD 5000). The DB->CB copy is still used for transfers. commit e2f623f1d6da9bc987582ff68d0471061ae44030 Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 28 13:55:59 2012 +0200 r600g: don't decompress depth or stencil if there isn't any commit 43e226b6efb77db2247741cc2057d9625a2cfa05 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:32:50 2012 +0200 r600g: optimize uploading depth textures Make it only copy the portion of a depth texture being uploaded and not the whole 2D layer. There is also a little code cleanup. commit b242adbe5cfa165b252064a1ea36f802d8251ef1 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:17:46 2012 +0200 r600g: remove needless wrapper r600_texture_depth_flush commit 611dd529425281d73f1f0ad2000362d4a5525a25 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:05:14 2012 +0200 r600g: init_flushed_depth_texture should be able to report errors commit 80755ff56317446a8c89e611edc1fdf320d6779b Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 14 17:06:27 2012 +0200 r600g: properly track which textures are depth This fixes the issue with have_depth_texture never being set to false. commit fe1fd675565231b49d3ac53d0b4bec39d8bc6781 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 03:10:37 2012 +0200 r600g: don't flush depth textures set as colorbuffers 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. commit 5a17d8318ec2c20bf86275044dc8f715105a88e7 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 02:14:18 2012 +0200 r600g: flush depth textures bound to vertex shaders This was missing/broken. There are also minor code cleanups. commit dee58f94af833906863b0ff2955b20f3ab407e63 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 01:54:24 2012 +0200 r600g: do fine-grained depth texture flushing - 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) Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 18:36:41 +00:00
checked_last_layer = last_layer < max_layer ? last_layer : max_layer;
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
radeonsi: More assorted depth/stencil changes ported from r600g. [ Squashed port of the following r600g commits: - Michel Dänzer ] commit 428e37c2da420f7dc14a2ea265f2387270f9bee1 Author: Marek Olšák <maraeo@gmail.com> Date: Tue Oct 2 22:02:54 2012 +0200 r600g: add in-place DB decompression and texturing with DB tiling The decompression is done in-place and only the compressed tiles are decompressed. Note: R6xx-R7xx can do that only with Z16 and Z32F. The texture unit is programmed to use non-displayable tiling and depth ordering of samples, so that it can fetch the texture in the native DB format. The latest version of the libdrm surface allocator is required for stencil texturing to work. The old one didn't create the mipmap tree correctly. We need a separate mipmap tree for stencil, because the stencil mipmap offsets are not really depth offsets/4. There are still some known bugs, but this should save some memory and it also improves performance a little bit in Lightsmark (especially with low resolutions; tested with Radeon HD 5000). The DB->CB copy is still used for transfers. commit e2f623f1d6da9bc987582ff68d0471061ae44030 Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 28 13:55:59 2012 +0200 r600g: don't decompress depth or stencil if there isn't any commit 43e226b6efb77db2247741cc2057d9625a2cfa05 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:32:50 2012 +0200 r600g: optimize uploading depth textures Make it only copy the portion of a depth texture being uploaded and not the whole 2D layer. There is also a little code cleanup. commit b242adbe5cfa165b252064a1ea36f802d8251ef1 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:17:46 2012 +0200 r600g: remove needless wrapper r600_texture_depth_flush commit 611dd529425281d73f1f0ad2000362d4a5525a25 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:05:14 2012 +0200 r600g: init_flushed_depth_texture should be able to report errors commit 80755ff56317446a8c89e611edc1fdf320d6779b Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 14 17:06:27 2012 +0200 r600g: properly track which textures are depth This fixes the issue with have_depth_texture never being set to false. commit fe1fd675565231b49d3ac53d0b4bec39d8bc6781 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 03:10:37 2012 +0200 r600g: don't flush depth textures set as colorbuffers 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. commit 5a17d8318ec2c20bf86275044dc8f715105a88e7 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 02:14:18 2012 +0200 r600g: flush depth textures bound to vertex shaders This was missing/broken. There are also minor code cleanups. commit dee58f94af833906863b0ff2955b20f3ab407e63 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 01:54:24 2012 +0200 r600g: do fine-grained depth texture flushing - 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) Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 18:36:41 +00:00
for (layer = first_layer; layer <= checked_last_layer; layer++) {
surf_tmpl.u.tex.first_layer = layer;
surf_tmpl.u.tex.last_layer = layer;
zsurf = rctx->context.create_surface(&rctx->context, &texture->resource.b.b, &surf_tmpl);
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
radeonsi: More assorted depth/stencil changes ported from r600g. [ Squashed port of the following r600g commits: - Michel Dänzer ] commit 428e37c2da420f7dc14a2ea265f2387270f9bee1 Author: Marek Olšák <maraeo@gmail.com> Date: Tue Oct 2 22:02:54 2012 +0200 r600g: add in-place DB decompression and texturing with DB tiling The decompression is done in-place and only the compressed tiles are decompressed. Note: R6xx-R7xx can do that only with Z16 and Z32F. The texture unit is programmed to use non-displayable tiling and depth ordering of samples, so that it can fetch the texture in the native DB format. The latest version of the libdrm surface allocator is required for stencil texturing to work. The old one didn't create the mipmap tree correctly. We need a separate mipmap tree for stencil, because the stencil mipmap offsets are not really depth offsets/4. There are still some known bugs, but this should save some memory and it also improves performance a little bit in Lightsmark (especially with low resolutions; tested with Radeon HD 5000). The DB->CB copy is still used for transfers. commit e2f623f1d6da9bc987582ff68d0471061ae44030 Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 28 13:55:59 2012 +0200 r600g: don't decompress depth or stencil if there isn't any commit 43e226b6efb77db2247741cc2057d9625a2cfa05 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:32:50 2012 +0200 r600g: optimize uploading depth textures Make it only copy the portion of a depth texture being uploaded and not the whole 2D layer. There is also a little code cleanup. commit b242adbe5cfa165b252064a1ea36f802d8251ef1 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:17:46 2012 +0200 r600g: remove needless wrapper r600_texture_depth_flush commit 611dd529425281d73f1f0ad2000362d4a5525a25 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:05:14 2012 +0200 r600g: init_flushed_depth_texture should be able to report errors commit 80755ff56317446a8c89e611edc1fdf320d6779b Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 14 17:06:27 2012 +0200 r600g: properly track which textures are depth This fixes the issue with have_depth_texture never being set to false. commit fe1fd675565231b49d3ac53d0b4bec39d8bc6781 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 03:10:37 2012 +0200 r600g: don't flush depth textures set as colorbuffers 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. commit 5a17d8318ec2c20bf86275044dc8f715105a88e7 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 02:14:18 2012 +0200 r600g: flush depth textures bound to vertex shaders This was missing/broken. There are also minor code cleanups. commit dee58f94af833906863b0ff2955b20f3ab407e63 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 01:54:24 2012 +0200 r600g: do fine-grained depth texture flushing - 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) Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 18:36:41 +00:00
r600_blitter_begin(&rctx->context, R600_DECOMPRESS);
util_blitter_custom_depth_stencil(rctx->blitter, zsurf, NULL, ~0,
rctx->custom_dsa_flush_inplace,
1.0f);
r600_blitter_end(&rctx->context);
pipe_surface_reference(&zsurf, NULL);
}
/* The texture will always be dirty if some layers aren't flushed.
* I don't think this case occurs often though. */
if (first_layer == 0 && last_layer == max_layer) {
texture->dirty_level_mask &= ~(1 << level);
radeonsi: More assorted depth/stencil changes ported from r600g. [ Squashed port of the following r600g commits: - Michel Dänzer ] commit 428e37c2da420f7dc14a2ea265f2387270f9bee1 Author: Marek Olšák <maraeo@gmail.com> Date: Tue Oct 2 22:02:54 2012 +0200 r600g: add in-place DB decompression and texturing with DB tiling The decompression is done in-place and only the compressed tiles are decompressed. Note: R6xx-R7xx can do that only with Z16 and Z32F. The texture unit is programmed to use non-displayable tiling and depth ordering of samples, so that it can fetch the texture in the native DB format. The latest version of the libdrm surface allocator is required for stencil texturing to work. The old one didn't create the mipmap tree correctly. We need a separate mipmap tree for stencil, because the stencil mipmap offsets are not really depth offsets/4. There are still some known bugs, but this should save some memory and it also improves performance a little bit in Lightsmark (especially with low resolutions; tested with Radeon HD 5000). The DB->CB copy is still used for transfers. commit e2f623f1d6da9bc987582ff68d0471061ae44030 Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 28 13:55:59 2012 +0200 r600g: don't decompress depth or stencil if there isn't any commit 43e226b6efb77db2247741cc2057d9625a2cfa05 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:32:50 2012 +0200 r600g: optimize uploading depth textures Make it only copy the portion of a depth texture being uploaded and not the whole 2D layer. There is also a little code cleanup. commit b242adbe5cfa165b252064a1ea36f802d8251ef1 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:17:46 2012 +0200 r600g: remove needless wrapper r600_texture_depth_flush commit 611dd529425281d73f1f0ad2000362d4a5525a25 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:05:14 2012 +0200 r600g: init_flushed_depth_texture should be able to report errors commit 80755ff56317446a8c89e611edc1fdf320d6779b Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 14 17:06:27 2012 +0200 r600g: properly track which textures are depth This fixes the issue with have_depth_texture never being set to false. commit fe1fd675565231b49d3ac53d0b4bec39d8bc6781 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 03:10:37 2012 +0200 r600g: don't flush depth textures set as colorbuffers 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. commit 5a17d8318ec2c20bf86275044dc8f715105a88e7 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 02:14:18 2012 +0200 r600g: flush depth textures bound to vertex shaders This was missing/broken. There are also minor code cleanups. commit dee58f94af833906863b0ff2955b20f3ab407e63 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 01:54:24 2012 +0200 r600g: do fine-grained depth texture flushing - 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) Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 18:36:41 +00:00
}
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
}
radeonsi: More assorted depth/stencil changes ported from r600g. [ Squashed port of the following r600g commits: - Michel Dänzer ] commit 428e37c2da420f7dc14a2ea265f2387270f9bee1 Author: Marek Olšák <maraeo@gmail.com> Date: Tue Oct 2 22:02:54 2012 +0200 r600g: add in-place DB decompression and texturing with DB tiling The decompression is done in-place and only the compressed tiles are decompressed. Note: R6xx-R7xx can do that only with Z16 and Z32F. The texture unit is programmed to use non-displayable tiling and depth ordering of samples, so that it can fetch the texture in the native DB format. The latest version of the libdrm surface allocator is required for stencil texturing to work. The old one didn't create the mipmap tree correctly. We need a separate mipmap tree for stencil, because the stencil mipmap offsets are not really depth offsets/4. There are still some known bugs, but this should save some memory and it also improves performance a little bit in Lightsmark (especially with low resolutions; tested with Radeon HD 5000). The DB->CB copy is still used for transfers. commit e2f623f1d6da9bc987582ff68d0471061ae44030 Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 28 13:55:59 2012 +0200 r600g: don't decompress depth or stencil if there isn't any commit 43e226b6efb77db2247741cc2057d9625a2cfa05 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:32:50 2012 +0200 r600g: optimize uploading depth textures Make it only copy the portion of a depth texture being uploaded and not the whole 2D layer. There is also a little code cleanup. commit b242adbe5cfa165b252064a1ea36f802d8251ef1 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:17:46 2012 +0200 r600g: remove needless wrapper r600_texture_depth_flush commit 611dd529425281d73f1f0ad2000362d4a5525a25 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:05:14 2012 +0200 r600g: init_flushed_depth_texture should be able to report errors commit 80755ff56317446a8c89e611edc1fdf320d6779b Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 14 17:06:27 2012 +0200 r600g: properly track which textures are depth This fixes the issue with have_depth_texture never being set to false. commit fe1fd675565231b49d3ac53d0b4bec39d8bc6781 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 03:10:37 2012 +0200 r600g: don't flush depth textures set as colorbuffers 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. commit 5a17d8318ec2c20bf86275044dc8f715105a88e7 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 02:14:18 2012 +0200 r600g: flush depth textures bound to vertex shaders This was missing/broken. There are also minor code cleanups. commit dee58f94af833906863b0ff2955b20f3ab407e63 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 01:54:24 2012 +0200 r600g: do fine-grained depth texture flushing - 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) Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 18:36:41 +00:00
}
void si_flush_depth_textures(struct r600_context *rctx,
struct r600_textures_info *textures)
{
unsigned i;
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
radeonsi: More assorted depth/stencil changes ported from r600g. [ Squashed port of the following r600g commits: - Michel Dänzer ] commit 428e37c2da420f7dc14a2ea265f2387270f9bee1 Author: Marek Olšák <maraeo@gmail.com> Date: Tue Oct 2 22:02:54 2012 +0200 r600g: add in-place DB decompression and texturing with DB tiling The decompression is done in-place and only the compressed tiles are decompressed. Note: R6xx-R7xx can do that only with Z16 and Z32F. The texture unit is programmed to use non-displayable tiling and depth ordering of samples, so that it can fetch the texture in the native DB format. The latest version of the libdrm surface allocator is required for stencil texturing to work. The old one didn't create the mipmap tree correctly. We need a separate mipmap tree for stencil, because the stencil mipmap offsets are not really depth offsets/4. There are still some known bugs, but this should save some memory and it also improves performance a little bit in Lightsmark (especially with low resolutions; tested with Radeon HD 5000). The DB->CB copy is still used for transfers. commit e2f623f1d6da9bc987582ff68d0471061ae44030 Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 28 13:55:59 2012 +0200 r600g: don't decompress depth or stencil if there isn't any commit 43e226b6efb77db2247741cc2057d9625a2cfa05 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:32:50 2012 +0200 r600g: optimize uploading depth textures Make it only copy the portion of a depth texture being uploaded and not the whole 2D layer. There is also a little code cleanup. commit b242adbe5cfa165b252064a1ea36f802d8251ef1 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:17:46 2012 +0200 r600g: remove needless wrapper r600_texture_depth_flush commit 611dd529425281d73f1f0ad2000362d4a5525a25 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:05:14 2012 +0200 r600g: init_flushed_depth_texture should be able to report errors commit 80755ff56317446a8c89e611edc1fdf320d6779b Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 14 17:06:27 2012 +0200 r600g: properly track which textures are depth This fixes the issue with have_depth_texture never being set to false. commit fe1fd675565231b49d3ac53d0b4bec39d8bc6781 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 03:10:37 2012 +0200 r600g: don't flush depth textures set as colorbuffers 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. commit 5a17d8318ec2c20bf86275044dc8f715105a88e7 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 02:14:18 2012 +0200 r600g: flush depth textures bound to vertex shaders This was missing/broken. There are also minor code cleanups. commit dee58f94af833906863b0ff2955b20f3ab407e63 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 01:54:24 2012 +0200 r600g: do fine-grained depth texture flushing - 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) Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 18:36:41 +00:00
for (i = 0; i < textures->n_views; ++i) {
struct pipe_sampler_view *view;
struct r600_texture *tex;
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
view = textures->views.views[i];
radeonsi: More assorted depth/stencil changes ported from r600g. [ Squashed port of the following r600g commits: - Michel Dänzer ] commit 428e37c2da420f7dc14a2ea265f2387270f9bee1 Author: Marek Olšák <maraeo@gmail.com> Date: Tue Oct 2 22:02:54 2012 +0200 r600g: add in-place DB decompression and texturing with DB tiling The decompression is done in-place and only the compressed tiles are decompressed. Note: R6xx-R7xx can do that only with Z16 and Z32F. The texture unit is programmed to use non-displayable tiling and depth ordering of samples, so that it can fetch the texture in the native DB format. The latest version of the libdrm surface allocator is required for stencil texturing to work. The old one didn't create the mipmap tree correctly. We need a separate mipmap tree for stencil, because the stencil mipmap offsets are not really depth offsets/4. There are still some known bugs, but this should save some memory and it also improves performance a little bit in Lightsmark (especially with low resolutions; tested with Radeon HD 5000). The DB->CB copy is still used for transfers. commit e2f623f1d6da9bc987582ff68d0471061ae44030 Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 28 13:55:59 2012 +0200 r600g: don't decompress depth or stencil if there isn't any commit 43e226b6efb77db2247741cc2057d9625a2cfa05 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:32:50 2012 +0200 r600g: optimize uploading depth textures Make it only copy the portion of a depth texture being uploaded and not the whole 2D layer. There is also a little code cleanup. commit b242adbe5cfa165b252064a1ea36f802d8251ef1 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:17:46 2012 +0200 r600g: remove needless wrapper r600_texture_depth_flush commit 611dd529425281d73f1f0ad2000362d4a5525a25 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:05:14 2012 +0200 r600g: init_flushed_depth_texture should be able to report errors commit 80755ff56317446a8c89e611edc1fdf320d6779b Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 14 17:06:27 2012 +0200 r600g: properly track which textures are depth This fixes the issue with have_depth_texture never being set to false. commit fe1fd675565231b49d3ac53d0b4bec39d8bc6781 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 03:10:37 2012 +0200 r600g: don't flush depth textures set as colorbuffers 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. commit 5a17d8318ec2c20bf86275044dc8f715105a88e7 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 02:14:18 2012 +0200 r600g: flush depth textures bound to vertex shaders This was missing/broken. There are also minor code cleanups. commit dee58f94af833906863b0ff2955b20f3ab407e63 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 01:54:24 2012 +0200 r600g: do fine-grained depth texture flushing - 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) Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 18:36:41 +00:00
if (!view) continue;
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
tex = (struct r600_texture *)view->texture;
radeonsi: More assorted depth/stencil changes ported from r600g. [ Squashed port of the following r600g commits: - Michel Dänzer ] commit 428e37c2da420f7dc14a2ea265f2387270f9bee1 Author: Marek Olšák <maraeo@gmail.com> Date: Tue Oct 2 22:02:54 2012 +0200 r600g: add in-place DB decompression and texturing with DB tiling The decompression is done in-place and only the compressed tiles are decompressed. Note: R6xx-R7xx can do that only with Z16 and Z32F. The texture unit is programmed to use non-displayable tiling and depth ordering of samples, so that it can fetch the texture in the native DB format. The latest version of the libdrm surface allocator is required for stencil texturing to work. The old one didn't create the mipmap tree correctly. We need a separate mipmap tree for stencil, because the stencil mipmap offsets are not really depth offsets/4. There are still some known bugs, but this should save some memory and it also improves performance a little bit in Lightsmark (especially with low resolutions; tested with Radeon HD 5000). The DB->CB copy is still used for transfers. commit e2f623f1d6da9bc987582ff68d0471061ae44030 Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 28 13:55:59 2012 +0200 r600g: don't decompress depth or stencil if there isn't any commit 43e226b6efb77db2247741cc2057d9625a2cfa05 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:32:50 2012 +0200 r600g: optimize uploading depth textures Make it only copy the portion of a depth texture being uploaded and not the whole 2D layer. There is also a little code cleanup. commit b242adbe5cfa165b252064a1ea36f802d8251ef1 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:17:46 2012 +0200 r600g: remove needless wrapper r600_texture_depth_flush commit 611dd529425281d73f1f0ad2000362d4a5525a25 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:05:14 2012 +0200 r600g: init_flushed_depth_texture should be able to report errors commit 80755ff56317446a8c89e611edc1fdf320d6779b Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 14 17:06:27 2012 +0200 r600g: properly track which textures are depth This fixes the issue with have_depth_texture never being set to false. commit fe1fd675565231b49d3ac53d0b4bec39d8bc6781 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 03:10:37 2012 +0200 r600g: don't flush depth textures set as colorbuffers 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. commit 5a17d8318ec2c20bf86275044dc8f715105a88e7 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 02:14:18 2012 +0200 r600g: flush depth textures bound to vertex shaders This was missing/broken. There are also minor code cleanups. commit dee58f94af833906863b0ff2955b20f3ab407e63 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 01:54:24 2012 +0200 r600g: do fine-grained depth texture flushing - 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) Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 18:36:41 +00:00
if (!tex->is_depth || tex->is_flushing_texture)
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
continue;
radeonsi: More assorted depth/stencil changes ported from r600g. [ Squashed port of the following r600g commits: - Michel Dänzer ] commit 428e37c2da420f7dc14a2ea265f2387270f9bee1 Author: Marek Olšák <maraeo@gmail.com> Date: Tue Oct 2 22:02:54 2012 +0200 r600g: add in-place DB decompression and texturing with DB tiling The decompression is done in-place and only the compressed tiles are decompressed. Note: R6xx-R7xx can do that only with Z16 and Z32F. The texture unit is programmed to use non-displayable tiling and depth ordering of samples, so that it can fetch the texture in the native DB format. The latest version of the libdrm surface allocator is required for stencil texturing to work. The old one didn't create the mipmap tree correctly. We need a separate mipmap tree for stencil, because the stencil mipmap offsets are not really depth offsets/4. There are still some known bugs, but this should save some memory and it also improves performance a little bit in Lightsmark (especially with low resolutions; tested with Radeon HD 5000). The DB->CB copy is still used for transfers. commit e2f623f1d6da9bc987582ff68d0471061ae44030 Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 28 13:55:59 2012 +0200 r600g: don't decompress depth or stencil if there isn't any commit 43e226b6efb77db2247741cc2057d9625a2cfa05 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:32:50 2012 +0200 r600g: optimize uploading depth textures Make it only copy the portion of a depth texture being uploaded and not the whole 2D layer. There is also a little code cleanup. commit b242adbe5cfa165b252064a1ea36f802d8251ef1 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:17:46 2012 +0200 r600g: remove needless wrapper r600_texture_depth_flush commit 611dd529425281d73f1f0ad2000362d4a5525a25 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:05:14 2012 +0200 r600g: init_flushed_depth_texture should be able to report errors commit 80755ff56317446a8c89e611edc1fdf320d6779b Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 14 17:06:27 2012 +0200 r600g: properly track which textures are depth This fixes the issue with have_depth_texture never being set to false. commit fe1fd675565231b49d3ac53d0b4bec39d8bc6781 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 03:10:37 2012 +0200 r600g: don't flush depth textures set as colorbuffers 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. commit 5a17d8318ec2c20bf86275044dc8f715105a88e7 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 02:14:18 2012 +0200 r600g: flush depth textures bound to vertex shaders This was missing/broken. There are also minor code cleanups. commit dee58f94af833906863b0ff2955b20f3ab407e63 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 01:54:24 2012 +0200 r600g: do fine-grained depth texture flushing - 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) Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 18:36:41 +00:00
si_blit_decompress_depth_in_place(rctx, tex,
view->u.tex.first_level, view->u.tex.last_level,
0, util_max_layer(&tex->resource.b.b, view->u.tex.first_level));
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
}
}
static void r600_clear(struct pipe_context *ctx, unsigned buffers,
const union pipe_color_union *color,
double depth, unsigned stencil)
{
struct r600_context *rctx = (struct r600_context *)ctx;
struct pipe_framebuffer_state *fb = &rctx->framebuffer;
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
r600_blitter_begin(ctx, R600_CLEAR);
util_blitter_clear(rctx->blitter, fb->width, fb->height,
buffers, color, depth, stencil);
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
r600_blitter_end(ctx);
}
static void r600_clear_render_target(struct pipe_context *ctx,
struct pipe_surface *dst,
const union pipe_color_union *color,
unsigned dstx, unsigned dsty,
unsigned width, unsigned height)
{
struct r600_context *rctx = (struct r600_context *)ctx;
r600_blitter_begin(ctx, R600_CLEAR_SURFACE);
util_blitter_clear_render_target(rctx->blitter, dst, color,
dstx, dsty, width, height);
r600_blitter_end(ctx);
}
static void r600_clear_depth_stencil(struct pipe_context *ctx,
struct pipe_surface *dst,
unsigned clear_flags,
double depth,
unsigned stencil,
unsigned dstx, unsigned dsty,
unsigned width, unsigned height)
{
struct r600_context *rctx = (struct r600_context *)ctx;
r600_blitter_begin(ctx, R600_CLEAR_SURFACE);
util_blitter_clear_depth_stencil(rctx->blitter, dst, clear_flags, depth, stencil,
dstx, dsty, width, height);
r600_blitter_end(ctx);
}
struct texture_orig_info {
unsigned format;
unsigned width0;
unsigned height0;
unsigned npix_x;
unsigned npix_y;
unsigned npix0_x;
unsigned npix0_y;
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
};
static void r600_compressed_to_blittable(struct pipe_resource *tex,
unsigned level,
struct texture_orig_info *orig)
{
struct r600_texture *rtex = (struct r600_texture*)tex;
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
unsigned pixsize = util_format_get_blocksize(rtex->real_format);
int new_format;
int new_height, new_width;
orig->format = tex->format;
orig->width0 = tex->width0;
orig->height0 = tex->height0;
orig->npix0_x = rtex->surface.level[0].npix_x;
orig->npix0_y = rtex->surface.level[0].npix_y;
orig->npix_x = rtex->surface.level[level].npix_x;
orig->npix_y = rtex->surface.level[level].npix_y;
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
if (pixsize == 8)
new_format = PIPE_FORMAT_R16G16B16A16_UINT; /* 64-bit block */
else
new_format = PIPE_FORMAT_R32G32B32A32_UINT; /* 128-bit block */
new_width = util_format_get_nblocksx(tex->format, orig->width0);
new_height = util_format_get_nblocksy(tex->format, orig->height0);
tex->width0 = new_width;
tex->height0 = new_height;
tex->format = new_format;
rtex->surface.level[0].npix_x = util_format_get_nblocksx(orig->format, orig->npix0_x);
rtex->surface.level[0].npix_y = util_format_get_nblocksy(orig->format, orig->npix0_y);
rtex->surface.level[level].npix_x = util_format_get_nblocksx(orig->format, orig->npix_x);
rtex->surface.level[level].npix_y = util_format_get_nblocksy(orig->format, orig->npix_y);
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
}
static void r600_change_format(struct pipe_resource *tex,
unsigned level,
struct texture_orig_info *orig,
enum pipe_format format)
{
struct r600_texture *rtex = (struct r600_texture*)tex;
orig->format = tex->format;
orig->width0 = tex->width0;
orig->height0 = tex->height0;
orig->npix0_x = rtex->surface.level[0].npix_x;
orig->npix0_y = rtex->surface.level[0].npix_y;
orig->npix_x = rtex->surface.level[level].npix_x;
orig->npix_y = rtex->surface.level[level].npix_y;
tex->format = format;
}
static void r600_reset_blittable_to_orig(struct pipe_resource *tex,
unsigned level,
struct texture_orig_info *orig)
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
{
struct r600_texture *rtex = (struct r600_texture*)tex;
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
tex->format = orig->format;
tex->width0 = orig->width0;
tex->height0 = orig->height0;
rtex->surface.level[0].npix_x = orig->npix0_x;
rtex->surface.level[0].npix_y = orig->npix0_y;
rtex->surface.level[level].npix_x = orig->npix_x;
rtex->surface.level[level].npix_y = orig->npix_y;
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
}
static void r600_resource_copy_region(struct pipe_context *ctx,
struct pipe_resource *dst,
unsigned dst_level,
unsigned dstx, unsigned dsty, unsigned dstz,
struct pipe_resource *src,
unsigned src_level,
const struct pipe_box *src_box)
{
struct r600_context *rctx = (struct r600_context *)ctx;
struct r600_texture *rsrc = (struct r600_texture*)src;
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
struct texture_orig_info orig_info[2];
struct pipe_box sbox;
const struct pipe_box *psbox = src_box;
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
boolean restore_orig[2];
memset(orig_info, 0, sizeof(orig_info));
/* Fallback for buffers. */
if (dst->target == PIPE_BUFFER && src->target == PIPE_BUFFER) {
util_resource_copy_region(ctx, dst, dst_level, dstx, dsty, dstz,
src, src_level, src_box);
return;
}
radeonsi: More assorted depth/stencil changes ported from r600g. [ Squashed port of the following r600g commits: - Michel Dänzer ] commit 428e37c2da420f7dc14a2ea265f2387270f9bee1 Author: Marek Olšák <maraeo@gmail.com> Date: Tue Oct 2 22:02:54 2012 +0200 r600g: add in-place DB decompression and texturing with DB tiling The decompression is done in-place and only the compressed tiles are decompressed. Note: R6xx-R7xx can do that only with Z16 and Z32F. The texture unit is programmed to use non-displayable tiling and depth ordering of samples, so that it can fetch the texture in the native DB format. The latest version of the libdrm surface allocator is required for stencil texturing to work. The old one didn't create the mipmap tree correctly. We need a separate mipmap tree for stencil, because the stencil mipmap offsets are not really depth offsets/4. There are still some known bugs, but this should save some memory and it also improves performance a little bit in Lightsmark (especially with low resolutions; tested with Radeon HD 5000). The DB->CB copy is still used for transfers. commit e2f623f1d6da9bc987582ff68d0471061ae44030 Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 28 13:55:59 2012 +0200 r600g: don't decompress depth or stencil if there isn't any commit 43e226b6efb77db2247741cc2057d9625a2cfa05 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:32:50 2012 +0200 r600g: optimize uploading depth textures Make it only copy the portion of a depth texture being uploaded and not the whole 2D layer. There is also a little code cleanup. commit b242adbe5cfa165b252064a1ea36f802d8251ef1 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:17:46 2012 +0200 r600g: remove needless wrapper r600_texture_depth_flush commit 611dd529425281d73f1f0ad2000362d4a5525a25 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:05:14 2012 +0200 r600g: init_flushed_depth_texture should be able to report errors commit 80755ff56317446a8c89e611edc1fdf320d6779b Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 14 17:06:27 2012 +0200 r600g: properly track which textures are depth This fixes the issue with have_depth_texture never being set to false. commit fe1fd675565231b49d3ac53d0b4bec39d8bc6781 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 03:10:37 2012 +0200 r600g: don't flush depth textures set as colorbuffers 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. commit 5a17d8318ec2c20bf86275044dc8f715105a88e7 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 02:14:18 2012 +0200 r600g: flush depth textures bound to vertex shaders This was missing/broken. There are also minor code cleanups. commit dee58f94af833906863b0ff2955b20f3ab407e63 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 01:54:24 2012 +0200 r600g: do fine-grained depth texture flushing - 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) Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 18:36:41 +00:00
/* This must be done before entering u_blitter to avoid recursion. */
if (rsrc->is_depth && !rsrc->is_flushing_texture) {
si_blit_decompress_depth_in_place(rctx, rsrc,
src_level, src_level,
src_box->z, src_box->z + src_box->depth - 1);
}
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
restore_orig[0] = restore_orig[1] = FALSE;
if (util_format_is_compressed(src->format) &&
util_format_is_compressed(dst->format)) {
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
r600_compressed_to_blittable(src, src_level, &orig_info[0]);
restore_orig[0] = TRUE;
sbox.x = util_format_get_nblocksx(orig_info[0].format, src_box->x);
sbox.y = util_format_get_nblocksy(orig_info[0].format, src_box->y);
sbox.z = src_box->z;
sbox.width = util_format_get_nblocksx(orig_info[0].format, src_box->width);
sbox.height = util_format_get_nblocksy(orig_info[0].format, src_box->height);
sbox.depth = src_box->depth;
psbox=&sbox;
r600_compressed_to_blittable(dst, dst_level, &orig_info[1]);
restore_orig[1] = TRUE;
/* translate the dst box as well */
dstx = util_format_get_nblocksx(orig_info[1].format, dstx);
dsty = util_format_get_nblocksy(orig_info[1].format, dsty);
} else if (!util_blitter_is_copy_supported(rctx->blitter, dst, src,
PIPE_MASK_RGBAZS)) {
unsigned blocksize = util_format_get_blocksize(src->format);
switch (blocksize) {
case 1:
r600_change_format(src, src_level, &orig_info[0],
PIPE_FORMAT_R8_UNORM);
r600_change_format(dst, dst_level, &orig_info[1],
PIPE_FORMAT_R8_UNORM);
break;
case 2:
r600_change_format(src, src_level, &orig_info[0],
PIPE_FORMAT_R8G8_UNORM);
r600_change_format(dst, dst_level, &orig_info[1],
PIPE_FORMAT_R8G8_UNORM);
break;
case 4:
r600_change_format(src, src_level, &orig_info[0],
PIPE_FORMAT_R8G8B8A8_UNORM);
r600_change_format(dst, dst_level, &orig_info[1],
PIPE_FORMAT_R8G8B8A8_UNORM);
break;
case 8:
r600_change_format(src, src_level, &orig_info[0],
PIPE_FORMAT_R16G16B16A16_UINT);
r600_change_format(dst, dst_level, &orig_info[1],
PIPE_FORMAT_R16G16B16A16_UINT);
break;
case 16:
r600_change_format(src, src_level, &orig_info[0],
PIPE_FORMAT_R32G32B32A32_UINT);
r600_change_format(dst, dst_level, &orig_info[1],
PIPE_FORMAT_R32G32B32A32_UINT);
break;
default:
fprintf(stderr, "Unhandled format %s with blocksize %u\n",
util_format_short_name(src->format), blocksize);
assert(0);
}
restore_orig[0] = TRUE;
restore_orig[1] = TRUE;
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
}
r600_blitter_begin(ctx, R600_COPY);
util_blitter_copy_texture(rctx->blitter, dst, dst_level, dstx, dsty, dstz,
src, src_level, psbox, PIPE_MASK_RGBAZS, TRUE);
r600_blitter_end(ctx);
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
if (restore_orig[0])
r600_reset_blittable_to_orig(src, src_level, &orig_info[0]);
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
if (restore_orig[1])
r600_reset_blittable_to_orig(dst, dst_level, &orig_info[1]);
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
}
static boolean is_simple_msaa_resolve(const struct pipe_blit_info *info)
{
unsigned dst_width = u_minify(info->dst.resource->width0, info->dst.level);
unsigned dst_height = u_minify(info->dst.resource->height0, info->dst.level);
struct r600_texture *dst = (struct r600_texture*)info->dst.resource;
unsigned dst_tile_mode = dst->surface.level[info->dst.level].mode;
bool dst_is_scanout = (dst->surface.flags & RADEON_SURF_SCANOUT) != 0;
return info->dst.resource->format == info->src.resource->format &&
info->dst.resource->format == info->dst.format &&
info->src.resource->format == info->src.format &&
!info->scissor_enable &&
info->mask == PIPE_MASK_RGBA &&
dst_width == info->src.resource->width0 &&
dst_height == info->src.resource->height0 &&
info->dst.box.x == 0 &&
info->dst.box.y == 0 &&
info->dst.box.width == dst_width &&
info->dst.box.height == dst_height &&
info->src.box.x == 0 &&
info->src.box.y == 0 &&
info->src.box.width == dst_width &&
info->src.box.height == dst_height &&
/* Dst must be tiled. If it's not, we have to use a temporary
* resource which is tiled. */
dst_tile_mode >= RADEON_SURF_MODE_1D &&
!dst_is_scanout;
}
/* For MSAA integer resolving to work, we change the format to NORM using this function. */
static enum pipe_format int_to_norm_format(enum pipe_format format)
{
switch (format) {
#define REPLACE_FORMAT_SIGN(format,sign) \
case PIPE_FORMAT_##format##_##sign##INT: \
return PIPE_FORMAT_##format##_##sign##NORM
#define REPLACE_FORMAT(format) \
REPLACE_FORMAT_SIGN(format, U); \
REPLACE_FORMAT_SIGN(format, S)
REPLACE_FORMAT_SIGN(B10G10R10A2, U);
REPLACE_FORMAT(R8);
REPLACE_FORMAT(R8G8);
REPLACE_FORMAT(R8G8B8X8);
REPLACE_FORMAT(R8G8B8A8);
REPLACE_FORMAT(A8);
REPLACE_FORMAT(I8);
REPLACE_FORMAT(L8);
REPLACE_FORMAT(L8A8);
REPLACE_FORMAT(R16);
REPLACE_FORMAT(R16G16);
REPLACE_FORMAT(R16G16B16X16);
REPLACE_FORMAT(R16G16B16A16);
REPLACE_FORMAT(A16);
REPLACE_FORMAT(I16);
REPLACE_FORMAT(L16);
REPLACE_FORMAT(L16A16);
#undef REPLACE_FORMAT
#undef REPLACE_FORMAT_SIGN
default:
return format;
}
}
static void si_msaa_color_resolve(struct pipe_context *ctx,
const struct pipe_blit_info *info)
{
struct r600_context *rctx = (struct r600_context *)ctx;
struct pipe_screen *screen = ctx->screen;
struct pipe_resource *tmp, templ;
struct pipe_blit_info blit;
unsigned sample_mask = ~0;
assert(info->src.level == 0);
assert(info->src.box.depth == 1);
assert(info->dst.box.depth == 1);
if (is_simple_msaa_resolve(info)) {
r600_blitter_begin(ctx, R600_COLOR_RESOLVE);
util_blitter_custom_resolve_color(rctx->blitter,
info->dst.resource, info->dst.level,
info->dst.box.z,
info->src.resource, info->src.box.z,
sample_mask, rctx->custom_blend_resolve,
int_to_norm_format(info->dst.format));
r600_blitter_end(ctx);
return;
}
/* resolve into a temporary texture, then blit */
templ.target = PIPE_TEXTURE_2D;
templ.format = info->src.resource->format;
templ.width0 = info->src.resource->width0;
templ.height0 = info->src.resource->height0;
templ.depth0 = 1;
templ.array_size = 1;
templ.last_level = 0;
templ.nr_samples = 0;
templ.usage = PIPE_USAGE_STATIC;
templ.bind = PIPE_BIND_RENDER_TARGET | PIPE_BIND_SAMPLER_VIEW;
templ.flags = R600_RESOURCE_FLAG_FORCE_TILING; /* dst must not have a linear layout */
tmp = screen->resource_create(screen, &templ);
/* resolve */
r600_blitter_begin(ctx, R600_COLOR_RESOLVE);
util_blitter_custom_resolve_color(rctx->blitter,
tmp, 0, 0,
info->src.resource, info->src.box.z,
sample_mask, rctx->custom_blend_resolve,
int_to_norm_format(tmp->format));
r600_blitter_end(ctx);
/* blit */
blit = *info;
blit.src.resource = tmp;
blit.src.box.z = 0;
r600_blitter_begin(ctx, R600_BLIT);
util_blitter_blit(rctx->blitter, &blit);
r600_blitter_end(ctx);
pipe_resource_reference(&tmp, NULL);
}
static void si_blit(struct pipe_context *ctx,
const struct pipe_blit_info *info)
{
struct r600_context *rctx = (struct r600_context*)ctx;
struct r600_texture *rsrc = (struct r600_texture*)info->src.resource;
if (info->src.resource->nr_samples > 1 &&
info->dst.resource->nr_samples <= 1 &&
!util_format_is_depth_or_stencil(info->src.resource->format) &&
!util_format_is_pure_integer(info->src.resource->format)) {
si_msaa_color_resolve(ctx, info);
return;
}
assert(util_blitter_is_blit_supported(rctx->blitter, info));
radeonsi: More assorted depth/stencil changes ported from r600g. [ Squashed port of the following r600g commits: - Michel Dänzer ] commit 428e37c2da420f7dc14a2ea265f2387270f9bee1 Author: Marek Olšák <maraeo@gmail.com> Date: Tue Oct 2 22:02:54 2012 +0200 r600g: add in-place DB decompression and texturing with DB tiling The decompression is done in-place and only the compressed tiles are decompressed. Note: R6xx-R7xx can do that only with Z16 and Z32F. The texture unit is programmed to use non-displayable tiling and depth ordering of samples, so that it can fetch the texture in the native DB format. The latest version of the libdrm surface allocator is required for stencil texturing to work. The old one didn't create the mipmap tree correctly. We need a separate mipmap tree for stencil, because the stencil mipmap offsets are not really depth offsets/4. There are still some known bugs, but this should save some memory and it also improves performance a little bit in Lightsmark (especially with low resolutions; tested with Radeon HD 5000). The DB->CB copy is still used for transfers. commit e2f623f1d6da9bc987582ff68d0471061ae44030 Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 28 13:55:59 2012 +0200 r600g: don't decompress depth or stencil if there isn't any commit 43e226b6efb77db2247741cc2057d9625a2cfa05 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:32:50 2012 +0200 r600g: optimize uploading depth textures Make it only copy the portion of a depth texture being uploaded and not the whole 2D layer. There is also a little code cleanup. commit b242adbe5cfa165b252064a1ea36f802d8251ef1 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:17:46 2012 +0200 r600g: remove needless wrapper r600_texture_depth_flush commit 611dd529425281d73f1f0ad2000362d4a5525a25 Author: Marek Olšák <maraeo@gmail.com> Date: Wed Jul 18 00:05:14 2012 +0200 r600g: init_flushed_depth_texture should be able to report errors commit 80755ff56317446a8c89e611edc1fdf320d6779b Author: Marek Olšák <maraeo@gmail.com> Date: Sat Jul 14 17:06:27 2012 +0200 r600g: properly track which textures are depth This fixes the issue with have_depth_texture never being set to false. commit fe1fd675565231b49d3ac53d0b4bec39d8bc6781 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 03:10:37 2012 +0200 r600g: don't flush depth textures set as colorbuffers 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. commit 5a17d8318ec2c20bf86275044dc8f715105a88e7 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 02:14:18 2012 +0200 r600g: flush depth textures bound to vertex shaders This was missing/broken. There are also minor code cleanups. commit dee58f94af833906863b0ff2955b20f3ab407e63 Author: Marek Olšák <maraeo@gmail.com> Date: Sun Jul 8 01:54:24 2012 +0200 r600g: do fine-grained depth texture flushing - 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) Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 18:36:41 +00:00
if (rsrc->is_depth && !rsrc->is_flushing_texture) {
si_blit_decompress_depth_in_place(rctx, rsrc,
info->src.level, info->src.level,
info->src.box.z,
info->src.box.z + info->src.box.depth - 1);
}
r600_blitter_begin(ctx, R600_BLIT);
util_blitter_blit(rctx->blitter, info);
r600_blitter_end(ctx);
}
void si_init_blit_functions(struct r600_context *rctx)
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
{
rctx->context.clear = r600_clear;
rctx->context.clear_render_target = r600_clear_render_target;
rctx->context.clear_depth_stencil = r600_clear_depth_stencil;
rctx->context.resource_copy_region = r600_resource_copy_region;
rctx->context.blit = si_blit;
radeonsi: initial WIP SI code This commit adds initial support for acceleration on SI chips. egltri is starting to work. The SI/R600 llvm backend is currently included in mesa but that may change in the future. The plan is to write a single gallium driver and use gallium to support X acceleration. This commit contains patches from: Tom Stellard <thomas.stellard@amd.com> Michel Dänzer <michel.daenzer@amd.com> Alex Deucher <alexander.deucher@amd.com> Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <maraeo@gmail.com> Date: Fri Feb 17 01:49:49 2012 +0100 gallium: remove unused winsys pointers in pipe_screen and pipe_context A winsys is already a private object of a driver. ====================================================================== radeonsi: Copy color clamping CAPs from r600 Not sure if the values of these CAPS are correct for radeonsi, but the same changed were made to r600g in commit: commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc Author: Marek Olšák <maraeo@gmail.com> Date: Mon Jan 23 03:11:17 2012 +0100 st/mesa: do vertex and fragment color clamping in shaders For ARB_color_buffer_float. Most hardware can't do it and st/mesa is the perfect place for a fallback. The exceptions are: - r500 (vertex clamp only) - nv50 (both) - nvc0 (both) - softpipe (both) We also have to take into account that r300 can do CLAMPED vertex colors only, while r600 can do UNCLAMPED vertex colors only. The difference can be expressed with the two new CAPs. ====================================================================== radeonsi: Remove PIPE_CAP_OUTPUT_READ This CAP was dropped in commit: commit 04e324008759282728a95a1394bac2c4c2a1a3f9 Author: Marek Olšák <maraeo@gmail.com> Date: Thu Feb 23 23:44:36 2012 +0100 gallium: remove PIPE_SHADER_CAP_OUTPUT_READ r600g is the only driver which has made use of it. The reason the CAP was added was to fix some piglit tests when the GLSL pass lower_output_reads didn't exist. However, not removing output reads breaks the fallback for glClampColorARB, which assumes outputs are not readable. The fix would be non-trivial and my personal preference is to remove the CAP, considering that reading outputs is uncommon and that we can now use lower_output_reads to fix the issue that the CAP was supposed to workaround in the first place. ====================================================================== radeonsi: Add missing parameters to rws->buffer_get_tiling() call This was changed in commit: commit c0c979eebc076b95cc8d18a013ce2968fe6311ad Author: Jerome Glisse <jglisse@redhat.com> Date: Mon Jan 30 17:22:13 2012 -0500 r600g: add support for common surface allocator for tiling v13 Tiled surface have all kind of alignment constraint that needs to be met. Instead of having all this code duplicated btw ddx and mesa use common code in libdrm_radeon this also ensure that both ddx and mesa compute those alignment in the same way. v2 fix evergreen v3 fix compressed texture and workaround cube texture issue by disabling 2D array mode for cubemap (need to check if r7xx and newer are also affected by the issue) v4 fix texture array v5 fix evergreen and newer, split surface values computation from mipmap tree generation so that we can get them directly from the ddx v6 final fix to evergreen tile split value v7 fix mipmap offset to avoid to use random value, use color view depth view to address different layer as hardware is doing some magic rotation depending on the layer v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on evergreen, align bytes per pixel to a multiple of a dword v9 fix handling of stencil on evergreen, half fix for compressed texture v10 fix evergreen compressed texture proper support for stencil tile split. Fix stencil issue when array mode was clear by the kernel, always program stencil bo. On evergreen depth buffer bo need to be big enough to hold depth buffer + stencil buffer as even with stencil disabled things get written there. v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen, old ddx overestimate those. Fix linear case when pitch*height < 64. Fix r300g. v12 Fix linear case when pitch*height < 64 for old path, adapt to libdrm API change v13 add libdrm check Signed-off-by: Jerome Glisse <jglisse@redhat.com> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <maraeo@gmail.com> Date: Mon Mar 5 13:45:00 2012 +0100 Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY" This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc. It was decided to refactor the transfer API instead of adding workarounds to address the performance issues. ====================================================================== radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT. Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90. ====================================================================== radeonsi: nuke the fallback for vertex and fragment color clamping Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853. ====================================================================== radeonsi: don't expose transform_feedback2 without kernel support Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48. ====================================================================== radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL. Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f. ====================================================================== radeonsi: set minimum point size to 1.0 for non-sprite non-aa points. Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc. ====================================================================== radeonsi: rework and consolidate stencilref state setting. Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070. ====================================================================== radeonsi: cleanup setting DB_SHADER_CONTROL. Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b. ====================================================================== radeonsi: Get rid of register masks. Ported from r600g commits 3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2. ====================================================================== radeonsi: get rid of r600_context_reg. Ported from r600g commits 9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f. ====================================================================== radeonsi: Fix regression from 'Get rid of register masks'. ====================================================================== radeonsi: optimize r600_resource_va. Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174. ====================================================================== radeonsi: remove u8,u16,u32,u64 types. Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5. ====================================================================== radeonsi: merge r600_context with r600_pipe_context. Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0. ====================================================================== radeonsi: Miscellaneous context cleanups. Ported from r600g commits e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888. ====================================================================== radeonsi: add a new simple API for state emission. Ported from r600g commits 621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e. ====================================================================== radeonsi: Also remove sbu_flags member of struct r600_reg. Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions, so some code needs to be disabled for now. ====================================================================== radeonsi: Miscellaneous simplifications. Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and b0337b679ad4c2feae59215104cfa60b58a619d5. ====================================================================== radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION. Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a. ====================================================================== radeonsi: Use a fake reloc to sleep for fences. Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4. ====================================================================== radeonsi: adapt to get_query_result interface change. Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.
2012-01-06 22:38:37 +00:00
}