summaryrefslogtreecommitdiffstats
path: root/src/gallium/drivers/radeonsi
diff options
context:
space:
mode:
authorTom Stellard <[email protected]>2012-01-06 17:38:37 -0500
committerTom Stellard <[email protected]>2012-04-13 10:32:06 -0400
commita75c6163e605f35b14f26930dd9227e4f337ec9e (patch)
tree0263219cbab9282896f874060bb03d445c4de891 /src/gallium/drivers/radeonsi
parente55cf4854d594eae9ac3f6abd24f4e616eea894f (diff)
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 <[email protected]> Michel Dänzer <[email protected]> Alex Deucher <[email protected]> Vadim Girlin <[email protected]> Signed-off-by: Alex Deucher <[email protected]> The following commits were squashed in: ====================================================================== radeonsi: Remove unused winsys pointer This was removed from r600g in commit: commit 96d882939d612fcc8332f107befec470ed4359de Author: Marek Olšák <[email protected]> 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 <[email protected]> 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 <[email protected]> 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 <[email protected]> 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 <[email protected]> ====================================================================== radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY This was removed in commit: commit 62f44f670bb0162e89fd4786af877f8da9ff607c Author: Marek Olšák <[email protected]> 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.
Diffstat (limited to 'src/gallium/drivers/radeonsi')
-rw-r--r--src/gallium/drivers/radeonsi/Android.mk38
-rw-r--r--src/gallium/drivers/radeonsi/Makefile24
-rw-r--r--src/gallium/drivers/radeonsi/Makefile.sources13
-rw-r--r--src/gallium/drivers/radeonsi/SConscript17
-rw-r--r--src/gallium/drivers/radeonsi/evergreen_hw_context.c561
-rw-r--r--src/gallium/drivers/radeonsi/evergreen_state.c2169
-rw-r--r--src/gallium/drivers/radeonsi/r600.h245
-rw-r--r--src/gallium/drivers/radeonsi/r600_blit.c379
-rw-r--r--src/gallium/drivers/radeonsi/r600_buffer.c282
-rw-r--r--src/gallium/drivers/radeonsi/r600_hw_context.c1151
-rw-r--r--src/gallium/drivers/radeonsi/r600_hw_context_priv.h76
-rw-r--r--src/gallium/drivers/radeonsi/r600_query.c130
-rw-r--r--src/gallium/drivers/radeonsi/r600_resource.c64
-rw-r--r--src/gallium/drivers/radeonsi/r600_resource.h105
-rw-r--r--src/gallium/drivers/radeonsi/r600_state_common.c899
-rw-r--r--src/gallium/drivers/radeonsi/r600_texture.c825
-rw-r--r--src/gallium/drivers/radeonsi/r600_translate.c54
-rw-r--r--src/gallium/drivers/radeonsi/radeonsi_pipe.c731
-rw-r--r--src/gallium/drivers/radeonsi/radeonsi_pipe.h490
-rw-r--r--src/gallium/drivers/radeonsi/radeonsi_public.h30
-rw-r--r--src/gallium/drivers/radeonsi/radeonsi_shader.c565
-rw-r--r--src/gallium/drivers/radeonsi/radeonsi_shader.h4
-rw-r--r--src/gallium/drivers/radeonsi/sid.h7668
23 files changed, 16520 insertions, 0 deletions
diff --git a/src/gallium/drivers/radeonsi/Android.mk b/src/gallium/drivers/radeonsi/Android.mk
new file mode 100644
index 00000000000..f7e01a3f658
--- /dev/null
+++ b/src/gallium/drivers/radeonsi/Android.mk
@@ -0,0 +1,38 @@
+# Mesa 3-D graphics library
+#
+# Copyright (C) 2010-2011 Chia-I Wu <[email protected]>
+# Copyright (C) 2010-2011 LunarG Inc.
+#
+# 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
+# the rights to use, copy, modify, merge, publish, distribute, sublicense,
+# 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 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 NONINFRINGEMENT. IN NO EVENT SHALL
+# THE AUTHORS OR COPYRIGHT HOLDERS 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.
+
+LOCAL_PATH := $(call my-dir)
+
+# get C_SOURCES
+include $(LOCAL_PATH)/Makefile.sources
+
+include $(CLEAR_VARS)
+
+LOCAL_SRC_FILES := $(C_SOURCES)
+
+LOCAL_C_INCLUDES :=
+
+LOCAL_MODULE := libmesa_pipe_radeonsi
+
+include $(GALLIUM_COMMON_MK)
+include $(BUILD_STATIC_LIBRARY)
diff --git a/src/gallium/drivers/radeonsi/Makefile b/src/gallium/drivers/radeonsi/Makefile
new file mode 100644
index 00000000000..90f6f472730
--- /dev/null
+++ b/src/gallium/drivers/radeonsi/Makefile
@@ -0,0 +1,24 @@
+TOP = ../../../..
+include $(TOP)/configs/current
+
+LIBNAME = radeonsi
+
+LIBRARY_INCLUDES = \
+ -I$(TOP)/include \
+ -I$(TOP)/src/gallium/drivers/radeon/
+
+
+# get C_SOURCES
+include Makefile.sources
+
+LIBRADEON = $(TOP)/src/gallium/drivers/radeon/libradeon.a
+
+EXTRA_OBJECTS = \
+ $(LIBRADEON)
+
+CFLAGS+=$(RADEON_CFLAGS)
+
+include ../../Makefile.template
+
+# FIXME: Remove when this driver is converted to automake.
+all: default
diff --git a/src/gallium/drivers/radeonsi/Makefile.sources b/src/gallium/drivers/radeonsi/Makefile.sources
new file mode 100644
index 00000000000..394cfe93e07
--- /dev/null
+++ b/src/gallium/drivers/radeonsi/Makefile.sources
@@ -0,0 +1,13 @@
+C_SOURCES := \
+ r600_blit.c \
+ r600_buffer.c \
+ r600_hw_context.c \
+ radeonsi_pipe.c \
+ r600_query.c \
+ r600_resource.c \
+ radeonsi_shader.c \
+ r600_texture.c \
+ evergreen_hw_context.c \
+ evergreen_state.c \
+ r600_translate.c \
+ r600_state_common.c
diff --git a/src/gallium/drivers/radeonsi/SConscript b/src/gallium/drivers/radeonsi/SConscript
new file mode 100644
index 00000000000..f2d2bec6e42
--- /dev/null
+++ b/src/gallium/drivers/radeonsi/SConscript
@@ -0,0 +1,17 @@
+Import('*')
+
+env = env.Clone()
+
+env.Append(CPPPATH = [
+ '#/include',
+ '#/src/gallium/drivers/radeon',
+])
+
+radeonsi = env.ConvenienceLibrary(
+ target = 'radeonsi',
+ source = env.ParseSourceList('Makefile.sources', 'C_SOURCES')
+ )
+
+env.Alias('radeonsi', radeonsi)
+
+Export('radeonsi')
diff --git a/src/gallium/drivers/radeonsi/evergreen_hw_context.c b/src/gallium/drivers/radeonsi/evergreen_hw_context.c
new file mode 100644
index 00000000000..549673f4a0b
--- /dev/null
+++ b/src/gallium/drivers/radeonsi/evergreen_hw_context.c
@@ -0,0 +1,561 @@
+/*
+ * Copyright 2010 Jerome Glisse <[email protected]>
+ *
+ * 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.
+ *
+ * Authors:
+ * Jerome Glisse
+ */
+#include "r600.h"
+#include "r600_hw_context_priv.h"
+#include "radeonsi_pipe.h"
+#include "sid.h"
+#include "util/u_memory.h"
+#include <errno.h>
+
+#define GROUP_FORCE_NEW_BLOCK 0
+
+static const struct r600_reg si_config_reg_list[] = {
+ {R_0088B0_VGT_VTX_VECT_EJECT_REG, REG_FLAG_FLUSH_CHANGE},
+ {R_0088C8_VGT_ESGS_RING_SIZE, REG_FLAG_FLUSH_CHANGE},
+ {R_0088CC_VGT_GSVS_RING_SIZE, REG_FLAG_FLUSH_CHANGE},
+ {R_008958_VGT_PRIMITIVE_TYPE, 0},
+ {R_008A14_PA_CL_ENHANCE, REG_FLAG_FLUSH_CHANGE},
+ {R_009100_SPI_CONFIG_CNTL, REG_FLAG_ENABLE_ALWAYS | REG_FLAG_FLUSH_CHANGE},
+ {R_00913C_SPI_CONFIG_CNTL_1, REG_FLAG_ENABLE_ALWAYS | REG_FLAG_FLUSH_CHANGE},
+};
+
+static const struct r600_reg si_context_reg_list[] = {
+ {R_028000_DB_RENDER_CONTROL, 0},
+ {R_028004_DB_COUNT_CONTROL, 0},
+ {R_028008_DB_DEPTH_VIEW, 0},
+ {R_02800C_DB_RENDER_OVERRIDE, 0},
+ {R_028010_DB_RENDER_OVERRIDE2, 0},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_028014_DB_HTILE_DATA_BASE, REG_FLAG_NEED_BO},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_028020_DB_DEPTH_BOUNDS_MIN, 0},
+ {R_028024_DB_DEPTH_BOUNDS_MAX, 0},
+ {R_028028_DB_STENCIL_CLEAR, 0},
+ {R_02802C_DB_DEPTH_CLEAR, 0},
+ {R_028030_PA_SC_SCREEN_SCISSOR_TL, 0},
+ {R_028034_PA_SC_SCREEN_SCISSOR_BR, 0},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_02803C_DB_DEPTH_INFO, 0},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_028040_DB_Z_INFO, 0},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_028044_DB_STENCIL_INFO, 0},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_028048_DB_Z_READ_BASE, REG_FLAG_NEED_BO},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_02804C_DB_STENCIL_READ_BASE, REG_FLAG_NEED_BO},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_028050_DB_Z_WRITE_BASE, REG_FLAG_NEED_BO},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_028054_DB_STENCIL_WRITE_BASE, REG_FLAG_NEED_BO},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_028058_DB_DEPTH_SIZE, 0},
+ {R_02805C_DB_DEPTH_SLICE, 0},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_028080_TA_BC_BASE_ADDR, REG_FLAG_NEED_BO},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_028200_PA_SC_WINDOW_OFFSET, 0},
+ {R_028204_PA_SC_WINDOW_SCISSOR_TL, 0},
+ {R_028208_PA_SC_WINDOW_SCISSOR_BR, 0},
+ {R_02820C_PA_SC_CLIPRECT_RULE, 0},
+ {R_028210_PA_SC_CLIPRECT_0_TL, 0},
+ {R_028214_PA_SC_CLIPRECT_0_BR, 0},
+ {R_028218_PA_SC_CLIPRECT_1_TL, 0},
+ {R_02821C_PA_SC_CLIPRECT_1_BR, 0},
+ {R_028220_PA_SC_CLIPRECT_2_TL, 0},
+ {R_028224_PA_SC_CLIPRECT_2_BR, 0},
+ {R_028228_PA_SC_CLIPRECT_3_TL, 0},
+ {R_02822C_PA_SC_CLIPRECT_3_BR, 0},
+ {R_028230_PA_SC_EDGERULE, 0},
+ {R_028234_PA_SU_HARDWARE_SCREEN_OFFSET, 0},
+ {R_028238_CB_TARGET_MASK, 0},
+ {R_02823C_CB_SHADER_MASK, 0},
+ {R_028240_PA_SC_GENERIC_SCISSOR_TL, 0},
+ {R_028244_PA_SC_GENERIC_SCISSOR_BR, 0},
+ {R_028250_PA_SC_VPORT_SCISSOR_0_TL, 0},
+ {R_028254_PA_SC_VPORT_SCISSOR_0_BR, 0},
+ {R_0282D0_PA_SC_VPORT_ZMIN_0, 0},
+ {R_0282D4_PA_SC_VPORT_ZMAX_0, 0},
+ {R_028350_PA_SC_RASTER_CONFIG, 0},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_028400_VGT_MAX_VTX_INDX, 0},
+ {R_028404_VGT_MIN_VTX_INDX, 0},
+ {R_028408_VGT_INDX_OFFSET, 0},
+ {R_02840C_VGT_MULTI_PRIM_IB_RESET_INDX, 0},
+ {R_028A94_VGT_MULTI_PRIM_IB_RESET_EN, 0},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_028414_CB_BLEND_RED, 0},
+ {R_028418_CB_BLEND_GREEN, 0},
+ {R_02841C_CB_BLEND_BLUE, 0},
+ {R_028420_CB_BLEND_ALPHA, 0},
+ {R_028430_DB_STENCILREFMASK, 0},
+ {R_028434_DB_STENCILREFMASK_BF, 0},
+ {R_02843C_PA_CL_VPORT_XSCALE_0, 0},
+ {R_028440_PA_CL_VPORT_XOFFSET_0, 0},
+ {R_028444_PA_CL_VPORT_YSCALE_0, 0},
+ {R_028448_PA_CL_VPORT_YOFFSET_0, 0},
+ {R_02844C_PA_CL_VPORT_ZSCALE_0, 0},
+ {R_028450_PA_CL_VPORT_ZOFFSET_0, 0},
+ {R_0285BC_PA_CL_UCP_0_X, 0},
+ {R_0285C0_PA_CL_UCP_0_Y, 0},
+ {R_0285C4_PA_CL_UCP_0_Z, 0},
+ {R_0285C8_PA_CL_UCP_0_W, 0},
+ {R_0285CC_PA_CL_UCP_1_X, 0},
+ {R_0285D0_PA_CL_UCP_1_Y, 0},
+ {R_0285D4_PA_CL_UCP_1_Z, 0},
+ {R_0285D8_PA_CL_UCP_1_W, 0},
+ {R_0285DC_PA_CL_UCP_2_X, 0},
+ {R_0285E0_PA_CL_UCP_2_Y, 0},
+ {R_0285E4_PA_CL_UCP_2_Z, 0},
+ {R_0285E8_PA_CL_UCP_2_W, 0},
+ {R_0285EC_PA_CL_UCP_3_X, 0},
+ {R_0285F0_PA_CL_UCP_3_Y, 0},
+ {R_0285F4_PA_CL_UCP_3_Z, 0},
+ {R_0285F8_PA_CL_UCP_3_W, 0},
+ {R_0285FC_PA_CL_UCP_4_X, 0},
+ {R_028600_PA_CL_UCP_4_Y, 0},
+ {R_028604_PA_CL_UCP_4_Z, 0},
+ {R_028608_PA_CL_UCP_4_W, 0},
+ {R_02860C_PA_CL_UCP_5_X, 0},
+ {R_028610_PA_CL_UCP_5_Y, 0},
+ {R_028614_PA_CL_UCP_5_Z, 0},
+ {R_028618_PA_CL_UCP_5_W, 0},
+ {R_028644_SPI_PS_INPUT_CNTL_0, 0},
+ {R_028648_SPI_PS_INPUT_CNTL_1, 0},
+ {R_02864C_SPI_PS_INPUT_CNTL_2, 0},
+ {R_028650_SPI_PS_INPUT_CNTL_3, 0},
+ {R_028654_SPI_PS_INPUT_CNTL_4, 0},
+ {R_028658_SPI_PS_INPUT_CNTL_5, 0},
+ {R_02865C_SPI_PS_INPUT_CNTL_6, 0},
+ {R_028660_SPI_PS_INPUT_CNTL_7, 0},
+ {R_028664_SPI_PS_INPUT_CNTL_8, 0},
+ {R_028668_SPI_PS_INPUT_CNTL_9, 0},
+ {R_02866C_SPI_PS_INPUT_CNTL_10, 0},
+ {R_028670_SPI_PS_INPUT_CNTL_11, 0},
+ {R_028674_SPI_PS_INPUT_CNTL_12, 0},
+ {R_028678_SPI_PS_INPUT_CNTL_13, 0},
+ {R_02867C_SPI_PS_INPUT_CNTL_14, 0},
+ {R_028680_SPI_PS_INPUT_CNTL_15, 0},
+ {R_028684_SPI_PS_INPUT_CNTL_16, 0},
+ {R_028688_SPI_PS_INPUT_CNTL_17, 0},
+ {R_02868C_SPI_PS_INPUT_CNTL_18, 0},
+ {R_028690_SPI_PS_INPUT_CNTL_19, 0},
+ {R_028694_SPI_PS_INPUT_CNTL_20, 0},
+ {R_028698_SPI_PS_INPUT_CNTL_21, 0},
+ {R_02869C_SPI_PS_INPUT_CNTL_22, 0},
+ {R_0286A0_SPI_PS_INPUT_CNTL_23, 0},
+ {R_0286A4_SPI_PS_INPUT_CNTL_24, 0},
+ {R_0286A8_SPI_PS_INPUT_CNTL_25, 0},
+ {R_0286AC_SPI_PS_INPUT_CNTL_26, 0},
+ {R_0286B0_SPI_PS_INPUT_CNTL_27, 0},
+ {R_0286B4_SPI_PS_INPUT_CNTL_28, 0},
+ {R_0286B8_SPI_PS_INPUT_CNTL_29, 0},
+ {R_0286BC_SPI_PS_INPUT_CNTL_30, 0},
+ {R_0286C0_SPI_PS_INPUT_CNTL_31, 0},
+ {R_0286C4_SPI_VS_OUT_CONFIG, 0},
+ {R_0286CC_SPI_PS_INPUT_ENA, 0},
+ {R_0286D0_SPI_PS_INPUT_ADDR, 0},
+ {R_0286D4_SPI_INTERP_CONTROL_0, 0},
+ {R_0286D8_SPI_PS_IN_CONTROL, 0},
+ {R_0286E0_SPI_BARYC_CNTL, 0},
+ {R_02870C_SPI_SHADER_POS_FORMAT, 0},
+ {R_028710_SPI_SHADER_Z_FORMAT, 0},
+ {R_028714_SPI_SHADER_COL_FORMAT, 0},
+ {R_028780_CB_BLEND0_CONTROL, 0},
+ {R_028784_CB_BLEND1_CONTROL, 0},
+ {R_028788_CB_BLEND2_CONTROL, 0},
+ {R_02878C_CB_BLEND3_CONTROL, 0},
+ {R_028790_CB_BLEND4_CONTROL, 0},
+ {R_028794_CB_BLEND5_CONTROL, 0},
+ {R_028798_CB_BLEND6_CONTROL, 0},
+ {R_02879C_CB_BLEND7_CONTROL, 0},
+ {R_0287D4_PA_CL_POINT_X_RAD, 0},
+ {R_0287D8_PA_CL_POINT_Y_RAD, 0},
+ {R_0287DC_PA_CL_POINT_SIZE, 0},
+ {R_0287E0_PA_CL_POINT_CULL_RAD, 0},
+ {R_028800_DB_DEPTH_CONTROL, 0},
+ {R_028804_DB_EQAA, 0},
+ {R_028808_CB_COLOR_CONTROL, 0},
+ {R_02880C_DB_SHADER_CONTROL, 0},
+ {R_028810_PA_CL_CLIP_CNTL, 0},
+ {R_028814_PA_SU_SC_MODE_CNTL, 0},
+ {R_028818_PA_CL_VTE_CNTL, 0},
+ {R_02881C_PA_CL_VS_OUT_CNTL, 0},
+ {R_028820_PA_CL_NANINF_CNTL, 0},
+ {R_028824_PA_SU_LINE_STIPPLE_CNTL, 0},
+ {R_028828_PA_SU_LINE_STIPPLE_SCALE, 0},
+ {R_02882C_PA_SU_PRIM_FILTER_CNTL, 0},
+ {R_028A00_PA_SU_POINT_SIZE, 0},
+ {R_028A04_PA_SU_POINT_MINMAX, 0},
+ {R_028A08_PA_SU_LINE_CNTL, 0},
+ {R_028A0C_PA_SC_LINE_STIPPLE, 0},
+ {R_028A10_VGT_OUTPUT_PATH_CNTL, 0},
+ {R_028A14_VGT_HOS_CNTL, 0},
+ {R_028A18_VGT_HOS_MAX_TESS_LEVEL, 0},
+ {R_028A1C_VGT_HOS_MIN_TESS_LEVEL, 0},
+ {R_028A20_VGT_HOS_REUSE_DEPTH, 0},
+ {R_028A24_VGT_GROUP_PRIM_TYPE, 0},
+ {R_028A28_VGT_GROUP_FIRST_DECR, 0},
+ {R_028A2C_VGT_GROUP_DECR, 0},
+ {R_028A30_VGT_GROUP_VECT_0_CNTL, 0},
+ {R_028A34_VGT_GROUP_VECT_1_CNTL, 0},
+ {R_028A38_VGT_GROUP_VECT_0_FMT_CNTL, 0},
+ {R_028A3C_VGT_GROUP_VECT_1_FMT_CNTL, 0},
+ {R_028A40_VGT_GS_MODE, 0},
+ {R_028A48_PA_SC_MODE_CNTL_0, 0},
+ {R_028A4C_PA_SC_MODE_CNTL_1, 0},
+ {R_028A50_VGT_ENHANCE, 0},
+ {R_028A54_VGT_GS_PER_ES, 0},
+ {R_028A58_VGT_ES_PER_GS, 0},
+ {R_028A5C_VGT_GS_PER_VS, 0},
+ {R_028A60_VGT_GSVS_RING_OFFSET_1, 0},
+ {R_028A64_VGT_GSVS_RING_OFFSET_2, 0},
+ {R_028A68_VGT_GSVS_RING_OFFSET_3, 0},
+ {R_028A6C_VGT_GS_OUT_PRIM_TYPE, 0},
+ {R_028A70_IA_ENHANCE, 0},
+ {R_028A84_VGT_PRIMITIVEID_EN, 0},
+ {R_028A8C_VGT_PRIMITIVEID_RESET, 0},
+ {R_028AA0_VGT_INSTANCE_STEP_RATE_0, 0},
+ {R_028AA4_VGT_INSTANCE_STEP_RATE_1, 0},
+ {R_028AA8_IA_MULTI_VGT_PARAM, 0},
+ {R_028AAC_VGT_ESGS_RING_ITEMSIZE, 0},
+ {R_028AB0_VGT_GSVS_RING_ITEMSIZE, 0},
+ {R_028AB4_VGT_REUSE_OFF, 0},
+ {R_028AB8_VGT_VTX_CNT_EN, 0},
+ {R_028ABC_DB_HTILE_SURFACE, 0},
+ {R_028AC0_DB_SRESULTS_COMPARE_STATE0, 0},
+ {R_028AC4_DB_SRESULTS_COMPARE_STATE1, 0},
+ {R_028AC8_DB_PRELOAD_CONTROL, 0},
+ {R_028B54_VGT_SHADER_STAGES_EN, 0},
+ {R_028B70_DB_ALPHA_TO_MASK, 0},
+ {R_028B78_PA_SU_POLY_OFFSET_DB_FMT_CNTL, 0},
+ {R_028B7C_PA_SU_POLY_OFFSET_CLAMP, 0},
+ {R_028B80_PA_SU_POLY_OFFSET_FRONT_SCALE, 0},
+ {R_028B84_PA_SU_POLY_OFFSET_FRONT_OFFSET, 0},
+ {R_028B88_PA_SU_POLY_OFFSET_BACK_SCALE, 0},
+ {R_028B8C_PA_SU_POLY_OFFSET_BACK_OFFSET, 0},
+ {R_028B94_VGT_STRMOUT_CONFIG, 0},
+ {R_028B98_VGT_STRMOUT_BUFFER_CONFIG, 0},
+ {R_028BD4_PA_SC_CENTROID_PRIORITY_0, 0},
+ {R_028BD8_PA_SC_CENTROID_PRIORITY_1, 0},
+ {R_028BDC_PA_SC_LINE_CNTL, 0},
+ {R_028BE0_PA_SC_AA_CONFIG, 0},
+ {R_028BE4_PA_SU_VTX_CNTL, 0},
+ {R_028BE8_PA_CL_GB_VERT_CLIP_ADJ, 0},
+ {R_028BEC_PA_CL_GB_VERT_DISC_ADJ, 0},
+ {R_028BF0_PA_CL_GB_HORZ_CLIP_ADJ, 0},
+ {R_028BF4_PA_CL_GB_HORZ_DISC_ADJ, 0},
+ {R_028BF8_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_0, 0},
+ {R_028BFC_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_1, 0},
+ {R_028C00_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_2, 0},
+ {R_028C04_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_3, 0},
+ {R_028C08_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_0, 0},
+ {R_028C0C_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_1, 0},
+ {R_028C10_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_2, 0},
+ {R_028C14_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_3, 0},
+ {R_028C18_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_0, 0},
+ {R_028C1C_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_1, 0},
+ {R_028C20_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_2, 0},
+ {R_028C24_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_3, 0},
+ {R_028C28_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_0, 0},
+ {R_028C2C_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_1, 0},
+ {R_028C30_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_2, 0},
+ {R_028C34_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_3, 0},
+ {R_028C38_PA_SC_AA_MASK_X0Y0_X1Y0, 0},
+ {R_028C3C_PA_SC_AA_MASK_X0Y1_X1Y1, 0},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_028C60_CB_COLOR0_BASE, REG_FLAG_NEED_BO},
+ {R_028C64_CB_COLOR0_PITCH, 0},
+ {R_028C68_CB_COLOR0_SLICE, 0},
+ {R_028C6C_CB_COLOR0_VIEW, 0},
+ {R_028C70_CB_COLOR0_INFO, REG_FLAG_NEED_BO},
+ {R_028C74_CB_COLOR0_ATTRIB, REG_FLAG_NEED_BO},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_028C9C_CB_COLOR1_BASE, REG_FLAG_NEED_BO},
+ {R_028CA0_CB_COLOR1_PITCH, 0},
+ {R_028CA4_CB_COLOR1_SLICE, 0},
+ {R_028CA8_CB_COLOR1_VIEW, 0},
+ {R_028CAC_CB_COLOR1_INFO, REG_FLAG_NEED_BO},
+ {R_028CB0_CB_COLOR1_ATTRIB, REG_FLAG_NEED_BO},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_028CD8_CB_COLOR2_BASE, REG_FLAG_NEED_BO},
+ {R_028CDC_CB_COLOR2_PITCH, 0},
+ {R_028CE0_CB_COLOR2_SLICE, 0},
+ {R_028CE4_CB_COLOR2_VIEW, 0},
+ {R_028CE8_CB_COLOR2_INFO, REG_FLAG_NEED_BO},
+ {R_028CEC_CB_COLOR2_ATTRIB, REG_FLAG_NEED_BO},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_028D14_CB_COLOR3_BASE, REG_FLAG_NEED_BO},
+ {R_028D18_CB_COLOR3_PITCH, 0},
+ {R_028D1C_CB_COLOR3_SLICE, 0},
+ {R_028D20_CB_COLOR3_VIEW, 0},
+ {R_028D24_CB_COLOR3_INFO, REG_FLAG_NEED_BO},
+ {R_028D28_CB_COLOR3_ATTRIB, REG_FLAG_NEED_BO},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_028D50_CB_COLOR4_BASE, REG_FLAG_NEED_BO},
+ {R_028D54_CB_COLOR4_PITCH, 0},
+ {R_028D58_CB_COLOR4_SLICE, 0},
+ {R_028D5C_CB_COLOR4_VIEW, 0},
+ {R_028D60_CB_COLOR4_INFO, REG_FLAG_NEED_BO},
+ {R_028D64_CB_COLOR4_ATTRIB, REG_FLAG_NEED_BO},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_028D8C_CB_COLOR5_BASE, REG_FLAG_NEED_BO},
+ {R_028D90_CB_COLOR5_PITCH, 0},
+ {R_028D94_CB_COLOR5_SLICE, 0},
+ {R_028D98_CB_COLOR5_VIEW, 0},
+ {R_028D9C_CB_COLOR5_INFO, REG_FLAG_NEED_BO},
+ {R_028DA0_CB_COLOR5_ATTRIB, REG_FLAG_NEED_BO},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_028DC8_CB_COLOR6_BASE, REG_FLAG_NEED_BO},
+ {R_028DCC_CB_COLOR6_PITCH, 0},
+ {R_028DD0_CB_COLOR6_SLICE, 0},
+ {R_028DD4_CB_COLOR6_VIEW, 0},
+ {R_028DD8_CB_COLOR6_INFO, REG_FLAG_NEED_BO},
+ {R_028DDC_CB_COLOR6_ATTRIB, REG_FLAG_NEED_BO},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_028E04_CB_COLOR7_BASE, REG_FLAG_NEED_BO},
+ {R_028E08_CB_COLOR7_PITCH, 0},
+ {R_028E0C_CB_COLOR7_SLICE, 0},
+ {R_028E10_CB_COLOR7_VIEW, 0},
+ {R_028E14_CB_COLOR7_INFO, REG_FLAG_NEED_BO},
+ {R_028E18_CB_COLOR7_ATTRIB, REG_FLAG_NEED_BO},
+};
+
+static const struct r600_reg si_sh_reg_list[] = {
+ {R_00B020_SPI_SHADER_PGM_LO_PS, REG_FLAG_NEED_BO},
+ {R_00B024_SPI_SHADER_PGM_HI_PS, REG_FLAG_NEED_BO},
+ {R_00B028_SPI_SHADER_PGM_RSRC1_PS, 0},
+ {R_00B02C_SPI_SHADER_PGM_RSRC2_PS, 0},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_00B030_SPI_SHADER_USER_DATA_PS_0, REG_FLAG_NEED_BO},
+ {R_00B034_SPI_SHADER_USER_DATA_PS_1, 0},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_00B038_SPI_SHADER_USER_DATA_PS_2, REG_FLAG_NEED_BO},
+ {R_00B03C_SPI_SHADER_USER_DATA_PS_3, 0},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_00B040_SPI_SHADER_USER_DATA_PS_4, REG_FLAG_NEED_BO},
+ {R_00B044_SPI_SHADER_USER_DATA_PS_5, 0},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_00B048_SPI_SHADER_USER_DATA_PS_6, REG_FLAG_NEED_BO},
+ {R_00B04C_SPI_SHADER_USER_DATA_PS_7, 0},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_00B050_SPI_SHADER_USER_DATA_PS_8, REG_FLAG_NEED_BO},
+ {R_00B054_SPI_SHADER_USER_DATA_PS_9, 0},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_00B058_SPI_SHADER_USER_DATA_PS_10, REG_FLAG_NEED_BO},
+ {R_00B05C_SPI_SHADER_USER_DATA_PS_11, 0},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_00B060_SPI_SHADER_USER_DATA_PS_12, REG_FLAG_NEED_BO},
+ {R_00B064_SPI_SHADER_USER_DATA_PS_13, 0},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_00B068_SPI_SHADER_USER_DATA_PS_14, REG_FLAG_NEED_BO},
+ {R_00B06C_SPI_SHADER_USER_DATA_PS_15, 0},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_00B120_SPI_SHADER_PGM_LO_VS, REG_FLAG_NEED_BO},
+ {R_00B124_SPI_SHADER_PGM_HI_VS, REG_FLAG_NEED_BO},
+ {R_00B128_SPI_SHADER_PGM_RSRC1_VS, 0},
+ {R_00B12C_SPI_SHADER_PGM_RSRC2_VS, 0},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_00B130_SPI_SHADER_USER_DATA_VS_0, REG_FLAG_NEED_BO},
+ {R_00B134_SPI_SHADER_USER_DATA_VS_1, 0},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_00B138_SPI_SHADER_USER_DATA_VS_2, REG_FLAG_NEED_BO},
+ {R_00B13C_SPI_SHADER_USER_DATA_VS_3, 0},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_00B140_SPI_SHADER_USER_DATA_VS_4, REG_FLAG_NEED_BO},
+ {R_00B144_SPI_SHADER_USER_DATA_VS_5, 0},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_00B148_SPI_SHADER_USER_DATA_VS_6, REG_FLAG_NEED_BO},
+ {R_00B14C_SPI_SHADER_USER_DATA_VS_7, 0},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_00B150_SPI_SHADER_USER_DATA_VS_8, REG_FLAG_NEED_BO},
+ {R_00B154_SPI_SHADER_USER_DATA_VS_9, 0},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_00B158_SPI_SHADER_USER_DATA_VS_10, REG_FLAG_NEED_BO},
+ {R_00B15C_SPI_SHADER_USER_DATA_VS_11, 0},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_00B160_SPI_SHADER_USER_DATA_VS_12, REG_FLAG_NEED_BO},
+ {R_00B164_SPI_SHADER_USER_DATA_VS_13, 0},
+ {GROUP_FORCE_NEW_BLOCK, 0},
+ {R_00B168_SPI_SHADER_USER_DATA_VS_14, REG_FLAG_NEED_BO},
+ {R_00B16C_SPI_SHADER_USER_DATA_VS_15, 0},
+};
+
+int si_context_init(struct r600_context *ctx)
+{
+ int r;
+
+ LIST_INITHEAD(&ctx->active_query_list);
+
+ /* init dirty list */
+ LIST_INITHEAD(&ctx->dirty);
+ LIST_INITHEAD(&ctx->enable_list);
+
+ ctx->range = calloc(NUM_RANGES, sizeof(struct r600_range));
+ if (!ctx->range) {
+ r = -ENOMEM;
+ goto out_err;
+ }
+
+ /* add blocks */
+ r = r600_context_add_block(ctx, si_config_reg_list,
+ Elements(si_config_reg_list), PKT3_SET_CONFIG_REG, SI_CONFIG_REG_OFFSET);
+ if (r)
+ goto out_err;
+ r = r600_context_add_block(ctx, si_context_reg_list,
+ Elements(si_context_reg_list), PKT3_SET_CONTEXT_REG, SI_CONTEXT_REG_OFFSET);
+ if (r)
+ goto out_err;
+ r = r600_context_add_block(ctx, si_sh_reg_list,
+ Elements(si_sh_reg_list), PKT3_SET_SH_REG, SI_SH_REG_OFFSET);
+ if (r)
+ goto out_err;
+
+
+ /* PS SAMPLER */
+ /* VS SAMPLER */
+
+ /* PS SAMPLER BORDER */
+ /* VS SAMPLER BORDER */
+
+ /* PS RESOURCES */
+ /* VS RESOURCES */
+
+ ctx->cs = ctx->ws->cs_create(ctx->ws);
+
+ r600_init_cs(ctx);
+ ctx->max_db = 8;
+ return 0;
+out_err:
+ r600_context_fini(ctx);
+ return r;
+}
+
+static inline void evergreen_context_ps_partial_flush(struct r600_context *ctx)
+{
+ struct radeon_winsys_cs *cs = ctx->cs;
+
+ if (!(ctx->flags & R600_CONTEXT_DRAW_PENDING))
+ return;
+
+ cs->buf[cs->cdw++] = PKT3(PKT3_EVENT_WRITE, 0, 0);
+ cs->buf[cs->cdw++] = EVENT_TYPE(EVENT_TYPE_PS_PARTIAL_FLUSH) | EVENT_INDEX(4);
+
+ ctx->flags &= ~R600_CONTEXT_DRAW_PENDING;
+}
+
+void evergreen_context_draw(struct r600_context *ctx, const struct r600_draw *draw)
+{
+ struct radeon_winsys_cs *cs = ctx->cs;
+ unsigned ndwords = 7;
+ uint32_t *pm4;
+ uint64_t va;
+
+ if (draw->indices) {
+ ndwords = 11;
+ }
+ if (ctx->num_cs_dw_queries_suspend)
+ ndwords += 6;
+
+ /* when increasing ndwords, bump the max limit too */
+ assert(ndwords <= R600_MAX_DRAW_CS_DWORDS);
+
+ /* queries need some special values
+ * (this is non-zero if any query is active) */
+ if (ctx->num_cs_dw_queries_suspend) {
+ pm4 = &cs->buf[cs->cdw];
+ pm4[0] = PKT3(PKT3_SET_CONTEXT_REG, 1, 0);
+ pm4[1] = (R_028004_DB_COUNT_CONTROL - SI_CONTEXT_REG_OFFSET) >> 2;
+ pm4[2] = S_028004_PERFECT_ZPASS_COUNTS(1);
+ pm4[3] = PKT3(PKT3_SET_CONTEXT_REG, 1, 0);
+ pm4[4] = (R_02800C_DB_RENDER_OVERRIDE - SI_CONTEXT_REG_OFFSET) >> 2;
+ pm4[5] = draw->db_render_override | S_02800C_NOOP_CULL_DISABLE(1);
+ cs->cdw += 6;
+ ndwords -= 6;
+ }
+
+ /* draw packet */
+ pm4 = &cs->buf[cs->cdw];
+ pm4[0] = PKT3(PKT3_INDEX_TYPE, 0, ctx->predicate_drawing);
+ pm4[1] = draw->vgt_index_type;
+ pm4[2] = PKT3(PKT3_NUM_INSTANCES, 0, ctx->predicate_drawing);
+ pm4[3] = draw->vgt_num_instances;
+ if (draw->indices) {
+ va = r600_resource_va(&ctx->screen->screen, (void*)draw->indices);
+ va += draw->indices_bo_offset;
+ pm4[4] = PKT3(PKT3_DRAW_INDEX, 3, ctx->predicate_drawing);
+ pm4[5] = va;
+ pm4[6] = (va >> 32UL) & 0xFF;
+ pm4[7] = draw->vgt_num_indices;
+ pm4[8] = draw->vgt_draw_initiator;
+ pm4[9] = PKT3(PKT3_NOP, 0, ctx->predicate_drawing);
+ pm4[10] = r600_context_bo_reloc(ctx, draw->indices, RADEON_USAGE_READ);
+ } else {
+ pm4[4] = PKT3(PKT3_DRAW_INDEX_AUTO, 1, ctx->predicate_drawing);
+ pm4[5] = draw->vgt_num_indices;
+ pm4[6] = draw->vgt_draw_initiator;
+ }
+ cs->cdw += ndwords;
+}
+
+void evergreen_flush_vgt_streamout(struct r600_context *ctx)
+{
+ struct radeon_winsys_cs *cs = ctx->cs;
+
+ cs->buf[cs->cdw++] = PKT3(PKT3_SET_CONFIG_REG, 1, 0);
+ cs->buf[cs->cdw++] = (R_0084FC_CP_STRMOUT_CNTL - SI_CONFIG_REG_OFFSET) >> 2;
+ cs->buf[cs->cdw++] = 0;
+
+ cs->buf[cs->cdw++] = PKT3(PKT3_EVENT_WRITE, 0, 0);
+ cs->buf[cs->cdw++] = EVENT_TYPE(EVENT_TYPE_SO_VGTSTREAMOUT_FLUSH) | EVENT_INDEX(0);
+
+ cs->buf[cs->cdw++] = PKT3(PKT3_WAIT_REG_MEM, 5, 0);
+ cs->buf[cs->cdw++] = WAIT_REG_MEM_EQUAL; /* wait until the register is equal to the reference value */
+ cs->buf[cs->cdw++] = R_0084FC_CP_STRMOUT_CNTL >> 2; /* register */
+ cs->buf[cs->cdw++] = 0;
+ cs->buf[cs->cdw++] = S_0084FC_OFFSET_UPDATE_DONE(1); /* reference value */
+ cs->buf[cs->cdw++] = S_0084FC_OFFSET_UPDATE_DONE(1); /* mask */
+ cs->buf[cs->cdw++] = 4; /* poll interval */
+}
+
+void evergreen_set_streamout_enable(struct r600_context *ctx, unsigned buffer_enable_bit)
+{
+ struct radeon_winsys_cs *cs = ctx->cs;
+
+ if (buffer_enable_bit) {
+ cs->buf[cs->cdw++] = PKT3(PKT3_SET_CONTEXT_REG, 1, 0);
+ cs->buf[cs->cdw++] = (R_028B94_VGT_STRMOUT_CONFIG - SI_CONTEXT_REG_OFFSET) >> 2;
+ cs->buf[cs->cdw++] = S_028B94_STREAMOUT_0_EN(1);
+
+ cs->buf[cs->cdw++] = PKT3(PKT3_SET_CONTEXT_REG, 1, 0);
+ cs->buf[cs->cdw++] = (R_028B98_VGT_STRMOUT_BUFFER_CONFIG - SI_CONTEXT_REG_OFFSET) >> 2;
+ cs->buf[cs->cdw++] = S_028B98_STREAM_0_BUFFER_EN(buffer_enable_bit);
+ } else {
+ cs->buf[cs->cdw++] = PKT3(PKT3_SET_CONTEXT_REG, 1, 0);
+ cs->buf[cs->cdw++] = (R_028B94_VGT_STRMOUT_CONFIG - SI_CONTEXT_REG_OFFSET) >> 2;
+ cs->buf[cs->cdw++] = S_028B94_STREAMOUT_0_EN(0);
+ }
+}
diff --git a/src/gallium/drivers/radeonsi/evergreen_state.c b/src/gallium/drivers/radeonsi/evergreen_state.c
new file mode 100644
index 00000000000..5049c7b2db6
--- /dev/null
+++ b/src/gallium/drivers/radeonsi/evergreen_state.c
@@ -0,0 +1,2169 @@
+/*
+ * Copyright 2010 Jerome Glisse <[email protected]>
+ *
+ * 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.
+ */
+
+/* TODO:
+ * - fix mask for depth control & cull for query
+ */
+#include <stdio.h>
+#include <errno.h>
+#include "pipe/p_defines.h"
+#include "pipe/p_state.h"
+#include "pipe/p_context.h"
+#include "tgsi/tgsi_scan.h"
+#include "tgsi/tgsi_parse.h"
+#include "tgsi/tgsi_util.h"
+#include "util/u_blitter.h"
+#include "util/u_double_list.h"
+#include "util/u_transfer.h"
+#include "util/u_surface.h"
+#include "util/u_pack_color.h"
+#include "util/u_memory.h"
+#include "util/u_inlines.h"
+#include "util/u_framebuffer.h"
+#include "pipebuffer/pb_buffer.h"
+#include "r600.h"
+#include "sid.h"
+#include "r600_resource.h"
+#include "radeonsi_pipe.h"
+
+static uint32_t si_translate_blend_function(int blend_func)
+{
+ switch (blend_func) {
+ case PIPE_BLEND_ADD:
+ return V_028780_COMB_DST_PLUS_SRC;
+ case PIPE_BLEND_SUBTRACT:
+ return V_028780_COMB_SRC_MINUS_DST;
+ case PIPE_BLEND_REVERSE_SUBTRACT:
+ return V_028780_COMB_DST_MINUS_SRC;
+ case PIPE_BLEND_MIN:
+ return V_028780_COMB_MIN_DST_SRC;
+ case PIPE_BLEND_MAX:
+ return V_028780_COMB_MAX_DST_SRC;
+ default:
+ R600_ERR("Unknown blend function %d\n", blend_func);
+ assert(0);
+ break;
+ }
+ return 0;
+}
+
+static uint32_t si_translate_blend_factor(int blend_fact)
+{
+ switch (blend_fact) {
+ case PIPE_BLENDFACTOR_ONE:
+ return V_028780_BLEND_ONE;
+ case PIPE_BLENDFACTOR_SRC_COLOR:
+ return V_028780_BLEND_SRC_COLOR;
+ case PIPE_BLENDFACTOR_SRC_ALPHA:
+ return V_028780_BLEND_SRC_ALPHA;
+ case PIPE_BLENDFACTOR_DST_ALPHA:
+ return V_028780_BLEND_DST_ALPHA;
+ case PIPE_BLENDFACTOR_DST_COLOR:
+ return V_028780_BLEND_DST_COLOR;
+ case PIPE_BLENDFACTOR_SRC_ALPHA_SATURATE:
+ return V_028780_BLEND_SRC_ALPHA_SATURATE;
+ case PIPE_BLENDFACTOR_CONST_COLOR:
+ return V_028780_BLEND_CONSTANT_COLOR;
+ case PIPE_BLENDFACTOR_CONST_ALPHA:
+ return V_028780_BLEND_CONSTANT_ALPHA;
+ case PIPE_BLENDFACTOR_ZERO:
+ return V_028780_BLEND_ZERO;
+ case PIPE_BLENDFACTOR_INV_SRC_COLOR:
+ return V_028780_BLEND_ONE_MINUS_SRC_COLOR;
+ case PIPE_BLENDFACTOR_INV_SRC_ALPHA:
+ return V_028780_BLEND_ONE_MINUS_SRC_ALPHA;
+ case PIPE_BLENDFACTOR_INV_DST_ALPHA:
+ return V_028780_BLEND_ONE_MINUS_DST_ALPHA;
+ case PIPE_BLENDFACTOR_INV_DST_COLOR:
+ return V_028780_BLEND_ONE_MINUS_DST_COLOR;
+ case PIPE_BLENDFACTOR_INV_CONST_COLOR:
+ return V_028780_BLEND_ONE_MINUS_CONSTANT_COLOR;
+ case PIPE_BLENDFACTOR_INV_CONST_ALPHA:
+ return V_028780_BLEND_ONE_MINUS_CONSTANT_ALPHA;
+ case PIPE_BLENDFACTOR_SRC1_COLOR:
+ return V_028780_BLEND_SRC1_COLOR;
+ case PIPE_BLENDFACTOR_SRC1_ALPHA:
+ return V_028780_BLEND_SRC1_ALPHA;
+ case PIPE_BLENDFACTOR_INV_SRC1_COLOR:
+ return V_028780_BLEND_INV_SRC1_COLOR;
+ case PIPE_BLENDFACTOR_INV_SRC1_ALPHA:
+ return V_028780_BLEND_INV_SRC1_ALPHA;
+ default:
+ R600_ERR("Bad blend factor %d not supported!\n", blend_fact);
+ assert(0);
+ break;
+ }
+ return 0;
+}
+
+#if 0
+static uint32_t r600_translate_stencil_op(int s_op)
+{
+ switch (s_op) {
+ case PIPE_STENCIL_OP_KEEP:
+ return V_028800_STENCIL_KEEP;
+ case PIPE_STENCIL_OP_ZERO:
+ return V_028800_STENCIL_ZERO;
+ case PIPE_STENCIL_OP_REPLACE:
+ return V_028800_STENCIL_REPLACE;
+ case PIPE_STENCIL_OP_INCR:
+ return V_028800_STENCIL_INCR;
+ case PIPE_STENCIL_OP_DECR:
+ return V_028800_STENCIL_DECR;
+ case PIPE_STENCIL_OP_INCR_WRAP:
+ return V_028800_STENCIL_INCR_WRAP;
+ case PIPE_STENCIL_OP_DECR_WRAP:
+ return V_028800_STENCIL_DECR_WRAP;
+ case PIPE_STENCIL_OP_INVERT:
+ return V_028800_STENCIL_INVERT;
+ default:
+ R600_ERR("Unknown stencil op %d", s_op);
+ assert(0);
+ break;
+ }
+ return 0;
+}
+#endif
+
+static uint32_t si_translate_fill(uint32_t func)
+{
+ switch(func) {
+ case PIPE_POLYGON_MODE_FILL:
+ return V_028814_X_DRAW_TRIANGLES;
+ case PIPE_POLYGON_MODE_LINE:
+ return V_028814_X_DRAW_LINES;
+ case PIPE_POLYGON_MODE_POINT:
+ return V_028814_X_DRAW_POINTS;
+ default:
+ assert(0);
+ return V_028814_X_DRAW_POINTS;
+ }
+}
+
+/* translates straight */
+static uint32_t si_translate_ds_func(int func)
+{
+ return func;
+}
+
+static unsigned si_tex_wrap(unsigned wrap)
+{
+ switch (wrap) {
+ default:
+ case PIPE_TEX_WRAP_REPEAT:
+ return V_008F30_SQ_TEX_WRAP;
+ case PIPE_TEX_WRAP_CLAMP:
+ return V_008F30_SQ_TEX_CLAMP_HALF_BORDER;
+ case PIPE_TEX_WRAP_CLAMP_TO_EDGE:
+ return V_008F30_SQ_TEX_CLAMP_LAST_TEXEL;
+ case PIPE_TEX_WRAP_CLAMP_TO_BORDER:
+ return V_008F30_SQ_TEX_CLAMP_BORDER;
+ case PIPE_TEX_WRAP_MIRROR_REPEAT:
+ return V_008F30_SQ_TEX_MIRROR;
+ case PIPE_TEX_WRAP_MIRROR_CLAMP:
+ return V_008F30_SQ_TEX_MIRROR_ONCE_HALF_BORDER;
+ case PIPE_TEX_WRAP_MIRROR_CLAMP_TO_EDGE:
+ return V_008F30_SQ_TEX_MIRROR_ONCE_LAST_TEXEL;
+ case PIPE_TEX_WRAP_MIRROR_CLAMP_TO_BORDER:
+ return V_008F30_SQ_TEX_MIRROR_ONCE_BORDER;
+ }
+}
+
+static unsigned si_tex_filter(unsigned filter)
+{
+ switch (filter) {
+ default:
+ case PIPE_TEX_FILTER_NEAREST:
+ return V_008F38_SQ_TEX_XY_FILTER_POINT;
+ case PIPE_TEX_FILTER_LINEAR:
+ return V_008F38_SQ_TEX_XY_FILTER_BILINEAR;
+ }
+}
+
+static unsigned si_tex_mipfilter(unsigned filter)
+{
+ switch (filter) {
+ case PIPE_TEX_MIPFILTER_NEAREST:
+ return V_008F38_SQ_TEX_Z_FILTER_POINT;
+ case PIPE_TEX_MIPFILTER_LINEAR:
+ return V_008F38_SQ_TEX_Z_FILTER_LINEAR;
+ default:
+ case PIPE_TEX_MIPFILTER_NONE:
+ return V_008F38_SQ_TEX_Z_FILTER_NONE;
+ }
+}
+
+static unsigned si_tex_compare(unsigned compare)
+{
+ switch (compare) {
+ default:
+ case PIPE_FUNC_NEVER:
+ return V_008F30_SQ_TEX_DEPTH_COMPARE_NEVER;
+ case PIPE_FUNC_LESS:
+ return V_008F30_SQ_TEX_DEPTH_COMPARE_LESS;
+ case PIPE_FUNC_EQUAL:
+ return V_008F30_SQ_TEX_DEPTH_COMPARE_EQUAL;
+ case PIPE_FUNC_LEQUAL:
+ return V_008F30_SQ_TEX_DEPTH_COMPARE_LESSEQUAL;
+ case PIPE_FUNC_GREATER:
+ return V_008F30_SQ_TEX_DEPTH_COMPARE_GREATER;
+ case PIPE_FUNC_NOTEQUAL:
+ return V_008F30_SQ_TEX_DEPTH_COMPARE_NOTEQUAL;
+ case PIPE_FUNC_GEQUAL:
+ return V_008F30_SQ_TEX_DEPTH_COMPARE_GREATEREQUAL;
+ case PIPE_FUNC_ALWAYS:
+ return V_008F30_SQ_TEX_DEPTH_COMPARE_ALWAYS;
+ }
+}
+
+static unsigned si_tex_dim(unsigned dim)
+{
+ switch (dim) {
+ default:
+ case PIPE_TEXTURE_1D:
+ return V_008F1C_SQ_RSRC_IMG_1D;
+ case PIPE_TEXTURE_1D_ARRAY:
+ return V_008F1C_SQ_RSRC_IMG_1D_ARRAY;
+ case PIPE_TEXTURE_2D:
+ case PIPE_TEXTURE_RECT:
+ return V_008F1C_SQ_RSRC_IMG_2D;
+ case PIPE_TEXTURE_2D_ARRAY:
+ return V_008F1C_SQ_RSRC_IMG_2D_ARRAY;
+ case PIPE_TEXTURE_3D:
+ return V_008F1C_SQ_RSRC_IMG_3D;
+ case PIPE_TEXTURE_CUBE:
+ return V_008F1C_SQ_RSRC_IMG_CUBE;
+ }
+}
+
+static uint32_t si_translate_dbformat(enum pipe_format format)
+{
+ switch (format) {
+ case PIPE_FORMAT_Z16_UNORM:
+ return V_028040_Z_16;
+ case PIPE_FORMAT_Z24X8_UNORM:
+ case PIPE_FORMAT_Z24_UNORM_S8_UINT:
+ return V_028040_Z_24; /* XXX no longer supported on SI */
+ case PIPE_FORMAT_Z32_FLOAT:
+ case PIPE_FORMAT_Z32_FLOAT_S8X24_UINT:
+ return V_028040_Z_32_FLOAT;
+ default:
+ return ~0U;
+ }
+}
+
+static uint32_t si_translate_colorswap(enum pipe_format format)
+{
+ switch (format) {
+ /* 8-bit buffers. */
+ case PIPE_FORMAT_L4A4_UNORM:
+ case PIPE_FORMAT_A4R4_UNORM:
+ return V_028C70_SWAP_ALT;
+
+ case PIPE_FORMAT_A8_UNORM:
+ case PIPE_FORMAT_A8_UINT:
+ case PIPE_FORMAT_A8_SINT:
+ case PIPE_FORMAT_R4A4_UNORM:
+ return V_028C70_SWAP_ALT_REV;
+ case PIPE_FORMAT_I8_UNORM:
+ case PIPE_FORMAT_L8_UNORM:
+ case PIPE_FORMAT_I8_UINT:
+ case PIPE_FORMAT_I8_SINT:
+ case PIPE_FORMAT_L8_UINT:
+ case PIPE_FORMAT_L8_SINT:
+ case PIPE_FORMAT_L8_SRGB:
+ case PIPE_FORMAT_R8_UNORM:
+ case PIPE_FORMAT_R8_SNORM:
+ case PIPE_FORMAT_R8_UINT:
+ case PIPE_FORMAT_R8_SINT:
+ return V_028C70_SWAP_STD;
+
+ /* 16-bit buffers. */
+ case PIPE_FORMAT_B5G6R5_UNORM:
+ return V_028C70_SWAP_STD_REV;
+
+ case PIPE_FORMAT_B5G5R5A1_UNORM:
+ case PIPE_FORMAT_B5G5R5X1_UNORM:
+ return V_028C70_SWAP_ALT;
+
+ case PIPE_FORMAT_B4G4R4A4_UNORM:
+ case PIPE_FORMAT_B4G4R4X4_UNORM:
+ return V_028C70_SWAP_ALT;
+
+ case PIPE_FORMAT_Z16_UNORM:
+ return V_028C70_SWAP_STD;
+
+ case PIPE_FORMAT_L8A8_UNORM:
+ case PIPE_FORMAT_L8A8_UINT:
+ case PIPE_FORMAT_L8A8_SINT:
+ case PIPE_FORMAT_L8A8_SRGB:
+ return V_028C70_SWAP_ALT;
+ case PIPE_FORMAT_R8G8_UNORM:
+ case PIPE_FORMAT_R8G8_UINT:
+ case PIPE_FORMAT_R8G8_SINT:
+ return V_028C70_SWAP_STD;
+
+ case PIPE_FORMAT_R16_UNORM:
+ case PIPE_FORMAT_R16_UINT:
+ case PIPE_FORMAT_R16_SINT:
+ case PIPE_FORMAT_R16_FLOAT:
+ return V_028C70_SWAP_STD;
+
+ /* 32-bit buffers. */
+ case PIPE_FORMAT_A8B8G8R8_SRGB:
+ return V_028C70_SWAP_STD_REV;
+ case PIPE_FORMAT_B8G8R8A8_SRGB:
+ return V_028C70_SWAP_ALT;
+
+ case PIPE_FORMAT_B8G8R8A8_UNORM:
+ case PIPE_FORMAT_B8G8R8X8_UNORM:
+ return V_028C70_SWAP_ALT;
+
+ case PIPE_FORMAT_A8R8G8B8_UNORM:
+ case PIPE_FORMAT_X8R8G8B8_UNORM:
+ return V_028C70_SWAP_ALT_REV;
+ case PIPE_FORMAT_R8G8B8A8_SNORM:
+ case PIPE_FORMAT_R8G8B8A8_UNORM:
+ case PIPE_FORMAT_R8G8B8A8_SSCALED:
+ case PIPE_FORMAT_R8G8B8A8_USCALED:
+ case PIPE_FORMAT_R8G8B8A8_SINT:
+ case PIPE_FORMAT_R8G8B8A8_UINT:
+ case PIPE_FORMAT_R8G8B8X8_UNORM:
+ return V_028C70_SWAP_STD;
+
+ case PIPE_FORMAT_A8B8G8R8_UNORM:
+ case PIPE_FORMAT_X8B8G8R8_UNORM:
+ /* case PIPE_FORMAT_R8SG8SB8UX8U_NORM: */
+ return V_028C70_SWAP_STD_REV;
+
+ case PIPE_FORMAT_Z24X8_UNORM:
+ case PIPE_FORMAT_Z24_UNORM_S8_UINT:
+ return V_028C70_SWAP_STD;
+
+ case PIPE_FORMAT_X8Z24_UNORM:
+ case PIPE_FORMAT_S8_UINT_Z24_UNORM:
+ return V_028C70_SWAP_STD;
+
+ case PIPE_FORMAT_R10G10B10A2_UNORM:
+ case PIPE_FORMAT_R10G10B10X2_SNORM:
+ case PIPE_FORMAT_R10SG10SB10SA2U_NORM:
+ return V_028C70_SWAP_STD;
+
+ case PIPE_FORMAT_B10G10R10A2_UNORM:
+ case PIPE_FORMAT_B10G10R10A2_UINT:
+ return V_028C70_SWAP_ALT;
+
+ case PIPE_FORMAT_R11G11B10_FLOAT:
+ case PIPE_FORMAT_R32_FLOAT:
+ case PIPE_FORMAT_R32_UINT:
+ case PIPE_FORMAT_R32_SINT:
+ case PIPE_FORMAT_Z32_FLOAT:
+ case PIPE_FORMAT_R16G16_FLOAT:
+ case PIPE_FORMAT_R16G16_UNORM:
+ case PIPE_FORMAT_R16G16_UINT:
+ case PIPE_FORMAT_R16G16_SINT:
+ return V_028C70_SWAP_STD;
+
+ /* 64-bit buffers. */
+ case PIPE_FORMAT_R32G32_FLOAT:
+ case PIPE_FORMAT_R32G32_UINT:
+ case PIPE_FORMAT_R32G32_SINT:
+ case PIPE_FORMAT_R16G16B16A16_UNORM:
+ case PIPE_FORMAT_R16G16B16A16_SNORM:
+ case PIPE_FORMAT_R16G16B16A16_USCALED:
+ case PIPE_FORMAT_R16G16B16A16_SSCALED:
+ case PIPE_FORMAT_R16G16B16A16_UINT:
+ case PIPE_FORMAT_R16G16B16A16_SINT:
+ case PIPE_FORMAT_R16G16B16A16_FLOAT:
+ case PIPE_FORMAT_Z32_FLOAT_S8X24_UINT:
+
+ /* 128-bit buffers. */
+ case PIPE_FORMAT_R32G32B32A32_FLOAT:
+ case PIPE_FORMAT_R32G32B32A32_SNORM:
+ case PIPE_FORMAT_R32G32B32A32_UNORM:
+ case PIPE_FORMAT_R32G32B32A32_SSCALED:
+ case PIPE_FORMAT_R32G32B32A32_USCALED:
+ case PIPE_FORMAT_R32G32B32A32_SINT:
+ case PIPE_FORMAT_R32G32B32A32_UINT:
+ return V_028C70_SWAP_STD;
+ default:
+ R600_ERR("unsupported colorswap format %d\n", format);
+ return ~0U;
+ }
+ return ~0U;
+}
+
+static uint32_t si_translate_colorformat(enum pipe_format format)
+{
+ switch (format) {
+ /* 8-bit buffers. */
+ case PIPE_FORMAT_A8_UNORM:
+ case PIPE_FORMAT_A8_UINT:
+ case PIPE_FORMAT_A8_SINT:
+ case PIPE_FORMAT_I8_UNORM:
+ case PIPE_FORMAT_I8_UINT:
+ case PIPE_FORMAT_I8_SINT:
+ case PIPE_FORMAT_L8_UNORM:
+ case PIPE_FORMAT_L8_UINT:
+ case PIPE_FORMAT_L8_SINT:
+ case PIPE_FORMAT_L8_SRGB:
+ case PIPE_FORMAT_R8_UNORM:
+ case PIPE_FORMAT_R8_SNORM:
+ case PIPE_FORMAT_R8_UINT:
+ case PIPE_FORMAT_R8_SINT:
+ return V_028C70_COLOR_8;
+
+ /* 16-bit buffers. */
+ case PIPE_FORMAT_B5G6R5_UNORM:
+ return V_028C70_COLOR_5_6_5;
+
+ case PIPE_FORMAT_B5G5R5A1_UNORM:
+ case PIPE_FORMAT_B5G5R5X1_UNORM:
+ return V_028C70_COLOR_1_5_5_5;
+
+ case PIPE_FORMAT_B4G4R4A4_UNORM:
+ case PIPE_FORMAT_B4G4R4X4_UNORM:
+ return V_028C70_COLOR_4_4_4_4;
+
+ case PIPE_FORMAT_L8A8_UNORM:
+ case PIPE_FORMAT_L8A8_UINT:
+ case PIPE_FORMAT_L8A8_SINT:
+ case PIPE_FORMAT_L8A8_SRGB:
+ case PIPE_FORMAT_R8G8_UNORM:
+ case PIPE_FORMAT_R8G8_UINT:
+ case PIPE_FORMAT_R8G8_SINT:
+ return V_028C70_COLOR_8_8;
+
+ case PIPE_FORMAT_Z16_UNORM:
+ case PIPE_FORMAT_R16_UNORM:
+ case PIPE_FORMAT_R16_UINT:
+ case PIPE_FORMAT_R16_SINT:
+ case PIPE_FORMAT_R16_FLOAT:
+ case PIPE_FORMAT_R16G16_FLOAT:
+ return V_028C70_COLOR_16;
+
+ /* 32-bit buffers. */
+ case PIPE_FORMAT_A8B8G8R8_SRGB:
+ case PIPE_FORMAT_A8B8G8R8_UNORM:
+ case PIPE_FORMAT_A8R8G8B8_UNORM:
+ case PIPE_FORMAT_B8G8R8A8_SRGB:
+ case PIPE_FORMAT_B8G8R8A8_UNORM:
+ case PIPE_FORMAT_B8G8R8X8_UNORM:
+ case PIPE_FORMAT_R8G8B8A8_SNORM:
+ case PIPE_FORMAT_R8G8B8A8_UNORM:
+ case PIPE_FORMAT_R8G8B8X8_UNORM:
+ case PIPE_FORMAT_R8SG8SB8UX8U_NORM:
+ case PIPE_FORMAT_X8B8G8R8_UNORM:
+ case PIPE_FORMAT_X8R8G8B8_UNORM:
+ case PIPE_FORMAT_R8G8B8_UNORM:
+ case PIPE_FORMAT_R8G8B8A8_SSCALED:
+ case PIPE_FORMAT_R8G8B8A8_USCALED:
+ case PIPE_FORMAT_R8G8B8A8_SINT:
+ case PIPE_FORMAT_R8G8B8A8_UINT:
+ return V_028C70_COLOR_8_8_8_8;
+
+ case PIPE_FORMAT_R10G10B10A2_UNORM:
+ case PIPE_FORMAT_R10G10B10X2_SNORM:
+ case PIPE_FORMAT_B10G10R10A2_UNORM:
+ case PIPE_FORMAT_B10G10R10A2_UINT:
+ case PIPE_FORMAT_R10SG10SB10SA2U_NORM:
+ return V_028C70_COLOR_2_10_10_10;
+
+ case PIPE_FORMAT_Z24X8_UNORM:
+ case PIPE_FORMAT_Z24_UNORM_S8_UINT:
+ return V_028C70_COLOR_8_24;
+
+ case PIPE_FORMAT_X8Z24_UNORM:
+ case PIPE_FORMAT_S8_UINT_Z24_UNORM:
+ return V_028C70_COLOR_24_8;
+
+ case PIPE_FORMAT_Z32_FLOAT_S8X24_UINT:
+ return V_028C70_COLOR_X24_8_32_FLOAT;
+
+ case PIPE_FORMAT_R32_FLOAT:
+ case PIPE_FORMAT_Z32_FLOAT:
+ return V_028C70_COLOR_32;
+
+ case PIPE_FORMAT_R16G16_SSCALED:
+ case PIPE_FORMAT_R16G16_UNORM:
+ case PIPE_FORMAT_R16G16_UINT:
+ case PIPE_FORMAT_R16G16_SINT:
+ return V_028C70_COLOR_16_16;
+
+ case PIPE_FORMAT_R11G11B10_FLOAT:
+ return V_028C70_COLOR_10_11_11;
+
+ /* 64-bit buffers. */
+ case PIPE_FORMAT_R16G16B16_USCALED:
+ case PIPE_FORMAT_R16G16B16_SSCALED:
+ case PIPE_FORMAT_R16G16B16A16_UINT:
+ case PIPE_FORMAT_R16G16B16A16_SINT:
+ case PIPE_FORMAT_R16G16B16A16_USCALED:
+ case PIPE_FORMAT_R16G16B16A16_SSCALED:
+ case PIPE_FORMAT_R16G16B16A16_UNORM:
+ case PIPE_FORMAT_R16G16B16A16_SNORM:
+ case PIPE_FORMAT_R16G16B16_FLOAT:
+ case PIPE_FORMAT_R16G16B16A16_FLOAT:
+ return V_028C70_COLOR_16_16_16_16;
+
+ case PIPE_FORMAT_R32G32_FLOAT:
+ case PIPE_FORMAT_R32G32_USCALED:
+ case PIPE_FORMAT_R32G32_SSCALED:
+ case PIPE_FORMAT_R32G32_SINT:
+ case PIPE_FORMAT_R32G32_UINT:
+ return V_028C70_COLOR_32_32;
+
+ /* 128-bit buffers. */
+ case PIPE_FORMAT_R32G32B32A32_SNORM:
+ case PIPE_FORMAT_R32G32B32A32_UNORM:
+ case PIPE_FORMAT_R32G32B32A32_SSCALED:
+ case PIPE_FORMAT_R32G32B32A32_USCALED:
+ case PIPE_FORMAT_R32G32B32A32_SINT:
+ case PIPE_FORMAT_R32G32B32A32_UINT:
+ case PIPE_FORMAT_R32G32B32A32_FLOAT:
+ return V_028C70_COLOR_32_32_32_32;
+
+ /* YUV buffers. */
+ case PIPE_FORMAT_UYVY:
+ case PIPE_FORMAT_YUYV:
+ /* 96-bit buffers. */
+ case PIPE_FORMAT_R32G32B32_FLOAT:
+ /* 8-bit buffers. */
+ case PIPE_FORMAT_L4A4_UNORM:
+ case PIPE_FORMAT_R4A4_UNORM:
+ case PIPE_FORMAT_A4R4_UNORM:
+ default:
+ return ~0U; /* Unsupported. */
+ }
+}
+
+static uint32_t si_colorformat_endian_swap(uint32_t colorformat)
+{
+ if (R600_BIG_ENDIAN) {
+ switch(colorformat) {
+ /* 8-bit buffers. */
+ case V_028C70_COLOR_8:
+ return V_028C70_ENDIAN_NONE;
+
+ /* 16-bit buffers. */
+ case V_028C70_COLOR_5_6_5:
+ case V_028C70_COLOR_1_5_5_5:
+ case V_028C70_COLOR_4_4_4_4:
+ case V_028C70_COLOR_16:
+ case V_028C70_COLOR_8_8:
+ return V_028C70_ENDIAN_8IN16;
+
+ /* 32-bit buffers. */
+ case V_028C70_COLOR_8_8_8_8:
+ case V_028C70_COLOR_2_10_10_10:
+ case V_028C70_COLOR_8_24:
+ case V_028C70_COLOR_24_8:
+ case V_028C70_COLOR_16_16:
+ return V_028C70_ENDIAN_8IN32;
+
+ /* 64-bit buffers. */
+ case V_028C70_COLOR_16_16_16_16:
+ return V_028C70_ENDIAN_8IN16;
+
+ case V_028C70_COLOR_32_32:
+ return V_028C70_ENDIAN_8IN32;
+
+ /* 128-bit buffers. */
+ case V_028C70_COLOR_32_32_32_32:
+ return V_028C70_ENDIAN_8IN32;
+ default:
+ return V_028C70_ENDIAN_NONE; /* Unsupported. */
+ }
+ } else {
+ return V_028C70_ENDIAN_NONE;
+ }
+}
+
+static uint32_t si_translate_texformat(struct pipe_screen *screen,
+ enum pipe_format format,
+ const struct util_format_description *desc,
+ int first_non_void)
+{
+ boolean uniform = TRUE;
+ int i;
+
+ /* Colorspace (return non-RGB formats directly). */
+ switch (desc->colorspace) {
+ /* Depth stencil formats */
+ case UTIL_FORMAT_COLORSPACE_ZS:
+ switch (format) {
+ case PIPE_FORMAT_Z16_UNORM:
+ return V_008F14_IMG_DATA_FORMAT_16;
+ case PIPE_FORMAT_X24S8_UINT:
+ case PIPE_FORMAT_Z24X8_UNORM:
+ case PIPE_FORMAT_Z24_UNORM_S8_UINT:
+ return V_008F14_IMG_DATA_FORMAT_24_8;
+ case PIPE_FORMAT_S8X24_UINT:
+ case PIPE_FORMAT_X8Z24_UNORM:
+ case PIPE_FORMAT_S8_UINT_Z24_UNORM:
+ return V_008F14_IMG_DATA_FORMAT_8_24;
+ case PIPE_FORMAT_S8_UINT:
+ return V_008F14_IMG_DATA_FORMAT_8;
+ case PIPE_FORMAT_Z32_FLOAT:
+ return V_008F14_IMG_DATA_FORMAT_32;
+ case PIPE_FORMAT_Z32_FLOAT_S8X24_UINT:
+ return V_008F14_IMG_DATA_FORMAT_X24_8_32;
+ default:
+ goto out_unknown;
+ }
+
+ case UTIL_FORMAT_COLORSPACE_YUV:
+ goto out_unknown; /* TODO */
+
+ case UTIL_FORMAT_COLORSPACE_SRGB:
+ break;
+
+ default:
+ break;
+ }
+
+ /* TODO compressed formats */
+
+ if (format == PIPE_FORMAT_R9G9B9E5_FLOAT) {
+ return V_008F14_IMG_DATA_FORMAT_5_9_9_9;
+ } else if (format == PIPE_FORMAT_R11G11B10_FLOAT) {
+ return V_008F14_IMG_DATA_FORMAT_10_11_11;
+ }
+
+ /* R8G8Bx_SNORM - TODO CxV8U8 */
+
+ /* See whether the components are of the same size. */
+ for (i = 1; i < desc->nr_channels; i++) {
+ uniform = uniform && desc->channel[0].size == desc->channel[i].size;
+ }
+
+ /* Non-uniform formats. */
+ if (!uniform) {
+ switch(desc->nr_channels) {
+ case 3:
+ if (desc->channel[0].size == 5 &&
+ desc->channel[1].size == 6 &&
+ desc->channel[2].size == 5) {
+ return V_008F14_IMG_DATA_FORMAT_5_6_5;
+ }
+ goto out_unknown;
+ case 4:
+ if (desc->channel[0].size == 5 &&
+ desc->channel[1].size == 5 &&
+ desc->channel[2].size == 5 &&
+ desc->channel[3].size == 1) {
+ return V_008F14_IMG_DATA_FORMAT_1_5_5_5;
+ }
+ if (desc->channel[0].size == 10 &&
+ desc->channel[1].size == 10 &&
+ desc->channel[2].size == 10 &&
+ desc->channel[3].size == 2) {
+ return V_008F14_IMG_DATA_FORMAT_2_10_10_10;
+ }
+ goto out_unknown;
+ }
+ goto out_unknown;
+ }
+
+ if (first_non_void < 0 || first_non_void > 3)
+ goto out_unknown;
+
+ /* uniform formats */
+ switch (desc->channel[first_non_void].size) {
+ case 4:
+ switch (desc->nr_channels) {
+ case 2:
+ return V_008F14_IMG_DATA_FORMAT_4_4;
+ case 4:
+ return V_008F14_IMG_DATA_FORMAT_4_4_4_4;
+ }
+ break;
+ case 8:
+ switch (desc->nr_channels) {
+ case 1:
+ return V_008F14_IMG_DATA_FORMAT_8;
+ case 2:
+ return V_008F14_IMG_DATA_FORMAT_8_8;
+ case 4:
+ return V_008F14_IMG_DATA_FORMAT_8_8_8_8;
+ }
+ break;
+ case 16:
+ switch (desc->nr_channels) {
+ case 1:
+ return V_008F14_IMG_DATA_FORMAT_16;
+ case 2:
+ return V_008F14_IMG_DATA_FORMAT_16_16;
+ case 4:
+ return V_008F14_IMG_DATA_FORMAT_16_16_16_16;
+ }
+ break;
+ case 32:
+ switch (desc->nr_channels) {
+ case 1:
+ return V_008F14_IMG_DATA_FORMAT_32;
+ case 2:
+ return V_008F14_IMG_DATA_FORMAT_32_32;
+ case 3:
+ return V_008F14_IMG_DATA_FORMAT_32_32_32;
+ case 4:
+ return V_008F14_IMG_DATA_FORMAT_32_32_32_32;
+ }
+ }
+
+out_unknown:
+ /* R600_ERR("Unable to handle texformat %d %s\n", format, util_format_name(format)); */
+ return ~0;
+}
+
+static bool si_is_sampler_format_supported(struct pipe_screen *screen, enum pipe_format format)
+{
+ return si_translate_texformat(screen, format, util_format_description(format),
+ util_format_get_first_non_void_channel(format)) != ~0U;
+}
+
+uint32_t si_translate_vertexformat(struct pipe_screen *screen,
+ enum pipe_format format,
+ const struct util_format_description *desc,
+ int first_non_void)
+{
+ uint32_t result = si_translate_texformat(screen, format, desc, first_non_void);
+
+ if (result == V_008F0C_BUF_DATA_FORMAT_INVALID ||
+ result > V_008F0C_BUF_DATA_FORMAT_32_32_32_32)
+ result = ~0;
+
+ return result;
+}
+
+static bool si_is_vertex_format_supported(struct pipe_screen *screen, enum pipe_format format)
+{
+ return si_translate_vertexformat(screen, format, util_format_description(format),
+ util_format_get_first_non_void_channel(format)) != ~0U;
+}
+
+static bool r600_is_colorbuffer_format_supported(enum pipe_format format)
+{
+ return si_translate_colorformat(format) != ~0U &&
+ si_translate_colorswap(format) != ~0U;
+}
+
+static bool r600_is_zs_format_supported(enum pipe_format format)
+{
+ return si_translate_dbformat(format) != ~0U;
+}
+
+boolean si_is_format_supported(struct pipe_screen *screen,
+ enum pipe_format format,
+ enum pipe_texture_target target,
+ unsigned sample_count,
+ unsigned usage)
+{
+ unsigned retval = 0;
+
+ if (target >= PIPE_MAX_TEXTURE_TYPES) {
+ R600_ERR("r600: unsupported texture type %d\n", target);
+ return FALSE;
+ }
+
+ if (!util_format_is_supported(format, usage))
+ return FALSE;
+
+ /* Multisample */
+ if (sample_count > 1)
+ return FALSE;
+
+ if ((usage & PIPE_BIND_SAMPLER_VIEW) &&
+ si_is_sampler_format_supported(screen, format)) {
+ retval |= PIPE_BIND_SAMPLER_VIEW;
+ }
+
+ if ((usage & (PIPE_BIND_RENDER_TARGET |
+ PIPE_BIND_DISPLAY_TARGET |
+ PIPE_BIND_SCANOUT |
+ PIPE_BIND_SHARED)) &&
+ r600_is_colorbuffer_format_supported(format)) {
+ retval |= usage &
+ (PIPE_BIND_RENDER_TARGET |
+ PIPE_BIND_DISPLAY_TARGET |
+ PIPE_BIND_SCANOUT |
+ PIPE_BIND_SHARED);
+ }
+
+ if ((usage & PIPE_BIND_DEPTH_STENCIL) &&
+ r600_is_zs_format_supported(format)) {
+ retval |= PIPE_BIND_DEPTH_STENCIL;
+ }
+
+ if ((usage & PIPE_BIND_VERTEX_BUFFER) &&
+ si_is_vertex_format_supported(screen, format)) {
+ retval |= PIPE_BIND_VERTEX_BUFFER;
+ }
+
+ if (usage & PIPE_BIND_TRANSFER_READ)
+ retval |= PIPE_BIND_TRANSFER_READ;
+ if (usage & PIPE_BIND_TRANSFER_WRITE)
+ retval |= PIPE_BIND_TRANSFER_WRITE;
+
+ return retval == usage;
+}
+
+static void evergreen_set_blend_color(struct pipe_context *ctx,
+ const struct pipe_blend_color *state)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct r600_pipe_state *rstate = CALLOC_STRUCT(r600_pipe_state);
+
+ if (rstate == NULL)
+ return;
+
+ rstate->id = R600_PIPE_STATE_BLEND_COLOR;
+ r600_pipe_state_add_reg(rstate, R_028414_CB_BLEND_RED, fui(state->color[0]), NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028418_CB_BLEND_GREEN, fui(state->color[1]), NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_02841C_CB_BLEND_BLUE, fui(state->color[2]), NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028420_CB_BLEND_ALPHA, fui(state->color[3]), NULL, 0);
+
+ free(rctx->states[R600_PIPE_STATE_BLEND_COLOR]);
+ rctx->states[R600_PIPE_STATE_BLEND_COLOR] = rstate;
+ r600_context_pipe_state_set(rctx, rstate);
+}
+
+static void *evergreen_create_blend_state(struct pipe_context *ctx,
+ const struct pipe_blend_state *state)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct r600_pipe_blend *blend = CALLOC_STRUCT(r600_pipe_blend);
+ struct r600_pipe_state *rstate;
+ uint32_t color_control, target_mask;
+ /* FIXME there is more then 8 framebuffer */
+ unsigned blend_cntl[8];
+
+ if (blend == NULL) {
+ return NULL;
+ }
+
+ rstate = &blend->rstate;
+
+ rstate->id = R600_PIPE_STATE_BLEND;
+
+ target_mask = 0;
+ color_control = S_028808_MODE(V_028808_CB_NORMAL);
+ if (state->logicop_enable) {
+ color_control |= S_028808_ROP3(state->logicop_func | (state->logicop_func << 4));
+ } else {
+ color_control |= S_028808_ROP3(0xcc);
+ }
+ /* we pretend 8 buffer are used, CB_SHADER_MASK will disable unused one */
+ if (state->independent_blend_enable) {
+ for (int i = 0; i < 8; i++) {
+ target_mask |= (state->rt[i].colormask << (4 * i));
+ }
+ } else {
+ for (int i = 0; i < 8; i++) {
+ target_mask |= (state->rt[0].colormask << (4 * i));
+ }
+ }
+ blend->cb_target_mask = target_mask;
+
+ r600_pipe_state_add_reg(rstate, R_028808_CB_COLOR_CONTROL,
+ color_control, NULL, 0);
+
+ r600_pipe_state_add_reg(rstate, R_028C38_PA_SC_AA_MASK_X0Y0_X1Y0, ~0, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028C3C_PA_SC_AA_MASK_X0Y1_X1Y1, ~0, NULL, 0);
+
+ for (int i = 0; i < 8; i++) {
+ /* state->rt entries > 0 only written if independent blending */
+ const int j = state->independent_blend_enable ? i : 0;
+
+ unsigned eqRGB = state->rt[j].rgb_func;
+ unsigned srcRGB = state->rt[j].rgb_src_factor;
+ unsigned dstRGB = state->rt[j].rgb_dst_factor;
+ unsigned eqA = state->rt[j].alpha_func;
+ unsigned srcA = state->rt[j].alpha_src_factor;
+ unsigned dstA = state->rt[j].alpha_dst_factor;
+
+ blend_cntl[i] = 0;
+ if (!state->rt[j].blend_enable)
+ continue;
+
+ blend_cntl[i] |= S_028780_ENABLE(1);
+ blend_cntl[i] |= S_028780_COLOR_COMB_FCN(si_translate_blend_function(eqRGB));
+ blend_cntl[i] |= S_028780_COLOR_SRCBLEND(si_translate_blend_factor(srcRGB));
+ blend_cntl[i] |= S_028780_COLOR_DESTBLEND(si_translate_blend_factor(dstRGB));
+
+ if (srcA != srcRGB || dstA != dstRGB || eqA != eqRGB) {
+ blend_cntl[i] |= S_028780_SEPARATE_ALPHA_BLEND(1);
+ blend_cntl[i] |= S_028780_ALPHA_COMB_FCN(si_translate_blend_function(eqA));
+ blend_cntl[i] |= S_028780_ALPHA_SRCBLEND(si_translate_blend_factor(srcA));
+ blend_cntl[i] |= S_028780_ALPHA_DESTBLEND(si_translate_blend_factor(dstA));
+ }
+ }
+ for (int i = 0; i < 8; i++) {
+ r600_pipe_state_add_reg(rstate, R_028780_CB_BLEND0_CONTROL + i * 4, blend_cntl[i], NULL, 0);
+ }
+
+ return rstate;
+}
+
+static void *evergreen_create_dsa_state(struct pipe_context *ctx,
+ const struct pipe_depth_stencil_alpha_state *state)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct r600_pipe_dsa *dsa = CALLOC_STRUCT(r600_pipe_dsa);
+ unsigned db_depth_control, alpha_test_control, alpha_ref;
+ unsigned db_render_override, db_render_control;
+ struct r600_pipe_state *rstate;
+
+ if (dsa == NULL) {
+ return NULL;
+ }
+
+ dsa->valuemask[0] = state->stencil[0].valuemask;
+ dsa->valuemask[1] = state->stencil[1].valuemask;
+ dsa->writemask[0] = state->stencil[0].writemask;
+ dsa->writemask[1] = state->stencil[1].writemask;
+
+ rstate = &dsa->rstate;
+
+ rstate->id = R600_PIPE_STATE_DSA;
+ db_depth_control = S_028800_Z_ENABLE(state->depth.enabled) |
+ S_028800_Z_WRITE_ENABLE(state->depth.writemask) |
+ S_028800_ZFUNC(state->depth.func);
+
+ /* stencil */
+ if (state->stencil[0].enabled) {
+ db_depth_control |= S_028800_STENCIL_ENABLE(1);
+ db_depth_control |= S_028800_STENCILFUNC(si_translate_ds_func(state->stencil[0].func));
+ //db_depth_control |= S_028800_STENCILFAIL(r600_translate_stencil_op(state->stencil[0].fail_op));
+ //db_depth_control |= S_028800_STENCILZPASS(r600_translate_stencil_op(state->stencil[0].zpass_op));
+ //db_depth_control |= S_028800_STENCILZFAIL(r600_translate_stencil_op(state->stencil[0].zfail_op));
+
+ if (state->stencil[1].enabled) {
+ db_depth_control |= S_028800_BACKFACE_ENABLE(1);
+ db_depth_control |= S_028800_STENCILFUNC_BF(si_translate_ds_func(state->stencil[1].func));
+ //db_depth_control |= S_028800_STENCILFAIL_BF(r600_translate_stencil_op(state->stencil[1].fail_op));
+ //db_depth_control |= S_028800_STENCILZPASS_BF(r600_translate_stencil_op(state->stencil[1].zpass_op));
+ //db_depth_control |= S_028800_STENCILZFAIL_BF(r600_translate_stencil_op(state->stencil[1].zfail_op));
+ }
+ }
+
+ /* alpha */
+ alpha_test_control = 0;
+ alpha_ref = 0;
+ if (state->alpha.enabled) {
+ //alpha_test_control = S_028410_ALPHA_FUNC(state->alpha.func);
+ //alpha_test_control |= S_028410_ALPHA_TEST_ENABLE(1);
+ alpha_ref = fui(state->alpha.ref_value);
+ }
+ dsa->alpha_ref = alpha_ref;
+
+ /* misc */
+ db_render_control = 0;
+ db_render_override = S_02800C_FORCE_HIZ_ENABLE(V_02800C_FORCE_DISABLE) |
+ S_02800C_FORCE_HIS_ENABLE0(V_02800C_FORCE_DISABLE) |
+ S_02800C_FORCE_HIS_ENABLE1(V_02800C_FORCE_DISABLE);
+ /* TODO db_render_override depends on query */
+ r600_pipe_state_add_reg(rstate, R_028020_DB_DEPTH_BOUNDS_MIN, 0x00000000, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028024_DB_DEPTH_BOUNDS_MAX, 0x00000000, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028028_DB_STENCIL_CLEAR, 0x00000000, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_02802C_DB_DEPTH_CLEAR, 0x3F800000, NULL, 0);
+ //r600_pipe_state_add_reg(rstate, R_028410_SX_ALPHA_TEST_CONTROL, alpha_test_control, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028800_DB_DEPTH_CONTROL, db_depth_control, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028000_DB_RENDER_CONTROL, db_render_control, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_02800C_DB_RENDER_OVERRIDE, db_render_override, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028AC0_DB_SRESULTS_COMPARE_STATE0, 0x0, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028AC4_DB_SRESULTS_COMPARE_STATE1, 0x0, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028AC8_DB_PRELOAD_CONTROL, 0x0, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028B70_DB_ALPHA_TO_MASK, 0x0000AA00, NULL, 0);
+ dsa->db_render_override = db_render_override;
+
+ return rstate;
+}
+
+static void *evergreen_create_rs_state(struct pipe_context *ctx,
+ const struct pipe_rasterizer_state *state)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct r600_pipe_rasterizer *rs = CALLOC_STRUCT(r600_pipe_rasterizer);
+ struct r600_pipe_state *rstate;
+ unsigned tmp;
+ unsigned prov_vtx = 1, polygon_dual_mode;
+ unsigned clip_rule;
+ float psize_min, psize_max;
+
+ if (rs == NULL) {
+ return NULL;
+ }
+
+ polygon_dual_mode = (state->fill_front != PIPE_POLYGON_MODE_FILL ||
+ state->fill_back != PIPE_POLYGON_MODE_FILL);
+
+ if (state->flatshade_first)
+ prov_vtx = 0;
+
+ rstate = &rs->rstate;
+ rs->flatshade = state->flatshade;
+ rs->sprite_coord_enable = state->sprite_coord_enable;
+ rs->pa_sc_line_stipple = state->line_stipple_enable ?
+ S_028A0C_LINE_PATTERN(state->line_stipple_pattern) |
+ S_028A0C_REPEAT_COUNT(state->line_stipple_factor) : 0;
+ rs->pa_su_sc_mode_cntl =
+ S_028814_PROVOKING_VTX_LAST(prov_vtx) |
+ S_028814_CULL_FRONT(state->rasterizer_discard || (state->cull_face & PIPE_FACE_FRONT) ? 1 : 0) |
+ S_028814_CULL_BACK(state->rasterizer_discard || (state->cull_face & PIPE_FACE_BACK) ? 1 : 0) |
+ S_028814_FACE(!state->front_ccw) |
+ S_028814_POLY_OFFSET_FRONT_ENABLE(state->offset_tri) |
+ S_028814_POLY_OFFSET_BACK_ENABLE(state->offset_tri) |
+ S_028814_POLY_OFFSET_PARA_ENABLE(state->offset_tri) |
+ S_028814_POLY_MODE(polygon_dual_mode) |
+ S_028814_POLYMODE_FRONT_PTYPE(si_translate_fill(state->fill_front)) |
+ S_028814_POLYMODE_BACK_PTYPE(si_translate_fill(state->fill_back));
+ rs->pa_cl_clip_cntl =
+ S_028810_PS_UCP_MODE(3) |
+ S_028810_ZCLIP_NEAR_DISABLE(!state->depth_clip) |
+ S_028810_ZCLIP_FAR_DISABLE(!state->depth_clip) |
+ S_028810_DX_LINEAR_ATTR_CLIP_ENA(1);
+ rs->pa_cl_vs_out_cntl =
+ S_02881C_USE_VTX_POINT_SIZE(state->point_size_per_vertex) |
+ S_02881C_VS_OUT_MISC_VEC_ENA(state->point_size_per_vertex);
+
+ clip_rule = state->scissor ? 0xAAAA : 0xFFFF;
+
+ /* offset */
+ rs->offset_units = state->offset_units;
+ rs->offset_scale = state->offset_scale * 12.0f;
+
+ rstate->id = R600_PIPE_STATE_RASTERIZER;
+ tmp = S_0286D4_FLAT_SHADE_ENA(state->flatshade);
+ if (state->sprite_coord_enable) {
+ tmp |= S_0286D4_PNT_SPRITE_ENA(1) |
+ S_0286D4_PNT_SPRITE_OVRD_X(V_0286D4_SPI_PNT_SPRITE_SEL_S) |
+ S_0286D4_PNT_SPRITE_OVRD_Y(V_0286D4_SPI_PNT_SPRITE_SEL_T) |
+ S_0286D4_PNT_SPRITE_OVRD_Z(V_0286D4_SPI_PNT_SPRITE_SEL_0) |
+ S_0286D4_PNT_SPRITE_OVRD_W(V_0286D4_SPI_PNT_SPRITE_SEL_1);
+ if (state->sprite_coord_mode != PIPE_SPRITE_COORD_UPPER_LEFT) {
+ tmp |= S_0286D4_PNT_SPRITE_TOP_1(1);
+ }
+ }
+ r600_pipe_state_add_reg(rstate, R_0286D4_SPI_INTERP_CONTROL_0, tmp, NULL, 0);
+
+ r600_pipe_state_add_reg(rstate, R_028820_PA_CL_NANINF_CNTL, 0x00000000, NULL, 0);
+ /* point size 12.4 fixed point */
+ tmp = (unsigned)(state->point_size * 8.0);
+ r600_pipe_state_add_reg(rstate, R_028A00_PA_SU_POINT_SIZE, S_028A00_HEIGHT(tmp) | S_028A00_WIDTH(tmp), NULL, 0);
+
+ if (state->point_size_per_vertex) {
+ psize_min = util_get_min_point_size(state);
+ psize_max = 8192;
+ } else {
+ /* Force the point size to be as if the vertex output was disabled. */
+ psize_min = state->point_size;
+ psize_max = state->point_size;
+ }
+ /* Divide by two, because 0.5 = 1 pixel. */
+ r600_pipe_state_add_reg(rstate, R_028A04_PA_SU_POINT_MINMAX,
+ S_028A04_MIN_SIZE(r600_pack_float_12p4(psize_min/2)) |
+ S_028A04_MAX_SIZE(r600_pack_float_12p4(psize_max/2)),
+ NULL, 0);
+
+ tmp = (unsigned)state->line_width * 8;
+ r600_pipe_state_add_reg(rstate, R_028A08_PA_SU_LINE_CNTL, S_028A08_WIDTH(tmp), NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028A48_PA_SC_MODE_CNTL_0,
+ S_028A48_LINE_STIPPLE_ENABLE(state->line_stipple_enable),
+ NULL, 0);
+
+ r600_pipe_state_add_reg(rstate, R_028BDC_PA_SC_LINE_CNTL, 0x00000400, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028BE4_PA_SU_VTX_CNTL,
+ S_028BE4_PIX_CENTER(state->gl_rasterization_rules),
+ NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028BE8_PA_CL_GB_VERT_CLIP_ADJ, 0x3F800000, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028BEC_PA_CL_GB_VERT_DISC_ADJ, 0x3F800000, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028BF0_PA_CL_GB_HORZ_CLIP_ADJ, 0x3F800000, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028BF4_PA_CL_GB_HORZ_DISC_ADJ, 0x3F800000, NULL, 0);
+
+ r600_pipe_state_add_reg(rstate, R_028B7C_PA_SU_POLY_OFFSET_CLAMP, fui(state->offset_clamp), NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_02820C_PA_SC_CLIPRECT_RULE, clip_rule, NULL, 0);
+ return rstate;
+}
+
+static void *si_create_sampler_state(struct pipe_context *ctx,
+ const struct pipe_sampler_state *state)
+{
+ struct si_pipe_sampler_state *rstate = CALLOC_STRUCT(si_pipe_sampler_state);
+ union util_color uc;
+ unsigned aniso_flag_offset = state->max_anisotropy > 1 ? 2 : 0;
+ unsigned border_color_type;
+
+ if (rstate == NULL) {
+ return NULL;
+ }
+
+ util_pack_color(state->border_color.f, PIPE_FORMAT_B8G8R8A8_UNORM, &uc);
+ switch (uc.ui) {
+ case 0x000000FF: /* opaque black */
+ border_color_type = 0;
+ break;
+ case 0x00000000: /* transparent black */
+ border_color_type = 1;
+ break;
+ case 0xFFFFFFFF: /* white */
+ border_color_type = 1;
+ break;
+ default: /* Use border color pointer */
+ border_color_type = 3;
+ }
+
+ rstate->val[0] = si_tex_wrap(state->wrap_s) |
+ si_tex_wrap(state->wrap_t) << 3 |
+ si_tex_wrap(state->wrap_r) << 6 |
+ (state->max_anisotropy & 0x7) << 9 | /* XXX */
+ si_tex_compare(state->compare_func) << 12 |
+ !state->normalized_coords << 15 |
+ aniso_flag_offset << 16 | /* XXX */
+ !state->seamless_cube_map << 28 |
+ si_tex_mipfilter(state->min_mip_filter) << 29;
+ rstate->val[1] = S_FIXED(CLAMP(state->min_lod, 0, 15), 8) |
+ S_FIXED(CLAMP(state->max_lod, 0, 15), 8) << 12;
+ rstate->val[2] = S_FIXED(CLAMP(state->lod_bias, -16, 16), 8) |
+ si_tex_filter(state->mag_img_filter) << 20 |
+ si_tex_filter(state->min_img_filter) << 22;
+ rstate->val[3] = border_color_type << 30;
+
+#if 0
+ if (border_color_type == 3) {
+ r600_pipe_state_add_reg_noblock(rstate, R_00A404_TD_PS_SAMPLER0_BORDER_RED, fui(state->border_color.f[0]), NULL, 0);
+ r600_pipe_state_add_reg_noblock(rstate, R_00A408_TD_PS_SAMPLER0_BORDER_GREEN, fui(state->border_color.f[1]), NULL, 0);
+ r600_pipe_state_add_reg_noblock(rstate, R_00A40C_TD_PS_SAMPLER0_BORDER_BLUE, fui(state->border_color.f[2]), NULL, 0);
+ r600_pipe_state_add_reg_noblock(rstate, R_00A410_TD_PS_SAMPLER0_BORDER_ALPHA, fui(state->border_color.f[3]), NULL, 0);
+ }
+#endif
+ return rstate;
+}
+
+static void si_delete_sampler_state(struct pipe_context *ctx,
+ void *state)
+{
+ free(state);
+}
+
+static unsigned si_map_swizzle(unsigned swizzle)
+{
+ switch (swizzle) {
+ case UTIL_FORMAT_SWIZZLE_Y:
+ return V_008F1C_SQ_SEL_Y;
+ case UTIL_FORMAT_SWIZZLE_Z:
+ return V_008F1C_SQ_SEL_Z;
+ case UTIL_FORMAT_SWIZZLE_W:
+ return V_008F1C_SQ_SEL_W;
+ case UTIL_FORMAT_SWIZZLE_0:
+ return V_008F1C_SQ_SEL_0;
+ case UTIL_FORMAT_SWIZZLE_1:
+ return V_008F1C_SQ_SEL_1;
+ default: /* UTIL_FORMAT_SWIZZLE_X */
+ return V_008F1C_SQ_SEL_X;
+ }
+}
+
+static struct pipe_sampler_view *evergreen_create_sampler_view(struct pipe_context *ctx,
+ struct pipe_resource *texture,
+ const struct pipe_sampler_view *state)
+{
+ struct si_pipe_sampler_view *view = CALLOC_STRUCT(si_pipe_sampler_view);
+ struct r600_resource_texture *tmp = (struct r600_resource_texture*)texture;
+ const struct util_format_description *desc = util_format_description(state->format);
+ unsigned format, num_format, endian;
+ uint32_t pitch = 0;
+ unsigned char state_swizzle[4], swizzle[4], array_mode = 0, tile_type = 0;
+ unsigned height, depth;
+ int first_non_void;
+ uint64_t va;
+
+ if (view == NULL)
+ return NULL;
+
+ /* initialize base object */
+ view->base = *state;
+ view->base.texture = NULL;
+ pipe_reference(NULL, &texture->reference);
+ view->base.texture = texture;
+ view->base.reference.count = 1;
+ view->base.context = ctx;
+
+ state_swizzle[0] = state->swizzle_r;
+ state_swizzle[1] = state->swizzle_g;
+ state_swizzle[2] = state->swizzle_b;
+ state_swizzle[3] = state->swizzle_a;
+ util_format_compose_swizzles(desc->swizzle, state_swizzle, swizzle);
+
+ first_non_void = util_format_get_first_non_void_channel(state->format);
+ switch (desc->channel[first_non_void].type) {
+ case UTIL_FORMAT_TYPE_FLOAT:
+ num_format = V_008F14_IMG_NUM_FORMAT_FLOAT;
+ break;
+ case UTIL_FORMAT_TYPE_FIXED:
+ num_format = V_008F14_IMG_NUM_FORMAT_USCALED; /* XXX */
+ break;
+ case UTIL_FORMAT_TYPE_SIGNED:
+ num_format = V_008F14_IMG_NUM_FORMAT_SNORM;
+ break;
+ case UTIL_FORMAT_TYPE_UNSIGNED:
+ default:
+ num_format = V_008F14_IMG_NUM_FORMAT_UNORM;
+ }
+
+ format = si_translate_texformat(ctx->screen, state->format, desc, first_non_void);
+ if (format == ~0) {
+ format = 0;
+ }
+
+ if (tmp->depth && !tmp->is_flushing_texture) {
+ r600_texture_depth_flush(ctx, texture, TRUE);
+ tmp = tmp->flushed_depth_texture;
+ }
+
+ endian = si_colorformat_endian_swap(format);
+
+ height = texture->height0;
+ depth = texture->depth0;
+
+ pitch = align(tmp->pitch_in_blocks[0] *
+ util_format_get_blockwidth(state->format), 8);
+ array_mode = tmp->array_mode[0];
+ tile_type = tmp->tile_type;
+
+ if (texture->target == PIPE_TEXTURE_1D_ARRAY) {
+ height = 1;
+ depth = texture->array_size;
+ } else if (texture->target == PIPE_TEXTURE_2D_ARRAY) {
+ depth = texture->array_size;
+ }
+
+ va = r600_resource_va(ctx->screen, texture);
+ view->state[0] = (va + tmp->offset[0]) >> 8;
+ view->state[1] = ((va + tmp->offset[0]) >> 40) & 0xff;
+ view->state[1] |= (S_008F14_DATA_FORMAT(format) |
+ S_008F14_NUM_FORMAT(num_format));
+ view->state[2] = (S_008F18_WIDTH(texture->width0 - 1) |
+ S_008F18_HEIGHT(height - 1));
+ view->state[3] = (S_008F1C_DST_SEL_X(si_map_swizzle(swizzle[0])) |
+ S_008F1C_DST_SEL_Y(si_map_swizzle(swizzle[1])) |
+ S_008F1C_DST_SEL_Z(si_map_swizzle(swizzle[2])) |
+ S_008F1C_DST_SEL_W(si_map_swizzle(swizzle[3])) |
+ S_008F1C_BASE_LEVEL(state->u.tex.first_level) |
+ S_008F1C_LAST_LEVEL(state->u.tex.last_level) |
+ S_008F1C_TYPE(si_tex_dim(texture->target)));
+ view->state[4] = (S_008F20_DEPTH(depth - 1) |
+ S_008F20_PITCH((pitch / 8) - 1));
+ view->state[5] = (S_008F24_BASE_ARRAY(state->u.tex.first_layer) |
+ S_008F24_LAST_ARRAY(state->u.tex.last_layer));
+ view->state[6] = 0;
+ view->state[7] = 0;
+
+ return &view->base;
+}
+
+static void evergreen_set_vs_sampler_view(struct pipe_context *ctx, unsigned count,
+ struct pipe_sampler_view **views)
+{
+}
+
+static void evergreen_set_ps_sampler_view(struct pipe_context *ctx, unsigned count,
+ struct pipe_sampler_view **views)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct si_pipe_sampler_view **resource = (struct si_pipe_sampler_view **)views;
+ struct r600_pipe_state *rstate = &rctx->ps_samplers.rstate;
+ struct r600_resource *bo;
+ int i;
+ int has_depth = 0;
+ uint64_t va;
+ char *ptr;
+
+ if (!count)
+ goto out;
+
+ r600_inval_texture_cache(rctx);
+
+ bo = (struct r600_resource*)
+ pipe_buffer_create(ctx->screen, PIPE_BIND_CUSTOM, PIPE_USAGE_IMMUTABLE,
+ count * sizeof(resource[0]->state));
+ ptr = rctx->ws->buffer_map(bo->buf, rctx->cs, PIPE_TRANSFER_WRITE);
+
+ for (i = 0; i < count; i++, ptr += sizeof(resource[0]->state)) {
+ pipe_sampler_view_reference(
+ (struct pipe_sampler_view **)&rctx->ps_samplers.views[i],
+ views[i]);
+
+ if (resource[i]) {
+ if (((struct r600_resource_texture *)resource[i]->base.texture)->depth)
+ has_depth = 1;
+
+ memcpy(ptr, resource[i]->state, sizeof(resource[0]->state));
+ } else
+ memset(ptr, 0, sizeof(resource[0]->state));
+ }
+
+ rctx->ws->buffer_unmap(bo->buf);
+
+ for (i = count; i < NUM_TEX_UNITS; i++) {
+ if (rctx->ps_samplers.views[i])
+ pipe_sampler_view_reference((struct pipe_sampler_view **)&rctx->ps_samplers.views[i], NULL);
+ }
+
+ va = r600_resource_va(ctx->screen, (void *)bo);
+ r600_pipe_state_add_reg(rstate, R_00B040_SPI_SHADER_USER_DATA_PS_4, va, bo, RADEON_USAGE_READ);
+ r600_pipe_state_add_reg(rstate, R_00B044_SPI_SHADER_USER_DATA_PS_5, va >> 32, NULL, 0);
+ r600_context_pipe_state_set(rctx, rstate);
+
+out:
+ rctx->have_depth_texture = has_depth;
+ rctx->ps_samplers.n_views = count;
+}
+
+static void evergreen_bind_ps_sampler(struct pipe_context *ctx, unsigned count, void **states)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct si_pipe_sampler_state **rstates = (struct si_pipe_sampler_state **)states;
+ struct r600_pipe_state *rstate = &rctx->ps_samplers.rstate;
+ struct r600_resource *bo;
+ uint64_t va;
+ char *ptr;
+ int i;
+
+ if (!count)
+ goto out;
+
+ r600_inval_texture_cache(rctx);
+
+ bo = (struct r600_resource*)
+ pipe_buffer_create(ctx->screen, PIPE_BIND_CUSTOM, PIPE_USAGE_IMMUTABLE,
+ count * sizeof(rstates[0]->val));
+ ptr = rctx->ws->buffer_map(bo->buf, rctx->cs, PIPE_TRANSFER_WRITE);
+
+ for (i = 0; i < count; i++, ptr += sizeof(rstates[0]->val)) {
+ memcpy(ptr, rstates[i]->val, sizeof(rstates[0]->val));
+ }
+
+ rctx->ws->buffer_unmap(bo->buf);
+
+ va = r600_resource_va(ctx->screen, (void *)bo);
+ r600_pipe_state_add_reg(rstate, R_00B038_SPI_SHADER_USER_DATA_PS_2, va, bo, RADEON_USAGE_READ);
+ r600_pipe_state_add_reg(rstate, R_00B03C_SPI_SHADER_USER_DATA_PS_3, va >> 32, NULL, 0);
+ r600_context_pipe_state_set(rctx, rstate);
+
+out:
+ rctx->ps_samplers.n_samplers = count;
+}
+
+static void evergreen_bind_vs_sampler(struct pipe_context *ctx, unsigned count, void **states)
+{
+}
+
+static void evergreen_set_clip_state(struct pipe_context *ctx,
+ const struct pipe_clip_state *state)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct r600_pipe_state *rstate = CALLOC_STRUCT(r600_pipe_state);
+
+ if (rstate == NULL)
+ return;
+
+ rctx->clip = *state;
+ rstate->id = R600_PIPE_STATE_CLIP;
+ for (int i = 0; i < 6; i++) {
+ r600_pipe_state_add_reg(rstate,
+ R_0285BC_PA_CL_UCP_0_X + i * 16,
+ fui(state->ucp[i][0]), NULL, 0);
+ r600_pipe_state_add_reg(rstate,
+ R_0285C0_PA_CL_UCP_0_Y + i * 16,
+ fui(state->ucp[i][1]) , NULL, 0);
+ r600_pipe_state_add_reg(rstate,
+ R_0285C4_PA_CL_UCP_0_Z + i * 16,
+ fui(state->ucp[i][2]), NULL, 0);
+ r600_pipe_state_add_reg(rstate,
+ R_0285C8_PA_CL_UCP_0_W + i * 16,
+ fui(state->ucp[i][3]), NULL, 0);
+ }
+
+ free(rctx->states[R600_PIPE_STATE_CLIP]);
+ rctx->states[R600_PIPE_STATE_CLIP] = rstate;
+ r600_context_pipe_state_set(rctx, rstate);
+}
+
+static void evergreen_set_polygon_stipple(struct pipe_context *ctx,
+ const struct pipe_poly_stipple *state)
+{
+}
+
+static void evergreen_set_sample_mask(struct pipe_context *pipe, unsigned sample_mask)
+{
+}
+
+static void evergreen_set_scissor_state(struct pipe_context *ctx,
+ const struct pipe_scissor_state *state)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct r600_pipe_state *rstate = CALLOC_STRUCT(r600_pipe_state);
+ uint32_t tl, br;
+
+ if (rstate == NULL)
+ return;
+
+ rstate->id = R600_PIPE_STATE_SCISSOR;
+ tl = S_028240_TL_X(state->minx) | S_028240_TL_Y(state->miny);
+ br = S_028244_BR_X(state->maxx) | S_028244_BR_Y(state->maxy);
+ r600_pipe_state_add_reg(rstate,
+ R_028210_PA_SC_CLIPRECT_0_TL, tl,
+ NULL, 0);
+ r600_pipe_state_add_reg(rstate,
+ R_028214_PA_SC_CLIPRECT_0_BR, br,
+ NULL, 0);
+ r600_pipe_state_add_reg(rstate,
+ R_028218_PA_SC_CLIPRECT_1_TL, tl,
+ NULL, 0);
+ r600_pipe_state_add_reg(rstate,
+ R_02821C_PA_SC_CLIPRECT_1_BR, br,
+ NULL, 0);
+ r600_pipe_state_add_reg(rstate,
+ R_028220_PA_SC_CLIPRECT_2_TL, tl,
+ NULL, 0);
+ r600_pipe_state_add_reg(rstate,
+ R_028224_PA_SC_CLIPRECT_2_BR, br,
+ NULL, 0);
+ r600_pipe_state_add_reg(rstate,
+ R_028228_PA_SC_CLIPRECT_3_TL, tl,
+ NULL, 0);
+ r600_pipe_state_add_reg(rstate,
+ R_02822C_PA_SC_CLIPRECT_3_BR, br,
+ NULL, 0);
+
+ free(rctx->states[R600_PIPE_STATE_SCISSOR]);
+ rctx->states[R600_PIPE_STATE_SCISSOR] = rstate;
+ r600_context_pipe_state_set(rctx, rstate);
+}
+
+static void evergreen_set_viewport_state(struct pipe_context *ctx,
+ const struct pipe_viewport_state *state)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct r600_pipe_state *rstate = CALLOC_STRUCT(r600_pipe_state);
+
+ if (rstate == NULL)
+ return;
+
+ rctx->viewport = *state;
+ rstate->id = R600_PIPE_STATE_VIEWPORT;
+ r600_pipe_state_add_reg(rstate, R_0282D0_PA_SC_VPORT_ZMIN_0, 0x00000000, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_0282D4_PA_SC_VPORT_ZMAX_0, 0x3F800000, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028350_PA_SC_RASTER_CONFIG, 0x00000000, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_02843C_PA_CL_VPORT_XSCALE_0, fui(state->scale[0]), NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028444_PA_CL_VPORT_YSCALE_0, fui(state->scale[1]), NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_02844C_PA_CL_VPORT_ZSCALE_0, fui(state->scale[2]), NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028440_PA_CL_VPORT_XOFFSET_0, fui(state->translate[0]), NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028448_PA_CL_VPORT_YOFFSET_0, fui(state->translate[1]), NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028450_PA_CL_VPORT_ZOFFSET_0, fui(state->translate[2]), NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028818_PA_CL_VTE_CNTL, 0x0000043F, NULL, 0);
+
+ free(rctx->states[R600_PIPE_STATE_VIEWPORT]);
+ rctx->states[R600_PIPE_STATE_VIEWPORT] = rstate;
+ r600_context_pipe_state_set(rctx, rstate);
+}
+
+static void evergreen_cb(struct r600_context *rctx, struct r600_pipe_state *rstate,
+ const struct pipe_framebuffer_state *state, int cb)
+{
+ struct r600_resource_texture *rtex;
+ struct r600_surface *surf;
+ unsigned level = state->cbufs[cb]->u.tex.level;
+ unsigned pitch, slice;
+ unsigned color_info;
+ unsigned format, swap, ntype, endian;
+ uint64_t offset;
+ unsigned tile_type;
+ const struct util_format_description *desc;
+ int i;
+ unsigned blend_clamp = 0, blend_bypass = 0;
+
+ surf = (struct r600_surface *)state->cbufs[cb];
+ rtex = (struct r600_resource_texture*)state->cbufs[cb]->texture;
+
+ if (rtex->depth)
+ rctx->have_depth_fb = TRUE;
+
+ if (rtex->depth && !rtex->is_flushing_texture) {
+ r600_texture_depth_flush(&rctx->context, state->cbufs[cb]->texture, TRUE);
+ rtex = rtex->flushed_depth_texture;
+ }
+
+ /* XXX quite sure for dx10+ hw don't need any offset hacks */
+ offset = r600_texture_get_offset(rtex,
+ level, state->cbufs[cb]->u.tex.first_layer);
+ pitch = rtex->pitch_in_blocks[level] / 8 - 1;
+ slice = rtex->pitch_in_blocks[level] * surf->aligned_height / 64 - 1;
+ desc = util_format_description(surf->base.format);
+ for (i = 0; i < 4; i++) {
+ if (desc->channel[i].type != UTIL_FORMAT_TYPE_VOID) {
+ break;
+ }
+ }
+ if (desc->channel[i].type == UTIL_FORMAT_TYPE_FLOAT) {
+ ntype = V_028C70_NUMBER_FLOAT;
+ } else {
+ ntype = V_028C70_NUMBER_UNORM;
+ if (desc->colorspace == UTIL_FORMAT_COLORSPACE_SRGB)
+ ntype = V_028C70_NUMBER_SRGB;
+ else if (desc->channel[i].type == UTIL_FORMAT_TYPE_SIGNED) {
+ if (desc->channel[i].normalized)
+ ntype = V_028C70_NUMBER_SNORM;
+ else if (desc->channel[i].pure_integer)
+ ntype = V_028C70_NUMBER_SINT;
+ } else if (desc->channel[i].type == UTIL_FORMAT_TYPE_UNSIGNED) {
+ if (desc->channel[i].normalized)
+ ntype = V_028C70_NUMBER_UNORM;
+ else if (desc->channel[i].pure_integer)
+ ntype = V_028C70_NUMBER_UINT;
+ }
+ }
+
+ format = si_translate_colorformat(surf->base.format);
+ swap = si_translate_colorswap(surf->base.format);
+ if (rtex->resource.b.b.b.usage == PIPE_USAGE_STAGING) {
+ endian = V_028C70_ENDIAN_NONE;
+ } else {
+ endian = si_colorformat_endian_swap(format);
+ }
+
+ /* blend clamp should be set for all NORM/SRGB types */
+ if (ntype == V_028C70_NUMBER_UNORM ||
+ ntype == V_028C70_NUMBER_SNORM ||
+ ntype == V_028C70_NUMBER_SRGB)
+ blend_clamp = 1;
+
+ /* set blend bypass according to docs if SINT/UINT or
+ 8/24 COLOR variants */
+ if (ntype == V_028C70_NUMBER_UINT || ntype == V_028C70_NUMBER_SINT ||
+ format == V_028C70_COLOR_8_24 || format == V_028C70_COLOR_24_8 ||
+ format == V_028C70_COLOR_X24_8_32_FLOAT) {
+ blend_clamp = 0;
+ blend_bypass = 1;
+ }
+
+ color_info = S_028C70_FORMAT(format) |
+ S_028C70_COMP_SWAP(swap) |
+ //S_028C70_ARRAY_MODE(rtex->array_mode[level]) |
+ S_028C70_BLEND_CLAMP(blend_clamp) |
+ S_028C70_BLEND_BYPASS(blend_bypass) |
+ S_028C70_NUMBER_TYPE(ntype) |
+ S_028C70_ENDIAN(endian);
+
+ color_info |= S_028C70_LINEAR_GENERAL(1);
+
+ rctx->alpha_ref_dirty = true;
+
+ offset += r600_resource_va(rctx->context.screen, state->cbufs[cb]->texture);
+ offset >>= 8;
+
+ /* FIXME handle enabling of CB beyond BASE8 which has different offset */
+ r600_pipe_state_add_reg(rstate,
+ R_028C60_CB_COLOR0_BASE + cb * 0x3C,
+ offset, &rtex->resource, RADEON_USAGE_READWRITE);
+ r600_pipe_state_add_reg(rstate,
+ R_028C64_CB_COLOR0_PITCH + cb * 0x3C,
+ S_028C64_TILE_MAX(pitch),
+ NULL, 0);
+ r600_pipe_state_add_reg(rstate,
+ R_028C68_CB_COLOR0_SLICE + cb * 0x3C,
+ S_028C68_TILE_MAX(slice),
+ NULL, 0);
+ r600_pipe_state_add_reg(rstate,
+ R_028C6C_CB_COLOR0_VIEW + cb * 0x3C,
+ 0x00000000, NULL, 0);
+ r600_pipe_state_add_reg(rstate,
+ R_028C70_CB_COLOR0_INFO + cb * 0x3C,
+ color_info, &rtex->resource, RADEON_USAGE_READWRITE);
+ r600_pipe_state_add_reg(rstate,
+ R_028C74_CB_COLOR0_ATTRIB + cb * 0x3C,
+ 0,
+ &rtex->resource, RADEON_USAGE_READWRITE);
+}
+
+static void evergreen_db(struct r600_context *rctx, struct r600_pipe_state *rstate,
+ const struct pipe_framebuffer_state *state)
+{
+ struct r600_resource_texture *rtex;
+ struct r600_surface *surf;
+ unsigned level, first_layer, pitch, slice, format, array_mode;
+ uint64_t offset;
+
+ if (state->zsbuf == NULL) {
+ r600_pipe_state_add_reg(rstate, R_028040_DB_Z_INFO, 0, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028044_DB_STENCIL_INFO, 0, NULL, 0);
+ return;
+ }
+
+ surf = (struct r600_surface *)state->zsbuf;
+ level = surf->base.u.tex.level;
+ rtex = (struct r600_resource_texture*)surf->base.texture;
+
+ /* XXX remove this once tiling is properly supported */
+ array_mode = 0;/*rtex->array_mode[level] ? rtex->array_mode[level] :
+ V_028C70_ARRAY_1D_TILED_THIN1;*/
+
+ first_layer = surf->base.u.tex.first_layer;
+ offset = r600_texture_get_offset(rtex, level, first_layer);
+ pitch = rtex->pitch_in_blocks[level] / 8 - 1;
+ slice = rtex->pitch_in_blocks[level] * surf->aligned_height / 64 - 1;
+ format = si_translate_dbformat(rtex->real_format);
+
+ offset += r600_resource_va(rctx->context.screen, surf->base.texture);
+ offset >>= 8;
+
+ r600_pipe_state_add_reg(rstate, R_028048_DB_Z_READ_BASE,
+ offset, &rtex->resource, RADEON_USAGE_READWRITE);
+ r600_pipe_state_add_reg(rstate, R_028050_DB_Z_WRITE_BASE,
+ offset, &rtex->resource, RADEON_USAGE_READWRITE);
+ r600_pipe_state_add_reg(rstate, R_028008_DB_DEPTH_VIEW, 0x00000000, NULL, 0);
+
+ if (rtex->stencil) {
+ uint64_t stencil_offset =
+ r600_texture_get_offset(rtex->stencil, level, first_layer);
+
+ stencil_offset += r600_resource_va(rctx->context.screen, (void*)rtex->stencil);
+ stencil_offset >>= 8;
+
+ r600_pipe_state_add_reg(rstate, R_02804C_DB_STENCIL_READ_BASE,
+ stencil_offset, &rtex->stencil->resource, RADEON_USAGE_READWRITE);
+ r600_pipe_state_add_reg(rstate, R_028054_DB_STENCIL_WRITE_BASE,
+ stencil_offset, &rtex->stencil->resource, RADEON_USAGE_READWRITE);
+ r600_pipe_state_add_reg(rstate, R_028044_DB_STENCIL_INFO,
+ 1, NULL, 0);
+ } else {
+ r600_pipe_state_add_reg(rstate, R_028044_DB_STENCIL_INFO,
+ 0, NULL, 0);
+ }
+
+ r600_pipe_state_add_reg(rstate, R_02803C_DB_DEPTH_INFO, 0x1, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028040_DB_Z_INFO,
+ /*S_028040_ARRAY_MODE(array_mode) |*/ S_028040_FORMAT(format),
+ NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028058_DB_DEPTH_SIZE,
+ S_028058_PITCH_TILE_MAX(pitch),
+ NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_02805C_DB_DEPTH_SLICE,
+ S_02805C_SLICE_TILE_MAX(slice),
+ NULL, 0);
+}
+
+static void evergreen_set_framebuffer_state(struct pipe_context *ctx,
+ const struct pipe_framebuffer_state *state)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct r600_pipe_state *rstate = CALLOC_STRUCT(r600_pipe_state);
+ uint32_t shader_mask, tl, br;
+ int tl_x, tl_y, br_x, br_y;
+
+ if (rstate == NULL)
+ return;
+
+ r600_flush_framebuffer(rctx, false);
+
+ /* unreference old buffer and reference new one */
+ rstate->id = R600_PIPE_STATE_FRAMEBUFFER;
+
+ util_copy_framebuffer_state(&rctx->framebuffer, state);
+
+ /* build states */
+ rctx->have_depth_fb = 0;
+ rctx->nr_cbufs = state->nr_cbufs;
+ for (int i = 0; i < state->nr_cbufs; i++) {
+ evergreen_cb(rctx, rstate, state, i);
+ }
+ evergreen_db(rctx, rstate, state);
+
+ shader_mask = 0;
+ for (int i = 0; i < state->nr_cbufs; i++) {
+ shader_mask |= 0xf << (i * 4);
+ }
+ tl_x = 0;
+ tl_y = 0;
+ br_x = state->width;
+ br_y = state->height;
+#if 0 /* These shouldn't be necessary on SI, see PA_SC_ENHANCE register */
+ /* EG hw workaround */
+ if (br_x == 0)
+ tl_x = 1;
+ if (br_y == 0)
+ tl_y = 1;
+ /* cayman hw workaround */
+ if (rctx->chip_class == CAYMAN) {
+ if (br_x == 1 && br_y == 1)
+ br_x = 2;
+ }
+#endif
+ tl = S_028240_TL_X(tl_x) | S_028240_TL_Y(tl_y);
+ br = S_028244_BR_X(br_x) | S_028244_BR_Y(br_y);
+
+ r600_pipe_state_add_reg(rstate,
+ R_028240_PA_SC_GENERIC_SCISSOR_TL, tl,
+ NULL, 0);
+ r600_pipe_state_add_reg(rstate,
+ R_028244_PA_SC_GENERIC_SCISSOR_BR, br,
+ NULL, 0);
+ r600_pipe_state_add_reg(rstate,
+ R_028250_PA_SC_VPORT_SCISSOR_0_TL, tl,
+ NULL, 0);
+ r600_pipe_state_add_reg(rstate,
+ R_028254_PA_SC_VPORT_SCISSOR_0_BR, br,
+ NULL, 0);
+ r600_pipe_state_add_reg(rstate,
+ R_028030_PA_SC_SCREEN_SCISSOR_TL, tl,
+ NULL, 0);
+ r600_pipe_state_add_reg(rstate,
+ R_028034_PA_SC_SCREEN_SCISSOR_BR, br,
+ NULL, 0);
+ r600_pipe_state_add_reg(rstate,
+ R_028204_PA_SC_WINDOW_SCISSOR_TL, tl,
+ NULL, 0);
+ r600_pipe_state_add_reg(rstate,
+ R_028208_PA_SC_WINDOW_SCISSOR_BR, br,
+ NULL, 0);
+ r600_pipe_state_add_reg(rstate,
+ R_028200_PA_SC_WINDOW_OFFSET, 0x00000000,
+ NULL, 0);
+ r600_pipe_state_add_reg(rstate,
+ R_028230_PA_SC_EDGERULE, 0xAAAAAAAA,
+ NULL, 0);
+
+ r600_pipe_state_add_reg(rstate, R_02823C_CB_SHADER_MASK,
+ shader_mask, NULL, 0);
+
+ r600_pipe_state_add_reg(rstate, R_028BE0_PA_SC_AA_CONFIG,
+ 0x00000000, NULL, 0);
+
+ free(rctx->states[R600_PIPE_STATE_FRAMEBUFFER]);
+ rctx->states[R600_PIPE_STATE_FRAMEBUFFER] = rstate;
+ r600_context_pipe_state_set(rctx, rstate);
+
+ if (state->zsbuf) {
+ cayman_polygon_offset_update(rctx);
+ }
+}
+
+void cayman_init_state_functions(struct r600_context *rctx)
+{
+ rctx->context.create_blend_state = evergreen_create_blend_state;
+ rctx->context.create_depth_stencil_alpha_state = evergreen_create_dsa_state;
+ rctx->context.create_fs_state = si_create_shader_state;
+ rctx->context.create_rasterizer_state = evergreen_create_rs_state;
+ rctx->context.create_sampler_state = si_create_sampler_state;
+ rctx->context.create_sampler_view = evergreen_create_sampler_view;
+ rctx->context.create_vertex_elements_state = si_create_vertex_elements;
+ rctx->context.create_vs_state = si_create_shader_state;
+ rctx->context.bind_blend_state = r600_bind_blend_state;
+ rctx->context.bind_depth_stencil_alpha_state = r600_bind_dsa_state;
+ rctx->context.bind_fragment_sampler_states = evergreen_bind_ps_sampler;
+ rctx->context.bind_fs_state = r600_bind_ps_shader;
+ rctx->context.bind_rasterizer_state = r600_bind_rs_state;
+ rctx->context.bind_vertex_elements_state = r600_bind_vertex_elements;
+ rctx->context.bind_vertex_sampler_states = evergreen_bind_vs_sampler;
+ rctx->context.bind_vs_state = r600_bind_vs_shader;
+ rctx->context.delete_blend_state = r600_delete_state;
+ rctx->context.delete_depth_stencil_alpha_state = r600_delete_state;
+ rctx->context.delete_fs_state = r600_delete_ps_shader;
+ rctx->context.delete_rasterizer_state = r600_delete_rs_state;
+ rctx->context.delete_sampler_state = si_delete_sampler_state;
+ rctx->context.delete_vertex_elements_state = r600_delete_vertex_element;
+ rctx->context.delete_vs_state = r600_delete_vs_shader;
+ rctx->context.set_blend_color = evergreen_set_blend_color;
+ rctx->context.set_clip_state = evergreen_set_clip_state;
+ rctx->context.set_constant_buffer = r600_set_constant_buffer;
+ rctx->context.set_fragment_sampler_views = evergreen_set_ps_sampler_view;
+ rctx->context.set_framebuffer_state = evergreen_set_framebuffer_state;
+ rctx->context.set_polygon_stipple = evergreen_set_polygon_stipple;
+ rctx->context.set_sample_mask = evergreen_set_sample_mask;
+ rctx->context.set_scissor_state = evergreen_set_scissor_state;
+ rctx->context.set_stencil_ref = r600_set_pipe_stencil_ref;
+ rctx->context.set_vertex_buffers = r600_set_vertex_buffers;
+ rctx->context.set_index_buffer = r600_set_index_buffer;
+ rctx->context.set_vertex_sampler_views = evergreen_set_vs_sampler_view;
+ rctx->context.set_viewport_state = evergreen_set_viewport_state;
+ rctx->context.sampler_view_destroy = r600_sampler_view_destroy;
+ rctx->context.redefine_user_buffer = u_default_redefine_user_buffer;
+ rctx->context.texture_barrier = r600_texture_barrier;
+ rctx->context.create_stream_output_target = r600_create_so_target;
+ rctx->context.stream_output_target_destroy = r600_so_target_destroy;
+ rctx->context.set_stream_output_targets = r600_set_so_targets;
+}
+
+void si_init_config(struct r600_context *rctx)
+{
+ struct r600_pipe_state *rstate = &rctx->config;
+ unsigned tmp;
+
+ r600_pipe_state_add_reg(rstate, R_028A4C_PA_SC_MODE_CNTL_1, 0x0, NULL, 0);
+
+ r600_pipe_state_add_reg(rstate, R_028A10_VGT_OUTPUT_PATH_CNTL, 0x0, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028A14_VGT_HOS_CNTL, 0x0, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028A18_VGT_HOS_MAX_TESS_LEVEL, 0x0, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028A1C_VGT_HOS_MIN_TESS_LEVEL, 0x0, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028A20_VGT_HOS_REUSE_DEPTH, 0x0, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028A24_VGT_GROUP_PRIM_TYPE, 0x0, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028A28_VGT_GROUP_FIRST_DECR, 0x0, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028A2C_VGT_GROUP_DECR, 0x0, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028A30_VGT_GROUP_VECT_0_CNTL, 0x0, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028A34_VGT_GROUP_VECT_1_CNTL, 0x0, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028A38_VGT_GROUP_VECT_0_FMT_CNTL, 0x0, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028A3C_VGT_GROUP_VECT_1_FMT_CNTL, 0x0, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028A40_VGT_GS_MODE, 0x0, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028A84_VGT_PRIMITIVEID_EN, 0x0, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028A8C_VGT_PRIMITIVEID_RESET, 0x0, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028B94_VGT_STRMOUT_CONFIG, 0x0, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028B98_VGT_STRMOUT_BUFFER_CONFIG, 0x0, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028AA8_IA_MULTI_VGT_PARAM, S_028AA8_SWITCH_ON_EOP(1) | S_028AA8_PARTIAL_VS_WAVE_ON(1) | S_028AA8_PRIMGROUP_SIZE(63), NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028AB4_VGT_REUSE_OFF, 0x00000000, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028AB8_VGT_VTX_CNT_EN, 0x0, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_008A14_PA_CL_ENHANCE, (3 << 1) | 1, NULL, 0);
+
+ r600_pipe_state_add_reg(rstate, R_028810_PA_CL_CLIP_CNTL, 0x0, NULL, 0);
+
+ r600_pipe_state_add_reg(rstate, R_028B54_VGT_SHADER_STAGES_EN, 0, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028BD4_PA_SC_CENTROID_PRIORITY_0, 0x76543210, NULL, 0);
+ r600_pipe_state_add_reg(rstate, R_028BD8_PA_SC_CENTROID_PRIORITY_1, 0xfedcba98, NULL, 0);
+
+ r600_pipe_state_add_reg(rstate, R_028804_DB_EQAA, 0x110000, NULL, 0);
+ r600_context_pipe_state_set(rctx, rstate);
+}
+
+void cayman_polygon_offset_update(struct r600_context *rctx)
+{
+ struct r600_pipe_state state;
+
+ state.id = R600_PIPE_STATE_POLYGON_OFFSET;
+ state.nregs = 0;
+ if (rctx->rasterizer && rctx->framebuffer.zsbuf) {
+ float offset_units = rctx->rasterizer->offset_units;
+ unsigned offset_db_fmt_cntl = 0, depth;
+
+ switch (rctx->framebuffer.zsbuf->texture->format) {
+ case PIPE_FORMAT_Z24X8_UNORM:
+ case PIPE_FORMAT_Z24_UNORM_S8_UINT:
+ depth = -24;
+ offset_units *= 2.0f;
+ break;
+ case PIPE_FORMAT_Z32_FLOAT:
+ case PIPE_FORMAT_Z32_FLOAT_S8X24_UINT:
+ depth = -23;
+ offset_units *= 1.0f;
+ offset_db_fmt_cntl |= S_028B78_POLY_OFFSET_DB_IS_FLOAT_FMT(1);
+ break;
+ case PIPE_FORMAT_Z16_UNORM:
+ depth = -16;
+ offset_units *= 4.0f;
+ break;
+ default:
+ return;
+ }
+ /* FIXME some of those reg can be computed with cso */
+ offset_db_fmt_cntl |= S_028B78_POLY_OFFSET_NEG_NUM_DB_BITS(depth);
+ r600_pipe_state_add_reg(&state,
+ R_028B80_PA_SU_POLY_OFFSET_FRONT_SCALE,
+ fui(rctx->rasterizer->offset_scale), NULL, 0);
+ r600_pipe_state_add_reg(&state,
+ R_028B84_PA_SU_POLY_OFFSET_FRONT_OFFSET,
+ fui(offset_units), NULL, 0);
+ r600_pipe_state_add_reg(&state,
+ R_028B88_PA_SU_POLY_OFFSET_BACK_SCALE,
+ fui(rctx->rasterizer->offset_scale), NULL, 0);
+ r600_pipe_state_add_reg(&state,
+ R_028B8C_PA_SU_POLY_OFFSET_BACK_OFFSET,
+ fui(offset_units), NULL, 0);
+ r600_pipe_state_add_reg(&state,
+ R_028B78_PA_SU_POLY_OFFSET_DB_FMT_CNTL,
+ offset_db_fmt_cntl, NULL, 0);
+ r600_context_pipe_state_set(rctx, &state);
+ }
+}
+
+void si_pipe_shader_ps(struct pipe_context *ctx, struct si_pipe_shader *shader)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct r600_pipe_state *rstate = &shader->rstate;
+ struct r600_shader *rshader = &shader->shader;
+ unsigned i, exports_ps, num_cout, spi_ps_in_control, db_shader_control;
+ int pos_index = -1, face_index = -1;
+ int ninterp = 0;
+ boolean have_linear = FALSE, have_centroid = FALSE, have_perspective = FALSE;
+ unsigned spi_baryc_cntl;
+ uint64_t va;
+
+ if (si_pipe_shader_create(ctx, shader))
+ return;
+
+ rstate->nregs = 0;
+
+ db_shader_control = S_02880C_Z_ORDER(V_02880C_EARLY_Z_THEN_LATE_Z);
+ for (i = 0; i < rshader->ninput; i++) {
+ /* evergreen NUM_INTERP only contains values interpolated into the LDS,
+ POSITION goes via GPRs from the SC so isn't counted */
+ if (rshader->input[i].name == TGSI_SEMANTIC_POSITION)
+ pos_index = i;
+ else if (rshader->input[i].name == TGSI_SEMANTIC_FACE)
+ face_index = i;
+ else {
+ ninterp++;
+ if (rshader->input[i].interpolate == TGSI_INTERPOLATE_LINEAR)
+ have_linear = TRUE;
+ if (rshader->input[i].interpolate == TGSI_INTERPOLATE_PERSPECTIVE)
+ have_perspective = TRUE;
+ if (rshader->input[i].centroid)
+ have_centroid = TRUE;
+ }
+ }
+
+ for (i = 0; i < rshader->noutput; i++) {
+ if (rshader->output[i].name == TGSI_SEMANTIC_POSITION)
+ db_shader_control |= S_02880C_Z_EXPORT_ENABLE(1);
+ if (rshader->output[i].name == TGSI_SEMANTIC_STENCIL)
+ db_shader_control |= 0; // XXX OP_VAL or TEST_VAL?
+ }
+ if (rshader->uses_kill)
+ db_shader_control |= S_02880C_KILL_ENABLE(1);
+
+ exports_ps = 0;
+ num_cout = 0;
+ for (i = 0; i < rshader->noutput; i++) {
+ if (rshader->output[i].name == TGSI_SEMANTIC_POSITION ||
+ rshader->output[i].name == TGSI_SEMANTIC_STENCIL)
+ exports_ps |= 1;
+ else if (rshader->output[i].name == TGSI_SEMANTIC_COLOR) {
+ if (rshader->fs_write_all)
+ num_cout = rshader->nr_cbufs;
+ else
+ num_cout++;
+ }
+ }
+ if (!exports_ps) {
+ /* always at least export 1 component per pixel */
+ exports_ps = 2;
+ }
+
+ if (ninterp == 0) {
+ ninterp = 1;
+ have_perspective = TRUE;
+ }
+
+ if (!have_perspective && !have_linear)
+ have_perspective = TRUE;
+
+ spi_ps_in_control = S_0286D8_NUM_INTERP(ninterp);
+
+ spi_baryc_cntl = 0;
+ if (have_perspective)
+ spi_baryc_cntl |= have_centroid ?
+ S_0286E0_PERSP_CENTROID_CNTL(1) : S_0286E0_PERSP_CENTER_CNTL(1);
+ if (have_linear)
+ spi_baryc_cntl |= have_centroid ?
+ S_0286E0_LINEAR_CENTROID_CNTL(1) : S_0286E0_LINEAR_CENTER_CNTL(1);
+
+ r600_pipe_state_add_reg(rstate,
+ R_0286E0_SPI_BARYC_CNTL,
+ spi_baryc_cntl,
+ NULL, 0);
+
+ r600_pipe_state_add_reg(rstate,
+ R_0286CC_SPI_PS_INPUT_ENA,
+ shader->spi_ps_input_ena,
+ NULL, 0);
+
+ r600_pipe_state_add_reg(rstate,
+ R_0286D0_SPI_PS_INPUT_ADDR,
+ shader->spi_ps_input_ena,
+ NULL, 0);
+
+ r600_pipe_state_add_reg(rstate,
+ R_0286D8_SPI_PS_IN_CONTROL,
+ spi_ps_in_control,
+ NULL, 0);
+
+ /* XXX: Depends on Z buffer format? */
+ r600_pipe_state_add_reg(rstate,
+ R_028710_SPI_SHADER_Z_FORMAT,
+ 0,
+ NULL, 0);
+
+ /* XXX: Depends on color buffer format? */
+ r600_pipe_state_add_reg(rstate,
+ R_028714_SPI_SHADER_COL_FORMAT,
+ S_028714_COL0_EXPORT_FORMAT(V_028714_SPI_SHADER_32_ABGR),
+ NULL, 0);
+
+ va = r600_resource_va(ctx->screen, (void *)shader->bo);
+ r600_pipe_state_add_reg(rstate,
+ R_00B020_SPI_SHADER_PGM_LO_PS,
+ va >> 8,
+ shader->bo, RADEON_USAGE_READ);
+ r600_pipe_state_add_reg(rstate,
+ R_00B024_SPI_SHADER_PGM_HI_PS,
+ va >> 40,
+ shader->bo, RADEON_USAGE_READ);
+
+ /* Last 2 reserved SGPRs are used for VCC */
+ /* XXX: Hard-coding 2 SGPRs for constant buffer */
+ r600_pipe_state_add_reg(rstate,
+ R_00B028_SPI_SHADER_PGM_RSRC1_PS,
+ S_00B028_VGPRS(shader->num_vgprs / 4) |
+ S_00B028_SGPRS((shader->num_sgprs + 2 + 2 + 1) / 8),
+ NULL, 0);
+ r600_pipe_state_add_reg(rstate,
+ R_00B02C_SPI_SHADER_PGM_RSRC2_PS,
+ S_00B02C_USER_SGPR(6),
+ NULL, 0);
+
+ r600_pipe_state_add_reg(rstate, R_02880C_DB_SHADER_CONTROL,
+ db_shader_control,
+ NULL, 0);
+
+ shader->sprite_coord_enable = rctx->sprite_coord_enable;
+}
+
+void si_pipe_shader_vs(struct pipe_context *ctx, struct si_pipe_shader *shader)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct r600_pipe_state *rstate = &shader->rstate;
+ struct r600_shader *rshader = &shader->shader;
+ unsigned nparams, i;
+ uint64_t va;
+
+ if (si_pipe_shader_create(ctx, shader))
+ return;
+
+ /* clear previous register */
+ rstate->nregs = 0;
+
+ /* Certain attributes (position, psize, etc.) don't count as params.
+ * VS is required to export at least one param and r600_shader_from_tgsi()
+ * takes care of adding a dummy export.
+ */
+ for (nparams = 0, i = 0 ; i < rshader->noutput; i++) {
+ if (rshader->output[i].name != TGSI_SEMANTIC_POSITION)
+ nparams++;
+ }
+ if (nparams < 1)
+ nparams = 1;
+
+ r600_pipe_state_add_reg(rstate,
+ R_0286C4_SPI_VS_OUT_CONFIG,
+ S_0286C4_VS_EXPORT_COUNT(nparams - 1),
+ NULL, 0);
+
+ r600_pipe_state_add_reg(rstate,
+ R_02870C_SPI_SHADER_POS_FORMAT,
+ S_02870C_POS0_EXPORT_FORMAT(V_02870C_SPI_SHADER_4COMP) |
+ S_02870C_POS1_EXPORT_FORMAT(V_02870C_SPI_SHADER_4COMP) |
+ S_02870C_POS2_EXPORT_FORMAT(V_02870C_SPI_SHADER_4COMP) |
+ S_02870C_POS3_EXPORT_FORMAT(V_02870C_SPI_SHADER_4COMP),
+ NULL, 0);
+
+ va = r600_resource_va(ctx->screen, (void *)shader->bo);
+ r600_pipe_state_add_reg(rstate,
+ R_00B120_SPI_SHADER_PGM_LO_VS,
+ va >> 8,
+ shader->bo, RADEON_USAGE_READ);
+ r600_pipe_state_add_reg(rstate,
+ R_00B124_SPI_SHADER_PGM_HI_VS,
+ va >> 40,
+ shader->bo, RADEON_USAGE_READ);
+
+ /* Last 2 reserved SGPRs are used for VCC */
+ /* XXX: Hard-coding 2 SGPRs for constant buffer */
+ r600_pipe_state_add_reg(rstate,
+ R_00B128_SPI_SHADER_PGM_RSRC1_VS,
+ S_00B128_VGPRS(shader->num_vgprs / 4) |
+ S_00B128_SGPRS((shader->num_sgprs + 2 + 2 + 2) / 8),
+ NULL, 0);
+ r600_pipe_state_add_reg(rstate,
+ R_00B12C_SPI_SHADER_PGM_RSRC2_VS,
+ S_00B12C_USER_SGPR(2 + 2),
+ NULL, 0);
+}
+
+void si_update_spi_map(struct r600_context *rctx)
+{
+ struct r600_shader *ps = &rctx->ps_shader->shader;
+ struct r600_shader *vs = &rctx->vs_shader->shader;
+ struct r600_pipe_state *rstate = &rctx->spi;
+ unsigned i, j, tmp;
+
+ rstate->nregs = 0;
+
+ for (i = 0; i < ps->ninput; i++) {
+ tmp = 0;
+
+ if (ps->input[i].name == TGSI_SEMANTIC_COLOR ||
+ ps->input[i].name == TGSI_SEMANTIC_BCOLOR ||
+ ps->input[i].name == TGSI_SEMANTIC_POSITION) {
+ tmp |= S_028644_FLAT_SHADE(1);
+ }
+
+ if (ps->input[i].name == TGSI_SEMANTIC_GENERIC &&
+ rctx->sprite_coord_enable & (1 << ps->input[i].sid)) {
+ tmp |= S_028644_PT_SPRITE_TEX(1);
+ }
+
+ for (j = 0; j < vs->noutput; j++) {
+ if (ps->input[i].name == vs->output[j].name &&
+ ps->input[i].sid == vs->output[j].sid) {
+ tmp |= S_028644_OFFSET(ps->input[i].sid);
+ break;
+ }
+ }
+
+ if (j == vs->noutput) {
+ /* No corresponding output found, load defaults into input */
+ tmp |= S_028644_OFFSET(0x20);
+ }
+
+ r600_pipe_state_add_reg(rstate, R_028644_SPI_PS_INPUT_CNTL_0 + i * 4,
+ tmp, NULL, 0);
+ }
+
+ if (rstate->nregs > 0)
+ r600_context_pipe_state_set(rctx, rstate);
+}
+
+void *cayman_create_db_flush_dsa(struct r600_context *rctx)
+{
+ struct pipe_depth_stencil_alpha_state dsa;
+ struct r600_pipe_state *rstate;
+
+ memset(&dsa, 0, sizeof(dsa));
+
+ rstate = rctx->context.create_depth_stencil_alpha_state(&rctx->context, &dsa);
+ r600_pipe_state_add_reg(rstate,
+ R_028000_DB_RENDER_CONTROL,
+ S_028000_DEPTH_COPY(1) |
+ S_028000_STENCIL_COPY(1) |
+ S_028000_COPY_CENTROID(1),
+ NULL, 0);
+ return rstate;
+}
diff --git a/src/gallium/drivers/radeonsi/r600.h b/src/gallium/drivers/radeonsi/r600.h
new file mode 100644
index 00000000000..56915ab966f
--- /dev/null
+++ b/src/gallium/drivers/radeonsi/r600.h
@@ -0,0 +1,245 @@
+/*
+ * Copyright 2010 Jerome Glisse <[email protected]>
+ *
+ * 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.
+ *
+ * Authors:
+ * Jerome Glisse
+ */
+#ifndef R600_H
+#define R600_H
+
+#include "../../winsys/radeon/drm/radeon_winsys.h"
+#include "util/u_double_list.h"
+#include "util/u_vbuf.h"
+
+#define R600_ERR(fmt, args...) \
+ fprintf(stderr, "EE %s:%d %s - "fmt, __FILE__, __LINE__, __func__, ##args)
+
+struct winsys_handle;
+
+enum radeon_family {
+ CHIP_UNKNOWN,
+ CHIP_CAYMAN,
+ CHIP_TAHITI,
+ CHIP_PITCAIRN,
+ CHIP_VERDE,
+ CHIP_LAST,
+};
+
+enum chip_class {
+ CAYMAN,
+ TAHITI,
+};
+
+struct r600_tiling_info {
+ unsigned num_channels;
+ unsigned num_banks;
+ unsigned group_bytes;
+};
+
+struct r600_resource {
+ struct u_vbuf_resource b;
+
+ /* Winsys objects. */
+ struct pb_buffer *buf;
+ struct radeon_winsys_cs_handle *cs_buf;
+
+ /* Resource state. */
+ unsigned domains;
+};
+
+/* R600/R700 STATES */
+#define R600_GROUP_MAX 16
+#define R600_BLOCK_MAX_BO 32
+#define R600_BLOCK_MAX_REG 128
+
+/* each range covers 9 bits of dword space = 512 dwords = 2k bytes */
+/* there is a block entry for each register so 512 blocks */
+/* we have no registers to read/write below 0x8000 (0x2000 in dw space) */
+/* we use some fake offsets at 0x40000 to do evergreen sampler borders so take 0x42000 as a max bound*/
+#define RANGE_OFFSET_START 0x8000
+#define HASH_SHIFT 9
+#define NUM_RANGES (0x42000 - RANGE_OFFSET_START) / (4 << HASH_SHIFT) /* 128 << 9 = 64k */
+
+#define CTX_RANGE_ID(offset) ((((offset - RANGE_OFFSET_START) >> 2) >> HASH_SHIFT) & 255)
+#define CTX_BLOCK_ID(offset) (((offset - RANGE_OFFSET_START) >> 2) & ((1 << HASH_SHIFT) - 1))
+
+struct r600_pipe_reg {
+ uint32_t value;
+ struct r600_block *block;
+ struct r600_resource *bo;
+ enum radeon_bo_usage bo_usage;
+ uint32_t id;
+};
+
+struct r600_pipe_state {
+ unsigned id;
+ unsigned nregs;
+ struct r600_pipe_reg regs[R600_BLOCK_MAX_REG];
+};
+
+#define R600_BLOCK_STATUS_ENABLED (1 << 0)
+#define R600_BLOCK_STATUS_DIRTY (1 << 1)
+
+struct r600_block_reloc {
+ struct r600_resource *bo;
+ enum radeon_bo_usage bo_usage;
+ unsigned bo_pm4_index;
+};
+
+struct r600_block {
+ struct list_head list;
+ struct list_head enable_list;
+ unsigned status;
+ unsigned flags;
+ unsigned start_offset;
+ unsigned pm4_ndwords;
+ unsigned nbo;
+ uint16_t nreg;
+ uint16_t nreg_dirty;
+ uint32_t *reg;
+ uint32_t pm4[R600_BLOCK_MAX_REG];
+ unsigned pm4_bo_index[R600_BLOCK_MAX_REG];
+ struct r600_block_reloc reloc[R600_BLOCK_MAX_BO];
+};
+
+struct r600_range {
+ struct r600_block **blocks;
+};
+
+struct r600_query {
+ union {
+ uint64_t u64;
+ boolean b;
+ struct pipe_query_data_so_statistics so;
+ } result;
+ /* The kind of query */
+ unsigned type;
+ /* Offset of the first result for current query */
+ unsigned results_start;
+ /* Offset of the next free result after current query data */
+ unsigned results_end;
+ /* Size of the result in memory for both begin_query and end_query,
+ * this can be one or two numbers, or it could even be a size of a structure. */
+ unsigned result_size;
+ /* The buffer where query results are stored. It's used as a ring,
+ * data blocks for current query are stored sequentially from
+ * results_start to results_end, with wrapping on the buffer end */
+ struct r600_resource *buffer;
+ /* The number of dwords for begin_query or end_query. */
+ unsigned num_cs_dw;
+ /* linked list of queries */
+ struct list_head list;
+};
+
+struct r600_so_target {
+ struct pipe_stream_output_target b;
+
+ /* The buffer where BUFFER_FILLED_SIZE is stored. */
+ struct r600_resource *filled_size;
+ unsigned stride;
+ unsigned so_index;
+};
+
+#define R600_CONTEXT_DRAW_PENDING (1 << 0)
+#define R600_CONTEXT_DST_CACHES_DIRTY (1 << 1)
+#define R600_CONTEXT_CHECK_EVENT_FLUSH (1 << 2)
+
+struct r600_draw {
+ uint32_t vgt_num_indices;
+ uint32_t vgt_num_instances;
+ uint32_t vgt_index_type;
+ uint32_t vgt_draw_initiator;
+ uint32_t indices_bo_offset;
+ unsigned db_render_override;
+ unsigned db_render_control;
+ struct r600_resource *indices;
+};
+
+struct r600_context;
+struct r600_screen;
+
+void r600_get_backend_mask(struct r600_context *ctx);
+void r600_context_fini(struct r600_context *ctx);
+void r600_context_pipe_state_set(struct r600_context *ctx, struct r600_pipe_state *state);
+void r600_context_flush(struct r600_context *ctx, unsigned flags);
+void r600_context_draw(struct r600_context *ctx, const struct r600_draw *draw);
+
+struct r600_query *r600_context_query_create(struct r600_context *ctx, unsigned query_type);
+void r600_context_query_destroy(struct r600_context *ctx, struct r600_query *query);
+boolean r600_context_query_result(struct r600_context *ctx,
+ struct r600_query *query,
+ boolean wait, void *vresult);
+void r600_query_begin(struct r600_context *ctx, struct r600_query *query);
+void r600_query_end(struct r600_context *ctx, struct r600_query *query);
+void r600_context_queries_suspend(struct r600_context *ctx);
+void r600_context_queries_resume(struct r600_context *ctx);
+void r600_query_predication(struct r600_context *ctx, struct r600_query *query, int operation,
+ int flag_wait);
+void r600_context_emit_fence(struct r600_context *ctx, struct r600_resource *fence,
+ unsigned offset, unsigned value);
+void r600_inval_shader_cache(struct r600_context *ctx);
+void r600_inval_texture_cache(struct r600_context *ctx);
+void r600_inval_vertex_cache(struct r600_context *ctx);
+void r600_flush_framebuffer(struct r600_context *ctx, bool flush_now);
+
+void r600_context_streamout_begin(struct r600_context *ctx);
+void r600_context_streamout_end(struct r600_context *ctx);
+void r600_context_draw_opaque_count(struct r600_context *ctx, struct r600_so_target *t);
+void r600_need_cs_space(struct r600_context *ctx, unsigned num_dw, boolean count_draw_in);
+void r600_context_block_emit_dirty(struct r600_context *ctx, struct r600_block *block);
+void r600_context_block_resource_emit_dirty(struct r600_context *ctx, struct r600_block *block);
+
+int si_context_init(struct r600_context *ctx);
+void evergreen_context_draw(struct r600_context *ctx, const struct r600_draw *draw);
+
+void _r600_pipe_state_add_reg(struct r600_context *ctx,
+ struct r600_pipe_state *state,
+ uint32_t offset, uint32_t value,
+ uint32_t range_id, uint32_t block_id,
+ struct r600_resource *bo,
+ enum radeon_bo_usage usage);
+
+void r600_pipe_state_add_reg_noblock(struct r600_pipe_state *state,
+ uint32_t offset, uint32_t value,
+ struct r600_resource *bo,
+ enum radeon_bo_usage usage);
+
+#define r600_pipe_state_add_reg(state, offset, value, bo, usage) _r600_pipe_state_add_reg(rctx, state, offset, value, CTX_RANGE_ID(offset), CTX_BLOCK_ID(offset), bo, usage)
+
+static inline void r600_pipe_state_mod_reg(struct r600_pipe_state *state,
+ uint32_t value)
+{
+ state->regs[state->nregs].value = value;
+ state->nregs++;
+}
+
+static inline void r600_pipe_state_mod_reg_bo(struct r600_pipe_state *state,
+ uint32_t value, struct r600_resource *bo,
+ enum radeon_bo_usage usage)
+{
+ state->regs[state->nregs].value = value;
+ state->regs[state->nregs].bo = bo;
+ state->regs[state->nregs].bo_usage = usage;
+ state->nregs++;
+}
+
+#endif
diff --git a/src/gallium/drivers/radeonsi/r600_blit.c b/src/gallium/drivers/radeonsi/r600_blit.c
new file mode 100644
index 00000000000..65158089acb
--- /dev/null
+++ b/src/gallium/drivers/radeonsi/r600_blit.c
@@ -0,0 +1,379 @@
+/*
+ * Copyright 2010 Jerome Glisse <[email protected]>
+ *
+ * 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"
+
+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_DECOMPRESS = R600_SAVE_FRAMEBUFFER | R600_DISABLE_RENDER_COND,
+};
+
+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->states[R600_PIPE_STATE_BLEND]);
+ util_blitter_save_depth_stencil_alpha(rctx->blitter, rctx->states[R600_PIPE_STATE_DSA]);
+ if (rctx->states[R600_PIPE_STATE_STENCIL_REF]) {
+ util_blitter_save_stencil_ref(rctx->blitter, &rctx->stencil_ref);
+ }
+ util_blitter_save_rasterizer(rctx->blitter, rctx->states[R600_PIPE_STATE_RASTERIZER]);
+ 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->states[R600_PIPE_STATE_VIEWPORT]) {
+ util_blitter_save_viewport(rctx->blitter, &rctx->viewport);
+ }
+ util_blitter_save_vertex_buffers(rctx->blitter,
+ rctx->vbuf_mgr->nr_vertex_buffers,
+ rctx->vbuf_mgr->vertex_buffer);
+ 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->ps_samplers.n_samplers,
+ (void**)rctx->ps_samplers.samplers);
+
+ util_blitter_save_fragment_sampler_views(
+ rctx->blitter, rctx->ps_samplers.n_views,
+ (struct pipe_sampler_view**)rctx->ps_samplers.views);
+ }
+
+ if ((op & R600_DISABLE_RENDER_COND) && rctx->current_render_cond) {
+ rctx->saved_render_cond = rctx->current_render_cond;
+ rctx->saved_render_cond_mode = rctx->current_render_cond_mode;
+ rctx->context.render_condition(&rctx->context, NULL, 0);
+ }
+
+}
+
+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_mode);
+ rctx->saved_render_cond = NULL;
+ }
+ r600_context_queries_resume(rctx);
+}
+
+static unsigned u_num_layers(struct pipe_resource *r, unsigned level)
+{
+ switch (r->target) {
+ case PIPE_TEXTURE_CUBE:
+ return 6;
+ case PIPE_TEXTURE_3D:
+ return u_minify(r->depth0, level);
+ case PIPE_TEXTURE_1D_ARRAY:
+ return r->array_size;
+ case PIPE_TEXTURE_2D_ARRAY:
+ return r->array_size;
+ default:
+ return 1;
+ }
+}
+
+void r600_blit_uncompress_depth(struct pipe_context *ctx, struct r600_resource_texture *texture)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ unsigned layer, level;
+ float depth = 1.0f;
+
+ if (!texture->dirty_db)
+ return;
+
+ for (level = 0; level <= texture->resource.b.b.b.last_level; level++) {
+ unsigned num_layers = u_num_layers(&texture->resource.b.b.b, level);
+
+ for (layer = 0; layer < num_layers; layer++) {
+ 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;
+ surf_tmpl.usage = PIPE_BIND_DEPTH_STENCIL;
+
+ zsurf = ctx->create_surface(ctx, &texture->resource.b.b.b, &surf_tmpl);
+
+ surf_tmpl.format = texture->flushed_depth_texture->real_format;
+ surf_tmpl.usage = PIPE_BIND_RENDER_TARGET;
+ cbsurf = ctx->create_surface(ctx,
+ (struct pipe_resource*)texture->flushed_depth_texture, &surf_tmpl);
+
+ r600_blitter_begin(ctx, R600_DECOMPRESS);
+ util_blitter_custom_depth_stencil(rctx->blitter, zsurf, cbsurf, rctx->custom_dsa_flush, depth);
+ r600_blitter_end(ctx);
+
+ pipe_surface_reference(&zsurf, NULL);
+ pipe_surface_reference(&cbsurf, NULL);
+ }
+ }
+
+ texture->dirty_db = FALSE;
+}
+
+void r600_flush_depth_textures(struct r600_context *rctx)
+{
+ unsigned int i;
+
+ /* FIXME: This handles fragment shader textures only. */
+
+ for (i = 0; i < rctx->ps_samplers.n_views; ++i) {
+ struct si_pipe_sampler_view *view;
+ struct r600_resource_texture *tex;
+
+ view = rctx->ps_samplers.views[i];
+ if (!view) continue;
+
+ tex = (struct r600_resource_texture *)view->base.texture;
+ if (!tex->depth)
+ continue;
+
+ if (tex->is_flushing_texture)
+ continue;
+
+ r600_blit_uncompress_depth(&rctx->context, tex);
+ }
+
+ /* also check CB here */
+ for (i = 0; i < rctx->framebuffer.nr_cbufs; i++) {
+ struct r600_resource_texture *tex;
+ tex = (struct r600_resource_texture *)rctx->framebuffer.cbufs[i]->texture;
+
+ if (!tex->depth)
+ continue;
+
+ if (tex->is_flushing_texture)
+ continue;
+
+ r600_blit_uncompress_depth(&rctx->context, tex);
+ }
+}
+
+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;
+
+ r600_blitter_begin(ctx, R600_CLEAR);
+ util_blitter_clear(rctx->blitter, fb->width, fb->height,
+ fb->nr_cbufs, buffers, fb->nr_cbufs ? fb->cbufs[0]->format : PIPE_FORMAT_NONE,
+ color, depth, stencil);
+ 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);
+}
+
+
+
+/* Copy a block of pixels from one surface to another using HW. */
+static void r600_hw_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;
+
+ r600_blitter_begin(ctx, R600_COPY);
+ util_blitter_copy_texture(rctx->blitter, dst, dst_level, dstx, dsty, dstz,
+ src, src_level, src_box, TRUE);
+ r600_blitter_end(ctx);
+}
+
+struct texture_orig_info {
+ unsigned format;
+ unsigned width0;
+ unsigned height0;
+};
+
+static void r600_compressed_to_blittable(struct pipe_resource *tex,
+ unsigned level,
+ struct texture_orig_info *orig)
+{
+ struct r600_resource_texture *rtex = (struct r600_resource_texture*)tex;
+ 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;
+
+ 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;
+}
+
+static void r600_reset_blittable_to_compressed(struct pipe_resource *tex,
+ struct texture_orig_info *orig)
+{
+ tex->format = orig->format;
+ tex->width0 = orig->width0;
+ tex->height0 = orig->height0;
+}
+
+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_resource_texture *rsrc = (struct r600_resource_texture*)src;
+ struct texture_orig_info orig_info[2];
+ struct pipe_box sbox;
+ const struct pipe_box *psbox;
+ 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;
+ }
+
+ if (rsrc->depth && !rsrc->is_flushing_texture)
+ r600_texture_depth_flush(ctx, src, FALSE);
+
+ restore_orig[0] = restore_orig[1] = FALSE;
+
+ if (util_format_is_compressed(src->format)) {
+ 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;
+ } else
+ psbox=src_box;
+
+ if (util_format_is_compressed(dst->format)) {
+ 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);
+ }
+
+ r600_hw_copy_region(ctx, dst, dst_level, dstx, dsty, dstz,
+ src, src_level, psbox);
+
+ if (restore_orig[0])
+ r600_reset_blittable_to_compressed(src, &orig_info[0]);
+
+ if (restore_orig[1])
+ r600_reset_blittable_to_compressed(dst, &orig_info[1]);
+}
+
+void r600_init_blit_functions(struct r600_context *rctx)
+{
+ 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;
+}
+
+void r600_blit_push_depth(struct pipe_context *ctx, struct r600_resource_texture *texture)
+{
+ struct pipe_box sbox;
+
+ sbox.x = sbox.y = sbox.z = 0;
+ sbox.width = texture->resource.b.b.b.width0;
+ sbox.height = texture->resource.b.b.b.height0;
+ /* XXX that might be wrong */
+ sbox.depth = 1;
+
+ r600_hw_copy_region(ctx, (struct pipe_resource *)texture, 0,
+ 0, 0, 0,
+ (struct pipe_resource *)texture->flushed_depth_texture, 0,
+ &sbox);
+}
diff --git a/src/gallium/drivers/radeonsi/r600_buffer.c b/src/gallium/drivers/radeonsi/r600_buffer.c
new file mode 100644
index 00000000000..bb885df8dcd
--- /dev/null
+++ b/src/gallium/drivers/radeonsi/r600_buffer.c
@@ -0,0 +1,282 @@
+/*
+ * Copyright 2010 Jerome Glisse <[email protected]>
+ *
+ * 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.
+ *
+ * Authors:
+ * Jerome Glisse
+ * Corbin Simpson <[email protected]>
+ */
+#include <byteswap.h>
+
+#include "pipe/p_screen.h"
+#include "util/u_format.h"
+#include "util/u_math.h"
+#include "util/u_inlines.h"
+#include "util/u_memory.h"
+#include "util/u_upload_mgr.h"
+
+#include "r600.h"
+#include "radeonsi_pipe.h"
+
+static void r600_buffer_destroy(struct pipe_screen *screen,
+ struct pipe_resource *buf)
+{
+ struct r600_screen *rscreen = (struct r600_screen*)screen;
+ struct r600_resource *rbuffer = r600_resource(buf);
+
+ pb_reference(&rbuffer->buf, NULL);
+ util_slab_free(&rscreen->pool_buffers, rbuffer);
+}
+
+static struct pipe_transfer *r600_get_transfer(struct pipe_context *ctx,
+ struct pipe_resource *resource,
+ unsigned level,
+ unsigned usage,
+ const struct pipe_box *box)
+{
+ struct r600_context *rctx = (struct r600_context*)ctx;
+ struct pipe_transfer *transfer = util_slab_alloc(&rctx->pool_transfers);
+
+ transfer->resource = resource;
+ transfer->level = level;
+ transfer->usage = usage;
+ transfer->box = *box;
+ transfer->stride = 0;
+ transfer->layer_stride = 0;
+ transfer->data = NULL;
+
+ /* Note strides are zero, this is ok for buffers, but not for
+ * textures 2d & higher at least.
+ */
+ return transfer;
+}
+
+static void *r600_buffer_transfer_map(struct pipe_context *pipe,
+ struct pipe_transfer *transfer)
+{
+ struct r600_resource *rbuffer = r600_resource(transfer->resource);
+ struct r600_context *rctx = (struct r600_context*)pipe;
+ uint8_t *data;
+
+ if (rbuffer->b.user_ptr)
+ return (uint8_t*)rbuffer->b.user_ptr + transfer->box.x;
+
+ data = rctx->ws->buffer_map(rbuffer->buf, rctx->cs, transfer->usage);
+ if (!data)
+ return NULL;
+
+ return (uint8_t*)data + transfer->box.x;
+}
+
+static void r600_buffer_transfer_unmap(struct pipe_context *pipe,
+ struct pipe_transfer *transfer)
+{
+ struct r600_resource *rbuffer = r600_resource(transfer->resource);
+ struct r600_context *rctx = (struct r600_context*)pipe;
+
+ if (rbuffer->b.user_ptr)
+ return;
+
+ rctx->ws->buffer_unmap(rbuffer->buf);
+}
+
+static void r600_buffer_transfer_flush_region(struct pipe_context *pipe,
+ struct pipe_transfer *transfer,
+ const struct pipe_box *box)
+{
+}
+
+static void r600_transfer_destroy(struct pipe_context *ctx,
+ struct pipe_transfer *transfer)
+{
+ struct r600_context *rctx = (struct r600_context*)ctx;
+ util_slab_free(&rctx->pool_transfers, transfer);
+}
+
+static void r600_buffer_transfer_inline_write(struct pipe_context *pipe,
+ struct pipe_resource *resource,
+ unsigned level,
+ unsigned usage,
+ const struct pipe_box *box,
+ const void *data,
+ unsigned stride,
+ unsigned layer_stride)
+{
+ struct r600_context *rctx = (struct r600_context*)pipe;
+ struct r600_resource *rbuffer = r600_resource(resource);
+ uint8_t *map = NULL;
+
+ assert(rbuffer->b.user_ptr == NULL);
+
+ map = rctx->ws->buffer_map(rbuffer->buf, rctx->cs,
+ PIPE_TRANSFER_WRITE | PIPE_TRANSFER_DISCARD_RANGE | usage);
+
+ memcpy(map + box->x, data, box->width);
+
+ rctx->ws->buffer_unmap(rbuffer->buf);
+}
+
+static const struct u_resource_vtbl r600_buffer_vtbl =
+{
+ u_default_resource_get_handle, /* get_handle */
+ r600_buffer_destroy, /* resource_destroy */
+ r600_get_transfer, /* get_transfer */
+ r600_transfer_destroy, /* transfer_destroy */
+ r600_buffer_transfer_map, /* transfer_map */
+ r600_buffer_transfer_flush_region, /* transfer_flush_region */
+ r600_buffer_transfer_unmap, /* transfer_unmap */
+ r600_buffer_transfer_inline_write /* transfer_inline_write */
+};
+
+bool r600_init_resource(struct r600_screen *rscreen,
+ struct r600_resource *res,
+ unsigned size, unsigned alignment,
+ unsigned bind, unsigned usage)
+{
+ uint32_t initial_domain, domains;
+
+ /* Staging resources particpate in transfers and blits only
+ * and are used for uploads and downloads from regular
+ * resources. We generate them internally for some transfers.
+ */
+ if (usage == PIPE_USAGE_STAGING) {
+ domains = RADEON_DOMAIN_GTT;
+ initial_domain = RADEON_DOMAIN_GTT;
+ } else {
+ domains = RADEON_DOMAIN_GTT | RADEON_DOMAIN_VRAM;
+
+ switch(usage) {
+ case PIPE_USAGE_DYNAMIC:
+ case PIPE_USAGE_STREAM:
+ case PIPE_USAGE_STAGING:
+ initial_domain = RADEON_DOMAIN_GTT;
+ break;
+ case PIPE_USAGE_DEFAULT:
+ case PIPE_USAGE_STATIC:
+ case PIPE_USAGE_IMMUTABLE:
+ default:
+ initial_domain = RADEON_DOMAIN_VRAM;
+ break;
+ }
+ }
+
+ res->buf = rscreen->ws->buffer_create(rscreen->ws, size, alignment, bind, initial_domain);
+ if (!res->buf) {
+ return false;
+ }
+
+ res->cs_buf = rscreen->ws->buffer_get_cs_handle(res->buf);
+ res->domains = domains;
+ return true;
+}
+
+struct pipe_resource *r600_buffer_create(struct pipe_screen *screen,
+ const struct pipe_resource *templ)
+{
+ struct r600_screen *rscreen = (struct r600_screen*)screen;
+ struct r600_resource *rbuffer;
+ /* XXX We probably want a different alignment for buffers and textures. */
+ unsigned alignment = 4096;
+
+ rbuffer = util_slab_alloc(&rscreen->pool_buffers);
+
+ rbuffer->b.b.b = *templ;
+ pipe_reference_init(&rbuffer->b.b.b.reference, 1);
+ rbuffer->b.b.b.screen = screen;
+ rbuffer->b.b.vtbl = &r600_buffer_vtbl;
+ rbuffer->b.user_ptr = NULL;
+
+ if (!r600_init_resource(rscreen, rbuffer, templ->width0, alignment, templ->bind, templ->usage)) {
+ util_slab_free(&rscreen->pool_buffers, rbuffer);
+ return NULL;
+ }
+ return &rbuffer->b.b.b;
+}
+
+struct pipe_resource *r600_user_buffer_create(struct pipe_screen *screen,
+ void *ptr, unsigned bytes,
+ unsigned bind)
+{
+ struct r600_screen *rscreen = (struct r600_screen*)screen;
+ struct r600_resource *rbuffer;
+
+ rbuffer = util_slab_alloc(&rscreen->pool_buffers);
+
+ pipe_reference_init(&rbuffer->b.b.b.reference, 1);
+ rbuffer->b.b.vtbl = &r600_buffer_vtbl;
+ rbuffer->b.b.b.screen = screen;
+ rbuffer->b.b.b.target = PIPE_BUFFER;
+ rbuffer->b.b.b.format = PIPE_FORMAT_R8_UNORM;
+ rbuffer->b.b.b.usage = PIPE_USAGE_IMMUTABLE;
+ rbuffer->b.b.b.bind = bind;
+ rbuffer->b.b.b.width0 = bytes;
+ rbuffer->b.b.b.height0 = 1;
+ rbuffer->b.b.b.depth0 = 1;
+ rbuffer->b.b.b.array_size = 1;
+ rbuffer->b.b.b.flags = 0;
+ rbuffer->b.user_ptr = ptr;
+ rbuffer->buf = NULL;
+ return &rbuffer->b.b.b;
+}
+
+void r600_upload_index_buffer(struct r600_context *rctx,
+ struct pipe_index_buffer *ib, unsigned count)
+{
+ struct r600_resource *rbuffer = r600_resource(ib->buffer);
+
+ u_upload_data(rctx->vbuf_mgr->uploader, 0, count * ib->index_size,
+ rbuffer->b.user_ptr, &ib->offset, &ib->buffer);
+}
+
+void r600_upload_const_buffer(struct r600_context *rctx, struct r600_resource **rbuffer,
+ uint32_t *const_offset)
+{
+ if ((*rbuffer)->b.user_ptr) {
+ uint8_t *ptr = (*rbuffer)->b.user_ptr;
+ unsigned size = (*rbuffer)->b.b.b.width0;
+
+ *rbuffer = NULL;
+
+ if (R600_BIG_ENDIAN) {
+ uint32_t *tmpPtr;
+ unsigned i;
+
+ if (!(tmpPtr = malloc(size))) {
+ R600_ERR("Failed to allocate BE swap buffer.\n");
+ return;
+ }
+
+ for (i = 0; i < size / 4; ++i) {
+ tmpPtr[i] = bswap_32(((uint32_t *)ptr)[i]);
+ }
+
+ u_upload_data(rctx->vbuf_mgr->uploader, 0, size, tmpPtr, const_offset,
+ (struct pipe_resource**)rbuffer);
+
+ free(tmpPtr);
+ } else {
+ u_upload_data(rctx->vbuf_mgr->uploader, 0, size, ptr, const_offset,
+ (struct pipe_resource**)rbuffer);
+ }
+ } else {
+ *const_offset = 0;
+ }
+}
diff --git a/src/gallium/drivers/radeonsi/r600_hw_context.c b/src/gallium/drivers/radeonsi/r600_hw_context.c
new file mode 100644
index 00000000000..494b0d34283
--- /dev/null
+++ b/src/gallium/drivers/radeonsi/r600_hw_context.c
@@ -0,0 +1,1151 @@
+/*
+ * Copyright 2010 Jerome Glisse <[email protected]>
+ *
+ * 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.
+ *
+ * Authors:
+ * Jerome Glisse
+ */
+#include "r600_hw_context_priv.h"
+#include "radeonsi_pipe.h"
+#include "sid.h"
+#include "util/u_memory.h"
+#include <errno.h>
+
+#define GROUP_FORCE_NEW_BLOCK 0
+
+/* Get backends mask */
+void r600_get_backend_mask(struct r600_context *ctx)
+{
+ struct radeon_winsys_cs *cs = ctx->cs;
+ struct r600_resource *buffer;
+ uint32_t *results;
+ unsigned num_backends = ctx->screen->info.r600_num_backends;
+ unsigned i, mask = 0;
+
+ /* if backend_map query is supported by the kernel */
+ if (ctx->screen->info.r600_backend_map_valid) {
+ unsigned num_tile_pipes = ctx->screen->info.r600_num_tile_pipes;
+ unsigned backend_map = ctx->screen->info.r600_backend_map;
+ unsigned item_width, item_mask;
+
+ if (ctx->chip_class >= CAYMAN) {
+ item_width = 4;
+ item_mask = 0x7;
+ }
+
+ while(num_tile_pipes--) {
+ i = backend_map & item_mask;
+ mask |= (1<<i);
+ backend_map >>= item_width;
+ }
+ if (mask != 0) {
+ ctx->backend_mask = mask;
+ return;
+ }
+ }
+
+ /* otherwise backup path for older kernels */
+
+ /* create buffer for event data */
+ buffer = (struct r600_resource*)
+ pipe_buffer_create(&ctx->screen->screen, PIPE_BIND_CUSTOM,
+ PIPE_USAGE_STAGING, ctx->max_db*16);
+ if (!buffer)
+ goto err;
+
+ /* initialize buffer with zeroes */
+ results = ctx->ws->buffer_map(buffer->buf, ctx->cs, PIPE_TRANSFER_WRITE);
+ if (results) {
+ uint64_t va = 0;
+
+ memset(results, 0, ctx->max_db * 4 * 4);
+ ctx->ws->buffer_unmap(buffer->buf);
+
+ /* emit EVENT_WRITE for ZPASS_DONE */
+ va = r600_resource_va(&ctx->screen->screen, (void *)buffer);
+ cs->buf[cs->cdw++] = PKT3(PKT3_EVENT_WRITE, 2, 0);
+ cs->buf[cs->cdw++] = EVENT_TYPE(EVENT_TYPE_ZPASS_DONE) | EVENT_INDEX(1);
+ cs->buf[cs->cdw++] = va;
+ cs->buf[cs->cdw++] = va >> 32;
+
+ cs->buf[cs->cdw++] = PKT3(PKT3_NOP, 0, 0);
+ cs->buf[cs->cdw++] = r600_context_bo_reloc(ctx, buffer, RADEON_USAGE_WRITE);
+
+ /* analyze results */
+ results = ctx->ws->buffer_map(buffer->buf, ctx->cs, PIPE_TRANSFER_READ);
+ if (results) {
+ for(i = 0; i < ctx->max_db; i++) {
+ /* at least highest bit will be set if backend is used */
+ if (results[i*4 + 1])
+ mask |= (1<<i);
+ }
+ ctx->ws->buffer_unmap(buffer->buf);
+ }
+ }
+
+ pipe_resource_reference((struct pipe_resource**)&buffer, NULL);
+
+ if (mask != 0) {
+ ctx->backend_mask = mask;
+ return;
+ }
+
+err:
+ /* fallback to old method - set num_backends lower bits to 1 */
+ ctx->backend_mask = (~((uint32_t)0))>>(32-num_backends);
+ return;
+}
+
+static inline void r600_context_ps_partial_flush(struct r600_context *ctx)
+{
+ struct radeon_winsys_cs *cs = ctx->cs;
+
+ if (!(ctx->flags & R600_CONTEXT_DRAW_PENDING))
+ return;
+
+ cs->buf[cs->cdw++] = PKT3(PKT3_EVENT_WRITE, 0, 0);
+ cs->buf[cs->cdw++] = EVENT_TYPE(EVENT_TYPE_PS_PARTIAL_FLUSH) | EVENT_INDEX(4);
+
+ ctx->flags &= ~R600_CONTEXT_DRAW_PENDING;
+}
+
+void r600_init_cs(struct r600_context *ctx)
+{
+ struct radeon_winsys_cs *cs = ctx->cs;
+
+ /* All asics require this one */
+ cs->buf[cs->cdw++] = PKT3(PKT3_CONTEXT_CONTROL, 1, 0);
+ cs->buf[cs->cdw++] = 0x80000000;
+ cs->buf[cs->cdw++] = 0x80000000;
+
+ ctx->init_dwords = cs->cdw;
+}
+
+static void r600_init_block(struct r600_context *ctx,
+ struct r600_block *block,
+ const struct r600_reg *reg, int index, int nreg,
+ unsigned opcode, unsigned offset_base)
+{
+ int i = index;
+ int j, n = nreg;
+
+ /* initialize block */
+ block->flags = 0;
+ block->status |= R600_BLOCK_STATUS_DIRTY; /* dirty all blocks at start */
+ block->start_offset = reg[i].offset;
+ block->pm4[block->pm4_ndwords++] = PKT3(opcode, n, 0);
+ block->pm4[block->pm4_ndwords++] = (block->start_offset - offset_base) >> 2;
+ block->reg = &block->pm4[block->pm4_ndwords];
+ block->pm4_ndwords += n;
+ block->nreg = n;
+ block->nreg_dirty = n;
+ LIST_INITHEAD(&block->list);
+ LIST_INITHEAD(&block->enable_list);
+
+ for (j = 0; j < n; j++) {
+ if (reg[i+j].flags & REG_FLAG_DIRTY_ALWAYS) {
+ block->flags |= REG_FLAG_DIRTY_ALWAYS;
+ }
+ if (reg[i+j].flags & REG_FLAG_ENABLE_ALWAYS) {
+ if (!(block->status & R600_BLOCK_STATUS_ENABLED)) {
+ block->status |= R600_BLOCK_STATUS_ENABLED;
+ LIST_ADDTAIL(&block->enable_list, &ctx->enable_list);
+ LIST_ADDTAIL(&block->list,&ctx->dirty);
+ }
+ }
+ if (reg[i+j].flags & REG_FLAG_FLUSH_CHANGE) {
+ block->flags |= REG_FLAG_FLUSH_CHANGE;
+ }
+
+ if (reg[i+j].flags & REG_FLAG_NEED_BO) {
+ block->nbo++;
+ assert(block->nbo < R600_BLOCK_MAX_BO);
+ block->pm4_bo_index[j] = block->nbo;
+ block->pm4[block->pm4_ndwords++] = PKT3(PKT3_NOP, 0, 0);
+ block->pm4[block->pm4_ndwords++] = 0x00000000;
+ block->reloc[block->nbo].bo_pm4_index = block->pm4_ndwords - 1;
+ }
+ }
+ /* check that we stay in limit */
+ assert(block->pm4_ndwords < R600_BLOCK_MAX_REG);
+}
+
+int r600_context_add_block(struct r600_context *ctx, const struct r600_reg *reg, unsigned nreg,
+ unsigned opcode, unsigned offset_base)
+{
+ struct r600_block *block;
+ struct r600_range *range;
+ int offset;
+
+ for (unsigned i = 0, n = 0; i < nreg; i += n) {
+ /* ignore new block balise */
+ if (reg[i].offset == GROUP_FORCE_NEW_BLOCK) {
+ n = 1;
+ continue;
+ }
+
+ /* register that need relocation are in their own group */
+ /* find number of consecutive registers */
+ n = 0;
+ offset = reg[i].offset;
+ while (reg[i + n].offset == offset) {
+ n++;
+ offset += 4;
+ if ((n + i) >= nreg)
+ break;
+ if (n >= (R600_BLOCK_MAX_REG - 2))
+ break;
+ }
+
+ /* allocate new block */
+ block = calloc(1, sizeof(struct r600_block));
+ if (block == NULL) {
+ return -ENOMEM;
+ }
+ ctx->nblocks++;
+ for (int j = 0; j < n; j++) {
+ range = &ctx->range[CTX_RANGE_ID(reg[i + j].offset)];
+ /* create block table if it doesn't exist */
+ if (!range->blocks)
+ range->blocks = calloc(1 << HASH_SHIFT, sizeof(void *));
+ if (!range->blocks)
+ return -1;
+
+ range->blocks[CTX_BLOCK_ID(reg[i + j].offset)] = block;
+ }
+
+ r600_init_block(ctx, block, reg, i, n, opcode, offset_base);
+
+ }
+ return 0;
+}
+
+
+/* initialize */
+void r600_context_fini(struct r600_context *ctx)
+{
+ struct r600_block *block;
+ struct r600_range *range;
+
+ for (int i = 0; i < NUM_RANGES; i++) {
+ if (!ctx->range[i].blocks)
+ continue;
+ for (int j = 0; j < (1 << HASH_SHIFT); j++) {
+ block = ctx->range[i].blocks[j];
+ if (block) {
+ for (int k = 0, offset = block->start_offset; k < block->nreg; k++, offset += 4) {
+ range = &ctx->range[CTX_RANGE_ID(offset)];
+ range->blocks[CTX_BLOCK_ID(offset)] = NULL;
+ }
+ for (int k = 1; k <= block->nbo; k++) {
+ pipe_resource_reference((struct pipe_resource**)&block->reloc[k].bo, NULL);
+ }
+ free(block);
+ }
+ }
+ free(ctx->range[i].blocks);
+ }
+ free(ctx->range);
+ free(ctx->blocks);
+ ctx->ws->cs_destroy(ctx->cs);
+}
+
+int r600_setup_block_table(struct r600_context *ctx)
+{
+ /* setup block table */
+ int c = 0;
+ ctx->blocks = calloc(ctx->nblocks, sizeof(void*));
+ if (!ctx->blocks)
+ return -ENOMEM;
+ for (int i = 0; i < NUM_RANGES; i++) {
+ if (!ctx->range[i].blocks)
+ continue;
+ for (int j = 0, add; j < (1 << HASH_SHIFT); j++) {
+ if (!ctx->range[i].blocks[j])
+ continue;
+
+ add = 1;
+ for (int k = 0; k < c; k++) {
+ if (ctx->blocks[k] == ctx->range[i].blocks[j]) {
+ add = 0;
+ break;
+ }
+ }
+ if (add) {
+ assert(c < ctx->nblocks);
+ ctx->blocks[c++] = ctx->range[i].blocks[j];
+ j += (ctx->range[i].blocks[j]->nreg) - 1;
+ }
+ }
+ }
+
+ return 0;
+}
+
+void r600_need_cs_space(struct r600_context *ctx, unsigned num_dw,
+ boolean count_draw_in)
+{
+ struct r600_atom *state;
+
+ /* The number of dwords we already used in the CS so far. */
+ num_dw += ctx->cs->cdw;
+
+ if (count_draw_in) {
+ /* The number of dwords all the dirty states would take. */
+ LIST_FOR_EACH_ENTRY(state, &ctx->dirty_states, head) {
+ num_dw += state->num_dw;
+ }
+
+ num_dw += ctx->pm4_dirty_cdwords;
+
+ /* The upper-bound of how much a draw command would take. */
+ num_dw += R600_MAX_DRAW_CS_DWORDS;
+ }
+
+ /* Count in queries_suspend. */
+ num_dw += ctx->num_cs_dw_queries_suspend;
+
+ /* Count in streamout_end at the end of CS. */
+ num_dw += ctx->num_cs_dw_streamout_end;
+
+ /* Count in render_condition(NULL) at the end of CS. */
+ if (ctx->predicate_drawing) {
+ num_dw += 3;
+ }
+
+ /* Count in framebuffer cache flushes at the end of CS. */
+ num_dw += 7; /* one SURFACE_SYNC and CACHE_FLUSH_AND_INV (r6xx-only) */
+
+ /* Save 16 dwords for the fence mechanism. */
+ num_dw += 16;
+
+ /* Flush if there's not enough space. */
+ if (num_dw > RADEON_MAX_CMDBUF_DWORDS) {
+ radeonsi_flush(&ctx->context, NULL, RADEON_FLUSH_ASYNC);
+ }
+}
+
+void r600_context_dirty_block(struct r600_context *ctx,
+ struct r600_block *block,
+ int dirty, int index)
+{
+ if ((index + 1) > block->nreg_dirty)
+ block->nreg_dirty = index + 1;
+
+ if ((dirty != (block->status & R600_BLOCK_STATUS_DIRTY)) || !(block->status & R600_BLOCK_STATUS_ENABLED)) {
+ block->status |= R600_BLOCK_STATUS_DIRTY;
+ ctx->pm4_dirty_cdwords += block->pm4_ndwords;
+ if (!(block->status & R600_BLOCK_STATUS_ENABLED)) {
+ block->status |= R600_BLOCK_STATUS_ENABLED;
+ LIST_ADDTAIL(&block->enable_list, &ctx->enable_list);
+ }
+ LIST_ADDTAIL(&block->list,&ctx->dirty);
+
+ if (block->flags & REG_FLAG_FLUSH_CHANGE) {
+ r600_context_ps_partial_flush(ctx);
+ }
+ }
+}
+
+void r600_context_pipe_state_set(struct r600_context *ctx, struct r600_pipe_state *state)
+{
+ struct r600_block *block;
+ int dirty;
+ for (int i = 0; i < state->nregs; i++) {
+ unsigned id, reloc_id;
+ struct r600_pipe_reg *reg = &state->regs[i];
+
+ block = reg->block;
+ id = reg->id;
+
+ dirty = block->status & R600_BLOCK_STATUS_DIRTY;
+
+ if (reg->value != block->reg[id]) {
+ block->reg[id] = reg->value;
+ dirty |= R600_BLOCK_STATUS_DIRTY;
+ }
+ if (block->flags & REG_FLAG_DIRTY_ALWAYS)
+ dirty |= R600_BLOCK_STATUS_DIRTY;
+ if (block->pm4_bo_index[id]) {
+ /* find relocation */
+ reloc_id = block->pm4_bo_index[id];
+ pipe_resource_reference((struct pipe_resource**)&block->reloc[reloc_id].bo, &reg->bo->b.b.b);
+ block->reloc[reloc_id].bo_usage = reg->bo_usage;
+ /* always force dirty for relocs for now */
+ dirty |= R600_BLOCK_STATUS_DIRTY;
+ }
+
+ if (dirty)
+ r600_context_dirty_block(ctx, block, dirty, id);
+ }
+}
+
+struct r600_resource *r600_context_reg_bo(struct r600_context *ctx, unsigned offset)
+{
+ struct r600_range *range;
+ struct r600_block *block;
+ unsigned id;
+
+ range = &ctx->range[CTX_RANGE_ID(offset)];
+ block = range->blocks[CTX_BLOCK_ID(offset)];
+ offset -= block->start_offset;
+ id = block->pm4_bo_index[offset >> 2];
+ if (block->reloc[id].bo) {
+ return block->reloc[id].bo;
+ }
+ return NULL;
+}
+
+void r600_context_block_emit_dirty(struct r600_context *ctx, struct r600_block *block)
+{
+ struct radeon_winsys_cs *cs = ctx->cs;
+ int optional = block->nbo == 0 && !(block->flags & REG_FLAG_DIRTY_ALWAYS);
+ int cp_dwords = block->pm4_ndwords, start_dword = 0;
+ int new_dwords = 0;
+ int nbo = block->nbo;
+
+ if (block->nreg_dirty == 0 && optional) {
+ goto out;
+ }
+
+ if (nbo) {
+ ctx->flags |= R600_CONTEXT_CHECK_EVENT_FLUSH;
+
+ for (int j = 0; j < block->nreg; j++) {
+ if (block->pm4_bo_index[j]) {
+ /* find relocation */
+ struct r600_block_reloc *reloc = &block->reloc[block->pm4_bo_index[j]];
+ block->pm4[reloc->bo_pm4_index] =
+ r600_context_bo_reloc(ctx, reloc->bo, reloc->bo_usage);
+ nbo--;
+ if (nbo == 0)
+ break;
+ }
+ }
+ ctx->flags &= ~R600_CONTEXT_CHECK_EVENT_FLUSH;
+ }
+
+ optional &= (block->nreg_dirty != block->nreg);
+ if (optional) {
+ new_dwords = block->nreg_dirty;
+ start_dword = cs->cdw;
+ cp_dwords = new_dwords + 2;
+ }
+ memcpy(&cs->buf[cs->cdw], block->pm4, cp_dwords * 4);
+ cs->cdw += cp_dwords;
+
+ if (optional) {
+ uint32_t newword;
+
+ newword = cs->buf[start_dword];
+ newword &= PKT_COUNT_C;
+ newword |= PKT_COUNT_S(new_dwords);
+ cs->buf[start_dword] = newword;
+ }
+out:
+ block->status ^= R600_BLOCK_STATUS_DIRTY;
+ block->nreg_dirty = 0;
+ LIST_DELINIT(&block->list);
+}
+
+void r600_inval_shader_cache(struct r600_context *ctx)
+{
+ ctx->atom_surface_sync.flush_flags |= S_0085F0_SH_ICACHE_ACTION_ENA(1);
+ ctx->atom_surface_sync.flush_flags |= S_0085F0_SH_KCACHE_ACTION_ENA(1);
+ r600_atom_dirty(ctx, &ctx->atom_surface_sync.atom);
+}
+
+void r600_inval_texture_cache(struct r600_context *ctx)
+{
+ ctx->atom_surface_sync.flush_flags |= S_0085F0_TC_ACTION_ENA(1);
+ r600_atom_dirty(ctx, &ctx->atom_surface_sync.atom);
+}
+
+void r600_inval_vertex_cache(struct r600_context *ctx)
+{
+ /* Some GPUs don't have the vertex cache and must use the texture cache instead. */
+ ctx->atom_surface_sync.flush_flags |= S_0085F0_TC_ACTION_ENA(1);
+ r600_atom_dirty(ctx, &ctx->atom_surface_sync.atom);
+}
+
+void r600_flush_framebuffer(struct r600_context *ctx, bool flush_now)
+{
+ if (!(ctx->flags & R600_CONTEXT_DST_CACHES_DIRTY))
+ return;
+
+ ctx->atom_surface_sync.flush_flags |=
+ r600_get_cb_flush_flags(ctx) |
+ (ctx->framebuffer.zsbuf ? S_0085F0_DB_ACTION_ENA(1) | S_0085F0_DB_DEST_BASE_ENA(1) : 0);
+
+ if (flush_now) {
+ r600_emit_atom(ctx, &ctx->atom_surface_sync.atom);
+ } else {
+ r600_atom_dirty(ctx, &ctx->atom_surface_sync.atom);
+ }
+
+ ctx->flags &= ~R600_CONTEXT_DST_CACHES_DIRTY;
+}
+
+void r600_context_flush(struct r600_context *ctx, unsigned flags)
+{
+ struct radeon_winsys_cs *cs = ctx->cs;
+ struct r600_block *enable_block = NULL;
+ bool queries_suspended = false;
+ bool streamout_suspended = false;
+
+ if (cs->cdw == ctx->init_dwords)
+ return;
+
+ /* suspend queries */
+ if (ctx->num_cs_dw_queries_suspend) {
+ r600_context_queries_suspend(ctx);
+ queries_suspended = true;
+ }
+
+ if (ctx->num_cs_dw_streamout_end) {
+ r600_context_streamout_end(ctx);
+ streamout_suspended = true;
+ }
+
+ r600_flush_framebuffer(ctx, true);
+
+ /* partial flush is needed to avoid lockups on some chips with user fences */
+ cs->buf[cs->cdw++] = PKT3(PKT3_EVENT_WRITE, 0, 0);
+ cs->buf[cs->cdw++] = EVENT_TYPE(EVENT_TYPE_PS_PARTIAL_FLUSH) | EVENT_INDEX(4);
+
+ /* Flush the CS. */
+ ctx->ws->cs_flush(ctx->cs, flags);
+
+ ctx->pm4_dirty_cdwords = 0;
+ ctx->flags = 0;
+
+ r600_init_cs(ctx);
+
+ if (streamout_suspended) {
+ ctx->streamout_start = TRUE;
+ ctx->streamout_append_bitmask = ~0;
+ }
+
+ /* resume queries */
+ if (queries_suspended) {
+ r600_context_queries_resume(ctx);
+ }
+
+ /* set all valid group as dirty so they get reemited on
+ * next draw command
+ */
+ LIST_FOR_EACH_ENTRY(enable_block, &ctx->enable_list, enable_list) {
+ if(!(enable_block->status & R600_BLOCK_STATUS_DIRTY)) {
+ LIST_ADDTAIL(&enable_block->list,&ctx->dirty);
+ enable_block->status |= R600_BLOCK_STATUS_DIRTY;
+ }
+ ctx->pm4_dirty_cdwords += enable_block->pm4_ndwords;
+ enable_block->nreg_dirty = enable_block->nreg;
+ }
+}
+
+void r600_context_emit_fence(struct r600_context *ctx, struct r600_resource *fence_bo, unsigned offset, unsigned value)
+{
+ struct radeon_winsys_cs *cs = ctx->cs;
+ uint64_t va;
+
+ r600_need_cs_space(ctx, 10, FALSE);
+
+ va = r600_resource_va(&ctx->screen->screen, (void*)fence_bo);
+ va = va + (offset << 2);
+
+ cs->buf[cs->cdw++] = PKT3(PKT3_EVENT_WRITE, 0, 0);
+ cs->buf[cs->cdw++] = EVENT_TYPE(EVENT_TYPE_PS_PARTIAL_FLUSH) | EVENT_INDEX(4);
+ cs->buf[cs->cdw++] = PKT3(PKT3_EVENT_WRITE_EOP, 4, 0);
+ cs->buf[cs->cdw++] = EVENT_TYPE(EVENT_TYPE_CACHE_FLUSH_AND_INV_TS_EVENT) | EVENT_INDEX(5);
+ cs->buf[cs->cdw++] = va & 0xFFFFFFFFUL; /* ADDRESS_LO */
+ /* DATA_SEL | INT_EN | ADDRESS_HI */
+ cs->buf[cs->cdw++] = (1 << 29) | (0 << 24) | ((va >> 32UL) & 0xFF);
+ cs->buf[cs->cdw++] = value; /* DATA_LO */
+ cs->buf[cs->cdw++] = 0; /* DATA_HI */
+ cs->buf[cs->cdw++] = PKT3(PKT3_NOP, 0, 0);
+ cs->buf[cs->cdw++] = r600_context_bo_reloc(ctx, fence_bo, RADEON_USAGE_WRITE);
+}
+
+static unsigned r600_query_read_result(char *map, unsigned start_index, unsigned end_index,
+ bool test_status_bit)
+{
+ uint32_t *current_result = (uint32_t*)map;
+ uint64_t start, end;
+
+ start = (uint64_t)current_result[start_index] |
+ (uint64_t)current_result[start_index+1] << 32;
+ end = (uint64_t)current_result[end_index] |
+ (uint64_t)current_result[end_index+1] << 32;
+
+ if (!test_status_bit ||
+ ((start & 0x8000000000000000UL) && (end & 0x8000000000000000UL))) {
+ return end - start;
+ }
+ return 0;
+}
+
+static boolean r600_query_result(struct r600_context *ctx, struct r600_query *query, boolean wait)
+{
+ unsigned results_base = query->results_start;
+ char *map;
+
+ map = ctx->ws->buffer_map(query->buffer->buf, ctx->cs,
+ PIPE_TRANSFER_READ |
+ (wait ? 0 : PIPE_TRANSFER_DONTBLOCK));
+ if (!map)
+ return FALSE;
+
+ /* count all results across all data blocks */
+ switch (query->type) {
+ case PIPE_QUERY_OCCLUSION_COUNTER:
+ while (results_base != query->results_end) {
+ query->result.u64 +=
+ r600_query_read_result(map + results_base, 0, 2, true);
+ results_base = (results_base + 16) % query->buffer->b.b.b.width0;
+ }
+ break;
+ case PIPE_QUERY_OCCLUSION_PREDICATE:
+ while (results_base != query->results_end) {
+ query->result.b = query->result.b ||
+ r600_query_read_result(map + results_base, 0, 2, true) != 0;
+ results_base = (results_base + 16) % query->buffer->b.b.b.width0;
+ }
+ break;
+ case PIPE_QUERY_TIME_ELAPSED:
+ while (results_base != query->results_end) {
+ query->result.u64 +=
+ r600_query_read_result(map + results_base, 0, 2, false);
+ results_base = (results_base + query->result_size) % query->buffer->b.b.b.width0;
+ }
+ break;
+ case PIPE_QUERY_PRIMITIVES_EMITTED:
+ /* SAMPLE_STREAMOUTSTATS stores this structure:
+ * {
+ * u64 NumPrimitivesWritten;
+ * u64 PrimitiveStorageNeeded;
+ * }
+ * We only need NumPrimitivesWritten here. */
+ while (results_base != query->results_end) {
+ query->result.u64 +=
+ r600_query_read_result(map + results_base, 2, 6, true);
+ results_base = (results_base + query->result_size) % query->buffer->b.b.b.width0;
+ }
+ break;
+ case PIPE_QUERY_PRIMITIVES_GENERATED:
+ /* Here we read PrimitiveStorageNeeded. */
+ while (results_base != query->results_end) {
+ query->result.u64 +=
+ r600_query_read_result(map + results_base, 0, 4, true);
+ results_base = (results_base + query->result_size) % query->buffer->b.b.b.width0;
+ }
+ break;
+ case PIPE_QUERY_SO_STATISTICS:
+ while (results_base != query->results_end) {
+ query->result.so.num_primitives_written +=
+ r600_query_read_result(map + results_base, 2, 6, true);
+ query->result.so.primitives_storage_needed +=
+ r600_query_read_result(map + results_base, 0, 4, true);
+ results_base = (results_base + query->result_size) % query->buffer->b.b.b.width0;
+ }
+ break;
+ case PIPE_QUERY_SO_OVERFLOW_PREDICATE:
+ while (results_base != query->results_end) {
+ query->result.b = query->result.b ||
+ r600_query_read_result(map + results_base, 2, 6, true) !=
+ r600_query_read_result(map + results_base, 0, 4, true);
+ results_base = (results_base + query->result_size) % query->buffer->b.b.b.width0;
+ }
+ break;
+ default:
+ assert(0);
+ }
+
+ query->results_start = query->results_end;
+ ctx->ws->buffer_unmap(query->buffer->buf);
+ return TRUE;
+}
+
+void r600_query_begin(struct r600_context *ctx, struct r600_query *query)
+{
+ struct radeon_winsys_cs *cs = ctx->cs;
+ unsigned new_results_end, i;
+ uint32_t *results;
+ uint64_t va;
+
+ r600_need_cs_space(ctx, query->num_cs_dw * 2, TRUE);
+
+ new_results_end = (query->results_end + query->result_size) % query->buffer->b.b.b.width0;
+
+ /* collect current results if query buffer is full */
+ if (new_results_end == query->results_start) {
+ r600_query_result(ctx, query, TRUE);
+ }
+
+ switch (query->type) {
+ case PIPE_QUERY_OCCLUSION_COUNTER:
+ case PIPE_QUERY_OCCLUSION_PREDICATE:
+ results = ctx->ws->buffer_map(query->buffer->buf, ctx->cs, PIPE_TRANSFER_WRITE);
+ if (results) {
+ results = (uint32_t*)((char*)results + query->results_end);
+ memset(results, 0, query->result_size);
+
+ /* Set top bits for unused backends */
+ for (i = 0; i < ctx->max_db; i++) {
+ if (!(ctx->backend_mask & (1<<i))) {
+ results[(i * 4)+1] = 0x80000000;
+ results[(i * 4)+3] = 0x80000000;
+ }
+ }
+ ctx->ws->buffer_unmap(query->buffer->buf);
+ }
+ break;
+ case PIPE_QUERY_TIME_ELAPSED:
+ break;
+ case PIPE_QUERY_PRIMITIVES_EMITTED:
+ case PIPE_QUERY_PRIMITIVES_GENERATED:
+ case PIPE_QUERY_SO_STATISTICS:
+ case PIPE_QUERY_SO_OVERFLOW_PREDICATE:
+ results = ctx->ws->buffer_map(query->buffer->buf, ctx->cs, PIPE_TRANSFER_WRITE);
+ results = (uint32_t*)((char*)results + query->results_end);
+ memset(results, 0, query->result_size);
+ ctx->ws->buffer_unmap(query->buffer->buf);
+ break;
+ default:
+ assert(0);
+ }
+
+ /* emit begin query */
+ va = r600_resource_va(&ctx->screen->screen, (void*)query->buffer);
+ va += query->results_end;
+
+ switch (query->type) {
+ case PIPE_QUERY_OCCLUSION_COUNTER:
+ case PIPE_QUERY_OCCLUSION_PREDICATE:
+ cs->buf[cs->cdw++] = PKT3(PKT3_EVENT_WRITE, 2, 0);
+ cs->buf[cs->cdw++] = EVENT_TYPE(EVENT_TYPE_ZPASS_DONE) | EVENT_INDEX(1);
+ cs->buf[cs->cdw++] = va;
+ cs->buf[cs->cdw++] = (va >> 32UL) & 0xFF;
+ break;
+ case PIPE_QUERY_PRIMITIVES_EMITTED:
+ case PIPE_QUERY_PRIMITIVES_GENERATED:
+ case PIPE_QUERY_SO_STATISTICS:
+ case PIPE_QUERY_SO_OVERFLOW_PREDICATE:
+ cs->buf[cs->cdw++] = PKT3(PKT3_EVENT_WRITE, 2, 0);
+ cs->buf[cs->cdw++] = EVENT_TYPE(EVENT_TYPE_SAMPLE_STREAMOUTSTATS) | EVENT_INDEX(3);
+ cs->buf[cs->cdw++] = query->results_end;
+ cs->buf[cs->cdw++] = 0;
+ break;
+ case PIPE_QUERY_TIME_ELAPSED:
+ cs->buf[cs->cdw++] = PKT3(PKT3_EVENT_WRITE_EOP, 4, 0);
+ cs->buf[cs->cdw++] = EVENT_TYPE(EVENT_TYPE_CACHE_FLUSH_AND_INV_TS_EVENT) | EVENT_INDEX(5);
+ cs->buf[cs->cdw++] = va;
+ cs->buf[cs->cdw++] = (3 << 29) | ((va >> 32UL) & 0xFF);
+ cs->buf[cs->cdw++] = 0;
+ cs->buf[cs->cdw++] = 0;
+ break;
+ default:
+ assert(0);
+ }
+ cs->buf[cs->cdw++] = PKT3(PKT3_NOP, 0, 0);
+ cs->buf[cs->cdw++] = r600_context_bo_reloc(ctx, query->buffer, RADEON_USAGE_WRITE);
+
+ ctx->num_cs_dw_queries_suspend += query->num_cs_dw;
+}
+
+void r600_query_end(struct r600_context *ctx, struct r600_query *query)
+{
+ struct radeon_winsys_cs *cs = ctx->cs;
+ uint64_t va;
+
+ va = r600_resource_va(&ctx->screen->screen, (void*)query->buffer);
+ /* emit end query */
+ switch (query->type) {
+ case PIPE_QUERY_OCCLUSION_COUNTER:
+ case PIPE_QUERY_OCCLUSION_PREDICATE:
+ va += query->results_end + 8;
+ cs->buf[cs->cdw++] = PKT3(PKT3_EVENT_WRITE, 2, 0);
+ cs->buf[cs->cdw++] = EVENT_TYPE(EVENT_TYPE_ZPASS_DONE) | EVENT_INDEX(1);
+ cs->buf[cs->cdw++] = va;
+ cs->buf[cs->cdw++] = (va >> 32UL) & 0xFF;
+ break;
+ case PIPE_QUERY_PRIMITIVES_EMITTED:
+ case PIPE_QUERY_PRIMITIVES_GENERATED:
+ case PIPE_QUERY_SO_STATISTICS:
+ case PIPE_QUERY_SO_OVERFLOW_PREDICATE:
+ cs->buf[cs->cdw++] = PKT3(PKT3_EVENT_WRITE, 2, 0);
+ cs->buf[cs->cdw++] = EVENT_TYPE(EVENT_TYPE_SAMPLE_STREAMOUTSTATS) | EVENT_INDEX(3);
+ cs->buf[cs->cdw++] = query->results_end + query->result_size/2;
+ cs->buf[cs->cdw++] = 0;
+ break;
+ case PIPE_QUERY_TIME_ELAPSED:
+ va += query->results_end + query->result_size/2;
+ cs->buf[cs->cdw++] = PKT3(PKT3_EVENT_WRITE_EOP, 4, 0);
+ cs->buf[cs->cdw++] = EVENT_TYPE(EVENT_TYPE_CACHE_FLUSH_AND_INV_TS_EVENT) | EVENT_INDEX(5);
+ cs->buf[cs->cdw++] = va;
+ cs->buf[cs->cdw++] = (3 << 29) | ((va >> 32UL) & 0xFF);
+ cs->buf[cs->cdw++] = 0;
+ cs->buf[cs->cdw++] = 0;
+ break;
+ default:
+ assert(0);
+ }
+ cs->buf[cs->cdw++] = PKT3(PKT3_NOP, 0, 0);
+ cs->buf[cs->cdw++] = r600_context_bo_reloc(ctx, query->buffer, RADEON_USAGE_WRITE);
+
+ query->results_end = (query->results_end + query->result_size) % query->buffer->b.b.b.width0;
+ ctx->num_cs_dw_queries_suspend -= query->num_cs_dw;
+}
+
+void r600_query_predication(struct r600_context *ctx, struct r600_query *query, int operation,
+ int flag_wait)
+{
+ struct radeon_winsys_cs *cs = ctx->cs;
+ uint64_t va;
+
+ if (operation == PREDICATION_OP_CLEAR) {
+ r600_need_cs_space(ctx, 3, FALSE);
+
+ cs->buf[cs->cdw++] = PKT3(PKT3_SET_PREDICATION, 1, 0);
+ cs->buf[cs->cdw++] = 0;
+ cs->buf[cs->cdw++] = PRED_OP(PREDICATION_OP_CLEAR);
+ } else {
+ unsigned results_base = query->results_start;
+ unsigned count;
+ uint32_t op;
+
+ /* find count of the query data blocks */
+ count = (query->buffer->b.b.b.width0 + query->results_end - query->results_start) % query->buffer->b.b.b.width0;
+ count /= query->result_size;
+
+ r600_need_cs_space(ctx, 5 * count, TRUE);
+
+ op = PRED_OP(operation) | PREDICATION_DRAW_VISIBLE |
+ (flag_wait ? PREDICATION_HINT_WAIT : PREDICATION_HINT_NOWAIT_DRAW);
+ va = r600_resource_va(&ctx->screen->screen, (void*)query->buffer);
+
+ /* emit predicate packets for all data blocks */
+ while (results_base != query->results_end) {
+ cs->buf[cs->cdw++] = PKT3(PKT3_SET_PREDICATION, 1, 0);
+ cs->buf[cs->cdw++] = (va + results_base) & 0xFFFFFFFFUL;
+ cs->buf[cs->cdw++] = op | (((va + results_base) >> 32UL) & 0xFF);
+ cs->buf[cs->cdw++] = PKT3(PKT3_NOP, 0, 0);
+ cs->buf[cs->cdw++] = r600_context_bo_reloc(ctx, query->buffer,
+ RADEON_USAGE_READ);
+ results_base = (results_base + query->result_size) % query->buffer->b.b.b.width0;
+
+ /* set CONTINUE bit for all packets except the first */
+ op |= PREDICATION_CONTINUE;
+ }
+ }
+}
+
+struct r600_query *r600_context_query_create(struct r600_context *ctx, unsigned query_type)
+{
+ struct r600_query *query;
+ unsigned buffer_size = 4096;
+
+ query = CALLOC_STRUCT(r600_query);
+ if (query == NULL)
+ return NULL;
+
+ query->type = query_type;
+
+ switch (query_type) {
+ case PIPE_QUERY_OCCLUSION_COUNTER:
+ case PIPE_QUERY_OCCLUSION_PREDICATE:
+ query->result_size = 16 * ctx->max_db;
+ query->num_cs_dw = 6;
+ break;
+ case PIPE_QUERY_TIME_ELAPSED:
+ query->result_size = 16;
+ query->num_cs_dw = 8;
+ break;
+ case PIPE_QUERY_PRIMITIVES_EMITTED:
+ case PIPE_QUERY_PRIMITIVES_GENERATED:
+ case PIPE_QUERY_SO_STATISTICS:
+ case PIPE_QUERY_SO_OVERFLOW_PREDICATE:
+ /* NumPrimitivesWritten, PrimitiveStorageNeeded. */
+ query->result_size = 32;
+ query->num_cs_dw = 6;
+ break;
+ default:
+ assert(0);
+ FREE(query);
+ return NULL;
+ }
+
+ /* adjust buffer size to simplify offsets wrapping math */
+ buffer_size -= buffer_size % query->result_size;
+
+ /* Queries are normally read by the CPU after
+ * being written by the gpu, hence staging is probably a good
+ * usage pattern.
+ */
+ query->buffer = (struct r600_resource*)
+ pipe_buffer_create(&ctx->screen->screen, PIPE_BIND_CUSTOM, PIPE_USAGE_STAGING, buffer_size);
+ if (!query->buffer) {
+ FREE(query);
+ return NULL;
+ }
+ return query;
+}
+
+void r600_context_query_destroy(struct r600_context *ctx, struct r600_query *query)
+{
+ pipe_resource_reference((struct pipe_resource**)&query->buffer, NULL);
+ free(query);
+}
+
+boolean r600_context_query_result(struct r600_context *ctx,
+ struct r600_query *query,
+ boolean wait, void *vresult)
+{
+ boolean *result_b = (boolean*)vresult;
+ uint64_t *result_u64 = (uint64_t*)vresult;
+ struct pipe_query_data_so_statistics *result_so =
+ (struct pipe_query_data_so_statistics*)vresult;
+
+ if (!r600_query_result(ctx, query, wait))
+ return FALSE;
+
+ switch (query->type) {
+ case PIPE_QUERY_OCCLUSION_COUNTER:
+ case PIPE_QUERY_PRIMITIVES_EMITTED:
+ case PIPE_QUERY_PRIMITIVES_GENERATED:
+ *result_u64 = query->result.u64;
+ break;
+ case PIPE_QUERY_OCCLUSION_PREDICATE:
+ case PIPE_QUERY_SO_OVERFLOW_PREDICATE:
+ *result_b = query->result.b;
+ break;
+ case PIPE_QUERY_TIME_ELAPSED:
+ *result_u64 = (1000000 * query->result.u64) / ctx->screen->info.r600_clock_crystal_freq;
+ break;
+ case PIPE_QUERY_SO_STATISTICS:
+ *result_so = query->result.so;
+ break;
+ default:
+ assert(0);
+ }
+ return TRUE;
+}
+
+void r600_context_queries_suspend(struct r600_context *ctx)
+{
+ struct r600_query *query;
+
+ LIST_FOR_EACH_ENTRY(query, &ctx->active_query_list, list) {
+ r600_query_end(ctx, query);
+ }
+ assert(ctx->num_cs_dw_queries_suspend == 0);
+}
+
+void r600_context_queries_resume(struct r600_context *ctx)
+{
+ struct r600_query *query;
+
+ assert(ctx->num_cs_dw_queries_suspend == 0);
+
+ LIST_FOR_EACH_ENTRY(query, &ctx->active_query_list, list) {
+ r600_query_begin(ctx, query);
+ }
+}
+
+void r600_context_streamout_begin(struct r600_context *ctx)
+{
+ struct radeon_winsys_cs *cs = ctx->cs;
+ struct r600_so_target **t = ctx->so_targets;
+ unsigned *strides = ctx->vs_shader_so_strides;
+ unsigned buffer_en, i;
+
+ buffer_en = (ctx->num_so_targets >= 1 && t[0] ? 1 : 0) |
+ (ctx->num_so_targets >= 2 && t[1] ? 2 : 0) |
+ (ctx->num_so_targets >= 3 && t[2] ? 4 : 0) |
+ (ctx->num_so_targets >= 4 && t[3] ? 8 : 0);
+
+ ctx->num_cs_dw_streamout_end =
+ 12 + /* flush_vgt_streamout */
+ util_bitcount(buffer_en) * 8 +
+ 3;
+
+ r600_need_cs_space(ctx,
+ 12 + /* flush_vgt_streamout */
+ 6 + /* enables */
+ util_bitcount(buffer_en & ctx->streamout_append_bitmask) * 8 +
+ util_bitcount(buffer_en & ~ctx->streamout_append_bitmask) * 6 +
+ ctx->num_cs_dw_streamout_end, TRUE);
+
+ if (ctx->chip_class >= CAYMAN) {
+ evergreen_flush_vgt_streamout(ctx);
+ evergreen_set_streamout_enable(ctx, buffer_en);
+ }
+
+ for (i = 0; i < ctx->num_so_targets; i++) {
+#if 0
+ if (t[i]) {
+ t[i]->stride = strides[i];
+ t[i]->so_index = i;
+
+ cs->buf[cs->cdw++] = PKT3(PKT3_SET_CONTEXT_REG, 3, 0);
+ cs->buf[cs->cdw++] = (R_028AD0_VGT_STRMOUT_BUFFER_SIZE_0 +
+ 16*i - SI_CONTEXT_REG_OFFSET) >> 2;
+ cs->buf[cs->cdw++] = (t[i]->b.buffer_offset +
+ t[i]->b.buffer_size) >> 2; /* BUFFER_SIZE (in DW) */
+ cs->buf[cs->cdw++] = strides[i] >> 2; /* VTX_STRIDE (in DW) */
+ cs->buf[cs->cdw++] = 0; /* BUFFER_BASE */
+
+ cs->buf[cs->cdw++] = PKT3(PKT3_NOP, 0, 0);
+ cs->buf[cs->cdw++] =
+ r600_context_bo_reloc(ctx, r600_resource(t[i]->b.buffer),
+ RADEON_USAGE_WRITE);
+
+ if (ctx->streamout_append_bitmask & (1 << i)) {
+ /* Append. */
+ cs->buf[cs->cdw++] = PKT3(PKT3_STRMOUT_BUFFER_UPDATE, 4, 0);
+ cs->buf[cs->cdw++] = STRMOUT_SELECT_BUFFER(i) |
+ STRMOUT_OFFSET_SOURCE(STRMOUT_OFFSET_FROM_MEM); /* control */
+ cs->buf[cs->cdw++] = 0; /* unused */
+ cs->buf[cs->cdw++] = 0; /* unused */
+ cs->buf[cs->cdw++] = 0; /* src address lo */
+ cs->buf[cs->cdw++] = 0; /* src address hi */
+
+ cs->buf[cs->cdw++] = PKT3(PKT3_NOP, 0, 0);
+ cs->buf[cs->cdw++] =
+ r600_context_bo_reloc(ctx, t[i]->filled_size,
+ RADEON_USAGE_READ);
+ } else {
+ /* Start from the beginning. */
+ cs->buf[cs->cdw++] = PKT3(PKT3_STRMOUT_BUFFER_UPDATE, 4, 0);
+ cs->buf[cs->cdw++] = STRMOUT_SELECT_BUFFER(i) |
+ STRMOUT_OFFSET_SOURCE(STRMOUT_OFFSET_FROM_PACKET); /* control */
+ cs->buf[cs->cdw++] = 0; /* unused */
+ cs->buf[cs->cdw++] = 0; /* unused */
+ cs->buf[cs->cdw++] = t[i]->b.buffer_offset >> 2; /* buffer offset in DW */
+ cs->buf[cs->cdw++] = 0; /* unused */
+ }
+ }
+#endif
+ }
+}
+
+void r600_context_streamout_end(struct r600_context *ctx)
+{
+ struct radeon_winsys_cs *cs = ctx->cs;
+ struct r600_so_target **t = ctx->so_targets;
+ unsigned i, flush_flags = 0;
+
+ evergreen_flush_vgt_streamout(ctx);
+
+ for (i = 0; i < ctx->num_so_targets; i++) {
+#if 0
+ if (t[i]) {
+ cs->buf[cs->cdw++] = PKT3(PKT3_STRMOUT_BUFFER_UPDATE, 4, 0);
+ cs->buf[cs->cdw++] = STRMOUT_SELECT_BUFFER(i) |
+ STRMOUT_OFFSET_SOURCE(STRMOUT_OFFSET_NONE) |
+ STRMOUT_STORE_BUFFER_FILLED_SIZE; /* control */
+ cs->buf[cs->cdw++] = 0; /* dst address lo */
+ cs->buf[cs->cdw++] = 0; /* dst address hi */
+ cs->buf[cs->cdw++] = 0; /* unused */
+ cs->buf[cs->cdw++] = 0; /* unused */
+
+ cs->buf[cs->cdw++] = PKT3(PKT3_NOP, 0, 0);
+ cs->buf[cs->cdw++] =
+ r600_context_bo_reloc(ctx, t[i]->filled_size,
+ RADEON_USAGE_WRITE);
+
+ flush_flags |= S_0085F0_SO0_DEST_BASE_ENA(1) << i;
+ }
+#endif
+ }
+
+ evergreen_set_streamout_enable(ctx, 0);
+
+ ctx->atom_surface_sync.flush_flags |= flush_flags;
+ r600_atom_dirty(ctx, &ctx->atom_surface_sync.atom);
+
+ ctx->num_cs_dw_streamout_end = 0;
+
+ /* XXX print some debug info */
+ for (i = 0; i < ctx->num_so_targets; i++) {
+ if (!t[i])
+ continue;
+
+ uint32_t *ptr = ctx->ws->buffer_map(t[i]->filled_size->buf, ctx->cs, RADEON_USAGE_READ);
+ printf("FILLED_SIZE%i: %u\n", i, *ptr);
+ ctx->ws->buffer_unmap(t[i]->filled_size->buf);
+ }
+}
+
+void r600_context_draw_opaque_count(struct r600_context *ctx, struct r600_so_target *t)
+{
+ struct radeon_winsys_cs *cs = ctx->cs;
+ r600_need_cs_space(ctx, 14 + 21, TRUE);
+
+ cs->buf[cs->cdw++] = PKT3(PKT3_SET_CONTEXT_REG, 1, 0);
+ cs->buf[cs->cdw++] = (R_028B28_VGT_STRMOUT_DRAW_OPAQUE_OFFSET - SI_CONTEXT_REG_OFFSET) >> 2;
+ cs->buf[cs->cdw++] = 0;
+
+ cs->buf[cs->cdw++] = PKT3(PKT3_SET_CONTEXT_REG, 1, 0);
+ cs->buf[cs->cdw++] = (R_028B30_VGT_STRMOUT_DRAW_OPAQUE_VERTEX_STRIDE - SI_CONTEXT_REG_OFFSET) >> 2;
+ cs->buf[cs->cdw++] = t->stride >> 2;
+
+#if 0
+ cs->buf[cs->cdw++] = PKT3(PKT3_COPY_DW, 4, 0);
+ cs->buf[cs->cdw++] = COPY_DW_SRC_IS_MEM | COPY_DW_DST_IS_REG;
+ cs->buf[cs->cdw++] = 0; /* src address lo */
+ cs->buf[cs->cdw++] = 0; /* src address hi */
+ cs->buf[cs->cdw++] = R_028B2C_VGT_STRMOUT_DRAW_OPAQUE_BUFFER_FILLED_SIZE >> 2; /* dst register */
+ cs->buf[cs->cdw++] = 0; /* unused */
+#endif
+
+ cs->buf[cs->cdw++] = PKT3(PKT3_NOP, 0, 0);
+ cs->buf[cs->cdw++] = r600_context_bo_reloc(ctx, t->filled_size, RADEON_USAGE_READ);
+
+#if 0 /* I have not found this useful yet. */
+ cs->buf[cs->cdw++] = PKT3(PKT3_COPY_DW, 4, 0);
+ cs->buf[cs->cdw++] = COPY_DW_SRC_IS_REG | COPY_DW_DST_IS_REG;
+ cs->buf[cs->cdw++] = R_028B2C_VGT_STRMOUT_DRAW_OPAQUE_BUFFER_FILLED_SIZE >> 2; /* src register */
+ cs->buf[cs->cdw++] = 0; /* unused */
+ cs->buf[cs->cdw++] = R_0085F4_CP_COHER_SIZE >> 2; /* dst register */
+ cs->buf[cs->cdw++] = 0; /* unused */
+
+ cs->buf[cs->cdw++] = PKT3(PKT3_SET_CONFIG_REG, 1, 0);
+ cs->buf[cs->cdw++] = (R_0085F0_CP_COHER_CNTL - SI_CONFIG_REG_OFFSET) >> 2;
+ cs->buf[cs->cdw++] = S_0085F0_SO0_DEST_BASE_ENA(1) << t->so_index;
+
+ cs->buf[cs->cdw++] = PKT3(PKT3_SET_CONFIG_REG, 1, 0);
+ cs->buf[cs->cdw++] = (R_0085F8_CP_COHER_BASE - SI_CONFIG_REG_OFFSET) >> 2;
+ cs->buf[cs->cdw++] = t->b.buffer_offset >> 2;
+
+ cs->buf[cs->cdw++] = PKT3(PKT3_NOP, 0, 0);
+ cs->buf[cs->cdw++] = r600_context_bo_reloc(ctx, (struct r600_resource*)t->b.buffer,
+ RADEON_USAGE_WRITE);
+
+ cs->buf[cs->cdw++] = PKT3(PKT3_WAIT_REG_MEM, 5, 0);
+ cs->buf[cs->cdw++] = WAIT_REG_MEM_EQUAL; /* wait until the register is equal to the reference value */
+ cs->buf[cs->cdw++] = R_0085FC_CP_COHER_STATUS >> 2; /* register */
+ cs->buf[cs->cdw++] = 0;
+ cs->buf[cs->cdw++] = 0; /* reference value */
+ cs->buf[cs->cdw++] = 0xffffffff; /* mask */
+ cs->buf[cs->cdw++] = 4; /* poll interval */
+#endif
+}
diff --git a/src/gallium/drivers/radeonsi/r600_hw_context_priv.h b/src/gallium/drivers/radeonsi/r600_hw_context_priv.h
new file mode 100644
index 00000000000..7d5394e9f58
--- /dev/null
+++ b/src/gallium/drivers/radeonsi/r600_hw_context_priv.h
@@ -0,0 +1,76 @@
+/*
+ * Copyright 2010 Jerome Glisse <[email protected]>
+ *
+ * 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.
+ *
+ * Authors:
+ * Jerome Glisse
+ */
+#ifndef R600_PRIV_H
+#define R600_PRIV_H
+
+#include "radeonsi_pipe.h"
+#include "util/u_hash_table.h"
+#include "os/os_thread.h"
+
+#define R600_MAX_DRAW_CS_DWORDS 17
+
+#define PKT_COUNT_C 0xC000FFFF
+#define PKT_COUNT_S(x) (((x) & 0x3FFF) << 16)
+
+/* these flags are used in register flags and added into block flags */
+#define REG_FLAG_NEED_BO 1
+#define REG_FLAG_DIRTY_ALWAYS 2
+#define REG_FLAG_RV6XX_SBU 4
+#define REG_FLAG_NOT_R600 8
+#define REG_FLAG_ENABLE_ALWAYS 16
+#define REG_FLAG_FLUSH_CHANGE 64
+
+struct r600_reg {
+ unsigned offset;
+ unsigned flags;
+};
+
+/*
+ * r600_hw_context.c
+ */
+struct r600_resource *r600_context_reg_bo(struct r600_context *ctx, unsigned offset);
+int r600_context_add_block(struct r600_context *ctx, const struct r600_reg *reg, unsigned nreg,
+ unsigned opcode, unsigned offset_base);
+void r600_context_dirty_block(struct r600_context *ctx, struct r600_block *block,
+ int dirty, int index);
+int r600_setup_block_table(struct r600_context *ctx);
+void r600_init_cs(struct r600_context *ctx);
+
+/*
+ * evergreen_hw_context.c
+ */
+void evergreen_flush_vgt_streamout(struct r600_context *ctx);
+void evergreen_set_streamout_enable(struct r600_context *ctx, unsigned buffer_enable_bit);
+
+
+static INLINE unsigned r600_context_bo_reloc(struct r600_context *ctx, struct r600_resource *rbo,
+ enum radeon_bo_usage usage)
+{
+ assert(usage);
+ return ctx->ws->cs_add_reloc(ctx->cs, rbo->cs_buf, usage, rbo->domains) * 4;
+}
+
+#endif
diff --git a/src/gallium/drivers/radeonsi/r600_query.c b/src/gallium/drivers/radeonsi/r600_query.c
new file mode 100644
index 00000000000..bbf7c046f57
--- /dev/null
+++ b/src/gallium/drivers/radeonsi/r600_query.c
@@ -0,0 +1,130 @@
+/*
+ * Copyright 2010 Jerome Glisse <[email protected]>
+ *
+ * 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 "radeonsi_pipe.h"
+#include "sid.h"
+
+static struct pipe_query *r600_create_query(struct pipe_context *ctx, unsigned query_type)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+
+ return (struct pipe_query*)r600_context_query_create(rctx, query_type);
+}
+
+static void r600_destroy_query(struct pipe_context *ctx, struct pipe_query *query)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+
+ r600_context_query_destroy(rctx, (struct r600_query *)query);
+}
+
+static void r600_begin_query(struct pipe_context *ctx, struct pipe_query *query)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct r600_query *rquery = (struct r600_query *)query;
+
+ memset(&rquery->result, 0, sizeof(rquery->result));
+ rquery->results_start = rquery->results_end;
+ r600_query_begin(rctx, (struct r600_query *)query);
+ LIST_ADDTAIL(&rquery->list, &rctx->active_query_list);
+}
+
+static void r600_end_query(struct pipe_context *ctx, struct pipe_query *query)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct r600_query *rquery = (struct r600_query *)query;
+
+ r600_query_end(rctx, rquery);
+ LIST_DELINIT(&rquery->list);
+}
+
+static boolean r600_get_query_result(struct pipe_context *ctx,
+ struct pipe_query *query,
+ boolean wait, union pipe_query_result *vresult)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct r600_query *rquery = (struct r600_query *)query;
+
+ return r600_context_query_result(rctx, rquery, wait, vresult);
+}
+
+static void r600_render_condition(struct pipe_context *ctx,
+ struct pipe_query *query,
+ uint mode)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct r600_query *rquery = (struct r600_query *)query;
+ int wait_flag = 0;
+
+ /* If we already have nonzero result, render unconditionally */
+ if (query != NULL && rquery->result.u64 != 0) {
+ if (rctx->current_render_cond) {
+ r600_render_condition(ctx, NULL, 0);
+ }
+ return;
+ }
+
+ rctx->current_render_cond = query;
+ rctx->current_render_cond_mode = mode;
+
+ if (query == NULL) {
+ if (rctx->predicate_drawing) {
+ rctx->predicate_drawing = false;
+ r600_query_predication(rctx, NULL, PREDICATION_OP_CLEAR, 1);
+ }
+ return;
+ }
+
+ if (mode == PIPE_RENDER_COND_WAIT ||
+ mode == PIPE_RENDER_COND_BY_REGION_WAIT) {
+ wait_flag = 1;
+ }
+
+ rctx->predicate_drawing = true;
+
+ switch (rquery->type) {
+ case PIPE_QUERY_OCCLUSION_COUNTER:
+ case PIPE_QUERY_OCCLUSION_PREDICATE:
+ r600_query_predication(rctx, rquery, PREDICATION_OP_ZPASS, wait_flag);
+ break;
+ case PIPE_QUERY_PRIMITIVES_EMITTED:
+ case PIPE_QUERY_PRIMITIVES_GENERATED:
+ case PIPE_QUERY_SO_STATISTICS:
+ case PIPE_QUERY_SO_OVERFLOW_PREDICATE:
+ r600_query_predication(rctx, rquery, PREDICATION_OP_PRIMCOUNT, wait_flag);
+ break;
+ default:
+ assert(0);
+ }
+}
+
+void r600_init_query_functions(struct r600_context *rctx)
+{
+ rctx->context.create_query = r600_create_query;
+ rctx->context.destroy_query = r600_destroy_query;
+ rctx->context.begin_query = r600_begin_query;
+ rctx->context.end_query = r600_end_query;
+ rctx->context.get_query_result = r600_get_query_result;
+
+ if (rctx->screen->info.r600_num_backends > 0)
+ rctx->context.render_condition = r600_render_condition;
+}
diff --git a/src/gallium/drivers/radeonsi/r600_resource.c b/src/gallium/drivers/radeonsi/r600_resource.c
new file mode 100644
index 00000000000..7bdf6d6bd5f
--- /dev/null
+++ b/src/gallium/drivers/radeonsi/r600_resource.c
@@ -0,0 +1,64 @@
+/*
+ * Copyright 2010 Marek Olšák <[email protected]
+ *
+ * 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 "radeonsi_pipe.h"
+
+static struct pipe_resource *r600_resource_create(struct pipe_screen *screen,
+ const struct pipe_resource *templ)
+{
+ if (templ->target == PIPE_BUFFER) {
+ return r600_buffer_create(screen, templ);
+ } else {
+ return r600_texture_create(screen, templ);
+ }
+}
+
+static struct pipe_resource *r600_resource_from_handle(struct pipe_screen * screen,
+ const struct pipe_resource *templ,
+ struct winsys_handle *whandle)
+{
+ if (templ->target == PIPE_BUFFER) {
+ return NULL;
+ } else {
+ return r600_texture_from_handle(screen, templ, whandle);
+ }
+}
+
+void r600_init_screen_resource_functions(struct pipe_screen *screen)
+{
+ screen->resource_create = r600_resource_create;
+ screen->resource_from_handle = r600_resource_from_handle;
+ screen->resource_get_handle = u_resource_get_handle_vtbl;
+ screen->resource_destroy = u_resource_destroy_vtbl;
+ screen->user_buffer_create = r600_user_buffer_create;
+}
+
+void r600_init_context_resource_functions(struct r600_context *r600)
+{
+ r600->context.get_transfer = u_get_transfer_vtbl;
+ r600->context.transfer_map = u_transfer_map_vtbl;
+ r600->context.transfer_flush_region = u_transfer_flush_region_vtbl;
+ r600->context.transfer_unmap = u_transfer_unmap_vtbl;
+ r600->context.transfer_destroy = u_transfer_destroy_vtbl;
+ r600->context.transfer_inline_write = u_transfer_inline_write_vtbl;
+}
diff --git a/src/gallium/drivers/radeonsi/r600_resource.h b/src/gallium/drivers/radeonsi/r600_resource.h
new file mode 100644
index 00000000000..d6f97b0d5a5
--- /dev/null
+++ b/src/gallium/drivers/radeonsi/r600_resource.h
@@ -0,0 +1,105 @@
+/*
+ * Copyright 2010 Marek Olšák <[email protected]
+ *
+ * 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.
+ */
+#ifndef R600_RESOURCE_H
+#define R600_RESOURCE_H
+
+#include "util/u_transfer.h"
+#include "util/u_vbuf.h"
+
+/* flag to indicate a resource is to be used as a transfer so should not be tiled */
+#define R600_RESOURCE_FLAG_TRANSFER PIPE_RESOURCE_FLAG_DRV_PRIV
+
+/* Texture transfer. */
+struct r600_transfer {
+ /* Base class. */
+ struct pipe_transfer transfer;
+ /* Buffer transfer. */
+ struct pipe_transfer *buffer_transfer;
+ unsigned offset;
+ struct pipe_resource *staging_texture;
+};
+
+struct r600_resource_texture {
+ struct r600_resource resource;
+
+ /* If this resource is a depth-stencil buffer on evergreen, this contains
+ * the depth part of the format. There is a separate stencil resource
+ * for the stencil buffer below. */
+ enum pipe_format real_format;
+
+ unsigned offset[PIPE_MAX_TEXTURE_LEVELS];
+ unsigned pitch_in_bytes[PIPE_MAX_TEXTURE_LEVELS]; /* transfer */
+ unsigned pitch_in_blocks[PIPE_MAX_TEXTURE_LEVELS]; /* texture resource */
+ unsigned layer_size[PIPE_MAX_TEXTURE_LEVELS];
+ unsigned array_mode[PIPE_MAX_TEXTURE_LEVELS];
+ unsigned pitch_override;
+ unsigned size;
+ unsigned tile_type;
+ unsigned depth;
+ unsigned dirty_db;
+ struct r600_resource_texture *stencil; /* Stencil is in a separate buffer on Evergreen. */
+ struct r600_resource_texture *flushed_depth_texture;
+ boolean is_flushing_texture;
+};
+
+#define R600_TEX_IS_TILED(tex, level) ((tex)->array_mode[level] != V_038000_ARRAY_LINEAR_GENERAL && (tex)->array_mode[level] != V_038000_ARRAY_LINEAR_ALIGNED)
+
+struct r600_surface {
+ struct pipe_surface base;
+ unsigned aligned_height;
+};
+
+void r600_init_screen_resource_functions(struct pipe_screen *screen);
+
+/* r600_texture */
+struct pipe_resource *r600_texture_create(struct pipe_screen *screen,
+ const struct pipe_resource *templ);
+struct pipe_resource *r600_texture_from_handle(struct pipe_screen *screen,
+ const struct pipe_resource *base,
+ struct winsys_handle *whandle);
+
+static INLINE struct r600_resource *r600_resource(struct pipe_resource *r)
+{
+ return (struct r600_resource*)r;
+}
+
+int r600_texture_depth_flush(struct pipe_context *ctx, struct pipe_resource *texture, boolean just_create);
+
+/* r600_texture.c texture transfer functions. */
+struct pipe_transfer* r600_texture_get_transfer(struct pipe_context *ctx,
+ struct pipe_resource *texture,
+ unsigned level,
+ unsigned usage,
+ const struct pipe_box *box);
+void r600_texture_transfer_destroy(struct pipe_context *ctx,
+ struct pipe_transfer *trans);
+void* r600_texture_transfer_map(struct pipe_context *ctx,
+ struct pipe_transfer* transfer);
+void r600_texture_transfer_unmap(struct pipe_context *ctx,
+ struct pipe_transfer* transfer);
+
+struct r600_context;
+
+void r600_upload_const_buffer(struct r600_context *rctx, struct r600_resource **rbuffer, uint32_t *offset);
+
+#endif
diff --git a/src/gallium/drivers/radeonsi/r600_state_common.c b/src/gallium/drivers/radeonsi/r600_state_common.c
new file mode 100644
index 00000000000..4ba83dec903
--- /dev/null
+++ b/src/gallium/drivers/radeonsi/r600_state_common.c
@@ -0,0 +1,899 @@
+/*
+ * Copyright 2010 Red Hat Inc.
+ * 2010 Jerome Glisse
+ *
+ * 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.
+ *
+ * Authors: Dave Airlie <[email protected]>
+ * Jerome Glisse <[email protected]>
+ */
+#include "util/u_blitter.h"
+#include "util/u_memory.h"
+#include "util/u_format.h"
+#include "pipebuffer/pb_buffer.h"
+#include "pipe/p_shader_tokens.h"
+#include "tgsi/tgsi_parse.h"
+#include "r600_hw_context_priv.h"
+#include "radeonsi_pipe.h"
+#include "sid.h"
+
+static void r600_emit_surface_sync(struct r600_context *rctx, struct r600_atom *atom)
+{
+ struct radeon_winsys_cs *cs = rctx->cs;
+ struct r600_atom_surface_sync *a = (struct r600_atom_surface_sync*)atom;
+
+ cs->buf[cs->cdw++] = PKT3(PKT3_SURFACE_SYNC, 3, 0);
+ cs->buf[cs->cdw++] = a->flush_flags; /* CP_COHER_CNTL */
+ cs->buf[cs->cdw++] = 0xffffffff; /* CP_COHER_SIZE */
+ cs->buf[cs->cdw++] = 0; /* CP_COHER_BASE */
+ cs->buf[cs->cdw++] = 0x0000000A; /* POLL_INTERVAL */
+
+ a->flush_flags = 0;
+}
+
+static void r600_emit_r6xx_flush_and_inv(struct r600_context *rctx, struct r600_atom *atom)
+{
+ struct radeon_winsys_cs *cs = rctx->cs;
+ cs->buf[cs->cdw++] = PKT3(PKT3_EVENT_WRITE, 0, 0);
+ cs->buf[cs->cdw++] = EVENT_TYPE(EVENT_TYPE_CACHE_FLUSH_AND_INV_EVENT) | EVENT_INDEX(0);
+}
+
+static void r600_init_atom(struct r600_atom *atom,
+ void (*emit)(struct r600_context *ctx, struct r600_atom *state),
+ unsigned num_dw,
+ enum r600_atom_flags flags)
+{
+ atom->emit = emit;
+ atom->num_dw = num_dw;
+ atom->flags = flags;
+}
+
+void r600_init_common_atoms(struct r600_context *rctx)
+{
+ r600_init_atom(&rctx->atom_surface_sync.atom, r600_emit_surface_sync, 5, EMIT_EARLY);
+ r600_init_atom(&rctx->atom_r6xx_flush_and_inv, r600_emit_r6xx_flush_and_inv, 2, EMIT_EARLY);
+}
+
+unsigned r600_get_cb_flush_flags(struct r600_context *rctx)
+{
+ unsigned flags = 0;
+
+ if (rctx->framebuffer.nr_cbufs) {
+ flags |= S_0085F0_CB_ACTION_ENA(1) |
+ (((1 << rctx->framebuffer.nr_cbufs) - 1) << S_0085F0_CB0_DEST_BASE_ENA_SHIFT);
+ }
+
+ return flags;
+}
+
+void r600_texture_barrier(struct pipe_context *ctx)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+
+ rctx->atom_surface_sync.flush_flags |= S_0085F0_TC_ACTION_ENA(1) | r600_get_cb_flush_flags(rctx);
+ r600_atom_dirty(rctx, &rctx->atom_surface_sync.atom);
+}
+
+static bool r600_conv_pipe_prim(unsigned pprim, unsigned *prim)
+{
+ static const int prim_conv[] = {
+ V_008958_DI_PT_POINTLIST,
+ V_008958_DI_PT_LINELIST,
+ V_008958_DI_PT_LINELOOP,
+ V_008958_DI_PT_LINESTRIP,
+ V_008958_DI_PT_TRILIST,
+ V_008958_DI_PT_TRISTRIP,
+ V_008958_DI_PT_TRIFAN,
+ V_008958_DI_PT_QUADLIST,
+ V_008958_DI_PT_QUADSTRIP,
+ V_008958_DI_PT_POLYGON,
+ -1,
+ -1,
+ -1,
+ -1
+ };
+
+ *prim = prim_conv[pprim];
+ if (*prim == -1) {
+ fprintf(stderr, "%s:%d unsupported %d\n", __func__, __LINE__, pprim);
+ return false;
+ }
+ return true;
+}
+
+/* common state between evergreen and r600 */
+void r600_bind_blend_state(struct pipe_context *ctx, void *state)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct r600_pipe_blend *blend = (struct r600_pipe_blend *)state;
+ struct r600_pipe_state *rstate;
+
+ if (state == NULL)
+ return;
+ rstate = &blend->rstate;
+ rctx->states[rstate->id] = rstate;
+ rctx->cb_target_mask = blend->cb_target_mask;
+ rctx->cb_color_control = blend->cb_color_control;
+
+ r600_context_pipe_state_set(rctx, rstate);
+}
+
+static void r600_set_stencil_ref(struct pipe_context *ctx,
+ const struct r600_stencil_ref *state)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct r600_pipe_state *rstate = CALLOC_STRUCT(r600_pipe_state);
+
+ if (rstate == NULL)
+ return;
+
+ rstate->id = R600_PIPE_STATE_STENCIL_REF;
+ r600_pipe_state_add_reg(rstate,
+ R_028430_DB_STENCILREFMASK,
+ S_028430_STENCILTESTVAL(state->ref_value[0]) |
+ S_028430_STENCILMASK(state->valuemask[0]) |
+ S_028430_STENCILWRITEMASK(state->writemask[0]),
+ NULL, 0);
+ r600_pipe_state_add_reg(rstate,
+ R_028434_DB_STENCILREFMASK_BF,
+ S_028434_STENCILTESTVAL_BF(state->ref_value[1]) |
+ S_028434_STENCILMASK_BF(state->valuemask[1]) |
+ S_028434_STENCILWRITEMASK_BF(state->writemask[1]),
+ NULL, 0);
+
+ free(rctx->states[R600_PIPE_STATE_STENCIL_REF]);
+ rctx->states[R600_PIPE_STATE_STENCIL_REF] = rstate;
+ r600_context_pipe_state_set(rctx, rstate);
+}
+
+void r600_set_pipe_stencil_ref(struct pipe_context *ctx,
+ const struct pipe_stencil_ref *state)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct r600_pipe_dsa *dsa = (struct r600_pipe_dsa*)rctx->states[R600_PIPE_STATE_DSA];
+ struct r600_stencil_ref ref;
+
+ rctx->stencil_ref = *state;
+
+ if (!dsa)
+ return;
+
+ ref.ref_value[0] = state->ref_value[0];
+ ref.ref_value[1] = state->ref_value[1];
+ ref.valuemask[0] = dsa->valuemask[0];
+ ref.valuemask[1] = dsa->valuemask[1];
+ ref.writemask[0] = dsa->writemask[0];
+ ref.writemask[1] = dsa->writemask[1];
+
+ r600_set_stencil_ref(ctx, &ref);
+}
+
+void r600_bind_dsa_state(struct pipe_context *ctx, void *state)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct r600_pipe_dsa *dsa = state;
+ struct r600_pipe_state *rstate;
+ struct r600_stencil_ref ref;
+
+ if (state == NULL)
+ return;
+ rstate = &dsa->rstate;
+ rctx->states[rstate->id] = rstate;
+ rctx->alpha_ref = dsa->alpha_ref;
+ rctx->alpha_ref_dirty = true;
+ r600_context_pipe_state_set(rctx, rstate);
+
+ ref.ref_value[0] = rctx->stencil_ref.ref_value[0];
+ ref.ref_value[1] = rctx->stencil_ref.ref_value[1];
+ ref.valuemask[0] = dsa->valuemask[0];
+ ref.valuemask[1] = dsa->valuemask[1];
+ ref.writemask[0] = dsa->writemask[0];
+ ref.writemask[1] = dsa->writemask[1];
+
+ r600_set_stencil_ref(ctx, &ref);
+}
+
+void r600_bind_rs_state(struct pipe_context *ctx, void *state)
+{
+ struct r600_pipe_rasterizer *rs = (struct r600_pipe_rasterizer *)state;
+ struct r600_context *rctx = (struct r600_context *)ctx;
+
+ if (state == NULL)
+ return;
+
+ rctx->sprite_coord_enable = rs->sprite_coord_enable;
+ rctx->pa_sc_line_stipple = rs->pa_sc_line_stipple;
+ rctx->pa_su_sc_mode_cntl = rs->pa_su_sc_mode_cntl;
+ rctx->pa_cl_clip_cntl = rs->pa_cl_clip_cntl;
+ rctx->pa_cl_vs_out_cntl = rs->pa_cl_vs_out_cntl;
+
+ rctx->rasterizer = rs;
+
+ rctx->states[rs->rstate.id] = &rs->rstate;
+ r600_context_pipe_state_set(rctx, &rs->rstate);
+
+ if (rctx->chip_class >= CAYMAN) {
+ cayman_polygon_offset_update(rctx);
+ }
+}
+
+void r600_delete_rs_state(struct pipe_context *ctx, void *state)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct r600_pipe_rasterizer *rs = (struct r600_pipe_rasterizer *)state;
+
+ if (rctx->rasterizer == rs) {
+ rctx->rasterizer = NULL;
+ }
+ if (rctx->states[rs->rstate.id] == &rs->rstate) {
+ rctx->states[rs->rstate.id] = NULL;
+ }
+ free(rs);
+}
+
+void r600_sampler_view_destroy(struct pipe_context *ctx,
+ struct pipe_sampler_view *state)
+{
+ struct r600_pipe_sampler_view *resource = (struct r600_pipe_sampler_view *)state;
+
+ pipe_resource_reference(&state->texture, NULL);
+ FREE(resource);
+}
+
+void r600_delete_state(struct pipe_context *ctx, void *state)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct r600_pipe_state *rstate = (struct r600_pipe_state *)state;
+
+ if (rctx->states[rstate->id] == rstate) {
+ rctx->states[rstate->id] = NULL;
+ }
+ for (int i = 0; i < rstate->nregs; i++) {
+ pipe_resource_reference((struct pipe_resource**)&rstate->regs[i].bo, NULL);
+ }
+ free(rstate);
+}
+
+void r600_bind_vertex_elements(struct pipe_context *ctx, void *state)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct r600_vertex_element *v = (struct r600_vertex_element*)state;
+
+ rctx->vertex_elements = v;
+ if (v) {
+ r600_inval_shader_cache(rctx);
+ u_vbuf_bind_vertex_elements(rctx->vbuf_mgr, state,
+ v->vmgr_elements);
+
+ rctx->states[v->rstate.id] = &v->rstate;
+ r600_context_pipe_state_set(rctx, &v->rstate);
+ }
+}
+
+void r600_delete_vertex_element(struct pipe_context *ctx, void *state)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct r600_vertex_element *v = (struct r600_vertex_element*)state;
+
+ if (rctx->states[v->rstate.id] == &v->rstate) {
+ rctx->states[v->rstate.id] = NULL;
+ }
+ if (rctx->vertex_elements == state)
+ rctx->vertex_elements = NULL;
+
+ u_vbuf_destroy_vertex_elements(rctx->vbuf_mgr, v->vmgr_elements);
+ FREE(state);
+}
+
+
+void r600_set_index_buffer(struct pipe_context *ctx,
+ const struct pipe_index_buffer *ib)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+
+ u_vbuf_set_index_buffer(rctx->vbuf_mgr, ib);
+}
+
+void r600_set_vertex_buffers(struct pipe_context *ctx, unsigned count,
+ const struct pipe_vertex_buffer *buffers)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+
+ u_vbuf_set_vertex_buffers(rctx->vbuf_mgr, count, buffers);
+}
+
+void *si_create_vertex_elements(struct pipe_context *ctx,
+ unsigned count,
+ const struct pipe_vertex_element *elements)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct r600_vertex_element *v = CALLOC_STRUCT(r600_vertex_element);
+
+ assert(count < 32);
+ if (!v)
+ return NULL;
+
+ v->count = count;
+ v->vmgr_elements =
+ u_vbuf_create_vertex_elements(rctx->vbuf_mgr, count,
+ elements, v->elements);
+
+ return v;
+}
+
+void *si_create_shader_state(struct pipe_context *ctx,
+ const struct pipe_shader_state *state)
+{
+ struct si_pipe_shader *shader = CALLOC_STRUCT(si_pipe_shader);
+
+ shader->tokens = tgsi_dup_tokens(state->tokens);
+ shader->so = state->stream_output;
+
+ return shader;
+}
+
+void r600_bind_ps_shader(struct pipe_context *ctx, void *state)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+
+ if (rctx->ps_shader != state)
+ rctx->shader_dirty = true;
+
+ /* TODO delete old shader */
+ rctx->ps_shader = (struct si_pipe_shader *)state;
+ if (state) {
+ r600_inval_shader_cache(rctx);
+ r600_context_pipe_state_set(rctx, &rctx->ps_shader->rstate);
+ }
+}
+
+void r600_bind_vs_shader(struct pipe_context *ctx, void *state)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+
+ if (rctx->vs_shader != state)
+ rctx->shader_dirty = true;
+
+ /* TODO delete old shader */
+ rctx->vs_shader = (struct si_pipe_shader *)state;
+ if (state) {
+ r600_inval_shader_cache(rctx);
+ r600_context_pipe_state_set(rctx, &rctx->vs_shader->rstate);
+ }
+}
+
+void r600_delete_ps_shader(struct pipe_context *ctx, void *state)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct si_pipe_shader *shader = (struct si_pipe_shader *)state;
+
+ if (rctx->ps_shader == shader) {
+ rctx->ps_shader = NULL;
+ }
+
+ free(shader->tokens);
+ si_pipe_shader_destroy(ctx, shader);
+ free(shader);
+}
+
+void r600_delete_vs_shader(struct pipe_context *ctx, void *state)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct si_pipe_shader *shader = (struct si_pipe_shader *)state;
+
+ if (rctx->vs_shader == shader) {
+ rctx->vs_shader = NULL;
+ }
+
+ free(shader->tokens);
+ si_pipe_shader_destroy(ctx, shader);
+ free(shader);
+}
+
+static void r600_update_alpha_ref(struct r600_context *rctx)
+{
+#if 0
+ unsigned alpha_ref;
+ struct r600_pipe_state rstate;
+
+ alpha_ref = rctx->alpha_ref;
+ rstate.nregs = 0;
+ if (rctx->export_16bpc)
+ alpha_ref &= ~0x1FFF;
+ r600_pipe_state_add_reg(&rstate, R_028438_SX_ALPHA_REF, alpha_ref, NULL, 0);
+
+ r600_context_pipe_state_set(rctx, &rstate);
+ rctx->alpha_ref_dirty = false;
+#endif
+}
+
+void r600_set_constant_buffer(struct pipe_context *ctx, uint shader, uint index,
+ struct pipe_resource *buffer)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct r600_resource *rbuffer = r600_resource(buffer);
+ struct r600_pipe_state *rstate;
+ uint64_t va_offset;
+ uint32_t offset;
+
+ /* Note that the state tracker can unbind constant buffers by
+ * passing NULL here.
+ */
+ if (buffer == NULL) {
+ return;
+ }
+
+ r600_inval_shader_cache(rctx);
+
+ r600_upload_const_buffer(rctx, &rbuffer, &offset);
+ va_offset = r600_resource_va(ctx->screen, (void*)rbuffer);
+ va_offset += offset;
+ //va_offset >>= 8;
+
+ switch (shader) {
+ case PIPE_SHADER_VERTEX:
+ rstate = &rctx->vs_const_buffer;
+ rstate->nregs = 0;
+ r600_pipe_state_add_reg(rstate,
+ R_00B138_SPI_SHADER_USER_DATA_VS_2,
+ va_offset, rbuffer, RADEON_USAGE_READ);
+ r600_pipe_state_add_reg(rstate,
+ R_00B13C_SPI_SHADER_USER_DATA_VS_3,
+ va_offset >> 32, NULL, 0);
+ break;
+ case PIPE_SHADER_FRAGMENT:
+ rstate = &rctx->ps_const_buffer;
+ rstate->nregs = 0;
+ r600_pipe_state_add_reg(rstate,
+ R_00B030_SPI_SHADER_USER_DATA_PS_0,
+ va_offset, rbuffer, RADEON_USAGE_READ);
+ r600_pipe_state_add_reg(rstate,
+ R_00B034_SPI_SHADER_USER_DATA_PS_1,
+ va_offset >> 32, NULL, 0);
+ break;
+ default:
+ R600_ERR("unsupported %d\n", shader);
+ return;
+ }
+
+ r600_context_pipe_state_set(rctx, rstate);
+
+ if (buffer != &rbuffer->b.b.b)
+ pipe_resource_reference((struct pipe_resource**)&rbuffer, NULL);
+}
+
+struct pipe_stream_output_target *
+r600_create_so_target(struct pipe_context *ctx,
+ struct pipe_resource *buffer,
+ unsigned buffer_offset,
+ unsigned buffer_size)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct r600_so_target *t;
+ void *ptr;
+
+ t = CALLOC_STRUCT(r600_so_target);
+ if (!t) {
+ return NULL;
+ }
+
+ t->b.reference.count = 1;
+ t->b.context = ctx;
+ pipe_resource_reference(&t->b.buffer, buffer);
+ t->b.buffer_offset = buffer_offset;
+ t->b.buffer_size = buffer_size;
+
+ t->filled_size = (struct r600_resource*)
+ pipe_buffer_create(ctx->screen, PIPE_BIND_CUSTOM, PIPE_USAGE_STATIC, 4);
+ ptr = rctx->ws->buffer_map(t->filled_size->buf, rctx->cs, PIPE_TRANSFER_WRITE);
+ memset(ptr, 0, t->filled_size->buf->size);
+ rctx->ws->buffer_unmap(t->filled_size->buf);
+
+ return &t->b;
+}
+
+void r600_so_target_destroy(struct pipe_context *ctx,
+ struct pipe_stream_output_target *target)
+{
+ struct r600_so_target *t = (struct r600_so_target*)target;
+ pipe_resource_reference(&t->b.buffer, NULL);
+ pipe_resource_reference((struct pipe_resource**)&t->filled_size, NULL);
+ FREE(t);
+}
+
+void r600_set_so_targets(struct pipe_context *ctx,
+ unsigned num_targets,
+ struct pipe_stream_output_target **targets,
+ unsigned append_bitmask)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ unsigned i;
+
+ /* Stop streamout. */
+ if (rctx->num_so_targets) {
+ r600_context_streamout_end(rctx);
+ }
+
+ /* Set the new targets. */
+ for (i = 0; i < num_targets; i++) {
+ pipe_so_target_reference((struct pipe_stream_output_target**)&rctx->so_targets[i], targets[i]);
+ }
+ for (; i < rctx->num_so_targets; i++) {
+ pipe_so_target_reference((struct pipe_stream_output_target**)&rctx->so_targets[i], NULL);
+ }
+
+ rctx->num_so_targets = num_targets;
+ rctx->streamout_start = num_targets != 0;
+ rctx->streamout_append_bitmask = append_bitmask;
+}
+
+static void r600_vertex_buffer_update(struct r600_context *rctx)
+{
+ struct pipe_context *ctx = &rctx->context;
+ struct r600_pipe_state *rstate = &rctx->vs_user_data;
+ struct r600_resource *rbuffer, *t_list_buffer;
+ struct pipe_vertex_buffer *vertex_buffer;
+ unsigned i, count, offset;
+ uint32_t *ptr;
+ uint64_t va;
+
+ r600_inval_vertex_cache(rctx);
+
+ if (rctx->vertex_elements->vbuffer_need_offset) {
+ /* one resource per vertex elements */
+ count = rctx->vertex_elements->count;
+ } else {
+ /* bind vertex buffer once */
+ count = rctx->vbuf_mgr->nr_real_vertex_buffers;
+ }
+ assert(count <= 256 / 4);
+
+ t_list_buffer = (struct r600_resource*)
+ pipe_buffer_create(ctx->screen, PIPE_BIND_CUSTOM,
+ PIPE_USAGE_IMMUTABLE, 4 * 4 * count);
+ if (t_list_buffer == NULL)
+ return;
+
+ ptr = (uint32_t*)rctx->ws->buffer_map(t_list_buffer->buf,
+ rctx->cs,
+ PIPE_TRANSFER_WRITE);
+
+ for (i = 0 ; i < count; i++, ptr += 4) {
+ struct pipe_vertex_element *velem = &rctx->vertex_elements->elements[i];
+ const struct util_format_description *desc;
+ unsigned data_format, num_format;
+ int first_non_void;
+
+ if (rctx->vertex_elements->vbuffer_need_offset) {
+ /* one resource per vertex elements */
+ unsigned vbuffer_index;
+ vbuffer_index = rctx->vertex_elements->elements[i].vertex_buffer_index;
+ vertex_buffer = &rctx->vbuf_mgr->real_vertex_buffer[vbuffer_index];
+ rbuffer = (struct r600_resource*)vertex_buffer->buffer;
+ offset = rctx->vertex_elements->vbuffer_offset[i];
+ } else {
+ /* bind vertex buffer once */
+ vertex_buffer = &rctx->vbuf_mgr->real_vertex_buffer[i];
+ rbuffer = (struct r600_resource*)vertex_buffer->buffer;
+ offset = 0;
+ }
+ if (vertex_buffer == NULL || rbuffer == NULL)
+ continue;
+ offset += vertex_buffer->buffer_offset;
+
+ va = r600_resource_va(ctx->screen, (void*)rbuffer);
+ va += offset;
+
+ desc = util_format_description(velem->src_format);
+ first_non_void = util_format_get_first_non_void_channel(velem->src_format);
+ data_format = si_translate_vertexformat(ctx->screen,
+ velem->src_format,
+ desc, first_non_void);
+
+ switch (desc->channel[first_non_void].type) {
+ case UTIL_FORMAT_TYPE_FIXED:
+ num_format = V_008F0C_BUF_NUM_FORMAT_USCALED; /* XXX */
+ break;
+ case UTIL_FORMAT_TYPE_SIGNED:
+ num_format = V_008F0C_BUF_NUM_FORMAT_SNORM;
+ break;
+ case UTIL_FORMAT_TYPE_UNSIGNED:
+ num_format = V_008F0C_BUF_NUM_FORMAT_UNORM;
+ break;
+ case UTIL_FORMAT_TYPE_FLOAT:
+ default:
+ num_format = V_008F14_IMG_NUM_FORMAT_FLOAT;
+ }
+
+ /* Fill in T# buffer resource description */
+ ptr[0] = va & 0xFFFFFFFF;
+ ptr[1] = ((va >> 32) & 0xFFFF) |
+ (vertex_buffer->stride & 0x3FFF) << 16;
+ ptr[2] = (vertex_buffer->buffer->width0 - offset) / vertex_buffer->stride;
+ /* XXX: Hardcoding RGBA */
+ ptr[3] = 4 | 5 << 3 | 6 << 6 | 7 << 9 |
+ num_format << 12 | data_format << 15;
+
+ r600_context_bo_reloc(rctx, rbuffer, RADEON_USAGE_READ);
+ }
+
+ rstate->nregs = 0;
+
+ va = r600_resource_va(ctx->screen, (void*)t_list_buffer);
+ r600_pipe_state_add_reg(rstate,
+ R_00B130_SPI_SHADER_USER_DATA_VS_0,
+ va, t_list_buffer, RADEON_USAGE_READ);
+ r600_pipe_state_add_reg(rstate,
+ R_00B134_SPI_SHADER_USER_DATA_VS_1,
+ va >> 32,
+ NULL, 0);
+
+ r600_context_pipe_state_set(rctx, rstate);
+}
+
+static void si_update_derived_state(struct r600_context *rctx)
+{
+ struct pipe_context * ctx = (struct pipe_context*)rctx;
+
+ if (!rctx->blitter->running) {
+ if (rctx->have_depth_fb || rctx->have_depth_texture)
+ r600_flush_depth_textures(rctx);
+ }
+
+ if (rctx->shader_dirty) {
+ si_pipe_shader_destroy(&rctx->context, rctx->vs_shader);
+ }
+
+ if (rctx->shader_dirty ||
+ (rctx->ps_shader->shader.fs_write_all &&
+ (rctx->ps_shader->shader.nr_cbufs != rctx->nr_cbufs)) ||
+ (rctx->sprite_coord_enable &&
+ (rctx->ps_shader->sprite_coord_enable != rctx->sprite_coord_enable))) {
+ si_pipe_shader_destroy(&rctx->context, rctx->ps_shader);
+ }
+
+ if (rctx->alpha_ref_dirty) {
+ r600_update_alpha_ref(rctx);
+ }
+
+ if (!rctx->vs_shader->bo) {
+ si_pipe_shader_vs(ctx, rctx->vs_shader);
+
+ r600_context_pipe_state_set(rctx, &rctx->vs_shader->rstate);
+ }
+
+ if (!rctx->ps_shader->bo) {
+ si_pipe_shader_ps(ctx, rctx->ps_shader);
+
+ r600_context_pipe_state_set(rctx, &rctx->ps_shader->rstate);
+ }
+
+ if (rctx->shader_dirty) {
+ si_update_spi_map(rctx);
+ rctx->shader_dirty = false;
+ }
+}
+
+void r600_draw_vbo(struct pipe_context *ctx, const struct pipe_draw_info *dinfo)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct r600_pipe_dsa *dsa = (struct r600_pipe_dsa*)rctx->states[R600_PIPE_STATE_DSA];
+ struct pipe_draw_info info = *dinfo;
+ struct r600_draw rdraw = {};
+ struct pipe_index_buffer ib = {};
+ unsigned prim, mask, ls_mask = 0;
+ struct r600_block *dirty_block = NULL, *next_block = NULL;
+ struct r600_atom *state = NULL, *next_state = NULL;
+ int i;
+
+ if ((!info.count && (info.indexed || !info.count_from_stream_output)) ||
+ (info.indexed && !rctx->vbuf_mgr->index_buffer.buffer) ||
+ !r600_conv_pipe_prim(info.mode, &prim)) {
+ return;
+ }
+
+ if (!rctx->ps_shader || !rctx->vs_shader)
+ return;
+
+ si_update_derived_state(rctx);
+
+ u_vbuf_draw_begin(rctx->vbuf_mgr, &info);
+ r600_vertex_buffer_update(rctx);
+
+ rdraw.vgt_num_indices = info.count;
+ rdraw.vgt_num_instances = info.instance_count;
+
+ if (info.indexed) {
+ /* Initialize the index buffer struct. */
+ pipe_resource_reference(&ib.buffer, rctx->vbuf_mgr->index_buffer.buffer);
+ ib.index_size = rctx->vbuf_mgr->index_buffer.index_size;
+ ib.offset = rctx->vbuf_mgr->index_buffer.offset + info.start * ib.index_size;
+
+ /* Translate or upload, if needed. */
+ r600_translate_index_buffer(rctx, &ib, info.count);
+
+ if (u_vbuf_resource(ib.buffer)->user_ptr) {
+ r600_upload_index_buffer(rctx, &ib, info.count);
+ }
+
+ /* Initialize the r600_draw struct with index buffer info. */
+ if (ib.index_size == 4) {
+ rdraw.vgt_index_type = V_028A7C_VGT_INDEX_32 |
+ (R600_BIG_ENDIAN ? V_028A7C_VGT_DMA_SWAP_32_BIT : 0);
+ } else {
+ rdraw.vgt_index_type = V_028A7C_VGT_INDEX_16 |
+ (R600_BIG_ENDIAN ? V_028A7C_VGT_DMA_SWAP_16_BIT : 0);
+ }
+ rdraw.indices = (struct r600_resource*)ib.buffer;
+ rdraw.indices_bo_offset = ib.offset;
+ rdraw.vgt_draw_initiator = V_0287F0_DI_SRC_SEL_DMA;
+ } else {
+ info.index_bias = info.start;
+ rdraw.vgt_draw_initiator = V_0287F0_DI_SRC_SEL_AUTO_INDEX;
+ if (info.count_from_stream_output) {
+ rdraw.vgt_draw_initiator |= S_0287F0_USE_OPAQUE(1);
+
+ r600_context_draw_opaque_count(rctx, (struct r600_so_target*)info.count_from_stream_output);
+ }
+ }
+
+ rctx->vs_shader_so_strides = rctx->vs_shader->so_strides;
+
+ mask = (1ULL << ((unsigned)rctx->framebuffer.nr_cbufs * 4)) - 1;
+
+ if (rctx->vgt.id != R600_PIPE_STATE_VGT) {
+ rctx->vgt.id = R600_PIPE_STATE_VGT;
+ rctx->vgt.nregs = 0;
+ r600_pipe_state_add_reg(&rctx->vgt, R_008958_VGT_PRIMITIVE_TYPE, prim, NULL, 0);
+ r600_pipe_state_add_reg(&rctx->vgt, R_028238_CB_TARGET_MASK, rctx->cb_target_mask & mask, NULL, 0);
+ r600_pipe_state_add_reg(&rctx->vgt, R_028400_VGT_MAX_VTX_INDX, ~0, NULL, 0);
+ r600_pipe_state_add_reg(&rctx->vgt, R_028404_VGT_MIN_VTX_INDX, 0, NULL, 0);
+ r600_pipe_state_add_reg(&rctx->vgt, R_028408_VGT_INDX_OFFSET, info.index_bias, NULL, 0);
+ r600_pipe_state_add_reg(&rctx->vgt, R_02840C_VGT_MULTI_PRIM_IB_RESET_INDX, info.restart_index, NULL, 0);
+ r600_pipe_state_add_reg(&rctx->vgt, R_028A94_VGT_MULTI_PRIM_IB_RESET_EN, info.primitive_restart, NULL, 0);
+#if 0
+ r600_pipe_state_add_reg(&rctx->vgt, R_03CFF0_SQ_VTX_BASE_VTX_LOC, 0, NULL, 0);
+ r600_pipe_state_add_reg(&rctx->vgt, R_03CFF4_SQ_VTX_START_INST_LOC, info.start_instance, NULL, 0);
+#endif
+ r600_pipe_state_add_reg(&rctx->vgt, R_028A0C_PA_SC_LINE_STIPPLE, 0, NULL, 0);
+ r600_pipe_state_add_reg(&rctx->vgt, R_028814_PA_SU_SC_MODE_CNTL, 0, NULL, 0);
+ r600_pipe_state_add_reg(&rctx->vgt, R_02881C_PA_CL_VS_OUT_CNTL, 0, NULL, 0);
+ r600_pipe_state_add_reg(&rctx->vgt, R_028810_PA_CL_CLIP_CNTL, 0x0, NULL, 0);
+ }
+
+ rctx->vgt.nregs = 0;
+ r600_pipe_state_mod_reg(&rctx->vgt, prim);
+ r600_pipe_state_mod_reg(&rctx->vgt, rctx->cb_target_mask & mask);
+ r600_pipe_state_mod_reg(&rctx->vgt, ~0);
+ r600_pipe_state_mod_reg(&rctx->vgt, 0);
+ r600_pipe_state_mod_reg(&rctx->vgt, info.index_bias);
+ r600_pipe_state_mod_reg(&rctx->vgt, info.restart_index);
+ r600_pipe_state_mod_reg(&rctx->vgt, info.primitive_restart);
+#if 0
+ r600_pipe_state_mod_reg(&rctx->vgt, 0);
+ r600_pipe_state_mod_reg(&rctx->vgt, info.start_instance);
+#endif
+
+ if (prim == V_008958_DI_PT_LINELIST)
+ ls_mask = 1;
+ else if (prim == V_008958_DI_PT_LINESTRIP)
+ ls_mask = 2;
+ r600_pipe_state_mod_reg(&rctx->vgt, S_028A0C_AUTO_RESET_CNTL(ls_mask) | rctx->pa_sc_line_stipple);
+
+ if (info.mode == PIPE_PRIM_QUADS || info.mode == PIPE_PRIM_QUAD_STRIP || info.mode == PIPE_PRIM_POLYGON) {
+ r600_pipe_state_mod_reg(&rctx->vgt, S_028814_PROVOKING_VTX_LAST(1) | rctx->pa_su_sc_mode_cntl);
+ } else {
+ r600_pipe_state_mod_reg(&rctx->vgt, rctx->pa_su_sc_mode_cntl);
+ }
+ r600_pipe_state_mod_reg(&rctx->vgt,
+ rctx->pa_cl_vs_out_cntl /*|
+ (rctx->rasterizer->clip_plane_enable & rctx->vs_shader->shader.clip_dist_write)*/);
+ r600_pipe_state_mod_reg(&rctx->vgt,
+ rctx->pa_cl_clip_cntl /*|
+ (rctx->vs_shader->shader.clip_dist_write ||
+ rctx->vs_shader->shader.vs_prohibit_ucps ?
+ 0 : rctx->rasterizer->clip_plane_enable & 0x3F)*/);
+
+ r600_context_pipe_state_set(rctx, &rctx->vgt);
+
+ rdraw.db_render_override = dsa->db_render_override;
+ rdraw.db_render_control = dsa->db_render_control;
+
+ /* Emit states. */
+ r600_need_cs_space(rctx, 0, TRUE);
+
+ LIST_FOR_EACH_ENTRY_SAFE(state, next_state, &rctx->dirty_states, head) {
+ r600_emit_atom(rctx, state);
+ }
+ LIST_FOR_EACH_ENTRY_SAFE(dirty_block, next_block, &rctx->dirty,list) {
+ r600_context_block_emit_dirty(rctx, dirty_block);
+ }
+ rctx->pm4_dirty_cdwords = 0;
+
+ /* Enable stream out if needed. */
+ if (rctx->streamout_start) {
+ r600_context_streamout_begin(rctx);
+ rctx->streamout_start = FALSE;
+ }
+
+ for (i = 0; i < NUM_TEX_UNITS; i++) {
+ if (rctx->ps_samplers.views[i])
+ r600_context_bo_reloc(rctx,
+ (struct r600_resource*)rctx->ps_samplers.views[i]->base.texture,
+ RADEON_USAGE_READ);
+ }
+
+ if (rctx->chip_class >= CAYMAN) {
+ evergreen_context_draw(rctx, &rdraw);
+ }
+
+ rctx->flags |= R600_CONTEXT_DST_CACHES_DIRTY | R600_CONTEXT_DRAW_PENDING;
+
+ if (rctx->framebuffer.zsbuf)
+ {
+ struct pipe_resource *tex = rctx->framebuffer.zsbuf->texture;
+ ((struct r600_resource_texture *)tex)->dirty_db = TRUE;
+ }
+
+ pipe_resource_reference(&ib.buffer, NULL);
+ u_vbuf_draw_end(rctx->vbuf_mgr);
+}
+
+void _r600_pipe_state_add_reg(struct r600_context *ctx,
+ struct r600_pipe_state *state,
+ uint32_t offset, uint32_t value,
+ uint32_t range_id, uint32_t block_id,
+ struct r600_resource *bo,
+ enum radeon_bo_usage usage)
+{
+ struct r600_range *range;
+ struct r600_block *block;
+
+ if (bo) assert(usage);
+
+ range = &ctx->range[range_id];
+ block = range->blocks[block_id];
+ state->regs[state->nregs].block = block;
+ state->regs[state->nregs].id = (offset - block->start_offset) >> 2;
+
+ state->regs[state->nregs].value = value;
+ state->regs[state->nregs].bo = bo;
+ state->regs[state->nregs].bo_usage = usage;
+
+ state->nregs++;
+ assert(state->nregs < R600_BLOCK_MAX_REG);
+}
+
+void r600_pipe_state_add_reg_noblock(struct r600_pipe_state *state,
+ uint32_t offset, uint32_t value,
+ struct r600_resource *bo,
+ enum radeon_bo_usage usage)
+{
+ if (bo) assert(usage);
+
+ state->regs[state->nregs].id = offset;
+ state->regs[state->nregs].block = NULL;
+ state->regs[state->nregs].value = value;
+ state->regs[state->nregs].bo = bo;
+ state->regs[state->nregs].bo_usage = usage;
+
+ state->nregs++;
+ assert(state->nregs < R600_BLOCK_MAX_REG);
+}
diff --git a/src/gallium/drivers/radeonsi/r600_texture.c b/src/gallium/drivers/radeonsi/r600_texture.c
new file mode 100644
index 00000000000..c9e1b832113
--- /dev/null
+++ b/src/gallium/drivers/radeonsi/r600_texture.c
@@ -0,0 +1,825 @@
+/*
+ * Copyright 2010 Jerome Glisse <[email protected]>
+ *
+ * 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.
+ *
+ * Authors:
+ * Jerome Glisse
+ * Corbin Simpson
+ */
+#include <errno.h>
+#include "pipe/p_screen.h"
+#include "util/u_format.h"
+#include "util/u_format_s3tc.h"
+#include "util/u_math.h"
+#include "util/u_inlines.h"
+#include "util/u_memory.h"
+#include "pipebuffer/pb_buffer.h"
+#include "radeonsi_pipe.h"
+#include "r600_resource.h"
+#include "sid.h"
+
+/* Copy from a full GPU texture to a transfer's staging one. */
+static void r600_copy_to_staging_texture(struct pipe_context *ctx, struct r600_transfer *rtransfer)
+{
+ struct pipe_transfer *transfer = (struct pipe_transfer*)rtransfer;
+ struct pipe_resource *texture = transfer->resource;
+
+ ctx->resource_copy_region(ctx, rtransfer->staging_texture,
+ 0, 0, 0, 0, texture, transfer->level,
+ &transfer->box);
+}
+
+
+/* Copy from a transfer's staging texture to a full GPU one. */
+static void r600_copy_from_staging_texture(struct pipe_context *ctx, struct r600_transfer *rtransfer)
+{
+ struct pipe_transfer *transfer = (struct pipe_transfer*)rtransfer;
+ struct pipe_resource *texture = transfer->resource;
+ struct pipe_box sbox;
+
+ sbox.x = sbox.y = sbox.z = 0;
+ sbox.width = transfer->box.width;
+ sbox.height = transfer->box.height;
+ /* XXX that might be wrong */
+ sbox.depth = 1;
+ ctx->resource_copy_region(ctx, texture, transfer->level,
+ transfer->box.x, transfer->box.y, transfer->box.z,
+ rtransfer->staging_texture,
+ 0, &sbox);
+}
+
+unsigned r600_texture_get_offset(struct r600_resource_texture *rtex,
+ unsigned level, unsigned layer)
+{
+ unsigned offset = rtex->offset[level];
+
+ switch (rtex->resource.b.b.b.target) {
+ case PIPE_TEXTURE_3D:
+ case PIPE_TEXTURE_CUBE:
+ default:
+ return offset + layer * rtex->layer_size[level];
+ }
+}
+
+static unsigned r600_get_block_alignment(struct pipe_screen *screen,
+ enum pipe_format format,
+ unsigned array_mode)
+{
+ struct r600_screen* rscreen = (struct r600_screen *)screen;
+ unsigned pixsize = util_format_get_blocksize(format);
+ int p_align;
+
+ switch(array_mode) {
+#if 0
+ case V_038000_ARRAY_1D_TILED_THIN1:
+ p_align = MAX2(8,
+ ((rscreen->tiling_info.group_bytes / 8 / pixsize)));
+ break;
+ case V_038000_ARRAY_2D_TILED_THIN1:
+ p_align = MAX2(rscreen->tiling_info.num_banks,
+ (((rscreen->tiling_info.group_bytes / 8 / pixsize)) *
+ rscreen->tiling_info.num_banks)) * 8;
+ break;
+ case V_038000_ARRAY_LINEAR_ALIGNED:
+ p_align = MAX2(64, rscreen->tiling_info.group_bytes / pixsize);
+ break;
+ case V_038000_ARRAY_LINEAR_GENERAL:
+#endif
+ default:
+ p_align = rscreen->tiling_info.group_bytes / pixsize;
+ break;
+ }
+ return p_align;
+}
+
+static unsigned r600_get_height_alignment(struct pipe_screen *screen,
+ unsigned array_mode)
+{
+ struct r600_screen* rscreen = (struct r600_screen *)screen;
+ int h_align;
+
+ switch (array_mode) {
+#if 0
+ case V_038000_ARRAY_2D_TILED_THIN1:
+ h_align = rscreen->tiling_info.num_channels * 8;
+ break;
+ case V_038000_ARRAY_1D_TILED_THIN1:
+ case V_038000_ARRAY_LINEAR_ALIGNED:
+ h_align = 8;
+ break;
+ case V_038000_ARRAY_LINEAR_GENERAL:
+#endif
+ default:
+ h_align = 1;
+ break;
+ }
+ return h_align;
+}
+
+static unsigned r600_get_base_alignment(struct pipe_screen *screen,
+ enum pipe_format format,
+ unsigned array_mode)
+{
+ struct r600_screen* rscreen = (struct r600_screen *)screen;
+ unsigned pixsize = util_format_get_blocksize(format);
+ int p_align = r600_get_block_alignment(screen, format, array_mode);
+ int h_align = r600_get_height_alignment(screen, array_mode);
+ int b_align;
+
+ switch (array_mode) {
+#if 0
+ case V_038000_ARRAY_2D_TILED_THIN1:
+ b_align = MAX2(rscreen->tiling_info.num_banks * rscreen->tiling_info.num_channels * 8 * 8 * pixsize,
+ p_align * pixsize * h_align);
+ break;
+ case V_038000_ARRAY_1D_TILED_THIN1:
+ case V_038000_ARRAY_LINEAR_ALIGNED:
+ case V_038000_ARRAY_LINEAR_GENERAL:
+#endif
+ default:
+ b_align = rscreen->tiling_info.group_bytes;
+ break;
+ }
+ return b_align;
+}
+
+static unsigned mip_minify(unsigned size, unsigned level)
+{
+ unsigned val;
+ val = u_minify(size, level);
+ if (level > 0)
+ val = util_next_power_of_two(val);
+ return val;
+}
+
+static unsigned r600_texture_get_nblocksx(struct pipe_screen *screen,
+ struct r600_resource_texture *rtex,
+ unsigned level)
+{
+ struct pipe_resource *ptex = &rtex->resource.b.b.b;
+ unsigned nblocksx, block_align, width;
+ unsigned blocksize = util_format_get_blocksize(rtex->real_format);
+
+ if (rtex->pitch_override)
+ return rtex->pitch_override / blocksize;
+
+ width = mip_minify(ptex->width0, level);
+ nblocksx = util_format_get_nblocksx(rtex->real_format, width);
+
+ block_align = r600_get_block_alignment(screen, rtex->real_format,
+ rtex->array_mode[level]);
+ nblocksx = align(nblocksx, block_align);
+ return nblocksx;
+}
+
+static unsigned r600_texture_get_nblocksy(struct pipe_screen *screen,
+ struct r600_resource_texture *rtex,
+ unsigned level)
+{
+ struct pipe_resource *ptex = &rtex->resource.b.b.b;
+ unsigned height, tile_height;
+
+ height = mip_minify(ptex->height0, level);
+ height = util_format_get_nblocksy(rtex->real_format, height);
+ tile_height = r600_get_height_alignment(screen,
+ rtex->array_mode[level]);
+
+ /* XXX Hack around an alignment issue. Less tests fail with this.
+ *
+ * The thing is depth-stencil buffers should be tiled, i.e.
+ * the alignment should be >=8. If I make them tiled, stencil starts
+ * working because it no longer overlaps with the depth buffer
+ * in memory, but texturing like drawpix-stencil breaks. */
+ if (util_format_is_depth_or_stencil(rtex->real_format) && tile_height < 8)
+ tile_height = 8;
+
+ height = align(height, tile_height);
+ return height;
+}
+
+static void r600_texture_set_array_mode(struct pipe_screen *screen,
+ struct r600_resource_texture *rtex,
+ unsigned level, unsigned array_mode)
+{
+ struct pipe_resource *ptex = &rtex->resource.b.b.b;
+
+ switch (array_mode) {
+#if 0
+ case V_0280A0_ARRAY_LINEAR_GENERAL:
+ case V_0280A0_ARRAY_LINEAR_ALIGNED:
+ case V_0280A0_ARRAY_1D_TILED_THIN1:
+#endif
+ default:
+ rtex->array_mode[level] = array_mode;
+ break;
+#if 0
+ case V_0280A0_ARRAY_2D_TILED_THIN1:
+ {
+ unsigned w, h, tile_height, tile_width;
+
+ tile_height = r600_get_height_alignment(screen, array_mode);
+ tile_width = r600_get_block_alignment(screen, rtex->real_format, array_mode);
+
+ w = mip_minify(ptex->width0, level);
+ h = mip_minify(ptex->height0, level);
+ if (w <= tile_width || h <= tile_height)
+ rtex->array_mode[level] = V_0280A0_ARRAY_1D_TILED_THIN1;
+ else
+ rtex->array_mode[level] = array_mode;
+ }
+ break;
+#endif
+ }
+}
+
+static void r600_setup_miptree(struct pipe_screen *screen,
+ struct r600_resource_texture *rtex,
+ unsigned array_mode)
+{
+ struct pipe_resource *ptex = &rtex->resource.b.b.b;
+ enum chip_class chipc = ((struct r600_screen*)screen)->chip_class;
+ unsigned size, layer_size, i, offset;
+ unsigned nblocksx, nblocksy;
+
+ for (i = 0, offset = 0; i <= ptex->last_level; i++) {
+ unsigned blocksize = util_format_get_blocksize(rtex->real_format);
+ unsigned base_align = r600_get_base_alignment(screen, rtex->real_format, array_mode);
+
+ r600_texture_set_array_mode(screen, rtex, i, array_mode);
+
+ nblocksx = r600_texture_get_nblocksx(screen, rtex, i);
+ nblocksy = r600_texture_get_nblocksy(screen, rtex, i);
+
+ if (chipc >= CAYMAN /*&& array_mode == V_038000_ARRAY_LINEAR_GENERAL*/)
+ layer_size = align(nblocksx, 64) * nblocksy * blocksize;
+ else
+ layer_size = nblocksx * nblocksy * blocksize;
+
+ if (ptex->target == PIPE_TEXTURE_CUBE) {
+ if (chipc >= CAYMAN)
+ size = layer_size * 8;
+ }
+ else if (ptex->target == PIPE_TEXTURE_3D)
+ size = layer_size * u_minify(ptex->depth0, i);
+ else
+ size = layer_size * ptex->array_size;
+
+ /* align base image and start of miptree */
+ if ((i == 0) || (i == 1))
+ offset = align(offset, base_align);
+ rtex->offset[i] = offset;
+ rtex->layer_size[i] = layer_size;
+ rtex->pitch_in_blocks[i] = nblocksx; /* CB talks in elements */
+ rtex->pitch_in_bytes[i] = nblocksx * blocksize;
+
+ offset += size;
+ }
+ rtex->size = offset;
+}
+
+/* Figure out whether u_blitter will fallback to a transfer operation.
+ * If so, don't use a staging resource.
+ */
+static boolean permit_hardware_blit(struct pipe_screen *screen,
+ const struct pipe_resource *res)
+{
+ unsigned bind;
+
+ if (util_format_is_depth_or_stencil(res->format))
+ bind = PIPE_BIND_DEPTH_STENCIL;
+ else
+ bind = PIPE_BIND_RENDER_TARGET;
+
+ /* hackaround for S3TC */
+ if (util_format_is_compressed(res->format))
+ return TRUE;
+
+ if (!screen->is_format_supported(screen,
+ res->format,
+ res->target,
+ res->nr_samples,
+ bind))
+ return FALSE;
+
+ if (!screen->is_format_supported(screen,
+ res->format,
+ res->target,
+ res->nr_samples,
+ PIPE_BIND_SAMPLER_VIEW))
+ return FALSE;
+
+ switch (res->usage) {
+ case PIPE_USAGE_STREAM:
+ case PIPE_USAGE_STAGING:
+ return FALSE;
+
+ default:
+ return TRUE;
+ }
+}
+
+static boolean r600_texture_get_handle(struct pipe_screen* screen,
+ struct pipe_resource *ptex,
+ struct winsys_handle *whandle)
+{
+ struct r600_resource_texture *rtex = (struct r600_resource_texture*)ptex;
+ struct r600_resource *resource = &rtex->resource;
+ struct r600_screen *rscreen = (struct r600_screen*)screen;
+
+ return rscreen->ws->buffer_get_handle(resource->buf,
+ rtex->pitch_in_bytes[0], whandle);
+}
+
+static void r600_texture_destroy(struct pipe_screen *screen,
+ struct pipe_resource *ptex)
+{
+ struct r600_resource_texture *rtex = (struct r600_resource_texture*)ptex;
+ struct r600_resource *resource = &rtex->resource;
+
+ if (rtex->flushed_depth_texture)
+ pipe_resource_reference((struct pipe_resource **)&rtex->flushed_depth_texture, NULL);
+
+ if (rtex->stencil)
+ pipe_resource_reference((struct pipe_resource **)&rtex->stencil, NULL);
+
+ pb_reference(&resource->buf, NULL);
+ FREE(rtex);
+}
+
+static const struct u_resource_vtbl r600_texture_vtbl =
+{
+ r600_texture_get_handle, /* get_handle */
+ r600_texture_destroy, /* resource_destroy */
+ r600_texture_get_transfer, /* get_transfer */
+ r600_texture_transfer_destroy, /* transfer_destroy */
+ r600_texture_transfer_map, /* transfer_map */
+ u_default_transfer_flush_region,/* transfer_flush_region */
+ r600_texture_transfer_unmap, /* transfer_unmap */
+ u_default_transfer_inline_write /* transfer_inline_write */
+};
+
+static struct r600_resource_texture *
+r600_texture_create_object(struct pipe_screen *screen,
+ const struct pipe_resource *base,
+ unsigned array_mode,
+ unsigned pitch_in_bytes_override,
+ unsigned max_buffer_size,
+ struct pb_buffer *buf,
+ boolean alloc_bo)
+{
+ struct r600_resource_texture *rtex;
+ struct r600_resource *resource;
+ struct r600_screen *rscreen = (struct r600_screen*)screen;
+
+ rtex = CALLOC_STRUCT(r600_resource_texture);
+ if (rtex == NULL)
+ return NULL;
+
+ resource = &rtex->resource;
+ resource->b.b.b = *base;
+ resource->b.b.vtbl = &r600_texture_vtbl;
+ pipe_reference_init(&resource->b.b.b.reference, 1);
+ resource->b.b.b.screen = screen;
+ rtex->pitch_override = pitch_in_bytes_override;
+ rtex->real_format = base->format;
+
+ /* We must split depth and stencil into two separate buffers on Evergreen. */
+ if (!(base->flags & R600_RESOURCE_FLAG_TRANSFER) &&
+ ((struct r600_screen*)screen)->chip_class >= CAYMAN &&
+ util_format_is_depth_and_stencil(base->format)) {
+ struct pipe_resource stencil;
+ unsigned stencil_pitch_override = 0;
+
+ switch (base->format) {
+ case PIPE_FORMAT_Z24_UNORM_S8_UINT:
+ rtex->real_format = PIPE_FORMAT_Z24X8_UNORM;
+ break;
+ case PIPE_FORMAT_S8_UINT_Z24_UNORM:
+ rtex->real_format = PIPE_FORMAT_X8Z24_UNORM;
+ break;
+ case PIPE_FORMAT_Z32_FLOAT_S8X24_UINT:
+ rtex->real_format = PIPE_FORMAT_Z32_FLOAT;
+ break;
+ default:
+ assert(0);
+ FREE(rtex);
+ return NULL;
+ }
+
+ /* Divide the pitch in bytes by 4 for stencil, because it has a smaller pixel size. */
+ if (pitch_in_bytes_override) {
+ assert(base->format == PIPE_FORMAT_Z24_UNORM_S8_UINT ||
+ base->format == PIPE_FORMAT_S8_UINT_Z24_UNORM);
+ stencil_pitch_override = pitch_in_bytes_override / 4;
+ }
+
+ /* Allocate the stencil buffer. */
+ stencil = *base;
+ stencil.format = PIPE_FORMAT_S8_UINT;
+ rtex->stencil = r600_texture_create_object(screen, &stencil, array_mode,
+ stencil_pitch_override,
+ max_buffer_size, NULL, FALSE);
+ if (!rtex->stencil) {
+ FREE(rtex);
+ return NULL;
+ }
+ /* Proceed in creating the depth buffer. */
+ }
+
+ /* only mark depth textures the HW can hit as depth textures */
+ if (util_format_is_depth_or_stencil(rtex->real_format) && permit_hardware_blit(screen, base))
+ rtex->depth = 1;
+
+ r600_setup_miptree(screen, rtex, array_mode);
+
+ /* If we initialized separate stencil for Evergreen. place it after depth. */
+ if (rtex->stencil) {
+ unsigned stencil_align, stencil_offset;
+
+ stencil_align = r600_get_base_alignment(screen, rtex->stencil->real_format, array_mode);
+ stencil_offset = align(rtex->size, stencil_align);
+
+ for (unsigned i = 0; i <= rtex->stencil->resource.b.b.b.last_level; i++)
+ rtex->stencil->offset[i] += stencil_offset;
+
+ rtex->size = stencil_offset + rtex->stencil->size;
+ }
+
+ /* Now create the backing buffer. */
+ if (!buf && alloc_bo) {
+ struct pipe_resource *ptex = &rtex->resource.b.b.b;
+ unsigned base_align = r600_get_base_alignment(screen, ptex->format, array_mode);
+
+ if (!r600_init_resource(rscreen, resource, rtex->size, base_align, base->bind, base->usage)) {
+ pipe_resource_reference((struct pipe_resource**)&rtex->stencil, NULL);
+ FREE(rtex);
+ return NULL;
+ }
+ } else if (buf) {
+ resource->buf = buf;
+ resource->cs_buf = rscreen->ws->buffer_get_cs_handle(buf);
+ resource->domains = RADEON_DOMAIN_GTT | RADEON_DOMAIN_VRAM;
+ }
+
+ if (rtex->stencil) {
+ pb_reference(&rtex->stencil->resource.buf, rtex->resource.buf);
+ rtex->stencil->resource.cs_buf = rtex->resource.cs_buf;
+ rtex->stencil->resource.domains = rtex->resource.domains;
+ }
+ return rtex;
+}
+
+DEBUG_GET_ONCE_BOOL_OPTION(tiling_enabled, "R600_TILING", FALSE);
+
+struct pipe_resource *r600_texture_create(struct pipe_screen *screen,
+ const struct pipe_resource *templ)
+{
+ struct r600_screen *rscreen = (struct r600_screen*)screen;
+ unsigned array_mode = 0;
+
+ if (!(templ->flags & R600_RESOURCE_FLAG_TRANSFER) &&
+ !(templ->bind & PIPE_BIND_SCANOUT)) {
+#if 0
+ if (util_format_is_compressed(templ->format)) {
+ array_mode = V_038000_ARRAY_1D_TILED_THIN1;
+ }
+ else if (debug_get_option_tiling_enabled() &&
+ rscreen->info.drm_minor >= 9 &&
+ permit_hardware_blit(screen, templ)) {
+ array_mode = V_038000_ARRAY_2D_TILED_THIN1;
+ }
+#endif
+ }
+
+ return (struct pipe_resource *)r600_texture_create_object(screen, templ, array_mode,
+ 0, 0, NULL, TRUE);
+}
+
+static struct pipe_surface *r600_create_surface(struct pipe_context *pipe,
+ struct pipe_resource *texture,
+ const struct pipe_surface *surf_tmpl)
+{
+ struct r600_resource_texture *rtex = (struct r600_resource_texture*)texture;
+ struct r600_surface *surface = CALLOC_STRUCT(r600_surface);
+ unsigned level = surf_tmpl->u.tex.level;
+
+ assert(surf_tmpl->u.tex.first_layer == surf_tmpl->u.tex.last_layer);
+ if (surface == NULL)
+ return NULL;
+ /* XXX no offset */
+/* offset = r600_texture_get_offset(rtex, level, surf_tmpl->u.tex.first_layer);*/
+ pipe_reference_init(&surface->base.reference, 1);
+ pipe_resource_reference(&surface->base.texture, texture);
+ surface->base.context = pipe;
+ surface->base.format = surf_tmpl->format;
+ surface->base.width = mip_minify(texture->width0, level);
+ surface->base.height = mip_minify(texture->height0, level);
+ surface->base.usage = surf_tmpl->usage;
+ surface->base.texture = texture;
+ surface->base.u.tex.first_layer = surf_tmpl->u.tex.first_layer;
+ surface->base.u.tex.last_layer = surf_tmpl->u.tex.last_layer;
+ surface->base.u.tex.level = level;
+
+ surface->aligned_height = r600_texture_get_nblocksy(pipe->screen,
+ rtex, level);
+ return &surface->base;
+}
+
+static void r600_surface_destroy(struct pipe_context *pipe,
+ struct pipe_surface *surface)
+{
+ pipe_resource_reference(&surface->texture, NULL);
+ FREE(surface);
+}
+
+struct pipe_resource *r600_texture_from_handle(struct pipe_screen *screen,
+ const struct pipe_resource *templ,
+ struct winsys_handle *whandle)
+{
+ struct r600_screen *rscreen = (struct r600_screen*)screen;
+ struct pb_buffer *buf = NULL;
+ unsigned stride = 0;
+ unsigned array_mode = 0;
+ enum radeon_bo_layout micro, macro;
+
+ /* Support only 2D textures without mipmaps */
+ if ((templ->target != PIPE_TEXTURE_2D && templ->target != PIPE_TEXTURE_RECT) ||
+ templ->depth0 != 1 || templ->last_level != 0)
+ return NULL;
+
+ buf = rscreen->ws->buffer_from_handle(rscreen->ws, whandle, &stride);
+ if (!buf)
+ return NULL;
+
+ rscreen->ws->buffer_get_tiling(buf, &micro, &macro, NULL, NULL, NULL, NULL, NULL);
+
+#if 0
+ if (macro == RADEON_LAYOUT_TILED)
+ array_mode = V_0280A0_ARRAY_2D_TILED_THIN1;
+ else if (micro == RADEON_LAYOUT_TILED)
+ array_mode = V_0280A0_ARRAY_1D_TILED_THIN1;
+ else
+#endif
+ array_mode = 0;
+
+ return (struct pipe_resource *)r600_texture_create_object(screen, templ, array_mode,
+ stride, 0, buf, FALSE);
+}
+
+int r600_texture_depth_flush(struct pipe_context *ctx,
+ struct pipe_resource *texture, boolean just_create)
+{
+ struct r600_resource_texture *rtex = (struct r600_resource_texture*)texture;
+ struct pipe_resource resource;
+
+ if (rtex->flushed_depth_texture)
+ goto out;
+
+ resource.target = texture->target;
+ resource.format = texture->format;
+ resource.width0 = texture->width0;
+ resource.height0 = texture->height0;
+ resource.depth0 = texture->depth0;
+ resource.array_size = texture->array_size;
+ resource.last_level = texture->last_level;
+ resource.nr_samples = texture->nr_samples;
+ resource.usage = PIPE_USAGE_DYNAMIC;
+ resource.bind = texture->bind | PIPE_BIND_DEPTH_STENCIL;
+ resource.flags = R600_RESOURCE_FLAG_TRANSFER | texture->flags;
+
+ rtex->flushed_depth_texture = (struct r600_resource_texture *)ctx->screen->resource_create(ctx->screen, &resource);
+ if (rtex->flushed_depth_texture == NULL) {
+ R600_ERR("failed to create temporary texture to hold untiled copy\n");
+ return -ENOMEM;
+ }
+
+ ((struct r600_resource_texture *)rtex->flushed_depth_texture)->is_flushing_texture = TRUE;
+out:
+ if (just_create)
+ return 0;
+
+ /* XXX: only do this if the depth texture has actually changed:
+ */
+ r600_blit_uncompress_depth(ctx, rtex);
+ return 0;
+}
+
+/* Needs adjustment for pixelformat:
+ */
+static INLINE unsigned u_box_volume( const struct pipe_box *box )
+{
+ return box->width * box->depth * box->height;
+};
+
+struct pipe_transfer* r600_texture_get_transfer(struct pipe_context *ctx,
+ struct pipe_resource *texture,
+ unsigned level,
+ unsigned usage,
+ const struct pipe_box *box)
+{
+ struct r600_resource_texture *rtex = (struct r600_resource_texture*)texture;
+ struct pipe_resource resource;
+ struct r600_transfer *trans;
+ int r;
+ boolean use_staging_texture = FALSE;
+
+#if 0
+ /* We cannot map a tiled texture directly because the data is
+ * in a different order, therefore we do detiling using a blit.
+ *
+ * Also, use a temporary in GTT memory for read transfers, as
+ * the CPU is much happier reading out of cached system memory
+ * than uncached VRAM.
+ */
+ if (R600_TEX_IS_TILED(rtex, level))
+ use_staging_texture = TRUE;
+#endif
+
+ if ((usage & PIPE_TRANSFER_READ) && u_box_volume(box) > 1024)
+ use_staging_texture = TRUE;
+
+ /* XXX: Use a staging texture for uploads if the underlying BO
+ * is busy. No interface for checking that currently? so do
+ * it eagerly whenever the transfer doesn't require a readback
+ * and might block.
+ */
+ if ((usage & PIPE_TRANSFER_WRITE) &&
+ !(usage & (PIPE_TRANSFER_READ |
+ PIPE_TRANSFER_DONTBLOCK |
+ PIPE_TRANSFER_UNSYNCHRONIZED)))
+ use_staging_texture = TRUE;
+
+ if (!permit_hardware_blit(ctx->screen, texture) ||
+ (texture->flags & R600_RESOURCE_FLAG_TRANSFER))
+ use_staging_texture = FALSE;
+
+ if (use_staging_texture && (usage & PIPE_TRANSFER_MAP_DIRECTLY))
+ return NULL;
+
+ trans = CALLOC_STRUCT(r600_transfer);
+ if (trans == NULL)
+ return NULL;
+ pipe_resource_reference(&trans->transfer.resource, texture);
+ trans->transfer.level = level;
+ trans->transfer.usage = usage;
+ trans->transfer.box = *box;
+ if (rtex->depth) {
+ /* XXX: only readback the rectangle which is being mapped?
+ */
+ /* XXX: when discard is true, no need to read back from depth texture
+ */
+ r = r600_texture_depth_flush(ctx, texture, FALSE);
+ if (r < 0) {
+ R600_ERR("failed to create temporary texture to hold untiled copy\n");
+ pipe_resource_reference(&trans->transfer.resource, NULL);
+ FREE(trans);
+ return NULL;
+ }
+ trans->transfer.stride = rtex->flushed_depth_texture->pitch_in_bytes[level];
+ trans->offset = r600_texture_get_offset(rtex->flushed_depth_texture, level, box->z);
+ return &trans->transfer;
+ } else if (use_staging_texture) {
+ resource.target = PIPE_TEXTURE_2D;
+ resource.format = texture->format;
+ resource.width0 = box->width;
+ resource.height0 = box->height;
+ resource.depth0 = 1;
+ resource.array_size = 1;
+ resource.last_level = 0;
+ resource.nr_samples = 0;
+ resource.usage = PIPE_USAGE_STAGING;
+ resource.bind = 0;
+ resource.flags = R600_RESOURCE_FLAG_TRANSFER;
+ /* For texture reading, the temporary (detiled) texture is used as
+ * a render target when blitting from a tiled texture. */
+ if (usage & PIPE_TRANSFER_READ) {
+ resource.bind |= PIPE_BIND_RENDER_TARGET;
+ }
+ /* For texture writing, the temporary texture is used as a sampler
+ * when blitting into a tiled texture. */
+ if (usage & PIPE_TRANSFER_WRITE) {
+ resource.bind |= PIPE_BIND_SAMPLER_VIEW;
+ }
+ /* Create the temporary texture. */
+ trans->staging_texture = ctx->screen->resource_create(ctx->screen, &resource);
+ if (trans->staging_texture == NULL) {
+ R600_ERR("failed to create temporary texture to hold untiled copy\n");
+ pipe_resource_reference(&trans->transfer.resource, NULL);
+ FREE(trans);
+ return NULL;
+ }
+
+ trans->transfer.stride =
+ ((struct r600_resource_texture *)trans->staging_texture)->pitch_in_bytes[0];
+ if (usage & PIPE_TRANSFER_READ) {
+ r600_copy_to_staging_texture(ctx, trans);
+ /* Always referenced in the blit. */
+ radeonsi_flush(ctx, NULL, 0);
+ }
+ return &trans->transfer;
+ }
+ trans->transfer.stride = rtex->pitch_in_bytes[level];
+ trans->transfer.layer_stride = rtex->layer_size[level];
+ trans->offset = r600_texture_get_offset(rtex, level, box->z);
+ return &trans->transfer;
+}
+
+void r600_texture_transfer_destroy(struct pipe_context *ctx,
+ struct pipe_transfer *transfer)
+{
+ struct r600_transfer *rtransfer = (struct r600_transfer*)transfer;
+ struct pipe_resource *texture = transfer->resource;
+ struct r600_resource_texture *rtex = (struct r600_resource_texture*)texture;
+
+ if (rtransfer->staging_texture) {
+ if (transfer->usage & PIPE_TRANSFER_WRITE) {
+ r600_copy_from_staging_texture(ctx, rtransfer);
+ }
+ pipe_resource_reference(&rtransfer->staging_texture, NULL);
+ }
+
+ if (rtex->depth && !rtex->is_flushing_texture) {
+ if ((transfer->usage & PIPE_TRANSFER_WRITE) && rtex->flushed_depth_texture)
+ r600_blit_push_depth(ctx, rtex);
+ }
+
+ pipe_resource_reference(&transfer->resource, NULL);
+ FREE(transfer);
+}
+
+void* r600_texture_transfer_map(struct pipe_context *ctx,
+ struct pipe_transfer* transfer)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct r600_transfer *rtransfer = (struct r600_transfer*)transfer;
+ struct pb_buffer *buf;
+ enum pipe_format format = transfer->resource->format;
+ unsigned offset = 0;
+ char *map;
+
+ if (rtransfer->staging_texture) {
+ buf = ((struct r600_resource *)rtransfer->staging_texture)->buf;
+ } else {
+ struct r600_resource_texture *rtex = (struct r600_resource_texture*)transfer->resource;
+
+ if (rtex->flushed_depth_texture)
+ buf = ((struct r600_resource *)rtex->flushed_depth_texture)->buf;
+ else
+ buf = ((struct r600_resource *)transfer->resource)->buf;
+
+ offset = rtransfer->offset +
+ transfer->box.y / util_format_get_blockheight(format) * transfer->stride +
+ transfer->box.x / util_format_get_blockwidth(format) * util_format_get_blocksize(format);
+ }
+
+ if (!(map = rctx->ws->buffer_map(buf, rctx->cs, transfer->usage))) {
+ return NULL;
+ }
+
+ return map + offset;
+}
+
+void r600_texture_transfer_unmap(struct pipe_context *ctx,
+ struct pipe_transfer* transfer)
+{
+ struct r600_transfer *rtransfer = (struct r600_transfer*)transfer;
+ struct r600_context *rctx = (struct r600_context*)ctx;
+ struct pb_buffer *buf;
+
+ if (rtransfer->staging_texture) {
+ buf = ((struct r600_resource *)rtransfer->staging_texture)->buf;
+ } else {
+ struct r600_resource_texture *rtex = (struct r600_resource_texture*)transfer->resource;
+
+ if (rtex->flushed_depth_texture) {
+ buf = ((struct r600_resource *)rtex->flushed_depth_texture)->buf;
+ } else {
+ buf = ((struct r600_resource *)transfer->resource)->buf;
+ }
+ }
+ rctx->ws->buffer_unmap(buf);
+}
+
+void r600_init_surface_functions(struct r600_context *r600)
+{
+ r600->context.create_surface = r600_create_surface;
+ r600->context.surface_destroy = r600_surface_destroy;
+}
diff --git a/src/gallium/drivers/radeonsi/r600_translate.c b/src/gallium/drivers/radeonsi/r600_translate.c
new file mode 100644
index 00000000000..6551044b553
--- /dev/null
+++ b/src/gallium/drivers/radeonsi/r600_translate.c
@@ -0,0 +1,54 @@
+/*
+ * Copyright 2010 Red Hat Inc.
+ *
+ * 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.
+ *
+ * Authors: Dave Airlie <[email protected]>
+ */
+
+#include "util/u_index_modify.h"
+#include "util/u_inlines.h"
+#include "util/u_upload_mgr.h"
+#include "radeonsi_pipe.h"
+
+
+void r600_translate_index_buffer(struct r600_context *r600,
+ struct pipe_index_buffer *ib,
+ unsigned count)
+{
+ struct pipe_resource *out_buffer = NULL;
+ unsigned out_offset;
+ void *ptr;
+
+ switch (ib->index_size) {
+ case 1:
+ u_upload_alloc(r600->vbuf_mgr->uploader, 0, count * 2,
+ &out_offset, &out_buffer, &ptr);
+
+ util_shorten_ubyte_elts_to_userptr(
+ &r600->context, ib->buffer, 0, ib->offset, count, ptr);
+
+ pipe_resource_reference(&ib->buffer, NULL);
+ ib->buffer = out_buffer;
+ ib->offset = out_offset;
+ ib->index_size = 2;
+ break;
+ }
+}
diff --git a/src/gallium/drivers/radeonsi/radeonsi_pipe.c b/src/gallium/drivers/radeonsi/radeonsi_pipe.c
new file mode 100644
index 00000000000..9e849525cc9
--- /dev/null
+++ b/src/gallium/drivers/radeonsi/radeonsi_pipe.c
@@ -0,0 +1,731 @@
+/*
+ * Copyright 2010 Jerome Glisse <[email protected]>
+ *
+ * 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 <stdio.h>
+#include <errno.h>
+#include "pipe/p_defines.h"
+#include "pipe/p_state.h"
+#include "pipe/p_context.h"
+#include "tgsi/tgsi_scan.h"
+#include "tgsi/tgsi_parse.h"
+#include "tgsi/tgsi_util.h"
+#include "util/u_blitter.h"
+#include "util/u_double_list.h"
+#include "util/u_format.h"
+#include "util/u_format_s3tc.h"
+#include "util/u_transfer.h"
+#include "util/u_surface.h"
+#include "util/u_pack_color.h"
+#include "util/u_memory.h"
+#include "util/u_inlines.h"
+#include "util/u_upload_mgr.h"
+#include "vl/vl_decoder.h"
+#include "vl/vl_video_buffer.h"
+#include "os/os_time.h"
+#include "pipebuffer/pb_buffer.h"
+#include "r600.h"
+#include "sid.h"
+#include "r600_resource.h"
+#include "radeonsi_pipe.h"
+#include "r600_hw_context_priv.h"
+
+/*
+ * pipe_context
+ */
+static struct r600_fence *r600_create_fence(struct r600_context *rctx)
+{
+ struct r600_screen *rscreen = rctx->screen;
+ struct r600_fence *fence = NULL;
+
+ pipe_mutex_lock(rscreen->fences.mutex);
+
+ if (!rscreen->fences.bo) {
+ /* Create the shared buffer object */
+ rscreen->fences.bo = (struct r600_resource*)
+ pipe_buffer_create(&rscreen->screen, PIPE_BIND_CUSTOM,
+ PIPE_USAGE_STAGING, 4096);
+ if (!rscreen->fences.bo) {
+ R600_ERR("r600: failed to create bo for fence objects\n");
+ goto out;
+ }
+ rscreen->fences.data = rctx->ws->buffer_map(rscreen->fences.bo->buf,
+ rctx->cs,
+ PIPE_TRANSFER_READ_WRITE);
+ }
+
+ if (!LIST_IS_EMPTY(&rscreen->fences.pool)) {
+ struct r600_fence *entry;
+
+ /* Try to find a freed fence that has been signalled */
+ LIST_FOR_EACH_ENTRY(entry, &rscreen->fences.pool, head) {
+ if (rscreen->fences.data[entry->index] != 0) {
+ LIST_DELINIT(&entry->head);
+ fence = entry;
+ break;
+ }
+ }
+ }
+
+ if (!fence) {
+ /* Allocate a new fence */
+ struct r600_fence_block *block;
+ unsigned index;
+
+ if ((rscreen->fences.next_index + 1) >= 1024) {
+ R600_ERR("r600: too many concurrent fences\n");
+ goto out;
+ }
+
+ index = rscreen->fences.next_index++;
+
+ if (!(index % FENCE_BLOCK_SIZE)) {
+ /* Allocate a new block */
+ block = CALLOC_STRUCT(r600_fence_block);
+ if (block == NULL)
+ goto out;
+
+ LIST_ADD(&block->head, &rscreen->fences.blocks);
+ } else {
+ block = LIST_ENTRY(struct r600_fence_block, rscreen->fences.blocks.next, head);
+ }
+
+ fence = &block->fences[index % FENCE_BLOCK_SIZE];
+ fence->index = index;
+ }
+
+ pipe_reference_init(&fence->reference, 1);
+
+ rscreen->fences.data[fence->index] = 0;
+ r600_context_emit_fence(rctx, rscreen->fences.bo, fence->index, 1);
+
+ /* Create a dummy BO so that fence_finish without a timeout can sleep waiting for completion */
+ fence->sleep_bo = (struct r600_resource*)
+ pipe_buffer_create(&rctx->screen->screen, PIPE_BIND_CUSTOM,
+ PIPE_USAGE_STAGING, 1);
+ /* Add the fence as a dummy relocation. */
+ r600_context_bo_reloc(rctx, fence->sleep_bo, RADEON_USAGE_READWRITE);
+
+out:
+ pipe_mutex_unlock(rscreen->fences.mutex);
+ return fence;
+}
+
+
+void radeonsi_flush(struct pipe_context *ctx, struct pipe_fence_handle **fence,
+ unsigned flags)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct r600_fence **rfence = (struct r600_fence**)fence;
+ struct pipe_query *render_cond = NULL;
+ unsigned render_cond_mode = 0;
+
+ if (rfence)
+ *rfence = r600_create_fence(rctx);
+
+ /* Disable render condition. */
+ if (rctx->current_render_cond) {
+ render_cond = rctx->current_render_cond;
+ render_cond_mode = rctx->current_render_cond_mode;
+ ctx->render_condition(ctx, NULL, 0);
+ }
+
+ r600_context_flush(rctx, flags);
+
+ /* Re-enable render condition. */
+ if (render_cond) {
+ ctx->render_condition(ctx, render_cond, render_cond_mode);
+ }
+}
+
+static void r600_flush_from_st(struct pipe_context *ctx,
+ struct pipe_fence_handle **fence)
+{
+ radeonsi_flush(ctx, fence, 0);
+}
+
+static void r600_flush_from_winsys(void *ctx, unsigned flags)
+{
+ radeonsi_flush((struct pipe_context*)ctx, NULL, flags);
+}
+
+static void r600_update_num_contexts(struct r600_screen *rscreen, int diff)
+{
+ pipe_mutex_lock(rscreen->mutex_num_contexts);
+ if (diff > 0) {
+ rscreen->num_contexts++;
+
+ if (rscreen->num_contexts > 1)
+ util_slab_set_thread_safety(&rscreen->pool_buffers,
+ UTIL_SLAB_MULTITHREADED);
+ } else {
+ rscreen->num_contexts--;
+
+ if (rscreen->num_contexts <= 1)
+ util_slab_set_thread_safety(&rscreen->pool_buffers,
+ UTIL_SLAB_SINGLETHREADED);
+ }
+ pipe_mutex_unlock(rscreen->mutex_num_contexts);
+}
+
+static void r600_destroy_context(struct pipe_context *context)
+{
+ struct r600_context *rctx = (struct r600_context *)context;
+
+ rctx->context.delete_depth_stencil_alpha_state(&rctx->context, rctx->custom_dsa_flush);
+ util_unreference_framebuffer_state(&rctx->framebuffer);
+
+ r600_context_fini(rctx);
+
+ util_blitter_destroy(rctx->blitter);
+
+ for (int i = 0; i < R600_PIPE_NSTATES; i++) {
+ free(rctx->states[i]);
+ }
+
+ u_vbuf_destroy(rctx->vbuf_mgr);
+ util_slab_destroy(&rctx->pool_transfers);
+
+ r600_update_num_contexts(rctx->screen, -1);
+
+ FREE(rctx);
+}
+
+static struct pipe_context *r600_create_context(struct pipe_screen *screen, void *priv)
+{
+ struct r600_context *rctx = CALLOC_STRUCT(r600_context);
+ struct r600_screen* rscreen = (struct r600_screen *)screen;
+
+ if (rctx == NULL)
+ return NULL;
+
+ r600_update_num_contexts(rscreen, 1);
+
+ rctx->context.screen = screen;
+ rctx->context.priv = priv;
+ rctx->context.destroy = r600_destroy_context;
+ rctx->context.flush = r600_flush_from_st;
+
+ /* Easy accessing of screen/winsys. */
+ rctx->screen = rscreen;
+ rctx->ws = rscreen->ws;
+ rctx->family = rscreen->family;
+ rctx->chip_class = rscreen->chip_class;
+
+ r600_init_blit_functions(rctx);
+ r600_init_query_functions(rctx);
+ r600_init_context_resource_functions(rctx);
+ r600_init_surface_functions(rctx);
+ rctx->context.draw_vbo = r600_draw_vbo;
+
+ rctx->context.create_video_decoder = vl_create_decoder;
+ rctx->context.create_video_buffer = vl_video_buffer_create;
+
+ r600_init_common_atoms(rctx);
+
+ switch (rctx->chip_class) {
+ case TAHITI:
+ cayman_init_state_functions(rctx);
+ if (si_context_init(rctx)) {
+ r600_destroy_context(&rctx->context);
+ return NULL;
+ }
+ si_init_config(rctx);
+ rctx->custom_dsa_flush = cayman_create_db_flush_dsa(rctx);
+ break;
+ default:
+ R600_ERR("Unsupported chip class %d.\n", rctx->chip_class);
+ r600_destroy_context(&rctx->context);
+ return NULL;
+ }
+
+ rctx->ws->cs_set_flush_callback(rctx->cs, r600_flush_from_winsys, rctx);
+
+ util_slab_create(&rctx->pool_transfers,
+ sizeof(struct pipe_transfer), 64,
+ UTIL_SLAB_SINGLETHREADED);
+
+ rctx->vbuf_mgr = u_vbuf_create(&rctx->context, 1024 * 1024, 256,
+ PIPE_BIND_VERTEX_BUFFER |
+ PIPE_BIND_INDEX_BUFFER |
+ PIPE_BIND_CONSTANT_BUFFER,
+ U_VERTEX_FETCH_DWORD_ALIGNED);
+ if (!rctx->vbuf_mgr) {
+ r600_destroy_context(&rctx->context);
+ return NULL;
+ }
+ rctx->vbuf_mgr->caps.format_fixed32 = 0;
+
+ rctx->blitter = util_blitter_create(&rctx->context);
+ if (rctx->blitter == NULL) {
+ r600_destroy_context(&rctx->context);
+ return NULL;
+ }
+
+ LIST_INITHEAD(&rctx->dirty_states);
+
+ r600_get_backend_mask(rctx); /* this emits commands and must be last */
+
+ return &rctx->context;
+}
+
+/*
+ * pipe_screen
+ */
+static const char* r600_get_vendor(struct pipe_screen* pscreen)
+{
+ return "X.Org";
+}
+
+static const char *r600_get_family_name(enum radeon_family family)
+{
+ switch(family) {
+ case CHIP_CAYMAN: return "AMD CAYMAN";
+ default: return "AMD unknown";
+ }
+}
+
+static const char* r600_get_name(struct pipe_screen* pscreen)
+{
+ struct r600_screen *rscreen = (struct r600_screen *)pscreen;
+
+ return r600_get_family_name(rscreen->family);
+}
+
+static int r600_get_param(struct pipe_screen* pscreen, enum pipe_cap param)
+{
+ struct r600_screen *rscreen = (struct r600_screen *)pscreen;
+ enum radeon_family family = rscreen->family;
+
+ switch (param) {
+ /* Supported features (boolean caps). */
+ case PIPE_CAP_NPOT_TEXTURES:
+ case PIPE_CAP_TWO_SIDED_STENCIL:
+ case PIPE_CAP_DUAL_SOURCE_BLEND:
+ case PIPE_CAP_ANISOTROPIC_FILTER:
+ case PIPE_CAP_POINT_SPRITE:
+ case PIPE_CAP_OCCLUSION_QUERY:
+ case PIPE_CAP_TEXTURE_SHADOW_MAP:
+ case PIPE_CAP_TEXTURE_MIRROR_CLAMP:
+ case PIPE_CAP_BLEND_EQUATION_SEPARATE:
+ case PIPE_CAP_TEXTURE_SWIZZLE:
+ case PIPE_CAP_DEPTHSTENCIL_CLEAR_SEPARATE:
+ case PIPE_CAP_DEPTH_CLIP_DISABLE:
+ case PIPE_CAP_SHADER_STENCIL_EXPORT:
+ case PIPE_CAP_VERTEX_ELEMENT_INSTANCE_DIVISOR:
+ case PIPE_CAP_MIXED_COLORBUFFER_FORMATS:
+ case PIPE_CAP_TGSI_FS_COORD_ORIGIN_UPPER_LEFT:
+ case PIPE_CAP_TGSI_FS_COORD_PIXEL_CENTER_HALF_INTEGER:
+ case PIPE_CAP_SM3:
+ case PIPE_CAP_SEAMLESS_CUBE_MAP:
+ case PIPE_CAP_PRIMITIVE_RESTART:
+ case PIPE_CAP_CONDITIONAL_RENDER:
+ case PIPE_CAP_TEXTURE_BARRIER:
+ case PIPE_CAP_INDEP_BLEND_ENABLE:
+ case PIPE_CAP_INDEP_BLEND_FUNC:
+ case PIPE_CAP_SEAMLESS_CUBE_MAP_PER_TEXTURE:
+ case PIPE_CAP_VERTEX_COLOR_UNCLAMPED:
+ return 1;
+
+ case PIPE_CAP_GLSL_FEATURE_LEVEL:
+ return debug_get_bool_option("R600_GLSL130", FALSE) ? 130 : 120;
+
+ /* Unsupported features. */
+ case PIPE_CAP_TGSI_INSTANCEID:
+ case PIPE_CAP_TGSI_FS_COORD_ORIGIN_LOWER_LEFT:
+ case PIPE_CAP_TGSI_FS_COORD_PIXEL_CENTER_INTEGER:
+ case PIPE_CAP_SCALED_RESOLVE:
+ case PIPE_CAP_TGSI_CAN_COMPACT_VARYINGS:
+ case PIPE_CAP_TGSI_CAN_COMPACT_CONSTANTS:
+ case PIPE_CAP_FRAGMENT_COLOR_CLAMPED:
+ case PIPE_CAP_VERTEX_COLOR_CLAMPED:
+ case PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION:
+ return 0;
+
+ /* Stream output. */
+ case PIPE_CAP_MAX_STREAM_OUTPUT_BUFFERS:
+ return debug_get_bool_option("R600_STREAMOUT", FALSE) ? 4 : 0;
+ case PIPE_CAP_STREAM_OUTPUT_PAUSE_RESUME:
+ return debug_get_bool_option("R600_STREAMOUT", FALSE) ? 1 : 0;
+ case PIPE_CAP_MAX_STREAM_OUTPUT_SEPARATE_COMPONENTS:
+ case PIPE_CAP_MAX_STREAM_OUTPUT_INTERLEAVED_COMPONENTS:
+ return 16*4;
+
+ /* Texturing. */
+ case PIPE_CAP_MAX_TEXTURE_2D_LEVELS:
+ case PIPE_CAP_MAX_TEXTURE_3D_LEVELS:
+ case PIPE_CAP_MAX_TEXTURE_CUBE_LEVELS:
+ return 15;
+ case PIPE_CAP_MAX_TEXTURE_ARRAY_LAYERS:
+ return rscreen->info.drm_minor >= 9 ? 16384 : 0;
+ case PIPE_CAP_MAX_COMBINED_SAMPLERS:
+ return 32;
+
+ /* Render targets. */
+ case PIPE_CAP_MAX_RENDER_TARGETS:
+ /* FIXME some r6xx are buggy and can only do 4 */
+ return 8;
+
+ /* Timer queries, present when the clock frequency is non zero. */
+ case PIPE_CAP_TIMER_QUERY:
+ return rscreen->info.r600_clock_crystal_freq != 0;
+
+ case PIPE_CAP_MIN_TEXEL_OFFSET:
+ return -8;
+
+ case PIPE_CAP_MAX_TEXEL_OFFSET:
+ return 7;
+ }
+ return 0;
+}
+
+static float r600_get_paramf(struct pipe_screen* pscreen,
+ enum pipe_capf param)
+{
+ struct r600_screen *rscreen = (struct r600_screen *)pscreen;
+ enum radeon_family family = rscreen->family;
+
+ switch (param) {
+ case PIPE_CAPF_MAX_LINE_WIDTH:
+ case PIPE_CAPF_MAX_LINE_WIDTH_AA:
+ case PIPE_CAPF_MAX_POINT_WIDTH:
+ case PIPE_CAPF_MAX_POINT_WIDTH_AA:
+ return 16384.0f;
+ case PIPE_CAPF_MAX_TEXTURE_ANISOTROPY:
+ return 16.0f;
+ case PIPE_CAPF_MAX_TEXTURE_LOD_BIAS:
+ return 16.0f;
+ case PIPE_CAPF_GUARD_BAND_LEFT:
+ case PIPE_CAPF_GUARD_BAND_TOP:
+ case PIPE_CAPF_GUARD_BAND_RIGHT:
+ case PIPE_CAPF_GUARD_BAND_BOTTOM:
+ return 0.0f;
+ }
+ return 0.0f;
+}
+
+static int r600_get_shader_param(struct pipe_screen* pscreen, unsigned shader, enum pipe_shader_cap param)
+{
+ struct r600_screen *rscreen = (struct r600_screen *)pscreen;
+ switch(shader)
+ {
+ case PIPE_SHADER_FRAGMENT:
+ case PIPE_SHADER_VERTEX:
+ break;
+ case PIPE_SHADER_GEOMETRY:
+ /* TODO: support and enable geometry programs */
+ return 0;
+ default:
+ /* TODO: support tessellation */
+ return 0;
+ }
+
+ /* TODO: all these should be fixed, since r600 surely supports much more! */
+ switch (param) {
+ case PIPE_SHADER_CAP_MAX_INSTRUCTIONS:
+ case PIPE_SHADER_CAP_MAX_ALU_INSTRUCTIONS:
+ case PIPE_SHADER_CAP_MAX_TEX_INSTRUCTIONS:
+ case PIPE_SHADER_CAP_MAX_TEX_INDIRECTIONS:
+ return 16384;
+ case PIPE_SHADER_CAP_MAX_CONTROL_FLOW_DEPTH:
+ return 8; /* FIXME */
+ case PIPE_SHADER_CAP_MAX_INPUTS:
+ if(shader == PIPE_SHADER_FRAGMENT)
+ return 34;
+ else
+ return 32;
+ case PIPE_SHADER_CAP_MAX_TEMPS:
+ return 256; /* Max native temporaries. */
+ case PIPE_SHADER_CAP_MAX_ADDRS:
+ /* FIXME Isn't this equal to TEMPS? */
+ return 1; /* Max native address registers */
+ case PIPE_SHADER_CAP_MAX_CONSTS:
+ return R600_MAX_CONST_BUFFER_SIZE;
+ case PIPE_SHADER_CAP_MAX_CONST_BUFFERS:
+ return R600_MAX_CONST_BUFFERS;
+ case PIPE_SHADER_CAP_MAX_PREDS:
+ return 0; /* FIXME */
+ case PIPE_SHADER_CAP_TGSI_CONT_SUPPORTED:
+ return 1;
+ case PIPE_SHADER_CAP_INDIRECT_INPUT_ADDR:
+ case PIPE_SHADER_CAP_INDIRECT_OUTPUT_ADDR:
+ case PIPE_SHADER_CAP_INDIRECT_TEMP_ADDR:
+ case PIPE_SHADER_CAP_INDIRECT_CONST_ADDR:
+ case PIPE_SHADER_CAP_INTEGERS:
+ return 0;
+ case PIPE_SHADER_CAP_SUBROUTINES:
+ return 0;
+ case PIPE_SHADER_CAP_MAX_TEXTURE_SAMPLERS:
+ return 16;
+ }
+ return 0;
+}
+
+static int r600_get_video_param(struct pipe_screen *screen,
+ enum pipe_video_profile profile,
+ enum pipe_video_cap param)
+{
+ switch (param) {
+ case PIPE_VIDEO_CAP_SUPPORTED:
+ return vl_profile_supported(screen, profile);
+ case PIPE_VIDEO_CAP_NPOT_TEXTURES:
+ return 1;
+ case PIPE_VIDEO_CAP_MAX_WIDTH:
+ case PIPE_VIDEO_CAP_MAX_HEIGHT:
+ return vl_video_buffer_max_size(screen);
+ case PIPE_VIDEO_CAP_PREFERED_FORMAT:
+ return PIPE_FORMAT_NV12;
+ default:
+ return 0;
+ }
+}
+
+static void r600_destroy_screen(struct pipe_screen* pscreen)
+{
+ struct r600_screen *rscreen = (struct r600_screen *)pscreen;
+
+ if (rscreen == NULL)
+ return;
+
+ if (rscreen->fences.bo) {
+ struct r600_fence_block *entry, *tmp;
+
+ LIST_FOR_EACH_ENTRY_SAFE(entry, tmp, &rscreen->fences.blocks, head) {
+ LIST_DEL(&entry->head);
+ FREE(entry);
+ }
+
+ rscreen->ws->buffer_unmap(rscreen->fences.bo->buf);
+ pipe_resource_reference((struct pipe_resource**)&rscreen->fences.bo, NULL);
+ }
+ pipe_mutex_destroy(rscreen->fences.mutex);
+
+ rscreen->ws->destroy(rscreen->ws);
+
+ util_slab_destroy(&rscreen->pool_buffers);
+ pipe_mutex_destroy(rscreen->mutex_num_contexts);
+ FREE(rscreen);
+}
+
+static void r600_fence_reference(struct pipe_screen *pscreen,
+ struct pipe_fence_handle **ptr,
+ struct pipe_fence_handle *fence)
+{
+ struct r600_fence **oldf = (struct r600_fence**)ptr;
+ struct r600_fence *newf = (struct r600_fence*)fence;
+
+ if (pipe_reference(&(*oldf)->reference, &newf->reference)) {
+ struct r600_screen *rscreen = (struct r600_screen *)pscreen;
+ pipe_mutex_lock(rscreen->fences.mutex);
+ pipe_resource_reference((struct pipe_resource**)&(*oldf)->sleep_bo, NULL);
+ LIST_ADDTAIL(&(*oldf)->head, &rscreen->fences.pool);
+ pipe_mutex_unlock(rscreen->fences.mutex);
+ }
+
+ *ptr = fence;
+}
+
+static boolean r600_fence_signalled(struct pipe_screen *pscreen,
+ struct pipe_fence_handle *fence)
+{
+ struct r600_screen *rscreen = (struct r600_screen *)pscreen;
+ struct r600_fence *rfence = (struct r600_fence*)fence;
+
+ return rscreen->fences.data[rfence->index];
+}
+
+static boolean r600_fence_finish(struct pipe_screen *pscreen,
+ struct pipe_fence_handle *fence,
+ uint64_t timeout)
+{
+ struct r600_screen *rscreen = (struct r600_screen *)pscreen;
+ struct r600_fence *rfence = (struct r600_fence*)fence;
+ int64_t start_time = 0;
+ unsigned spins = 0;
+
+ if (timeout != PIPE_TIMEOUT_INFINITE) {
+ start_time = os_time_get();
+
+ /* Convert to microseconds. */
+ timeout /= 1000;
+ }
+
+ while (rscreen->fences.data[rfence->index] == 0) {
+ /* Special-case infinite timeout - wait for the dummy BO to become idle */
+ if (timeout == PIPE_TIMEOUT_INFINITE) {
+ rscreen->ws->buffer_wait(rfence->sleep_bo->buf, RADEON_USAGE_READWRITE);
+ break;
+ }
+
+ /* The dummy BO will be busy until the CS including the fence has completed, or
+ * the GPU is reset. Don't bother continuing to spin when the BO is idle. */
+ if (!rscreen->ws->buffer_is_busy(rfence->sleep_bo->buf, RADEON_USAGE_READWRITE))
+ break;
+
+ if (++spins % 256)
+ continue;
+#ifdef PIPE_OS_UNIX
+ sched_yield();
+#else
+ os_time_sleep(10);
+#endif
+ if (timeout != PIPE_TIMEOUT_INFINITE &&
+ os_time_get() - start_time >= timeout) {
+ break;
+ }
+ }
+
+ return rscreen->fences.data[rfence->index] != 0;
+}
+
+static int evergreen_interpret_tiling(struct r600_screen *rscreen, uint32_t tiling_config)
+{
+ switch (tiling_config & 0xf) {
+ case 0:
+ rscreen->tiling_info.num_channels = 1;
+ break;
+ case 1:
+ rscreen->tiling_info.num_channels = 2;
+ break;
+ case 2:
+ rscreen->tiling_info.num_channels = 4;
+ break;
+ case 3:
+ rscreen->tiling_info.num_channels = 8;
+ break;
+ default:
+ return -EINVAL;
+ }
+
+ switch ((tiling_config & 0xf0) >> 4) {
+ case 0:
+ rscreen->tiling_info.num_banks = 4;
+ break;
+ case 1:
+ rscreen->tiling_info.num_banks = 8;
+ break;
+ case 2:
+ rscreen->tiling_info.num_banks = 16;
+ break;
+ default:
+ return -EINVAL;
+ }
+
+ switch ((tiling_config & 0xf00) >> 8) {
+ case 0:
+ rscreen->tiling_info.group_bytes = 256;
+ break;
+ case 1:
+ rscreen->tiling_info.group_bytes = 512;
+ break;
+ default:
+ return -EINVAL;
+ }
+ return 0;
+}
+
+static int r600_init_tiling(struct r600_screen *rscreen)
+{
+ uint32_t tiling_config = rscreen->info.r600_tiling_config;
+
+ /* set default group bytes, overridden by tiling info ioctl */
+ rscreen->tiling_info.group_bytes = 512;
+
+ if (!tiling_config)
+ return 0;
+
+ return evergreen_interpret_tiling(rscreen, tiling_config);
+}
+
+static unsigned radeon_family_from_device(unsigned device)
+{
+ switch (device) {
+#define CHIPSET(pciid, name, family) case pciid: return CHIP_##family;
+#include "pci_ids/radeonsi_pci_ids.h"
+#undef CHIPSET
+ default:
+ return CHIP_UNKNOWN;
+ }
+}
+
+struct pipe_screen *radeonsi_screen_create(struct radeon_winsys *ws)
+{
+ struct r600_screen *rscreen = CALLOC_STRUCT(r600_screen);
+ if (rscreen == NULL) {
+ return NULL;
+ }
+
+ rscreen->ws = ws;
+ ws->query_info(ws, &rscreen->info);
+
+ rscreen->family = radeon_family_from_device(rscreen->info.pci_id);
+ if (rscreen->family == CHIP_UNKNOWN) {
+ fprintf(stderr, "r600: Unknown chipset 0x%04X\n", rscreen->info.pci_id);
+ FREE(rscreen);
+ return NULL;
+ }
+
+ /* setup class */
+ if (rscreen->family >= CHIP_TAHITI) {
+ rscreen->chip_class = TAHITI;
+ } else {
+ fprintf(stderr, "r600: Unsupported family %d\n", rscreen->family);
+ FREE(rscreen);
+ return NULL;
+ }
+
+ if (r600_init_tiling(rscreen)) {
+ FREE(rscreen);
+ return NULL;
+ }
+
+ rscreen->screen.destroy = r600_destroy_screen;
+ rscreen->screen.get_name = r600_get_name;
+ rscreen->screen.get_vendor = r600_get_vendor;
+ rscreen->screen.get_param = r600_get_param;
+ rscreen->screen.get_shader_param = r600_get_shader_param;
+ rscreen->screen.get_paramf = r600_get_paramf;
+ rscreen->screen.get_video_param = r600_get_video_param;
+ rscreen->screen.is_format_supported = si_is_format_supported;
+ rscreen->screen.is_video_format_supported = vl_video_buffer_is_format_supported;
+ rscreen->screen.context_create = r600_create_context;
+ rscreen->screen.fence_reference = r600_fence_reference;
+ rscreen->screen.fence_signalled = r600_fence_signalled;
+ rscreen->screen.fence_finish = r600_fence_finish;
+ r600_init_screen_resource_functions(&rscreen->screen);
+
+ util_format_s3tc_init();
+
+ util_slab_create(&rscreen->pool_buffers,
+ sizeof(struct r600_resource), 64,
+ UTIL_SLAB_SINGLETHREADED);
+
+ pipe_mutex_init(rscreen->mutex_num_contexts);
+
+ rscreen->fences.bo = NULL;
+ rscreen->fences.data = NULL;
+ rscreen->fences.next_index = 0;
+ LIST_INITHEAD(&rscreen->fences.pool);
+ LIST_INITHEAD(&rscreen->fences.blocks);
+ pipe_mutex_init(rscreen->fences.mutex);
+
+ return &rscreen->screen;
+}
diff --git a/src/gallium/drivers/radeonsi/radeonsi_pipe.h b/src/gallium/drivers/radeonsi/radeonsi_pipe.h
new file mode 100644
index 00000000000..f4a1219d860
--- /dev/null
+++ b/src/gallium/drivers/radeonsi/radeonsi_pipe.h
@@ -0,0 +1,490 @@
+/*
+ * Copyright 2010 Jerome Glisse <[email protected]>
+ *
+ * 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.
+ *
+ * Authors:
+ * Jerome Glisse
+ */
+#ifndef RADEONSI_PIPE_H
+#define RADEONSI_PIPE_H
+
+#include "../../winsys/radeon/drm/radeon_winsys.h"
+
+#include "pipe/p_state.h"
+#include "pipe/p_screen.h"
+#include "pipe/p_context.h"
+#include "util/u_format.h"
+#include "util/u_math.h"
+#include "util/u_slab.h"
+#include "util/u_vbuf.h"
+#include "r600.h"
+#include "radeonsi_public.h"
+#include "r600_resource.h"
+
+#define R600_MAX_CONST_BUFFERS 1
+#define R600_MAX_CONST_BUFFER_SIZE 4096
+
+#ifdef PIPE_ARCH_BIG_ENDIAN
+#define R600_BIG_ENDIAN 1
+#else
+#define R600_BIG_ENDIAN 0
+#endif
+
+enum r600_atom_flags {
+ /* When set, atoms are added at the beginning of the dirty list
+ * instead of the end. */
+ EMIT_EARLY = (1 << 0)
+};
+
+/* This encapsulates a state or an operation which can emitted into the GPU
+ * command stream. It's not limited to states only, it can be used for anything
+ * that wants to write commands into the CS (e.g. cache flushes). */
+struct r600_atom {
+ void (*emit)(struct r600_context *ctx, struct r600_atom *state);
+
+ unsigned num_dw;
+ enum r600_atom_flags flags;
+ bool dirty;
+
+ struct list_head head;
+};
+
+struct r600_atom_surface_sync {
+ struct r600_atom atom;
+ unsigned flush_flags; /* CP_COHER_CNTL */
+};
+
+enum r600_pipe_state_id {
+ R600_PIPE_STATE_BLEND = 0,
+ R600_PIPE_STATE_BLEND_COLOR,
+ R600_PIPE_STATE_CONFIG,
+ R600_PIPE_STATE_SEAMLESS_CUBEMAP,
+ R600_PIPE_STATE_CLIP,
+ R600_PIPE_STATE_SCISSOR,
+ R600_PIPE_STATE_VIEWPORT,
+ R600_PIPE_STATE_RASTERIZER,
+ R600_PIPE_STATE_VGT,
+ R600_PIPE_STATE_FRAMEBUFFER,
+ R600_PIPE_STATE_DSA,
+ R600_PIPE_STATE_STENCIL_REF,
+ R600_PIPE_STATE_PS_SHADER,
+ R600_PIPE_STATE_VS_SHADER,
+ R600_PIPE_STATE_CONSTANT,
+ R600_PIPE_STATE_SAMPLER,
+ R600_PIPE_STATE_RESOURCE,
+ R600_PIPE_STATE_POLYGON_OFFSET,
+ R600_PIPE_NSTATES
+};
+
+struct r600_pipe_fences {
+ struct r600_resource *bo;
+ unsigned *data;
+ unsigned next_index;
+ /* linked list of preallocated blocks */
+ struct list_head blocks;
+ /* linked list of freed fences */
+ struct list_head pool;
+ pipe_mutex mutex;
+};
+
+struct r600_screen {
+ struct pipe_screen screen;
+ struct radeon_winsys *ws;
+ unsigned family;
+ enum chip_class chip_class;
+ struct radeon_info info;
+ struct r600_tiling_info tiling_info;
+ struct util_slab_mempool pool_buffers;
+ struct r600_pipe_fences fences;
+
+ unsigned num_contexts;
+
+ /* for thread-safe write accessing to num_contexts */
+ pipe_mutex mutex_num_contexts;
+};
+
+struct si_pipe_sampler_view {
+ struct pipe_sampler_view base;
+ uint32_t state[8];
+};
+
+struct si_pipe_sampler_state {
+ uint32_t val[4];
+};
+
+struct r600_pipe_rasterizer {
+ struct r600_pipe_state rstate;
+ boolean flatshade;
+ unsigned sprite_coord_enable;
+ unsigned pa_sc_line_stipple;
+ unsigned pa_su_sc_mode_cntl;
+ unsigned pa_cl_clip_cntl;
+ unsigned pa_cl_vs_out_cntl;
+ float offset_units;
+ float offset_scale;
+};
+
+struct r600_pipe_blend {
+ struct r600_pipe_state rstate;
+ unsigned cb_target_mask;
+ unsigned cb_color_control;
+};
+
+struct r600_pipe_dsa {
+ struct r600_pipe_state rstate;
+ unsigned alpha_ref;
+ unsigned db_render_override;
+ unsigned db_render_control;
+ ubyte valuemask[2];
+ ubyte writemask[2];
+};
+
+struct r600_vertex_element
+{
+ unsigned count;
+ struct pipe_vertex_element elements[PIPE_MAX_ATTRIBS];
+ struct u_vbuf_elements *vmgr_elements;
+ unsigned fs_size;
+ struct r600_pipe_state rstate;
+ /* if offset is to big for fetch instructio we need to alterate
+ * offset of vertex buffer, record here the offset need to add
+ */
+ unsigned vbuffer_need_offset;
+ unsigned vbuffer_offset[PIPE_MAX_ATTRIBS];
+};
+
+struct r600_shader_io {
+ unsigned name;
+ unsigned gpr;
+ unsigned done;
+ int sid;
+ unsigned interpolate;
+ boolean centroid;
+ unsigned lds_pos; /* for evergreen */
+};
+
+struct r600_shader {
+ unsigned ninput;
+ unsigned noutput;
+ struct r600_shader_io input[32];
+ struct r600_shader_io output[32];
+ boolean uses_kill;
+ boolean fs_write_all;
+ unsigned nr_cbufs;
+};
+
+struct si_pipe_shader {
+ struct r600_shader shader;
+ struct r600_pipe_state rstate;
+ struct r600_resource *bo;
+ struct r600_vertex_element vertex_elements;
+ struct tgsi_token *tokens;
+ unsigned num_sgprs;
+ unsigned num_vgprs;
+ unsigned spi_ps_input_ena;
+ unsigned sprite_coord_enable;
+ struct pipe_stream_output_info so;
+ unsigned so_strides[4];
+};
+
+/* needed for blitter save */
+#define NUM_TEX_UNITS 16
+
+struct r600_textures_info {
+ struct r600_pipe_state rstate;
+ struct si_pipe_sampler_view *views[NUM_TEX_UNITS];
+ struct si_pipe_sampler_state *samplers[NUM_TEX_UNITS];
+ unsigned n_views;
+ unsigned n_samplers;
+ bool samplers_dirty;
+ bool is_array_sampler[NUM_TEX_UNITS];
+};
+
+struct r600_fence {
+ struct pipe_reference reference;
+ unsigned index; /* in the shared bo */
+ struct r600_resource *sleep_bo;
+ struct list_head head;
+};
+
+#define FENCE_BLOCK_SIZE 16
+
+struct r600_fence_block {
+ struct r600_fence fences[FENCE_BLOCK_SIZE];
+ struct list_head head;
+};
+
+#define R600_CONSTANT_ARRAY_SIZE 256
+#define R600_RESOURCE_ARRAY_SIZE 160
+
+struct r600_stencil_ref
+{
+ ubyte ref_value[2];
+ ubyte valuemask[2];
+ ubyte writemask[2];
+};
+
+struct r600_context {
+ struct pipe_context context;
+ struct blitter_context *blitter;
+ enum radeon_family family;
+ enum chip_class chip_class;
+ void *custom_dsa_flush;
+ struct r600_screen *screen;
+ struct radeon_winsys *ws;
+ struct r600_pipe_state *states[R600_PIPE_NSTATES];
+ struct r600_vertex_element *vertex_elements;
+ struct pipe_framebuffer_state framebuffer;
+ unsigned cb_target_mask;
+ unsigned cb_color_control;
+ unsigned pa_sc_line_stipple;
+ unsigned pa_su_sc_mode_cntl;
+ unsigned pa_cl_clip_cntl;
+ unsigned pa_cl_vs_out_cntl;
+ /* for saving when using blitter */
+ struct pipe_stencil_ref stencil_ref;
+ struct pipe_viewport_state viewport;
+ struct pipe_clip_state clip;
+ struct r600_pipe_state config;
+ struct si_pipe_shader *ps_shader;
+ struct si_pipe_shader *vs_shader;
+ struct r600_pipe_state vs_const_buffer;
+ struct r600_pipe_state vs_user_data;
+ struct r600_pipe_state ps_const_buffer;
+ struct r600_pipe_rasterizer *rasterizer;
+ struct r600_pipe_state vgt;
+ struct r600_pipe_state spi;
+ struct pipe_query *current_render_cond;
+ unsigned current_render_cond_mode;
+ struct pipe_query *saved_render_cond;
+ unsigned saved_render_cond_mode;
+ /* shader information */
+ unsigned sprite_coord_enable;
+ boolean export_16bpc;
+ unsigned alpha_ref;
+ boolean alpha_ref_dirty;
+ unsigned nr_cbufs;
+ struct r600_textures_info vs_samplers;
+ struct r600_textures_info ps_samplers;
+ boolean shader_dirty;
+
+ struct u_vbuf *vbuf_mgr;
+ struct util_slab_mempool pool_transfers;
+ boolean have_depth_texture, have_depth_fb;
+
+ unsigned default_ps_gprs, default_vs_gprs;
+
+ /* States based on r600_state. */
+ struct list_head dirty_states;
+ struct r600_atom_surface_sync atom_surface_sync;
+ struct r600_atom atom_r6xx_flush_and_inv;
+
+ /* Below are variables from the old r600_context.
+ */
+ struct radeon_winsys_cs *cs;
+
+ struct r600_range *range;
+ unsigned nblocks;
+ struct r600_block **blocks;
+ struct list_head dirty;
+ struct list_head enable_list;
+ unsigned pm4_dirty_cdwords;
+ unsigned ctx_pm4_ndwords;
+ unsigned init_dwords;
+
+ /* The list of active queries. Only one query of each type can be active. */
+ struct list_head active_query_list;
+ unsigned num_cs_dw_queries_suspend;
+ unsigned num_cs_dw_streamout_end;
+
+ unsigned backend_mask;
+ unsigned max_db; /* for OQ */
+ unsigned flags;
+ boolean predicate_drawing;
+
+ unsigned num_so_targets;
+ struct r600_so_target *so_targets[PIPE_MAX_SO_BUFFERS];
+ boolean streamout_start;
+ unsigned streamout_append_bitmask;
+ unsigned *vs_so_stride_in_dw;
+ unsigned *vs_shader_so_strides;
+};
+
+static INLINE void r600_emit_atom(struct r600_context *rctx, struct r600_atom *atom)
+{
+ atom->emit(rctx, atom);
+ atom->dirty = false;
+ if (atom->head.next && atom->head.prev)
+ LIST_DELINIT(&atom->head);
+}
+
+static INLINE void r600_atom_dirty(struct r600_context *rctx, struct r600_atom *state)
+{
+ if (!state->dirty) {
+ if (state->flags & EMIT_EARLY) {
+ LIST_ADD(&state->head, &rctx->dirty_states);
+ } else {
+ LIST_ADDTAIL(&state->head, &rctx->dirty_states);
+ }
+ state->dirty = true;
+ }
+}
+
+/* evergreen_state.c */
+void cayman_init_state_functions(struct r600_context *rctx);
+void si_init_config(struct r600_context *rctx);
+void si_pipe_shader_ps(struct pipe_context *ctx, struct si_pipe_shader *shader);
+void si_pipe_shader_vs(struct pipe_context *ctx, struct si_pipe_shader *shader);
+void si_update_spi_map(struct r600_context *rctx);
+void *cayman_create_db_flush_dsa(struct r600_context *rctx);
+void cayman_polygon_offset_update(struct r600_context *rctx);
+uint32_t si_translate_vertexformat(struct pipe_screen *screen,
+ enum pipe_format format,
+ const struct util_format_description *desc,
+ int first_non_void);
+boolean si_is_format_supported(struct pipe_screen *screen,
+ enum pipe_format format,
+ enum pipe_texture_target target,
+ unsigned sample_count,
+ unsigned usage);
+
+/* r600_blit.c */
+void r600_init_blit_functions(struct r600_context *rctx);
+void r600_blit_uncompress_depth(struct pipe_context *ctx, struct r600_resource_texture *texture);
+void r600_blit_push_depth(struct pipe_context *ctx, struct r600_resource_texture *texture);
+void r600_flush_depth_textures(struct r600_context *rctx);
+
+/* r600_buffer.c */
+bool r600_init_resource(struct r600_screen *rscreen,
+ struct r600_resource *res,
+ unsigned size, unsigned alignment,
+ unsigned bind, unsigned usage);
+struct pipe_resource *r600_buffer_create(struct pipe_screen *screen,
+ const struct pipe_resource *templ);
+struct pipe_resource *r600_user_buffer_create(struct pipe_screen *screen,
+ void *ptr, unsigned bytes,
+ unsigned bind);
+void r600_upload_index_buffer(struct r600_context *rctx,
+ struct pipe_index_buffer *ib, unsigned count);
+
+
+/* r600_pipe.c */
+void radeonsi_flush(struct pipe_context *ctx, struct pipe_fence_handle **fence,
+ unsigned flags);
+
+/* r600_query.c */
+void r600_init_query_functions(struct r600_context *rctx);
+
+/* r600_resource.c */
+void r600_init_context_resource_functions(struct r600_context *r600);
+
+/* radeonsi_shader.c */
+int si_pipe_shader_create(struct pipe_context *ctx, struct si_pipe_shader *shader);
+void si_pipe_shader_destroy(struct pipe_context *ctx, struct si_pipe_shader *shader);
+
+/* r600_texture.c */
+void r600_init_screen_texture_functions(struct pipe_screen *screen);
+void r600_init_surface_functions(struct r600_context *r600);
+unsigned r600_texture_get_offset(struct r600_resource_texture *rtex,
+ unsigned level, unsigned layer);
+
+/* r600_translate.c */
+void r600_translate_index_buffer(struct r600_context *r600,
+ struct pipe_index_buffer *ib,
+ unsigned count);
+
+/* r600_state_common.c */
+void r600_init_common_atoms(struct r600_context *rctx);
+unsigned r600_get_cb_flush_flags(struct r600_context *rctx);
+void r600_texture_barrier(struct pipe_context *ctx);
+void r600_set_index_buffer(struct pipe_context *ctx,
+ const struct pipe_index_buffer *ib);
+void r600_set_vertex_buffers(struct pipe_context *ctx, unsigned count,
+ const struct pipe_vertex_buffer *buffers);
+void *si_create_vertex_elements(struct pipe_context *ctx,
+ unsigned count,
+ const struct pipe_vertex_element *elements);
+void r600_delete_vertex_element(struct pipe_context *ctx, void *state);
+void r600_bind_blend_state(struct pipe_context *ctx, void *state);
+void r600_bind_dsa_state(struct pipe_context *ctx, void *state);
+void r600_bind_rs_state(struct pipe_context *ctx, void *state);
+void r600_delete_rs_state(struct pipe_context *ctx, void *state);
+void r600_sampler_view_destroy(struct pipe_context *ctx,
+ struct pipe_sampler_view *state);
+void r600_delete_state(struct pipe_context *ctx, void *state);
+void r600_bind_vertex_elements(struct pipe_context *ctx, void *state);
+void *si_create_shader_state(struct pipe_context *ctx,
+ const struct pipe_shader_state *state);
+void r600_bind_ps_shader(struct pipe_context *ctx, void *state);
+void r600_bind_vs_shader(struct pipe_context *ctx, void *state);
+void r600_delete_ps_shader(struct pipe_context *ctx, void *state);
+void r600_delete_vs_shader(struct pipe_context *ctx, void *state);
+void r600_set_constant_buffer(struct pipe_context *ctx, uint shader, uint index,
+ struct pipe_resource *buffer);
+struct pipe_stream_output_target *
+r600_create_so_target(struct pipe_context *ctx,
+ struct pipe_resource *buffer,
+ unsigned buffer_offset,
+ unsigned buffer_size);
+void r600_so_target_destroy(struct pipe_context *ctx,
+ struct pipe_stream_output_target *target);
+void r600_set_so_targets(struct pipe_context *ctx,
+ unsigned num_targets,
+ struct pipe_stream_output_target **targets,
+ unsigned append_bitmask);
+void r600_set_pipe_stencil_ref(struct pipe_context *ctx,
+ const struct pipe_stencil_ref *state);
+void r600_draw_vbo(struct pipe_context *ctx, const struct pipe_draw_info *info);
+
+/*
+ * common helpers
+ */
+static INLINE uint32_t S_FIXED(float value, uint32_t frac_bits)
+{
+ return value * (1 << frac_bits);
+}
+#define ALIGN_DIVUP(x, y) (((x) + (y) - 1) / (y))
+
+static inline unsigned r600_tex_aniso_filter(unsigned filter)
+{
+ if (filter <= 1) return 0;
+ if (filter <= 2) return 1;
+ if (filter <= 4) return 2;
+ if (filter <= 8) return 3;
+ /* else */ return 4;
+}
+
+/* 12.4 fixed-point */
+static INLINE unsigned r600_pack_float_12p4(float x)
+{
+ return x <= 0 ? 0 :
+ x >= 4096 ? 0xffff : x * 16;
+}
+
+static INLINE uint64_t r600_resource_va(struct pipe_screen *screen, struct pipe_resource *resource)
+{
+ struct r600_screen *rscreen = (struct r600_screen*)screen;
+ struct r600_resource *rresource = (struct r600_resource*)resource;
+
+ return rscreen->ws->buffer_get_virtual_address(rresource->cs_buf);
+}
+
+#endif
diff --git a/src/gallium/drivers/radeonsi/radeonsi_public.h b/src/gallium/drivers/radeonsi/radeonsi_public.h
new file mode 100644
index 00000000000..5dcec0fc93b
--- /dev/null
+++ b/src/gallium/drivers/radeonsi/radeonsi_public.h
@@ -0,0 +1,30 @@
+/*
+ * Copyright 2010 Jerome Glisse <[email protected]>
+ *
+ * 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.
+ */
+#ifndef RADEONSI_PUBLIC_H
+#define RADEONSI_PUBLIC_H
+
+struct radeon_winsys;
+
+struct pipe_screen *radeonsi_screen_create(struct radeon_winsys *ws);
+
+#endif
diff --git a/src/gallium/drivers/radeonsi/radeonsi_shader.c b/src/gallium/drivers/radeonsi/radeonsi_shader.c
new file mode 100644
index 00000000000..50f2e39314f
--- /dev/null
+++ b/src/gallium/drivers/radeonsi/radeonsi_shader.c
@@ -0,0 +1,565 @@
+
+#include "gallivm/lp_bld_tgsi_action.h"
+#include "gallivm/lp_bld_const.h"
+#include "gallivm/lp_bld_intr.h"
+#include "gallivm/lp_bld_tgsi.h"
+#include "radeon_llvm.h"
+#include "tgsi/tgsi_info.h"
+#include "tgsi/tgsi_parse.h"
+#include "tgsi/tgsi_scan.h"
+#include "tgsi/tgsi_dump.h"
+
+#include "radeonsi_pipe.h"
+#include "radeonsi_shader.h"
+#include "sid.h"
+
+#include <assert.h>
+#include <errno.h>
+#include <stdio.h>
+
+/*
+static ps_remap_inputs(
+ struct tgsi_llvm_context * tl_ctx,
+ unsigned tgsi_index,
+ unsigned tgsi_chan)
+{
+ :
+}
+
+struct si_input
+{
+ struct list_head head;
+ unsigned tgsi_index;
+ unsigned tgsi_chan;
+ unsigned order;
+};
+*/
+
+
+struct si_shader_context
+{
+ struct radeon_llvm_context radeon_bld;
+ struct r600_context *rctx;
+ struct tgsi_parse_context parse;
+ struct tgsi_token * tokens;
+ struct si_pipe_shader *shader;
+ unsigned type; /* TGSI_PROCESSOR_* specifies the type of shader. */
+/* unsigned num_inputs; */
+/* struct list_head inputs; */
+/* unsigned * input_mappings *//* From TGSI to SI hw */
+/* struct tgsi_shader_info info;*/
+};
+
+static struct si_shader_context * si_shader_context(
+ struct lp_build_tgsi_context * bld_base)
+{
+ return (struct si_shader_context *)bld_base;
+}
+
+
+#define PERSPECTIVE_BASE 0
+#define LINEAR_BASE 9
+
+#define SAMPLE_OFFSET 0
+#define CENTER_OFFSET 2
+#define CENTROID_OFSET 4
+
+#define USE_SGPR_MAX_SUFFIX_LEN 5
+
+enum sgpr_type {
+ SGPR_I32,
+ SGPR_I64,
+ SGPR_PTR_V4I32,
+ SGPR_PTR_V8I32
+};
+
+static LLVMValueRef use_sgpr(
+ struct gallivm_state * gallivm,
+ enum sgpr_type type,
+ unsigned sgpr)
+{
+ LLVMValueRef sgpr_index;
+ LLVMValueRef sgpr_value;
+ LLVMTypeRef ret_type;
+
+ sgpr_index = lp_build_const_int32(gallivm, sgpr);
+
+ if (type == SGPR_I32) {
+ ret_type = LLVMInt32TypeInContext(gallivm->context);
+ return lp_build_intrinsic_unary(gallivm->builder,
+ "llvm.SI.use.sgpr.i32",
+ ret_type, sgpr_index);
+ }
+
+ ret_type = LLVMInt64TypeInContext(gallivm->context);
+ sgpr_value = lp_build_intrinsic_unary(gallivm->builder,
+ "llvm.SI.use.sgpr.i64",
+ ret_type, sgpr_index);
+
+ switch (type) {
+ case SGPR_I64:
+ return sgpr_value;
+ case SGPR_PTR_V4I32:
+ ret_type = LLVMInt32TypeInContext(gallivm->context);
+ ret_type = LLVMVectorType(ret_type, 4);
+ ret_type = LLVMPointerType(ret_type,
+ 0 /*XXX: Specify address space*/);
+ return LLVMBuildIntToPtr(gallivm->builder, sgpr_value,
+ ret_type, "");
+ case SGPR_PTR_V8I32:
+ ret_type = LLVMInt32TypeInContext(gallivm->context);
+ ret_type = LLVMVectorType(ret_type, 8);
+ ret_type = LLVMPointerType(ret_type,
+ 0 /*XXX: Specify address space*/);
+ return LLVMBuildIntToPtr(gallivm->builder, sgpr_value,
+ ret_type, "");
+ default:
+ assert(!"Unsupported SGPR type in use_sgpr()");
+ return NULL;
+ }
+}
+
+static void declare_input_vs(
+ struct si_shader_context * si_shader_ctx,
+ unsigned input_index,
+ const struct tgsi_full_declaration *decl)
+{
+ LLVMValueRef t_list_ptr;
+ LLVMValueRef t_offset;
+ LLVMValueRef attribute_offset;
+ LLVMValueRef buffer_index_reg;
+ LLVMValueRef args[4];
+ LLVMTypeRef vec4_type;
+ LLVMValueRef input;
+ struct lp_build_context * uint = &si_shader_ctx->radeon_bld.soa.bld_base.uint_bld;
+ struct lp_build_context * base = &si_shader_ctx->radeon_bld.soa.bld_base.base;
+ struct r600_context *rctx = si_shader_ctx->rctx;
+ struct pipe_vertex_element *velem = &rctx->vertex_elements->elements[input_index];
+ unsigned chan;
+
+ /* XXX: Communicate with the rest of the driver about which SGPR the T#
+ * list pointer is going to be stored in. Hard code to SGPR[0-1] for
+ * now */
+ t_list_ptr = use_sgpr(base->gallivm, SGPR_I64, 0);
+
+ t_offset = lp_build_const_int32(base->gallivm,
+ 4 * velem->vertex_buffer_index);
+ attribute_offset = lp_build_const_int32(base->gallivm, velem->src_offset);
+
+ /* Load the buffer index is always, which is always stored in VGPR0
+ * for Vertex Shaders */
+ buffer_index_reg = lp_build_intrinsic(base->gallivm->builder,
+ "llvm.SI.vs.load.buffer.index", uint->elem_type, NULL, 0);
+
+ vec4_type = LLVMVectorType(base->elem_type, 4);
+ args[0] = t_list_ptr;
+ args[1] = t_offset;
+ args[2] = attribute_offset;
+ args[3] = buffer_index_reg;
+ input = lp_build_intrinsic(base->gallivm->builder,
+ "llvm.SI.vs.load.input", vec4_type, args, 4);
+
+ /* Break up the vec4 into individual components */
+ for (chan = 0; chan < 4; chan++) {
+ LLVMValueRef llvm_chan = lp_build_const_int32(base->gallivm, chan);
+ /* XXX: Use a helper function for this. There is one in
+ * tgsi_llvm.c. */
+ si_shader_ctx->radeon_bld.inputs[radeon_llvm_reg_index_soa(input_index, chan)] =
+ LLVMBuildExtractElement(base->gallivm->builder,
+ input, llvm_chan, "");
+ }
+}
+
+static void declare_input_fs(
+ struct si_shader_context * si_shader_ctx,
+ unsigned input_index,
+ const struct tgsi_full_declaration *decl)
+{
+ const char * intr_name;
+ unsigned chan;
+ struct lp_build_context * base =
+ &si_shader_ctx->radeon_bld.soa.bld_base.base;
+ struct gallivm_state * gallivm = base->gallivm;
+
+ /* This value is:
+ * [15:0] NewPrimMask (Bit mask for each quad. It is set it the
+ * quad begins a new primitive. Bit 0 always needs
+ * to be unset)
+ * [32:16] ParamOffset
+ *
+ */
+ LLVMValueRef params = use_sgpr(base->gallivm, SGPR_I32, 6);
+
+
+ /* XXX: Is this the input_index? */
+ LLVMValueRef attr_number = lp_build_const_int32(gallivm, input_index);
+
+ /* XXX: Handle all possible interpolation modes */
+ switch (decl->Declaration.Interpolate) {
+ case TGSI_INTERPOLATE_COLOR:
+ if (si_shader_ctx->rctx->rasterizer->flatshade)
+ intr_name = "llvm.SI.fs.interp.constant";
+ else
+ intr_name = "llvm.SI.fs.interp.linear.center";
+ break;
+ case TGSI_INTERPOLATE_CONSTANT:
+ intr_name = "llvm.SI.fs.interp.constant";
+ break;
+ case TGSI_INTERPOLATE_LINEAR:
+ intr_name = "llvm.SI.fs.interp.linear.center";
+ break;
+ default:
+ fprintf(stderr, "Warning: Unhandled interpolation mode.\n");
+ return;
+ }
+
+ /* XXX: Could there be more than TGSI_NUM_CHANNELS (4) ? */
+ for (chan = 0; chan < TGSI_NUM_CHANNELS; chan++) {
+ LLVMValueRef args[3];
+ LLVMValueRef llvm_chan = lp_build_const_int32(gallivm, chan);
+ unsigned soa_index = radeon_llvm_reg_index_soa(input_index, chan);
+ LLVMTypeRef input_type = LLVMFloatTypeInContext(gallivm->context);
+ args[0] = llvm_chan;
+ args[1] = attr_number;
+ args[2] = params;
+ si_shader_ctx->radeon_bld.inputs[soa_index] =
+ lp_build_intrinsic(gallivm->builder, intr_name,
+ input_type, args, 3);
+ }
+}
+
+static void declare_input(
+ struct radeon_llvm_context * radeon_bld,
+ unsigned input_index,
+ const struct tgsi_full_declaration *decl)
+{
+ struct si_shader_context * si_shader_ctx =
+ si_shader_context(&radeon_bld->soa.bld_base);
+ if (si_shader_ctx->type == TGSI_PROCESSOR_VERTEX) {
+ declare_input_vs(si_shader_ctx, input_index, decl);
+ } else if (si_shader_ctx->type == TGSI_PROCESSOR_FRAGMENT) {
+ declare_input_fs(si_shader_ctx, input_index, decl);
+ } else {
+ fprintf(stderr, "Warning: Unsupported shader type,\n");
+ }
+}
+
+static LLVMValueRef fetch_constant(
+ struct lp_build_tgsi_context * bld_base,
+ const struct tgsi_full_src_register *reg,
+ enum tgsi_opcode_type type,
+ unsigned swizzle)
+{
+ struct lp_build_context * base = &bld_base->base;
+
+ LLVMValueRef const_ptr;
+ LLVMValueRef offset;
+
+ /* XXX: Assume the pointer to the constant buffer is being stored in
+ * SGPR[2:3] */
+ const_ptr = use_sgpr(base->gallivm, SGPR_I64, 1);
+
+ /* XXX: This assumes that the constant buffer is not packed, so
+ * CONST[0].x will have an offset of 0 and CONST[1].x will have an
+ * offset of 4. */
+ offset = lp_build_const_int32(base->gallivm,
+ (reg->Register.Index * 4) + swizzle);
+
+ return lp_build_intrinsic_binary(base->gallivm->builder,
+ "llvm.SI.load.const", base->elem_type, const_ptr, offset);
+}
+
+
+/* Declare some intrinsics with the correct attributes */
+static void si_llvm_emit_prologue(struct lp_build_tgsi_context * bld_base)
+{
+ LLVMValueRef function;
+ struct gallivm_state * gallivm = bld_base->base.gallivm;
+
+ LLVMTypeRef i64 = LLVMInt64TypeInContext(gallivm->context);
+ LLVMTypeRef i32 = LLVMInt32TypeInContext(gallivm->context);
+
+ /* declare i32 @llvm.SI.use.sgpr.i32(i32) */
+ function = lp_declare_intrinsic(gallivm->module, "llvm.SI.use.sgpr.i32",
+ i32, &i32, 1);
+ LLVMAddFunctionAttr(function, LLVMReadNoneAttribute);
+
+ /* declare i64 @llvm.SI.use.sgpr.i64(i32) */
+ function = lp_declare_intrinsic(gallivm->module, "llvm.SI.use.sgpr.i64",
+ i64, &i32, 1);
+ LLVMAddFunctionAttr(function, LLVMReadNoneAttribute);
+}
+
+/* XXX: This is partially implemented for VS only at this point. It is not complete */
+static void si_llvm_emit_epilogue(struct lp_build_tgsi_context * bld_base)
+{
+ struct si_shader_context * si_shader_ctx = si_shader_context(bld_base);
+ struct r600_shader * shader = &si_shader_ctx->shader->shader;
+ struct lp_build_context * base = &bld_base->base;
+ struct lp_build_context * uint =
+ &si_shader_ctx->radeon_bld.soa.bld_base.uint_bld;
+ struct tgsi_parse_context *parse = &si_shader_ctx->parse;
+ LLVMValueRef last_args[9] = { 0 };
+
+ while (!tgsi_parse_end_of_tokens(parse)) {
+ /* XXX: component_bits controls which components of the output
+ * registers actually get exported. (e.g bit 0 means export
+ * X component, bit 1 means export Y component, etc.) I'm
+ * hard coding this to 0xf for now. In the future, we might
+ * want to do something else. */
+ unsigned component_bits = 0xf;
+ unsigned chan;
+ struct tgsi_full_declaration *d =
+ &parse->FullToken.FullDeclaration;
+ LLVMValueRef args[9];
+ unsigned target;
+ unsigned index;
+ unsigned color_count = 0;
+ unsigned param_count = 0;
+ int i;
+
+ tgsi_parse_token(parse);
+ if (parse->FullToken.Token.Type != TGSI_TOKEN_TYPE_DECLARATION)
+ continue;
+
+ switch (d->Declaration.File) {
+ case TGSI_FILE_INPUT:
+ i = shader->ninput++;
+ shader->input[i].name = d->Semantic.Name;
+ shader->input[i].sid = d->Semantic.Index;
+ shader->input[i].interpolate = d->Declaration.Interpolate;
+ shader->input[i].centroid = d->Declaration.Centroid;
+ break;
+ case TGSI_FILE_OUTPUT:
+ i = shader->noutput++;
+ shader->output[i].name = d->Semantic.Name;
+ shader->output[i].sid = d->Semantic.Index;
+ shader->output[i].interpolate = d->Declaration.Interpolate;
+ break;
+ }
+
+ if (d->Declaration.File != TGSI_FILE_OUTPUT)
+ continue;
+
+ for (index = d->Range.First; index <= d->Range.Last; index++) {
+ for (chan = 0; chan < 4; chan++ ) {
+ LLVMValueRef out_ptr =
+ si_shader_ctx->radeon_bld.soa.outputs
+ [index][chan];
+ /* +5 because the first output value will be
+ * the 6th argument to the intrinsic. */
+ args[chan + 5]= LLVMBuildLoad(
+ base->gallivm->builder, out_ptr, "");
+ }
+
+ /* XXX: We probably need to keep track of the output
+ * values, so we know what we are passing to the next
+ * stage. */
+
+ /* Select the correct target */
+ switch(d->Semantic.Name) {
+ case TGSI_SEMANTIC_POSITION:
+ target = V_008DFC_SQ_EXP_POS;
+ break;
+ case TGSI_SEMANTIC_COLOR:
+ if (si_shader_ctx->type == TGSI_PROCESSOR_VERTEX) {
+ target = V_008DFC_SQ_EXP_PARAM + param_count;
+ param_count++;
+ } else {
+ target = V_008DFC_SQ_EXP_MRT + color_count;
+ color_count++;
+ }
+ break;
+ case TGSI_SEMANTIC_GENERIC:
+ target = V_008DFC_SQ_EXP_PARAM + param_count;
+ param_count++;
+ break;
+ default:
+ target = 0;
+ fprintf(stderr,
+ "Warning: SI unhandled output type:%d\n",
+ d->Semantic.Name);
+ }
+
+ /* Specify which components to enable */
+ args[0] = lp_build_const_int32(base->gallivm,
+ component_bits);
+
+ /* Specify whether the EXEC mask represents the valid mask */
+ args[1] = lp_build_const_int32(base->gallivm, 0);
+
+ /* Specify whether this is the last export */
+ args[2] = lp_build_const_int32(base->gallivm, 0);
+
+ /* Specify the target we are exporting */
+ args[3] = lp_build_const_int32(base->gallivm, target);
+
+ /* Set COMPR flag to zero to export data as 32-bit */
+ args[4] = uint->zero;
+
+ if (si_shader_ctx->type == TGSI_PROCESSOR_VERTEX ?
+ (d->Semantic.Name == TGSI_SEMANTIC_POSITION) :
+ (d->Semantic.Name == TGSI_SEMANTIC_COLOR)) {
+ if (last_args[0]) {
+ lp_build_intrinsic(base->gallivm->builder,
+ "llvm.SI.export",
+ LLVMVoidTypeInContext(base->gallivm->context),
+ last_args, 9);
+ }
+
+ memcpy(last_args, args, sizeof(args));
+ } else {
+ lp_build_intrinsic(base->gallivm->builder,
+ "llvm.SI.export",
+ LLVMVoidTypeInContext(base->gallivm->context),
+ args, 9);
+ }
+
+ }
+ }
+
+ /* Specify whether the EXEC mask represents the valid mask */
+ last_args[1] = lp_build_const_int32(base->gallivm,
+ si_shader_ctx->type == TGSI_PROCESSOR_FRAGMENT);
+
+ /* Specify that this is the last export */
+ last_args[2] = lp_build_const_int32(base->gallivm, 1);
+
+ lp_build_intrinsic(base->gallivm->builder,
+ "llvm.SI.export",
+ LLVMVoidTypeInContext(base->gallivm->context),
+ last_args, 9);
+
+/* XXX: Look up what this function does */
+/* ctx->shader->output[i].spi_sid = r600_spi_sid(&ctx->shader->output[i]);*/
+}
+
+static void tex_fetch_args(
+ struct lp_build_tgsi_context * bld_base,
+ struct lp_build_emit_data * emit_data)
+{
+ /* WriteMask */
+ emit_data->args[0] = lp_build_const_int32(bld_base->base.gallivm,
+ emit_data->inst->Dst[0].Register.WriteMask);
+
+ /* Coordinates */
+ /* XXX: Not all sample instructions need 4 address arguments. */
+ emit_data->args[1] = lp_build_emit_fetch(bld_base, emit_data->inst,
+ 0, LP_CHAN_ALL);
+
+ /* Resource */
+ emit_data->args[2] = use_sgpr(bld_base->base.gallivm, SGPR_I64, 2);
+ emit_data->args[3] = lp_build_const_int32(bld_base->base.gallivm,
+ 32 * emit_data->inst->Src[2].Register.Index);
+
+ /* Sampler */
+ emit_data->args[4] = use_sgpr(bld_base->base.gallivm, SGPR_I64, 1);
+ emit_data->args[5] = lp_build_const_int32(bld_base->base.gallivm,
+ 16 * emit_data->inst->Src[2].Register.Index);
+
+ /* Dimensions */
+ /* XXX: We might want to pass this information to the shader at some. */
+/* emit_data->args[4] = lp_build_const_int32(bld_base->base.gallivm,
+ emit_data->inst->Texture.Texture);
+*/
+
+ emit_data->arg_count = 6;
+ /* XXX: To optimize, we could use a float or v2f32, if the last bits of
+ * the writemask are clear */
+ emit_data->dst_type = LLVMVectorType(
+ LLVMFloatTypeInContext(bld_base->base.gallivm->context),
+ 4);
+}
+
+static const struct lp_build_tgsi_action tex_action = {
+ .fetch_args = tex_fetch_args,
+ .emit = lp_build_tgsi_intrinsic,
+ .intr_name = "llvm.SI.sample"
+};
+
+
+int si_pipe_shader_create(
+ struct pipe_context *ctx,
+ struct si_pipe_shader *shader)
+{
+ struct r600_context *rctx = (struct r600_context*)ctx;
+ struct si_shader_context si_shader_ctx;
+ struct tgsi_shader_info shader_info;
+ struct lp_build_tgsi_context * bld_base;
+ LLVMModuleRef mod;
+ unsigned char * inst_bytes;
+ unsigned inst_byte_count;
+ unsigned i;
+
+ radeon_llvm_context_init(&si_shader_ctx.radeon_bld);
+ bld_base = &si_shader_ctx.radeon_bld.soa.bld_base;
+
+ tgsi_scan_shader(shader->tokens, &shader_info);
+ bld_base->info = &shader_info;
+ bld_base->emit_fetch_funcs[TGSI_FILE_CONSTANT] = fetch_constant;
+ bld_base->emit_prologue = si_llvm_emit_prologue;
+ bld_base->emit_epilogue = si_llvm_emit_epilogue;
+
+ bld_base->op_actions[TGSI_OPCODE_TEX] = tex_action;
+
+ si_shader_ctx.radeon_bld.load_input = declare_input;
+ si_shader_ctx.tokens = shader->tokens;
+ tgsi_parse_init(&si_shader_ctx.parse, si_shader_ctx.tokens);
+ si_shader_ctx.shader = shader;
+ si_shader_ctx.type = si_shader_ctx.parse.FullHeader.Processor.Processor;
+ si_shader_ctx.rctx = rctx;
+
+ shader->shader.nr_cbufs = rctx->nr_cbufs;
+
+ lp_build_tgsi_llvm(bld_base, shader->tokens);
+
+ radeon_llvm_finalize_module(&si_shader_ctx.radeon_bld);
+
+ mod = bld_base->base.gallivm->module;
+ tgsi_dump(shader->tokens, 0);
+ LLVMDumpModule(mod);
+ radeon_llvm_compile(mod, &inst_bytes, &inst_byte_count, "SI", 1 /* dump */);
+ fprintf(stderr, "SI CODE:\n");
+ for (i = 0; i < inst_byte_count; i+=4 ) {
+ fprintf(stderr, "%02x%02x%02x%02x\n", inst_bytes[i + 3],
+ inst_bytes[i + 2], inst_bytes[i + 1],
+ inst_bytes[i]);
+ }
+
+ shader->num_sgprs = util_le32_to_cpu(*(uint32_t*)inst_bytes);
+ shader->num_vgprs = util_le32_to_cpu(*(uint32_t*)(inst_bytes + 4));
+ shader->spi_ps_input_ena = util_le32_to_cpu(*(uint32_t*)(inst_bytes + 8));
+
+ tgsi_parse_free(&si_shader_ctx.parse);
+
+ /* copy new shader */
+ if (shader->bo == NULL) {
+ uint32_t *ptr;
+
+ shader->bo = (struct r600_resource*)
+ pipe_buffer_create(ctx->screen, PIPE_BIND_CUSTOM, PIPE_USAGE_IMMUTABLE, inst_byte_count);
+ if (shader->bo == NULL) {
+ return -ENOMEM;
+ }
+ ptr = (uint32_t*)rctx->ws->buffer_map(shader->bo->buf, rctx->cs, PIPE_TRANSFER_WRITE);
+ if (0 /*R600_BIG_ENDIAN*/) {
+ for (i = 0; i < (inst_byte_count-12)/4; ++i) {
+ ptr[i] = util_bswap32(*(uint32_t*)(inst_bytes+12 + i*4));
+ }
+ } else {
+ memcpy(ptr, inst_bytes + 12, inst_byte_count - 12);
+ }
+ rctx->ws->buffer_unmap(shader->bo->buf);
+ }
+
+ free(inst_bytes);
+
+ return 0;
+}
+
+void si_pipe_shader_destroy(struct pipe_context *ctx, struct si_pipe_shader *shader)
+{
+ pipe_resource_reference((struct pipe_resource**)&shader->bo, NULL);
+
+ memset(&shader->shader,0,sizeof(struct r600_shader));
+}
diff --git a/src/gallium/drivers/radeonsi/radeonsi_shader.h b/src/gallium/drivers/radeonsi/radeonsi_shader.h
new file mode 100644
index 00000000000..cd742f57da1
--- /dev/null
+++ b/src/gallium/drivers/radeonsi/radeonsi_shader.h
@@ -0,0 +1,4 @@
+
+struct tgsi_token;
+
+void si_test(struct tgsi_token * token, unsigned type);
diff --git a/src/gallium/drivers/radeonsi/sid.h b/src/gallium/drivers/radeonsi/sid.h
new file mode 100644
index 00000000000..325445c97e0
--- /dev/null
+++ b/src/gallium/drivers/radeonsi/sid.h
@@ -0,0 +1,7668 @@
+/*
+ * Southern Islands Register documentation
+ *
+ * Copyright (C) 2011 Advanced Micro Devices, Inc.
+ *
+ * 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
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * 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 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 NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) 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.
+ */
+
+#ifndef SID_H
+#define SID_H
+
+/* si values */
+#define SI_CONFIG_REG_OFFSET 0x00008000
+#define SI_CONFIG_REG_END 0x0000B000
+#define SI_SH_REG_OFFSET 0x0000B000
+#define SI_SH_REG_END 0x0000C000
+#define SI_CONTEXT_REG_OFFSET 0x00028000
+#define SI_CONTEXT_REG_END 0x00029000
+
+#define EVENT_TYPE_PS_PARTIAL_FLUSH 0x10
+#define EVENT_TYPE_CACHE_FLUSH_AND_INV_TS_EVENT 0x14
+#define EVENT_TYPE_ZPASS_DONE 0x15
+#define EVENT_TYPE_CACHE_FLUSH_AND_INV_EVENT 0x16
+#define EVENT_TYPE_SO_VGTSTREAMOUT_FLUSH 0x1f
+#define EVENT_TYPE_SAMPLE_STREAMOUTSTATS 0x20
+#define EVENT_TYPE(x) ((x) << 0)
+#define EVENT_INDEX(x) ((x) << 8)
+ /* 0 - any non-TS event
+ * 1 - ZPASS_DONE
+ * 2 - SAMPLE_PIPELINESTAT
+ * 3 - SAMPLE_STREAMOUTSTAT*
+ * 4 - *S_PARTIAL_FLUSH
+ * 5 - TS events
+ */
+
+#define PREDICATION_OP_CLEAR 0x0
+#define PREDICATION_OP_ZPASS 0x1
+#define PREDICATION_OP_PRIMCOUNT 0x2
+
+#define PRED_OP(x) ((x) << 16)
+
+#define PREDICATION_CONTINUE (1 << 31)
+
+#define PREDICATION_HINT_WAIT (0 << 12)
+#define PREDICATION_HINT_NOWAIT_DRAW (1 << 12)
+
+#define PREDICATION_DRAW_NOT_VISIBLE (0 << 8)
+#define PREDICATION_DRAW_VISIBLE (1 << 8)
+
+#define R600_TEXEL_PITCH_ALIGNMENT_MASK 0x7
+
+#define PKT3_NOP 0x10
+#define PKT3_SET_PREDICATION 0x20
+#define PKT3_COND_EXEC 0x22
+#define PKT3_PRED_EXEC 0x23
+#define PKT3_START_3D_CMDBUF 0x24
+#define PKT3_DRAW_INDEX_2 0x27
+#define PKT3_CONTEXT_CONTROL 0x28
+#define PKT3_INDEX_TYPE 0x2A
+#define PKT3_DRAW_INDEX 0x2B
+#define PKT3_DRAW_INDEX_AUTO 0x2D
+#define PKT3_DRAW_INDEX_IMMD 0x2E
+#define PKT3_NUM_INSTANCES 0x2F
+#define PKT3_STRMOUT_BUFFER_UPDATE 0x34
+#define PKT3_MEM_SEMAPHORE 0x39
+#define PKT3_MPEG_INDEX 0x3A
+#define PKT3_WAIT_REG_MEM 0x3C
+#define WAIT_REG_MEM_EQUAL 3
+#define PKT3_MEM_WRITE 0x3D
+#define PKT3_INDIRECT_BUFFER 0x32
+#define PKT3_SURFACE_SYNC 0x43
+#define PKT3_ME_INITIALIZE 0x44
+#define PKT3_COND_WRITE 0x45
+#define PKT3_EVENT_WRITE 0x46
+#define PKT3_EVENT_WRITE_EOP 0x47
+#define PKT3_EVENT_WRITE_EOS 0x48
+#define PKT3_ONE_REG_WRITE 0x57
+#define PKT3_SET_CONFIG_REG 0x68
+#define PKT3_SET_CONTEXT_REG 0x69
+#define PKT3_SET_SH_REG 0x76
+
+#define PKT_TYPE_S(x) (((x) & 0x3) << 30)
+#define PKT_TYPE_G(x) (((x) >> 30) & 0x3)
+#define PKT_TYPE_C 0x3FFFFFFF
+#define PKT_COUNT_S(x) (((x) & 0x3FFF) << 16)
+#define PKT_COUNT_G(x) (((x) >> 16) & 0x3FFF)
+#define PKT_COUNT_C 0xC000FFFF
+#define PKT0_BASE_INDEX_S(x) (((x) & 0xFFFF) << 0)
+#define PKT0_BASE_INDEX_G(x) (((x) >> 0) & 0xFFFF)
+#define PKT0_BASE_INDEX_C 0xFFFF0000
+#define PKT3_IT_OPCODE_S(x) (((x) & 0xFF) << 8)
+#define PKT3_IT_OPCODE_G(x) (((x) >> 8) & 0xFF)
+#define PKT3_IT_OPCODE_C 0xFFFF00FF
+#define PKT3_PREDICATE(x) (((x) >> 0) & 0x1)
+#define PKT0(index, count) (PKT_TYPE_S(0) | PKT0_BASE_INDEX_S(index) | PKT_COUNT_S(count))
+#define PKT3(op, count, predicate) (PKT_TYPE_S(3) | PKT3_IT_OPCODE_S(op) | PKT_COUNT_S(count) | PKT3_PREDICATE(predicate))
+
+#define R_0084FC_CP_STRMOUT_CNTL 0x0084FC
+#define S_0084FC_OFFSET_UPDATE_DONE(x) (((x) & 0x1) << 0)
+#define R_0085F0_CP_COHER_CNTL 0x0085F0
+#define S_0085F0_DEST_BASE_0_ENA(x) (((x) & 0x1) << 0)
+#define G_0085F0_DEST_BASE_0_ENA(x) (((x) >> 0) & 0x1)
+#define C_0085F0_DEST_BASE_0_ENA 0xFFFFFFFE
+#define S_0085F0_DEST_BASE_1_ENA(x) (((x) & 0x1) << 1)
+#define G_0085F0_DEST_BASE_1_ENA(x) (((x) >> 1) & 0x1)
+#define C_0085F0_DEST_BASE_1_ENA 0xFFFFFFFD
+#define S_0085F0_CB0_DEST_BASE_ENA_SHIFT 6
+#define S_0085F0_CB0_DEST_BASE_ENA(x) (((x) & 0x1) << 6)
+#define G_0085F0_CB0_DEST_BASE_ENA(x) (((x) >> 6) & 0x1)
+#define C_0085F0_CB0_DEST_BASE_ENA 0xFFFFFFBF
+#define S_0085F0_CB1_DEST_BASE_ENA(x) (((x) & 0x1) << 7)
+#define G_0085F0_CB1_DEST_BASE_ENA(x) (((x) >> 7) & 0x1)
+#define C_0085F0_CB1_DEST_BASE_ENA 0xFFFFFF7F
+#define S_0085F0_CB2_DEST_BASE_ENA(x) (((x) & 0x1) << 8)
+#define G_0085F0_CB2_DEST_BASE_ENA(x) (((x) >> 8) & 0x1)
+#define C_0085F0_CB2_DEST_BASE_ENA 0xFFFFFEFF
+#define S_0085F0_CB3_DEST_BASE_ENA(x) (((x) & 0x1) << 9)
+#define G_0085F0_CB3_DEST_BASE_ENA(x) (((x) >> 9) & 0x1)
+#define C_0085F0_CB3_DEST_BASE_ENA 0xFFFFFDFF
+#define S_0085F0_CB4_DEST_BASE_ENA(x) (((x) & 0x1) << 10)
+#define G_0085F0_CB4_DEST_BASE_ENA(x) (((x) >> 10) & 0x1)
+#define C_0085F0_CB4_DEST_BASE_ENA 0xFFFFFBFF
+#define S_0085F0_CB5_DEST_BASE_ENA(x) (((x) & 0x1) << 11)
+#define G_0085F0_CB5_DEST_BASE_ENA(x) (((x) >> 11) & 0x1)
+#define C_0085F0_CB5_DEST_BASE_ENA 0xFFFFF7FF
+#define S_0085F0_CB6_DEST_BASE_ENA(x) (((x) & 0x1) << 12)
+#define G_0085F0_CB6_DEST_BASE_ENA(x) (((x) >> 12) & 0x1)
+#define C_0085F0_CB6_DEST_BASE_ENA 0xFFFFEFFF
+#define S_0085F0_CB7_DEST_BASE_ENA(x) (((x) & 0x1) << 13)
+#define G_0085F0_CB7_DEST_BASE_ENA(x) (((x) >> 13) & 0x1)
+#define C_0085F0_CB7_DEST_BASE_ENA 0xFFFFDFFF
+#define S_0085F0_DB_DEST_BASE_ENA(x) (((x) & 0x1) << 14)
+#define G_0085F0_DB_DEST_BASE_ENA(x) (((x) >> 14) & 0x1)
+#define C_0085F0_DB_DEST_BASE_ENA 0xFFFFBFFF
+#define S_0085F0_DEST_BASE_2_ENA(x) (((x) & 0x1) << 19)
+#define G_0085F0_DEST_BASE_2_ENA(x) (((x) >> 19) & 0x1)
+#define C_0085F0_DEST_BASE_2_ENA 0xFFF7FFFF
+#define S_0085F0_DEST_BASE_3_ENA(x) (((x) & 0x1) << 21)
+#define G_0085F0_DEST_BASE_3_ENA(x) (((x) >> 21) & 0x1)
+#define C_0085F0_DEST_BASE_3_ENA 0xFFDFFFFF
+#define S_0085F0_TCL1_ACTION_ENA(x) (((x) & 0x1) << 22)
+#define G_0085F0_TCL1_ACTION_ENA(x) (((x) >> 22) & 0x1)
+#define C_0085F0_TCL1_ACTION_ENA 0xFFBFFFFF
+#define S_0085F0_TC_ACTION_ENA(x) (((x) & 0x1) << 23)
+#define G_0085F0_TC_ACTION_ENA(x) (((x) >> 23) & 0x1)
+#define C_0085F0_TC_ACTION_ENA 0xFF7FFFFF
+#define S_0085F0_CB_ACTION_ENA(x) (((x) & 0x1) << 25)
+#define G_0085F0_CB_ACTION_ENA(x) (((x) >> 25) & 0x1)
+#define C_0085F0_CB_ACTION_ENA 0xFDFFFFFF
+#define S_0085F0_DB_ACTION_ENA(x) (((x) & 0x1) << 26)
+#define G_0085F0_DB_ACTION_ENA(x) (((x) >> 26) & 0x1)
+#define C_0085F0_DB_ACTION_ENA 0xFBFFFFFF
+#define S_0085F0_SH_KCACHE_ACTION_ENA(x) (((x) & 0x1) << 27)
+#define G_0085F0_SH_KCACHE_ACTION_ENA(x) (((x) >> 27) & 0x1)
+#define C_0085F0_SH_KCACHE_ACTION_ENA 0xF7FFFFFF
+#define S_0085F0_SH_ICACHE_ACTION_ENA(x) (((x) & 0x1) << 29)
+#define G_0085F0_SH_ICACHE_ACTION_ENA(x) (((x) >> 29) & 0x1)
+#define C_0085F0_SH_ICACHE_ACTION_ENA 0xDFFFFFFF
+#define R_0085F4_CP_COHER_SIZE 0x0085F4
+#define R_0085F8_CP_COHER_BASE 0x0085F8
+#define R_0088B0_VGT_VTX_VECT_EJECT_REG 0x0088B0
+#define S_0088B0_PRIM_COUNT(x) (((x) & 0x3FF) << 0)
+#define G_0088B0_PRIM_COUNT(x) (((x) >> 0) & 0x3FF)
+#define C_0088B0_PRIM_COUNT 0xFFFFFC00
+#define R_0088C4_VGT_CACHE_INVALIDATION 0x0088C4
+#define S_0088C4_VS_NO_EXTRA_BUFFER(x) (((x) & 0x1) << 5)
+#define G_0088C4_VS_NO_EXTRA_BUFFER(x) (((x) >> 5) & 0x1)
+#define C_0088C4_VS_NO_EXTRA_BUFFER 0xFFFFFFDF
+#define S_0088C4_STREAMOUT_FULL_FLUSH(x) (((x) & 0x1) << 13)
+#define G_0088C4_STREAMOUT_FULL_FLUSH(x) (((x) >> 13) & 0x1)
+#define C_0088C4_STREAMOUT_FULL_FLUSH 0xFFFFDFFF
+#define S_0088C4_ES_LIMIT(x) (((x) & 0x1F) << 16)
+#define G_0088C4_ES_LIMIT(x) (((x) >> 16) & 0x1F)
+#define C_0088C4_ES_LIMIT 0xFFE0FFFF
+#define R_0088C8_VGT_ESGS_RING_SIZE 0x0088C8
+#define R_0088CC_VGT_GSVS_RING_SIZE 0x0088CC
+#define R_0088D4_VGT_GS_VERTEX_REUSE 0x0088D4
+#define S_0088D4_VERT_REUSE(x) (((x) & 0x1F) << 0)
+#define G_0088D4_VERT_REUSE(x) (((x) >> 0) & 0x1F)
+#define C_0088D4_VERT_REUSE 0xFFFFFFE0
+#define R_008958_VGT_PRIMITIVE_TYPE 0x008958
+#define S_008958_PRIM_TYPE(x) (((x) & 0x3F) << 0)
+#define G_008958_PRIM_TYPE(x) (((x) >> 0) & 0x3F)
+#define C_008958_PRIM_TYPE 0xFFFFFFC0
+#define V_008958_DI_PT_NONE 0x00
+#define V_008958_DI_PT_POINTLIST 0x01
+#define V_008958_DI_PT_LINELIST 0x02
+#define V_008958_DI_PT_LINESTRIP 0x03
+#define V_008958_DI_PT_TRILIST 0x04
+#define V_008958_DI_PT_TRIFAN 0x05
+#define V_008958_DI_PT_TRISTRIP 0x06
+#define V_008958_DI_PT_UNUSED_0 0x07
+#define V_008958_DI_PT_UNUSED_1 0x08
+#define V_008958_DI_PT_PATCH 0x09
+#define V_008958_DI_PT_LINELIST_ADJ 0x0A
+#define V_008958_DI_PT_LINESTRIP_ADJ 0x0B
+#define V_008958_DI_PT_TRILIST_ADJ 0x0C
+#define V_008958_DI_PT_TRISTRIP_ADJ 0x0D
+#define V_008958_DI_PT_UNUSED_3 0x0E
+#define V_008958_DI_PT_UNUSED_4 0x0F
+#define V_008958_DI_PT_TRI_WITH_WFLAGS 0x10
+#define V_008958_DI_PT_RECTLIST 0x11
+#define V_008958_DI_PT_LINELOOP 0x12
+#define V_008958_DI_PT_QUADLIST 0x13
+#define V_008958_DI_PT_QUADSTRIP 0x14
+#define V_008958_DI_PT_POLYGON 0x15
+#define V_008958_DI_PT_2D_COPY_RECT_LIST_V0 0x16
+#define V_008958_DI_PT_2D_COPY_RECT_LIST_V1 0x17
+#define V_008958_DI_PT_2D_COPY_RECT_LIST_V2 0x18
+#define V_008958_DI_PT_2D_COPY_RECT_LIST_V3 0x19
+#define V_008958_DI_PT_2D_FILL_RECT_LIST 0x1A
+#define V_008958_DI_PT_2D_LINE_STRIP 0x1B
+#define V_008958_DI_PT_2D_TRI_STRIP 0x1C
+#define R_00895C_VGT_INDEX_TYPE 0x00895C
+#define S_00895C_INDEX_TYPE(x) (((x) & 0x03) << 0)
+#define G_00895C_INDEX_TYPE(x) (((x) >> 0) & 0x03)
+#define C_00895C_INDEX_TYPE 0xFFFFFFFC
+#define V_00895C_DI_INDEX_SIZE_16_BIT 0x00
+#define V_00895C_DI_INDEX_SIZE_32_BIT 0x01
+#define R_008960_VGT_STRMOUT_BUFFER_FILLED_SIZE_0 0x008960
+#define R_008964_VGT_STRMOUT_BUFFER_FILLED_SIZE_1 0x008964
+#define R_008968_VGT_STRMOUT_BUFFER_FILLED_SIZE_2 0x008968
+#define R_00896C_VGT_STRMOUT_BUFFER_FILLED_SIZE_3 0x00896C
+#define R_008970_VGT_NUM_INDICES 0x008970
+#define R_008974_VGT_NUM_INSTANCES 0x008974
+#define R_008988_VGT_TF_RING_SIZE 0x008988
+#define S_008988_SIZE(x) (((x) & 0xFFFF) << 0)
+#define G_008988_SIZE(x) (((x) >> 0) & 0xFFFF)
+#define C_008988_SIZE 0xFFFF0000
+#define R_0089B0_VGT_HS_OFFCHIP_PARAM 0x0089B0
+#define S_0089B0_OFFCHIP_BUFFERING(x) (((x) & 0x7F) << 0)
+#define G_0089B0_OFFCHIP_BUFFERING(x) (((x) >> 0) & 0x7F)
+#define C_0089B0_OFFCHIP_BUFFERING 0xFFFFFF80
+#define R_0089B8_VGT_TF_MEMORY_BASE 0x0089B8
+#define R_008A14_PA_CL_ENHANCE 0x008A14
+#define S_008A14_CLIP_VTX_REORDER_ENA(x) (((x) & 0x1) << 0)
+#define G_008A14_CLIP_VTX_REORDER_ENA(x) (((x) >> 0) & 0x1)
+#define C_008A14_CLIP_VTX_REORDER_ENA 0xFFFFFFFE
+#define S_008A14_NUM_CLIP_SEQ(x) (((x) & 0x03) << 1)
+#define G_008A14_NUM_CLIP_SEQ(x) (((x) >> 1) & 0x03)
+#define C_008A14_NUM_CLIP_SEQ 0xFFFFFFF9
+#define S_008A14_CLIPPED_PRIM_SEQ_STALL(x) (((x) & 0x1) << 3)
+#define G_008A14_CLIPPED_PRIM_SEQ_STALL(x) (((x) >> 3) & 0x1)
+#define C_008A14_CLIPPED_PRIM_SEQ_STALL 0xFFFFFFF7
+#define S_008A14_VE_NAN_PROC_DISABLE(x) (((x) & 0x1) << 4)
+#define G_008A14_VE_NAN_PROC_DISABLE(x) (((x) >> 4) & 0x1)
+#define C_008A14_VE_NAN_PROC_DISABLE 0xFFFFFFEF
+#define R_008A60_PA_SU_LINE_STIPPLE_VALUE 0x008A60
+#define S_008A60_LINE_STIPPLE_VALUE(x) (((x) & 0xFFFFFF) << 0)
+#define G_008A60_LINE_STIPPLE_VALUE(x) (((x) >> 0) & 0xFFFFFF)
+#define C_008A60_LINE_STIPPLE_VALUE 0xFF000000
+#define R_008B10_PA_SC_LINE_STIPPLE_STATE 0x008B10
+#define S_008B10_CURRENT_PTR(x) (((x) & 0x0F) << 0)
+#define G_008B10_CURRENT_PTR(x) (((x) >> 0) & 0x0F)
+#define C_008B10_CURRENT_PTR 0xFFFFFFF0
+#define S_008B10_CURRENT_COUNT(x) (((x) & 0xFF) << 8)
+#define G_008B10_CURRENT_COUNT(x) (((x) >> 8) & 0xFF)
+#define C_008B10_CURRENT_COUNT 0xFFFF00FF
+#define R_008BF0_PA_SC_ENHANCE 0x008BF0
+#define S_008BF0_ENABLE_PA_SC_OUT_OF_ORDER(x) (((x) & 0x1) << 0)
+#define G_008BF0_ENABLE_PA_SC_OUT_OF_ORDER(x) (((x) >> 0) & 0x1)
+#define C_008BF0_ENABLE_PA_SC_OUT_OF_ORDER 0xFFFFFFFE
+#define S_008BF0_DISABLE_SC_DB_TILE_FIX(x) (((x) & 0x1) << 1)
+#define G_008BF0_DISABLE_SC_DB_TILE_FIX(x) (((x) >> 1) & 0x1)
+#define C_008BF0_DISABLE_SC_DB_TILE_FIX 0xFFFFFFFD
+#define S_008BF0_DISABLE_AA_MASK_FULL_FIX(x) (((x) & 0x1) << 2)
+#define G_008BF0_DISABLE_AA_MASK_FULL_FIX(x) (((x) >> 2) & 0x1)
+#define C_008BF0_DISABLE_AA_MASK_FULL_FIX 0xFFFFFFFB
+#define S_008BF0_ENABLE_1XMSAA_SAMPLE_LOCATIONS(x) (((x) & 0x1) << 3)
+#define G_008BF0_ENABLE_1XMSAA_SAMPLE_LOCATIONS(x) (((x) >> 3) & 0x1)
+#define C_008BF0_ENABLE_1XMSAA_SAMPLE_LOCATIONS 0xFFFFFFF7
+#define S_008BF0_ENABLE_1XMSAA_SAMPLE_LOC_CENTROID(x) (((x) & 0x1) << 4)
+#define G_008BF0_ENABLE_1XMSAA_SAMPLE_LOC_CENTROID(x) (((x) >> 4) & 0x1)
+#define C_008BF0_ENABLE_1XMSAA_SAMPLE_LOC_CENTROID 0xFFFFFFEF
+#define S_008BF0_DISABLE_SCISSOR_FIX(x) (((x) & 0x1) << 5)
+#define G_008BF0_DISABLE_SCISSOR_FIX(x) (((x) >> 5) & 0x1)
+#define C_008BF0_DISABLE_SCISSOR_FIX 0xFFFFFFDF
+#define S_008BF0_DISABLE_PW_BUBBLE_COLLAPSE(x) (((x) & 0x03) << 6)
+#define G_008BF0_DISABLE_PW_BUBBLE_COLLAPSE(x) (((x) >> 6) & 0x03)
+#define C_008BF0_DISABLE_PW_BUBBLE_COLLAPSE 0xFFFFFF3F
+#define S_008BF0_SEND_UNLIT_STILES_TO_PACKER(x) (((x) & 0x1) << 8)
+#define G_008BF0_SEND_UNLIT_STILES_TO_PACKER(x) (((x) >> 8) & 0x1)
+#define C_008BF0_SEND_UNLIT_STILES_TO_PACKER 0xFFFFFEFF
+#define S_008BF0_DISABLE_DUALGRAD_PERF_OPTIMIZATION(x) (((x) & 0x1) << 9)
+#define G_008BF0_DISABLE_DUALGRAD_PERF_OPTIMIZATION(x) (((x) >> 9) & 0x1)
+#define C_008BF0_DISABLE_DUALGRAD_PERF_OPTIMIZATION 0xFFFFFDFF
+#define R_008C08_SQC_CACHES 0x008C08
+#define S_008C08_INST_INVALIDATE(x) (((x) & 0x1) << 0)
+#define G_008C08_INST_INVALIDATE(x) (((x) >> 0) & 0x1)
+#define C_008C08_INST_INVALIDATE 0xFFFFFFFE
+#define S_008C08_DATA_INVALIDATE(x) (((x) & 0x1) << 1)
+#define G_008C08_DATA_INVALIDATE(x) (((x) >> 1) & 0x1)
+#define C_008C08_DATA_INVALIDATE 0xFFFFFFFD
+#define R_008C0C_SQ_RANDOM_WAVE_PRI 0x008C0C
+#define S_008C0C_RET(x) (((x) & 0x7F) << 0)
+#define G_008C0C_RET(x) (((x) >> 0) & 0x7F)
+#define C_008C0C_RET 0xFFFFFF80
+#define S_008C0C_RUI(x) (((x) & 0x07) << 7)
+#define G_008C0C_RUI(x) (((x) >> 7) & 0x07)
+#define C_008C0C_RUI 0xFFFFFC7F
+#define S_008C0C_RNG(x) (((x) & 0x7FF) << 10)
+#define G_008C0C_RNG(x) (((x) >> 10) & 0x7FF)
+#define C_008C0C_RNG 0xFFE003FF
+#if 0
+#define R_008DFC_SQ_INST 0x008DFC
+#define R_008DFC_SQ_VOP1 0x008DFC
+#define S_008DFC_SRC0(x) (((x) & 0x1FF) << 0)
+#define G_008DFC_SRC0(x) (((x) >> 0) & 0x1FF)
+#define C_008DFC_SRC0 0xFFFFFE00
+#define V_008DFC_SQ_SGPR 0x00
+#define V_008DFC_SQ_VCC_LO 0x6A
+#define V_008DFC_SQ_VCC_HI 0x6B
+#define V_008DFC_SQ_TBA_LO 0x6C
+#define V_008DFC_SQ_TBA_HI 0x6D
+#define V_008DFC_SQ_TMA_LO 0x6E
+#define V_008DFC_SQ_TMA_HI 0x6F
+#define V_008DFC_SQ_TTMP0 0x70
+#define V_008DFC_SQ_TTMP1 0x71
+#define V_008DFC_SQ_TTMP2 0x72
+#define V_008DFC_SQ_TTMP3 0x73
+#define V_008DFC_SQ_TTMP4 0x74
+#define V_008DFC_SQ_TTMP5 0x75
+#define V_008DFC_SQ_TTMP6 0x76
+#define V_008DFC_SQ_TTMP7 0x77
+#define V_008DFC_SQ_TTMP8 0x78
+#define V_008DFC_SQ_TTMP9 0x79
+#define V_008DFC_SQ_TTMP10 0x7A
+#define V_008DFC_SQ_TTMP11 0x7B
+#define V_008DFC_SQ_M0 0x7C
+#define V_008DFC_SQ_EXEC_LO 0x7E
+#define V_008DFC_SQ_EXEC_HI 0x7F
+#define V_008DFC_SQ_SRC_0 0x80
+#define V_008DFC_SQ_SRC_1_INT 0x81
+#define V_008DFC_SQ_SRC_2_INT 0x82
+#define V_008DFC_SQ_SRC_3_INT 0x83
+#define V_008DFC_SQ_SRC_4_INT 0x84
+#define V_008DFC_SQ_SRC_5_INT 0x85
+#define V_008DFC_SQ_SRC_6_INT 0x86
+#define V_008DFC_SQ_SRC_7_INT 0x87
+#define V_008DFC_SQ_SRC_8_INT 0x88
+#define V_008DFC_SQ_SRC_9_INT 0x89
+#define V_008DFC_SQ_SRC_10_INT 0x8A
+#define V_008DFC_SQ_SRC_11_INT 0x8B
+#define V_008DFC_SQ_SRC_12_INT 0x8C
+#define V_008DFC_SQ_SRC_13_INT 0x8D
+#define V_008DFC_SQ_SRC_14_INT 0x8E
+#define V_008DFC_SQ_SRC_15_INT 0x8F
+#define V_008DFC_SQ_SRC_16_INT 0x90
+#define V_008DFC_SQ_SRC_17_INT 0x91
+#define V_008DFC_SQ_SRC_18_INT 0x92
+#define V_008DFC_SQ_SRC_19_INT 0x93
+#define V_008DFC_SQ_SRC_20_INT 0x94
+#define V_008DFC_SQ_SRC_21_INT 0x95
+#define V_008DFC_SQ_SRC_22_INT 0x96
+#define V_008DFC_SQ_SRC_23_INT 0x97
+#define V_008DFC_SQ_SRC_24_INT 0x98
+#define V_008DFC_SQ_SRC_25_INT 0x99
+#define V_008DFC_SQ_SRC_26_INT 0x9A
+#define V_008DFC_SQ_SRC_27_INT 0x9B
+#define V_008DFC_SQ_SRC_28_INT 0x9C
+#define V_008DFC_SQ_SRC_29_INT 0x9D
+#define V_008DFC_SQ_SRC_30_INT 0x9E
+#define V_008DFC_SQ_SRC_31_INT 0x9F
+#define V_008DFC_SQ_SRC_32_INT 0xA0
+#define V_008DFC_SQ_SRC_33_INT 0xA1
+#define V_008DFC_SQ_SRC_34_INT 0xA2
+#define V_008DFC_SQ_SRC_35_INT 0xA3
+#define V_008DFC_SQ_SRC_36_INT 0xA4
+#define V_008DFC_SQ_SRC_37_INT 0xA5
+#define V_008DFC_SQ_SRC_38_INT 0xA6
+#define V_008DFC_SQ_SRC_39_INT 0xA7
+#define V_008DFC_SQ_SRC_40_INT 0xA8
+#define V_008DFC_SQ_SRC_41_INT 0xA9
+#define V_008DFC_SQ_SRC_42_INT 0xAA
+#define V_008DFC_SQ_SRC_43_INT 0xAB
+#define V_008DFC_SQ_SRC_44_INT 0xAC
+#define V_008DFC_SQ_SRC_45_INT 0xAD
+#define V_008DFC_SQ_SRC_46_INT 0xAE
+#define V_008DFC_SQ_SRC_47_INT 0xAF
+#define V_008DFC_SQ_SRC_48_INT 0xB0
+#define V_008DFC_SQ_SRC_49_INT 0xB1
+#define V_008DFC_SQ_SRC_50_INT 0xB2
+#define V_008DFC_SQ_SRC_51_INT 0xB3
+#define V_008DFC_SQ_SRC_52_INT 0xB4
+#define V_008DFC_SQ_SRC_53_INT 0xB5
+#define V_008DFC_SQ_SRC_54_INT 0xB6
+#define V_008DFC_SQ_SRC_55_INT 0xB7
+#define V_008DFC_SQ_SRC_56_INT 0xB8
+#define V_008DFC_SQ_SRC_57_INT 0xB9
+#define V_008DFC_SQ_SRC_58_INT 0xBA
+#define V_008DFC_SQ_SRC_59_INT 0xBB
+#define V_008DFC_SQ_SRC_60_INT 0xBC
+#define V_008DFC_SQ_SRC_61_INT 0xBD
+#define V_008DFC_SQ_SRC_62_INT 0xBE
+#define V_008DFC_SQ_SRC_63_INT 0xBF
+#define V_008DFC_SQ_SRC_64_INT 0xC0
+#define V_008DFC_SQ_SRC_M_1_INT 0xC1
+#define V_008DFC_SQ_SRC_M_2_INT 0xC2
+#define V_008DFC_SQ_SRC_M_3_INT 0xC3
+#define V_008DFC_SQ_SRC_M_4_INT 0xC4
+#define V_008DFC_SQ_SRC_M_5_INT 0xC5
+#define V_008DFC_SQ_SRC_M_6_INT 0xC6
+#define V_008DFC_SQ_SRC_M_7_INT 0xC7
+#define V_008DFC_SQ_SRC_M_8_INT 0xC8
+#define V_008DFC_SQ_SRC_M_9_INT 0xC9
+#define V_008DFC_SQ_SRC_M_10_INT 0xCA
+#define V_008DFC_SQ_SRC_M_11_INT 0xCB
+#define V_008DFC_SQ_SRC_M_12_INT 0xCC
+#define V_008DFC_SQ_SRC_M_13_INT 0xCD
+#define V_008DFC_SQ_SRC_M_14_INT 0xCE
+#define V_008DFC_SQ_SRC_M_15_INT 0xCF
+#define V_008DFC_SQ_SRC_M_16_INT 0xD0
+#define V_008DFC_SQ_SRC_0_5 0xF0
+#define V_008DFC_SQ_SRC_M_0_5 0xF1
+#define V_008DFC_SQ_SRC_1 0xF2
+#define V_008DFC_SQ_SRC_M_1 0xF3
+#define V_008DFC_SQ_SRC_2 0xF4
+#define V_008DFC_SQ_SRC_M_2 0xF5
+#define V_008DFC_SQ_SRC_4 0xF6
+#define V_008DFC_SQ_SRC_M_4 0xF7
+#define V_008DFC_SQ_SRC_VCCZ 0xFB
+#define V_008DFC_SQ_SRC_EXECZ 0xFC
+#define V_008DFC_SQ_SRC_SCC 0xFD
+#define V_008DFC_SQ_SRC_LDS_DIRECT 0xFE
+#define V_008DFC_SQ_SRC_VGPR 0x100
+#define S_008DFC_OP(x) (((x) & 0xFF) << 9)
+#define G_008DFC_OP(x) (((x) >> 9) & 0xFF)
+#define C_008DFC_OP 0xFFFE01FF
+#define V_008DFC_SQ_V_NOP 0x00
+#define V_008DFC_SQ_V_MOV_B32 0x01
+#define V_008DFC_SQ_V_READFIRSTLANE_B32 0x02
+#define V_008DFC_SQ_V_CVT_I32_F64 0x03
+#define V_008DFC_SQ_V_CVT_F64_I32 0x04
+#define V_008DFC_SQ_V_CVT_F32_I32 0x05
+#define V_008DFC_SQ_V_CVT_F32_U32 0x06
+#define V_008DFC_SQ_V_CVT_U32_F32 0x07
+#define V_008DFC_SQ_V_CVT_I32_F32 0x08
+#define V_008DFC_SQ_V_MOV_FED_B32 0x09
+#define V_008DFC_SQ_V_CVT_F16_F32 0x0A
+#define V_008DFC_SQ_V_CVT_F32_F16 0x0B
+#define V_008DFC_SQ_V_CVT_RPI_I32_F32 0x0C
+#define V_008DFC_SQ_V_CVT_FLR_I32_F32 0x0D
+#define V_008DFC_SQ_V_CVT_OFF_F32_I4 0x0E
+#define V_008DFC_SQ_V_CVT_F32_F64 0x0F
+#define V_008DFC_SQ_V_CVT_F64_F32 0x10
+#define V_008DFC_SQ_V_CVT_F32_UBYTE0 0x11
+#define V_008DFC_SQ_V_CVT_F32_UBYTE1 0x12
+#define V_008DFC_SQ_V_CVT_F32_UBYTE2 0x13
+#define V_008DFC_SQ_V_CVT_F32_UBYTE3 0x14
+#define V_008DFC_SQ_V_CVT_U32_F64 0x15
+#define V_008DFC_SQ_V_CVT_F64_U32 0x16
+#define V_008DFC_SQ_V_FRACT_F32 0x20
+#define V_008DFC_SQ_V_TRUNC_F32 0x21
+#define V_008DFC_SQ_V_CEIL_F32 0x22
+#define V_008DFC_SQ_V_RNDNE_F32 0x23
+#define V_008DFC_SQ_V_FLOOR_F32 0x24
+#define V_008DFC_SQ_V_EXP_F32 0x25
+#define V_008DFC_SQ_V_LOG_CLAMP_F32 0x26
+#define V_008DFC_SQ_V_LOG_F32 0x27
+#define V_008DFC_SQ_V_RCP_CLAMP_F32 0x28
+#define V_008DFC_SQ_V_RCP_LEGACY_F32 0x29
+#define V_008DFC_SQ_V_RCP_F32 0x2A
+#define V_008DFC_SQ_V_RCP_IFLAG_F32 0x2B
+#define V_008DFC_SQ_V_RSQ_CLAMP_F32 0x2C
+#define V_008DFC_SQ_V_RSQ_LEGACY_F32 0x2D
+#define V_008DFC_SQ_V_RSQ_F32 0x2E
+#define V_008DFC_SQ_V_RCP_F64 0x2F
+#define V_008DFC_SQ_V_RCP_CLAMP_F64 0x30
+#define V_008DFC_SQ_V_RSQ_F64 0x31
+#define V_008DFC_SQ_V_RSQ_CLAMP_F64 0x32
+#define V_008DFC_SQ_V_SQRT_F32 0x33
+#define V_008DFC_SQ_V_SQRT_F64 0x34
+#define V_008DFC_SQ_V_SIN_F32 0x35
+#define V_008DFC_SQ_V_COS_F32 0x36
+#define V_008DFC_SQ_V_NOT_B32 0x37
+#define V_008DFC_SQ_V_BFREV_B32 0x38
+#define V_008DFC_SQ_V_FFBH_U32 0x39
+#define V_008DFC_SQ_V_FFBL_B32 0x3A
+#define V_008DFC_SQ_V_FFBH_I32 0x3B
+#define V_008DFC_SQ_V_FREXP_EXP_I32_F64 0x3C
+#define V_008DFC_SQ_V_FREXP_MANT_F64 0x3D
+#define V_008DFC_SQ_V_FRACT_F64 0x3E
+#define V_008DFC_SQ_V_FREXP_EXP_I32_F32 0x3F
+#define V_008DFC_SQ_V_FREXP_MANT_F32 0x40
+#define V_008DFC_SQ_V_CLREXCP 0x41
+#define V_008DFC_SQ_V_MOVRELD_B32 0x42
+#define V_008DFC_SQ_V_MOVRELS_B32 0x43
+#define V_008DFC_SQ_V_MOVRELSD_B32 0x44
+#define S_008DFC_VDST(x) (((x) & 0xFF) << 17)
+#define G_008DFC_VDST(x) (((x) >> 17) & 0xFF)
+#define C_008DFC_VDST 0xFE01FFFF
+#define V_008DFC_SQ_VGPR 0x00
+#define S_008DFC_ENCODING(x) (((x) & 0x7F) << 25)
+#define G_008DFC_ENCODING(x) (((x) >> 25) & 0x7F)
+#define C_008DFC_ENCODING 0x01FFFFFF
+#define V_008DFC_SQ_ENC_VOP1_FIELD 0x3F
+#define R_008DFC_SQ_MIMG_1 0x008DFC
+#define S_008DFC_VADDR(x) (((x) & 0xFF) << 0)
+#define G_008DFC_VADDR(x) (((x) >> 0) & 0xFF)
+#define C_008DFC_VADDR 0xFFFFFF00
+#define V_008DFC_SQ_VGPR 0x00
+#define S_008DFC_VDATA(x) (((x) & 0xFF) << 8)
+#define G_008DFC_VDATA(x) (((x) >> 8) & 0xFF)
+#define C_008DFC_VDATA 0xFFFF00FF
+#define V_008DFC_SQ_VGPR 0x00
+#define S_008DFC_SRSRC(x) (((x) & 0x1F) << 16)
+#define G_008DFC_SRSRC(x) (((x) >> 16) & 0x1F)
+#define C_008DFC_SRSRC 0xFFE0FFFF
+#define S_008DFC_SSAMP(x) (((x) & 0x1F) << 21)
+#define G_008DFC_SSAMP(x) (((x) >> 21) & 0x1F)
+#define C_008DFC_SSAMP 0xFC1FFFFF
+#define R_008DFC_SQ_VOP3_1 0x008DFC
+#define S_008DFC_SRC0(x) (((x) & 0x1FF) << 0)
+#define G_008DFC_SRC0(x) (((x) >> 0) & 0x1FF)
+#define C_008DFC_SRC0 0xFFFFFE00
+#define V_008DFC_SQ_SGPR 0x00
+#define V_008DFC_SQ_VCC_LO 0x6A
+#define V_008DFC_SQ_VCC_HI 0x6B
+#define V_008DFC_SQ_TBA_LO 0x6C
+#define V_008DFC_SQ_TBA_HI 0x6D
+#define V_008DFC_SQ_TMA_LO 0x6E
+#define V_008DFC_SQ_TMA_HI 0x6F
+#define V_008DFC_SQ_TTMP0 0x70
+#define V_008DFC_SQ_TTMP1 0x71
+#define V_008DFC_SQ_TTMP2 0x72
+#define V_008DFC_SQ_TTMP3 0x73
+#define V_008DFC_SQ_TTMP4 0x74
+#define V_008DFC_SQ_TTMP5 0x75
+#define V_008DFC_SQ_TTMP6 0x76
+#define V_008DFC_SQ_TTMP7 0x77
+#define V_008DFC_SQ_TTMP8 0x78
+#define V_008DFC_SQ_TTMP9 0x79
+#define V_008DFC_SQ_TTMP10 0x7A
+#define V_008DFC_SQ_TTMP11 0x7B
+#define V_008DFC_SQ_M0 0x7C
+#define V_008DFC_SQ_EXEC_LO 0x7E
+#define V_008DFC_SQ_EXEC_HI 0x7F
+#define V_008DFC_SQ_SRC_0 0x80
+#define V_008DFC_SQ_SRC_1_INT 0x81
+#define V_008DFC_SQ_SRC_2_INT 0x82
+#define V_008DFC_SQ_SRC_3_INT 0x83
+#define V_008DFC_SQ_SRC_4_INT 0x84
+#define V_008DFC_SQ_SRC_5_INT 0x85
+#define V_008DFC_SQ_SRC_6_INT 0x86
+#define V_008DFC_SQ_SRC_7_INT 0x87
+#define V_008DFC_SQ_SRC_8_INT 0x88
+#define V_008DFC_SQ_SRC_9_INT 0x89
+#define V_008DFC_SQ_SRC_10_INT 0x8A
+#define V_008DFC_SQ_SRC_11_INT 0x8B
+#define V_008DFC_SQ_SRC_12_INT 0x8C
+#define V_008DFC_SQ_SRC_13_INT 0x8D
+#define V_008DFC_SQ_SRC_14_INT 0x8E
+#define V_008DFC_SQ_SRC_15_INT 0x8F
+#define V_008DFC_SQ_SRC_16_INT 0x90
+#define V_008DFC_SQ_SRC_17_INT 0x91
+#define V_008DFC_SQ_SRC_18_INT 0x92
+#define V_008DFC_SQ_SRC_19_INT 0x93
+#define V_008DFC_SQ_SRC_20_INT 0x94
+#define V_008DFC_SQ_SRC_21_INT 0x95
+#define V_008DFC_SQ_SRC_22_INT 0x96
+#define V_008DFC_SQ_SRC_23_INT 0x97
+#define V_008DFC_SQ_SRC_24_INT 0x98
+#define V_008DFC_SQ_SRC_25_INT 0x99
+#define V_008DFC_SQ_SRC_26_INT 0x9A
+#define V_008DFC_SQ_SRC_27_INT 0x9B
+#define V_008DFC_SQ_SRC_28_INT 0x9C
+#define V_008DFC_SQ_SRC_29_INT 0x9D
+#define V_008DFC_SQ_SRC_30_INT 0x9E
+#define V_008DFC_SQ_SRC_31_INT 0x9F
+#define V_008DFC_SQ_SRC_32_INT 0xA0
+#define V_008DFC_SQ_SRC_33_INT 0xA1
+#define V_008DFC_SQ_SRC_34_INT 0xA2
+#define V_008DFC_SQ_SRC_35_INT 0xA3
+#define V_008DFC_SQ_SRC_36_INT 0xA4
+#define V_008DFC_SQ_SRC_37_INT 0xA5
+#define V_008DFC_SQ_SRC_38_INT 0xA6
+#define V_008DFC_SQ_SRC_39_INT 0xA7
+#define V_008DFC_SQ_SRC_40_INT 0xA8
+#define V_008DFC_SQ_SRC_41_INT 0xA9
+#define V_008DFC_SQ_SRC_42_INT 0xAA
+#define V_008DFC_SQ_SRC_43_INT 0xAB
+#define V_008DFC_SQ_SRC_44_INT 0xAC
+#define V_008DFC_SQ_SRC_45_INT 0xAD
+#define V_008DFC_SQ_SRC_46_INT 0xAE
+#define V_008DFC_SQ_SRC_47_INT 0xAF
+#define V_008DFC_SQ_SRC_48_INT 0xB0
+#define V_008DFC_SQ_SRC_49_INT 0xB1
+#define V_008DFC_SQ_SRC_50_INT 0xB2
+#define V_008DFC_SQ_SRC_51_INT 0xB3
+#define V_008DFC_SQ_SRC_52_INT 0xB4
+#define V_008DFC_SQ_SRC_53_INT 0xB5
+#define V_008DFC_SQ_SRC_54_INT 0xB6
+#define V_008DFC_SQ_SRC_55_INT 0xB7
+#define V_008DFC_SQ_SRC_56_INT 0xB8
+#define V_008DFC_SQ_SRC_57_INT 0xB9
+#define V_008DFC_SQ_SRC_58_INT 0xBA
+#define V_008DFC_SQ_SRC_59_INT 0xBB
+#define V_008DFC_SQ_SRC_60_INT 0xBC
+#define V_008DFC_SQ_SRC_61_INT 0xBD
+#define V_008DFC_SQ_SRC_62_INT 0xBE
+#define V_008DFC_SQ_SRC_63_INT 0xBF
+#define V_008DFC_SQ_SRC_64_INT 0xC0
+#define V_008DFC_SQ_SRC_M_1_INT 0xC1
+#define V_008DFC_SQ_SRC_M_2_INT 0xC2
+#define V_008DFC_SQ_SRC_M_3_INT 0xC3
+#define V_008DFC_SQ_SRC_M_4_INT 0xC4
+#define V_008DFC_SQ_SRC_M_5_INT 0xC5
+#define V_008DFC_SQ_SRC_M_6_INT 0xC6
+#define V_008DFC_SQ_SRC_M_7_INT 0xC7
+#define V_008DFC_SQ_SRC_M_8_INT 0xC8
+#define V_008DFC_SQ_SRC_M_9_INT 0xC9
+#define V_008DFC_SQ_SRC_M_10_INT 0xCA
+#define V_008DFC_SQ_SRC_M_11_INT 0xCB
+#define V_008DFC_SQ_SRC_M_12_INT 0xCC
+#define V_008DFC_SQ_SRC_M_13_INT 0xCD
+#define V_008DFC_SQ_SRC_M_14_INT 0xCE
+#define V_008DFC_SQ_SRC_M_15_INT 0xCF
+#define V_008DFC_SQ_SRC_M_16_INT 0xD0
+#define V_008DFC_SQ_SRC_0_5 0xF0
+#define V_008DFC_SQ_SRC_M_0_5 0xF1
+#define V_008DFC_SQ_SRC_1 0xF2
+#define V_008DFC_SQ_SRC_M_1 0xF3
+#define V_008DFC_SQ_SRC_2 0xF4
+#define V_008DFC_SQ_SRC_M_2 0xF5
+#define V_008DFC_SQ_SRC_4 0xF6
+#define V_008DFC_SQ_SRC_M_4 0xF7
+#define V_008DFC_SQ_SRC_VCCZ 0xFB
+#define V_008DFC_SQ_SRC_EXECZ 0xFC
+#define V_008DFC_SQ_SRC_SCC 0xFD
+#define V_008DFC_SQ_SRC_LDS_DIRECT 0xFE
+#define V_008DFC_SQ_SRC_VGPR 0x100
+#define S_008DFC_SRC1(x) (((x) & 0x1FF) << 9)
+#define G_008DFC_SRC1(x) (((x) >> 9) & 0x1FF)
+#define C_008DFC_SRC1 0xFFFC01FF
+#define V_008DFC_SQ_SGPR 0x00
+#define V_008DFC_SQ_VCC_LO 0x6A
+#define V_008DFC_SQ_VCC_HI 0x6B
+#define V_008DFC_SQ_TBA_LO 0x6C
+#define V_008DFC_SQ_TBA_HI 0x6D
+#define V_008DFC_SQ_TMA_LO 0x6E
+#define V_008DFC_SQ_TMA_HI 0x6F
+#define V_008DFC_SQ_TTMP0 0x70
+#define V_008DFC_SQ_TTMP1 0x71
+#define V_008DFC_SQ_TTMP2 0x72
+#define V_008DFC_SQ_TTMP3 0x73
+#define V_008DFC_SQ_TTMP4 0x74
+#define V_008DFC_SQ_TTMP5 0x75
+#define V_008DFC_SQ_TTMP6 0x76
+#define V_008DFC_SQ_TTMP7 0x77
+#define V_008DFC_SQ_TTMP8 0x78
+#define V_008DFC_SQ_TTMP9 0x79
+#define V_008DFC_SQ_TTMP10 0x7A
+#define V_008DFC_SQ_TTMP11 0x7B
+#define V_008DFC_SQ_M0 0x7C
+#define V_008DFC_SQ_EXEC_LO 0x7E
+#define V_008DFC_SQ_EXEC_HI 0x7F
+#define V_008DFC_SQ_SRC_0 0x80
+#define V_008DFC_SQ_SRC_1_INT 0x81
+#define V_008DFC_SQ_SRC_2_INT 0x82
+#define V_008DFC_SQ_SRC_3_INT 0x83
+#define V_008DFC_SQ_SRC_4_INT 0x84
+#define V_008DFC_SQ_SRC_5_INT 0x85
+#define V_008DFC_SQ_SRC_6_INT 0x86
+#define V_008DFC_SQ_SRC_7_INT 0x87
+#define V_008DFC_SQ_SRC_8_INT 0x88
+#define V_008DFC_SQ_SRC_9_INT 0x89
+#define V_008DFC_SQ_SRC_10_INT 0x8A
+#define V_008DFC_SQ_SRC_11_INT 0x8B
+#define V_008DFC_SQ_SRC_12_INT 0x8C
+#define V_008DFC_SQ_SRC_13_INT 0x8D
+#define V_008DFC_SQ_SRC_14_INT 0x8E
+#define V_008DFC_SQ_SRC_15_INT 0x8F
+#define V_008DFC_SQ_SRC_16_INT 0x90
+#define V_008DFC_SQ_SRC_17_INT 0x91
+#define V_008DFC_SQ_SRC_18_INT 0x92
+#define V_008DFC_SQ_SRC_19_INT 0x93
+#define V_008DFC_SQ_SRC_20_INT 0x94
+#define V_008DFC_SQ_SRC_21_INT 0x95
+#define V_008DFC_SQ_SRC_22_INT 0x96
+#define V_008DFC_SQ_SRC_23_INT 0x97
+#define V_008DFC_SQ_SRC_24_INT 0x98
+#define V_008DFC_SQ_SRC_25_INT 0x99
+#define V_008DFC_SQ_SRC_26_INT 0x9A
+#define V_008DFC_SQ_SRC_27_INT 0x9B
+#define V_008DFC_SQ_SRC_28_INT 0x9C
+#define V_008DFC_SQ_SRC_29_INT 0x9D
+#define V_008DFC_SQ_SRC_30_INT 0x9E
+#define V_008DFC_SQ_SRC_31_INT 0x9F
+#define V_008DFC_SQ_SRC_32_INT 0xA0
+#define V_008DFC_SQ_SRC_33_INT 0xA1
+#define V_008DFC_SQ_SRC_34_INT 0xA2
+#define V_008DFC_SQ_SRC_35_INT 0xA3
+#define V_008DFC_SQ_SRC_36_INT 0xA4
+#define V_008DFC_SQ_SRC_37_INT 0xA5
+#define V_008DFC_SQ_SRC_38_INT 0xA6
+#define V_008DFC_SQ_SRC_39_INT 0xA7
+#define V_008DFC_SQ_SRC_40_INT 0xA8
+#define V_008DFC_SQ_SRC_41_INT 0xA9
+#define V_008DFC_SQ_SRC_42_INT 0xAA
+#define V_008DFC_SQ_SRC_43_INT 0xAB
+#define V_008DFC_SQ_SRC_44_INT 0xAC
+#define V_008DFC_SQ_SRC_45_INT 0xAD
+#define V_008DFC_SQ_SRC_46_INT 0xAE
+#define V_008DFC_SQ_SRC_47_INT 0xAF
+#define V_008DFC_SQ_SRC_48_INT 0xB0
+#define V_008DFC_SQ_SRC_49_INT 0xB1
+#define V_008DFC_SQ_SRC_50_INT 0xB2
+#define V_008DFC_SQ_SRC_51_INT 0xB3
+#define V_008DFC_SQ_SRC_52_INT 0xB4
+#define V_008DFC_SQ_SRC_53_INT 0xB5
+#define V_008DFC_SQ_SRC_54_INT 0xB6
+#define V_008DFC_SQ_SRC_55_INT 0xB7
+#define V_008DFC_SQ_SRC_56_INT 0xB8
+#define V_008DFC_SQ_SRC_57_INT 0xB9
+#define V_008DFC_SQ_SRC_58_INT 0xBA
+#define V_008DFC_SQ_SRC_59_INT 0xBB
+#define V_008DFC_SQ_SRC_60_INT 0xBC
+#define V_008DFC_SQ_SRC_61_INT 0xBD
+#define V_008DFC_SQ_SRC_62_INT 0xBE
+#define V_008DFC_SQ_SRC_63_INT 0xBF
+#define V_008DFC_SQ_SRC_64_INT 0xC0
+#define V_008DFC_SQ_SRC_M_1_INT 0xC1
+#define V_008DFC_SQ_SRC_M_2_INT 0xC2
+#define V_008DFC_SQ_SRC_M_3_INT 0xC3
+#define V_008DFC_SQ_SRC_M_4_INT 0xC4
+#define V_008DFC_SQ_SRC_M_5_INT 0xC5
+#define V_008DFC_SQ_SRC_M_6_INT 0xC6
+#define V_008DFC_SQ_SRC_M_7_INT 0xC7
+#define V_008DFC_SQ_SRC_M_8_INT 0xC8
+#define V_008DFC_SQ_SRC_M_9_INT 0xC9
+#define V_008DFC_SQ_SRC_M_10_INT 0xCA
+#define V_008DFC_SQ_SRC_M_11_INT 0xCB
+#define V_008DFC_SQ_SRC_M_12_INT 0xCC
+#define V_008DFC_SQ_SRC_M_13_INT 0xCD
+#define V_008DFC_SQ_SRC_M_14_INT 0xCE
+#define V_008DFC_SQ_SRC_M_15_INT 0xCF
+#define V_008DFC_SQ_SRC_M_16_INT 0xD0
+#define V_008DFC_SQ_SRC_0_5 0xF0
+#define V_008DFC_SQ_SRC_M_0_5 0xF1
+#define V_008DFC_SQ_SRC_1 0xF2
+#define V_008DFC_SQ_SRC_M_1 0xF3
+#define V_008DFC_SQ_SRC_2 0xF4
+#define V_008DFC_SQ_SRC_M_2 0xF5
+#define V_008DFC_SQ_SRC_4 0xF6
+#define V_008DFC_SQ_SRC_M_4 0xF7
+#define V_008DFC_SQ_SRC_VCCZ 0xFB
+#define V_008DFC_SQ_SRC_EXECZ 0xFC
+#define V_008DFC_SQ_SRC_SCC 0xFD
+#define V_008DFC_SQ_SRC_LDS_DIRECT 0xFE
+#define V_008DFC_SQ_SRC_VGPR 0x100
+#define S_008DFC_SRC2(x) (((x) & 0x1FF) << 18)
+#define G_008DFC_SRC2(x) (((x) >> 18) & 0x1FF)
+#define C_008DFC_SRC2 0xF803FFFF
+#define V_008DFC_SQ_SGPR 0x00
+#define V_008DFC_SQ_VCC_LO 0x6A
+#define V_008DFC_SQ_VCC_HI 0x6B
+#define V_008DFC_SQ_TBA_LO 0x6C
+#define V_008DFC_SQ_TBA_HI 0x6D
+#define V_008DFC_SQ_TMA_LO 0x6E
+#define V_008DFC_SQ_TMA_HI 0x6F
+#define V_008DFC_SQ_TTMP0 0x70
+#define V_008DFC_SQ_TTMP1 0x71
+#define V_008DFC_SQ_TTMP2 0x72
+#define V_008DFC_SQ_TTMP3 0x73
+#define V_008DFC_SQ_TTMP4 0x74
+#define V_008DFC_SQ_TTMP5 0x75
+#define V_008DFC_SQ_TTMP6 0x76
+#define V_008DFC_SQ_TTMP7 0x77
+#define V_008DFC_SQ_TTMP8 0x78
+#define V_008DFC_SQ_TTMP9 0x79
+#define V_008DFC_SQ_TTMP10 0x7A
+#define V_008DFC_SQ_TTMP11 0x7B
+#define V_008DFC_SQ_M0 0x7C
+#define V_008DFC_SQ_EXEC_LO 0x7E
+#define V_008DFC_SQ_EXEC_HI 0x7F
+#define V_008DFC_SQ_SRC_0 0x80
+#define V_008DFC_SQ_SRC_1_INT 0x81
+#define V_008DFC_SQ_SRC_2_INT 0x82
+#define V_008DFC_SQ_SRC_3_INT 0x83
+#define V_008DFC_SQ_SRC_4_INT 0x84
+#define V_008DFC_SQ_SRC_5_INT 0x85
+#define V_008DFC_SQ_SRC_6_INT 0x86
+#define V_008DFC_SQ_SRC_7_INT 0x87
+#define V_008DFC_SQ_SRC_8_INT 0x88
+#define V_008DFC_SQ_SRC_9_INT 0x89
+#define V_008DFC_SQ_SRC_10_INT 0x8A
+#define V_008DFC_SQ_SRC_11_INT 0x8B
+#define V_008DFC_SQ_SRC_12_INT 0x8C
+#define V_008DFC_SQ_SRC_13_INT 0x8D
+#define V_008DFC_SQ_SRC_14_INT 0x8E
+#define V_008DFC_SQ_SRC_15_INT 0x8F
+#define V_008DFC_SQ_SRC_16_INT 0x90
+#define V_008DFC_SQ_SRC_17_INT 0x91
+#define V_008DFC_SQ_SRC_18_INT 0x92
+#define V_008DFC_SQ_SRC_19_INT 0x93
+#define V_008DFC_SQ_SRC_20_INT 0x94
+#define V_008DFC_SQ_SRC_21_INT 0x95
+#define V_008DFC_SQ_SRC_22_INT 0x96
+#define V_008DFC_SQ_SRC_23_INT 0x97
+#define V_008DFC_SQ_SRC_24_INT 0x98
+#define V_008DFC_SQ_SRC_25_INT 0x99
+#define V_008DFC_SQ_SRC_26_INT 0x9A
+#define V_008DFC_SQ_SRC_27_INT 0x9B
+#define V_008DFC_SQ_SRC_28_INT 0x9C
+#define V_008DFC_SQ_SRC_29_INT 0x9D
+#define V_008DFC_SQ_SRC_30_INT 0x9E
+#define V_008DFC_SQ_SRC_31_INT 0x9F
+#define V_008DFC_SQ_SRC_32_INT 0xA0
+#define V_008DFC_SQ_SRC_33_INT 0xA1
+#define V_008DFC_SQ_SRC_34_INT 0xA2
+#define V_008DFC_SQ_SRC_35_INT 0xA3
+#define V_008DFC_SQ_SRC_36_INT 0xA4
+#define V_008DFC_SQ_SRC_37_INT 0xA5
+#define V_008DFC_SQ_SRC_38_INT 0xA6
+#define V_008DFC_SQ_SRC_39_INT 0xA7
+#define V_008DFC_SQ_SRC_40_INT 0xA8
+#define V_008DFC_SQ_SRC_41_INT 0xA9
+#define V_008DFC_SQ_SRC_42_INT 0xAA
+#define V_008DFC_SQ_SRC_43_INT 0xAB
+#define V_008DFC_SQ_SRC_44_INT 0xAC
+#define V_008DFC_SQ_SRC_45_INT 0xAD
+#define V_008DFC_SQ_SRC_46_INT 0xAE
+#define V_008DFC_SQ_SRC_47_INT 0xAF
+#define V_008DFC_SQ_SRC_48_INT 0xB0
+#define V_008DFC_SQ_SRC_49_INT 0xB1
+#define V_008DFC_SQ_SRC_50_INT 0xB2
+#define V_008DFC_SQ_SRC_51_INT 0xB3
+#define V_008DFC_SQ_SRC_52_INT 0xB4
+#define V_008DFC_SQ_SRC_53_INT 0xB5
+#define V_008DFC_SQ_SRC_54_INT 0xB6
+#define V_008DFC_SQ_SRC_55_INT 0xB7
+#define V_008DFC_SQ_SRC_56_INT 0xB8
+#define V_008DFC_SQ_SRC_57_INT 0xB9
+#define V_008DFC_SQ_SRC_58_INT 0xBA
+#define V_008DFC_SQ_SRC_59_INT 0xBB
+#define V_008DFC_SQ_SRC_60_INT 0xBC
+#define V_008DFC_SQ_SRC_61_INT 0xBD
+#define V_008DFC_SQ_SRC_62_INT 0xBE
+#define V_008DFC_SQ_SRC_63_INT 0xBF
+#define V_008DFC_SQ_SRC_64_INT 0xC0
+#define V_008DFC_SQ_SRC_M_1_INT 0xC1
+#define V_008DFC_SQ_SRC_M_2_INT 0xC2
+#define V_008DFC_SQ_SRC_M_3_INT 0xC3
+#define V_008DFC_SQ_SRC_M_4_INT 0xC4
+#define V_008DFC_SQ_SRC_M_5_INT 0xC5
+#define V_008DFC_SQ_SRC_M_6_INT 0xC6
+#define V_008DFC_SQ_SRC_M_7_INT 0xC7
+#define V_008DFC_SQ_SRC_M_8_INT 0xC8
+#define V_008DFC_SQ_SRC_M_9_INT 0xC9
+#define V_008DFC_SQ_SRC_M_10_INT 0xCA
+#define V_008DFC_SQ_SRC_M_11_INT 0xCB
+#define V_008DFC_SQ_SRC_M_12_INT 0xCC
+#define V_008DFC_SQ_SRC_M_13_INT 0xCD
+#define V_008DFC_SQ_SRC_M_14_INT 0xCE
+#define V_008DFC_SQ_SRC_M_15_INT 0xCF
+#define V_008DFC_SQ_SRC_M_16_INT 0xD0
+#define V_008DFC_SQ_SRC_0_5 0xF0
+#define V_008DFC_SQ_SRC_M_0_5 0xF1
+#define V_008DFC_SQ_SRC_1 0xF2
+#define V_008DFC_SQ_SRC_M_1 0xF3
+#define V_008DFC_SQ_SRC_2 0xF4
+#define V_008DFC_SQ_SRC_M_2 0xF5
+#define V_008DFC_SQ_SRC_4 0xF6
+#define V_008DFC_SQ_SRC_M_4 0xF7
+#define V_008DFC_SQ_SRC_VCCZ 0xFB
+#define V_008DFC_SQ_SRC_EXECZ 0xFC
+#define V_008DFC_SQ_SRC_SCC 0xFD
+#define V_008DFC_SQ_SRC_LDS_DIRECT 0xFE
+#define V_008DFC_SQ_SRC_VGPR 0x100
+#define S_008DFC_OMOD(x) (((x) & 0x03) << 27)
+#define G_008DFC_OMOD(x) (((x) >> 27) & 0x03)
+#define C_008DFC_OMOD 0xE7FFFFFF
+#define V_008DFC_SQ_OMOD_OFF 0x00
+#define V_008DFC_SQ_OMOD_M2 0x01
+#define V_008DFC_SQ_OMOD_M4 0x02
+#define V_008DFC_SQ_OMOD_D2 0x03
+#define S_008DFC_NEG(x) (((x) & 0x07) << 29)
+#define G_008DFC_NEG(x) (((x) >> 29) & 0x07)
+#define C_008DFC_NEG 0x1FFFFFFF
+#define R_008DFC_SQ_MUBUF_1 0x008DFC
+#define S_008DFC_VADDR(x) (((x) & 0xFF) << 0)
+#define G_008DFC_VADDR(x) (((x) >> 0) & 0xFF)
+#define C_008DFC_VADDR 0xFFFFFF00
+#define V_008DFC_SQ_VGPR 0x00
+#define S_008DFC_VDATA(x) (((x) & 0xFF) << 8)
+#define G_008DFC_VDATA(x) (((x) >> 8) & 0xFF)
+#define C_008DFC_VDATA 0xFFFF00FF
+#define V_008DFC_SQ_VGPR 0x00
+#define S_008DFC_SRSRC(x) (((x) & 0x1F) << 16)
+#define G_008DFC_SRSRC(x) (((x) >> 16) & 0x1F)
+#define C_008DFC_SRSRC 0xFFE0FFFF
+#define S_008DFC_SLC(x) (((x) & 0x1) << 22)
+#define G_008DFC_SLC(x) (((x) >> 22) & 0x1)
+#define C_008DFC_SLC 0xFFBFFFFF
+#define S_008DFC_TFE(x) (((x) & 0x1) << 23)
+#define G_008DFC_TFE(x) (((x) >> 23) & 0x1)
+#define C_008DFC_TFE 0xFF7FFFFF
+#define S_008DFC_SOFFSET(x) (((x) & 0xFF) << 24)
+#define G_008DFC_SOFFSET(x) (((x) >> 24) & 0xFF)
+#define C_008DFC_SOFFSET 0x00FFFFFF
+#define V_008DFC_SQ_SGPR 0x00
+#define V_008DFC_SQ_VCC_LO 0x6A
+#define V_008DFC_SQ_VCC_HI 0x6B
+#define V_008DFC_SQ_TBA_LO 0x6C
+#define V_008DFC_SQ_TBA_HI 0x6D
+#define V_008DFC_SQ_TMA_LO 0x6E
+#define V_008DFC_SQ_TMA_HI 0x6F
+#define V_008DFC_SQ_TTMP0 0x70
+#define V_008DFC_SQ_TTMP1 0x71
+#define V_008DFC_SQ_TTMP2 0x72
+#define V_008DFC_SQ_TTMP3 0x73
+#define V_008DFC_SQ_TTMP4 0x74
+#define V_008DFC_SQ_TTMP5 0x75
+#define V_008DFC_SQ_TTMP6 0x76
+#define V_008DFC_SQ_TTMP7 0x77
+#define V_008DFC_SQ_TTMP8 0x78
+#define V_008DFC_SQ_TTMP9 0x79
+#define V_008DFC_SQ_TTMP10 0x7A
+#define V_008DFC_SQ_TTMP11 0x7B
+#define V_008DFC_SQ_M0 0x7C
+#define V_008DFC_SQ_EXEC_LO 0x7E
+#define V_008DFC_SQ_EXEC_HI 0x7F
+#define V_008DFC_SQ_SRC_0 0x80
+#define V_008DFC_SQ_SRC_1_INT 0x81
+#define V_008DFC_SQ_SRC_2_INT 0x82
+#define V_008DFC_SQ_SRC_3_INT 0x83
+#define V_008DFC_SQ_SRC_4_INT 0x84
+#define V_008DFC_SQ_SRC_5_INT 0x85
+#define V_008DFC_SQ_SRC_6_INT 0x86
+#define V_008DFC_SQ_SRC_7_INT 0x87
+#define V_008DFC_SQ_SRC_8_INT 0x88
+#define V_008DFC_SQ_SRC_9_INT 0x89
+#define V_008DFC_SQ_SRC_10_INT 0x8A
+#define V_008DFC_SQ_SRC_11_INT 0x8B
+#define V_008DFC_SQ_SRC_12_INT 0x8C
+#define V_008DFC_SQ_SRC_13_INT 0x8D
+#define V_008DFC_SQ_SRC_14_INT 0x8E
+#define V_008DFC_SQ_SRC_15_INT 0x8F
+#define V_008DFC_SQ_SRC_16_INT 0x90
+#define V_008DFC_SQ_SRC_17_INT 0x91
+#define V_008DFC_SQ_SRC_18_INT 0x92
+#define V_008DFC_SQ_SRC_19_INT 0x93
+#define V_008DFC_SQ_SRC_20_INT 0x94
+#define V_008DFC_SQ_SRC_21_INT 0x95
+#define V_008DFC_SQ_SRC_22_INT 0x96
+#define V_008DFC_SQ_SRC_23_INT 0x97
+#define V_008DFC_SQ_SRC_24_INT 0x98
+#define V_008DFC_SQ_SRC_25_INT 0x99
+#define V_008DFC_SQ_SRC_26_INT 0x9A
+#define V_008DFC_SQ_SRC_27_INT 0x9B
+#define V_008DFC_SQ_SRC_28_INT 0x9C
+#define V_008DFC_SQ_SRC_29_INT 0x9D
+#define V_008DFC_SQ_SRC_30_INT 0x9E
+#define V_008DFC_SQ_SRC_31_INT 0x9F
+#define V_008DFC_SQ_SRC_32_INT 0xA0
+#define V_008DFC_SQ_SRC_33_INT 0xA1
+#define V_008DFC_SQ_SRC_34_INT 0xA2
+#define V_008DFC_SQ_SRC_35_INT 0xA3
+#define V_008DFC_SQ_SRC_36_INT 0xA4
+#define V_008DFC_SQ_SRC_37_INT 0xA5
+#define V_008DFC_SQ_SRC_38_INT 0xA6
+#define V_008DFC_SQ_SRC_39_INT 0xA7
+#define V_008DFC_SQ_SRC_40_INT 0xA8
+#define V_008DFC_SQ_SRC_41_INT 0xA9
+#define V_008DFC_SQ_SRC_42_INT 0xAA
+#define V_008DFC_SQ_SRC_43_INT 0xAB
+#define V_008DFC_SQ_SRC_44_INT 0xAC
+#define V_008DFC_SQ_SRC_45_INT 0xAD
+#define V_008DFC_SQ_SRC_46_INT 0xAE
+#define V_008DFC_SQ_SRC_47_INT 0xAF
+#define V_008DFC_SQ_SRC_48_INT 0xB0
+#define V_008DFC_SQ_SRC_49_INT 0xB1
+#define V_008DFC_SQ_SRC_50_INT 0xB2
+#define V_008DFC_SQ_SRC_51_INT 0xB3
+#define V_008DFC_SQ_SRC_52_INT 0xB4
+#define V_008DFC_SQ_SRC_53_INT 0xB5
+#define V_008DFC_SQ_SRC_54_INT 0xB6
+#define V_008DFC_SQ_SRC_55_INT 0xB7
+#define V_008DFC_SQ_SRC_56_INT 0xB8
+#define V_008DFC_SQ_SRC_57_INT 0xB9
+#define V_008DFC_SQ_SRC_58_INT 0xBA
+#define V_008DFC_SQ_SRC_59_INT 0xBB
+#define V_008DFC_SQ_SRC_60_INT 0xBC
+#define V_008DFC_SQ_SRC_61_INT 0xBD
+#define V_008DFC_SQ_SRC_62_INT 0xBE
+#define V_008DFC_SQ_SRC_63_INT 0xBF
+#define V_008DFC_SQ_SRC_64_INT 0xC0
+#define V_008DFC_SQ_SRC_M_1_INT 0xC1
+#define V_008DFC_SQ_SRC_M_2_INT 0xC2
+#define V_008DFC_SQ_SRC_M_3_INT 0xC3
+#define V_008DFC_SQ_SRC_M_4_INT 0xC4
+#define V_008DFC_SQ_SRC_M_5_INT 0xC5
+#define V_008DFC_SQ_SRC_M_6_INT 0xC6
+#define V_008DFC_SQ_SRC_M_7_INT 0xC7
+#define V_008DFC_SQ_SRC_M_8_INT 0xC8
+#define V_008DFC_SQ_SRC_M_9_INT 0xC9
+#define V_008DFC_SQ_SRC_M_10_INT 0xCA
+#define V_008DFC_SQ_SRC_M_11_INT 0xCB
+#define V_008DFC_SQ_SRC_M_12_INT 0xCC
+#define V_008DFC_SQ_SRC_M_13_INT 0xCD
+#define V_008DFC_SQ_SRC_M_14_INT 0xCE
+#define V_008DFC_SQ_SRC_M_15_INT 0xCF
+#define V_008DFC_SQ_SRC_M_16_INT 0xD0
+#define V_008DFC_SQ_SRC_0_5 0xF0
+#define V_008DFC_SQ_SRC_M_0_5 0xF1
+#define V_008DFC_SQ_SRC_1 0xF2
+#define V_008DFC_SQ_SRC_M_1 0xF3
+#define V_008DFC_SQ_SRC_2 0xF4
+#define V_008DFC_SQ_SRC_M_2 0xF5
+#define V_008DFC_SQ_SRC_4 0xF6
+#define V_008DFC_SQ_SRC_M_4 0xF7
+#define V_008DFC_SQ_SRC_VCCZ 0xFB
+#define V_008DFC_SQ_SRC_EXECZ 0xFC
+#define V_008DFC_SQ_SRC_SCC 0xFD
+#define V_008DFC_SQ_SRC_LDS_DIRECT 0xFE
+#define R_008DFC_SQ_DS_0 0x008DFC
+#define S_008DFC_OFFSET0(x) (((x) & 0xFF) << 0)
+#define G_008DFC_OFFSET0(x) (((x) >> 0) & 0xFF)
+#define C_008DFC_OFFSET0 0xFFFFFF00
+#define S_008DFC_OFFSET1(x) (((x) & 0xFF) << 8)
+#define G_008DFC_OFFSET1(x) (((x) >> 8) & 0xFF)
+#define C_008DFC_OFFSET1 0xFFFF00FF
+#define S_008DFC_GDS(x) (((x) & 0x1) << 17)
+#define G_008DFC_GDS(x) (((x) >> 17) & 0x1)
+#define C_008DFC_GDS 0xFFFDFFFF
+#define S_008DFC_OP(x) (((x) & 0xFF) << 18)
+#define G_008DFC_OP(x) (((x) >> 18) & 0xFF)
+#define C_008DFC_OP 0xFC03FFFF
+#define V_008DFC_SQ_DS_ADD_U32 0x00
+#define V_008DFC_SQ_DS_SUB_U32 0x01
+#define V_008DFC_SQ_DS_RSUB_U32 0x02
+#define V_008DFC_SQ_DS_INC_U32 0x03
+#define V_008DFC_SQ_DS_DEC_U32 0x04
+#define V_008DFC_SQ_DS_MIN_I32 0x05
+#define V_008DFC_SQ_DS_MAX_I32 0x06
+#define V_008DFC_SQ_DS_MIN_U32 0x07
+#define V_008DFC_SQ_DS_MAX_U32 0x08
+#define V_008DFC_SQ_DS_AND_B32 0x09
+#define V_008DFC_SQ_DS_OR_B32 0x0A
+#define V_008DFC_SQ_DS_XOR_B32 0x0B
+#define V_008DFC_SQ_DS_MSKOR_B32 0x0C
+#define V_008DFC_SQ_DS_WRITE_B32 0x0D
+#define V_008DFC_SQ_DS_WRITE2_B32 0x0E
+#define V_008DFC_SQ_DS_WRITE2ST64_B32 0x0F
+#define V_008DFC_SQ_DS_CMPST_B32 0x10
+#define V_008DFC_SQ_DS_CMPST_F32 0x11
+#define V_008DFC_SQ_DS_MIN_F32 0x12
+#define V_008DFC_SQ_DS_MAX_F32 0x13
+#define V_008DFC_SQ_DS_GWS_INIT 0x19
+#define V_008DFC_SQ_DS_GWS_SEMA_V 0x1A
+#define V_008DFC_SQ_DS_GWS_SEMA_BR 0x1B
+#define V_008DFC_SQ_DS_GWS_SEMA_P 0x1C
+#define V_008DFC_SQ_DS_GWS_BARRIER 0x1D
+#define V_008DFC_SQ_DS_WRITE_B8 0x1E
+#define V_008DFC_SQ_DS_WRITE_B16 0x1F
+#define V_008DFC_SQ_DS_ADD_RTN_U32 0x20
+#define V_008DFC_SQ_DS_SUB_RTN_U32 0x21
+#define V_008DFC_SQ_DS_RSUB_RTN_U32 0x22
+#define V_008DFC_SQ_DS_INC_RTN_U32 0x23
+#define V_008DFC_SQ_DS_DEC_RTN_U32 0x24
+#define V_008DFC_SQ_DS_MIN_RTN_I32 0x25
+#define V_008DFC_SQ_DS_MAX_RTN_I32 0x26
+#define V_008DFC_SQ_DS_MIN_RTN_U32 0x27
+#define V_008DFC_SQ_DS_MAX_RTN_U32 0x28
+#define V_008DFC_SQ_DS_AND_RTN_B32 0x29
+#define V_008DFC_SQ_DS_OR_RTN_B32 0x2A
+#define V_008DFC_SQ_DS_XOR_RTN_B32 0x2B
+#define V_008DFC_SQ_DS_MSKOR_RTN_B32 0x2C
+#define V_008DFC_SQ_DS_WRXCHG_RTN_B32 0x2D
+#define V_008DFC_SQ_DS_WRXCHG2_RTN_B32 0x2E
+#define V_008DFC_SQ_DS_WRXCHG2ST64_RTN_B32 0x2F
+#define V_008DFC_SQ_DS_CMPST_RTN_B32 0x30
+#define V_008DFC_SQ_DS_CMPST_RTN_F32 0x31
+#define V_008DFC_SQ_DS_MIN_RTN_F32 0x32
+#define V_008DFC_SQ_DS_MAX_RTN_F32 0x33
+#define V_008DFC_SQ_DS_SWIZZLE_B32 0x35
+#define V_008DFC_SQ_DS_READ_B32 0x36
+#define V_008DFC_SQ_DS_READ2_B32 0x37
+#define V_008DFC_SQ_DS_READ2ST64_B32 0x38
+#define V_008DFC_SQ_DS_READ_I8 0x39
+#define V_008DFC_SQ_DS_READ_U8 0x3A
+#define V_008DFC_SQ_DS_READ_I16 0x3B
+#define V_008DFC_SQ_DS_READ_U16 0x3C
+#define V_008DFC_SQ_DS_CONSUME 0x3D
+#define V_008DFC_SQ_DS_APPEND 0x3E
+#define V_008DFC_SQ_DS_ORDERED_COUNT 0x3F
+#define V_008DFC_SQ_DS_ADD_U64 0x40
+#define V_008DFC_SQ_DS_SUB_U64 0x41
+#define V_008DFC_SQ_DS_RSUB_U64 0x42
+#define V_008DFC_SQ_DS_INC_U64 0x43
+#define V_008DFC_SQ_DS_DEC_U64 0x44
+#define V_008DFC_SQ_DS_MIN_I64 0x45
+#define V_008DFC_SQ_DS_MAX_I64 0x46
+#define V_008DFC_SQ_DS_MIN_U64 0x47
+#define V_008DFC_SQ_DS_MAX_U64 0x48
+#define V_008DFC_SQ_DS_AND_B64 0x49
+#define V_008DFC_SQ_DS_OR_B64 0x4A
+#define V_008DFC_SQ_DS_XOR_B64 0x4B
+#define V_008DFC_SQ_DS_MSKOR_B64 0x4C
+#define V_008DFC_SQ_DS_WRITE_B64 0x4D
+#define V_008DFC_SQ_DS_WRITE2_B64 0x4E
+#define V_008DFC_SQ_DS_WRITE2ST64_B64 0x4F
+#define V_008DFC_SQ_DS_CMPST_B64 0x50
+#define V_008DFC_SQ_DS_CMPST_F64 0x51
+#define V_008DFC_SQ_DS_MIN_F64 0x52
+#define V_008DFC_SQ_DS_MAX_F64 0x53
+#define V_008DFC_SQ_DS_ADD_RTN_U64 0x60
+#define V_008DFC_SQ_DS_SUB_RTN_U64 0x61
+#define V_008DFC_SQ_DS_RSUB_RTN_U64 0x62
+#define V_008DFC_SQ_DS_INC_RTN_U64 0x63
+#define V_008DFC_SQ_DS_DEC_RTN_U64 0x64
+#define V_008DFC_SQ_DS_MIN_RTN_I64 0x65
+#define V_008DFC_SQ_DS_MAX_RTN_I64 0x66
+#define V_008DFC_SQ_DS_MIN_RTN_U64 0x67
+#define V_008DFC_SQ_DS_MAX_RTN_U64 0x68
+#define V_008DFC_SQ_DS_AND_RTN_B64 0x69
+#define V_008DFC_SQ_DS_OR_RTN_B64 0x6A
+#define V_008DFC_SQ_DS_XOR_RTN_B64 0x6B
+#define V_008DFC_SQ_DS_MSKOR_RTN_B64 0x6C
+#define V_008DFC_SQ_DS_WRXCHG_RTN_B64 0x6D
+#define V_008DFC_SQ_DS_WRXCHG2_RTN_B64 0x6E
+#define V_008DFC_SQ_DS_WRXCHG2ST64_RTN_B64 0x6F
+#define V_008DFC_SQ_DS_CMPST_RTN_B64 0x70
+#define V_008DFC_SQ_DS_CMPST_RTN_F64 0x71
+#define V_008DFC_SQ_DS_MIN_RTN_F64 0x72
+#define V_008DFC_SQ_DS_MAX_RTN_F64 0x73
+#define V_008DFC_SQ_DS_READ_B64 0x76
+#define V_008DFC_SQ_DS_READ2_B64 0x77
+#define V_008DFC_SQ_DS_READ2ST64_B64 0x78
+#define V_008DFC_SQ_DS_ADD_SRC2_U32 0x80
+#define V_008DFC_SQ_DS_SUB_SRC2_U32 0x81
+#define V_008DFC_SQ_DS_RSUB_SRC2_U32 0x82
+#define V_008DFC_SQ_DS_INC_SRC2_U32 0x83
+#define V_008DFC_SQ_DS_DEC_SRC2_U32 0x84
+#define V_008DFC_SQ_DS_MIN_SRC2_I32 0x85
+#define V_008DFC_SQ_DS_MAX_SRC2_I32 0x86
+#define V_008DFC_SQ_DS_MIN_SRC2_U32 0x87
+#define V_008DFC_SQ_DS_MAX_SRC2_U32 0x88
+#define V_008DFC_SQ_DS_AND_SRC2_B32 0x89
+#define V_008DFC_SQ_DS_OR_SRC2_B32 0x8A
+#define V_008DFC_SQ_DS_XOR_SRC2_B32 0x8B
+#define V_008DFC_SQ_DS_WRITE_SRC2_B32 0x8D
+#define V_008DFC_SQ_DS_MIN_SRC2_F32 0x92
+#define V_008DFC_SQ_DS_MAX_SRC2_F32 0x93
+#define V_008DFC_SQ_DS_ADD_SRC2_U64 0xC0
+#define V_008DFC_SQ_DS_SUB_SRC2_U64 0xC1
+#define V_008DFC_SQ_DS_RSUB_SRC2_U64 0xC2
+#define V_008DFC_SQ_DS_INC_SRC2_U64 0xC3
+#define V_008DFC_SQ_DS_DEC_SRC2_U64 0xC4
+#define V_008DFC_SQ_DS_MIN_SRC2_I64 0xC5
+#define V_008DFC_SQ_DS_MAX_SRC2_I64 0xC6
+#define V_008DFC_SQ_DS_MIN_SRC2_U64 0xC7
+#define V_008DFC_SQ_DS_MAX_SRC2_U64 0xC8
+#define V_008DFC_SQ_DS_AND_SRC2_B64 0xC9
+#define V_008DFC_SQ_DS_OR_SRC2_B64 0xCA
+#define V_008DFC_SQ_DS_XOR_SRC2_B64 0xCB
+#define V_008DFC_SQ_DS_WRITE_SRC2_B64 0xCD
+#define V_008DFC_SQ_DS_MIN_SRC2_F64 0xD2
+#define V_008DFC_SQ_DS_MAX_SRC2_F64 0xD3
+#define S_008DFC_ENCODING(x) (((x) & 0x3F) << 26)
+#define G_008DFC_ENCODING(x) (((x) >> 26) & 0x3F)
+#define C_008DFC_ENCODING 0x03FFFFFF
+#define V_008DFC_SQ_ENC_DS_FIELD 0x36
+#define R_008DFC_SQ_SOPC 0x008DFC
+#define S_008DFC_SSRC0(x) (((x) & 0xFF) << 0)
+#define G_008DFC_SSRC0(x) (((x) >> 0) & 0xFF)
+#define C_008DFC_SSRC0 0xFFFFFF00
+#define V_008DFC_SQ_SGPR 0x00
+#define V_008DFC_SQ_VCC_LO 0x6A
+#define V_008DFC_SQ_VCC_HI 0x6B
+#define V_008DFC_SQ_TBA_LO 0x6C
+#define V_008DFC_SQ_TBA_HI 0x6D
+#define V_008DFC_SQ_TMA_LO 0x6E
+#define V_008DFC_SQ_TMA_HI 0x6F
+#define V_008DFC_SQ_TTMP0 0x70
+#define V_008DFC_SQ_TTMP1 0x71
+#define V_008DFC_SQ_TTMP2 0x72
+#define V_008DFC_SQ_TTMP3 0x73
+#define V_008DFC_SQ_TTMP4 0x74
+#define V_008DFC_SQ_TTMP5 0x75
+#define V_008DFC_SQ_TTMP6 0x76
+#define V_008DFC_SQ_TTMP7 0x77
+#define V_008DFC_SQ_TTMP8 0x78
+#define V_008DFC_SQ_TTMP9 0x79
+#define V_008DFC_SQ_TTMP10 0x7A
+#define V_008DFC_SQ_TTMP11 0x7B
+#define V_008DFC_SQ_M0 0x7C
+#define V_008DFC_SQ_EXEC_LO 0x7E
+#define V_008DFC_SQ_EXEC_HI 0x7F
+#define V_008DFC_SQ_SRC_0 0x80
+#define V_008DFC_SQ_SRC_1_INT 0x81
+#define V_008DFC_SQ_SRC_2_INT 0x82
+#define V_008DFC_SQ_SRC_3_INT 0x83
+#define V_008DFC_SQ_SRC_4_INT 0x84
+#define V_008DFC_SQ_SRC_5_INT 0x85
+#define V_008DFC_SQ_SRC_6_INT 0x86
+#define V_008DFC_SQ_SRC_7_INT 0x87
+#define V_008DFC_SQ_SRC_8_INT 0x88
+#define V_008DFC_SQ_SRC_9_INT 0x89
+#define V_008DFC_SQ_SRC_10_INT 0x8A
+#define V_008DFC_SQ_SRC_11_INT 0x8B
+#define V_008DFC_SQ_SRC_12_INT 0x8C
+#define V_008DFC_SQ_SRC_13_INT 0x8D
+#define V_008DFC_SQ_SRC_14_INT 0x8E
+#define V_008DFC_SQ_SRC_15_INT 0x8F
+#define V_008DFC_SQ_SRC_16_INT 0x90
+#define V_008DFC_SQ_SRC_17_INT 0x91
+#define V_008DFC_SQ_SRC_18_INT 0x92
+#define V_008DFC_SQ_SRC_19_INT 0x93
+#define V_008DFC_SQ_SRC_20_INT 0x94
+#define V_008DFC_SQ_SRC_21_INT 0x95
+#define V_008DFC_SQ_SRC_22_INT 0x96
+#define V_008DFC_SQ_SRC_23_INT 0x97
+#define V_008DFC_SQ_SRC_24_INT 0x98
+#define V_008DFC_SQ_SRC_25_INT 0x99
+#define V_008DFC_SQ_SRC_26_INT 0x9A
+#define V_008DFC_SQ_SRC_27_INT 0x9B
+#define V_008DFC_SQ_SRC_28_INT 0x9C
+#define V_008DFC_SQ_SRC_29_INT 0x9D
+#define V_008DFC_SQ_SRC_30_INT 0x9E
+#define V_008DFC_SQ_SRC_31_INT 0x9F
+#define V_008DFC_SQ_SRC_32_INT 0xA0
+#define V_008DFC_SQ_SRC_33_INT 0xA1
+#define V_008DFC_SQ_SRC_34_INT 0xA2
+#define V_008DFC_SQ_SRC_35_INT 0xA3
+#define V_008DFC_SQ_SRC_36_INT 0xA4
+#define V_008DFC_SQ_SRC_37_INT 0xA5
+#define V_008DFC_SQ_SRC_38_INT 0xA6
+#define V_008DFC_SQ_SRC_39_INT 0xA7
+#define V_008DFC_SQ_SRC_40_INT 0xA8
+#define V_008DFC_SQ_SRC_41_INT 0xA9
+#define V_008DFC_SQ_SRC_42_INT 0xAA
+#define V_008DFC_SQ_SRC_43_INT 0xAB
+#define V_008DFC_SQ_SRC_44_INT 0xAC
+#define V_008DFC_SQ_SRC_45_INT 0xAD
+#define V_008DFC_SQ_SRC_46_INT 0xAE
+#define V_008DFC_SQ_SRC_47_INT 0xAF
+#define V_008DFC_SQ_SRC_48_INT 0xB0
+#define V_008DFC_SQ_SRC_49_INT 0xB1
+#define V_008DFC_SQ_SRC_50_INT 0xB2
+#define V_008DFC_SQ_SRC_51_INT 0xB3
+#define V_008DFC_SQ_SRC_52_INT 0xB4
+#define V_008DFC_SQ_SRC_53_INT 0xB5
+#define V_008DFC_SQ_SRC_54_INT 0xB6
+#define V_008DFC_SQ_SRC_55_INT 0xB7
+#define V_008DFC_SQ_SRC_56_INT 0xB8
+#define V_008DFC_SQ_SRC_57_INT 0xB9
+#define V_008DFC_SQ_SRC_58_INT 0xBA
+#define V_008DFC_SQ_SRC_59_INT 0xBB
+#define V_008DFC_SQ_SRC_60_INT 0xBC
+#define V_008DFC_SQ_SRC_61_INT 0xBD
+#define V_008DFC_SQ_SRC_62_INT 0xBE
+#define V_008DFC_SQ_SRC_63_INT 0xBF
+#define V_008DFC_SQ_SRC_64_INT 0xC0
+#define V_008DFC_SQ_SRC_M_1_INT 0xC1
+#define V_008DFC_SQ_SRC_M_2_INT 0xC2
+#define V_008DFC_SQ_SRC_M_3_INT 0xC3
+#define V_008DFC_SQ_SRC_M_4_INT 0xC4
+#define V_008DFC_SQ_SRC_M_5_INT 0xC5
+#define V_008DFC_SQ_SRC_M_6_INT 0xC6
+#define V_008DFC_SQ_SRC_M_7_INT 0xC7
+#define V_008DFC_SQ_SRC_M_8_INT 0xC8
+#define V_008DFC_SQ_SRC_M_9_INT 0xC9
+#define V_008DFC_SQ_SRC_M_10_INT 0xCA
+#define V_008DFC_SQ_SRC_M_11_INT 0xCB
+#define V_008DFC_SQ_SRC_M_12_INT 0xCC
+#define V_008DFC_SQ_SRC_M_13_INT 0xCD
+#define V_008DFC_SQ_SRC_M_14_INT 0xCE
+#define V_008DFC_SQ_SRC_M_15_INT 0xCF
+#define V_008DFC_SQ_SRC_M_16_INT 0xD0
+#define V_008DFC_SQ_SRC_0_5 0xF0
+#define V_008DFC_SQ_SRC_M_0_5 0xF1
+#define V_008DFC_SQ_SRC_1 0xF2
+#define V_008DFC_SQ_SRC_M_1 0xF3
+#define V_008DFC_SQ_SRC_2 0xF4
+#define V_008DFC_SQ_SRC_M_2 0xF5
+#define V_008DFC_SQ_SRC_4 0xF6
+#define V_008DFC_SQ_SRC_M_4 0xF7
+#define V_008DFC_SQ_SRC_VCCZ 0xFB
+#define V_008DFC_SQ_SRC_EXECZ 0xFC
+#define V_008DFC_SQ_SRC_SCC 0xFD
+#define V_008DFC_SQ_SRC_LDS_DIRECT 0xFE
+#define S_008DFC_SSRC1(x) (((x) & 0xFF) << 8)
+#define G_008DFC_SSRC1(x) (((x) >> 8) & 0xFF)
+#define C_008DFC_SSRC1 0xFFFF00FF
+#define V_008DFC_SQ_SGPR 0x00
+#define V_008DFC_SQ_VCC_LO 0x6A
+#define V_008DFC_SQ_VCC_HI 0x6B
+#define V_008DFC_SQ_TBA_LO 0x6C
+#define V_008DFC_SQ_TBA_HI 0x6D
+#define V_008DFC_SQ_TMA_LO 0x6E
+#define V_008DFC_SQ_TMA_HI 0x6F
+#define V_008DFC_SQ_TTMP0 0x70
+#define V_008DFC_SQ_TTMP1 0x71
+#define V_008DFC_SQ_TTMP2 0x72
+#define V_008DFC_SQ_TTMP3 0x73
+#define V_008DFC_SQ_TTMP4 0x74
+#define V_008DFC_SQ_TTMP5 0x75
+#define V_008DFC_SQ_TTMP6 0x76
+#define V_008DFC_SQ_TTMP7 0x77
+#define V_008DFC_SQ_TTMP8 0x78
+#define V_008DFC_SQ_TTMP9 0x79
+#define V_008DFC_SQ_TTMP10 0x7A
+#define V_008DFC_SQ_TTMP11 0x7B
+#define V_008DFC_SQ_M0 0x7C
+#define V_008DFC_SQ_EXEC_LO 0x7E
+#define V_008DFC_SQ_EXEC_HI 0x7F
+#define V_008DFC_SQ_SRC_0 0x80
+#define V_008DFC_SQ_SRC_1_INT 0x81
+#define V_008DFC_SQ_SRC_2_INT 0x82
+#define V_008DFC_SQ_SRC_3_INT 0x83
+#define V_008DFC_SQ_SRC_4_INT 0x84
+#define V_008DFC_SQ_SRC_5_INT 0x85
+#define V_008DFC_SQ_SRC_6_INT 0x86
+#define V_008DFC_SQ_SRC_7_INT 0x87
+#define V_008DFC_SQ_SRC_8_INT 0x88
+#define V_008DFC_SQ_SRC_9_INT 0x89
+#define V_008DFC_SQ_SRC_10_INT 0x8A
+#define V_008DFC_SQ_SRC_11_INT 0x8B
+#define V_008DFC_SQ_SRC_12_INT 0x8C
+#define V_008DFC_SQ_SRC_13_INT 0x8D
+#define V_008DFC_SQ_SRC_14_INT 0x8E
+#define V_008DFC_SQ_SRC_15_INT 0x8F
+#define V_008DFC_SQ_SRC_16_INT 0x90
+#define V_008DFC_SQ_SRC_17_INT 0x91
+#define V_008DFC_SQ_SRC_18_INT 0x92
+#define V_008DFC_SQ_SRC_19_INT 0x93
+#define V_008DFC_SQ_SRC_20_INT 0x94
+#define V_008DFC_SQ_SRC_21_INT 0x95
+#define V_008DFC_SQ_SRC_22_INT 0x96
+#define V_008DFC_SQ_SRC_23_INT 0x97
+#define V_008DFC_SQ_SRC_24_INT 0x98
+#define V_008DFC_SQ_SRC_25_INT 0x99
+#define V_008DFC_SQ_SRC_26_INT 0x9A
+#define V_008DFC_SQ_SRC_27_INT 0x9B
+#define V_008DFC_SQ_SRC_28_INT 0x9C
+#define V_008DFC_SQ_SRC_29_INT 0x9D
+#define V_008DFC_SQ_SRC_30_INT 0x9E
+#define V_008DFC_SQ_SRC_31_INT 0x9F
+#define V_008DFC_SQ_SRC_32_INT 0xA0
+#define V_008DFC_SQ_SRC_33_INT 0xA1
+#define V_008DFC_SQ_SRC_34_INT 0xA2
+#define V_008DFC_SQ_SRC_35_INT 0xA3
+#define V_008DFC_SQ_SRC_36_INT 0xA4
+#define V_008DFC_SQ_SRC_37_INT 0xA5
+#define V_008DFC_SQ_SRC_38_INT 0xA6
+#define V_008DFC_SQ_SRC_39_INT 0xA7
+#define V_008DFC_SQ_SRC_40_INT 0xA8
+#define V_008DFC_SQ_SRC_41_INT 0xA9
+#define V_008DFC_SQ_SRC_42_INT 0xAA
+#define V_008DFC_SQ_SRC_43_INT 0xAB
+#define V_008DFC_SQ_SRC_44_INT 0xAC
+#define V_008DFC_SQ_SRC_45_INT 0xAD
+#define V_008DFC_SQ_SRC_46_INT 0xAE
+#define V_008DFC_SQ_SRC_47_INT 0xAF
+#define V_008DFC_SQ_SRC_48_INT 0xB0
+#define V_008DFC_SQ_SRC_49_INT 0xB1
+#define V_008DFC_SQ_SRC_50_INT 0xB2
+#define V_008DFC_SQ_SRC_51_INT 0xB3
+#define V_008DFC_SQ_SRC_52_INT 0xB4
+#define V_008DFC_SQ_SRC_53_INT 0xB5
+#define V_008DFC_SQ_SRC_54_INT 0xB6
+#define V_008DFC_SQ_SRC_55_INT 0xB7
+#define V_008DFC_SQ_SRC_56_INT 0xB8
+#define V_008DFC_SQ_SRC_57_INT 0xB9
+#define V_008DFC_SQ_SRC_58_INT 0xBA
+#define V_008DFC_SQ_SRC_59_INT 0xBB
+#define V_008DFC_SQ_SRC_60_INT 0xBC
+#define V_008DFC_SQ_SRC_61_INT 0xBD
+#define V_008DFC_SQ_SRC_62_INT 0xBE
+#define V_008DFC_SQ_SRC_63_INT 0xBF
+#define V_008DFC_SQ_SRC_64_INT 0xC0
+#define V_008DFC_SQ_SRC_M_1_INT 0xC1
+#define V_008DFC_SQ_SRC_M_2_INT 0xC2
+#define V_008DFC_SQ_SRC_M_3_INT 0xC3
+#define V_008DFC_SQ_SRC_M_4_INT 0xC4
+#define V_008DFC_SQ_SRC_M_5_INT 0xC5
+#define V_008DFC_SQ_SRC_M_6_INT 0xC6
+#define V_008DFC_SQ_SRC_M_7_INT 0xC7
+#define V_008DFC_SQ_SRC_M_8_INT 0xC8
+#define V_008DFC_SQ_SRC_M_9_INT 0xC9
+#define V_008DFC_SQ_SRC_M_10_INT 0xCA
+#define V_008DFC_SQ_SRC_M_11_INT 0xCB
+#define V_008DFC_SQ_SRC_M_12_INT 0xCC
+#define V_008DFC_SQ_SRC_M_13_INT 0xCD
+#define V_008DFC_SQ_SRC_M_14_INT 0xCE
+#define V_008DFC_SQ_SRC_M_15_INT 0xCF
+#define V_008DFC_SQ_SRC_M_16_INT 0xD0
+#define V_008DFC_SQ_SRC_0_5 0xF0
+#define V_008DFC_SQ_SRC_M_0_5 0xF1
+#define V_008DFC_SQ_SRC_1 0xF2
+#define V_008DFC_SQ_SRC_M_1 0xF3
+#define V_008DFC_SQ_SRC_2 0xF4
+#define V_008DFC_SQ_SRC_M_2 0xF5
+#define V_008DFC_SQ_SRC_4 0xF6
+#define V_008DFC_SQ_SRC_M_4 0xF7
+#define V_008DFC_SQ_SRC_VCCZ 0xFB
+#define V_008DFC_SQ_SRC_EXECZ 0xFC
+#define V_008DFC_SQ_SRC_SCC 0xFD
+#define V_008DFC_SQ_SRC_LDS_DIRECT 0xFE
+#define S_008DFC_OP(x) (((x) & 0x7F) << 16)
+#define G_008DFC_OP(x) (((x) >> 16) & 0x7F)
+#define C_008DFC_OP 0xFF80FFFF
+#define V_008DFC_SQ_S_CMP_EQ_I32 0x00
+#define V_008DFC_SQ_S_CMP_LG_I32 0x01
+#define V_008DFC_SQ_S_CMP_GT_I32 0x02
+#define V_008DFC_SQ_S_CMP_GE_I32 0x03
+#define V_008DFC_SQ_S_CMP_LT_I32 0x04
+#define V_008DFC_SQ_S_CMP_LE_I32 0x05
+#define V_008DFC_SQ_S_CMP_EQ_U32 0x06
+#define V_008DFC_SQ_S_CMP_LG_U32 0x07
+#define V_008DFC_SQ_S_CMP_GT_U32 0x08
+#define V_008DFC_SQ_S_CMP_GE_U32 0x09
+#define V_008DFC_SQ_S_CMP_LT_U32 0x0A
+#define V_008DFC_SQ_S_CMP_LE_U32 0x0B
+#define V_008DFC_SQ_S_BITCMP0_B32 0x0C
+#define V_008DFC_SQ_S_BITCMP1_B32 0x0D
+#define V_008DFC_SQ_S_BITCMP0_B64 0x0E
+#define V_008DFC_SQ_S_BITCMP1_B64 0x0F
+#define V_008DFC_SQ_S_SETVSKIP 0x10
+#define S_008DFC_ENCODING(x) (((x) & 0x1FF) << 23)
+#define G_008DFC_ENCODING(x) (((x) >> 23) & 0x1FF)
+#define C_008DFC_ENCODING 0x007FFFFF
+#define V_008DFC_SQ_ENC_SOPC_FIELD 0x17E
+#endif
+#define R_008DFC_SQ_EXP_0 0x008DFC
+#define S_008DFC_EN(x) (((x) & 0x0F) << 0)
+#define G_008DFC_EN(x) (((x) >> 0) & 0x0F)
+#define C_008DFC_EN 0xFFFFFFF0
+#define S_008DFC_TGT(x) (((x) & 0x3F) << 4)
+#define G_008DFC_TGT(x) (((x) >> 4) & 0x3F)
+#define C_008DFC_TGT 0xFFFFFC0F
+#define V_008DFC_SQ_EXP_MRT 0x00
+#define V_008DFC_SQ_EXP_MRTZ 0x08
+#define V_008DFC_SQ_EXP_NULL 0x09
+#define V_008DFC_SQ_EXP_POS 0x0C
+#define V_008DFC_SQ_EXP_PARAM 0x20
+#define S_008DFC_COMPR(x) (((x) & 0x1) << 10)
+#define G_008DFC_COMPR(x) (((x) >> 10) & 0x1)
+#define C_008DFC_COMPR 0xFFFFFBFF
+#define S_008DFC_DONE(x) (((x) & 0x1) << 11)
+#define G_008DFC_DONE(x) (((x) >> 11) & 0x1)
+#define C_008DFC_DONE 0xFFFFF7FF
+#define S_008DFC_VM(x) (((x) & 0x1) << 12)
+#define G_008DFC_VM(x) (((x) >> 12) & 0x1)
+#define C_008DFC_VM 0xFFFFEFFF
+#define S_008DFC_ENCODING(x) (((x) & 0x3F) << 26)
+#define G_008DFC_ENCODING(x) (((x) >> 26) & 0x3F)
+#define C_008DFC_ENCODING 0x03FFFFFF
+#define V_008DFC_SQ_ENC_EXP_FIELD 0x3E
+#if 0
+#define R_008DFC_SQ_MIMG_0 0x008DFC
+#define S_008DFC_DMASK(x) (((x) & 0x0F) << 8)
+#define G_008DFC_DMASK(x) (((x) >> 8) & 0x0F)
+#define C_008DFC_DMASK 0xFFFFF0FF
+#define S_008DFC_UNORM(x) (((x) & 0x1) << 12)
+#define G_008DFC_UNORM(x) (((x) >> 12) & 0x1)
+#define C_008DFC_UNORM 0xFFFFEFFF
+#define S_008DFC_GLC(x) (((x) & 0x1) << 13)
+#define G_008DFC_GLC(x) (((x) >> 13) & 0x1)
+#define C_008DFC_GLC 0xFFFFDFFF
+#define S_008DFC_DA(x) (((x) & 0x1) << 14)
+#define G_008DFC_DA(x) (((x) >> 14) & 0x1)
+#define C_008DFC_DA 0xFFFFBFFF
+#define S_008DFC_R128(x) (((x) & 0x1) << 15)
+#define G_008DFC_R128(x) (((x) >> 15) & 0x1)
+#define C_008DFC_R128 0xFFFF7FFF
+#define S_008DFC_TFE(x) (((x) & 0x1) << 16)
+#define G_008DFC_TFE(x) (((x) >> 16) & 0x1)
+#define C_008DFC_TFE 0xFFFEFFFF
+#define S_008DFC_LWE(x) (((x) & 0x1) << 17)
+#define G_008DFC_LWE(x) (((x) >> 17) & 0x1)
+#define C_008DFC_LWE 0xFFFDFFFF
+#define S_008DFC_OP(x) (((x) & 0x7F) << 18)
+#define G_008DFC_OP(x) (((x) >> 18) & 0x7F)
+#define C_008DFC_OP 0xFE03FFFF
+#define V_008DFC_SQ_IMAGE_LOAD 0x00
+#define V_008DFC_SQ_IMAGE_LOAD_MIP 0x01
+#define V_008DFC_SQ_IMAGE_LOAD_PCK 0x02
+#define V_008DFC_SQ_IMAGE_LOAD_PCK_SGN 0x03
+#define V_008DFC_SQ_IMAGE_LOAD_MIP_PCK 0x04
+#define V_008DFC_SQ_IMAGE_LOAD_MIP_PCK_SGN 0x05
+#define V_008DFC_SQ_IMAGE_STORE 0x08
+#define V_008DFC_SQ_IMAGE_STORE_MIP 0x09
+#define V_008DFC_SQ_IMAGE_STORE_PCK 0x0A
+#define V_008DFC_SQ_IMAGE_STORE_MIP_PCK 0x0B
+#define V_008DFC_SQ_IMAGE_GET_RESINFO 0x0E
+#define V_008DFC_SQ_IMAGE_ATOMIC_SWAP 0x0F
+#define V_008DFC_SQ_IMAGE_ATOMIC_CMPSWAP 0x10
+#define V_008DFC_SQ_IMAGE_ATOMIC_ADD 0x11
+#define V_008DFC_SQ_IMAGE_ATOMIC_SUB 0x12
+#define V_008DFC_SQ_IMAGE_ATOMIC_RSUB 0x13
+#define V_008DFC_SQ_IMAGE_ATOMIC_SMIN 0x14
+#define V_008DFC_SQ_IMAGE_ATOMIC_UMIN 0x15
+#define V_008DFC_SQ_IMAGE_ATOMIC_SMAX 0x16
+#define V_008DFC_SQ_IMAGE_ATOMIC_UMAX 0x17
+#define V_008DFC_SQ_IMAGE_ATOMIC_AND 0x18
+#define V_008DFC_SQ_IMAGE_ATOMIC_OR 0x19
+#define V_008DFC_SQ_IMAGE_ATOMIC_XOR 0x1A
+#define V_008DFC_SQ_IMAGE_ATOMIC_INC 0x1B
+#define V_008DFC_SQ_IMAGE_ATOMIC_DEC 0x1C
+#define V_008DFC_SQ_IMAGE_ATOMIC_FCMPSWAP 0x1D
+#define V_008DFC_SQ_IMAGE_ATOMIC_FMIN 0x1E
+#define V_008DFC_SQ_IMAGE_ATOMIC_FMAX 0x1F
+#define V_008DFC_SQ_IMAGE_SAMPLE 0x20
+#define V_008DFC_SQ_IMAGE_SAMPLE_CL 0x21
+#define V_008DFC_SQ_IMAGE_SAMPLE_D 0x22
+#define V_008DFC_SQ_IMAGE_SAMPLE_D_CL 0x23
+#define V_008DFC_SQ_IMAGE_SAMPLE_L 0x24
+#define V_008DFC_SQ_IMAGE_SAMPLE_B 0x25
+#define V_008DFC_SQ_IMAGE_SAMPLE_B_CL 0x26
+#define V_008DFC_SQ_IMAGE_SAMPLE_LZ 0x27
+#define V_008DFC_SQ_IMAGE_SAMPLE_C 0x28
+#define V_008DFC_SQ_IMAGE_SAMPLE_C_CL 0x29
+#define V_008DFC_SQ_IMAGE_SAMPLE_C_D 0x2A
+#define V_008DFC_SQ_IMAGE_SAMPLE_C_D_CL 0x2B
+#define V_008DFC_SQ_IMAGE_SAMPLE_C_L 0x2C
+#define V_008DFC_SQ_IMAGE_SAMPLE_C_B 0x2D
+#define V_008DFC_SQ_IMAGE_SAMPLE_C_B_CL 0x2E
+#define V_008DFC_SQ_IMAGE_SAMPLE_C_LZ 0x2F
+#define V_008DFC_SQ_IMAGE_SAMPLE_O 0x30
+#define V_008DFC_SQ_IMAGE_SAMPLE_CL_O 0x31
+#define V_008DFC_SQ_IMAGE_SAMPLE_D_O 0x32
+#define V_008DFC_SQ_IMAGE_SAMPLE_D_CL_O 0x33
+#define V_008DFC_SQ_IMAGE_SAMPLE_L_O 0x34
+#define V_008DFC_SQ_IMAGE_SAMPLE_B_O 0x35
+#define V_008DFC_SQ_IMAGE_SAMPLE_B_CL_O 0x36
+#define V_008DFC_SQ_IMAGE_SAMPLE_LZ_O 0x37
+#define V_008DFC_SQ_IMAGE_SAMPLE_C_O 0x38
+#define V_008DFC_SQ_IMAGE_SAMPLE_C_CL_O 0x39
+#define V_008DFC_SQ_IMAGE_SAMPLE_C_D_O 0x3A
+#define V_008DFC_SQ_IMAGE_SAMPLE_C_D_CL_O 0x3B
+#define V_008DFC_SQ_IMAGE_SAMPLE_C_L_O 0x3C
+#define V_008DFC_SQ_IMAGE_SAMPLE_C_B_O 0x3D
+#define V_008DFC_SQ_IMAGE_SAMPLE_C_B_CL_O 0x3E
+#define V_008DFC_SQ_IMAGE_SAMPLE_C_LZ_O 0x3F
+#define V_008DFC_SQ_IMAGE_GATHER4 0x40
+#define V_008DFC_SQ_IMAGE_GATHER4_CL 0x41
+#define V_008DFC_SQ_IMAGE_GATHER4_L 0x44
+#define V_008DFC_SQ_IMAGE_GATHER4_B 0x45
+#define V_008DFC_SQ_IMAGE_GATHER4_B_CL 0x46
+#define V_008DFC_SQ_IMAGE_GATHER4_LZ 0x47
+#define V_008DFC_SQ_IMAGE_GATHER4_C 0x48
+#define V_008DFC_SQ_IMAGE_GATHER4_C_CL 0x49
+#define V_008DFC_SQ_IMAGE_GATHER4_C_L 0x4C
+#define V_008DFC_SQ_IMAGE_GATHER4_C_B 0x4D
+#define V_008DFC_SQ_IMAGE_GATHER4_C_B_CL 0x4E
+#define V_008DFC_SQ_IMAGE_GATHER4_C_LZ 0x4F
+#define V_008DFC_SQ_IMAGE_GATHER4_O 0x50
+#define V_008DFC_SQ_IMAGE_GATHER4_CL_O 0x51
+#define V_008DFC_SQ_IMAGE_GATHER4_L_O 0x54
+#define V_008DFC_SQ_IMAGE_GATHER4_B_O 0x55
+#define V_008DFC_SQ_IMAGE_GATHER4_B_CL_O 0x56
+#define V_008DFC_SQ_IMAGE_GATHER4_LZ_O 0x57
+#define V_008DFC_SQ_IMAGE_GATHER4_C_O 0x58
+#define V_008DFC_SQ_IMAGE_GATHER4_C_CL_O 0x59
+#define V_008DFC_SQ_IMAGE_GATHER4_C_L_O 0x5C
+#define V_008DFC_SQ_IMAGE_GATHER4_C_B_O 0x5D
+#define V_008DFC_SQ_IMAGE_GATHER4_C_B_CL_O 0x5E
+#define V_008DFC_SQ_IMAGE_GATHER4_C_LZ_O 0x5F
+#define V_008DFC_SQ_IMAGE_GET_LOD 0x60
+#define V_008DFC_SQ_IMAGE_SAMPLE_CD 0x68
+#define V_008DFC_SQ_IMAGE_SAMPLE_CD_CL 0x69
+#define V_008DFC_SQ_IMAGE_SAMPLE_C_CD 0x6A
+#define V_008DFC_SQ_IMAGE_SAMPLE_C_CD_CL 0x6B
+#define V_008DFC_SQ_IMAGE_SAMPLE_CD_O 0x6C
+#define V_008DFC_SQ_IMAGE_SAMPLE_CD_CL_O 0x6D
+#define V_008DFC_SQ_IMAGE_SAMPLE_C_CD_O 0x6E
+#define V_008DFC_SQ_IMAGE_SAMPLE_C_CD_CL_O 0x6F
+#define V_008DFC_SQ_IMAGE_RSRC256 0x7E
+#define V_008DFC_SQ_IMAGE_SAMPLER 0x7F
+#define S_008DFC_SLC(x) (((x) & 0x1) << 25)
+#define G_008DFC_SLC(x) (((x) >> 25) & 0x1)
+#define C_008DFC_SLC 0xFDFFFFFF
+#define S_008DFC_ENCODING(x) (((x) & 0x3F) << 26)
+#define G_008DFC_ENCODING(x) (((x) >> 26) & 0x3F)
+#define C_008DFC_ENCODING 0x03FFFFFF
+#define V_008DFC_SQ_ENC_MIMG_FIELD 0x3C
+#define R_008DFC_SQ_SOPP 0x008DFC
+#define S_008DFC_SIMM16(x) (((x) & 0xFFFF) << 0)
+#define G_008DFC_SIMM16(x) (((x) >> 0) & 0xFFFF)
+#define C_008DFC_SIMM16 0xFFFF0000
+#define S_008DFC_OP(x) (((x) & 0x7F) << 16)
+#define G_008DFC_OP(x) (((x) >> 16) & 0x7F)
+#define C_008DFC_OP 0xFF80FFFF
+#define V_008DFC_SQ_S_NOP 0x00
+#define V_008DFC_SQ_S_ENDPGM 0x01
+#define V_008DFC_SQ_S_BRANCH 0x02
+#define V_008DFC_SQ_S_CBRANCH_SCC0 0x04
+#define V_008DFC_SQ_S_CBRANCH_SCC1 0x05
+#define V_008DFC_SQ_S_CBRANCH_VCCZ 0x06
+#define V_008DFC_SQ_S_CBRANCH_VCCNZ 0x07
+#define V_008DFC_SQ_S_CBRANCH_EXECZ 0x08
+#define V_008DFC_SQ_S_CBRANCH_EXECNZ 0x09
+#define V_008DFC_SQ_S_BARRIER 0x0A
+#define V_008DFC_SQ_S_WAITCNT 0x0C
+#define V_008DFC_SQ_S_SETHALT 0x0D
+#define V_008DFC_SQ_S_SLEEP 0x0E
+#define V_008DFC_SQ_S_SETPRIO 0x0F
+#define V_008DFC_SQ_S_SENDMSG 0x10
+#define V_008DFC_SQ_S_SENDMSGHALT 0x11
+#define V_008DFC_SQ_S_TRAP 0x12
+#define V_008DFC_SQ_S_ICACHE_INV 0x13
+#define V_008DFC_SQ_S_INCPERFLEVEL 0x14
+#define V_008DFC_SQ_S_DECPERFLEVEL 0x15
+#define V_008DFC_SQ_S_TTRACEDATA 0x16
+#define S_008DFC_ENCODING(x) (((x) & 0x1FF) << 23)
+#define G_008DFC_ENCODING(x) (((x) >> 23) & 0x1FF)
+#define C_008DFC_ENCODING 0x007FFFFF
+#define V_008DFC_SQ_ENC_SOPP_FIELD 0x17F
+#define R_008DFC_SQ_VINTRP 0x008DFC
+#define S_008DFC_VSRC(x) (((x) & 0xFF) << 0)
+#define G_008DFC_VSRC(x) (((x) >> 0) & 0xFF)
+#define C_008DFC_VSRC 0xFFFFFF00
+#define V_008DFC_SQ_VGPR 0x00
+#define S_008DFC_ATTRCHAN(x) (((x) & 0x03) << 8)
+#define G_008DFC_ATTRCHAN(x) (((x) >> 8) & 0x03)
+#define C_008DFC_ATTRCHAN 0xFFFFFCFF
+#define V_008DFC_SQ_CHAN_X 0x00
+#define V_008DFC_SQ_CHAN_Y 0x01
+#define V_008DFC_SQ_CHAN_Z 0x02
+#define V_008DFC_SQ_CHAN_W 0x03
+#define S_008DFC_ATTR(x) (((x) & 0x3F) << 10)
+#define G_008DFC_ATTR(x) (((x) >> 10) & 0x3F)
+#define C_008DFC_ATTR 0xFFFF03FF
+#define V_008DFC_SQ_ATTR 0x00
+#define S_008DFC_OP(x) (((x) & 0x03) << 16)
+#define G_008DFC_OP(x) (((x) >> 16) & 0x03)
+#define C_008DFC_OP 0xFFFCFFFF
+#define V_008DFC_SQ_V_INTERP_P1_F32 0x00
+#define V_008DFC_SQ_V_INTERP_P2_F32 0x01
+#define V_008DFC_SQ_V_INTERP_MOV_F32 0x02
+#define S_008DFC_VDST(x) (((x) & 0xFF) << 18)
+#define G_008DFC_VDST(x) (((x) >> 18) & 0xFF)
+#define C_008DFC_VDST 0xFC03FFFF
+#define V_008DFC_SQ_VGPR 0x00
+#define S_008DFC_ENCODING(x) (((x) & 0x3F) << 26)
+#define G_008DFC_ENCODING(x) (((x) >> 26) & 0x3F)
+#define C_008DFC_ENCODING 0x03FFFFFF
+#define V_008DFC_SQ_ENC_VINTRP_FIELD 0x32
+#define R_008DFC_SQ_MTBUF_0 0x008DFC
+#define S_008DFC_OFFSET(x) (((x) & 0xFFF) << 0)
+#define G_008DFC_OFFSET(x) (((x) >> 0) & 0xFFF)
+#define C_008DFC_OFFSET 0xFFFFF000
+#define S_008DFC_OFFEN(x) (((x) & 0x1) << 12)
+#define G_008DFC_OFFEN(x) (((x) >> 12) & 0x1)
+#define C_008DFC_OFFEN 0xFFFFEFFF
+#define S_008DFC_IDXEN(x) (((x) & 0x1) << 13)
+#define G_008DFC_IDXEN(x) (((x) >> 13) & 0x1)
+#define C_008DFC_IDXEN 0xFFFFDFFF
+#define S_008DFC_GLC(x) (((x) & 0x1) << 14)
+#define G_008DFC_GLC(x) (((x) >> 14) & 0x1)
+#define C_008DFC_GLC 0xFFFFBFFF
+#define S_008DFC_ADDR64(x) (((x) & 0x1) << 15)
+#define G_008DFC_ADDR64(x) (((x) >> 15) & 0x1)
+#define C_008DFC_ADDR64 0xFFFF7FFF
+#define S_008DFC_OP(x) (((x) & 0x07) << 16)
+#define G_008DFC_OP(x) (((x) >> 16) & 0x07)
+#define C_008DFC_OP 0xFFF8FFFF
+#define V_008DFC_SQ_TBUFFER_LOAD_FORMAT_X 0x00
+#define V_008DFC_SQ_TBUFFER_LOAD_FORMAT_XY 0x01
+#define V_008DFC_SQ_TBUFFER_LOAD_FORMAT_XYZ 0x02
+#define V_008DFC_SQ_TBUFFER_LOAD_FORMAT_XYZW 0x03
+#define V_008DFC_SQ_TBUFFER_STORE_FORMAT_X 0x04
+#define V_008DFC_SQ_TBUFFER_STORE_FORMAT_XY 0x05
+#define V_008DFC_SQ_TBUFFER_STORE_FORMAT_XYZ 0x06
+#define V_008DFC_SQ_TBUFFER_STORE_FORMAT_XYZW 0x07
+#define S_008DFC_DFMT(x) (((x) & 0x0F) << 19)
+#define G_008DFC_DFMT(x) (((x) >> 19) & 0x0F)
+#define C_008DFC_DFMT 0xFF87FFFF
+#define S_008DFC_NFMT(x) (((x) & 0x07) << 23)
+#define G_008DFC_NFMT(x) (((x) >> 23) & 0x07)
+#define C_008DFC_NFMT 0xFC7FFFFF
+#define S_008DFC_ENCODING(x) (((x) & 0x3F) << 26)
+#define G_008DFC_ENCODING(x) (((x) >> 26) & 0x3F)
+#define C_008DFC_ENCODING 0x03FFFFFF
+#define V_008DFC_SQ_ENC_MTBUF_FIELD 0x3A
+#define R_008DFC_SQ_SMRD 0x008DFC
+#define S_008DFC_OFFSET(x) (((x) & 0xFF) << 0)
+#define G_008DFC_OFFSET(x) (((x) >> 0) & 0xFF)
+#define C_008DFC_OFFSET 0xFFFFFF00
+#define V_008DFC_SQ_SGPR 0x00
+#define V_008DFC_SQ_VCC_LO 0x6A
+#define V_008DFC_SQ_VCC_HI 0x6B
+#define V_008DFC_SQ_TBA_LO 0x6C
+#define V_008DFC_SQ_TBA_HI 0x6D
+#define V_008DFC_SQ_TMA_LO 0x6E
+#define V_008DFC_SQ_TMA_HI 0x6F
+#define V_008DFC_SQ_TTMP0 0x70
+#define V_008DFC_SQ_TTMP1 0x71
+#define V_008DFC_SQ_TTMP2 0x72
+#define V_008DFC_SQ_TTMP3 0x73
+#define V_008DFC_SQ_TTMP4 0x74
+#define V_008DFC_SQ_TTMP5 0x75
+#define V_008DFC_SQ_TTMP6 0x76
+#define V_008DFC_SQ_TTMP7 0x77
+#define V_008DFC_SQ_TTMP8 0x78
+#define V_008DFC_SQ_TTMP9 0x79
+#define V_008DFC_SQ_TTMP10 0x7A
+#define V_008DFC_SQ_TTMP11 0x7B
+#define S_008DFC_IMM(x) (((x) & 0x1) << 8)
+#define G_008DFC_IMM(x) (((x) >> 8) & 0x1)
+#define C_008DFC_IMM 0xFFFFFEFF
+#define S_008DFC_SBASE(x) (((x) & 0x3F) << 9)
+#define G_008DFC_SBASE(x) (((x) >> 9) & 0x3F)
+#define C_008DFC_SBASE 0xFFFF81FF
+#define S_008DFC_SDST(x) (((x) & 0x7F) << 15)
+#define G_008DFC_SDST(x) (((x) >> 15) & 0x7F)
+#define C_008DFC_SDST 0xFFC07FFF
+#define V_008DFC_SQ_SGPR 0x00
+#define V_008DFC_SQ_VCC_LO 0x6A
+#define V_008DFC_SQ_VCC_HI 0x6B
+#define V_008DFC_SQ_TBA_LO 0x6C
+#define V_008DFC_SQ_TBA_HI 0x6D
+#define V_008DFC_SQ_TMA_LO 0x6E
+#define V_008DFC_SQ_TMA_HI 0x6F
+#define V_008DFC_SQ_TTMP0 0x70
+#define V_008DFC_SQ_TTMP1 0x71
+#define V_008DFC_SQ_TTMP2 0x72
+#define V_008DFC_SQ_TTMP3 0x73
+#define V_008DFC_SQ_TTMP4 0x74
+#define V_008DFC_SQ_TTMP5 0x75
+#define V_008DFC_SQ_TTMP6 0x76
+#define V_008DFC_SQ_TTMP7 0x77
+#define V_008DFC_SQ_TTMP8 0x78
+#define V_008DFC_SQ_TTMP9 0x79
+#define V_008DFC_SQ_TTMP10 0x7A
+#define V_008DFC_SQ_TTMP11 0x7B
+#define V_008DFC_SQ_M0 0x7C
+#define V_008DFC_SQ_EXEC_LO 0x7E
+#define V_008DFC_SQ_EXEC_HI 0x7F
+#define S_008DFC_OP(x) (((x) & 0x1F) << 22)
+#define G_008DFC_OP(x) (((x) >> 22) & 0x1F)
+#define C_008DFC_OP 0xF83FFFFF
+#define V_008DFC_SQ_S_LOAD_DWORD 0x00
+#define V_008DFC_SQ_S_LOAD_DWORDX2 0x01
+#define V_008DFC_SQ_S_LOAD_DWORDX4 0x02
+#define V_008DFC_SQ_S_LOAD_DWORDX8 0x03
+#define V_008DFC_SQ_S_LOAD_DWORDX16 0x04
+#define V_008DFC_SQ_S_BUFFER_LOAD_DWORD 0x08
+#define V_008DFC_SQ_S_BUFFER_LOAD_DWORDX2 0x09
+#define V_008DFC_SQ_S_BUFFER_LOAD_DWORDX4 0x0A
+#define V_008DFC_SQ_S_BUFFER_LOAD_DWORDX8 0x0B
+#define V_008DFC_SQ_S_BUFFER_LOAD_DWORDX16 0x0C
+#define V_008DFC_SQ_S_MEMTIME 0x1E
+#define V_008DFC_SQ_S_DCACHE_INV 0x1F
+#define S_008DFC_ENCODING(x) (((x) & 0x1F) << 27)
+#define G_008DFC_ENCODING(x) (((x) >> 27) & 0x1F)
+#define C_008DFC_ENCODING 0x07FFFFFF
+#define V_008DFC_SQ_ENC_SMRD_FIELD 0x18
+#define R_008DFC_SQ_EXP_1 0x008DFC
+#define S_008DFC_VSRC0(x) (((x) & 0xFF) << 0)
+#define G_008DFC_VSRC0(x) (((x) >> 0) & 0xFF)
+#define C_008DFC_VSRC0 0xFFFFFF00
+#define V_008DFC_SQ_VGPR 0x00
+#define S_008DFC_VSRC1(x) (((x) & 0xFF) << 8)
+#define G_008DFC_VSRC1(x) (((x) >> 8) & 0xFF)
+#define C_008DFC_VSRC1 0xFFFF00FF
+#define V_008DFC_SQ_VGPR 0x00
+#define S_008DFC_VSRC2(x) (((x) & 0xFF) << 16)
+#define G_008DFC_VSRC2(x) (((x) >> 16) & 0xFF)
+#define C_008DFC_VSRC2 0xFF00FFFF
+#define V_008DFC_SQ_VGPR 0x00
+#define S_008DFC_VSRC3(x) (((x) & 0xFF) << 24)
+#define G_008DFC_VSRC3(x) (((x) >> 24) & 0xFF)
+#define C_008DFC_VSRC3 0x00FFFFFF
+#define V_008DFC_SQ_VGPR 0x00
+#define R_008DFC_SQ_DS_1 0x008DFC
+#define S_008DFC_ADDR(x) (((x) & 0xFF) << 0)
+#define G_008DFC_ADDR(x) (((x) >> 0) & 0xFF)
+#define C_008DFC_ADDR 0xFFFFFF00
+#define V_008DFC_SQ_VGPR 0x00
+#define S_008DFC_DATA0(x) (((x) & 0xFF) << 8)
+#define G_008DFC_DATA0(x) (((x) >> 8) & 0xFF)
+#define C_008DFC_DATA0 0xFFFF00FF
+#define V_008DFC_SQ_VGPR 0x00
+#define S_008DFC_DATA1(x) (((x) & 0xFF) << 16)
+#define G_008DFC_DATA1(x) (((x) >> 16) & 0xFF)
+#define C_008DFC_DATA1 0xFF00FFFF
+#define V_008DFC_SQ_VGPR 0x00
+#define S_008DFC_VDST(x) (((x) & 0xFF) << 24)
+#define G_008DFC_VDST(x) (((x) >> 24) & 0xFF)
+#define C_008DFC_VDST 0x00FFFFFF
+#define V_008DFC_SQ_VGPR 0x00
+#define R_008DFC_SQ_VOPC 0x008DFC
+#define S_008DFC_SRC0(x) (((x) & 0x1FF) << 0)
+#define G_008DFC_SRC0(x) (((x) >> 0) & 0x1FF)
+#define C_008DFC_SRC0 0xFFFFFE00
+#define V_008DFC_SQ_SGPR 0x00
+#define V_008DFC_SQ_VCC_LO 0x6A
+#define V_008DFC_SQ_VCC_HI 0x6B
+#define V_008DFC_SQ_TBA_LO 0x6C
+#define V_008DFC_SQ_TBA_HI 0x6D
+#define V_008DFC_SQ_TMA_LO 0x6E
+#define V_008DFC_SQ_TMA_HI 0x6F
+#define V_008DFC_SQ_TTMP0 0x70
+#define V_008DFC_SQ_TTMP1 0x71
+#define V_008DFC_SQ_TTMP2 0x72
+#define V_008DFC_SQ_TTMP3 0x73
+#define V_008DFC_SQ_TTMP4 0x74
+#define V_008DFC_SQ_TTMP5 0x75
+#define V_008DFC_SQ_TTMP6 0x76
+#define V_008DFC_SQ_TTMP7 0x77
+#define V_008DFC_SQ_TTMP8 0x78
+#define V_008DFC_SQ_TTMP9 0x79
+#define V_008DFC_SQ_TTMP10 0x7A
+#define V_008DFC_SQ_TTMP11 0x7B
+#define V_008DFC_SQ_M0 0x7C
+#define V_008DFC_SQ_EXEC_LO 0x7E
+#define V_008DFC_SQ_EXEC_HI 0x7F
+#define V_008DFC_SQ_SRC_0 0x80
+#define V_008DFC_SQ_SRC_1_INT 0x81
+#define V_008DFC_SQ_SRC_2_INT 0x82
+#define V_008DFC_SQ_SRC_3_INT 0x83
+#define V_008DFC_SQ_SRC_4_INT 0x84
+#define V_008DFC_SQ_SRC_5_INT 0x85
+#define V_008DFC_SQ_SRC_6_INT 0x86
+#define V_008DFC_SQ_SRC_7_INT 0x87
+#define V_008DFC_SQ_SRC_8_INT 0x88
+#define V_008DFC_SQ_SRC_9_INT 0x89
+#define V_008DFC_SQ_SRC_10_INT 0x8A
+#define V_008DFC_SQ_SRC_11_INT 0x8B
+#define V_008DFC_SQ_SRC_12_INT 0x8C
+#define V_008DFC_SQ_SRC_13_INT 0x8D
+#define V_008DFC_SQ_SRC_14_INT 0x8E
+#define V_008DFC_SQ_SRC_15_INT 0x8F
+#define V_008DFC_SQ_SRC_16_INT 0x90
+#define V_008DFC_SQ_SRC_17_INT 0x91
+#define V_008DFC_SQ_SRC_18_INT 0x92
+#define V_008DFC_SQ_SRC_19_INT 0x93
+#define V_008DFC_SQ_SRC_20_INT 0x94
+#define V_008DFC_SQ_SRC_21_INT 0x95
+#define V_008DFC_SQ_SRC_22_INT 0x96
+#define V_008DFC_SQ_SRC_23_INT 0x97
+#define V_008DFC_SQ_SRC_24_INT 0x98
+#define V_008DFC_SQ_SRC_25_INT 0x99
+#define V_008DFC_SQ_SRC_26_INT 0x9A
+#define V_008DFC_SQ_SRC_27_INT 0x9B
+#define V_008DFC_SQ_SRC_28_INT 0x9C
+#define V_008DFC_SQ_SRC_29_INT 0x9D
+#define V_008DFC_SQ_SRC_30_INT 0x9E
+#define V_008DFC_SQ_SRC_31_INT 0x9F
+#define V_008DFC_SQ_SRC_32_INT 0xA0
+#define V_008DFC_SQ_SRC_33_INT 0xA1
+#define V_008DFC_SQ_SRC_34_INT 0xA2
+#define V_008DFC_SQ_SRC_35_INT 0xA3
+#define V_008DFC_SQ_SRC_36_INT 0xA4
+#define V_008DFC_SQ_SRC_37_INT 0xA5
+#define V_008DFC_SQ_SRC_38_INT 0xA6
+#define V_008DFC_SQ_SRC_39_INT 0xA7
+#define V_008DFC_SQ_SRC_40_INT 0xA8
+#define V_008DFC_SQ_SRC_41_INT 0xA9
+#define V_008DFC_SQ_SRC_42_INT 0xAA
+#define V_008DFC_SQ_SRC_43_INT 0xAB
+#define V_008DFC_SQ_SRC_44_INT 0xAC
+#define V_008DFC_SQ_SRC_45_INT 0xAD
+#define V_008DFC_SQ_SRC_46_INT 0xAE
+#define V_008DFC_SQ_SRC_47_INT 0xAF
+#define V_008DFC_SQ_SRC_48_INT 0xB0
+#define V_008DFC_SQ_SRC_49_INT 0xB1
+#define V_008DFC_SQ_SRC_50_INT 0xB2
+#define V_008DFC_SQ_SRC_51_INT 0xB3
+#define V_008DFC_SQ_SRC_52_INT 0xB4
+#define V_008DFC_SQ_SRC_53_INT 0xB5
+#define V_008DFC_SQ_SRC_54_INT 0xB6
+#define V_008DFC_SQ_SRC_55_INT 0xB7
+#define V_008DFC_SQ_SRC_56_INT 0xB8
+#define V_008DFC_SQ_SRC_57_INT 0xB9
+#define V_008DFC_SQ_SRC_58_INT 0xBA
+#define V_008DFC_SQ_SRC_59_INT 0xBB
+#define V_008DFC_SQ_SRC_60_INT 0xBC
+#define V_008DFC_SQ_SRC_61_INT 0xBD
+#define V_008DFC_SQ_SRC_62_INT 0xBE
+#define V_008DFC_SQ_SRC_63_INT 0xBF
+#define V_008DFC_SQ_SRC_64_INT 0xC0
+#define V_008DFC_SQ_SRC_M_1_INT 0xC1
+#define V_008DFC_SQ_SRC_M_2_INT 0xC2
+#define V_008DFC_SQ_SRC_M_3_INT 0xC3
+#define V_008DFC_SQ_SRC_M_4_INT 0xC4
+#define V_008DFC_SQ_SRC_M_5_INT 0xC5
+#define V_008DFC_SQ_SRC_M_6_INT 0xC6
+#define V_008DFC_SQ_SRC_M_7_INT 0xC7
+#define V_008DFC_SQ_SRC_M_8_INT 0xC8
+#define V_008DFC_SQ_SRC_M_9_INT 0xC9
+#define V_008DFC_SQ_SRC_M_10_INT 0xCA
+#define V_008DFC_SQ_SRC_M_11_INT 0xCB
+#define V_008DFC_SQ_SRC_M_12_INT 0xCC
+#define V_008DFC_SQ_SRC_M_13_INT 0xCD
+#define V_008DFC_SQ_SRC_M_14_INT 0xCE
+#define V_008DFC_SQ_SRC_M_15_INT 0xCF
+#define V_008DFC_SQ_SRC_M_16_INT 0xD0
+#define V_008DFC_SQ_SRC_0_5 0xF0
+#define V_008DFC_SQ_SRC_M_0_5 0xF1
+#define V_008DFC_SQ_SRC_1 0xF2
+#define V_008DFC_SQ_SRC_M_1 0xF3
+#define V_008DFC_SQ_SRC_2 0xF4
+#define V_008DFC_SQ_SRC_M_2 0xF5
+#define V_008DFC_SQ_SRC_4 0xF6
+#define V_008DFC_SQ_SRC_M_4 0xF7
+#define V_008DFC_SQ_SRC_VCCZ 0xFB
+#define V_008DFC_SQ_SRC_EXECZ 0xFC
+#define V_008DFC_SQ_SRC_SCC 0xFD
+#define V_008DFC_SQ_SRC_LDS_DIRECT 0xFE
+#define V_008DFC_SQ_SRC_VGPR 0x100
+#define S_008DFC_VSRC1(x) (((x) & 0xFF) << 9)
+#define G_008DFC_VSRC1(x) (((x) >> 9) & 0xFF)
+#define C_008DFC_VSRC1 0xFFFE01FF
+#define V_008DFC_SQ_VGPR 0x00
+#define S_008DFC_OP(x) (((x) & 0xFF) << 17)
+#define G_008DFC_OP(x) (((x) >> 17) & 0xFF)
+#define C_008DFC_OP 0xFE01FFFF
+#define V_008DFC_SQ_V_CMP_F_F32 0x00
+#define V_008DFC_SQ_V_CMP_LT_F32 0x01
+#define V_008DFC_SQ_V_CMP_EQ_F32 0x02
+#define V_008DFC_SQ_V_CMP_LE_F32 0x03
+#define V_008DFC_SQ_V_CMP_GT_F32 0x04
+#define V_008DFC_SQ_V_CMP_LG_F32 0x05
+#define V_008DFC_SQ_V_CMP_GE_F32 0x06
+#define V_008DFC_SQ_V_CMP_O_F32 0x07
+#define V_008DFC_SQ_V_CMP_U_F32 0x08
+#define V_008DFC_SQ_V_CMP_NGE_F32 0x09
+#define V_008DFC_SQ_V_CMP_NLG_F32 0x0A
+#define V_008DFC_SQ_V_CMP_NGT_F32 0x0B
+#define V_008DFC_SQ_V_CMP_NLE_F32 0x0C
+#define V_008DFC_SQ_V_CMP_NEQ_F32 0x0D
+#define V_008DFC_SQ_V_CMP_NLT_F32 0x0E
+#define V_008DFC_SQ_V_CMP_TRU_F32 0x0F
+#define V_008DFC_SQ_V_CMPX_F_F32 0x10
+#define V_008DFC_SQ_V_CMPX_LT_F32 0x11
+#define V_008DFC_SQ_V_CMPX_EQ_F32 0x12
+#define V_008DFC_SQ_V_CMPX_LE_F32 0x13
+#define V_008DFC_SQ_V_CMPX_GT_F32 0x14
+#define V_008DFC_SQ_V_CMPX_LG_F32 0x15
+#define V_008DFC_SQ_V_CMPX_GE_F32 0x16
+#define V_008DFC_SQ_V_CMPX_O_F32 0x17
+#define V_008DFC_SQ_V_CMPX_U_F32 0x18
+#define V_008DFC_SQ_V_CMPX_NGE_F32 0x19
+#define V_008DFC_SQ_V_CMPX_NLG_F32 0x1A
+#define V_008DFC_SQ_V_CMPX_NGT_F32 0x1B
+#define V_008DFC_SQ_V_CMPX_NLE_F32 0x1C
+#define V_008DFC_SQ_V_CMPX_NEQ_F32 0x1D
+#define V_008DFC_SQ_V_CMPX_NLT_F32 0x1E
+#define V_008DFC_SQ_V_CMPX_TRU_F32 0x1F
+#define V_008DFC_SQ_V_CMP_F_F64 0x20
+#define V_008DFC_SQ_V_CMP_LT_F64 0x21
+#define V_008DFC_SQ_V_CMP_EQ_F64 0x22
+#define V_008DFC_SQ_V_CMP_LE_F64 0x23
+#define V_008DFC_SQ_V_CMP_GT_F64 0x24
+#define V_008DFC_SQ_V_CMP_LG_F64 0x25
+#define V_008DFC_SQ_V_CMP_GE_F64 0x26
+#define V_008DFC_SQ_V_CMP_O_F64 0x27
+#define V_008DFC_SQ_V_CMP_U_F64 0x28
+#define V_008DFC_SQ_V_CMP_NGE_F64 0x29
+#define V_008DFC_SQ_V_CMP_NLG_F64 0x2A
+#define V_008DFC_SQ_V_CMP_NGT_F64 0x2B
+#define V_008DFC_SQ_V_CMP_NLE_F64 0x2C
+#define V_008DFC_SQ_V_CMP_NEQ_F64 0x2D
+#define V_008DFC_SQ_V_CMP_NLT_F64 0x2E
+#define V_008DFC_SQ_V_CMP_TRU_F64 0x2F
+#define V_008DFC_SQ_V_CMPX_F_F64 0x30
+#define V_008DFC_SQ_V_CMPX_LT_F64 0x31
+#define V_008DFC_SQ_V_CMPX_EQ_F64 0x32
+#define V_008DFC_SQ_V_CMPX_LE_F64 0x33
+#define V_008DFC_SQ_V_CMPX_GT_F64 0x34
+#define V_008DFC_SQ_V_CMPX_LG_F64 0x35
+#define V_008DFC_SQ_V_CMPX_GE_F64 0x36
+#define V_008DFC_SQ_V_CMPX_O_F64 0x37
+#define V_008DFC_SQ_V_CMPX_U_F64 0x38
+#define V_008DFC_SQ_V_CMPX_NGE_F64 0x39
+#define V_008DFC_SQ_V_CMPX_NLG_F64 0x3A
+#define V_008DFC_SQ_V_CMPX_NGT_F64 0x3B
+#define V_008DFC_SQ_V_CMPX_NLE_F64 0x3C
+#define V_008DFC_SQ_V_CMPX_NEQ_F64 0x3D
+#define V_008DFC_SQ_V_CMPX_NLT_F64 0x3E
+#define V_008DFC_SQ_V_CMPX_TRU_F64 0x3F
+#define V_008DFC_SQ_V_CMPS_F_F32 0x40
+#define V_008DFC_SQ_V_CMPS_LT_F32 0x41
+#define V_008DFC_SQ_V_CMPS_EQ_F32 0x42
+#define V_008DFC_SQ_V_CMPS_LE_F32 0x43
+#define V_008DFC_SQ_V_CMPS_GT_F32 0x44
+#define V_008DFC_SQ_V_CMPS_LG_F32 0x45
+#define V_008DFC_SQ_V_CMPS_GE_F32 0x46
+#define V_008DFC_SQ_V_CMPS_O_F32 0x47
+#define V_008DFC_SQ_V_CMPS_U_F32 0x48
+#define V_008DFC_SQ_V_CMPS_NGE_F32 0x49
+#define V_008DFC_SQ_V_CMPS_NLG_F32 0x4A
+#define V_008DFC_SQ_V_CMPS_NGT_F32 0x4B
+#define V_008DFC_SQ_V_CMPS_NLE_F32 0x4C
+#define V_008DFC_SQ_V_CMPS_NEQ_F32 0x4D
+#define V_008DFC_SQ_V_CMPS_NLT_F32 0x4E
+#define V_008DFC_SQ_V_CMPS_TRU_F32 0x4F
+#define V_008DFC_SQ_V_CMPSX_F_F32 0x50
+#define V_008DFC_SQ_V_CMPSX_LT_F32 0x51
+#define V_008DFC_SQ_V_CMPSX_EQ_F32 0x52
+#define V_008DFC_SQ_V_CMPSX_LE_F32 0x53
+#define V_008DFC_SQ_V_CMPSX_GT_F32 0x54
+#define V_008DFC_SQ_V_CMPSX_LG_F32 0x55
+#define V_008DFC_SQ_V_CMPSX_GE_F32 0x56
+#define V_008DFC_SQ_V_CMPSX_O_F32 0x57
+#define V_008DFC_SQ_V_CMPSX_U_F32 0x58
+#define V_008DFC_SQ_V_CMPSX_NGE_F32 0x59
+#define V_008DFC_SQ_V_CMPSX_NLG_F32 0x5A
+#define V_008DFC_SQ_V_CMPSX_NGT_F32 0x5B
+#define V_008DFC_SQ_V_CMPSX_NLE_F32 0x5C
+#define V_008DFC_SQ_V_CMPSX_NEQ_F32 0x5D
+#define V_008DFC_SQ_V_CMPSX_NLT_F32 0x5E
+#define V_008DFC_SQ_V_CMPSX_TRU_F32 0x5F
+#define V_008DFC_SQ_V_CMPS_F_F64 0x60
+#define V_008DFC_SQ_V_CMPS_LT_F64 0x61
+#define V_008DFC_SQ_V_CMPS_EQ_F64 0x62
+#define V_008DFC_SQ_V_CMPS_LE_F64 0x63
+#define V_008DFC_SQ_V_CMPS_GT_F64 0x64
+#define V_008DFC_SQ_V_CMPS_LG_F64 0x65
+#define V_008DFC_SQ_V_CMPS_GE_F64 0x66
+#define V_008DFC_SQ_V_CMPS_O_F64 0x67
+#define V_008DFC_SQ_V_CMPS_U_F64 0x68
+#define V_008DFC_SQ_V_CMPS_NGE_F64 0x69
+#define V_008DFC_SQ_V_CMPS_NLG_F64 0x6A
+#define V_008DFC_SQ_V_CMPS_NGT_F64 0x6B
+#define V_008DFC_SQ_V_CMPS_NLE_F64 0x6C
+#define V_008DFC_SQ_V_CMPS_NEQ_F64 0x6D
+#define V_008DFC_SQ_V_CMPS_NLT_F64 0x6E
+#define V_008DFC_SQ_V_CMPS_TRU_F64 0x6F
+#define V_008DFC_SQ_V_CMPSX_F_F64 0x70
+#define V_008DFC_SQ_V_CMPSX_LT_F64 0x71
+#define V_008DFC_SQ_V_CMPSX_EQ_F64 0x72
+#define V_008DFC_SQ_V_CMPSX_LE_F64 0x73
+#define V_008DFC_SQ_V_CMPSX_GT_F64 0x74
+#define V_008DFC_SQ_V_CMPSX_LG_F64 0x75
+#define V_008DFC_SQ_V_CMPSX_GE_F64 0x76
+#define V_008DFC_SQ_V_CMPSX_O_F64 0x77
+#define V_008DFC_SQ_V_CMPSX_U_F64 0x78
+#define V_008DFC_SQ_V_CMPSX_NGE_F64 0x79
+#define V_008DFC_SQ_V_CMPSX_NLG_F64 0x7A
+#define V_008DFC_SQ_V_CMPSX_NGT_F64 0x7B
+#define V_008DFC_SQ_V_CMPSX_NLE_F64 0x7C
+#define V_008DFC_SQ_V_CMPSX_NEQ_F64 0x7D
+#define V_008DFC_SQ_V_CMPSX_NLT_F64 0x7E
+#define V_008DFC_SQ_V_CMPSX_TRU_F64 0x7F
+#define V_008DFC_SQ_V_CMP_F_I32 0x80
+#define V_008DFC_SQ_V_CMP_LT_I32 0x81
+#define V_008DFC_SQ_V_CMP_EQ_I32 0x82
+#define V_008DFC_SQ_V_CMP_LE_I32 0x83
+#define V_008DFC_SQ_V_CMP_GT_I32 0x84
+#define V_008DFC_SQ_V_CMP_NE_I32 0x85
+#define V_008DFC_SQ_V_CMP_GE_I32 0x86
+#define V_008DFC_SQ_V_CMP_T_I32 0x87
+#define V_008DFC_SQ_V_CMP_CLASS_F32 0x88
+#define V_008DFC_SQ_V_CMPX_F_I32 0x90
+#define V_008DFC_SQ_V_CMPX_LT_I32 0x91
+#define V_008DFC_SQ_V_CMPX_EQ_I32 0x92
+#define V_008DFC_SQ_V_CMPX_LE_I32 0x93
+#define V_008DFC_SQ_V_CMPX_GT_I32 0x94
+#define V_008DFC_SQ_V_CMPX_NE_I32 0x95
+#define V_008DFC_SQ_V_CMPX_GE_I32 0x96
+#define V_008DFC_SQ_V_CMPX_T_I32 0x97
+#define V_008DFC_SQ_V_CMPX_CLASS_F32 0x98
+#define V_008DFC_SQ_V_CMP_F_I64 0xA0
+#define V_008DFC_SQ_V_CMP_LT_I64 0xA1
+#define V_008DFC_SQ_V_CMP_EQ_I64 0xA2
+#define V_008DFC_SQ_V_CMP_LE_I64 0xA3
+#define V_008DFC_SQ_V_CMP_GT_I64 0xA4
+#define V_008DFC_SQ_V_CMP_NE_I64 0xA5
+#define V_008DFC_SQ_V_CMP_GE_I64 0xA6
+#define V_008DFC_SQ_V_CMP_T_I64 0xA7
+#define V_008DFC_SQ_V_CMP_CLASS_F64 0xA8
+#define V_008DFC_SQ_V_CMPX_F_I64 0xB0
+#define V_008DFC_SQ_V_CMPX_LT_I64 0xB1
+#define V_008DFC_SQ_V_CMPX_EQ_I64 0xB2
+#define V_008DFC_SQ_V_CMPX_LE_I64 0xB3
+#define V_008DFC_SQ_V_CMPX_GT_I64 0xB4
+#define V_008DFC_SQ_V_CMPX_NE_I64 0xB5
+#define V_008DFC_SQ_V_CMPX_GE_I64 0xB6
+#define V_008DFC_SQ_V_CMPX_T_I64 0xB7
+#define V_008DFC_SQ_V_CMPX_CLASS_F64 0xB8
+#define V_008DFC_SQ_V_CMP_F_U32 0xC0
+#define V_008DFC_SQ_V_CMP_LT_U32 0xC1
+#define V_008DFC_SQ_V_CMP_EQ_U32 0xC2
+#define V_008DFC_SQ_V_CMP_LE_U32 0xC3
+#define V_008DFC_SQ_V_CMP_GT_U32 0xC4
+#define V_008DFC_SQ_V_CMP_NE_U32 0xC5
+#define V_008DFC_SQ_V_CMP_GE_U32 0xC6
+#define V_008DFC_SQ_V_CMP_T_U32 0xC7
+#define V_008DFC_SQ_V_CMPX_F_U32 0xD0
+#define V_008DFC_SQ_V_CMPX_LT_U32 0xD1
+#define V_008DFC_SQ_V_CMPX_EQ_U32 0xD2
+#define V_008DFC_SQ_V_CMPX_LE_U32 0xD3
+#define V_008DFC_SQ_V_CMPX_GT_U32 0xD4
+#define V_008DFC_SQ_V_CMPX_NE_U32 0xD5
+#define V_008DFC_SQ_V_CMPX_GE_U32 0xD6
+#define V_008DFC_SQ_V_CMPX_T_U32 0xD7
+#define V_008DFC_SQ_V_CMP_F_U64 0xE0
+#define V_008DFC_SQ_V_CMP_LT_U64 0xE1
+#define V_008DFC_SQ_V_CMP_EQ_U64 0xE2
+#define V_008DFC_SQ_V_CMP_LE_U64 0xE3
+#define V_008DFC_SQ_V_CMP_GT_U64 0xE4
+#define V_008DFC_SQ_V_CMP_NE_U64 0xE5
+#define V_008DFC_SQ_V_CMP_GE_U64 0xE6
+#define V_008DFC_SQ_V_CMP_T_U64 0xE7
+#define V_008DFC_SQ_V_CMPX_F_U64 0xF0
+#define V_008DFC_SQ_V_CMPX_LT_U64 0xF1
+#define V_008DFC_SQ_V_CMPX_EQ_U64 0xF2
+#define V_008DFC_SQ_V_CMPX_LE_U64 0xF3
+#define V_008DFC_SQ_V_CMPX_GT_U64 0xF4
+#define V_008DFC_SQ_V_CMPX_NE_U64 0xF5
+#define V_008DFC_SQ_V_CMPX_GE_U64 0xF6
+#define V_008DFC_SQ_V_CMPX_T_U64 0xF7
+#define S_008DFC_ENCODING(x) (((x) & 0x7F) << 25)
+#define G_008DFC_ENCODING(x) (((x) >> 25) & 0x7F)
+#define C_008DFC_ENCODING 0x01FFFFFF
+#define V_008DFC_SQ_ENC_VOPC_FIELD 0x3E
+#define R_008DFC_SQ_SOP1 0x008DFC
+#define S_008DFC_SSRC0(x) (((x) & 0xFF) << 0)
+#define G_008DFC_SSRC0(x) (((x) >> 0) & 0xFF)
+#define C_008DFC_SSRC0 0xFFFFFF00
+#define V_008DFC_SQ_SGPR 0x00
+#define V_008DFC_SQ_VCC_LO 0x6A
+#define V_008DFC_SQ_VCC_HI 0x6B
+#define V_008DFC_SQ_TBA_LO 0x6C
+#define V_008DFC_SQ_TBA_HI 0x6D
+#define V_008DFC_SQ_TMA_LO 0x6E
+#define V_008DFC_SQ_TMA_HI 0x6F
+#define V_008DFC_SQ_TTMP0 0x70
+#define V_008DFC_SQ_TTMP1 0x71
+#define V_008DFC_SQ_TTMP2 0x72
+#define V_008DFC_SQ_TTMP3 0x73
+#define V_008DFC_SQ_TTMP4 0x74
+#define V_008DFC_SQ_TTMP5 0x75
+#define V_008DFC_SQ_TTMP6 0x76
+#define V_008DFC_SQ_TTMP7 0x77
+#define V_008DFC_SQ_TTMP8 0x78
+#define V_008DFC_SQ_TTMP9 0x79
+#define V_008DFC_SQ_TTMP10 0x7A
+#define V_008DFC_SQ_TTMP11 0x7B
+#define V_008DFC_SQ_M0 0x7C
+#define V_008DFC_SQ_EXEC_LO 0x7E
+#define V_008DFC_SQ_EXEC_HI 0x7F
+#define V_008DFC_SQ_SRC_0 0x80
+#define V_008DFC_SQ_SRC_1_INT 0x81
+#define V_008DFC_SQ_SRC_2_INT 0x82
+#define V_008DFC_SQ_SRC_3_INT 0x83
+#define V_008DFC_SQ_SRC_4_INT 0x84
+#define V_008DFC_SQ_SRC_5_INT 0x85
+#define V_008DFC_SQ_SRC_6_INT 0x86
+#define V_008DFC_SQ_SRC_7_INT 0x87
+#define V_008DFC_SQ_SRC_8_INT 0x88
+#define V_008DFC_SQ_SRC_9_INT 0x89
+#define V_008DFC_SQ_SRC_10_INT 0x8A
+#define V_008DFC_SQ_SRC_11_INT 0x8B
+#define V_008DFC_SQ_SRC_12_INT 0x8C
+#define V_008DFC_SQ_SRC_13_INT 0x8D
+#define V_008DFC_SQ_SRC_14_INT 0x8E
+#define V_008DFC_SQ_SRC_15_INT 0x8F
+#define V_008DFC_SQ_SRC_16_INT 0x90
+#define V_008DFC_SQ_SRC_17_INT 0x91
+#define V_008DFC_SQ_SRC_18_INT 0x92
+#define V_008DFC_SQ_SRC_19_INT 0x93
+#define V_008DFC_SQ_SRC_20_INT 0x94
+#define V_008DFC_SQ_SRC_21_INT 0x95
+#define V_008DFC_SQ_SRC_22_INT 0x96
+#define V_008DFC_SQ_SRC_23_INT 0x97
+#define V_008DFC_SQ_SRC_24_INT 0x98
+#define V_008DFC_SQ_SRC_25_INT 0x99
+#define V_008DFC_SQ_SRC_26_INT 0x9A
+#define V_008DFC_SQ_SRC_27_INT 0x9B
+#define V_008DFC_SQ_SRC_28_INT 0x9C
+#define V_008DFC_SQ_SRC_29_INT 0x9D
+#define V_008DFC_SQ_SRC_30_INT 0x9E
+#define V_008DFC_SQ_SRC_31_INT 0x9F
+#define V_008DFC_SQ_SRC_32_INT 0xA0
+#define V_008DFC_SQ_SRC_33_INT 0xA1
+#define V_008DFC_SQ_SRC_34_INT 0xA2
+#define V_008DFC_SQ_SRC_35_INT 0xA3
+#define V_008DFC_SQ_SRC_36_INT 0xA4
+#define V_008DFC_SQ_SRC_37_INT 0xA5
+#define V_008DFC_SQ_SRC_38_INT 0xA6
+#define V_008DFC_SQ_SRC_39_INT 0xA7
+#define V_008DFC_SQ_SRC_40_INT 0xA8
+#define V_008DFC_SQ_SRC_41_INT 0xA9
+#define V_008DFC_SQ_SRC_42_INT 0xAA
+#define V_008DFC_SQ_SRC_43_INT 0xAB
+#define V_008DFC_SQ_SRC_44_INT 0xAC
+#define V_008DFC_SQ_SRC_45_INT 0xAD
+#define V_008DFC_SQ_SRC_46_INT 0xAE
+#define V_008DFC_SQ_SRC_47_INT 0xAF
+#define V_008DFC_SQ_SRC_48_INT 0xB0
+#define V_008DFC_SQ_SRC_49_INT 0xB1
+#define V_008DFC_SQ_SRC_50_INT 0xB2
+#define V_008DFC_SQ_SRC_51_INT 0xB3
+#define V_008DFC_SQ_SRC_52_INT 0xB4
+#define V_008DFC_SQ_SRC_53_INT 0xB5
+#define V_008DFC_SQ_SRC_54_INT 0xB6
+#define V_008DFC_SQ_SRC_55_INT 0xB7
+#define V_008DFC_SQ_SRC_56_INT 0xB8
+#define V_008DFC_SQ_SRC_57_INT 0xB9
+#define V_008DFC_SQ_SRC_58_INT 0xBA
+#define V_008DFC_SQ_SRC_59_INT 0xBB
+#define V_008DFC_SQ_SRC_60_INT 0xBC
+#define V_008DFC_SQ_SRC_61_INT 0xBD
+#define V_008DFC_SQ_SRC_62_INT 0xBE
+#define V_008DFC_SQ_SRC_63_INT 0xBF
+#define V_008DFC_SQ_SRC_64_INT 0xC0
+#define V_008DFC_SQ_SRC_M_1_INT 0xC1
+#define V_008DFC_SQ_SRC_M_2_INT 0xC2
+#define V_008DFC_SQ_SRC_M_3_INT 0xC3
+#define V_008DFC_SQ_SRC_M_4_INT 0xC4
+#define V_008DFC_SQ_SRC_M_5_INT 0xC5
+#define V_008DFC_SQ_SRC_M_6_INT 0xC6
+#define V_008DFC_SQ_SRC_M_7_INT 0xC7
+#define V_008DFC_SQ_SRC_M_8_INT 0xC8
+#define V_008DFC_SQ_SRC_M_9_INT 0xC9
+#define V_008DFC_SQ_SRC_M_10_INT 0xCA
+#define V_008DFC_SQ_SRC_M_11_INT 0xCB
+#define V_008DFC_SQ_SRC_M_12_INT 0xCC
+#define V_008DFC_SQ_SRC_M_13_INT 0xCD
+#define V_008DFC_SQ_SRC_M_14_INT 0xCE
+#define V_008DFC_SQ_SRC_M_15_INT 0xCF
+#define V_008DFC_SQ_SRC_M_16_INT 0xD0
+#define V_008DFC_SQ_SRC_0_5 0xF0
+#define V_008DFC_SQ_SRC_M_0_5 0xF1
+#define V_008DFC_SQ_SRC_1 0xF2
+#define V_008DFC_SQ_SRC_M_1 0xF3
+#define V_008DFC_SQ_SRC_2 0xF4
+#define V_008DFC_SQ_SRC_M_2 0xF5
+#define V_008DFC_SQ_SRC_4 0xF6
+#define V_008DFC_SQ_SRC_M_4 0xF7
+#define V_008DFC_SQ_SRC_VCCZ 0xFB
+#define V_008DFC_SQ_SRC_EXECZ 0xFC
+#define V_008DFC_SQ_SRC_SCC 0xFD
+#define V_008DFC_SQ_SRC_LDS_DIRECT 0xFE
+#define S_008DFC_OP(x) (((x) & 0xFF) << 8)
+#define G_008DFC_OP(x) (((x) >> 8) & 0xFF)
+#define C_008DFC_OP 0xFFFF00FF
+#define V_008DFC_SQ_S_MOV_B32 0x03
+#define V_008DFC_SQ_S_MOV_B64 0x04
+#define V_008DFC_SQ_S_CMOV_B32 0x05
+#define V_008DFC_SQ_S_CMOV_B64 0x06
+#define V_008DFC_SQ_S_NOT_B32 0x07
+#define V_008DFC_SQ_S_NOT_B64 0x08
+#define V_008DFC_SQ_S_WQM_B32 0x09
+#define V_008DFC_SQ_S_WQM_B64 0x0A
+#define V_008DFC_SQ_S_BREV_B32 0x0B
+#define V_008DFC_SQ_S_BREV_B64 0x0C
+#define V_008DFC_SQ_S_BCNT0_I32_B32 0x0D
+#define V_008DFC_SQ_S_BCNT0_I32_B64 0x0E
+#define V_008DFC_SQ_S_BCNT1_I32_B32 0x0F
+#define V_008DFC_SQ_S_BCNT1_I32_B64 0x10
+#define V_008DFC_SQ_S_FF0_I32_B32 0x11
+#define V_008DFC_SQ_S_FF0_I32_B64 0x12
+#define V_008DFC_SQ_S_FF1_I32_B32 0x13
+#define V_008DFC_SQ_S_FF1_I32_B64 0x14
+#define V_008DFC_SQ_S_FLBIT_I32_B32 0x15
+#define V_008DFC_SQ_S_FLBIT_I32_B64 0x16
+#define V_008DFC_SQ_S_FLBIT_I32 0x17
+#define V_008DFC_SQ_S_FLBIT_I32_I64 0x18
+#define V_008DFC_SQ_S_SEXT_I32_I8 0x19
+#define V_008DFC_SQ_S_SEXT_I32_I16 0x1A
+#define V_008DFC_SQ_S_BITSET0_B32 0x1B
+#define V_008DFC_SQ_S_BITSET0_B64 0x1C
+#define V_008DFC_SQ_S_BITSET1_B32 0x1D
+#define V_008DFC_SQ_S_BITSET1_B64 0x1E
+#define V_008DFC_SQ_S_GETPC_B64 0x1F
+#define V_008DFC_SQ_S_SETPC_B64 0x20
+#define V_008DFC_SQ_S_SWAPPC_B64 0x21
+#define V_008DFC_SQ_S_RFE_B64 0x22
+#define V_008DFC_SQ_S_AND_SAVEEXEC_B64 0x24
+#define V_008DFC_SQ_S_OR_SAVEEXEC_B64 0x25
+#define V_008DFC_SQ_S_XOR_SAVEEXEC_B64 0x26
+#define V_008DFC_SQ_S_ANDN2_SAVEEXEC_B64 0x27
+#define V_008DFC_SQ_S_ORN2_SAVEEXEC_B64 0x28
+#define V_008DFC_SQ_S_NAND_SAVEEXEC_B64 0x29
+#define V_008DFC_SQ_S_NOR_SAVEEXEC_B64 0x2A
+#define V_008DFC_SQ_S_XNOR_SAVEEXEC_B64 0x2B
+#define V_008DFC_SQ_S_QUADMASK_B32 0x2C
+#define V_008DFC_SQ_S_QUADMASK_B64 0x2D
+#define V_008DFC_SQ_S_MOVRELS_B32 0x2E
+#define V_008DFC_SQ_S_MOVRELS_B64 0x2F
+#define V_008DFC_SQ_S_MOVRELD_B32 0x30
+#define V_008DFC_SQ_S_MOVRELD_B64 0x31
+#define V_008DFC_SQ_S_CBRANCH_JOIN 0x32
+#define V_008DFC_SQ_S_MOV_REGRD_B32 0x33
+#define V_008DFC_SQ_S_ABS_I32 0x34
+#define V_008DFC_SQ_S_MOV_FED_B32 0x35
+#define S_008DFC_SDST(x) (((x) & 0x7F) << 16)
+#define G_008DFC_SDST(x) (((x) >> 16) & 0x7F)
+#define C_008DFC_SDST 0xFF80FFFF
+#define V_008DFC_SQ_SGPR 0x00
+#define V_008DFC_SQ_VCC_LO 0x6A
+#define V_008DFC_SQ_VCC_HI 0x6B
+#define V_008DFC_SQ_TBA_LO 0x6C
+#define V_008DFC_SQ_TBA_HI 0x6D
+#define V_008DFC_SQ_TMA_LO 0x6E
+#define V_008DFC_SQ_TMA_HI 0x6F
+#define V_008DFC_SQ_TTMP0 0x70
+#define V_008DFC_SQ_TTMP1 0x71
+#define V_008DFC_SQ_TTMP2 0x72
+#define V_008DFC_SQ_TTMP3 0x73
+#define V_008DFC_SQ_TTMP4 0x74
+#define V_008DFC_SQ_TTMP5 0x75
+#define V_008DFC_SQ_TTMP6 0x76
+#define V_008DFC_SQ_TTMP7 0x77
+#define V_008DFC_SQ_TTMP8 0x78
+#define V_008DFC_SQ_TTMP9 0x79
+#define V_008DFC_SQ_TTMP10 0x7A
+#define V_008DFC_SQ_TTMP11 0x7B
+#define V_008DFC_SQ_M0 0x7C
+#define V_008DFC_SQ_EXEC_LO 0x7E
+#define V_008DFC_SQ_EXEC_HI 0x7F
+#define S_008DFC_ENCODING(x) (((x) & 0x1FF) << 23)
+#define G_008DFC_ENCODING(x) (((x) >> 23) & 0x1FF)
+#define C_008DFC_ENCODING 0x007FFFFF
+#define V_008DFC_SQ_ENC_SOP1_FIELD 0x17D
+#define R_008DFC_SQ_MTBUF_1 0x008DFC
+#define S_008DFC_VADDR(x) (((x) & 0xFF) << 0)
+#define G_008DFC_VADDR(x) (((x) >> 0) & 0xFF)
+#define C_008DFC_VADDR 0xFFFFFF00
+#define V_008DFC_SQ_VGPR 0x00
+#define S_008DFC_VDATA(x) (((x) & 0xFF) << 8)
+#define G_008DFC_VDATA(x) (((x) >> 8) & 0xFF)
+#define C_008DFC_VDATA 0xFFFF00FF
+#define V_008DFC_SQ_VGPR 0x00
+#define S_008DFC_SRSRC(x) (((x) & 0x1F) << 16)
+#define G_008DFC_SRSRC(x) (((x) >> 16) & 0x1F)
+#define C_008DFC_SRSRC 0xFFE0FFFF
+#define S_008DFC_SLC(x) (((x) & 0x1) << 22)
+#define G_008DFC_SLC(x) (((x) >> 22) & 0x1)
+#define C_008DFC_SLC 0xFFBFFFFF
+#define S_008DFC_TFE(x) (((x) & 0x1) << 23)
+#define G_008DFC_TFE(x) (((x) >> 23) & 0x1)
+#define C_008DFC_TFE 0xFF7FFFFF
+#define S_008DFC_SOFFSET(x) (((x) & 0xFF) << 24)
+#define G_008DFC_SOFFSET(x) (((x) >> 24) & 0xFF)
+#define C_008DFC_SOFFSET 0x00FFFFFF
+#define V_008DFC_SQ_SGPR 0x00
+#define V_008DFC_SQ_VCC_LO 0x6A
+#define V_008DFC_SQ_VCC_HI 0x6B
+#define V_008DFC_SQ_TBA_LO 0x6C
+#define V_008DFC_SQ_TBA_HI 0x6D
+#define V_008DFC_SQ_TMA_LO 0x6E
+#define V_008DFC_SQ_TMA_HI 0x6F
+#define V_008DFC_SQ_TTMP0 0x70
+#define V_008DFC_SQ_TTMP1 0x71
+#define V_008DFC_SQ_TTMP2 0x72
+#define V_008DFC_SQ_TTMP3 0x73
+#define V_008DFC_SQ_TTMP4 0x74
+#define V_008DFC_SQ_TTMP5 0x75
+#define V_008DFC_SQ_TTMP6 0x76
+#define V_008DFC_SQ_TTMP7 0x77
+#define V_008DFC_SQ_TTMP8 0x78
+#define V_008DFC_SQ_TTMP9 0x79
+#define V_008DFC_SQ_TTMP10 0x7A
+#define V_008DFC_SQ_TTMP11 0x7B
+#define V_008DFC_SQ_M0 0x7C
+#define V_008DFC_SQ_EXEC_LO 0x7E
+#define V_008DFC_SQ_EXEC_HI 0x7F
+#define V_008DFC_SQ_SRC_0 0x80
+#define V_008DFC_SQ_SRC_1_INT 0x81
+#define V_008DFC_SQ_SRC_2_INT 0x82
+#define V_008DFC_SQ_SRC_3_INT 0x83
+#define V_008DFC_SQ_SRC_4_INT 0x84
+#define V_008DFC_SQ_SRC_5_INT 0x85
+#define V_008DFC_SQ_SRC_6_INT 0x86
+#define V_008DFC_SQ_SRC_7_INT 0x87
+#define V_008DFC_SQ_SRC_8_INT 0x88
+#define V_008DFC_SQ_SRC_9_INT 0x89
+#define V_008DFC_SQ_SRC_10_INT 0x8A
+#define V_008DFC_SQ_SRC_11_INT 0x8B
+#define V_008DFC_SQ_SRC_12_INT 0x8C
+#define V_008DFC_SQ_SRC_13_INT 0x8D
+#define V_008DFC_SQ_SRC_14_INT 0x8E
+#define V_008DFC_SQ_SRC_15_INT 0x8F
+#define V_008DFC_SQ_SRC_16_INT 0x90
+#define V_008DFC_SQ_SRC_17_INT 0x91
+#define V_008DFC_SQ_SRC_18_INT 0x92
+#define V_008DFC_SQ_SRC_19_INT 0x93
+#define V_008DFC_SQ_SRC_20_INT 0x94
+#define V_008DFC_SQ_SRC_21_INT 0x95
+#define V_008DFC_SQ_SRC_22_INT 0x96
+#define V_008DFC_SQ_SRC_23_INT 0x97
+#define V_008DFC_SQ_SRC_24_INT 0x98
+#define V_008DFC_SQ_SRC_25_INT 0x99
+#define V_008DFC_SQ_SRC_26_INT 0x9A
+#define V_008DFC_SQ_SRC_27_INT 0x9B
+#define V_008DFC_SQ_SRC_28_INT 0x9C
+#define V_008DFC_SQ_SRC_29_INT 0x9D
+#define V_008DFC_SQ_SRC_30_INT 0x9E
+#define V_008DFC_SQ_SRC_31_INT 0x9F
+#define V_008DFC_SQ_SRC_32_INT 0xA0
+#define V_008DFC_SQ_SRC_33_INT 0xA1
+#define V_008DFC_SQ_SRC_34_INT 0xA2
+#define V_008DFC_SQ_SRC_35_INT 0xA3
+#define V_008DFC_SQ_SRC_36_INT 0xA4
+#define V_008DFC_SQ_SRC_37_INT 0xA5
+#define V_008DFC_SQ_SRC_38_INT 0xA6
+#define V_008DFC_SQ_SRC_39_INT 0xA7
+#define V_008DFC_SQ_SRC_40_INT 0xA8
+#define V_008DFC_SQ_SRC_41_INT 0xA9
+#define V_008DFC_SQ_SRC_42_INT 0xAA
+#define V_008DFC_SQ_SRC_43_INT 0xAB
+#define V_008DFC_SQ_SRC_44_INT 0xAC
+#define V_008DFC_SQ_SRC_45_INT 0xAD
+#define V_008DFC_SQ_SRC_46_INT 0xAE
+#define V_008DFC_SQ_SRC_47_INT 0xAF
+#define V_008DFC_SQ_SRC_48_INT 0xB0
+#define V_008DFC_SQ_SRC_49_INT 0xB1
+#define V_008DFC_SQ_SRC_50_INT 0xB2
+#define V_008DFC_SQ_SRC_51_INT 0xB3
+#define V_008DFC_SQ_SRC_52_INT 0xB4
+#define V_008DFC_SQ_SRC_53_INT 0xB5
+#define V_008DFC_SQ_SRC_54_INT 0xB6
+#define V_008DFC_SQ_SRC_55_INT 0xB7
+#define V_008DFC_SQ_SRC_56_INT 0xB8
+#define V_008DFC_SQ_SRC_57_INT 0xB9
+#define V_008DFC_SQ_SRC_58_INT 0xBA
+#define V_008DFC_SQ_SRC_59_INT 0xBB
+#define V_008DFC_SQ_SRC_60_INT 0xBC
+#define V_008DFC_SQ_SRC_61_INT 0xBD
+#define V_008DFC_SQ_SRC_62_INT 0xBE
+#define V_008DFC_SQ_SRC_63_INT 0xBF
+#define V_008DFC_SQ_SRC_64_INT 0xC0
+#define V_008DFC_SQ_SRC_M_1_INT 0xC1
+#define V_008DFC_SQ_SRC_M_2_INT 0xC2
+#define V_008DFC_SQ_SRC_M_3_INT 0xC3
+#define V_008DFC_SQ_SRC_M_4_INT 0xC4
+#define V_008DFC_SQ_SRC_M_5_INT 0xC5
+#define V_008DFC_SQ_SRC_M_6_INT 0xC6
+#define V_008DFC_SQ_SRC_M_7_INT 0xC7
+#define V_008DFC_SQ_SRC_M_8_INT 0xC8
+#define V_008DFC_SQ_SRC_M_9_INT 0xC9
+#define V_008DFC_SQ_SRC_M_10_INT 0xCA
+#define V_008DFC_SQ_SRC_M_11_INT 0xCB
+#define V_008DFC_SQ_SRC_M_12_INT 0xCC
+#define V_008DFC_SQ_SRC_M_13_INT 0xCD
+#define V_008DFC_SQ_SRC_M_14_INT 0xCE
+#define V_008DFC_SQ_SRC_M_15_INT 0xCF
+#define V_008DFC_SQ_SRC_M_16_INT 0xD0
+#define V_008DFC_SQ_SRC_0_5 0xF0
+#define V_008DFC_SQ_SRC_M_0_5 0xF1
+#define V_008DFC_SQ_SRC_1 0xF2
+#define V_008DFC_SQ_SRC_M_1 0xF3
+#define V_008DFC_SQ_SRC_2 0xF4
+#define V_008DFC_SQ_SRC_M_2 0xF5
+#define V_008DFC_SQ_SRC_4 0xF6
+#define V_008DFC_SQ_SRC_M_4 0xF7
+#define V_008DFC_SQ_SRC_VCCZ 0xFB
+#define V_008DFC_SQ_SRC_EXECZ 0xFC
+#define V_008DFC_SQ_SRC_SCC 0xFD
+#define V_008DFC_SQ_SRC_LDS_DIRECT 0xFE
+#define R_008DFC_SQ_SOP2 0x008DFC
+#define S_008DFC_SSRC0(x) (((x) & 0xFF) << 0)
+#define G_008DFC_SSRC0(x) (((x) >> 0) & 0xFF)
+#define C_008DFC_SSRC0 0xFFFFFF00
+#define V_008DFC_SQ_SGPR 0x00
+#define V_008DFC_SQ_VCC_LO 0x6A
+#define V_008DFC_SQ_VCC_HI 0x6B
+#define V_008DFC_SQ_TBA_LO 0x6C
+#define V_008DFC_SQ_TBA_HI 0x6D
+#define V_008DFC_SQ_TMA_LO 0x6E
+#define V_008DFC_SQ_TMA_HI 0x6F
+#define V_008DFC_SQ_TTMP0 0x70
+#define V_008DFC_SQ_TTMP1 0x71
+#define V_008DFC_SQ_TTMP2 0x72
+#define V_008DFC_SQ_TTMP3 0x73
+#define V_008DFC_SQ_TTMP4 0x74
+#define V_008DFC_SQ_TTMP5 0x75
+#define V_008DFC_SQ_TTMP6 0x76
+#define V_008DFC_SQ_TTMP7 0x77
+#define V_008DFC_SQ_TTMP8 0x78
+#define V_008DFC_SQ_TTMP9 0x79
+#define V_008DFC_SQ_TTMP10 0x7A
+#define V_008DFC_SQ_TTMP11 0x7B
+#define V_008DFC_SQ_M0 0x7C
+#define V_008DFC_SQ_EXEC_LO 0x7E
+#define V_008DFC_SQ_EXEC_HI 0x7F
+#define V_008DFC_SQ_SRC_0 0x80
+#define V_008DFC_SQ_SRC_1_INT 0x81
+#define V_008DFC_SQ_SRC_2_INT 0x82
+#define V_008DFC_SQ_SRC_3_INT 0x83
+#define V_008DFC_SQ_SRC_4_INT 0x84
+#define V_008DFC_SQ_SRC_5_INT 0x85
+#define V_008DFC_SQ_SRC_6_INT 0x86
+#define V_008DFC_SQ_SRC_7_INT 0x87
+#define V_008DFC_SQ_SRC_8_INT 0x88
+#define V_008DFC_SQ_SRC_9_INT 0x89
+#define V_008DFC_SQ_SRC_10_INT 0x8A
+#define V_008DFC_SQ_SRC_11_INT 0x8B
+#define V_008DFC_SQ_SRC_12_INT 0x8C
+#define V_008DFC_SQ_SRC_13_INT 0x8D
+#define V_008DFC_SQ_SRC_14_INT 0x8E
+#define V_008DFC_SQ_SRC_15_INT 0x8F
+#define V_008DFC_SQ_SRC_16_INT 0x90
+#define V_008DFC_SQ_SRC_17_INT 0x91
+#define V_008DFC_SQ_SRC_18_INT 0x92
+#define V_008DFC_SQ_SRC_19_INT 0x93
+#define V_008DFC_SQ_SRC_20_INT 0x94
+#define V_008DFC_SQ_SRC_21_INT 0x95
+#define V_008DFC_SQ_SRC_22_INT 0x96
+#define V_008DFC_SQ_SRC_23_INT 0x97
+#define V_008DFC_SQ_SRC_24_INT 0x98
+#define V_008DFC_SQ_SRC_25_INT 0x99
+#define V_008DFC_SQ_SRC_26_INT 0x9A
+#define V_008DFC_SQ_SRC_27_INT 0x9B
+#define V_008DFC_SQ_SRC_28_INT 0x9C
+#define V_008DFC_SQ_SRC_29_INT 0x9D
+#define V_008DFC_SQ_SRC_30_INT 0x9E
+#define V_008DFC_SQ_SRC_31_INT 0x9F
+#define V_008DFC_SQ_SRC_32_INT 0xA0
+#define V_008DFC_SQ_SRC_33_INT 0xA1
+#define V_008DFC_SQ_SRC_34_INT 0xA2
+#define V_008DFC_SQ_SRC_35_INT 0xA3
+#define V_008DFC_SQ_SRC_36_INT 0xA4
+#define V_008DFC_SQ_SRC_37_INT 0xA5
+#define V_008DFC_SQ_SRC_38_INT 0xA6
+#define V_008DFC_SQ_SRC_39_INT 0xA7
+#define V_008DFC_SQ_SRC_40_INT 0xA8
+#define V_008DFC_SQ_SRC_41_INT 0xA9
+#define V_008DFC_SQ_SRC_42_INT 0xAA
+#define V_008DFC_SQ_SRC_43_INT 0xAB
+#define V_008DFC_SQ_SRC_44_INT 0xAC
+#define V_008DFC_SQ_SRC_45_INT 0xAD
+#define V_008DFC_SQ_SRC_46_INT 0xAE
+#define V_008DFC_SQ_SRC_47_INT 0xAF
+#define V_008DFC_SQ_SRC_48_INT 0xB0
+#define V_008DFC_SQ_SRC_49_INT 0xB1
+#define V_008DFC_SQ_SRC_50_INT 0xB2
+#define V_008DFC_SQ_SRC_51_INT 0xB3
+#define V_008DFC_SQ_SRC_52_INT 0xB4
+#define V_008DFC_SQ_SRC_53_INT 0xB5
+#define V_008DFC_SQ_SRC_54_INT 0xB6
+#define V_008DFC_SQ_SRC_55_INT 0xB7
+#define V_008DFC_SQ_SRC_56_INT 0xB8
+#define V_008DFC_SQ_SRC_57_INT 0xB9
+#define V_008DFC_SQ_SRC_58_INT 0xBA
+#define V_008DFC_SQ_SRC_59_INT 0xBB
+#define V_008DFC_SQ_SRC_60_INT 0xBC
+#define V_008DFC_SQ_SRC_61_INT 0xBD
+#define V_008DFC_SQ_SRC_62_INT 0xBE
+#define V_008DFC_SQ_SRC_63_INT 0xBF
+#define V_008DFC_SQ_SRC_64_INT 0xC0
+#define V_008DFC_SQ_SRC_M_1_INT 0xC1
+#define V_008DFC_SQ_SRC_M_2_INT 0xC2
+#define V_008DFC_SQ_SRC_M_3_INT 0xC3
+#define V_008DFC_SQ_SRC_M_4_INT 0xC4
+#define V_008DFC_SQ_SRC_M_5_INT 0xC5
+#define V_008DFC_SQ_SRC_M_6_INT 0xC6
+#define V_008DFC_SQ_SRC_M_7_INT 0xC7
+#define V_008DFC_SQ_SRC_M_8_INT 0xC8
+#define V_008DFC_SQ_SRC_M_9_INT 0xC9
+#define V_008DFC_SQ_SRC_M_10_INT 0xCA
+#define V_008DFC_SQ_SRC_M_11_INT 0xCB
+#define V_008DFC_SQ_SRC_M_12_INT 0xCC
+#define V_008DFC_SQ_SRC_M_13_INT 0xCD
+#define V_008DFC_SQ_SRC_M_14_INT 0xCE
+#define V_008DFC_SQ_SRC_M_15_INT 0xCF
+#define V_008DFC_SQ_SRC_M_16_INT 0xD0
+#define V_008DFC_SQ_SRC_0_5 0xF0
+#define V_008DFC_SQ_SRC_M_0_5 0xF1
+#define V_008DFC_SQ_SRC_1 0xF2
+#define V_008DFC_SQ_SRC_M_1 0xF3
+#define V_008DFC_SQ_SRC_2 0xF4
+#define V_008DFC_SQ_SRC_M_2 0xF5
+#define V_008DFC_SQ_SRC_4 0xF6
+#define V_008DFC_SQ_SRC_M_4 0xF7
+#define V_008DFC_SQ_SRC_VCCZ 0xFB
+#define V_008DFC_SQ_SRC_EXECZ 0xFC
+#define V_008DFC_SQ_SRC_SCC 0xFD
+#define V_008DFC_SQ_SRC_LDS_DIRECT 0xFE
+#define S_008DFC_SSRC1(x) (((x) & 0xFF) << 8)
+#define G_008DFC_SSRC1(x) (((x) >> 8) & 0xFF)
+#define C_008DFC_SSRC1 0xFFFF00FF
+#define V_008DFC_SQ_SGPR 0x00
+#define V_008DFC_SQ_VCC_LO 0x6A
+#define V_008DFC_SQ_VCC_HI 0x6B
+#define V_008DFC_SQ_TBA_LO 0x6C
+#define V_008DFC_SQ_TBA_HI 0x6D
+#define V_008DFC_SQ_TMA_LO 0x6E
+#define V_008DFC_SQ_TMA_HI 0x6F
+#define V_008DFC_SQ_TTMP0 0x70
+#define V_008DFC_SQ_TTMP1 0x71
+#define V_008DFC_SQ_TTMP2 0x72
+#define V_008DFC_SQ_TTMP3 0x73
+#define V_008DFC_SQ_TTMP4 0x74
+#define V_008DFC_SQ_TTMP5 0x75
+#define V_008DFC_SQ_TTMP6 0x76
+#define V_008DFC_SQ_TTMP7 0x77
+#define V_008DFC_SQ_TTMP8 0x78
+#define V_008DFC_SQ_TTMP9 0x79
+#define V_008DFC_SQ_TTMP10 0x7A
+#define V_008DFC_SQ_TTMP11 0x7B
+#define V_008DFC_SQ_M0 0x7C
+#define V_008DFC_SQ_EXEC_LO 0x7E
+#define V_008DFC_SQ_EXEC_HI 0x7F
+#define V_008DFC_SQ_SRC_0 0x80
+#define V_008DFC_SQ_SRC_1_INT 0x81
+#define V_008DFC_SQ_SRC_2_INT 0x82
+#define V_008DFC_SQ_SRC_3_INT 0x83
+#define V_008DFC_SQ_SRC_4_INT 0x84
+#define V_008DFC_SQ_SRC_5_INT 0x85
+#define V_008DFC_SQ_SRC_6_INT 0x86
+#define V_008DFC_SQ_SRC_7_INT 0x87
+#define V_008DFC_SQ_SRC_8_INT 0x88
+#define V_008DFC_SQ_SRC_9_INT 0x89
+#define V_008DFC_SQ_SRC_10_INT 0x8A
+#define V_008DFC_SQ_SRC_11_INT 0x8B
+#define V_008DFC_SQ_SRC_12_INT 0x8C
+#define V_008DFC_SQ_SRC_13_INT 0x8D
+#define V_008DFC_SQ_SRC_14_INT 0x8E
+#define V_008DFC_SQ_SRC_15_INT 0x8F
+#define V_008DFC_SQ_SRC_16_INT 0x90
+#define V_008DFC_SQ_SRC_17_INT 0x91
+#define V_008DFC_SQ_SRC_18_INT 0x92
+#define V_008DFC_SQ_SRC_19_INT 0x93
+#define V_008DFC_SQ_SRC_20_INT 0x94
+#define V_008DFC_SQ_SRC_21_INT 0x95
+#define V_008DFC_SQ_SRC_22_INT 0x96
+#define V_008DFC_SQ_SRC_23_INT 0x97
+#define V_008DFC_SQ_SRC_24_INT 0x98
+#define V_008DFC_SQ_SRC_25_INT 0x99
+#define V_008DFC_SQ_SRC_26_INT 0x9A
+#define V_008DFC_SQ_SRC_27_INT 0x9B
+#define V_008DFC_SQ_SRC_28_INT 0x9C
+#define V_008DFC_SQ_SRC_29_INT 0x9D
+#define V_008DFC_SQ_SRC_30_INT 0x9E
+#define V_008DFC_SQ_SRC_31_INT 0x9F
+#define V_008DFC_SQ_SRC_32_INT 0xA0
+#define V_008DFC_SQ_SRC_33_INT 0xA1
+#define V_008DFC_SQ_SRC_34_INT 0xA2
+#define V_008DFC_SQ_SRC_35_INT 0xA3
+#define V_008DFC_SQ_SRC_36_INT 0xA4
+#define V_008DFC_SQ_SRC_37_INT 0xA5
+#define V_008DFC_SQ_SRC_38_INT 0xA6
+#define V_008DFC_SQ_SRC_39_INT 0xA7
+#define V_008DFC_SQ_SRC_40_INT 0xA8
+#define V_008DFC_SQ_SRC_41_INT 0xA9
+#define V_008DFC_SQ_SRC_42_INT 0xAA
+#define V_008DFC_SQ_SRC_43_INT 0xAB
+#define V_008DFC_SQ_SRC_44_INT 0xAC
+#define V_008DFC_SQ_SRC_45_INT 0xAD
+#define V_008DFC_SQ_SRC_46_INT 0xAE
+#define V_008DFC_SQ_SRC_47_INT 0xAF
+#define V_008DFC_SQ_SRC_48_INT 0xB0
+#define V_008DFC_SQ_SRC_49_INT 0xB1
+#define V_008DFC_SQ_SRC_50_INT 0xB2
+#define V_008DFC_SQ_SRC_51_INT 0xB3
+#define V_008DFC_SQ_SRC_52_INT 0xB4
+#define V_008DFC_SQ_SRC_53_INT 0xB5
+#define V_008DFC_SQ_SRC_54_INT 0xB6
+#define V_008DFC_SQ_SRC_55_INT 0xB7
+#define V_008DFC_SQ_SRC_56_INT 0xB8
+#define V_008DFC_SQ_SRC_57_INT 0xB9
+#define V_008DFC_SQ_SRC_58_INT 0xBA
+#define V_008DFC_SQ_SRC_59_INT 0xBB
+#define V_008DFC_SQ_SRC_60_INT 0xBC
+#define V_008DFC_SQ_SRC_61_INT 0xBD
+#define V_008DFC_SQ_SRC_62_INT 0xBE
+#define V_008DFC_SQ_SRC_63_INT 0xBF
+#define V_008DFC_SQ_SRC_64_INT 0xC0
+#define V_008DFC_SQ_SRC_M_1_INT 0xC1
+#define V_008DFC_SQ_SRC_M_2_INT 0xC2
+#define V_008DFC_SQ_SRC_M_3_INT 0xC3
+#define V_008DFC_SQ_SRC_M_4_INT 0xC4
+#define V_008DFC_SQ_SRC_M_5_INT 0xC5
+#define V_008DFC_SQ_SRC_M_6_INT 0xC6
+#define V_008DFC_SQ_SRC_M_7_INT 0xC7
+#define V_008DFC_SQ_SRC_M_8_INT 0xC8
+#define V_008DFC_SQ_SRC_M_9_INT 0xC9
+#define V_008DFC_SQ_SRC_M_10_INT 0xCA
+#define V_008DFC_SQ_SRC_M_11_INT 0xCB
+#define V_008DFC_SQ_SRC_M_12_INT 0xCC
+#define V_008DFC_SQ_SRC_M_13_INT 0xCD
+#define V_008DFC_SQ_SRC_M_14_INT 0xCE
+#define V_008DFC_SQ_SRC_M_15_INT 0xCF
+#define V_008DFC_SQ_SRC_M_16_INT 0xD0
+#define V_008DFC_SQ_SRC_0_5 0xF0
+#define V_008DFC_SQ_SRC_M_0_5 0xF1
+#define V_008DFC_SQ_SRC_1 0xF2
+#define V_008DFC_SQ_SRC_M_1 0xF3
+#define V_008DFC_SQ_SRC_2 0xF4
+#define V_008DFC_SQ_SRC_M_2 0xF5
+#define V_008DFC_SQ_SRC_4 0xF6
+#define V_008DFC_SQ_SRC_M_4 0xF7
+#define V_008DFC_SQ_SRC_VCCZ 0xFB
+#define V_008DFC_SQ_SRC_EXECZ 0xFC
+#define V_008DFC_SQ_SRC_SCC 0xFD
+#define V_008DFC_SQ_SRC_LDS_DIRECT 0xFE
+#define S_008DFC_SDST(x) (((x) & 0x7F) << 16)
+#define G_008DFC_SDST(x) (((x) >> 16) & 0x7F)
+#define C_008DFC_SDST 0xFF80FFFF
+#define V_008DFC_SQ_SGPR 0x00
+#define V_008DFC_SQ_VCC_LO 0x6A
+#define V_008DFC_SQ_VCC_HI 0x6B
+#define V_008DFC_SQ_TBA_LO 0x6C
+#define V_008DFC_SQ_TBA_HI 0x6D
+#define V_008DFC_SQ_TMA_LO 0x6E
+#define V_008DFC_SQ_TMA_HI 0x6F
+#define V_008DFC_SQ_TTMP0 0x70
+#define V_008DFC_SQ_TTMP1 0x71
+#define V_008DFC_SQ_TTMP2 0x72
+#define V_008DFC_SQ_TTMP3 0x73
+#define V_008DFC_SQ_TTMP4 0x74
+#define V_008DFC_SQ_TTMP5 0x75
+#define V_008DFC_SQ_TTMP6 0x76
+#define V_008DFC_SQ_TTMP7 0x77
+#define V_008DFC_SQ_TTMP8 0x78
+#define V_008DFC_SQ_TTMP9 0x79
+#define V_008DFC_SQ_TTMP10 0x7A
+#define V_008DFC_SQ_TTMP11 0x7B
+#define V_008DFC_SQ_M0 0x7C
+#define V_008DFC_SQ_EXEC_LO 0x7E
+#define V_008DFC_SQ_EXEC_HI 0x7F
+#define S_008DFC_OP(x) (((x) & 0x7F) << 23)
+#define G_008DFC_OP(x) (((x) >> 23) & 0x7F)
+#define C_008DFC_OP 0xC07FFFFF
+#define V_008DFC_SQ_S_ADD_U32 0x00
+#define V_008DFC_SQ_S_SUB_U32 0x01
+#define V_008DFC_SQ_S_ADD_I32 0x02
+#define V_008DFC_SQ_S_SUB_I32 0x03
+#define V_008DFC_SQ_S_ADDC_U32 0x04
+#define V_008DFC_SQ_S_SUBB_U32 0x05
+#define V_008DFC_SQ_S_MIN_I32 0x06
+#define V_008DFC_SQ_S_MIN_U32 0x07
+#define V_008DFC_SQ_S_MAX_I32 0x08
+#define V_008DFC_SQ_S_MAX_U32 0x09
+#define V_008DFC_SQ_S_CSELECT_B32 0x0A
+#define V_008DFC_SQ_S_CSELECT_B64 0x0B
+#define V_008DFC_SQ_S_AND_B32 0x0E
+#define V_008DFC_SQ_S_AND_B64 0x0F
+#define V_008DFC_SQ_S_OR_B32 0x10
+#define V_008DFC_SQ_S_OR_B64 0x11
+#define V_008DFC_SQ_S_XOR_B32 0x12
+#define V_008DFC_SQ_S_XOR_B64 0x13
+#define V_008DFC_SQ_S_ANDN2_B32 0x14
+#define V_008DFC_SQ_S_ANDN2_B64 0x15
+#define V_008DFC_SQ_S_ORN2_B32 0x16
+#define V_008DFC_SQ_S_ORN2_B64 0x17
+#define V_008DFC_SQ_S_NAND_B32 0x18
+#define V_008DFC_SQ_S_NAND_B64 0x19
+#define V_008DFC_SQ_S_NOR_B32 0x1A
+#define V_008DFC_SQ_S_NOR_B64 0x1B
+#define V_008DFC_SQ_S_XNOR_B32 0x1C
+#define V_008DFC_SQ_S_XNOR_B64 0x1D
+#define V_008DFC_SQ_S_LSHL_B32 0x1E
+#define V_008DFC_SQ_S_LSHL_B64 0x1F
+#define V_008DFC_SQ_S_LSHR_B32 0x20
+#define V_008DFC_SQ_S_LSHR_B64 0x21
+#define V_008DFC_SQ_S_ASHR_I32 0x22
+#define V_008DFC_SQ_S_ASHR_I64 0x23
+#define V_008DFC_SQ_S_BFM_B32 0x24
+#define V_008DFC_SQ_S_BFM_B64 0x25
+#define V_008DFC_SQ_S_MUL_I32 0x26
+#define V_008DFC_SQ_S_BFE_U32 0x27
+#define V_008DFC_SQ_S_BFE_I32 0x28
+#define V_008DFC_SQ_S_BFE_U64 0x29
+#define V_008DFC_SQ_S_BFE_I64 0x2A
+#define V_008DFC_SQ_S_CBRANCH_G_FORK 0x2B
+#define V_008DFC_SQ_S_ABSDIFF_I32 0x2C
+#define S_008DFC_ENCODING(x) (((x) & 0x03) << 30)
+#define G_008DFC_ENCODING(x) (((x) >> 30) & 0x03)
+#define C_008DFC_ENCODING 0x3FFFFFFF
+#define V_008DFC_SQ_ENC_SOP2_FIELD 0x02
+#define R_008DFC_SQ_SOPK 0x008DFC
+#define S_008DFC_SIMM16(x) (((x) & 0xFFFF) << 0)
+#define G_008DFC_SIMM16(x) (((x) >> 0) & 0xFFFF)
+#define C_008DFC_SIMM16 0xFFFF0000
+#define S_008DFC_SDST(x) (((x) & 0x7F) << 16)
+#define G_008DFC_SDST(x) (((x) >> 16) & 0x7F)
+#define C_008DFC_SDST 0xFF80FFFF
+#define V_008DFC_SQ_SGPR 0x00
+#define V_008DFC_SQ_VCC_LO 0x6A
+#define V_008DFC_SQ_VCC_HI 0x6B
+#define V_008DFC_SQ_TBA_LO 0x6C
+#define V_008DFC_SQ_TBA_HI 0x6D
+#define V_008DFC_SQ_TMA_LO 0x6E
+#define V_008DFC_SQ_TMA_HI 0x6F
+#define V_008DFC_SQ_TTMP0 0x70
+#define V_008DFC_SQ_TTMP1 0x71
+#define V_008DFC_SQ_TTMP2 0x72
+#define V_008DFC_SQ_TTMP3 0x73
+#define V_008DFC_SQ_TTMP4 0x74
+#define V_008DFC_SQ_TTMP5 0x75
+#define V_008DFC_SQ_TTMP6 0x76
+#define V_008DFC_SQ_TTMP7 0x77
+#define V_008DFC_SQ_TTMP8 0x78
+#define V_008DFC_SQ_TTMP9 0x79
+#define V_008DFC_SQ_TTMP10 0x7A
+#define V_008DFC_SQ_TTMP11 0x7B
+#define V_008DFC_SQ_M0 0x7C
+#define V_008DFC_SQ_EXEC_LO 0x7E
+#define V_008DFC_SQ_EXEC_HI 0x7F
+#define S_008DFC_OP(x) (((x) & 0x1F) << 23)
+#define G_008DFC_OP(x) (((x) >> 23) & 0x1F)
+#define C_008DFC_OP 0xF07FFFFF
+#define V_008DFC_SQ_S_MOVK_I32 0x00
+#define V_008DFC_SQ_S_CMOVK_I32 0x02
+#define V_008DFC_SQ_S_CMPK_EQ_I32 0x03
+#define V_008DFC_SQ_S_CMPK_LG_I32 0x04
+#define V_008DFC_SQ_S_CMPK_GT_I32 0x05
+#define V_008DFC_SQ_S_CMPK_GE_I32 0x06
+#define V_008DFC_SQ_S_CMPK_LT_I32 0x07
+#define V_008DFC_SQ_S_CMPK_LE_I32 0x08
+#define V_008DFC_SQ_S_CMPK_EQ_U32 0x09
+#define V_008DFC_SQ_S_CMPK_LG_U32 0x0A
+#define V_008DFC_SQ_S_CMPK_GT_U32 0x0B
+#define V_008DFC_SQ_S_CMPK_GE_U32 0x0C
+#define V_008DFC_SQ_S_CMPK_LT_U32 0x0D
+#define V_008DFC_SQ_S_CMPK_LE_U32 0x0E
+#define V_008DFC_SQ_S_ADDK_I32 0x0F
+#define V_008DFC_SQ_S_MULK_I32 0x10
+#define V_008DFC_SQ_S_CBRANCH_I_FORK 0x11
+#define V_008DFC_SQ_S_GETREG_B32 0x12
+#define V_008DFC_SQ_S_SETREG_B32 0x13
+#define V_008DFC_SQ_S_GETREG_REGRD_B32 0x14
+#define V_008DFC_SQ_S_SETREG_IMM32_B32 0x15
+#define S_008DFC_ENCODING(x) (((x) & 0x0F) << 28)
+#define G_008DFC_ENCODING(x) (((x) >> 28) & 0x0F)
+#define C_008DFC_ENCODING 0x0FFFFFFF
+#define V_008DFC_SQ_ENC_SOPK_FIELD 0x0B
+#define R_008DFC_SQ_VOP3_0 0x008DFC
+#define S_008DFC_VDST(x) (((x) & 0xFF) << 0)
+#define G_008DFC_VDST(x) (((x) >> 0) & 0xFF)
+#define C_008DFC_VDST 0xFFFFFF00
+#define V_008DFC_SQ_VGPR 0x00
+#define S_008DFC_ABS(x) (((x) & 0x07) << 8)
+#define G_008DFC_ABS(x) (((x) >> 8) & 0x07)
+#define C_008DFC_ABS 0xFFFFF8FF
+#define S_008DFC_CLAMP(x) (((x) & 0x1) << 11)
+#define G_008DFC_CLAMP(x) (((x) >> 11) & 0x1)
+#define C_008DFC_CLAMP 0xFFFFF7FF
+#define S_008DFC_OP(x) (((x) & 0x1FF) << 17)
+#define G_008DFC_OP(x) (((x) >> 17) & 0x1FF)
+#define C_008DFC_OP 0xFC01FFFF
+#define V_008DFC_SQ_V_OPC_OFFSET 0x00
+#define V_008DFC_SQ_V_OP2_OFFSET 0x100
+#define V_008DFC_SQ_V_MAD_LEGACY_F32 0x140
+#define V_008DFC_SQ_V_MAD_F32 0x141
+#define V_008DFC_SQ_V_MAD_I32_I24 0x142
+#define V_008DFC_SQ_V_MAD_U32_U24 0x143
+#define V_008DFC_SQ_V_CUBEID_F32 0x144
+#define V_008DFC_SQ_V_CUBESC_F32 0x145
+#define V_008DFC_SQ_V_CUBETC_F32 0x146
+#define V_008DFC_SQ_V_CUBEMA_F32 0x147
+#define V_008DFC_SQ_V_BFE_U32 0x148
+#define V_008DFC_SQ_V_BFE_I32 0x149
+#define V_008DFC_SQ_V_BFI_B32 0x14A
+#define V_008DFC_SQ_V_FMA_F32 0x14B
+#define V_008DFC_SQ_V_FMA_F64 0x14C
+#define V_008DFC_SQ_V_LERP_U8 0x14D
+#define V_008DFC_SQ_V_ALIGNBIT_B32 0x14E
+#define V_008DFC_SQ_V_ALIGNBYTE_B32 0x14F
+#define V_008DFC_SQ_V_MULLIT_F32 0x150
+#define V_008DFC_SQ_V_MIN3_F32 0x151
+#define V_008DFC_SQ_V_MIN3_I32 0x152
+#define V_008DFC_SQ_V_MIN3_U32 0x153
+#define V_008DFC_SQ_V_MAX3_F32 0x154
+#define V_008DFC_SQ_V_MAX3_I32 0x155
+#define V_008DFC_SQ_V_MAX3_U32 0x156
+#define V_008DFC_SQ_V_MED3_F32 0x157
+#define V_008DFC_SQ_V_MED3_I32 0x158
+#define V_008DFC_SQ_V_MED3_U32 0x159
+#define V_008DFC_SQ_V_SAD_U8 0x15A
+#define V_008DFC_SQ_V_SAD_HI_U8 0x15B
+#define V_008DFC_SQ_V_SAD_U16 0x15C
+#define V_008DFC_SQ_V_SAD_U32 0x15D
+#define V_008DFC_SQ_V_CVT_PK_U8_F32 0x15E
+#define V_008DFC_SQ_V_DIV_FIXUP_F32 0x15F
+#define V_008DFC_SQ_V_DIV_FIXUP_F64 0x160
+#define V_008DFC_SQ_V_LSHL_B64 0x161
+#define V_008DFC_SQ_V_LSHR_B64 0x162
+#define V_008DFC_SQ_V_ASHR_I64 0x163
+#define V_008DFC_SQ_V_ADD_F64 0x164
+#define V_008DFC_SQ_V_MUL_F64 0x165
+#define V_008DFC_SQ_V_MIN_F64 0x166
+#define V_008DFC_SQ_V_MAX_F64 0x167
+#define V_008DFC_SQ_V_LDEXP_F64 0x168
+#define V_008DFC_SQ_V_MUL_LO_U32 0x169
+#define V_008DFC_SQ_V_MUL_HI_U32 0x16A
+#define V_008DFC_SQ_V_MUL_LO_I32 0x16B
+#define V_008DFC_SQ_V_MUL_HI_I32 0x16C
+#define V_008DFC_SQ_V_DIV_SCALE_F32 0x16D
+#define V_008DFC_SQ_V_DIV_SCALE_F64 0x16E
+#define V_008DFC_SQ_V_DIV_FMAS_F32 0x16F
+#define V_008DFC_SQ_V_DIV_FMAS_F64 0x170
+#define V_008DFC_SQ_V_MSAD_U8 0x171
+#define V_008DFC_SQ_V_QSAD_U8 0x172
+#define V_008DFC_SQ_V_MQSAD_U8 0x173
+#define V_008DFC_SQ_V_TRIG_PREOP_F64 0x174
+#define V_008DFC_SQ_V_OP1_OFFSET 0x180
+#define S_008DFC_ENCODING(x) (((x) & 0x3F) << 26)
+#define G_008DFC_ENCODING(x) (((x) >> 26) & 0x3F)
+#define C_008DFC_ENCODING 0x03FFFFFF
+#define V_008DFC_SQ_ENC_VOP3_FIELD 0x34
+#define R_008DFC_SQ_VOP2 0x008DFC
+#define S_008DFC_SRC0(x) (((x) & 0x1FF) << 0)
+#define G_008DFC_SRC0(x) (((x) >> 0) & 0x1FF)
+#define C_008DFC_SRC0 0xFFFFFE00
+#define V_008DFC_SQ_SGPR 0x00
+#define V_008DFC_SQ_VCC_LO 0x6A
+#define V_008DFC_SQ_VCC_HI 0x6B
+#define V_008DFC_SQ_TBA_LO 0x6C
+#define V_008DFC_SQ_TBA_HI 0x6D
+#define V_008DFC_SQ_TMA_LO 0x6E
+#define V_008DFC_SQ_TMA_HI 0x6F
+#define V_008DFC_SQ_TTMP0 0x70
+#define V_008DFC_SQ_TTMP1 0x71
+#define V_008DFC_SQ_TTMP2 0x72
+#define V_008DFC_SQ_TTMP3 0x73
+#define V_008DFC_SQ_TTMP4 0x74
+#define V_008DFC_SQ_TTMP5 0x75
+#define V_008DFC_SQ_TTMP6 0x76
+#define V_008DFC_SQ_TTMP7 0x77
+#define V_008DFC_SQ_TTMP8 0x78
+#define V_008DFC_SQ_TTMP9 0x79
+#define V_008DFC_SQ_TTMP10 0x7A
+#define V_008DFC_SQ_TTMP11 0x7B
+#define V_008DFC_SQ_M0 0x7C
+#define V_008DFC_SQ_EXEC_LO 0x7E
+#define V_008DFC_SQ_EXEC_HI 0x7F
+#define V_008DFC_SQ_SRC_0 0x80
+#define V_008DFC_SQ_SRC_1_INT 0x81
+#define V_008DFC_SQ_SRC_2_INT 0x82
+#define V_008DFC_SQ_SRC_3_INT 0x83
+#define V_008DFC_SQ_SRC_4_INT 0x84
+#define V_008DFC_SQ_SRC_5_INT 0x85
+#define V_008DFC_SQ_SRC_6_INT 0x86
+#define V_008DFC_SQ_SRC_7_INT 0x87
+#define V_008DFC_SQ_SRC_8_INT 0x88
+#define V_008DFC_SQ_SRC_9_INT 0x89
+#define V_008DFC_SQ_SRC_10_INT 0x8A
+#define V_008DFC_SQ_SRC_11_INT 0x8B
+#define V_008DFC_SQ_SRC_12_INT 0x8C
+#define V_008DFC_SQ_SRC_13_INT 0x8D
+#define V_008DFC_SQ_SRC_14_INT 0x8E
+#define V_008DFC_SQ_SRC_15_INT 0x8F
+#define V_008DFC_SQ_SRC_16_INT 0x90
+#define V_008DFC_SQ_SRC_17_INT 0x91
+#define V_008DFC_SQ_SRC_18_INT 0x92
+#define V_008DFC_SQ_SRC_19_INT 0x93
+#define V_008DFC_SQ_SRC_20_INT 0x94
+#define V_008DFC_SQ_SRC_21_INT 0x95
+#define V_008DFC_SQ_SRC_22_INT 0x96
+#define V_008DFC_SQ_SRC_23_INT 0x97
+#define V_008DFC_SQ_SRC_24_INT 0x98
+#define V_008DFC_SQ_SRC_25_INT 0x99
+#define V_008DFC_SQ_SRC_26_INT 0x9A
+#define V_008DFC_SQ_SRC_27_INT 0x9B
+#define V_008DFC_SQ_SRC_28_INT 0x9C
+#define V_008DFC_SQ_SRC_29_INT 0x9D
+#define V_008DFC_SQ_SRC_30_INT 0x9E
+#define V_008DFC_SQ_SRC_31_INT 0x9F
+#define V_008DFC_SQ_SRC_32_INT 0xA0
+#define V_008DFC_SQ_SRC_33_INT 0xA1
+#define V_008DFC_SQ_SRC_34_INT 0xA2
+#define V_008DFC_SQ_SRC_35_INT 0xA3
+#define V_008DFC_SQ_SRC_36_INT 0xA4
+#define V_008DFC_SQ_SRC_37_INT 0xA5
+#define V_008DFC_SQ_SRC_38_INT 0xA6
+#define V_008DFC_SQ_SRC_39_INT 0xA7
+#define V_008DFC_SQ_SRC_40_INT 0xA8
+#define V_008DFC_SQ_SRC_41_INT 0xA9
+#define V_008DFC_SQ_SRC_42_INT 0xAA
+#define V_008DFC_SQ_SRC_43_INT 0xAB
+#define V_008DFC_SQ_SRC_44_INT 0xAC
+#define V_008DFC_SQ_SRC_45_INT 0xAD
+#define V_008DFC_SQ_SRC_46_INT 0xAE
+#define V_008DFC_SQ_SRC_47_INT 0xAF
+#define V_008DFC_SQ_SRC_48_INT 0xB0
+#define V_008DFC_SQ_SRC_49_INT 0xB1
+#define V_008DFC_SQ_SRC_50_INT 0xB2
+#define V_008DFC_SQ_SRC_51_INT 0xB3
+#define V_008DFC_SQ_SRC_52_INT 0xB4
+#define V_008DFC_SQ_SRC_53_INT 0xB5
+#define V_008DFC_SQ_SRC_54_INT 0xB6
+#define V_008DFC_SQ_SRC_55_INT 0xB7
+#define V_008DFC_SQ_SRC_56_INT 0xB8
+#define V_008DFC_SQ_SRC_57_INT 0xB9
+#define V_008DFC_SQ_SRC_58_INT 0xBA
+#define V_008DFC_SQ_SRC_59_INT 0xBB
+#define V_008DFC_SQ_SRC_60_INT 0xBC
+#define V_008DFC_SQ_SRC_61_INT 0xBD
+#define V_008DFC_SQ_SRC_62_INT 0xBE
+#define V_008DFC_SQ_SRC_63_INT 0xBF
+#define V_008DFC_SQ_SRC_64_INT 0xC0
+#define V_008DFC_SQ_SRC_M_1_INT 0xC1
+#define V_008DFC_SQ_SRC_M_2_INT 0xC2
+#define V_008DFC_SQ_SRC_M_3_INT 0xC3
+#define V_008DFC_SQ_SRC_M_4_INT 0xC4
+#define V_008DFC_SQ_SRC_M_5_INT 0xC5
+#define V_008DFC_SQ_SRC_M_6_INT 0xC6
+#define V_008DFC_SQ_SRC_M_7_INT 0xC7
+#define V_008DFC_SQ_SRC_M_8_INT 0xC8
+#define V_008DFC_SQ_SRC_M_9_INT 0xC9
+#define V_008DFC_SQ_SRC_M_10_INT 0xCA
+#define V_008DFC_SQ_SRC_M_11_INT 0xCB
+#define V_008DFC_SQ_SRC_M_12_INT 0xCC
+#define V_008DFC_SQ_SRC_M_13_INT 0xCD
+#define V_008DFC_SQ_SRC_M_14_INT 0xCE
+#define V_008DFC_SQ_SRC_M_15_INT 0xCF
+#define V_008DFC_SQ_SRC_M_16_INT 0xD0
+#define V_008DFC_SQ_SRC_0_5 0xF0
+#define V_008DFC_SQ_SRC_M_0_5 0xF1
+#define V_008DFC_SQ_SRC_1 0xF2
+#define V_008DFC_SQ_SRC_M_1 0xF3
+#define V_008DFC_SQ_SRC_2 0xF4
+#define V_008DFC_SQ_SRC_M_2 0xF5
+#define V_008DFC_SQ_SRC_4 0xF6
+#define V_008DFC_SQ_SRC_M_4 0xF7
+#define V_008DFC_SQ_SRC_VCCZ 0xFB
+#define V_008DFC_SQ_SRC_EXECZ 0xFC
+#define V_008DFC_SQ_SRC_SCC 0xFD
+#define V_008DFC_SQ_SRC_LDS_DIRECT 0xFE
+#define V_008DFC_SQ_SRC_VGPR 0x100
+#define S_008DFC_VSRC1(x) (((x) & 0xFF) << 9)
+#define G_008DFC_VSRC1(x) (((x) >> 9) & 0xFF)
+#define C_008DFC_VSRC1 0xFFFE01FF
+#define V_008DFC_SQ_VGPR 0x00
+#define S_008DFC_VDST(x) (((x) & 0xFF) << 17)
+#define G_008DFC_VDST(x) (((x) >> 17) & 0xFF)
+#define C_008DFC_VDST 0xFE01FFFF
+#define V_008DFC_SQ_VGPR 0x00
+#define S_008DFC_OP(x) (((x) & 0x3F) << 25)
+#define G_008DFC_OP(x) (((x) >> 25) & 0x3F)
+#define C_008DFC_OP 0x81FFFFFF
+#define V_008DFC_SQ_V_CNDMASK_B32 0x00
+#define V_008DFC_SQ_V_READLANE_B32 0x01
+#define V_008DFC_SQ_V_WRITELANE_B32 0x02
+#define V_008DFC_SQ_V_ADD_F32 0x03
+#define V_008DFC_SQ_V_SUB_F32 0x04
+#define V_008DFC_SQ_V_SUBREV_F32 0x05
+#define V_008DFC_SQ_V_MAC_LEGACY_F32 0x06
+#define V_008DFC_SQ_V_MUL_LEGACY_F32 0x07
+#define V_008DFC_SQ_V_MUL_F32 0x08
+#define V_008DFC_SQ_V_MUL_I32_I24 0x09
+#define V_008DFC_SQ_V_MUL_HI_I32_I24 0x0A
+#define V_008DFC_SQ_V_MUL_U32_U24 0x0B
+#define V_008DFC_SQ_V_MUL_HI_U32_U24 0x0C
+#define V_008DFC_SQ_V_MIN_LEGACY_F32 0x0D
+#define V_008DFC_SQ_V_MAX_LEGACY_F32 0x0E
+#define V_008DFC_SQ_V_MIN_F32 0x0F
+#define V_008DFC_SQ_V_MAX_F32 0x10
+#define V_008DFC_SQ_V_MIN_I32 0x11
+#define V_008DFC_SQ_V_MAX_I32 0x12
+#define V_008DFC_SQ_V_MIN_U32 0x13
+#define V_008DFC_SQ_V_MAX_U32 0x14
+#define V_008DFC_SQ_V_LSHR_B32 0x15
+#define V_008DFC_SQ_V_LSHRREV_B32 0x16
+#define V_008DFC_SQ_V_ASHR_I32 0x17
+#define V_008DFC_SQ_V_ASHRREV_I32 0x18
+#define V_008DFC_SQ_V_LSHL_B32 0x19
+#define V_008DFC_SQ_V_LSHLREV_B32 0x1A
+#define V_008DFC_SQ_V_AND_B32 0x1B
+#define V_008DFC_SQ_V_OR_B32 0x1C
+#define V_008DFC_SQ_V_XOR_B32 0x1D
+#define V_008DFC_SQ_V_BFM_B32 0x1E
+#define V_008DFC_SQ_V_MAC_F32 0x1F
+#define V_008DFC_SQ_V_MADMK_F32 0x20
+#define V_008DFC_SQ_V_MADAK_F32 0x21
+#define V_008DFC_SQ_V_BCNT_U32_B32 0x22
+#define V_008DFC_SQ_V_MBCNT_LO_U32_B32 0x23
+#define V_008DFC_SQ_V_MBCNT_HI_U32_B32 0x24
+#define V_008DFC_SQ_V_ADD_I32 0x25
+#define V_008DFC_SQ_V_SUB_I32 0x26
+#define V_008DFC_SQ_V_SUBREV_I32 0x27
+#define V_008DFC_SQ_V_ADDC_U32 0x28
+#define V_008DFC_SQ_V_SUBB_U32 0x29
+#define V_008DFC_SQ_V_SUBBREV_U32 0x2A
+#define V_008DFC_SQ_V_LDEXP_F32 0x2B
+#define V_008DFC_SQ_V_CVT_PKACCUM_U8_F32 0x2C
+#define V_008DFC_SQ_V_CVT_PKNORM_I16_F32 0x2D
+#define V_008DFC_SQ_V_CVT_PKNORM_U16_F32 0x2E
+#define V_008DFC_SQ_V_CVT_PKRTZ_F16_F32 0x2F
+#define V_008DFC_SQ_V_CVT_PK_U16_U32 0x30
+#define V_008DFC_SQ_V_CVT_PK_I16_I32 0x31
+#define S_008DFC_ENCODING(x) (((x) & 0x1) << 31)
+#define G_008DFC_ENCODING(x) (((x) >> 31) & 0x1)
+#define C_008DFC_ENCODING 0x7FFFFFFF
+#define R_008DFC_SQ_VOP3_0_SDST_ENC 0x008DFC
+#define S_008DFC_VDST(x) (((x) & 0xFF) << 0)
+#define G_008DFC_VDST(x) (((x) >> 0) & 0xFF)
+#define C_008DFC_VDST 0xFFFFFF00
+#define V_008DFC_SQ_VGPR 0x00
+#define S_008DFC_SDST(x) (((x) & 0x7F) << 8)
+#define G_008DFC_SDST(x) (((x) >> 8) & 0x7F)
+#define C_008DFC_SDST 0xFFFF80FF
+#define V_008DFC_SQ_SGPR 0x00
+#define V_008DFC_SQ_VCC_LO 0x6A
+#define V_008DFC_SQ_VCC_HI 0x6B
+#define V_008DFC_SQ_TBA_LO 0x6C
+#define V_008DFC_SQ_TBA_HI 0x6D
+#define V_008DFC_SQ_TMA_LO 0x6E
+#define V_008DFC_SQ_TMA_HI 0x6F
+#define V_008DFC_SQ_TTMP0 0x70
+#define V_008DFC_SQ_TTMP1 0x71
+#define V_008DFC_SQ_TTMP2 0x72
+#define V_008DFC_SQ_TTMP3 0x73
+#define V_008DFC_SQ_TTMP4 0x74
+#define V_008DFC_SQ_TTMP5 0x75
+#define V_008DFC_SQ_TTMP6 0x76
+#define V_008DFC_SQ_TTMP7 0x77
+#define V_008DFC_SQ_TTMP8 0x78
+#define V_008DFC_SQ_TTMP9 0x79
+#define V_008DFC_SQ_TTMP10 0x7A
+#define V_008DFC_SQ_TTMP11 0x7B
+#define S_008DFC_OP(x) (((x) & 0x1FF) << 17)
+#define G_008DFC_OP(x) (((x) >> 17) & 0x1FF)
+#define C_008DFC_OP 0xFC01FFFF
+#define V_008DFC_SQ_V_OPC_OFFSET 0x00
+#define V_008DFC_SQ_V_OP2_OFFSET 0x100
+#define V_008DFC_SQ_V_MAD_LEGACY_F32 0x140
+#define V_008DFC_SQ_V_MAD_F32 0x141
+#define V_008DFC_SQ_V_MAD_I32_I24 0x142
+#define V_008DFC_SQ_V_MAD_U32_U24 0x143
+#define V_008DFC_SQ_V_CUBEID_F32 0x144
+#define V_008DFC_SQ_V_CUBESC_F32 0x145
+#define V_008DFC_SQ_V_CUBETC_F32 0x146
+#define V_008DFC_SQ_V_CUBEMA_F32 0x147
+#define V_008DFC_SQ_V_BFE_U32 0x148
+#define V_008DFC_SQ_V_BFE_I32 0x149
+#define V_008DFC_SQ_V_BFI_B32 0x14A
+#define V_008DFC_SQ_V_FMA_F32 0x14B
+#define V_008DFC_SQ_V_FMA_F64 0x14C
+#define V_008DFC_SQ_V_LERP_U8 0x14D
+#define V_008DFC_SQ_V_ALIGNBIT_B32 0x14E
+#define V_008DFC_SQ_V_ALIGNBYTE_B32 0x14F
+#define V_008DFC_SQ_V_MULLIT_F32 0x150
+#define V_008DFC_SQ_V_MIN3_F32 0x151
+#define V_008DFC_SQ_V_MIN3_I32 0x152
+#define V_008DFC_SQ_V_MIN3_U32 0x153
+#define V_008DFC_SQ_V_MAX3_F32 0x154
+#define V_008DFC_SQ_V_MAX3_I32 0x155
+#define V_008DFC_SQ_V_MAX3_U32 0x156
+#define V_008DFC_SQ_V_MED3_F32 0x157
+#define V_008DFC_SQ_V_MED3_I32 0x158
+#define V_008DFC_SQ_V_MED3_U32 0x159
+#define V_008DFC_SQ_V_SAD_U8 0x15A
+#define V_008DFC_SQ_V_SAD_HI_U8 0x15B
+#define V_008DFC_SQ_V_SAD_U16 0x15C
+#define V_008DFC_SQ_V_SAD_U32 0x15D
+#define V_008DFC_SQ_V_CVT_PK_U8_F32 0x15E
+#define V_008DFC_SQ_V_DIV_FIXUP_F32 0x15F
+#define V_008DFC_SQ_V_DIV_FIXUP_F64 0x160
+#define V_008DFC_SQ_V_LSHL_B64 0x161
+#define V_008DFC_SQ_V_LSHR_B64 0x162
+#define V_008DFC_SQ_V_ASHR_I64 0x163
+#define V_008DFC_SQ_V_ADD_F64 0x164
+#define V_008DFC_SQ_V_MUL_F64 0x165
+#define V_008DFC_SQ_V_MIN_F64 0x166
+#define V_008DFC_SQ_V_MAX_F64 0x167
+#define V_008DFC_SQ_V_LDEXP_F64 0x168
+#define V_008DFC_SQ_V_MUL_LO_U32 0x169
+#define V_008DFC_SQ_V_MUL_HI_U32 0x16A
+#define V_008DFC_SQ_V_MUL_LO_I32 0x16B
+#define V_008DFC_SQ_V_MUL_HI_I32 0x16C
+#define V_008DFC_SQ_V_DIV_SCALE_F32 0x16D
+#define V_008DFC_SQ_V_DIV_SCALE_F64 0x16E
+#define V_008DFC_SQ_V_DIV_FMAS_F32 0x16F
+#define V_008DFC_SQ_V_DIV_FMAS_F64 0x170
+#define V_008DFC_SQ_V_MSAD_U8 0x171
+#define V_008DFC_SQ_V_QSAD_U8 0x172
+#define V_008DFC_SQ_V_MQSAD_U8 0x173
+#define V_008DFC_SQ_V_TRIG_PREOP_F64 0x174
+#define V_008DFC_SQ_V_OP1_OFFSET 0x180
+#define S_008DFC_ENCODING(x) (((x) & 0x3F) << 26)
+#define G_008DFC_ENCODING(x) (((x) >> 26) & 0x3F)
+#define C_008DFC_ENCODING 0x03FFFFFF
+#define V_008DFC_SQ_ENC_VOP3_FIELD 0x34
+#define R_008DFC_SQ_MUBUF_0 0x008DFC
+#define S_008DFC_OFFSET(x) (((x) & 0xFFF) << 0)
+#define G_008DFC_OFFSET(x) (((x) >> 0) & 0xFFF)
+#define C_008DFC_OFFSET 0xFFFFF000
+#define S_008DFC_OFFEN(x) (((x) & 0x1) << 12)
+#define G_008DFC_OFFEN(x) (((x) >> 12) & 0x1)
+#define C_008DFC_OFFEN 0xFFFFEFFF
+#define S_008DFC_IDXEN(x) (((x) & 0x1) << 13)
+#define G_008DFC_IDXEN(x) (((x) >> 13) & 0x1)
+#define C_008DFC_IDXEN 0xFFFFDFFF
+#define S_008DFC_GLC(x) (((x) & 0x1) << 14)
+#define G_008DFC_GLC(x) (((x) >> 14) & 0x1)
+#define C_008DFC_GLC 0xFFFFBFFF
+#define S_008DFC_ADDR64(x) (((x) & 0x1) << 15)
+#define G_008DFC_ADDR64(x) (((x) >> 15) & 0x1)
+#define C_008DFC_ADDR64 0xFFFF7FFF
+#define S_008DFC_LDS(x) (((x) & 0x1) << 16)
+#define G_008DFC_LDS(x) (((x) >> 16) & 0x1)
+#define C_008DFC_LDS 0xFFFEFFFF
+#define S_008DFC_OP(x) (((x) & 0x7F) << 18)
+#define G_008DFC_OP(x) (((x) >> 18) & 0x7F)
+#define C_008DFC_OP 0xFE03FFFF
+#define V_008DFC_SQ_BUFFER_LOAD_FORMAT_X 0x00
+#define V_008DFC_SQ_BUFFER_LOAD_FORMAT_XY 0x01
+#define V_008DFC_SQ_BUFFER_LOAD_FORMAT_XYZ 0x02
+#define V_008DFC_SQ_BUFFER_LOAD_FORMAT_XYZW 0x03
+#define V_008DFC_SQ_BUFFER_STORE_FORMAT_X 0x04
+#define V_008DFC_SQ_BUFFER_STORE_FORMAT_XY 0x05
+#define V_008DFC_SQ_BUFFER_STORE_FORMAT_XYZ 0x06
+#define V_008DFC_SQ_BUFFER_STORE_FORMAT_XYZW 0x07
+#define V_008DFC_SQ_BUFFER_LOAD_UBYTE 0x08
+#define V_008DFC_SQ_BUFFER_LOAD_SBYTE 0x09
+#define V_008DFC_SQ_BUFFER_LOAD_USHORT 0x0A
+#define V_008DFC_SQ_BUFFER_LOAD_SSHORT 0x0B
+#define V_008DFC_SQ_BUFFER_LOAD_DWORD 0x0C
+#define V_008DFC_SQ_BUFFER_LOAD_DWORDX2 0x0D
+#define V_008DFC_SQ_BUFFER_LOAD_DWORDX4 0x0E
+#define V_008DFC_SQ_BUFFER_STORE_BYTE 0x18
+#define V_008DFC_SQ_BUFFER_STORE_SHORT 0x1A
+#define V_008DFC_SQ_BUFFER_STORE_DWORD 0x1C
+#define V_008DFC_SQ_BUFFER_STORE_DWORDX2 0x1D
+#define V_008DFC_SQ_BUFFER_STORE_DWORDX4 0x1E
+#define V_008DFC_SQ_BUFFER_ATOMIC_SWAP 0x30
+#define V_008DFC_SQ_BUFFER_ATOMIC_CMPSWAP 0x31
+#define V_008DFC_SQ_BUFFER_ATOMIC_ADD 0x32
+#define V_008DFC_SQ_BUFFER_ATOMIC_SUB 0x33
+#define V_008DFC_SQ_BUFFER_ATOMIC_RSUB 0x34
+#define V_008DFC_SQ_BUFFER_ATOMIC_SMIN 0x35
+#define V_008DFC_SQ_BUFFER_ATOMIC_UMIN 0x36
+#define V_008DFC_SQ_BUFFER_ATOMIC_SMAX 0x37
+#define V_008DFC_SQ_BUFFER_ATOMIC_UMAX 0x38
+#define V_008DFC_SQ_BUFFER_ATOMIC_AND 0x39
+#define V_008DFC_SQ_BUFFER_ATOMIC_OR 0x3A
+#define V_008DFC_SQ_BUFFER_ATOMIC_XOR 0x3B
+#define V_008DFC_SQ_BUFFER_ATOMIC_INC 0x3C
+#define V_008DFC_SQ_BUFFER_ATOMIC_DEC 0x3D
+#define V_008DFC_SQ_BUFFER_ATOMIC_FCMPSWAP 0x3E
+#define V_008DFC_SQ_BUFFER_ATOMIC_FMIN 0x3F
+#define V_008DFC_SQ_BUFFER_ATOMIC_FMAX 0x40
+#define V_008DFC_SQ_BUFFER_ATOMIC_SWAP_X2 0x50
+#define V_008DFC_SQ_BUFFER_ATOMIC_CMPSWAP_X2 0x51
+#define V_008DFC_SQ_BUFFER_ATOMIC_ADD_X2 0x52
+#define V_008DFC_SQ_BUFFER_ATOMIC_SUB_X2 0x53
+#define V_008DFC_SQ_BUFFER_ATOMIC_RSUB_X2 0x54
+#define V_008DFC_SQ_BUFFER_ATOMIC_SMIN_X2 0x55
+#define V_008DFC_SQ_BUFFER_ATOMIC_UMIN_X2 0x56
+#define V_008DFC_SQ_BUFFER_ATOMIC_SMAX_X2 0x57
+#define V_008DFC_SQ_BUFFER_ATOMIC_UMAX_X2 0x58
+#define V_008DFC_SQ_BUFFER_ATOMIC_AND_X2 0x59
+#define V_008DFC_SQ_BUFFER_ATOMIC_OR_X2 0x5A
+#define V_008DFC_SQ_BUFFER_ATOMIC_XOR_X2 0x5B
+#define V_008DFC_SQ_BUFFER_ATOMIC_INC_X2 0x5C
+#define V_008DFC_SQ_BUFFER_ATOMIC_DEC_X2 0x5D
+#define V_008DFC_SQ_BUFFER_ATOMIC_FCMPSWAP_X2 0x5E
+#define V_008DFC_SQ_BUFFER_ATOMIC_FMIN_X2 0x5F
+#define V_008DFC_SQ_BUFFER_ATOMIC_FMAX_X2 0x60
+#define V_008DFC_SQ_BUFFER_WBINVL1_SC 0x70
+#define V_008DFC_SQ_BUFFER_WBINVL1 0x71
+#define S_008DFC_ENCODING(x) (((x) & 0x3F) << 26)
+#define G_008DFC_ENCODING(x) (((x) >> 26) & 0x3F)
+#define C_008DFC_ENCODING 0x03FFFFFF
+#define V_008DFC_SQ_ENC_MUBUF_FIELD 0x38
+#endif
+#define R_008F00_SQ_BUF_RSRC_WORD0 0x008F00
+#define R_008F04_SQ_BUF_RSRC_WORD1 0x008F04
+#define S_008F04_BASE_ADDRESS_HI(x) (((x) & 0xFFFF) << 0)
+#define G_008F04_BASE_ADDRESS_HI(x) (((x) >> 0) & 0xFFFF)
+#define C_008F04_BASE_ADDRESS_HI 0xFFFF0000
+#define S_008F04_STRIDE(x) (((x) & 0x3FFF) << 16)
+#define G_008F04_STRIDE(x) (((x) >> 16) & 0x3FFF)
+#define C_008F04_STRIDE 0xC000FFFF
+#define S_008F04_CACHE_SWIZZLE(x) (((x) & 0x1) << 30)
+#define G_008F04_CACHE_SWIZZLE(x) (((x) >> 30) & 0x1)
+#define C_008F04_CACHE_SWIZZLE 0xBFFFFFFF
+#define S_008F04_SWIZZLE_ENABLE(x) (((x) & 0x1) << 31)
+#define G_008F04_SWIZZLE_ENABLE(x) (((x) >> 31) & 0x1)
+#define C_008F04_SWIZZLE_ENABLE 0x7FFFFFFF
+#define R_008F08_SQ_BUF_RSRC_WORD2 0x008F08
+#define R_008F0C_SQ_BUF_RSRC_WORD3 0x008F0C
+#define S_008F0C_DST_SEL_X(x) (((x) & 0x07) << 0)
+#define G_008F0C_DST_SEL_X(x) (((x) >> 0) & 0x07)
+#define C_008F0C_DST_SEL_X 0xFFFFFFF8
+#define V_008F0C_SQ_SEL_0 0x00
+#define V_008F0C_SQ_SEL_1 0x01
+#define V_008F0C_SQ_SEL_RESERVED_0 0x02
+#define V_008F0C_SQ_SEL_RESERVED_1 0x03
+#define V_008F0C_SQ_SEL_X 0x04
+#define V_008F0C_SQ_SEL_Y 0x05
+#define V_008F0C_SQ_SEL_Z 0x06
+#define V_008F0C_SQ_SEL_W 0x07
+#define S_008F0C_DST_SEL_Y(x) (((x) & 0x07) << 3)
+#define G_008F0C_DST_SEL_Y(x) (((x) >> 3) & 0x07)
+#define C_008F0C_DST_SEL_Y 0xFFFFFFC7
+#define V_008F0C_SQ_SEL_0 0x00
+#define V_008F0C_SQ_SEL_1 0x01
+#define V_008F0C_SQ_SEL_RESERVED_0 0x02
+#define V_008F0C_SQ_SEL_RESERVED_1 0x03
+#define V_008F0C_SQ_SEL_X 0x04
+#define V_008F0C_SQ_SEL_Y 0x05
+#define V_008F0C_SQ_SEL_Z 0x06
+#define V_008F0C_SQ_SEL_W 0x07
+#define S_008F0C_DST_SEL_Z(x) (((x) & 0x07) << 6)
+#define G_008F0C_DST_SEL_Z(x) (((x) >> 6) & 0x07)
+#define C_008F0C_DST_SEL_Z 0xFFFFFE3F
+#define V_008F0C_SQ_SEL_0 0x00
+#define V_008F0C_SQ_SEL_1 0x01
+#define V_008F0C_SQ_SEL_RESERVED_0 0x02
+#define V_008F0C_SQ_SEL_RESERVED_1 0x03
+#define V_008F0C_SQ_SEL_X 0x04
+#define V_008F0C_SQ_SEL_Y 0x05
+#define V_008F0C_SQ_SEL_Z 0x06
+#define V_008F0C_SQ_SEL_W 0x07
+#define S_008F0C_DST_SEL_W(x) (((x) & 0x07) << 9)
+#define G_008F0C_DST_SEL_W(x) (((x) >> 9) & 0x07)
+#define C_008F0C_DST_SEL_W 0xFFFFF1FF
+#define V_008F0C_SQ_SEL_0 0x00
+#define V_008F0C_SQ_SEL_1 0x01
+#define V_008F0C_SQ_SEL_RESERVED_0 0x02
+#define V_008F0C_SQ_SEL_RESERVED_1 0x03
+#define V_008F0C_SQ_SEL_X 0x04
+#define V_008F0C_SQ_SEL_Y 0x05
+#define V_008F0C_SQ_SEL_Z 0x06
+#define V_008F0C_SQ_SEL_W 0x07
+#define S_008F0C_NUM_FORMAT(x) (((x) & 0x07) << 12)
+#define G_008F0C_NUM_FORMAT(x) (((x) >> 12) & 0x07)
+#define C_008F0C_NUM_FORMAT 0xFFFF8FFF
+#define V_008F0C_BUF_NUM_FORMAT_UNORM 0x00
+#define V_008F0C_BUF_NUM_FORMAT_SNORM 0x01
+#define V_008F0C_BUF_NUM_FORMAT_USCALED 0x02
+#define V_008F0C_BUF_NUM_FORMAT_SSCALED 0x03
+#define V_008F0C_BUF_NUM_FORMAT_UINT 0x04
+#define V_008F0C_BUF_NUM_FORMAT_SINT 0x05
+#define V_008F0C_BUF_NUM_FORMAT_SNORM_OGL 0x06
+#define V_008F0C_BUF_NUM_FORMAT_FLOAT 0x07
+#define S_008F0C_DATA_FORMAT(x) (((x) & 0x0F) << 15)
+#define G_008F0C_DATA_FORMAT(x) (((x) >> 15) & 0x0F)
+#define C_008F0C_DATA_FORMAT 0xFFF87FFF
+#define V_008F0C_BUF_DATA_FORMAT_INVALID 0x00
+#define V_008F0C_BUF_DATA_FORMAT_8 0x01
+#define V_008F0C_BUF_DATA_FORMAT_16 0x02
+#define V_008F0C_BUF_DATA_FORMAT_8_8 0x03
+#define V_008F0C_BUF_DATA_FORMAT_32 0x04
+#define V_008F0C_BUF_DATA_FORMAT_16_16 0x05
+#define V_008F0C_BUF_DATA_FORMAT_10_11_11 0x06
+#define V_008F0C_BUF_DATA_FORMAT_11_11_10 0x07
+#define V_008F0C_BUF_DATA_FORMAT_10_10_10_2 0x08
+#define V_008F0C_BUF_DATA_FORMAT_2_10_10_10 0x09
+#define V_008F0C_BUF_DATA_FORMAT_8_8_8_8 0x0A
+#define V_008F0C_BUF_DATA_FORMAT_32_32 0x0B
+#define V_008F0C_BUF_DATA_FORMAT_16_16_16_16 0x0C
+#define V_008F0C_BUF_DATA_FORMAT_32_32_32 0x0D
+#define V_008F0C_BUF_DATA_FORMAT_32_32_32_32 0x0E
+#define V_008F0C_BUF_DATA_FORMAT_RESERVED_15 0x0F
+#define S_008F0C_ELEMENT_SIZE(x) (((x) & 0x03) << 19)
+#define G_008F0C_ELEMENT_SIZE(x) (((x) >> 19) & 0x03)
+#define C_008F0C_ELEMENT_SIZE 0xFFE7FFFF
+#define S_008F0C_INDEX_STRIDE(x) (((x) & 0x03) << 21)
+#define G_008F0C_INDEX_STRIDE(x) (((x) >> 21) & 0x03)
+#define C_008F0C_INDEX_STRIDE 0xFF9FFFFF
+#define S_008F0C_ADD_TID_ENABLE(x) (((x) & 0x1) << 23)
+#define G_008F0C_ADD_TID_ENABLE(x) (((x) >> 23) & 0x1)
+#define C_008F0C_ADD_TID_ENABLE 0xFF7FFFFF
+#define S_008F0C_HASH_ENABLE(x) (((x) & 0x1) << 25)
+#define G_008F0C_HASH_ENABLE(x) (((x) >> 25) & 0x1)
+#define C_008F0C_HASH_ENABLE 0xFDFFFFFF
+#define S_008F0C_HEAP(x) (((x) & 0x1) << 26)
+#define G_008F0C_HEAP(x) (((x) >> 26) & 0x1)
+#define C_008F0C_HEAP 0xFBFFFFFF
+#define S_008F0C_TYPE(x) (((x) & 0x03) << 30)
+#define G_008F0C_TYPE(x) (((x) >> 30) & 0x03)
+#define C_008F0C_TYPE 0x3FFFFFFF
+#define V_008F0C_SQ_RSRC_BUF 0x00
+#define V_008F0C_SQ_RSRC_BUF_RSVD_1 0x01
+#define V_008F0C_SQ_RSRC_BUF_RSVD_2 0x02
+#define V_008F0C_SQ_RSRC_BUF_RSVD_3 0x03
+#define R_008F10_SQ_IMG_RSRC_WORD0 0x008F10
+#define R_008F14_SQ_IMG_RSRC_WORD1 0x008F14
+#define S_008F14_BASE_ADDRESS_HI(x) (((x) & 0xFF) << 0)
+#define G_008F14_BASE_ADDRESS_HI(x) (((x) >> 0) & 0xFF)
+#define C_008F14_BASE_ADDRESS_HI 0xFFFFFF00
+#define S_008F14_MIN_LOD(x) (((x) & 0xFFF) << 8)
+#define G_008F14_MIN_LOD(x) (((x) >> 8) & 0xFFF)
+#define C_008F14_MIN_LOD 0xFFF000FF
+#define S_008F14_DATA_FORMAT(x) (((x) & 0x3F) << 20)
+#define G_008F14_DATA_FORMAT(x) (((x) >> 20) & 0x3F)
+#define C_008F14_DATA_FORMAT 0xFC0FFFFF
+#define V_008F14_IMG_DATA_FORMAT_INVALID 0x00
+#define V_008F14_IMG_DATA_FORMAT_8 0x01
+#define V_008F14_IMG_DATA_FORMAT_16 0x02
+#define V_008F14_IMG_DATA_FORMAT_8_8 0x03
+#define V_008F14_IMG_DATA_FORMAT_32 0x04
+#define V_008F14_IMG_DATA_FORMAT_16_16 0x05
+#define V_008F14_IMG_DATA_FORMAT_10_11_11 0x06
+#define V_008F14_IMG_DATA_FORMAT_11_11_10 0x07
+#define V_008F14_IMG_DATA_FORMAT_10_10_10_2 0x08
+#define V_008F14_IMG_DATA_FORMAT_2_10_10_10 0x09
+#define V_008F14_IMG_DATA_FORMAT_8_8_8_8 0x0A
+#define V_008F14_IMG_DATA_FORMAT_32_32 0x0B
+#define V_008F14_IMG_DATA_FORMAT_16_16_16_16 0x0C
+#define V_008F14_IMG_DATA_FORMAT_32_32_32 0x0D
+#define V_008F14_IMG_DATA_FORMAT_32_32_32_32 0x0E
+#define V_008F14_IMG_DATA_FORMAT_RESERVED_15 0x0F
+#define V_008F14_IMG_DATA_FORMAT_5_6_5 0x10
+#define V_008F14_IMG_DATA_FORMAT_1_5_5_5 0x11
+#define V_008F14_IMG_DATA_FORMAT_5_5_5_1 0x12
+#define V_008F14_IMG_DATA_FORMAT_4_4_4_4 0x13
+#define V_008F14_IMG_DATA_FORMAT_8_24 0x14
+#define V_008F14_IMG_DATA_FORMAT_24_8 0x15
+#define V_008F14_IMG_DATA_FORMAT_X24_8_32 0x16
+#define V_008F14_IMG_DATA_FORMAT_RESERVED_23 0x17
+#define V_008F14_IMG_DATA_FORMAT_RESERVED_24 0x18
+#define V_008F14_IMG_DATA_FORMAT_RESERVED_25 0x19
+#define V_008F14_IMG_DATA_FORMAT_RESERVED_26 0x1A
+#define V_008F14_IMG_DATA_FORMAT_RESERVED_27 0x1B
+#define V_008F14_IMG_DATA_FORMAT_RESERVED_28 0x1C
+#define V_008F14_IMG_DATA_FORMAT_RESERVED_29 0x1D
+#define V_008F14_IMG_DATA_FORMAT_RESERVED_30 0x1E
+#define V_008F14_IMG_DATA_FORMAT_RESERVED_31 0x1F
+#define V_008F14_IMG_DATA_FORMAT_GB_GR 0x20
+#define V_008F14_IMG_DATA_FORMAT_BG_RG 0x21
+#define V_008F14_IMG_DATA_FORMAT_5_9_9_9 0x22
+#define V_008F14_IMG_DATA_FORMAT_RESERVED_42 0x2A
+#define V_008F14_IMG_DATA_FORMAT_RESERVED_43 0x2B
+#define V_008F14_IMG_DATA_FORMAT_FMASK8_S2_F1 0x2C
+#define V_008F14_IMG_DATA_FORMAT_FMASK8_S4_F1 0x2D
+#define V_008F14_IMG_DATA_FORMAT_FMASK8_S8_F1 0x2E
+#define V_008F14_IMG_DATA_FORMAT_FMASK8_S2_F2 0x2F
+#define V_008F14_IMG_DATA_FORMAT_FMASK8_S4_F2 0x30
+#define V_008F14_IMG_DATA_FORMAT_FMASK8_S4_F4 0x31
+#define V_008F14_IMG_DATA_FORMAT_FMASK16_S16_F1 0x32
+#define V_008F14_IMG_DATA_FORMAT_FMASK16_S8_F2 0x33
+#define V_008F14_IMG_DATA_FORMAT_FMASK32_S16_F2 0x34
+#define V_008F14_IMG_DATA_FORMAT_FMASK32_S8_F4 0x35
+#define V_008F14_IMG_DATA_FORMAT_FMASK32_S8_F8 0x36
+#define V_008F14_IMG_DATA_FORMAT_FMASK64_S16_F4 0x37
+#define V_008F14_IMG_DATA_FORMAT_FMASK64_S16_F8 0x38
+#define V_008F14_IMG_DATA_FORMAT_4_4 0x39
+#define V_008F14_IMG_DATA_FORMAT_6_5_5 0x3A
+#define V_008F14_IMG_DATA_FORMAT_1 0x3B
+#define V_008F14_IMG_DATA_FORMAT_1_REVERSED 0x3C
+#define V_008F14_IMG_DATA_FORMAT_32_AS_8 0x3D
+#define V_008F14_IMG_DATA_FORMAT_32_AS_8_8 0x3E
+#define V_008F14_IMG_DATA_FORMAT_32_AS_32_32_32_32 0x3F
+#define S_008F14_NUM_FORMAT(x) (((x) & 0x0F) << 26)
+#define G_008F14_NUM_FORMAT(x) (((x) >> 26) & 0x0F)
+#define C_008F14_NUM_FORMAT 0xC3FFFFFF
+#define V_008F14_IMG_NUM_FORMAT_UNORM 0x00
+#define V_008F14_IMG_NUM_FORMAT_SNORM 0x01
+#define V_008F14_IMG_NUM_FORMAT_USCALED 0x02
+#define V_008F14_IMG_NUM_FORMAT_SSCALED 0x03
+#define V_008F14_IMG_NUM_FORMAT_UINT 0x04
+#define V_008F14_IMG_NUM_FORMAT_SINT 0x05
+#define V_008F14_IMG_NUM_FORMAT_SNORM_OGL 0x06
+#define V_008F14_IMG_NUM_FORMAT_FLOAT 0x07
+#define V_008F14_IMG_NUM_FORMAT_RESERVED_8 0x08
+#define V_008F14_IMG_NUM_FORMAT_SRGB 0x09
+#define V_008F14_IMG_NUM_FORMAT_UBNORM 0x0A
+#define V_008F14_IMG_NUM_FORMAT_UBNORM_OGL 0x0B
+#define V_008F14_IMG_NUM_FORMAT_UBINT 0x0C
+#define V_008F14_IMG_NUM_FORMAT_UBSCALED 0x0D
+#define V_008F14_IMG_NUM_FORMAT_RESERVED_14 0x0E
+#define V_008F14_IMG_NUM_FORMAT_RESERVED_15 0x0F
+#define R_008F18_SQ_IMG_RSRC_WORD2 0x008F18
+#define S_008F18_WIDTH(x) (((x) & 0x3FFF) << 0)
+#define G_008F18_WIDTH(x) (((x) >> 0) & 0x3FFF)
+#define C_008F18_WIDTH 0xFFFFC000
+#define S_008F18_HEIGHT(x) (((x) & 0x3FFF) << 14)
+#define G_008F18_HEIGHT(x) (((x) >> 14) & 0x3FFF)
+#define C_008F18_HEIGHT 0xF0003FFF
+#define S_008F18_PERF_MOD(x) (((x) & 0x07) << 28)
+#define G_008F18_PERF_MOD(x) (((x) >> 28) & 0x07)
+#define C_008F18_PERF_MOD 0x8FFFFFFF
+#define S_008F18_INTERLACED(x) (((x) & 0x1) << 31)
+#define G_008F18_INTERLACED(x) (((x) >> 31) & 0x1)
+#define C_008F18_INTERLACED 0x7FFFFFFF
+#define R_008F1C_SQ_IMG_RSRC_WORD3 0x008F1C
+#define S_008F1C_DST_SEL_X(x) (((x) & 0x07) << 0)
+#define G_008F1C_DST_SEL_X(x) (((x) >> 0) & 0x07)
+#define C_008F1C_DST_SEL_X 0xFFFFFFF8
+#define V_008F1C_SQ_SEL_0 0x00
+#define V_008F1C_SQ_SEL_1 0x01
+#define V_008F1C_SQ_SEL_RESERVED_0 0x02
+#define V_008F1C_SQ_SEL_RESERVED_1 0x03
+#define V_008F1C_SQ_SEL_X 0x04
+#define V_008F1C_SQ_SEL_Y 0x05
+#define V_008F1C_SQ_SEL_Z 0x06
+#define V_008F1C_SQ_SEL_W 0x07
+#define S_008F1C_DST_SEL_Y(x) (((x) & 0x07) << 3)
+#define G_008F1C_DST_SEL_Y(x) (((x) >> 3) & 0x07)
+#define C_008F1C_DST_SEL_Y 0xFFFFFFC7
+#define V_008F1C_SQ_SEL_0 0x00
+#define V_008F1C_SQ_SEL_1 0x01
+#define V_008F1C_SQ_SEL_RESERVED_0 0x02
+#define V_008F1C_SQ_SEL_RESERVED_1 0x03
+#define V_008F1C_SQ_SEL_X 0x04
+#define V_008F1C_SQ_SEL_Y 0x05
+#define V_008F1C_SQ_SEL_Z 0x06
+#define V_008F1C_SQ_SEL_W 0x07
+#define S_008F1C_DST_SEL_Z(x) (((x) & 0x07) << 6)
+#define G_008F1C_DST_SEL_Z(x) (((x) >> 6) & 0x07)
+#define C_008F1C_DST_SEL_Z 0xFFFFFE3F
+#define V_008F1C_SQ_SEL_0 0x00
+#define V_008F1C_SQ_SEL_1 0x01
+#define V_008F1C_SQ_SEL_RESERVED_0 0x02
+#define V_008F1C_SQ_SEL_RESERVED_1 0x03
+#define V_008F1C_SQ_SEL_X 0x04
+#define V_008F1C_SQ_SEL_Y 0x05
+#define V_008F1C_SQ_SEL_Z 0x06
+#define V_008F1C_SQ_SEL_W 0x07
+#define S_008F1C_DST_SEL_W(x) (((x) & 0x07) << 9)
+#define G_008F1C_DST_SEL_W(x) (((x) >> 9) & 0x07)
+#define C_008F1C_DST_SEL_W 0xFFFFF1FF
+#define V_008F1C_SQ_SEL_0 0x00
+#define V_008F1C_SQ_SEL_1 0x01
+#define V_008F1C_SQ_SEL_RESERVED_0 0x02
+#define V_008F1C_SQ_SEL_RESERVED_1 0x03
+#define V_008F1C_SQ_SEL_X 0x04
+#define V_008F1C_SQ_SEL_Y 0x05
+#define V_008F1C_SQ_SEL_Z 0x06
+#define V_008F1C_SQ_SEL_W 0x07
+#define S_008F1C_BASE_LEVEL(x) (((x) & 0x0F) << 12)
+#define G_008F1C_BASE_LEVEL(x) (((x) >> 12) & 0x0F)
+#define C_008F1C_BASE_LEVEL 0xFFFF0FFF
+#define S_008F1C_LAST_LEVEL(x) (((x) & 0x0F) << 16)
+#define G_008F1C_LAST_LEVEL(x) (((x) >> 16) & 0x0F)
+#define C_008F1C_LAST_LEVEL 0xFFF0FFFF
+#define S_008F1C_TILING_INDEX(x) (((x) & 0x1F) << 20)
+#define G_008F1C_TILING_INDEX(x) (((x) >> 20) & 0x1F)
+#define C_008F1C_TILING_INDEX 0xFE0FFFFF
+#define S_008F1C_POW2_PAD(x) (((x) & 0x1) << 25)
+#define G_008F1C_POW2_PAD(x) (((x) >> 25) & 0x1)
+#define C_008F1C_POW2_PAD 0xFDFFFFFF
+#define S_008F1C_TYPE(x) (((x) & 0x0F) << 28)
+#define G_008F1C_TYPE(x) (((x) >> 28) & 0x0F)
+#define C_008F1C_TYPE 0x0FFFFFFF
+#define V_008F1C_SQ_RSRC_IMG_RSVD_0 0x00
+#define V_008F1C_SQ_RSRC_IMG_RSVD_1 0x01
+#define V_008F1C_SQ_RSRC_IMG_RSVD_2 0x02
+#define V_008F1C_SQ_RSRC_IMG_RSVD_3 0x03
+#define V_008F1C_SQ_RSRC_IMG_RSVD_4 0x04
+#define V_008F1C_SQ_RSRC_IMG_RSVD_5 0x05
+#define V_008F1C_SQ_RSRC_IMG_RSVD_6 0x06
+#define V_008F1C_SQ_RSRC_IMG_RSVD_7 0x07
+#define V_008F1C_SQ_RSRC_IMG_1D 0x08
+#define V_008F1C_SQ_RSRC_IMG_2D 0x09
+#define V_008F1C_SQ_RSRC_IMG_3D 0x0A
+#define V_008F1C_SQ_RSRC_IMG_CUBE 0x0B
+#define V_008F1C_SQ_RSRC_IMG_1D_ARRAY 0x0C
+#define V_008F1C_SQ_RSRC_IMG_2D_ARRAY 0x0D
+#define V_008F1C_SQ_RSRC_IMG_2D_MSAA 0x0E
+#define V_008F1C_SQ_RSRC_IMG_2D_MSAA_ARRAY 0x0F
+#define R_008F20_SQ_IMG_RSRC_WORD4 0x008F20
+#define S_008F20_DEPTH(x) (((x) & 0x1FFF) << 0)
+#define G_008F20_DEPTH(x) (((x) >> 0) & 0x1FFF)
+#define C_008F20_DEPTH 0xFFFFE000
+#define S_008F20_PITCH(x) (((x) & 0x3FFF) << 13)
+#define G_008F20_PITCH(x) (((x) >> 13) & 0x3FFF)
+#define C_008F20_PITCH 0xF8001FFF
+#define R_008F24_SQ_IMG_RSRC_WORD5 0x008F24
+#define S_008F24_BASE_ARRAY(x) (((x) & 0x1FFF) << 0)
+#define G_008F24_BASE_ARRAY(x) (((x) >> 0) & 0x1FFF)
+#define C_008F24_BASE_ARRAY 0xFFFFE000
+#define S_008F24_LAST_ARRAY(x) (((x) & 0x1FFF) << 13)
+#define G_008F24_LAST_ARRAY(x) (((x) >> 13) & 0x1FFF)
+#define C_008F24_LAST_ARRAY 0xFC001FFF
+#define R_008F28_SQ_IMG_RSRC_WORD6 0x008F28
+#define S_008F28_MIN_LOD_WARN(x) (((x) & 0xFFF) << 0)
+#define G_008F28_MIN_LOD_WARN(x) (((x) >> 0) & 0xFFF)
+#define C_008F28_MIN_LOD_WARN 0xFFFFF000
+#define R_008F2C_SQ_IMG_RSRC_WORD7 0x008F2C
+#define R_008F30_SQ_IMG_SAMP_WORD0 0x008F30
+#define S_008F30_CLAMP_X(x) (((x) & 0x07) << 0)
+#define G_008F30_CLAMP_X(x) (((x) >> 0) & 0x07)
+#define C_008F30_CLAMP_X 0xFFFFFFF8
+#define V_008F30_SQ_TEX_WRAP 0x00
+#define V_008F30_SQ_TEX_MIRROR 0x01
+#define V_008F30_SQ_TEX_CLAMP_LAST_TEXEL 0x02
+#define V_008F30_SQ_TEX_MIRROR_ONCE_LAST_TEXEL 0x03
+#define V_008F30_SQ_TEX_CLAMP_HALF_BORDER 0x04
+#define V_008F30_SQ_TEX_MIRROR_ONCE_HALF_BORDER 0x05
+#define V_008F30_SQ_TEX_CLAMP_BORDER 0x06
+#define V_008F30_SQ_TEX_MIRROR_ONCE_BORDER 0x07
+#define S_008F30_CLAMP_Y(x) (((x) & 0x07) << 3)
+#define G_008F30_CLAMP_Y(x) (((x) >> 3) & 0x07)
+#define C_008F30_CLAMP_Y 0xFFFFFFC7
+#define V_008F30_SQ_TEX_WRAP 0x00
+#define V_008F30_SQ_TEX_MIRROR 0x01
+#define V_008F30_SQ_TEX_CLAMP_LAST_TEXEL 0x02
+#define V_008F30_SQ_TEX_MIRROR_ONCE_LAST_TEXEL 0x03
+#define V_008F30_SQ_TEX_CLAMP_HALF_BORDER 0x04
+#define V_008F30_SQ_TEX_MIRROR_ONCE_HALF_BORDER 0x05
+#define V_008F30_SQ_TEX_CLAMP_BORDER 0x06
+#define V_008F30_SQ_TEX_MIRROR_ONCE_BORDER 0x07
+#define S_008F30_CLAMP_Z(x) (((x) & 0x07) << 6)
+#define G_008F30_CLAMP_Z(x) (((x) >> 6) & 0x07)
+#define C_008F30_CLAMP_Z 0xFFFFFE3F
+#define V_008F30_SQ_TEX_WRAP 0x00
+#define V_008F30_SQ_TEX_MIRROR 0x01
+#define V_008F30_SQ_TEX_CLAMP_LAST_TEXEL 0x02
+#define V_008F30_SQ_TEX_MIRROR_ONCE_LAST_TEXEL 0x03
+#define V_008F30_SQ_TEX_CLAMP_HALF_BORDER 0x04
+#define V_008F30_SQ_TEX_MIRROR_ONCE_HALF_BORDER 0x05
+#define V_008F30_SQ_TEX_CLAMP_BORDER 0x06
+#define V_008F30_SQ_TEX_MIRROR_ONCE_BORDER 0x07
+#define S_008F30_DEPTH_COMPARE_FUNC(x) (((x) & 0x07) << 12)
+#define G_008F30_DEPTH_COMPARE_FUNC(x) (((x) >> 12) & 0x07)
+#define C_008F30_DEPTH_COMPARE_FUNC 0xFFFF8FFF
+#define V_008F30_SQ_TEX_DEPTH_COMPARE_NEVER 0x00
+#define V_008F30_SQ_TEX_DEPTH_COMPARE_LESS 0x01
+#define V_008F30_SQ_TEX_DEPTH_COMPARE_EQUAL 0x02
+#define V_008F30_SQ_TEX_DEPTH_COMPARE_LESSEQUAL 0x03
+#define V_008F30_SQ_TEX_DEPTH_COMPARE_GREATER 0x04
+#define V_008F30_SQ_TEX_DEPTH_COMPARE_NOTEQUAL 0x05
+#define V_008F30_SQ_TEX_DEPTH_COMPARE_GREATEREQUAL 0x06
+#define V_008F30_SQ_TEX_DEPTH_COMPARE_ALWAYS 0x07
+#define S_008F30_FORCE_UNNORMALIZED(x) (((x) & 0x1) << 15)
+#define G_008F30_FORCE_UNNORMALIZED(x) (((x) >> 15) & 0x1)
+#define C_008F30_FORCE_UNNORMALIZED 0xFFFF7FFF
+#define S_008F30_MC_COORD_TRUNC(x) (((x) & 0x1) << 19)
+#define G_008F30_MC_COORD_TRUNC(x) (((x) >> 19) & 0x1)
+#define C_008F30_MC_COORD_TRUNC 0xFFF7FFFF
+#define S_008F30_FORCE_DEGAMMA(x) (((x) & 0x1) << 20)
+#define G_008F30_FORCE_DEGAMMA(x) (((x) >> 20) & 0x1)
+#define C_008F30_FORCE_DEGAMMA 0xFFEFFFFF
+#define S_008F30_TRUNC_COORD(x) (((x) & 0x1) << 27)
+#define G_008F30_TRUNC_COORD(x) (((x) >> 27) & 0x1)
+#define C_008F30_TRUNC_COORD 0xF7FFFFFF
+#define S_008F30_DISABLE_CUBE_WRAP(x) (((x) & 0x1) << 28)
+#define G_008F30_DISABLE_CUBE_WRAP(x) (((x) >> 28) & 0x1)
+#define C_008F30_DISABLE_CUBE_WRAP 0xEFFFFFFF
+#define S_008F30_FILTER_MODE(x) (((x) & 0x03) << 29)
+#define G_008F30_FILTER_MODE(x) (((x) >> 29) & 0x03)
+#define C_008F30_FILTER_MODE 0x9FFFFFFF
+#define R_008F34_SQ_IMG_SAMP_WORD1 0x008F34
+#define S_008F34_MIN_LOD(x) (((x) & 0xFFF) << 0)
+#define G_008F34_MIN_LOD(x) (((x) >> 0) & 0xFFF)
+#define C_008F34_MIN_LOD 0xFFFFF000
+#define S_008F34_MAX_LOD(x) (((x) & 0xFFF) << 12)
+#define G_008F34_MAX_LOD(x) (((x) >> 12) & 0xFFF)
+#define C_008F34_MAX_LOD 0xFF000FFF
+#define S_008F34_PERF_MIP(x) (((x) & 0x0F) << 24)
+#define G_008F34_PERF_MIP(x) (((x) >> 24) & 0x0F)
+#define C_008F34_PERF_MIP 0xF0FFFFFF
+#define S_008F34_PERF_Z(x) (((x) & 0x0F) << 28)
+#define G_008F34_PERF_Z(x) (((x) >> 28) & 0x0F)
+#define C_008F34_PERF_Z 0x0FFFFFFF
+#define R_008F38_SQ_IMG_SAMP_WORD2 0x008F38
+#define S_008F38_LOD_BIAS(x) (((x) & 0x3FFF) << 0)
+#define G_008F38_LOD_BIAS(x) (((x) >> 0) & 0x3FFF)
+#define C_008F38_LOD_BIAS 0xFFFFC000
+#define S_008F38_LOD_BIAS_SEC(x) (((x) & 0x3F) << 14)
+#define G_008F38_LOD_BIAS_SEC(x) (((x) >> 14) & 0x3F)
+#define C_008F38_LOD_BIAS_SEC 0xFFF03FFF
+#define S_008F38_XY_MAG_FILTER(x) (((x) & 0x03) << 20)
+#define G_008F38_XY_MAG_FILTER(x) (((x) >> 20) & 0x03)
+#define C_008F38_XY_MAG_FILTER 0xFFCFFFFF
+#define V_008F38_SQ_TEX_XY_FILTER_POINT 0x00
+#define V_008F38_SQ_TEX_XY_FILTER_BILINEAR 0x01
+#define S_008F38_XY_MIN_FILTER(x) (((x) & 0x03) << 22)
+#define G_008F38_XY_MIN_FILTER(x) (((x) >> 22) & 0x03)
+#define C_008F38_XY_MIN_FILTER 0xFF3FFFFF
+#define V_008F38_SQ_TEX_XY_FILTER_POINT 0x00
+#define V_008F38_SQ_TEX_XY_FILTER_BILINEAR 0x01
+#define S_008F38_Z_FILTER(x) (((x) & 0x03) << 24)
+#define G_008F38_Z_FILTER(x) (((x) >> 24) & 0x03)
+#define C_008F38_Z_FILTER 0xFCFFFFFF
+#define V_008F38_SQ_TEX_Z_FILTER_NONE 0x00
+#define V_008F38_SQ_TEX_Z_FILTER_POINT 0x01
+#define V_008F38_SQ_TEX_Z_FILTER_LINEAR 0x02
+#define S_008F38_MIP_FILTER(x) (((x) & 0x03) << 26)
+#define G_008F38_MIP_FILTER(x) (((x) >> 26) & 0x03)
+#define C_008F38_MIP_FILTER 0xF3FFFFFF
+#define V_008F38_SQ_TEX_Z_FILTER_NONE 0x00
+#define V_008F38_SQ_TEX_Z_FILTER_POINT 0x01
+#define V_008F38_SQ_TEX_Z_FILTER_LINEAR 0x02
+#define S_008F38_MIP_POINT_PRECLAMP(x) (((x) & 0x1) << 28)
+#define G_008F38_MIP_POINT_PRECLAMP(x) (((x) >> 28) & 0x1)
+#define C_008F38_MIP_POINT_PRECLAMP 0xEFFFFFFF
+#define S_008F38_DISABLE_LSB_CEIL(x) (((x) & 0x1) << 29)
+#define G_008F38_DISABLE_LSB_CEIL(x) (((x) >> 29) & 0x1)
+#define C_008F38_DISABLE_LSB_CEIL 0xDFFFFFFF
+#define S_008F38_FILTER_PREC_FIX(x) (((x) & 0x1) << 30)
+#define G_008F38_FILTER_PREC_FIX(x) (((x) >> 30) & 0x1)
+#define C_008F38_FILTER_PREC_FIX 0xBFFFFFFF
+#define R_008F3C_SQ_IMG_SAMP_WORD3 0x008F3C
+#define S_008F3C_BORDER_COLOR_PTR(x) (((x) & 0xFFF) << 0)
+#define G_008F3C_BORDER_COLOR_PTR(x) (((x) >> 0) & 0xFFF)
+#define C_008F3C_BORDER_COLOR_PTR 0xFFFFF000
+#define S_008F3C_BORDER_COLOR_TYPE(x) (((x) & 0x03) << 30)
+#define G_008F3C_BORDER_COLOR_TYPE(x) (((x) >> 30) & 0x03)
+#define C_008F3C_BORDER_COLOR_TYPE 0x3FFFFFFF
+#define V_008F3C_SQ_TEX_BORDER_COLOR_TRANS_BLACK 0x00
+#define V_008F3C_SQ_TEX_BORDER_COLOR_OPAQUE_BLACK 0x01
+#define V_008F3C_SQ_TEX_BORDER_COLOR_OPAQUE_WHITE 0x02
+#define V_008F3C_SQ_TEX_BORDER_COLOR_REGISTER 0x03
+#define R_0090DC_SPI_DYN_GPR_LOCK_EN 0x0090DC
+#define S_0090DC_VS_LOW_THRESHOLD(x) (((x) & 0x0F) << 0)
+#define G_0090DC_VS_LOW_THRESHOLD(x) (((x) >> 0) & 0x0F)
+#define C_0090DC_VS_LOW_THRESHOLD 0xFFFFFFF0
+#define S_0090DC_GS_LOW_THRESHOLD(x) (((x) & 0x0F) << 4)
+#define G_0090DC_GS_LOW_THRESHOLD(x) (((x) >> 4) & 0x0F)
+#define C_0090DC_GS_LOW_THRESHOLD 0xFFFFFF0F
+#define S_0090DC_ES_LOW_THRESHOLD(x) (((x) & 0x0F) << 8)
+#define G_0090DC_ES_LOW_THRESHOLD(x) (((x) >> 8) & 0x0F)
+#define C_0090DC_ES_LOW_THRESHOLD 0xFFFFF0FF
+#define S_0090DC_HS_LOW_THRESHOLD(x) (((x) & 0x0F) << 12)
+#define G_0090DC_HS_LOW_THRESHOLD(x) (((x) >> 12) & 0x0F)
+#define C_0090DC_HS_LOW_THRESHOLD 0xFFFF0FFF
+#define S_0090DC_LS_LOW_THRESHOLD(x) (((x) & 0x0F) << 16)
+#define G_0090DC_LS_LOW_THRESHOLD(x) (((x) >> 16) & 0x0F)
+#define C_0090DC_LS_LOW_THRESHOLD 0xFFF0FFFF
+#define R_0090E0_SPI_STATIC_THREAD_MGMT_1 0x0090E0
+#define S_0090E0_PS_CU_EN(x) (((x) & 0xFFFF) << 0)
+#define G_0090E0_PS_CU_EN(x) (((x) >> 0) & 0xFFFF)
+#define C_0090E0_PS_CU_EN 0xFFFF0000
+#define S_0090E0_VS_CU_EN(x) (((x) & 0xFFFF) << 16)
+#define G_0090E0_VS_CU_EN(x) (((x) >> 16) & 0xFFFF)
+#define C_0090E0_VS_CU_EN 0x0000FFFF
+#define R_0090E4_SPI_STATIC_THREAD_MGMT_2 0x0090E4
+#define S_0090E4_GS_CU_EN(x) (((x) & 0xFFFF) << 0)
+#define G_0090E4_GS_CU_EN(x) (((x) >> 0) & 0xFFFF)
+#define C_0090E4_GS_CU_EN 0xFFFF0000
+#define S_0090E4_ES_CU_EN(x) (((x) & 0xFFFF) << 16)
+#define G_0090E4_ES_CU_EN(x) (((x) >> 16) & 0xFFFF)
+#define C_0090E4_ES_CU_EN 0x0000FFFF
+#define R_0090E8_SPI_STATIC_THREAD_MGMT_3 0x0090E8
+#define S_0090E8_LSHS_CU_EN(x) (((x) & 0xFFFF) << 0)
+#define G_0090E8_LSHS_CU_EN(x) (((x) >> 0) & 0xFFFF)
+#define C_0090E8_LSHS_CU_EN 0xFFFF0000
+#define R_0090EC_SPI_PS_MAX_WAVE_ID 0x0090EC
+#define S_0090EC_MAX_WAVE_ID(x) (((x) & 0xFFF) << 0)
+#define G_0090EC_MAX_WAVE_ID(x) (((x) >> 0) & 0xFFF)
+#define C_0090EC_MAX_WAVE_ID 0xFFFFF000
+#define R_0090F0_SPI_ARB_PRIORITY 0x0090F0
+#define S_0090F0_RING_ORDER_TS0(x) (((x) & 0x07) << 0)
+#define G_0090F0_RING_ORDER_TS0(x) (((x) >> 0) & 0x07)
+#define C_0090F0_RING_ORDER_TS0 0xFFFFFFF8
+#define V_0090F0_X_R0 0x00
+#define S_0090F0_RING_ORDER_TS1(x) (((x) & 0x07) << 3)
+#define G_0090F0_RING_ORDER_TS1(x) (((x) >> 3) & 0x07)
+#define C_0090F0_RING_ORDER_TS1 0xFFFFFFC7
+#define S_0090F0_RING_ORDER_TS2(x) (((x) & 0x07) << 6)
+#define G_0090F0_RING_ORDER_TS2(x) (((x) >> 6) & 0x07)
+#define C_0090F0_RING_ORDER_TS2 0xFFFFFE3F
+#define R_0090F4_SPI_ARB_CYCLES_0 0x0090F4
+#define S_0090F4_TS0_DURATION(x) (((x) & 0xFFFF) << 0)
+#define G_0090F4_TS0_DURATION(x) (((x) >> 0) & 0xFFFF)
+#define C_0090F4_TS0_DURATION 0xFFFF0000
+#define S_0090F4_TS1_DURATION(x) (((x) & 0xFFFF) << 16)
+#define G_0090F4_TS1_DURATION(x) (((x) >> 16) & 0xFFFF)
+#define C_0090F4_TS1_DURATION 0x0000FFFF
+#define R_0090F8_SPI_ARB_CYCLES_1 0x0090F8
+#define S_0090F8_TS2_DURATION(x) (((x) & 0xFFFF) << 0)
+#define G_0090F8_TS2_DURATION(x) (((x) >> 0) & 0xFFFF)
+#define C_0090F8_TS2_DURATION 0xFFFF0000
+#define R_009100_SPI_CONFIG_CNTL 0x009100
+#define S_009100_GPR_WRITE_PRIORITY(x) (((x) & 0x1FFFFF) << 0)
+#define G_009100_GPR_WRITE_PRIORITY(x) (((x) >> 0) & 0x1FFFFF)
+#define C_009100_GPR_WRITE_PRIORITY 0xFFE00000
+#define S_009100_EXP_PRIORITY_ORDER(x) (((x) & 0x07) << 21)
+#define G_009100_EXP_PRIORITY_ORDER(x) (((x) >> 21) & 0x07)
+#define C_009100_EXP_PRIORITY_ORDER 0xFF1FFFFF
+#define S_009100_ENABLE_SQG_TOP_EVENTS(x) (((x) & 0x1) << 24)
+#define G_009100_ENABLE_SQG_TOP_EVENTS(x) (((x) >> 24) & 0x1)
+#define C_009100_ENABLE_SQG_TOP_EVENTS 0xFEFFFFFF
+#define S_009100_ENABLE_SQG_BOP_EVENTS(x) (((x) & 0x1) << 25)
+#define G_009100_ENABLE_SQG_BOP_EVENTS(x) (((x) >> 25) & 0x1)
+#define C_009100_ENABLE_SQG_BOP_EVENTS 0xFDFFFFFF
+#define S_009100_RSRC_MGMT_RESET(x) (((x) & 0x1) << 26)
+#define G_009100_RSRC_MGMT_RESET(x) (((x) >> 26) & 0x1)
+#define C_009100_RSRC_MGMT_RESET 0xFBFFFFFF
+#define R_00913C_SPI_CONFIG_CNTL_1 0x00913C
+#define S_00913C_VTX_DONE_DELAY(x) (((x) & 0x0F) << 0)
+#define G_00913C_VTX_DONE_DELAY(x) (((x) >> 0) & 0x0F)
+#define C_00913C_VTX_DONE_DELAY 0xFFFFFFF0
+#define V_00913C_X_DELAY_14_CLKS 0x00
+#define V_00913C_X_DELAY_16_CLKS 0x01
+#define V_00913C_X_DELAY_18_CLKS 0x02
+#define V_00913C_X_DELAY_20_CLKS 0x03
+#define V_00913C_X_DELAY_22_CLKS 0x04
+#define V_00913C_X_DELAY_24_CLKS 0x05
+#define V_00913C_X_DELAY_26_CLKS 0x06
+#define V_00913C_X_DELAY_28_CLKS 0x07
+#define V_00913C_X_DELAY_30_CLKS 0x08
+#define V_00913C_X_DELAY_32_CLKS 0x09
+#define V_00913C_X_DELAY_34_CLKS 0x0A
+#define V_00913C_X_DELAY_4_CLKS 0x0B
+#define V_00913C_X_DELAY_6_CLKS 0x0C
+#define V_00913C_X_DELAY_8_CLKS 0x0D
+#define V_00913C_X_DELAY_10_CLKS 0x0E
+#define V_00913C_X_DELAY_12_CLKS 0x0F
+#define S_00913C_INTERP_ONE_PRIM_PER_ROW(x) (((x) & 0x1) << 4)
+#define G_00913C_INTERP_ONE_PRIM_PER_ROW(x) (((x) >> 4) & 0x1)
+#define C_00913C_INTERP_ONE_PRIM_PER_ROW 0xFFFFFFEF
+#define S_00913C_PC_LIMIT_ENABLE(x) (((x) & 0x1) << 6)
+#define G_00913C_PC_LIMIT_ENABLE(x) (((x) >> 6) & 0x1)
+#define C_00913C_PC_LIMIT_ENABLE 0xFFFFFFBF
+#define S_00913C_PC_LIMIT_STRICT(x) (((x) & 0x1) << 7)
+#define G_00913C_PC_LIMIT_STRICT(x) (((x) >> 7) & 0x1)
+#define C_00913C_PC_LIMIT_STRICT 0xFFFFFF7F
+#define S_00913C_PC_LIMIT_SIZE(x) (((x) & 0xFFFF) << 16)
+#define G_00913C_PC_LIMIT_SIZE(x) (((x) >> 16) & 0xFFFF)
+#define C_00913C_PC_LIMIT_SIZE 0x0000FFFF
+#define R_00936C_SPI_RESOURCE_RESERVE_CU_AB_0 0x00936C
+#define S_00936C_TYPE_A(x) (((x) & 0x0F) << 0)
+#define G_00936C_TYPE_A(x) (((x) >> 0) & 0x0F)
+#define C_00936C_TYPE_A 0xFFFFFFF0
+#define S_00936C_VGPR_A(x) (((x) & 0x07) << 4)
+#define G_00936C_VGPR_A(x) (((x) >> 4) & 0x07)
+#define C_00936C_VGPR_A 0xFFFFFF8F
+#define S_00936C_SGPR_A(x) (((x) & 0x07) << 7)
+#define G_00936C_SGPR_A(x) (((x) >> 7) & 0x07)
+#define C_00936C_SGPR_A 0xFFFFFC7F
+#define S_00936C_LDS_A(x) (((x) & 0x07) << 10)
+#define G_00936C_LDS_A(x) (((x) >> 10) & 0x07)
+#define C_00936C_LDS_A 0xFFFFE3FF
+#define S_00936C_WAVES_A(x) (((x) & 0x03) << 13)
+#define G_00936C_WAVES_A(x) (((x) >> 13) & 0x03)
+#define C_00936C_WAVES_A 0xFFFF9FFF
+#define S_00936C_EN_A(x) (((x) & 0x1) << 15)
+#define G_00936C_EN_A(x) (((x) >> 15) & 0x1)
+#define C_00936C_EN_A 0xFFFF7FFF
+#define S_00936C_TYPE_B(x) (((x) & 0x0F) << 16)
+#define G_00936C_TYPE_B(x) (((x) >> 16) & 0x0F)
+#define C_00936C_TYPE_B 0xFFF0FFFF
+#define S_00936C_VGPR_B(x) (((x) & 0x07) << 20)
+#define G_00936C_VGPR_B(x) (((x) >> 20) & 0x07)
+#define C_00936C_VGPR_B 0xFF8FFFFF
+#define S_00936C_SGPR_B(x) (((x) & 0x07) << 23)
+#define G_00936C_SGPR_B(x) (((x) >> 23) & 0x07)
+#define C_00936C_SGPR_B 0xFC7FFFFF
+#define S_00936C_LDS_B(x) (((x) & 0x07) << 26)
+#define G_00936C_LDS_B(x) (((x) >> 26) & 0x07)
+#define C_00936C_LDS_B 0xE3FFFFFF
+#define S_00936C_WAVES_B(x) (((x) & 0x03) << 29)
+#define G_00936C_WAVES_B(x) (((x) >> 29) & 0x03)
+#define C_00936C_WAVES_B 0x9FFFFFFF
+#define S_00936C_EN_B(x) (((x) & 0x1) << 31)
+#define G_00936C_EN_B(x) (((x) >> 31) & 0x1)
+#define C_00936C_EN_B 0x7FFFFFFF
+#define R_00950C_TA_CS_BC_BASE_ADDR 0x00950C
+#define R_009858_DB_SUBTILE_CONTROL 0x009858
+#define S_009858_MSAA1_X(x) (((x) & 0x03) << 0)
+#define G_009858_MSAA1_X(x) (((x) >> 0) & 0x03)
+#define C_009858_MSAA1_X 0xFFFFFFFC
+#define S_009858_MSAA1_Y(x) (((x) & 0x03) << 2)
+#define G_009858_MSAA1_Y(x) (((x) >> 2) & 0x03)
+#define C_009858_MSAA1_Y 0xFFFFFFF3
+#define S_009858_MSAA2_X(x) (((x) & 0x03) << 4)
+#define G_009858_MSAA2_X(x) (((x) >> 4) & 0x03)
+#define C_009858_MSAA2_X 0xFFFFFFCF
+#define S_009858_MSAA2_Y(x) (((x) & 0x03) << 6)
+#define G_009858_MSAA2_Y(x) (((x) >> 6) & 0x03)
+#define C_009858_MSAA2_Y 0xFFFFFF3F
+#define S_009858_MSAA4_X(x) (((x) & 0x03) << 8)
+#define G_009858_MSAA4_X(x) (((x) >> 8) & 0x03)
+#define C_009858_MSAA4_X 0xFFFFFCFF
+#define S_009858_MSAA4_Y(x) (((x) & 0x03) << 10)
+#define G_009858_MSAA4_Y(x) (((x) >> 10) & 0x03)
+#define C_009858_MSAA4_Y 0xFFFFF3FF
+#define S_009858_MSAA8_X(x) (((x) & 0x03) << 12)
+#define G_009858_MSAA8_X(x) (((x) >> 12) & 0x03)
+#define C_009858_MSAA8_X 0xFFFFCFFF
+#define S_009858_MSAA8_Y(x) (((x) & 0x03) << 14)
+#define G_009858_MSAA8_Y(x) (((x) >> 14) & 0x03)
+#define C_009858_MSAA8_Y 0xFFFF3FFF
+#define S_009858_MSAA16_X(x) (((x) & 0x03) << 16)
+#define G_009858_MSAA16_X(x) (((x) >> 16) & 0x03)
+#define C_009858_MSAA16_X 0xFFFCFFFF
+#define S_009858_MSAA16_Y(x) (((x) & 0x03) << 18)
+#define G_009858_MSAA16_Y(x) (((x) >> 18) & 0x03)
+#define C_009858_MSAA16_Y 0xFFF3FFFF
+#define R_009910_GB_TILE_MODE0 0x009910
+#define S_009910_MICRO_TILE_MODE(x) (((x) & 0x03) << 0)
+#define G_009910_MICRO_TILE_MODE(x) (((x) >> 0) & 0x03)
+#define C_009910_MICRO_TILE_MODE 0xFFFFFFFC
+#define V_009910_ADDR_SURF_DISPLAY_MICRO_TILING 0x00
+#define V_009910_ADDR_SURF_THIN_MICRO_TILING 0x01
+#define V_009910_ADDR_SURF_DEPTH_MICRO_TILING 0x02
+#define V_009910_ADDR_SURF_THICK_MICRO_TILING 0x03
+#define S_009910_ARRAY_MODE(x) (((x) & 0x0F) << 2)
+#define G_009910_ARRAY_MODE(x) (((x) >> 2) & 0x0F)
+#define C_009910_ARRAY_MODE 0xFFFFFFC3
+#define V_009910_ARRAY_LINEAR_GENERAL 0x00
+#define V_009910_ARRAY_LINEAR_ALIGNED 0x01
+#define V_009910_ARRAY_1D_TILED_THIN1 0x02
+#define V_009910_ARRAY_1D_TILED_THICK 0x03
+#define V_009910_ARRAY_2D_TILED_THIN1 0x04
+#define V_009910_ARRAY_2D_TILED_THICK 0x07
+#define V_009910_ARRAY_2D_TILED_XTHICK 0x08
+#define V_009910_ARRAY_3D_TILED_THIN1 0x0C
+#define V_009910_ARRAY_3D_TILED_THICK 0x0D
+#define V_009910_ARRAY_3D_TILED_XTHICK 0x0E
+#define V_009910_ARRAY_POWER_SAVE 0x0F
+#define S_009910_PIPE_CONFIG(x) (((x) & 0x1F) << 6)
+#define G_009910_PIPE_CONFIG(x) (((x) >> 6) & 0x1F)
+#define C_009910_PIPE_CONFIG 0xFFFFF83F
+#define V_009910_ADDR_SURF_P2 0x00
+#define V_009910_ADDR_SURF_P2_RESERVED0 0x01
+#define V_009910_ADDR_SURF_P2_RESERVED1 0x02
+#define V_009910_ADDR_SURF_P2_RESERVED2 0x03
+#define V_009910_X_ADDR_SURF_P4_8X16 0x04
+#define V_009910_X_ADDR_SURF_P4_16X16 0x05
+#define V_009910_X_ADDR_SURF_P4_16X32 0x06
+#define V_009910_X_ADDR_SURF_P4_32X32 0x07
+#define V_009910_X_ADDR_SURF_P8_16X16_8X16 0x08
+#define V_009910_X_ADDR_SURF_P8_16X32_8X16 0x09
+#define V_009910_X_ADDR_SURF_P8_32X32_8X16 0x0A
+#define V_009910_X_ADDR_SURF_P8_16X32_16X16 0x0B
+#define V_009910_X_ADDR_SURF_P8_32X32_16X16 0x0C
+#define V_009910_X_ADDR_SURF_P8_32X32_16X32 0x0D
+#define V_009910_X_ADDR_SURF_P8_32X64_32X32 0x0E
+#define S_009910_TILE_SPLIT(x) (((x) & 0x07) << 11)
+#define G_009910_TILE_SPLIT(x) (((x) >> 11) & 0x07)
+#define C_009910_TILE_SPLIT 0xFFFFC7FF
+#define V_009910_ADDR_SURF_TILE_SPLIT_64B 0x00
+#define V_009910_ADDR_SURF_TILE_SPLIT_128B 0x01
+#define V_009910_ADDR_SURF_TILE_SPLIT_256B 0x02
+#define V_009910_ADDR_SURF_TILE_SPLIT_512B 0x03
+#define V_009910_ADDR_SURF_TILE_SPLIT_1KB 0x04
+#define V_009910_ADDR_SURF_TILE_SPLIT_2KB 0x05
+#define V_009910_ADDR_SURF_TILE_SPLIT_4KB 0x06
+#define S_009910_BANK_WIDTH(x) (((x) & 0x03) << 14)
+#define G_009910_BANK_WIDTH(x) (((x) >> 14) & 0x03)
+#define C_009910_BANK_WIDTH 0xFFFF3FFF
+#define V_009910_ADDR_SURF_BANK_WIDTH_1 0x00
+#define V_009910_ADDR_SURF_BANK_WIDTH_2 0x01
+#define V_009910_ADDR_SURF_BANK_WIDTH_4 0x02
+#define V_009910_ADDR_SURF_BANK_WIDTH_8 0x03
+#define S_009910_BANK_HEIGHT(x) (((x) & 0x03) << 16)
+#define G_009910_BANK_HEIGHT(x) (((x) >> 16) & 0x03)
+#define C_009910_BANK_HEIGHT 0xFFFCFFFF
+#define V_009910_ADDR_SURF_BANK_HEIGHT_1 0x00
+#define V_009910_ADDR_SURF_BANK_HEIGHT_2 0x01
+#define V_009910_ADDR_SURF_BANK_HEIGHT_4 0x02
+#define V_009910_ADDR_SURF_BANK_HEIGHT_8 0x03
+#define S_009910_MACRO_TILE_ASPECT(x) (((x) & 0x03) << 18)
+#define G_009910_MACRO_TILE_ASPECT(x) (((x) >> 18) & 0x03)
+#define C_009910_MACRO_TILE_ASPECT 0xFFF3FFFF
+#define V_009910_ADDR_SURF_MACRO_ASPECT_1 0x00
+#define V_009910_ADDR_SURF_MACRO_ASPECT_2 0x01
+#define V_009910_ADDR_SURF_MACRO_ASPECT_4 0x02
+#define V_009910_ADDR_SURF_MACRO_ASPECT_8 0x03
+#define S_009910_NUM_BANKS(x) (((x) & 0x03) << 20)
+#define G_009910_NUM_BANKS(x) (((x) >> 20) & 0x03)
+#define C_009910_NUM_BANKS 0xFFCFFFFF
+#define V_009910_ADDR_SURF_2_BANK 0x00
+#define V_009910_ADDR_SURF_4_BANK 0x01
+#define V_009910_ADDR_SURF_8_BANK 0x02
+#define V_009910_ADDR_SURF_16_BANK 0x03
+#define R_00B020_SPI_SHADER_PGM_LO_PS 0x00B020
+#define R_00B024_SPI_SHADER_PGM_HI_PS 0x00B024
+#define S_00B024_MEM_BASE(x) (((x) & 0xFF) << 0)
+#define G_00B024_MEM_BASE(x) (((x) >> 0) & 0xFF)
+#define C_00B024_MEM_BASE 0xFFFFFF00
+#define R_00B028_SPI_SHADER_PGM_RSRC1_PS 0x00B028
+#define S_00B028_VGPRS(x) (((x) & 0x3F) << 0)
+#define G_00B028_VGPRS(x) (((x) >> 0) & 0x3F)
+#define C_00B028_VGPRS 0xFFFFFFC0
+#define S_00B028_SGPRS(x) (((x) & 0x0F) << 6)
+#define G_00B028_SGPRS(x) (((x) >> 6) & 0x0F)
+#define C_00B028_SGPRS 0xFFFFFC3F
+#define S_00B028_PRIORITY(x) (((x) & 0x03) << 10)
+#define G_00B028_PRIORITY(x) (((x) >> 10) & 0x03)
+#define C_00B028_PRIORITY 0xFFFFF3FF
+#define S_00B028_FLOAT_MODE(x) (((x) & 0xFF) << 12)
+#define G_00B028_FLOAT_MODE(x) (((x) >> 12) & 0xFF)
+#define C_00B028_FLOAT_MODE 0xFFF00FFF
+#define S_00B028_PRIV(x) (((x) & 0x1) << 20)
+#define G_00B028_PRIV(x) (((x) >> 20) & 0x1)
+#define C_00B028_PRIV 0xFFEFFFFF
+#define S_00B028_DX10_CLAMP(x) (((x) & 0x1) << 21)
+#define G_00B028_DX10_CLAMP(x) (((x) >> 21) & 0x1)
+#define C_00B028_DX10_CLAMP 0xFFDFFFFF
+#define S_00B028_DEBUG_MODE(x) (((x) & 0x1) << 22)
+#define G_00B028_DEBUG_MODE(x) (((x) >> 22) & 0x1)
+#define C_00B028_DEBUG_MODE 0xFFBFFFFF
+#define S_00B028_IEEE_MODE(x) (((x) & 0x1) << 23)
+#define G_00B028_IEEE_MODE(x) (((x) >> 23) & 0x1)
+#define C_00B028_IEEE_MODE 0xFF7FFFFF
+#define S_00B028_CU_GROUP_DISABLE(x) (((x) & 0x1) << 24)
+#define G_00B028_CU_GROUP_DISABLE(x) (((x) >> 24) & 0x1)
+#define C_00B028_CU_GROUP_DISABLE 0xFEFFFFFF
+#define R_00B02C_SPI_SHADER_PGM_RSRC2_PS 0x00B02C
+#define S_00B02C_SCRATCH_EN(x) (((x) & 0x1) << 0)
+#define G_00B02C_SCRATCH_EN(x) (((x) >> 0) & 0x1)
+#define C_00B02C_SCRATCH_EN 0xFFFFFFFE
+#define S_00B02C_USER_SGPR(x) (((x) & 0x1F) << 1)
+#define G_00B02C_USER_SGPR(x) (((x) >> 1) & 0x1F)
+#define C_00B02C_USER_SGPR 0xFFFFFFC1
+#define S_00B02C_WAVE_CNT_EN(x) (((x) & 0x1) << 7)
+#define G_00B02C_WAVE_CNT_EN(x) (((x) >> 7) & 0x1)
+#define C_00B02C_WAVE_CNT_EN 0xFFFFFF7F
+#define S_00B02C_EXTRA_LDS_SIZE(x) (((x) & 0xFF) << 8)
+#define G_00B02C_EXTRA_LDS_SIZE(x) (((x) >> 8) & 0xFF)
+#define C_00B02C_EXTRA_LDS_SIZE 0xFFFF00FF
+#define S_00B02C_EXCP_EN(x) (((x) & 0x7F) << 16)
+#define G_00B02C_EXCP_EN(x) (((x) >> 16) & 0x7F)
+#define C_00B02C_EXCP_EN 0xFF80FFFF
+#define R_00B030_SPI_SHADER_USER_DATA_PS_0 0x00B030
+#define R_00B034_SPI_SHADER_USER_DATA_PS_1 0x00B034
+#define R_00B038_SPI_SHADER_USER_DATA_PS_2 0x00B038
+#define R_00B03C_SPI_SHADER_USER_DATA_PS_3 0x00B03C
+#define R_00B040_SPI_SHADER_USER_DATA_PS_4 0x00B040
+#define R_00B044_SPI_SHADER_USER_DATA_PS_5 0x00B044
+#define R_00B048_SPI_SHADER_USER_DATA_PS_6 0x00B048
+#define R_00B04C_SPI_SHADER_USER_DATA_PS_7 0x00B04C
+#define R_00B050_SPI_SHADER_USER_DATA_PS_8 0x00B050
+#define R_00B054_SPI_SHADER_USER_DATA_PS_9 0x00B054
+#define R_00B058_SPI_SHADER_USER_DATA_PS_10 0x00B058
+#define R_00B05C_SPI_SHADER_USER_DATA_PS_11 0x00B05C
+#define R_00B060_SPI_SHADER_USER_DATA_PS_12 0x00B060
+#define R_00B064_SPI_SHADER_USER_DATA_PS_13 0x00B064
+#define R_00B068_SPI_SHADER_USER_DATA_PS_14 0x00B068
+#define R_00B06C_SPI_SHADER_USER_DATA_PS_15 0x00B06C
+#define R_00B120_SPI_SHADER_PGM_LO_VS 0x00B120
+#define R_00B124_SPI_SHADER_PGM_HI_VS 0x00B124
+#define S_00B124_MEM_BASE(x) (((x) & 0xFF) << 0)
+#define G_00B124_MEM_BASE(x) (((x) >> 0) & 0xFF)
+#define C_00B124_MEM_BASE 0xFFFFFF00
+#define R_00B128_SPI_SHADER_PGM_RSRC1_VS 0x00B128
+#define S_00B128_VGPRS(x) (((x) & 0x3F) << 0)
+#define G_00B128_VGPRS(x) (((x) >> 0) & 0x3F)
+#define C_00B128_VGPRS 0xFFFFFFC0
+#define S_00B128_SGPRS(x) (((x) & 0x0F) << 6)
+#define G_00B128_SGPRS(x) (((x) >> 6) & 0x0F)
+#define C_00B128_SGPRS 0xFFFFFC3F
+#define S_00B128_PRIORITY(x) (((x) & 0x03) << 10)
+#define G_00B128_PRIORITY(x) (((x) >> 10) & 0x03)
+#define C_00B128_PRIORITY 0xFFFFF3FF
+#define S_00B128_FLOAT_MODE(x) (((x) & 0xFF) << 12)
+#define G_00B128_FLOAT_MODE(x) (((x) >> 12) & 0xFF)
+#define C_00B128_FLOAT_MODE 0xFFF00FFF
+#define S_00B128_PRIV(x) (((x) & 0x1) << 20)
+#define G_00B128_PRIV(x) (((x) >> 20) & 0x1)
+#define C_00B128_PRIV 0xFFEFFFFF
+#define S_00B128_DX10_CLAMP(x) (((x) & 0x1) << 21)
+#define G_00B128_DX10_CLAMP(x) (((x) >> 21) & 0x1)
+#define C_00B128_DX10_CLAMP 0xFFDFFFFF
+#define S_00B128_DEBUG_MODE(x) (((x) & 0x1) << 22)
+#define G_00B128_DEBUG_MODE(x) (((x) >> 22) & 0x1)
+#define C_00B128_DEBUG_MODE 0xFFBFFFFF
+#define S_00B128_IEEE_MODE(x) (((x) & 0x1) << 23)
+#define G_00B128_IEEE_MODE(x) (((x) >> 23) & 0x1)
+#define C_00B128_IEEE_MODE 0xFF7FFFFF
+#define S_00B128_VGPR_COMP_CNT(x) (((x) & 0x03) << 24)
+#define G_00B128_VGPR_COMP_CNT(x) (((x) >> 24) & 0x03)
+#define C_00B128_VGPR_COMP_CNT 0xFCFFFFFF
+#define S_00B128_CU_GROUP_ENABLE(x) (((x) & 0x1) << 26)
+#define G_00B128_CU_GROUP_ENABLE(x) (((x) >> 26) & 0x1)
+#define C_00B128_CU_GROUP_ENABLE 0xFBFFFFFF
+#define R_00B12C_SPI_SHADER_PGM_RSRC2_VS 0x00B12C
+#define S_00B12C_SCRATCH_EN(x) (((x) & 0x1) << 0)
+#define G_00B12C_SCRATCH_EN(x) (((x) >> 0) & 0x1)
+#define C_00B12C_SCRATCH_EN 0xFFFFFFFE
+#define S_00B12C_USER_SGPR(x) (((x) & 0x1F) << 1)
+#define G_00B12C_USER_SGPR(x) (((x) >> 1) & 0x1F)
+#define C_00B12C_USER_SGPR 0xFFFFFFC1
+#define S_00B12C_OC_LDS_EN(x) (((x) & 0x1) << 7)
+#define G_00B12C_OC_LDS_EN(x) (((x) >> 7) & 0x1)
+#define C_00B12C_OC_LDS_EN 0xFFFFFF7F
+#define S_00B12C_SO_BASE0_EN(x) (((x) & 0x1) << 8)
+#define G_00B12C_SO_BASE0_EN(x) (((x) >> 8) & 0x1)
+#define C_00B12C_SO_BASE0_EN 0xFFFFFEFF
+#define S_00B12C_SO_BASE1_EN(x) (((x) & 0x1) << 9)
+#define G_00B12C_SO_BASE1_EN(x) (((x) >> 9) & 0x1)
+#define C_00B12C_SO_BASE1_EN 0xFFFFFDFF
+#define S_00B12C_SO_BASE2_EN(x) (((x) & 0x1) << 10)
+#define G_00B12C_SO_BASE2_EN(x) (((x) >> 10) & 0x1)
+#define C_00B12C_SO_BASE2_EN 0xFFFFFBFF
+#define S_00B12C_SO_BASE3_EN(x) (((x) & 0x1) << 11)
+#define G_00B12C_SO_BASE3_EN(x) (((x) >> 11) & 0x1)
+#define C_00B12C_SO_BASE3_EN 0xFFFFF7FF
+#define S_00B12C_SO_EN(x) (((x) & 0x1) << 12)
+#define G_00B12C_SO_EN(x) (((x) >> 12) & 0x1)
+#define C_00B12C_SO_EN 0xFFFFEFFF
+#define S_00B12C_EXCP_EN(x) (((x) & 0x7F) << 13)
+#define G_00B12C_EXCP_EN(x) (((x) >> 13) & 0x7F)
+#define C_00B12C_EXCP_EN 0xFFF01FFF
+#define R_00B130_SPI_SHADER_USER_DATA_VS_0 0x00B130
+#define R_00B134_SPI_SHADER_USER_DATA_VS_1 0x00B134
+#define R_00B138_SPI_SHADER_USER_DATA_VS_2 0x00B138
+#define R_00B13C_SPI_SHADER_USER_DATA_VS_3 0x00B13C
+#define R_00B140_SPI_SHADER_USER_DATA_VS_4 0x00B140
+#define R_00B144_SPI_SHADER_USER_DATA_VS_5 0x00B144
+#define R_00B148_SPI_SHADER_USER_DATA_VS_6 0x00B148
+#define R_00B14C_SPI_SHADER_USER_DATA_VS_7 0x00B14C
+#define R_00B150_SPI_SHADER_USER_DATA_VS_8 0x00B150
+#define R_00B154_SPI_SHADER_USER_DATA_VS_9 0x00B154
+#define R_00B158_SPI_SHADER_USER_DATA_VS_10 0x00B158
+#define R_00B15C_SPI_SHADER_USER_DATA_VS_11 0x00B15C
+#define R_00B160_SPI_SHADER_USER_DATA_VS_12 0x00B160
+#define R_00B164_SPI_SHADER_USER_DATA_VS_13 0x00B164
+#define R_00B168_SPI_SHADER_USER_DATA_VS_14 0x00B168
+#define R_00B16C_SPI_SHADER_USER_DATA_VS_15 0x00B16C
+#define R_00B220_SPI_SHADER_PGM_LO_GS 0x00B220
+#define R_00B224_SPI_SHADER_PGM_HI_GS 0x00B224
+#define S_00B224_MEM_BASE(x) (((x) & 0xFF) << 0)
+#define G_00B224_MEM_BASE(x) (((x) >> 0) & 0xFF)
+#define C_00B224_MEM_BASE 0xFFFFFF00
+#define R_00B228_SPI_SHADER_PGM_RSRC1_GS 0x00B228
+#define S_00B228_VGPRS(x) (((x) & 0x3F) << 0)
+#define G_00B228_VGPRS(x) (((x) >> 0) & 0x3F)
+#define C_00B228_VGPRS 0xFFFFFFC0
+#define S_00B228_SGPRS(x) (((x) & 0x0F) << 6)
+#define G_00B228_SGPRS(x) (((x) >> 6) & 0x0F)
+#define C_00B228_SGPRS 0xFFFFFC3F
+#define S_00B228_PRIORITY(x) (((x) & 0x03) << 10)
+#define G_00B228_PRIORITY(x) (((x) >> 10) & 0x03)
+#define C_00B228_PRIORITY 0xFFFFF3FF
+#define S_00B228_FLOAT_MODE(x) (((x) & 0xFF) << 12)
+#define G_00B228_FLOAT_MODE(x) (((x) >> 12) & 0xFF)
+#define C_00B228_FLOAT_MODE 0xFFF00FFF
+#define S_00B228_PRIV(x) (((x) & 0x1) << 20)
+#define G_00B228_PRIV(x) (((x) >> 20) & 0x1)
+#define C_00B228_PRIV 0xFFEFFFFF
+#define S_00B228_DX10_CLAMP(x) (((x) & 0x1) << 21)
+#define G_00B228_DX10_CLAMP(x) (((x) >> 21) & 0x1)
+#define C_00B228_DX10_CLAMP 0xFFDFFFFF
+#define S_00B228_DEBUG_MODE(x) (((x) & 0x1) << 22)
+#define G_00B228_DEBUG_MODE(x) (((x) >> 22) & 0x1)
+#define C_00B228_DEBUG_MODE 0xFFBFFFFF
+#define S_00B228_IEEE_MODE(x) (((x) & 0x1) << 23)
+#define G_00B228_IEEE_MODE(x) (((x) >> 23) & 0x1)
+#define C_00B228_IEEE_MODE 0xFF7FFFFF
+#define S_00B228_CU_GROUP_ENABLE(x) (((x) & 0x1) << 24)
+#define G_00B228_CU_GROUP_ENABLE(x) (((x) >> 24) & 0x1)
+#define C_00B228_CU_GROUP_ENABLE 0xFEFFFFFF
+#define R_00B22C_SPI_SHADER_PGM_RSRC2_GS 0x00B22C
+#define S_00B22C_SCRATCH_EN(x) (((x) & 0x1) << 0)
+#define G_00B22C_SCRATCH_EN(x) (((x) >> 0) & 0x1)
+#define C_00B22C_SCRATCH_EN 0xFFFFFFFE
+#define S_00B22C_USER_SGPR(x) (((x) & 0x1F) << 1)
+#define G_00B22C_USER_SGPR(x) (((x) >> 1) & 0x1F)
+#define C_00B22C_USER_SGPR 0xFFFFFFC1
+#define S_00B22C_EXCP_EN(x) (((x) & 0x7F) << 7)
+#define G_00B22C_EXCP_EN(x) (((x) >> 7) & 0x7F)
+#define C_00B22C_EXCP_EN 0xFFFFC07F
+#define R_00B230_SPI_SHADER_USER_DATA_GS_0 0x00B230
+#define R_00B320_SPI_SHADER_PGM_LO_ES 0x00B320
+#define R_00B324_SPI_SHADER_PGM_HI_ES 0x00B324
+#define S_00B324_MEM_BASE(x) (((x) & 0xFF) << 0)
+#define G_00B324_MEM_BASE(x) (((x) >> 0) & 0xFF)
+#define C_00B324_MEM_BASE 0xFFFFFF00
+#define R_00B328_SPI_SHADER_PGM_RSRC1_ES 0x00B328
+#define S_00B328_VGPRS(x) (((x) & 0x3F) << 0)
+#define G_00B328_VGPRS(x) (((x) >> 0) & 0x3F)
+#define C_00B328_VGPRS 0xFFFFFFC0
+#define S_00B328_SGPRS(x) (((x) & 0x0F) << 6)
+#define G_00B328_SGPRS(x) (((x) >> 6) & 0x0F)
+#define C_00B328_SGPRS 0xFFFFFC3F
+#define S_00B328_PRIORITY(x) (((x) & 0x03) << 10)
+#define G_00B328_PRIORITY(x) (((x) >> 10) & 0x03)
+#define C_00B328_PRIORITY 0xFFFFF3FF
+#define S_00B328_FLOAT_MODE(x) (((x) & 0xFF) << 12)
+#define G_00B328_FLOAT_MODE(x) (((x) >> 12) & 0xFF)
+#define C_00B328_FLOAT_MODE 0xFFF00FFF
+#define S_00B328_PRIV(x) (((x) & 0x1) << 20)
+#define G_00B328_PRIV(x) (((x) >> 20) & 0x1)
+#define C_00B328_PRIV 0xFFEFFFFF
+#define S_00B328_DX10_CLAMP(x) (((x) & 0x1) << 21)
+#define G_00B328_DX10_CLAMP(x) (((x) >> 21) & 0x1)
+#define C_00B328_DX10_CLAMP 0xFFDFFFFF
+#define S_00B328_DEBUG_MODE(x) (((x) & 0x1) << 22)
+#define G_00B328_DEBUG_MODE(x) (((x) >> 22) & 0x1)
+#define C_00B328_DEBUG_MODE 0xFFBFFFFF
+#define S_00B328_IEEE_MODE(x) (((x) & 0x1) << 23)
+#define G_00B328_IEEE_MODE(x) (((x) >> 23) & 0x1)
+#define C_00B328_IEEE_MODE 0xFF7FFFFF
+#define S_00B328_VGPR_COMP_CNT(x) (((x) & 0x03) << 24)
+#define G_00B328_VGPR_COMP_CNT(x) (((x) >> 24) & 0x03)
+#define C_00B328_VGPR_COMP_CNT 0xFCFFFFFF
+#define S_00B328_CU_GROUP_ENABLE(x) (((x) & 0x1) << 26)
+#define G_00B328_CU_GROUP_ENABLE(x) (((x) >> 26) & 0x1)
+#define C_00B328_CU_GROUP_ENABLE 0xFBFFFFFF
+#define R_00B32C_SPI_SHADER_PGM_RSRC2_ES 0x00B32C
+#define S_00B32C_SCRATCH_EN(x) (((x) & 0x1) << 0)
+#define G_00B32C_SCRATCH_EN(x) (((x) >> 0) & 0x1)
+#define C_00B32C_SCRATCH_EN 0xFFFFFFFE
+#define S_00B32C_USER_SGPR(x) (((x) & 0x1F) << 1)
+#define G_00B32C_USER_SGPR(x) (((x) >> 1) & 0x1F)
+#define C_00B32C_USER_SGPR 0xFFFFFFC1
+#define S_00B32C_OC_LDS_EN(x) (((x) & 0x1) << 7)
+#define G_00B32C_OC_LDS_EN(x) (((x) >> 7) & 0x1)
+#define C_00B32C_OC_LDS_EN 0xFFFFFF7F
+#define S_00B32C_EXCP_EN(x) (((x) & 0x7F) << 8)
+#define G_00B32C_EXCP_EN(x) (((x) >> 8) & 0x7F)
+#define C_00B32C_EXCP_EN 0xFFFF80FF
+#define R_00B330_SPI_SHADER_USER_DATA_ES_0 0x00B330
+#define R_00B420_SPI_SHADER_PGM_LO_HS 0x00B420
+#define R_00B424_SPI_SHADER_PGM_HI_HS 0x00B424
+#define S_00B424_MEM_BASE(x) (((x) & 0xFF) << 0)
+#define G_00B424_MEM_BASE(x) (((x) >> 0) & 0xFF)
+#define C_00B424_MEM_BASE 0xFFFFFF00
+#define R_00B428_SPI_SHADER_PGM_RSRC1_HS 0x00B428
+#define S_00B428_VGPRS(x) (((x) & 0x3F) << 0)
+#define G_00B428_VGPRS(x) (((x) >> 0) & 0x3F)
+#define C_00B428_VGPRS 0xFFFFFFC0
+#define S_00B428_SGPRS(x) (((x) & 0x0F) << 6)
+#define G_00B428_SGPRS(x) (((x) >> 6) & 0x0F)
+#define C_00B428_SGPRS 0xFFFFFC3F
+#define S_00B428_PRIORITY(x) (((x) & 0x03) << 10)
+#define G_00B428_PRIORITY(x) (((x) >> 10) & 0x03)
+#define C_00B428_PRIORITY 0xFFFFF3FF
+#define S_00B428_FLOAT_MODE(x) (((x) & 0xFF) << 12)
+#define G_00B428_FLOAT_MODE(x) (((x) >> 12) & 0xFF)
+#define C_00B428_FLOAT_MODE 0xFFF00FFF
+#define S_00B428_PRIV(x) (((x) & 0x1) << 20)
+#define G_00B428_PRIV(x) (((x) >> 20) & 0x1)
+#define C_00B428_PRIV 0xFFEFFFFF
+#define S_00B428_DX10_CLAMP(x) (((x) & 0x1) << 21)
+#define G_00B428_DX10_CLAMP(x) (((x) >> 21) & 0x1)
+#define C_00B428_DX10_CLAMP 0xFFDFFFFF
+#define S_00B428_DEBUG_MODE(x) (((x) & 0x1) << 22)
+#define G_00B428_DEBUG_MODE(x) (((x) >> 22) & 0x1)
+#define C_00B428_DEBUG_MODE 0xFFBFFFFF
+#define S_00B428_IEEE_MODE(x) (((x) & 0x1) << 23)
+#define G_00B428_IEEE_MODE(x) (((x) >> 23) & 0x1)
+#define C_00B428_IEEE_MODE 0xFF7FFFFF
+#define R_00B42C_SPI_SHADER_PGM_RSRC2_HS 0x00B42C
+#define S_00B42C_SCRATCH_EN(x) (((x) & 0x1) << 0)
+#define G_00B42C_SCRATCH_EN(x) (((x) >> 0) & 0x1)
+#define C_00B42C_SCRATCH_EN 0xFFFFFFFE
+#define S_00B42C_USER_SGPR(x) (((x) & 0x1F) << 1)
+#define G_00B42C_USER_SGPR(x) (((x) >> 1) & 0x1F)
+#define C_00B42C_USER_SGPR 0xFFFFFFC1
+#define S_00B42C_OC_LDS_EN(x) (((x) & 0x1) << 7)
+#define G_00B42C_OC_LDS_EN(x) (((x) >> 7) & 0x1)
+#define C_00B42C_OC_LDS_EN 0xFFFFFF7F
+#define S_00B42C_TG_SIZE_EN(x) (((x) & 0x1) << 8)
+#define G_00B42C_TG_SIZE_EN(x) (((x) >> 8) & 0x1)
+#define C_00B42C_TG_SIZE_EN 0xFFFFFEFF
+#define S_00B42C_EXCP_EN(x) (((x) & 0x7F) << 9)
+#define G_00B42C_EXCP_EN(x) (((x) >> 9) & 0x7F)
+#define C_00B42C_EXCP_EN 0xFFFF01FF
+#define R_00B430_SPI_SHADER_USER_DATA_HS_0 0x00B430
+#define R_00B520_SPI_SHADER_PGM_LO_LS 0x00B520
+#define R_00B524_SPI_SHADER_PGM_HI_LS 0x00B524
+#define S_00B524_MEM_BASE(x) (((x) & 0xFF) << 0)
+#define G_00B524_MEM_BASE(x) (((x) >> 0) & 0xFF)
+#define C_00B524_MEM_BASE 0xFFFFFF00
+#define R_00B528_SPI_SHADER_PGM_RSRC1_LS 0x00B528
+#define S_00B528_VGPRS(x) (((x) & 0x3F) << 0)
+#define G_00B528_VGPRS(x) (((x) >> 0) & 0x3F)
+#define C_00B528_VGPRS 0xFFFFFFC0
+#define S_00B528_SGPRS(x) (((x) & 0x0F) << 6)
+#define G_00B528_SGPRS(x) (((x) >> 6) & 0x0F)
+#define C_00B528_SGPRS 0xFFFFFC3F
+#define S_00B528_PRIORITY(x) (((x) & 0x03) << 10)
+#define G_00B528_PRIORITY(x) (((x) >> 10) & 0x03)
+#define C_00B528_PRIORITY 0xFFFFF3FF
+#define S_00B528_FLOAT_MODE(x) (((x) & 0xFF) << 12)
+#define G_00B528_FLOAT_MODE(x) (((x) >> 12) & 0xFF)
+#define C_00B528_FLOAT_MODE 0xFFF00FFF
+#define S_00B528_PRIV(x) (((x) & 0x1) << 20)
+#define G_00B528_PRIV(x) (((x) >> 20) & 0x1)
+#define C_00B528_PRIV 0xFFEFFFFF
+#define S_00B528_DX10_CLAMP(x) (((x) & 0x1) << 21)
+#define G_00B528_DX10_CLAMP(x) (((x) >> 21) & 0x1)
+#define C_00B528_DX10_CLAMP 0xFFDFFFFF
+#define S_00B528_DEBUG_MODE(x) (((x) & 0x1) << 22)
+#define G_00B528_DEBUG_MODE(x) (((x) >> 22) & 0x1)
+#define C_00B528_DEBUG_MODE 0xFFBFFFFF
+#define S_00B528_IEEE_MODE(x) (((x) & 0x1) << 23)
+#define G_00B528_IEEE_MODE(x) (((x) >> 23) & 0x1)
+#define C_00B528_IEEE_MODE 0xFF7FFFFF
+#define S_00B528_VGPR_COMP_CNT(x) (((x) & 0x03) << 24)
+#define G_00B528_VGPR_COMP_CNT(x) (((x) >> 24) & 0x03)
+#define C_00B528_VGPR_COMP_CNT 0xFCFFFFFF
+#define R_00B52C_SPI_SHADER_PGM_RSRC2_LS 0x00B52C
+#define S_00B52C_SCRATCH_EN(x) (((x) & 0x1) << 0)
+#define G_00B52C_SCRATCH_EN(x) (((x) >> 0) & 0x1)
+#define C_00B52C_SCRATCH_EN 0xFFFFFFFE
+#define S_00B52C_USER_SGPR(x) (((x) & 0x1F) << 1)
+#define G_00B52C_USER_SGPR(x) (((x) >> 1) & 0x1F)
+#define C_00B52C_USER_SGPR 0xFFFFFFC1
+#define S_00B52C_LDS_SIZE(x) (((x) & 0x1FF) << 7)
+#define G_00B52C_LDS_SIZE(x) (((x) >> 7) & 0x1FF)
+#define C_00B52C_LDS_SIZE 0xFFFF007F
+#define S_00B52C_EXCP_EN(x) (((x) & 0x7F) << 16)
+#define G_00B52C_EXCP_EN(x) (((x) >> 16) & 0x7F)
+#define C_00B52C_EXCP_EN 0xFF80FFFF
+#define R_00B530_SPI_SHADER_USER_DATA_LS_0 0x00B530
+#define R_00B800_COMPUTE_DISPATCH_INITIATOR 0x00B800
+#define S_00B800_COMPUTE_SHADER_EN(x) (((x) & 0x1) << 0)
+#define G_00B800_COMPUTE_SHADER_EN(x) (((x) >> 0) & 0x1)
+#define C_00B800_COMPUTE_SHADER_EN 0xFFFFFFFE
+#define S_00B800_PARTIAL_TG_EN(x) (((x) & 0x1) << 1)
+#define G_00B800_PARTIAL_TG_EN(x) (((x) >> 1) & 0x1)
+#define C_00B800_PARTIAL_TG_EN 0xFFFFFFFD
+#define S_00B800_FORCE_START_AT_000(x) (((x) & 0x1) << 2)
+#define G_00B800_FORCE_START_AT_000(x) (((x) >> 2) & 0x1)
+#define C_00B800_FORCE_START_AT_000 0xFFFFFFFB
+#define S_00B800_ORDERED_APPEND_ENBL(x) (((x) & 0x1) << 3)
+#define G_00B800_ORDERED_APPEND_ENBL(x) (((x) >> 3) & 0x1)
+#define C_00B800_ORDERED_APPEND_ENBL 0xFFFFFFF7
+#define R_00B804_COMPUTE_DIM_X 0x00B804
+#define R_00B808_COMPUTE_DIM_Y 0x00B808
+#define R_00B80C_COMPUTE_DIM_Z 0x00B80C
+#define R_00B810_COMPUTE_START_X 0x00B810
+#define R_00B814_COMPUTE_START_Y 0x00B814
+#define R_00B818_COMPUTE_START_Z 0x00B818
+#define R_00B81C_COMPUTE_NUM_THREAD_X 0x00B81C
+#define S_00B81C_NUM_THREAD_FULL(x) (((x) & 0xFFFF) << 0)
+#define G_00B81C_NUM_THREAD_FULL(x) (((x) >> 0) & 0xFFFF)
+#define C_00B81C_NUM_THREAD_FULL 0xFFFF0000
+#define S_00B81C_NUM_THREAD_PARTIAL(x) (((x) & 0xFFFF) << 16)
+#define G_00B81C_NUM_THREAD_PARTIAL(x) (((x) >> 16) & 0xFFFF)
+#define C_00B81C_NUM_THREAD_PARTIAL 0x0000FFFF
+#define R_00B820_COMPUTE_NUM_THREAD_Y 0x00B820
+#define S_00B820_NUM_THREAD_FULL(x) (((x) & 0xFFFF) << 0)
+#define G_00B820_NUM_THREAD_FULL(x) (((x) >> 0) & 0xFFFF)
+#define C_00B820_NUM_THREAD_FULL 0xFFFF0000
+#define S_00B820_NUM_THREAD_PARTIAL(x) (((x) & 0xFFFF) << 16)
+#define G_00B820_NUM_THREAD_PARTIAL(x) (((x) >> 16) & 0xFFFF)
+#define C_00B820_NUM_THREAD_PARTIAL 0x0000FFFF
+#define R_00B824_COMPUTE_NUM_THREAD_Z 0x00B824
+#define S_00B824_NUM_THREAD_FULL(x) (((x) & 0xFFFF) << 0)
+#define G_00B824_NUM_THREAD_FULL(x) (((x) >> 0) & 0xFFFF)
+#define C_00B824_NUM_THREAD_FULL 0xFFFF0000
+#define S_00B824_NUM_THREAD_PARTIAL(x) (((x) & 0xFFFF) << 16)
+#define G_00B824_NUM_THREAD_PARTIAL(x) (((x) >> 16) & 0xFFFF)
+#define C_00B824_NUM_THREAD_PARTIAL 0x0000FFFF
+#define R_00B82C_COMPUTE_MAX_WAVE_ID 0x00B82C
+#define S_00B82C_MAX_WAVE_ID(x) (((x) & 0xFFF) << 0)
+#define G_00B82C_MAX_WAVE_ID(x) (((x) >> 0) & 0xFFF)
+#define C_00B82C_MAX_WAVE_ID 0xFFFFF000
+#define R_00B830_COMPUTE_PGM_LO 0x00B830
+#define R_00B834_COMPUTE_PGM_HI 0x00B834
+#define S_00B834_DATA(x) (((x) & 0xFF) << 0)
+#define G_00B834_DATA(x) (((x) >> 0) & 0xFF)
+#define C_00B834_DATA 0xFFFFFF00
+#define R_00B848_COMPUTE_PGM_RSRC1 0x00B848
+#define S_00B848_VGPRS(x) (((x) & 0x3F) << 0)
+#define G_00B848_VGPRS(x) (((x) >> 0) & 0x3F)
+#define C_00B848_VGPRS 0xFFFFFFC0
+#define S_00B848_SGPRS(x) (((x) & 0x0F) << 6)
+#define G_00B848_SGPRS(x) (((x) >> 6) & 0x0F)
+#define C_00B848_SGPRS 0xFFFFFC3F
+#define S_00B848_PRIORITY(x) (((x) & 0x03) << 10)
+#define G_00B848_PRIORITY(x) (((x) >> 10) & 0x03)
+#define C_00B848_PRIORITY 0xFFFFF3FF
+#define S_00B848_FLOAT_MODE(x) (((x) & 0xFF) << 12)
+#define G_00B848_FLOAT_MODE(x) (((x) >> 12) & 0xFF)
+#define C_00B848_FLOAT_MODE 0xFFF00FFF
+#define S_00B848_PRIV(x) (((x) & 0x1) << 20)
+#define G_00B848_PRIV(x) (((x) >> 20) & 0x1)
+#define C_00B848_PRIV 0xFFEFFFFF
+#define S_00B848_DX10_CLAMP(x) (((x) & 0x1) << 21)
+#define G_00B848_DX10_CLAMP(x) (((x) >> 21) & 0x1)
+#define C_00B848_DX10_CLAMP 0xFFDFFFFF
+#define S_00B848_DEBUG_MODE(x) (((x) & 0x1) << 22)
+#define G_00B848_DEBUG_MODE(x) (((x) >> 22) & 0x1)
+#define C_00B848_DEBUG_MODE 0xFFBFFFFF
+#define S_00B848_IEEE_MODE(x) (((x) & 0x1) << 23)
+#define G_00B848_IEEE_MODE(x) (((x) >> 23) & 0x1)
+#define C_00B848_IEEE_MODE 0xFF7FFFFF
+#define R_00B84C_COMPUTE_PGM_RSRC2 0x00B84C
+#define S_00B84C_SCRATCH_EN(x) (((x) & 0x1) << 0)
+#define G_00B84C_SCRATCH_EN(x) (((x) >> 0) & 0x1)
+#define C_00B84C_SCRATCH_EN 0xFFFFFFFE
+#define S_00B84C_USER_SGPR(x) (((x) & 0x1F) << 1)
+#define G_00B84C_USER_SGPR(x) (((x) >> 1) & 0x1F)
+#define C_00B84C_USER_SGPR 0xFFFFFFC1
+#define S_00B84C_TGID_X_EN(x) (((x) & 0x1) << 7)
+#define G_00B84C_TGID_X_EN(x) (((x) >> 7) & 0x1)
+#define C_00B84C_TGID_X_EN 0xFFFFFF7F
+#define S_00B84C_TGID_Y_EN(x) (((x) & 0x1) << 8)
+#define G_00B84C_TGID_Y_EN(x) (((x) >> 8) & 0x1)
+#define C_00B84C_TGID_Y_EN 0xFFFFFEFF
+#define S_00B84C_TGID_Z_EN(x) (((x) & 0x1) << 9)
+#define G_00B84C_TGID_Z_EN(x) (((x) >> 9) & 0x1)
+#define C_00B84C_TGID_Z_EN 0xFFFFFDFF
+#define S_00B84C_TG_SIZE_EN(x) (((x) & 0x1) << 10)
+#define G_00B84C_TG_SIZE_EN(x) (((x) >> 10) & 0x1)
+#define C_00B84C_TG_SIZE_EN 0xFFFFFBFF
+#define S_00B84C_TIDIG_COMP_CNT(x) (((x) & 0x03) << 11)
+#define G_00B84C_TIDIG_COMP_CNT(x) (((x) >> 11) & 0x03)
+#define C_00B84C_TIDIG_COMP_CNT 0xFFFFE7FF
+#define S_00B84C_LDS_SIZE(x) (((x) & 0x1FF) << 15)
+#define G_00B84C_LDS_SIZE(x) (((x) >> 15) & 0x1FF)
+#define C_00B84C_LDS_SIZE 0xFF007FFF
+#define S_00B84C_EXCP_EN(x) (((x) & 0x7F) << 24)
+#define G_00B84C_EXCP_EN(x) (((x) >> 24) & 0x7F)
+#define C_00B84C_EXCP_EN 0x80FFFFFF
+#define R_00B854_COMPUTE_RESOURCE_LIMITS 0x00B854
+#define S_00B854_WAVES_PER_SH(x) (((x) & 0x3F) << 0)
+#define G_00B854_WAVES_PER_SH(x) (((x) >> 0) & 0x3F)
+#define C_00B854_WAVES_PER_SH 0xFFFFFFC0
+#define S_00B854_TG_PER_CU(x) (((x) & 0x0F) << 12)
+#define G_00B854_TG_PER_CU(x) (((x) >> 12) & 0x0F)
+#define C_00B854_TG_PER_CU 0xFFFF0FFF
+#define S_00B854_LOCK_THRESHOLD(x) (((x) & 0x3F) << 16)
+#define G_00B854_LOCK_THRESHOLD(x) (((x) >> 16) & 0x3F)
+#define C_00B854_LOCK_THRESHOLD 0xFFC0FFFF
+#define S_00B854_SIMD_DEST_CNTL(x) (((x) & 0x1) << 22)
+#define G_00B854_SIMD_DEST_CNTL(x) (((x) >> 22) & 0x1)
+#define C_00B854_SIMD_DEST_CNTL 0xFFBFFFFF
+#define R_00B858_COMPUTE_STATIC_THREAD_MGMT_SE0 0x00B858
+#define S_00B858_SH0_CU_EN(x) (((x) & 0xFFFF) << 0)
+#define G_00B858_SH0_CU_EN(x) (((x) >> 0) & 0xFFFF)
+#define C_00B858_SH0_CU_EN 0xFFFF0000
+#define S_00B858_SH1_CU_EN(x) (((x) & 0xFFFF) << 16)
+#define G_00B858_SH1_CU_EN(x) (((x) >> 16) & 0xFFFF)
+#define C_00B858_SH1_CU_EN 0x0000FFFF
+#define R_00B85C_COMPUTE_STATIC_THREAD_MGMT_SE1 0x00B85C
+#define S_00B85C_SH0_CU_EN(x) (((x) & 0xFFFF) << 0)
+#define G_00B85C_SH0_CU_EN(x) (((x) >> 0) & 0xFFFF)
+#define C_00B85C_SH0_CU_EN 0xFFFF0000
+#define S_00B85C_SH1_CU_EN(x) (((x) & 0xFFFF) << 16)
+#define G_00B85C_SH1_CU_EN(x) (((x) >> 16) & 0xFFFF)
+#define C_00B85C_SH1_CU_EN 0x0000FFFF
+#define R_00B860_COMPUTE_TMPRING_SIZE 0x00B860
+#define S_00B860_WAVES(x) (((x) & 0xFFF) << 0)
+#define G_00B860_WAVES(x) (((x) >> 0) & 0xFFF)
+#define C_00B860_WAVES 0xFFFFF000
+#define S_00B860_WAVESIZE(x) (((x) & 0x1FFF) << 12)
+#define G_00B860_WAVESIZE(x) (((x) >> 12) & 0x1FFF)
+#define C_00B860_WAVESIZE 0xFE000FFF
+#define R_00B900_COMPUTE_USER_DATA_0 0x00B900
+#define R_028000_DB_RENDER_CONTROL 0x028000
+#define S_028000_DEPTH_CLEAR_ENABLE(x) (((x) & 0x1) << 0)
+#define G_028000_DEPTH_CLEAR_ENABLE(x) (((x) >> 0) & 0x1)
+#define C_028000_DEPTH_CLEAR_ENABLE 0xFFFFFFFE
+#define S_028000_STENCIL_CLEAR_ENABLE(x) (((x) & 0x1) << 1)
+#define G_028000_STENCIL_CLEAR_ENABLE(x) (((x) >> 1) & 0x1)
+#define C_028000_STENCIL_CLEAR_ENABLE 0xFFFFFFFD
+#define S_028000_DEPTH_COPY(x) (((x) & 0x1) << 2)
+#define G_028000_DEPTH_COPY(x) (((x) >> 2) & 0x1)
+#define C_028000_DEPTH_COPY 0xFFFFFFFB
+#define S_028000_STENCIL_COPY(x) (((x) & 0x1) << 3)
+#define G_028000_STENCIL_COPY(x) (((x) >> 3) & 0x1)
+#define C_028000_STENCIL_COPY 0xFFFFFFF7
+#define S_028000_RESUMMARIZE_ENABLE(x) (((x) & 0x1) << 4)
+#define G_028000_RESUMMARIZE_ENABLE(x) (((x) >> 4) & 0x1)
+#define C_028000_RESUMMARIZE_ENABLE 0xFFFFFFEF
+#define S_028000_STENCIL_COMPRESS_DISABLE(x) (((x) & 0x1) << 5)
+#define G_028000_STENCIL_COMPRESS_DISABLE(x) (((x) >> 5) & 0x1)
+#define C_028000_STENCIL_COMPRESS_DISABLE 0xFFFFFFDF
+#define S_028000_DEPTH_COMPRESS_DISABLE(x) (((x) & 0x1) << 6)
+#define G_028000_DEPTH_COMPRESS_DISABLE(x) (((x) >> 6) & 0x1)
+#define C_028000_DEPTH_COMPRESS_DISABLE 0xFFFFFFBF
+#define S_028000_COPY_CENTROID(x) (((x) & 0x1) << 7)
+#define G_028000_COPY_CENTROID(x) (((x) >> 7) & 0x1)
+#define C_028000_COPY_CENTROID 0xFFFFFF7F
+#define S_028000_COPY_SAMPLE(x) (((x) & 0x0F) << 8)
+#define G_028000_COPY_SAMPLE(x) (((x) >> 8) & 0x0F)
+#define C_028000_COPY_SAMPLE 0xFFFFF0FF
+#define R_028004_DB_COUNT_CONTROL 0x028004
+#define S_028004_ZPASS_INCREMENT_DISABLE(x) (((x) & 0x1) << 0)
+#define G_028004_ZPASS_INCREMENT_DISABLE(x) (((x) >> 0) & 0x1)
+#define C_028004_ZPASS_INCREMENT_DISABLE 0xFFFFFFFE
+#define S_028004_PERFECT_ZPASS_COUNTS(x) (((x) & 0x1) << 1)
+#define G_028004_PERFECT_ZPASS_COUNTS(x) (((x) >> 1) & 0x1)
+#define C_028004_PERFECT_ZPASS_COUNTS 0xFFFFFFFD
+#define S_028004_SAMPLE_RATE(x) (((x) & 0x07) << 4)
+#define G_028004_SAMPLE_RATE(x) (((x) >> 4) & 0x07)
+#define C_028004_SAMPLE_RATE 0xFFFFFF8F
+#define R_028008_DB_DEPTH_VIEW 0x028008
+#define S_028008_SLICE_START(x) (((x) & 0x7FF) << 0)
+#define G_028008_SLICE_START(x) (((x) >> 0) & 0x7FF)
+#define C_028008_SLICE_START 0xFFFFF800
+#define S_028008_SLICE_MAX(x) (((x) & 0x7FF) << 13)
+#define G_028008_SLICE_MAX(x) (((x) >> 13) & 0x7FF)
+#define C_028008_SLICE_MAX 0xFF001FFF
+#define S_028008_Z_READ_ONLY(x) (((x) & 0x1) << 24)
+#define G_028008_Z_READ_ONLY(x) (((x) >> 24) & 0x1)
+#define C_028008_Z_READ_ONLY 0xFEFFFFFF
+#define S_028008_STENCIL_READ_ONLY(x) (((x) & 0x1) << 25)
+#define G_028008_STENCIL_READ_ONLY(x) (((x) >> 25) & 0x1)
+#define C_028008_STENCIL_READ_ONLY 0xFDFFFFFF
+#define R_02800C_DB_RENDER_OVERRIDE 0x02800C
+#define S_02800C_FORCE_HIZ_ENABLE(x) (((x) & 0x03) << 0)
+#define G_02800C_FORCE_HIZ_ENABLE(x) (((x) >> 0) & 0x03)
+#define C_02800C_FORCE_HIZ_ENABLE 0xFFFFFFFC
+#define V_02800C_FORCE_OFF 0x00
+#define V_02800C_FORCE_ENABLE 0x01
+#define V_02800C_FORCE_DISABLE 0x02
+#define V_02800C_FORCE_RESERVED 0x03
+#define S_02800C_FORCE_HIS_ENABLE0(x) (((x) & 0x03) << 2)
+#define G_02800C_FORCE_HIS_ENABLE0(x) (((x) >> 2) & 0x03)
+#define C_02800C_FORCE_HIS_ENABLE0 0xFFFFFFF3
+#define V_02800C_FORCE_OFF 0x00
+#define V_02800C_FORCE_ENABLE 0x01
+#define V_02800C_FORCE_DISABLE 0x02
+#define V_02800C_FORCE_RESERVED 0x03
+#define S_02800C_FORCE_HIS_ENABLE1(x) (((x) & 0x03) << 4)
+#define G_02800C_FORCE_HIS_ENABLE1(x) (((x) >> 4) & 0x03)
+#define C_02800C_FORCE_HIS_ENABLE1 0xFFFFFFCF
+#define V_02800C_FORCE_OFF 0x00
+#define V_02800C_FORCE_ENABLE 0x01
+#define V_02800C_FORCE_DISABLE 0x02
+#define V_02800C_FORCE_RESERVED 0x03
+#define S_02800C_FORCE_SHADER_Z_ORDER(x) (((x) & 0x1) << 6)
+#define G_02800C_FORCE_SHADER_Z_ORDER(x) (((x) >> 6) & 0x1)
+#define C_02800C_FORCE_SHADER_Z_ORDER 0xFFFFFFBF
+#define S_02800C_FAST_Z_DISABLE(x) (((x) & 0x1) << 7)
+#define G_02800C_FAST_Z_DISABLE(x) (((x) >> 7) & 0x1)
+#define C_02800C_FAST_Z_DISABLE 0xFFFFFF7F
+#define S_02800C_FAST_STENCIL_DISABLE(x) (((x) & 0x1) << 8)
+#define G_02800C_FAST_STENCIL_DISABLE(x) (((x) >> 8) & 0x1)
+#define C_02800C_FAST_STENCIL_DISABLE 0xFFFFFEFF
+#define S_02800C_NOOP_CULL_DISABLE(x) (((x) & 0x1) << 9)
+#define G_02800C_NOOP_CULL_DISABLE(x) (((x) >> 9) & 0x1)
+#define C_02800C_NOOP_CULL_DISABLE 0xFFFFFDFF
+#define S_02800C_FORCE_COLOR_KILL(x) (((x) & 0x1) << 10)
+#define G_02800C_FORCE_COLOR_KILL(x) (((x) >> 10) & 0x1)
+#define C_02800C_FORCE_COLOR_KILL 0xFFFFFBFF
+#define S_02800C_FORCE_Z_READ(x) (((x) & 0x1) << 11)
+#define G_02800C_FORCE_Z_READ(x) (((x) >> 11) & 0x1)
+#define C_02800C_FORCE_Z_READ 0xFFFFF7FF
+#define S_02800C_FORCE_STENCIL_READ(x) (((x) & 0x1) << 12)
+#define G_02800C_FORCE_STENCIL_READ(x) (((x) >> 12) & 0x1)
+#define C_02800C_FORCE_STENCIL_READ 0xFFFFEFFF
+#define S_02800C_FORCE_FULL_Z_RANGE(x) (((x) & 0x03) << 13)
+#define G_02800C_FORCE_FULL_Z_RANGE(x) (((x) >> 13) & 0x03)
+#define C_02800C_FORCE_FULL_Z_RANGE 0xFFFF9FFF
+#define V_02800C_FORCE_OFF 0x00
+#define V_02800C_FORCE_ENABLE 0x01
+#define V_02800C_FORCE_DISABLE 0x02
+#define V_02800C_FORCE_RESERVED 0x03
+#define S_02800C_FORCE_QC_SMASK_CONFLICT(x) (((x) & 0x1) << 15)
+#define G_02800C_FORCE_QC_SMASK_CONFLICT(x) (((x) >> 15) & 0x1)
+#define C_02800C_FORCE_QC_SMASK_CONFLICT 0xFFFF7FFF
+#define S_02800C_DISABLE_VIEWPORT_CLAMP(x) (((x) & 0x1) << 16)
+#define G_02800C_DISABLE_VIEWPORT_CLAMP(x) (((x) >> 16) & 0x1)
+#define C_02800C_DISABLE_VIEWPORT_CLAMP 0xFFFEFFFF
+#define S_02800C_IGNORE_SC_ZRANGE(x) (((x) & 0x1) << 17)
+#define G_02800C_IGNORE_SC_ZRANGE(x) (((x) >> 17) & 0x1)
+#define C_02800C_IGNORE_SC_ZRANGE 0xFFFDFFFF
+#define S_02800C_DISABLE_FULLY_COVERED(x) (((x) & 0x1) << 18)
+#define G_02800C_DISABLE_FULLY_COVERED(x) (((x) >> 18) & 0x1)
+#define C_02800C_DISABLE_FULLY_COVERED 0xFFFBFFFF
+#define S_02800C_FORCE_Z_LIMIT_SUMM(x) (((x) & 0x03) << 19)
+#define G_02800C_FORCE_Z_LIMIT_SUMM(x) (((x) >> 19) & 0x03)
+#define C_02800C_FORCE_Z_LIMIT_SUMM 0xFFE7FFFF
+#define V_02800C_FORCE_SUMM_OFF 0x00
+#define V_02800C_FORCE_SUMM_MINZ 0x01
+#define V_02800C_FORCE_SUMM_MAXZ 0x02
+#define V_02800C_FORCE_SUMM_BOTH 0x03
+#define S_02800C_MAX_TILES_IN_DTT(x) (((x) & 0x1F) << 21)
+#define G_02800C_MAX_TILES_IN_DTT(x) (((x) >> 21) & 0x1F)
+#define C_02800C_MAX_TILES_IN_DTT 0xFC1FFFFF
+#define S_02800C_DISABLE_TILE_RATE_TILES(x) (((x) & 0x1) << 26)
+#define G_02800C_DISABLE_TILE_RATE_TILES(x) (((x) >> 26) & 0x1)
+#define C_02800C_DISABLE_TILE_RATE_TILES 0xFBFFFFFF
+#define S_02800C_FORCE_Z_DIRTY(x) (((x) & 0x1) << 27)
+#define G_02800C_FORCE_Z_DIRTY(x) (((x) >> 27) & 0x1)
+#define C_02800C_FORCE_Z_DIRTY 0xF7FFFFFF
+#define S_02800C_FORCE_STENCIL_DIRTY(x) (((x) & 0x1) << 28)
+#define G_02800C_FORCE_STENCIL_DIRTY(x) (((x) >> 28) & 0x1)
+#define C_02800C_FORCE_STENCIL_DIRTY 0xEFFFFFFF
+#define S_02800C_FORCE_Z_VALID(x) (((x) & 0x1) << 29)
+#define G_02800C_FORCE_Z_VALID(x) (((x) >> 29) & 0x1)
+#define C_02800C_FORCE_Z_VALID 0xDFFFFFFF
+#define S_02800C_FORCE_STENCIL_VALID(x) (((x) & 0x1) << 30)
+#define G_02800C_FORCE_STENCIL_VALID(x) (((x) >> 30) & 0x1)
+#define C_02800C_FORCE_STENCIL_VALID 0xBFFFFFFF
+#define S_02800C_PRESERVE_COMPRESSION(x) (((x) & 0x1) << 31)
+#define G_02800C_PRESERVE_COMPRESSION(x) (((x) >> 31) & 0x1)
+#define C_02800C_PRESERVE_COMPRESSION 0x7FFFFFFF
+#define R_028010_DB_RENDER_OVERRIDE2 0x028010
+#define S_028010_PARTIAL_SQUAD_LAUNCH_CONTROL(x) (((x) & 0x03) << 0)
+#define G_028010_PARTIAL_SQUAD_LAUNCH_CONTROL(x) (((x) >> 0) & 0x03)
+#define C_028010_PARTIAL_SQUAD_LAUNCH_CONTROL 0xFFFFFFFC
+#define V_028010_PSLC_AUTO 0x00
+#define V_028010_PSLC_ON_HANG_ONLY 0x01
+#define V_028010_PSLC_ASAP 0x02
+#define V_028010_PSLC_COUNTDOWN 0x03
+#define S_028010_PARTIAL_SQUAD_LAUNCH_COUNTDOWN(x) (((x) & 0x07) << 2)
+#define G_028010_PARTIAL_SQUAD_LAUNCH_COUNTDOWN(x) (((x) >> 2) & 0x07)
+#define C_028010_PARTIAL_SQUAD_LAUNCH_COUNTDOWN 0xFFFFFFE3
+#define S_028010_DISABLE_ZMASK_EXPCLEAR_OPTIMIZATIO(x) (((x) & 0x1) << 5)
+#define G_028010_DISABLE_ZMASK_EXPCLEAR_OPTIMIZATIO(x) (((x) >> 5) & 0x1)
+#define C_028010_DISABLE_ZMASK_EXPCLEAR_OPTIMIZATIO 0xFFFFFFDF
+#define S_028010_DISABLE_SMEM_EXPCLEAR_OPTIMIZATION(x) (((x) & 0x1) << 6)
+#define G_028010_DISABLE_SMEM_EXPCLEAR_OPTIMIZATION(x) (((x) >> 6) & 0x1)
+#define C_028010_DISABLE_SMEM_EXPCLEAR_OPTIMIZATION 0xFFFFFFBF
+#define S_028010_DISABLE_COLOR_ON_VALIDATION(x) (((x) & 0x1) << 7)
+#define G_028010_DISABLE_COLOR_ON_VALIDATION(x) (((x) >> 7) & 0x1)
+#define C_028010_DISABLE_COLOR_ON_VALIDATION 0xFFFFFF7F
+#define S_028010_DECOMPRESS_Z_ON_FLUSH(x) (((x) & 0x1) << 8)
+#define G_028010_DECOMPRESS_Z_ON_FLUSH(x) (((x) >> 8) & 0x1)
+#define C_028010_DECOMPRESS_Z_ON_FLUSH 0xFFFFFEFF
+#define S_028010_DISABLE_REG_SNOOP(x) (((x) & 0x1) << 9)
+#define G_028010_DISABLE_REG_SNOOP(x) (((x) >> 9) & 0x1)
+#define C_028010_DISABLE_REG_SNOOP 0xFFFFFDFF
+#define S_028010_DEPTH_BOUNDS_HIER_DEPTH_DISABLE(x) (((x) & 0x1) << 10)
+#define G_028010_DEPTH_BOUNDS_HIER_DEPTH_DISABLE(x) (((x) >> 10) & 0x1)
+#define C_028010_DEPTH_BOUNDS_HIER_DEPTH_DISABLE 0xFFFFFBFF
+#define R_028014_DB_HTILE_DATA_BASE 0x028014
+#define R_028020_DB_DEPTH_BOUNDS_MIN 0x028020
+#define R_028024_DB_DEPTH_BOUNDS_MAX 0x028024
+#define R_028028_DB_STENCIL_CLEAR 0x028028
+#define S_028028_CLEAR(x) (((x) & 0xFF) << 0)
+#define G_028028_CLEAR(x) (((x) >> 0) & 0xFF)
+#define C_028028_CLEAR 0xFFFFFF00
+#define R_02802C_DB_DEPTH_CLEAR 0x02802C
+#define R_028030_PA_SC_SCREEN_SCISSOR_TL 0x028030
+#define S_028030_TL_X(x) (((x) & 0xFFFF) << 0)
+#define G_028030_TL_X(x) (((x) >> 0) & 0xFFFF)
+#define C_028030_TL_X 0xFFFF0000
+#define S_028030_TL_Y(x) (((x) & 0xFFFF) << 16)
+#define G_028030_TL_Y(x) (((x) >> 16) & 0xFFFF)
+#define C_028030_TL_Y 0x0000FFFF
+#define R_028034_PA_SC_SCREEN_SCISSOR_BR 0x028034
+#define S_028034_BR_X(x) (((x) & 0xFFFF) << 0)
+#define G_028034_BR_X(x) (((x) >> 0) & 0xFFFF)
+#define C_028034_BR_X 0xFFFF0000
+#define S_028034_BR_Y(x) (((x) & 0xFFFF) << 16)
+#define G_028034_BR_Y(x) (((x) >> 16) & 0xFFFF)
+#define C_028034_BR_Y 0x0000FFFF
+#define R_02803C_DB_DEPTH_INFO 0x02803C
+#define S_02803C_ADDR5_SWIZZLE_MASK(x) (((x) & 0x0F) << 0)
+#define G_02803C_ADDR5_SWIZZLE_MASK(x) (((x) >> 0) & 0x0F)
+#define C_02803C_ADDR5_SWIZZLE_MASK 0xFFFFFFF0
+#define R_028040_DB_Z_INFO 0x028040
+#define S_028040_FORMAT(x) (((x) & 0x03) << 0)
+#define G_028040_FORMAT(x) (((x) >> 0) & 0x03)
+#define C_028040_FORMAT 0xFFFFFFFC
+#define V_028040_Z_INVALID 0x00
+#define V_028040_Z_16 0x01
+#define V_028040_Z_24 0x02 /* deprecated */
+#define V_028040_Z_32_FLOAT 0x03
+#define S_028040_NUM_SAMPLES(x) (((x) & 0x03) << 2)
+#define G_028040_NUM_SAMPLES(x) (((x) >> 2) & 0x03)
+#define C_028040_NUM_SAMPLES 0xFFFFFFF3
+#define S_028040_TILE_MODE_INDEX(x) (((x) & 0x07) << 20)
+#define G_028040_TILE_MODE_INDEX(x) (((x) >> 20) & 0x07)
+#define C_028040_TILE_MODE_INDEX 0xFF8FFFFF
+#define S_028040_ALLOW_EXPCLEAR(x) (((x) & 0x1) << 27)
+#define G_028040_ALLOW_EXPCLEAR(x) (((x) >> 27) & 0x1)
+#define C_028040_ALLOW_EXPCLEAR 0xF7FFFFFF
+#define S_028040_READ_SIZE(x) (((x) & 0x1) << 28)
+#define G_028040_READ_SIZE(x) (((x) >> 28) & 0x1)
+#define C_028040_READ_SIZE 0xEFFFFFFF
+#define S_028040_TILE_SURFACE_ENABLE(x) (((x) & 0x1) << 29)
+#define G_028040_TILE_SURFACE_ENABLE(x) (((x) >> 29) & 0x1)
+#define C_028040_TILE_SURFACE_ENABLE 0xDFFFFFFF
+#define S_028040_ZRANGE_PRECISION(x) (((x) & 0x1) << 31)
+#define G_028040_ZRANGE_PRECISION(x) (((x) >> 31) & 0x1)
+#define C_028040_ZRANGE_PRECISION 0x7FFFFFFF
+#define R_028044_DB_STENCIL_INFO 0x028044
+#define S_028044_FORMAT(x) (((x) & 0x1) << 0)
+#define G_028044_FORMAT(x) (((x) >> 0) & 0x1)
+#define C_028044_FORMAT 0xFFFFFFFE
+#define S_028044_TILE_MODE_INDEX(x) (((x) & 0x07) << 20)
+#define G_028044_TILE_MODE_INDEX(x) (((x) >> 20) & 0x07)
+#define C_028044_TILE_MODE_INDEX 0xFF8FFFFF
+#define S_028044_ALLOW_EXPCLEAR(x) (((x) & 0x1) << 27)
+#define G_028044_ALLOW_EXPCLEAR(x) (((x) >> 27) & 0x1)
+#define C_028044_ALLOW_EXPCLEAR 0xF7FFFFFF
+#define S_028044_TILE_STENCIL_DISABLE(x) (((x) & 0x1) << 29)
+#define G_028044_TILE_STENCIL_DISABLE(x) (((x) >> 29) & 0x1)
+#define C_028044_TILE_STENCIL_DISABLE 0xDFFFFFFF
+#define R_028048_DB_Z_READ_BASE 0x028048
+#define R_02804C_DB_STENCIL_READ_BASE 0x02804C
+#define R_028050_DB_Z_WRITE_BASE 0x028050
+#define R_028054_DB_STENCIL_WRITE_BASE 0x028054
+#define R_028058_DB_DEPTH_SIZE 0x028058
+#define S_028058_PITCH_TILE_MAX(x) (((x) & 0x7FF) << 0)
+#define G_028058_PITCH_TILE_MAX(x) (((x) >> 0) & 0x7FF)
+#define C_028058_PITCH_TILE_MAX 0xFFFFF800
+#define S_028058_HEIGHT_TILE_MAX(x) (((x) & 0x7FF) << 11)
+#define G_028058_HEIGHT_TILE_MAX(x) (((x) >> 11) & 0x7FF)
+#define C_028058_HEIGHT_TILE_MAX 0xFFC007FF
+#define R_02805C_DB_DEPTH_SLICE 0x02805C
+#define S_02805C_SLICE_TILE_MAX(x) (((x) & 0x3FFFFF) << 0)
+#define G_02805C_SLICE_TILE_MAX(x) (((x) >> 0) & 0x3FFFFF)
+#define C_02805C_SLICE_TILE_MAX 0xFFC00000
+#define R_028080_TA_BC_BASE_ADDR 0x028080
+#define R_028200_PA_SC_WINDOW_OFFSET 0x028200
+#define S_028200_WINDOW_X_OFFSET(x) (((x) & 0xFFFF) << 0)
+#define G_028200_WINDOW_X_OFFSET(x) (((x) >> 0) & 0xFFFF)
+#define C_028200_WINDOW_X_OFFSET 0xFFFF0000
+#define S_028200_WINDOW_Y_OFFSET(x) (((x) & 0xFFFF) << 16)
+#define G_028200_WINDOW_Y_OFFSET(x) (((x) >> 16) & 0xFFFF)
+#define C_028200_WINDOW_Y_OFFSET 0x0000FFFF
+#define R_028204_PA_SC_WINDOW_SCISSOR_TL 0x028204
+#define S_028204_TL_X(x) (((x) & 0x7FFF) << 0)
+#define G_028204_TL_X(x) (((x) >> 0) & 0x7FFF)
+#define C_028204_TL_X 0xFFFF8000
+#define S_028204_TL_Y(x) (((x) & 0x7FFF) << 16)
+#define G_028204_TL_Y(x) (((x) >> 16) & 0x7FFF)
+#define C_028204_TL_Y 0x8000FFFF
+#define S_028204_WINDOW_OFFSET_DISABLE(x) (((x) & 0x1) << 31)
+#define G_028204_WINDOW_OFFSET_DISABLE(x) (((x) >> 31) & 0x1)
+#define C_028204_WINDOW_OFFSET_DISABLE 0x7FFFFFFF
+#define R_028208_PA_SC_WINDOW_SCISSOR_BR 0x028208
+#define S_028208_BR_X(x) (((x) & 0x7FFF) << 0)
+#define G_028208_BR_X(x) (((x) >> 0) & 0x7FFF)
+#define C_028208_BR_X 0xFFFF8000
+#define S_028208_BR_Y(x) (((x) & 0x7FFF) << 16)
+#define G_028208_BR_Y(x) (((x) >> 16) & 0x7FFF)
+#define C_028208_BR_Y 0x8000FFFF
+#define R_02820C_PA_SC_CLIPRECT_RULE 0x02820C
+#define S_02820C_CLIP_RULE(x) (((x) & 0xFFFF) << 0)
+#define G_02820C_CLIP_RULE(x) (((x) >> 0) & 0xFFFF)
+#define C_02820C_CLIP_RULE 0xFFFF0000
+#define R_028210_PA_SC_CLIPRECT_0_TL 0x028210
+#define S_028210_TL_X(x) (((x) & 0x7FFF) << 0)
+#define G_028210_TL_X(x) (((x) >> 0) & 0x7FFF)
+#define C_028210_TL_X 0xFFFF8000
+#define S_028210_TL_Y(x) (((x) & 0x7FFF) << 16)
+#define G_028210_TL_Y(x) (((x) >> 16) & 0x7FFF)
+#define C_028210_TL_Y 0x8000FFFF
+#define R_028214_PA_SC_CLIPRECT_0_BR 0x028214
+#define S_028214_BR_X(x) (((x) & 0x7FFF) << 0)
+#define G_028214_BR_X(x) (((x) >> 0) & 0x7FFF)
+#define C_028214_BR_X 0xFFFF8000
+#define S_028214_BR_Y(x) (((x) & 0x7FFF) << 16)
+#define G_028214_BR_Y(x) (((x) >> 16) & 0x7FFF)
+#define C_028214_BR_Y 0x8000FFFF
+#define R_028218_PA_SC_CLIPRECT_1_TL 0x028218
+#define R_02821C_PA_SC_CLIPRECT_1_BR 0x02821C
+#define R_028220_PA_SC_CLIPRECT_2_TL 0x028220
+#define R_028224_PA_SC_CLIPRECT_2_BR 0x028224
+#define R_028228_PA_SC_CLIPRECT_3_TL 0x028228
+#define R_02822C_PA_SC_CLIPRECT_3_BR 0x02822C
+#define R_028230_PA_SC_EDGERULE 0x028230
+#define S_028230_ER_TRI(x) (((x) & 0x0F) << 0)
+#define G_028230_ER_TRI(x) (((x) >> 0) & 0x0F)
+#define C_028230_ER_TRI 0xFFFFFFF0
+#define S_028230_ER_POINT(x) (((x) & 0x0F) << 4)
+#define G_028230_ER_POINT(x) (((x) >> 4) & 0x0F)
+#define C_028230_ER_POINT 0xFFFFFF0F
+#define S_028230_ER_RECT(x) (((x) & 0x0F) << 8)
+#define G_028230_ER_RECT(x) (((x) >> 8) & 0x0F)
+#define C_028230_ER_RECT 0xFFFFF0FF
+#define S_028230_ER_LINE_LR(x) (((x) & 0x3F) << 12)
+#define G_028230_ER_LINE_LR(x) (((x) >> 12) & 0x3F)
+#define C_028230_ER_LINE_LR 0xFFFC0FFF
+#define S_028230_ER_LINE_RL(x) (((x) & 0x3F) << 18)
+#define G_028230_ER_LINE_RL(x) (((x) >> 18) & 0x3F)
+#define C_028230_ER_LINE_RL 0xFF03FFFF
+#define S_028230_ER_LINE_TB(x) (((x) & 0x0F) << 24)
+#define G_028230_ER_LINE_TB(x) (((x) >> 24) & 0x0F)
+#define C_028230_ER_LINE_TB 0xF0FFFFFF
+#define S_028230_ER_LINE_BT(x) (((x) & 0x0F) << 28)
+#define G_028230_ER_LINE_BT(x) (((x) >> 28) & 0x0F)
+#define C_028230_ER_LINE_BT 0x0FFFFFFF
+#define R_028234_PA_SU_HARDWARE_SCREEN_OFFSET 0x028234
+#define S_028234_HW_SCREEN_OFFSET_X(x) (((x) & 0x1FF) << 0)
+#define G_028234_HW_SCREEN_OFFSET_X(x) (((x) >> 0) & 0x1FF)
+#define C_028234_HW_SCREEN_OFFSET_X 0xFFFFFE00
+#define S_028234_HW_SCREEN_OFFSET_Y(x) (((x) & 0x1FF) << 16)
+#define G_028234_HW_SCREEN_OFFSET_Y(x) (((x) >> 16) & 0x1FF)
+#define C_028234_HW_SCREEN_OFFSET_Y 0xFE00FFFF
+#define R_028238_CB_TARGET_MASK 0x028238
+#define S_028238_TARGET0_ENABLE(x) (((x) & 0x0F) << 0)
+#define G_028238_TARGET0_ENABLE(x) (((x) >> 0) & 0x0F)
+#define C_028238_TARGET0_ENABLE 0xFFFFFFF0
+#define S_028238_TARGET1_ENABLE(x) (((x) & 0x0F) << 4)
+#define G_028238_TARGET1_ENABLE(x) (((x) >> 4) & 0x0F)
+#define C_028238_TARGET1_ENABLE 0xFFFFFF0F
+#define S_028238_TARGET2_ENABLE(x) (((x) & 0x0F) << 8)
+#define G_028238_TARGET2_ENABLE(x) (((x) >> 8) & 0x0F)
+#define C_028238_TARGET2_ENABLE 0xFFFFF0FF
+#define S_028238_TARGET3_ENABLE(x) (((x) & 0x0F) << 12)
+#define G_028238_TARGET3_ENABLE(x) (((x) >> 12) & 0x0F)
+#define C_028238_TARGET3_ENABLE 0xFFFF0FFF
+#define S_028238_TARGET4_ENABLE(x) (((x) & 0x0F) << 16)
+#define G_028238_TARGET4_ENABLE(x) (((x) >> 16) & 0x0F)
+#define C_028238_TARGET4_ENABLE 0xFFF0FFFF
+#define S_028238_TARGET5_ENABLE(x) (((x) & 0x0F) << 20)
+#define G_028238_TARGET5_ENABLE(x) (((x) >> 20) & 0x0F)
+#define C_028238_TARGET5_ENABLE 0xFF0FFFFF
+#define S_028238_TARGET6_ENABLE(x) (((x) & 0x0F) << 24)
+#define G_028238_TARGET6_ENABLE(x) (((x) >> 24) & 0x0F)
+#define C_028238_TARGET6_ENABLE 0xF0FFFFFF
+#define S_028238_TARGET7_ENABLE(x) (((x) & 0x0F) << 28)
+#define G_028238_TARGET7_ENABLE(x) (((x) >> 28) & 0x0F)
+#define C_028238_TARGET7_ENABLE 0x0FFFFFFF
+#define R_02823C_CB_SHADER_MASK 0x02823C
+#define S_02823C_OUTPUT0_ENABLE(x) (((x) & 0x0F) << 0)
+#define G_02823C_OUTPUT0_ENABLE(x) (((x) >> 0) & 0x0F)
+#define C_02823C_OUTPUT0_ENABLE 0xFFFFFFF0
+#define S_02823C_OUTPUT1_ENABLE(x) (((x) & 0x0F) << 4)
+#define G_02823C_OUTPUT1_ENABLE(x) (((x) >> 4) & 0x0F)
+#define C_02823C_OUTPUT1_ENABLE 0xFFFFFF0F
+#define S_02823C_OUTPUT2_ENABLE(x) (((x) & 0x0F) << 8)
+#define G_02823C_OUTPUT2_ENABLE(x) (((x) >> 8) & 0x0F)
+#define C_02823C_OUTPUT2_ENABLE 0xFFFFF0FF
+#define S_02823C_OUTPUT3_ENABLE(x) (((x) & 0x0F) << 12)
+#define G_02823C_OUTPUT3_ENABLE(x) (((x) >> 12) & 0x0F)
+#define C_02823C_OUTPUT3_ENABLE 0xFFFF0FFF
+#define S_02823C_OUTPUT4_ENABLE(x) (((x) & 0x0F) << 16)
+#define G_02823C_OUTPUT4_ENABLE(x) (((x) >> 16) & 0x0F)
+#define C_02823C_OUTPUT4_ENABLE 0xFFF0FFFF
+#define S_02823C_OUTPUT5_ENABLE(x) (((x) & 0x0F) << 20)
+#define G_02823C_OUTPUT5_ENABLE(x) (((x) >> 20) & 0x0F)
+#define C_02823C_OUTPUT5_ENABLE 0xFF0FFFFF
+#define S_02823C_OUTPUT6_ENABLE(x) (((x) & 0x0F) << 24)
+#define G_02823C_OUTPUT6_ENABLE(x) (((x) >> 24) & 0x0F)
+#define C_02823C_OUTPUT6_ENABLE 0xF0FFFFFF
+#define S_02823C_OUTPUT7_ENABLE(x) (((x) & 0x0F) << 28)
+#define G_02823C_OUTPUT7_ENABLE(x) (((x) >> 28) & 0x0F)
+#define C_02823C_OUTPUT7_ENABLE 0x0FFFFFFF
+#define R_028240_PA_SC_GENERIC_SCISSOR_TL 0x028240
+#define S_028240_TL_X(x) (((x) & 0x7FFF) << 0)
+#define G_028240_TL_X(x) (((x) >> 0) & 0x7FFF)
+#define C_028240_TL_X 0xFFFF8000
+#define S_028240_TL_Y(x) (((x) & 0x7FFF) << 16)
+#define G_028240_TL_Y(x) (((x) >> 16) & 0x7FFF)
+#define C_028240_TL_Y 0x8000FFFF
+#define S_028240_WINDOW_OFFSET_DISABLE(x) (((x) & 0x1) << 31)
+#define G_028240_WINDOW_OFFSET_DISABLE(x) (((x) >> 31) & 0x1)
+#define C_028240_WINDOW_OFFSET_DISABLE 0x7FFFFFFF
+#define R_028244_PA_SC_GENERIC_SCISSOR_BR 0x028244
+#define S_028244_BR_X(x) (((x) & 0x7FFF) << 0)
+#define G_028244_BR_X(x) (((x) >> 0) & 0x7FFF)
+#define C_028244_BR_X 0xFFFF8000
+#define S_028244_BR_Y(x) (((x) & 0x7FFF) << 16)
+#define G_028244_BR_Y(x) (((x) >> 16) & 0x7FFF)
+#define C_028244_BR_Y 0x8000FFFF
+#define R_028250_PA_SC_VPORT_SCISSOR_0_TL 0x028250
+#define S_028250_TL_X(x) (((x) & 0x7FFF) << 0)
+#define G_028250_TL_X(x) (((x) >> 0) & 0x7FFF)
+#define C_028250_TL_X 0xFFFF8000
+#define S_028250_TL_Y(x) (((x) & 0x7FFF) << 16)
+#define G_028250_TL_Y(x) (((x) >> 16) & 0x7FFF)
+#define C_028250_TL_Y 0x8000FFFF
+#define S_028250_WINDOW_OFFSET_DISABLE(x) (((x) & 0x1) << 31)
+#define G_028250_WINDOW_OFFSET_DISABLE(x) (((x) >> 31) & 0x1)
+#define C_028250_WINDOW_OFFSET_DISABLE 0x7FFFFFFF
+#define R_028254_PA_SC_VPORT_SCISSOR_0_BR 0x028254
+#define S_028254_BR_X(x) (((x) & 0x7FFF) << 0)
+#define G_028254_BR_X(x) (((x) >> 0) & 0x7FFF)
+#define C_028254_BR_X 0xFFFF8000
+#define S_028254_BR_Y(x) (((x) & 0x7FFF) << 16)
+#define G_028254_BR_Y(x) (((x) >> 16) & 0x7FFF)
+#define C_028254_BR_Y 0x8000FFFF
+#define R_0282D0_PA_SC_VPORT_ZMIN_0 0x0282D0
+#define R_0282D4_PA_SC_VPORT_ZMAX_0 0x0282D4
+#define R_028350_PA_SC_RASTER_CONFIG 0x028350
+#define S_028350_RB_MAP_PKR0(x) (((x) & 0x03) << 0)
+#define G_028350_RB_MAP_PKR0(x) (((x) >> 0) & 0x03)
+#define C_028350_RB_MAP_PKR0 0xFFFFFFFC
+#define V_028350_RASTER_CONFIG_RB_MAP_0 0x00
+#define V_028350_RASTER_CONFIG_RB_MAP_1 0x01
+#define V_028350_RASTER_CONFIG_RB_MAP_2 0x02
+#define V_028350_RASTER_CONFIG_RB_MAP_3 0x03
+#define S_028350_RB_MAP_PKR1(x) (((x) & 0x03) << 2)
+#define G_028350_RB_MAP_PKR1(x) (((x) >> 2) & 0x03)
+#define C_028350_RB_MAP_PKR1 0xFFFFFFF3
+#define V_028350_RASTER_CONFIG_RB_MAP_0 0x00
+#define V_028350_RASTER_CONFIG_RB_MAP_1 0x01
+#define V_028350_RASTER_CONFIG_RB_MAP_2 0x02
+#define V_028350_RASTER_CONFIG_RB_MAP_3 0x03
+#define S_028350_RB_XSEL2(x) (((x) & 0x03) << 4)
+#define G_028350_RB_XSEL2(x) (((x) >> 4) & 0x03)
+#define C_028350_RB_XSEL2 0xFFFFFFCF
+#define V_028350_RASTER_CONFIG_RB_XSEL2_0 0x00
+#define V_028350_RASTER_CONFIG_RB_XSEL2_1 0x01
+#define V_028350_RASTER_CONFIG_RB_XSEL2_2 0x02
+#define V_028350_RASTER_CONFIG_RB_XSEL2_3 0x03
+#define S_028350_RB_XSEL(x) (((x) & 0x1) << 6)
+#define G_028350_RB_XSEL(x) (((x) >> 6) & 0x1)
+#define C_028350_RB_XSEL 0xFFFFFFBF
+#define S_028350_RB_YSEL(x) (((x) & 0x1) << 7)
+#define G_028350_RB_YSEL(x) (((x) >> 7) & 0x1)
+#define C_028350_RB_YSEL 0xFFFFFF7F
+#define S_028350_PKR_MAP(x) (((x) & 0x03) << 8)
+#define G_028350_PKR_MAP(x) (((x) >> 8) & 0x03)
+#define C_028350_PKR_MAP 0xFFFFFCFF
+#define V_028350_RASTER_CONFIG_PKR_MAP_0 0x00
+#define V_028350_RASTER_CONFIG_PKR_MAP_1 0x01
+#define V_028350_RASTER_CONFIG_PKR_MAP_2 0x02
+#define V_028350_RASTER_CONFIG_PKR_MAP_3 0x03
+#define S_028350_PKR_XSEL(x) (((x) & 0x03) << 10)
+#define G_028350_PKR_XSEL(x) (((x) >> 10) & 0x03)
+#define C_028350_PKR_XSEL 0xFFFFF3FF
+#define V_028350_RASTER_CONFIG_PKR_XSEL_0 0x00
+#define V_028350_RASTER_CONFIG_PKR_XSEL_1 0x01
+#define V_028350_RASTER_CONFIG_PKR_XSEL_2 0x02
+#define V_028350_RASTER_CONFIG_PKR_XSEL_3 0x03
+#define S_028350_PKR_YSEL(x) (((x) & 0x03) << 12)
+#define G_028350_PKR_YSEL(x) (((x) >> 12) & 0x03)
+#define C_028350_PKR_YSEL 0xFFFFCFFF
+#define V_028350_RASTER_CONFIG_PKR_YSEL_0 0x00
+#define V_028350_RASTER_CONFIG_PKR_YSEL_1 0x01
+#define V_028350_RASTER_CONFIG_PKR_YSEL_2 0x02
+#define V_028350_RASTER_CONFIG_PKR_YSEL_3 0x03
+#define S_028350_SC_MAP(x) (((x) & 0x03) << 16)
+#define G_028350_SC_MAP(x) (((x) >> 16) & 0x03)
+#define C_028350_SC_MAP 0xFFFCFFFF
+#define V_028350_RASTER_CONFIG_SC_MAP_0 0x00
+#define V_028350_RASTER_CONFIG_SC_MAP_1 0x01
+#define V_028350_RASTER_CONFIG_SC_MAP_2 0x02
+#define V_028350_RASTER_CONFIG_SC_MAP_3 0x03
+#define S_028350_SC_XSEL(x) (((x) & 0x03) << 18)
+#define G_028350_SC_XSEL(x) (((x) >> 18) & 0x03)
+#define C_028350_SC_XSEL 0xFFF3FFFF
+#define V_028350_RASTER_CONFIG_SC_XSEL_8_WIDE_TILE 0x00
+#define V_028350_RASTER_CONFIG_SC_XSEL_16_WIDE_TILE 0x01
+#define V_028350_RASTER_CONFIG_SC_XSEL_32_WIDE_TILE 0x02
+#define V_028350_RASTER_CONFIG_SC_XSEL_64_WIDE_TILE 0x03
+#define S_028350_SC_YSEL(x) (((x) & 0x03) << 20)
+#define G_028350_SC_YSEL(x) (((x) >> 20) & 0x03)
+#define C_028350_SC_YSEL 0xFFCFFFFF
+#define V_028350_RASTER_CONFIG_SC_YSEL_8_WIDE_TILE 0x00
+#define V_028350_RASTER_CONFIG_SC_YSEL_16_WIDE_TILE 0x01
+#define V_028350_RASTER_CONFIG_SC_YSEL_32_WIDE_TILE 0x02
+#define V_028350_RASTER_CONFIG_SC_YSEL_64_WIDE_TILE 0x03
+#define S_028350_SE_MAP(x) (((x) & 0x03) << 24)
+#define G_028350_SE_MAP(x) (((x) >> 24) & 0x03)
+#define C_028350_SE_MAP 0xFCFFFFFF
+#define V_028350_RASTER_CONFIG_SE_MAP_0 0x00
+#define V_028350_RASTER_CONFIG_SE_MAP_1 0x01
+#define V_028350_RASTER_CONFIG_SE_MAP_2 0x02
+#define V_028350_RASTER_CONFIG_SE_MAP_3 0x03
+#define S_028350_SE_XSEL(x) (((x) & 0x03) << 26)
+#define G_028350_SE_XSEL(x) (((x) >> 26) & 0x03)
+#define C_028350_SE_XSEL 0xF3FFFFFF
+#define V_028350_RASTER_CONFIG_SE_XSEL_8_WIDE_TILE 0x00
+#define V_028350_RASTER_CONFIG_SE_XSEL_16_WIDE_TILE 0x01
+#define V_028350_RASTER_CONFIG_SE_XSEL_32_WIDE_TILE 0x02
+#define V_028350_RASTER_CONFIG_SE_XSEL_64_WIDE_TILE 0x03
+#define S_028350_SE_YSEL(x) (((x) & 0x03) << 28)
+#define G_028350_SE_YSEL(x) (((x) >> 28) & 0x03)
+#define C_028350_SE_YSEL 0xCFFFFFFF
+#define V_028350_RASTER_CONFIG_SE_YSEL_8_WIDE_TILE 0x00
+#define V_028350_RASTER_CONFIG_SE_YSEL_16_WIDE_TILE 0x01
+#define V_028350_RASTER_CONFIG_SE_YSEL_32_WIDE_TILE 0x02
+#define V_028350_RASTER_CONFIG_SE_YSEL_64_WIDE_TILE 0x03
+#define R_028400_VGT_MAX_VTX_INDX 0x028400
+#define R_028404_VGT_MIN_VTX_INDX 0x028404
+#define R_028408_VGT_INDX_OFFSET 0x028408
+#define R_02840C_VGT_MULTI_PRIM_IB_RESET_INDX 0x02840C
+#define R_028414_CB_BLEND_RED 0x028414
+#define R_028418_CB_BLEND_GREEN 0x028418
+#define R_02841C_CB_BLEND_BLUE 0x02841C
+#define R_028420_CB_BLEND_ALPHA 0x028420
+#define R_02842C_DB_STENCIL_CONTROL 0x02842C
+#define S_02842C_STENCILFAIL(x) (((x) & 0x0F) << 0)
+#define G_02842C_STENCILFAIL(x) (((x) >> 0) & 0x0F)
+#define C_02842C_STENCILFAIL 0xFFFFFFF0
+#define V_02842C_STENCIL_KEEP 0x00
+#define V_02842C_STENCIL_ZERO 0x01
+#define V_02842C_STENCIL_ONES 0x02
+#define V_02842C_STENCIL_REPLACE_TEST 0x03
+#define V_02842C_STENCIL_REPLACE_OP 0x04
+#define V_02842C_STENCIL_ADD_CLAMP 0x05
+#define V_02842C_STENCIL_SUB_CLAMP 0x06
+#define V_02842C_STENCIL_INVERT 0x07
+#define V_02842C_STENCIL_ADD_WRAP 0x08
+#define V_02842C_STENCIL_SUB_WRAP 0x09
+#define V_02842C_STENCIL_AND 0x0A
+#define V_02842C_STENCIL_OR 0x0B
+#define V_02842C_STENCIL_XOR 0x0C
+#define V_02842C_STENCIL_NAND 0x0D
+#define V_02842C_STENCIL_NOR 0x0E
+#define V_02842C_STENCIL_XNOR 0x0F
+#define S_02842C_STENCILZPASS(x) (((x) & 0x0F) << 4)
+#define G_02842C_STENCILZPASS(x) (((x) >> 4) & 0x0F)
+#define C_02842C_STENCILZPASS 0xFFFFFF0F
+#define V_02842C_STENCIL_KEEP 0x00
+#define V_02842C_STENCIL_ZERO 0x01
+#define V_02842C_STENCIL_ONES 0x02
+#define V_02842C_STENCIL_REPLACE_TEST 0x03
+#define V_02842C_STENCIL_REPLACE_OP 0x04
+#define V_02842C_STENCIL_ADD_CLAMP 0x05
+#define V_02842C_STENCIL_SUB_CLAMP 0x06
+#define V_02842C_STENCIL_INVERT 0x07
+#define V_02842C_STENCIL_ADD_WRAP 0x08
+#define V_02842C_STENCIL_SUB_WRAP 0x09
+#define V_02842C_STENCIL_AND 0x0A
+#define V_02842C_STENCIL_OR 0x0B
+#define V_02842C_STENCIL_XOR 0x0C
+#define V_02842C_STENCIL_NAND 0x0D
+#define V_02842C_STENCIL_NOR 0x0E
+#define V_02842C_STENCIL_XNOR 0x0F
+#define S_02842C_STENCILZFAIL(x) (((x) & 0x0F) << 8)
+#define G_02842C_STENCILZFAIL(x) (((x) >> 8) & 0x0F)
+#define C_02842C_STENCILZFAIL 0xFFFFF0FF
+#define V_02842C_STENCIL_KEEP 0x00
+#define V_02842C_STENCIL_ZERO 0x01
+#define V_02842C_STENCIL_ONES 0x02
+#define V_02842C_STENCIL_REPLACE_TEST 0x03
+#define V_02842C_STENCIL_REPLACE_OP 0x04
+#define V_02842C_STENCIL_ADD_CLAMP 0x05
+#define V_02842C_STENCIL_SUB_CLAMP 0x06
+#define V_02842C_STENCIL_INVERT 0x07
+#define V_02842C_STENCIL_ADD_WRAP 0x08
+#define V_02842C_STENCIL_SUB_WRAP 0x09
+#define V_02842C_STENCIL_AND 0x0A
+#define V_02842C_STENCIL_OR 0x0B
+#define V_02842C_STENCIL_XOR 0x0C
+#define V_02842C_STENCIL_NAND 0x0D
+#define V_02842C_STENCIL_NOR 0x0E
+#define V_02842C_STENCIL_XNOR 0x0F
+#define S_02842C_STENCILFAIL_BF(x) (((x) & 0x0F) << 12)
+#define G_02842C_STENCILFAIL_BF(x) (((x) >> 12) & 0x0F)
+#define C_02842C_STENCILFAIL_BF 0xFFFF0FFF
+#define V_02842C_STENCIL_KEEP 0x00
+#define V_02842C_STENCIL_ZERO 0x01
+#define V_02842C_STENCIL_ONES 0x02
+#define V_02842C_STENCIL_REPLACE_TEST 0x03
+#define V_02842C_STENCIL_REPLACE_OP 0x04
+#define V_02842C_STENCIL_ADD_CLAMP 0x05
+#define V_02842C_STENCIL_SUB_CLAMP 0x06
+#define V_02842C_STENCIL_INVERT 0x07
+#define V_02842C_STENCIL_ADD_WRAP 0x08
+#define V_02842C_STENCIL_SUB_WRAP 0x09
+#define V_02842C_STENCIL_AND 0x0A
+#define V_02842C_STENCIL_OR 0x0B
+#define V_02842C_STENCIL_XOR 0x0C
+#define V_02842C_STENCIL_NAND 0x0D
+#define V_02842C_STENCIL_NOR 0x0E
+#define V_02842C_STENCIL_XNOR 0x0F
+#define S_02842C_STENCILZPASS_BF(x) (((x) & 0x0F) << 16)
+#define G_02842C_STENCILZPASS_BF(x) (((x) >> 16) & 0x0F)
+#define C_02842C_STENCILZPASS_BF 0xFFF0FFFF
+#define V_02842C_STENCIL_KEEP 0x00
+#define V_02842C_STENCIL_ZERO 0x01
+#define V_02842C_STENCIL_ONES 0x02
+#define V_02842C_STENCIL_REPLACE_TEST 0x03
+#define V_02842C_STENCIL_REPLACE_OP 0x04
+#define V_02842C_STENCIL_ADD_CLAMP 0x05
+#define V_02842C_STENCIL_SUB_CLAMP 0x06
+#define V_02842C_STENCIL_INVERT 0x07
+#define V_02842C_STENCIL_ADD_WRAP 0x08
+#define V_02842C_STENCIL_SUB_WRAP 0x09
+#define V_02842C_STENCIL_AND 0x0A
+#define V_02842C_STENCIL_OR 0x0B
+#define V_02842C_STENCIL_XOR 0x0C
+#define V_02842C_STENCIL_NAND 0x0D
+#define V_02842C_STENCIL_NOR 0x0E
+#define V_02842C_STENCIL_XNOR 0x0F
+#define S_02842C_STENCILZFAIL_BF(x) (((x) & 0x0F) << 20)
+#define G_02842C_STENCILZFAIL_BF(x) (((x) >> 20) & 0x0F)
+#define C_02842C_STENCILZFAIL_BF 0xFF0FFFFF
+#define V_02842C_STENCIL_KEEP 0x00
+#define V_02842C_STENCIL_ZERO 0x01
+#define V_02842C_STENCIL_ONES 0x02
+#define V_02842C_STENCIL_REPLACE_TEST 0x03
+#define V_02842C_STENCIL_REPLACE_OP 0x04
+#define V_02842C_STENCIL_ADD_CLAMP 0x05
+#define V_02842C_STENCIL_SUB_CLAMP 0x06
+#define V_02842C_STENCIL_INVERT 0x07
+#define V_02842C_STENCIL_ADD_WRAP 0x08
+#define V_02842C_STENCIL_SUB_WRAP 0x09
+#define V_02842C_STENCIL_AND 0x0A
+#define V_02842C_STENCIL_OR 0x0B
+#define V_02842C_STENCIL_XOR 0x0C
+#define V_02842C_STENCIL_NAND 0x0D
+#define V_02842C_STENCIL_NOR 0x0E
+#define V_02842C_STENCIL_XNOR 0x0F
+#define R_028430_DB_STENCILREFMASK 0x028430
+#define S_028430_STENCILTESTVAL(x) (((x) & 0xFF) << 0)
+#define G_028430_STENCILTESTVAL(x) (((x) >> 0) & 0xFF)
+#define C_028430_STENCILTESTVAL 0xFFFFFF00
+#define S_028430_STENCILMASK(x) (((x) & 0xFF) << 8)
+#define G_028430_STENCILMASK(x) (((x) >> 8) & 0xFF)
+#define C_028430_STENCILMASK 0xFFFF00FF
+#define S_028430_STENCILWRITEMASK(x) (((x) & 0xFF) << 16)
+#define G_028430_STENCILWRITEMASK(x) (((x) >> 16) & 0xFF)
+#define C_028430_STENCILWRITEMASK 0xFF00FFFF
+#define S_028430_STENCILOPVAL(x) (((x) & 0xFF) << 24)
+#define G_028430_STENCILOPVAL(x) (((x) >> 24) & 0xFF)
+#define C_028430_STENCILOPVAL 0x00FFFFFF
+#define R_028434_DB_STENCILREFMASK_BF 0x028434
+#define S_028434_STENCILTESTVAL_BF(x) (((x) & 0xFF) << 0)
+#define G_028434_STENCILTESTVAL_BF(x) (((x) >> 0) & 0xFF)
+#define C_028434_STENCILTESTVAL_BF 0xFFFFFF00
+#define S_028434_STENCILMASK_BF(x) (((x) & 0xFF) << 8)
+#define G_028434_STENCILMASK_BF(x) (((x) >> 8) & 0xFF)
+#define C_028434_STENCILMASK_BF 0xFFFF00FF
+#define S_028434_STENCILWRITEMASK_BF(x) (((x) & 0xFF) << 16)
+#define G_028434_STENCILWRITEMASK_BF(x) (((x) >> 16) & 0xFF)
+#define C_028434_STENCILWRITEMASK_BF 0xFF00FFFF
+#define S_028434_STENCILOPVAL_BF(x) (((x) & 0xFF) << 24)
+#define G_028434_STENCILOPVAL_BF(x) (((x) >> 24) & 0xFF)
+#define C_028434_STENCILOPVAL_BF 0x00FFFFFF
+#define R_02843C_PA_CL_VPORT_XSCALE_0 0x02843C
+#define R_028440_PA_CL_VPORT_XOFFSET_0 0x028440
+#define R_028444_PA_CL_VPORT_YSCALE_0 0x028444
+#define R_028448_PA_CL_VPORT_YOFFSET_0 0x028448
+#define R_02844C_PA_CL_VPORT_ZSCALE_0 0x02844C
+#define R_028450_PA_CL_VPORT_ZOFFSET_0 0x028450
+#define R_0285BC_PA_CL_UCP_0_X 0x0285BC
+#define R_0285C0_PA_CL_UCP_0_Y 0x0285C0
+#define R_0285C4_PA_CL_UCP_0_Z 0x0285C4
+#define R_0285C8_PA_CL_UCP_0_W 0x0285C8
+#define R_0285CC_PA_CL_UCP_1_X 0x0285CC
+#define R_0285D0_PA_CL_UCP_1_Y 0x0285D0
+#define R_0285D4_PA_CL_UCP_1_Z 0x0285D4
+#define R_0285D8_PA_CL_UCP_1_W 0x0285D8
+#define R_0285DC_PA_CL_UCP_2_X 0x0285DC
+#define R_0285E0_PA_CL_UCP_2_Y 0x0285E0
+#define R_0285E4_PA_CL_UCP_2_Z 0x0285E4
+#define R_0285E8_PA_CL_UCP_2_W 0x0285E8
+#define R_0285EC_PA_CL_UCP_3_X 0x0285EC
+#define R_0285F0_PA_CL_UCP_3_Y 0x0285F0
+#define R_0285F4_PA_CL_UCP_3_Z 0x0285F4
+#define R_0285F8_PA_CL_UCP_3_W 0x0285F8
+#define R_0285FC_PA_CL_UCP_4_X 0x0285FC
+#define R_028600_PA_CL_UCP_4_Y 0x028600
+#define R_028604_PA_CL_UCP_4_Z 0x028604
+#define R_028608_PA_CL_UCP_4_W 0x028608
+#define R_02860C_PA_CL_UCP_5_X 0x02860C
+#define R_028610_PA_CL_UCP_5_Y 0x028610
+#define R_028614_PA_CL_UCP_5_Z 0x028614
+#define R_028618_PA_CL_UCP_5_W 0x028618
+#define R_028644_SPI_PS_INPUT_CNTL_0 0x028644
+#define S_028644_OFFSET(x) (((x) & 0x3F) << 0)
+#define G_028644_OFFSET(x) (((x) >> 0) & 0x3F)
+#define C_028644_OFFSET 0xFFFFFFC0
+#define S_028644_DEFAULT_VAL(x) (((x) & 0x03) << 8)
+#define G_028644_DEFAULT_VAL(x) (((x) >> 8) & 0x03)
+#define C_028644_DEFAULT_VAL 0xFFFFFCFF
+#define V_028644_X_0_0F 0x00
+#define S_028644_FLAT_SHADE(x) (((x) & 0x1) << 10)
+#define G_028644_FLAT_SHADE(x) (((x) >> 10) & 0x1)
+#define C_028644_FLAT_SHADE 0xFFFFFBFF
+#define S_028644_CYL_WRAP(x) (((x) & 0x0F) << 13)
+#define G_028644_CYL_WRAP(x) (((x) >> 13) & 0x0F)
+#define C_028644_CYL_WRAP 0xFFFE1FFF
+#define S_028644_PT_SPRITE_TEX(x) (((x) & 0x1) << 17)
+#define G_028644_PT_SPRITE_TEX(x) (((x) >> 17) & 0x1)
+#define C_028644_PT_SPRITE_TEX 0xFFFDFFFF
+#define R_028648_SPI_PS_INPUT_CNTL_1 0x028648
+#define R_02864C_SPI_PS_INPUT_CNTL_2 0x02864C
+#define R_028650_SPI_PS_INPUT_CNTL_3 0x028650
+#define R_028654_SPI_PS_INPUT_CNTL_4 0x028654
+#define R_028658_SPI_PS_INPUT_CNTL_5 0x028658
+#define R_02865C_SPI_PS_INPUT_CNTL_6 0x02865C
+#define R_028660_SPI_PS_INPUT_CNTL_7 0x028660
+#define R_028664_SPI_PS_INPUT_CNTL_8 0x028664
+#define R_028668_SPI_PS_INPUT_CNTL_9 0x028668
+#define R_02866C_SPI_PS_INPUT_CNTL_10 0x02866C
+#define R_028670_SPI_PS_INPUT_CNTL_11 0x028670
+#define R_028674_SPI_PS_INPUT_CNTL_12 0x028674
+#define R_028678_SPI_PS_INPUT_CNTL_13 0x028678
+#define R_02867C_SPI_PS_INPUT_CNTL_14 0x02867C
+#define R_028680_SPI_PS_INPUT_CNTL_15 0x028680
+#define R_028684_SPI_PS_INPUT_CNTL_16 0x028684
+#define R_028688_SPI_PS_INPUT_CNTL_17 0x028688
+#define R_02868C_SPI_PS_INPUT_CNTL_18 0x02868C
+#define R_028690_SPI_PS_INPUT_CNTL_19 0x028690
+#define R_028694_SPI_PS_INPUT_CNTL_20 0x028694
+#define R_028698_SPI_PS_INPUT_CNTL_21 0x028698
+#define R_02869C_SPI_PS_INPUT_CNTL_22 0x02869C
+#define R_0286A0_SPI_PS_INPUT_CNTL_23 0x0286A0
+#define R_0286A4_SPI_PS_INPUT_CNTL_24 0x0286A4
+#define R_0286A8_SPI_PS_INPUT_CNTL_25 0x0286A8
+#define R_0286AC_SPI_PS_INPUT_CNTL_26 0x0286AC
+#define R_0286B0_SPI_PS_INPUT_CNTL_27 0x0286B0
+#define R_0286B4_SPI_PS_INPUT_CNTL_28 0x0286B4
+#define R_0286B8_SPI_PS_INPUT_CNTL_29 0x0286B8
+#define R_0286BC_SPI_PS_INPUT_CNTL_30 0x0286BC
+#define R_0286C0_SPI_PS_INPUT_CNTL_31 0x0286C0
+#define R_0286C4_SPI_VS_OUT_CONFIG 0x0286C4
+#define S_0286C4_VS_EXPORT_COUNT(x) (((x) & 0x1F) << 1)
+#define G_0286C4_VS_EXPORT_COUNT(x) (((x) >> 1) & 0x1F)
+#define C_0286C4_VS_EXPORT_COUNT 0xFFFFFFC1
+#define S_0286C4_VS_HALF_PACK(x) (((x) & 0x1) << 6)
+#define G_0286C4_VS_HALF_PACK(x) (((x) >> 6) & 0x1)
+#define C_0286C4_VS_HALF_PACK 0xFFFFFFBF
+#define S_0286C4_VS_EXPORTS_FOG(x) (((x) & 0x1) << 7)
+#define G_0286C4_VS_EXPORTS_FOG(x) (((x) >> 7) & 0x1)
+#define C_0286C4_VS_EXPORTS_FOG 0xFFFFFF7F
+#define S_0286C4_VS_OUT_FOG_VEC_ADDR(x) (((x) & 0x1F) << 8)
+#define G_0286C4_VS_OUT_FOG_VEC_ADDR(x) (((x) >> 8) & 0x1F)
+#define C_0286C4_VS_OUT_FOG_VEC_ADDR 0xFFFFE0FF
+#define R_0286CC_SPI_PS_INPUT_ENA 0x0286CC
+#define S_0286CC_PERSP_SAMPLE_ENA(x) (((x) & 0x1) << 0)
+#define G_0286CC_PERSP_SAMPLE_ENA(x) (((x) >> 0) & 0x1)
+#define C_0286CC_PERSP_SAMPLE_ENA 0xFFFFFFFE
+#define S_0286CC_PERSP_CENTER_ENA(x) (((x) & 0x1) << 1)
+#define G_0286CC_PERSP_CENTER_ENA(x) (((x) >> 1) & 0x1)
+#define C_0286CC_PERSP_CENTER_ENA 0xFFFFFFFD
+#define S_0286CC_PERSP_CENTROID_ENA(x) (((x) & 0x1) << 2)
+#define G_0286CC_PERSP_CENTROID_ENA(x) (((x) >> 2) & 0x1)
+#define C_0286CC_PERSP_CENTROID_ENA 0xFFFFFFFB
+#define S_0286CC_PERSP_PULL_MODEL_ENA(x) (((x) & 0x1) << 3)
+#define G_0286CC_PERSP_PULL_MODEL_ENA(x) (((x) >> 3) & 0x1)
+#define C_0286CC_PERSP_PULL_MODEL_ENA 0xFFFFFFF7
+#define S_0286CC_LINEAR_SAMPLE_ENA(x) (((x) & 0x1) << 4)
+#define G_0286CC_LINEAR_SAMPLE_ENA(x) (((x) >> 4) & 0x1)
+#define C_0286CC_LINEAR_SAMPLE_ENA 0xFFFFFFEF
+#define S_0286CC_LINEAR_CENTER_ENA(x) (((x) & 0x1) << 5)
+#define G_0286CC_LINEAR_CENTER_ENA(x) (((x) >> 5) & 0x1)
+#define C_0286CC_LINEAR_CENTER_ENA 0xFFFFFFDF
+#define S_0286CC_LINEAR_CENTROID_ENA(x) (((x) & 0x1) << 6)
+#define G_0286CC_LINEAR_CENTROID_ENA(x) (((x) >> 6) & 0x1)
+#define C_0286CC_LINEAR_CENTROID_ENA 0xFFFFFFBF
+#define S_0286CC_LINE_STIPPLE_TEX_ENA(x) (((x) & 0x1) << 7)
+#define G_0286CC_LINE_STIPPLE_TEX_ENA(x) (((x) >> 7) & 0x1)
+#define C_0286CC_LINE_STIPPLE_TEX_ENA 0xFFFFFF7F
+#define S_0286CC_POS_X_FLOAT_ENA(x) (((x) & 0x1) << 8)
+#define G_0286CC_POS_X_FLOAT_ENA(x) (((x) >> 8) & 0x1)
+#define C_0286CC_POS_X_FLOAT_ENA 0xFFFFFEFF
+#define S_0286CC_POS_Y_FLOAT_ENA(x) (((x) & 0x1) << 9)
+#define G_0286CC_POS_Y_FLOAT_ENA(x) (((x) >> 9) & 0x1)
+#define C_0286CC_POS_Y_FLOAT_ENA 0xFFFFFDFF
+#define S_0286CC_POS_Z_FLOAT_ENA(x) (((x) & 0x1) << 10)
+#define G_0286CC_POS_Z_FLOAT_ENA(x) (((x) >> 10) & 0x1)
+#define C_0286CC_POS_Z_FLOAT_ENA 0xFFFFFBFF
+#define S_0286CC_POS_W_FLOAT_ENA(x) (((x) & 0x1) << 11)
+#define G_0286CC_POS_W_FLOAT_ENA(x) (((x) >> 11) & 0x1)
+#define C_0286CC_POS_W_FLOAT_ENA 0xFFFFF7FF
+#define S_0286CC_FRONT_FACE_ENA(x) (((x) & 0x1) << 12)
+#define G_0286CC_FRONT_FACE_ENA(x) (((x) >> 12) & 0x1)
+#define C_0286CC_FRONT_FACE_ENA 0xFFFFEFFF
+#define S_0286CC_ANCILLARY_ENA(x) (((x) & 0x1) << 13)
+#define G_0286CC_ANCILLARY_ENA(x) (((x) >> 13) & 0x1)
+#define C_0286CC_ANCILLARY_ENA 0xFFFFDFFF
+#define S_0286CC_SAMPLE_COVERAGE_ENA(x) (((x) & 0x1) << 14)
+#define G_0286CC_SAMPLE_COVERAGE_ENA(x) (((x) >> 14) & 0x1)
+#define C_0286CC_SAMPLE_COVERAGE_ENA 0xFFFFBFFF
+#define S_0286CC_POS_FIXED_PT_ENA(x) (((x) & 0x1) << 15)
+#define G_0286CC_POS_FIXED_PT_ENA(x) (((x) >> 15) & 0x1)
+#define C_0286CC_POS_FIXED_PT_ENA 0xFFFF7FFF
+#define R_0286D0_SPI_PS_INPUT_ADDR 0x0286D0
+#define S_0286D0_PERSP_SAMPLE_ENA(x) (((x) & 0x1) << 0)
+#define G_0286D0_PERSP_SAMPLE_ENA(x) (((x) >> 0) & 0x1)
+#define C_0286D0_PERSP_SAMPLE_ENA 0xFFFFFFFE
+#define S_0286D0_PERSP_CENTER_ENA(x) (((x) & 0x1) << 1)
+#define G_0286D0_PERSP_CENTER_ENA(x) (((x) >> 1) & 0x1)
+#define C_0286D0_PERSP_CENTER_ENA 0xFFFFFFFD
+#define S_0286D0_PERSP_CENTROID_ENA(x) (((x) & 0x1) << 2)
+#define G_0286D0_PERSP_CENTROID_ENA(x) (((x) >> 2) & 0x1)
+#define C_0286D0_PERSP_CENTROID_ENA 0xFFFFFFFB
+#define S_0286D0_PERSP_PULL_MODEL_ENA(x) (((x) & 0x1) << 3)
+#define G_0286D0_PERSP_PULL_MODEL_ENA(x) (((x) >> 3) & 0x1)
+#define C_0286D0_PERSP_PULL_MODEL_ENA 0xFFFFFFF7
+#define S_0286D0_LINEAR_SAMPLE_ENA(x) (((x) & 0x1) << 4)
+#define G_0286D0_LINEAR_SAMPLE_ENA(x) (((x) >> 4) & 0x1)
+#define C_0286D0_LINEAR_SAMPLE_ENA 0xFFFFFFEF
+#define S_0286D0_LINEAR_CENTER_ENA(x) (((x) & 0x1) << 5)
+#define G_0286D0_LINEAR_CENTER_ENA(x) (((x) >> 5) & 0x1)
+#define C_0286D0_LINEAR_CENTER_ENA 0xFFFFFFDF
+#define S_0286D0_LINEAR_CENTROID_ENA(x) (((x) & 0x1) << 6)
+#define G_0286D0_LINEAR_CENTROID_ENA(x) (((x) >> 6) & 0x1)
+#define C_0286D0_LINEAR_CENTROID_ENA 0xFFFFFFBF
+#define S_0286D0_LINE_STIPPLE_TEX_ENA(x) (((x) & 0x1) << 7)
+#define G_0286D0_LINE_STIPPLE_TEX_ENA(x) (((x) >> 7) & 0x1)
+#define C_0286D0_LINE_STIPPLE_TEX_ENA 0xFFFFFF7F
+#define S_0286D0_POS_X_FLOAT_ENA(x) (((x) & 0x1) << 8)
+#define G_0286D0_POS_X_FLOAT_ENA(x) (((x) >> 8) & 0x1)
+#define C_0286D0_POS_X_FLOAT_ENA 0xFFFFFEFF
+#define S_0286D0_POS_Y_FLOAT_ENA(x) (((x) & 0x1) << 9)
+#define G_0286D0_POS_Y_FLOAT_ENA(x) (((x) >> 9) & 0x1)
+#define C_0286D0_POS_Y_FLOAT_ENA 0xFFFFFDFF
+#define S_0286D0_POS_Z_FLOAT_ENA(x) (((x) & 0x1) << 10)
+#define G_0286D0_POS_Z_FLOAT_ENA(x) (((x) >> 10) & 0x1)
+#define C_0286D0_POS_Z_FLOAT_ENA 0xFFFFFBFF
+#define S_0286D0_POS_W_FLOAT_ENA(x) (((x) & 0x1) << 11)
+#define G_0286D0_POS_W_FLOAT_ENA(x) (((x) >> 11) & 0x1)
+#define C_0286D0_POS_W_FLOAT_ENA 0xFFFFF7FF
+#define S_0286D0_FRONT_FACE_ENA(x) (((x) & 0x1) << 12)
+#define G_0286D0_FRONT_FACE_ENA(x) (((x) >> 12) & 0x1)
+#define C_0286D0_FRONT_FACE_ENA 0xFFFFEFFF
+#define S_0286D0_ANCILLARY_ENA(x) (((x) & 0x1) << 13)
+#define G_0286D0_ANCILLARY_ENA(x) (((x) >> 13) & 0x1)
+#define C_0286D0_ANCILLARY_ENA 0xFFFFDFFF
+#define S_0286D0_SAMPLE_COVERAGE_ENA(x) (((x) & 0x1) << 14)
+#define G_0286D0_SAMPLE_COVERAGE_ENA(x) (((x) >> 14) & 0x1)
+#define C_0286D0_SAMPLE_COVERAGE_ENA 0xFFFFBFFF
+#define S_0286D0_POS_FIXED_PT_ENA(x) (((x) & 0x1) << 15)
+#define G_0286D0_POS_FIXED_PT_ENA(x) (((x) >> 15) & 0x1)
+#define C_0286D0_POS_FIXED_PT_ENA 0xFFFF7FFF
+#define R_0286D4_SPI_INTERP_CONTROL_0 0x0286D4
+#define S_0286D4_FLAT_SHADE_ENA(x) (((x) & 0x1) << 0)
+#define G_0286D4_FLAT_SHADE_ENA(x) (((x) >> 0) & 0x1)
+#define C_0286D4_FLAT_SHADE_ENA 0xFFFFFFFE
+#define S_0286D4_PNT_SPRITE_ENA(x) (((x) & 0x1) << 1)
+#define G_0286D4_PNT_SPRITE_ENA(x) (((x) >> 1) & 0x1)
+#define C_0286D4_PNT_SPRITE_ENA 0xFFFFFFFD
+#define S_0286D4_PNT_SPRITE_OVRD_X(x) (((x) & 0x07) << 2)
+#define G_0286D4_PNT_SPRITE_OVRD_X(x) (((x) >> 2) & 0x07)
+#define C_0286D4_PNT_SPRITE_OVRD_X 0xFFFFFFE3
+#define V_0286D4_SPI_PNT_SPRITE_SEL_0 0x00
+#define V_0286D4_SPI_PNT_SPRITE_SEL_1 0x01
+#define V_0286D4_SPI_PNT_SPRITE_SEL_S 0x02
+#define V_0286D4_SPI_PNT_SPRITE_SEL_T 0x03
+#define V_0286D4_SPI_PNT_SPRITE_SEL_NONE 0x04
+#define S_0286D4_PNT_SPRITE_OVRD_Y(x) (((x) & 0x07) << 5)
+#define G_0286D4_PNT_SPRITE_OVRD_Y(x) (((x) >> 5) & 0x07)
+#define C_0286D4_PNT_SPRITE_OVRD_Y 0xFFFFFF1F
+#define V_0286D4_SPI_PNT_SPRITE_SEL_0 0x00
+#define V_0286D4_SPI_PNT_SPRITE_SEL_1 0x01
+#define V_0286D4_SPI_PNT_SPRITE_SEL_S 0x02
+#define V_0286D4_SPI_PNT_SPRITE_SEL_T 0x03
+#define V_0286D4_SPI_PNT_SPRITE_SEL_NONE 0x04
+#define S_0286D4_PNT_SPRITE_OVRD_Z(x) (((x) & 0x07) << 8)
+#define G_0286D4_PNT_SPRITE_OVRD_Z(x) (((x) >> 8) & 0x07)
+#define C_0286D4_PNT_SPRITE_OVRD_Z 0xFFFFF8FF
+#define V_0286D4_SPI_PNT_SPRITE_SEL_0 0x00
+#define V_0286D4_SPI_PNT_SPRITE_SEL_1 0x01
+#define V_0286D4_SPI_PNT_SPRITE_SEL_S 0x02
+#define V_0286D4_SPI_PNT_SPRITE_SEL_T 0x03
+#define V_0286D4_SPI_PNT_SPRITE_SEL_NONE 0x04
+#define S_0286D4_PNT_SPRITE_OVRD_W(x) (((x) & 0x07) << 11)
+#define G_0286D4_PNT_SPRITE_OVRD_W(x) (((x) >> 11) & 0x07)
+#define C_0286D4_PNT_SPRITE_OVRD_W 0xFFFFC7FF
+#define V_0286D4_SPI_PNT_SPRITE_SEL_0 0x00
+#define V_0286D4_SPI_PNT_SPRITE_SEL_1 0x01
+#define V_0286D4_SPI_PNT_SPRITE_SEL_S 0x02
+#define V_0286D4_SPI_PNT_SPRITE_SEL_T 0x03
+#define V_0286D4_SPI_PNT_SPRITE_SEL_NONE 0x04
+#define S_0286D4_PNT_SPRITE_TOP_1(x) (((x) & 0x1) << 14)
+#define G_0286D4_PNT_SPRITE_TOP_1(x) (((x) >> 14) & 0x1)
+#define C_0286D4_PNT_SPRITE_TOP_1 0xFFFFBFFF
+#define R_0286D8_SPI_PS_IN_CONTROL 0x0286D8
+#define S_0286D8_NUM_INTERP(x) (((x) & 0x3F) << 0)
+#define G_0286D8_NUM_INTERP(x) (((x) >> 0) & 0x3F)
+#define C_0286D8_NUM_INTERP 0xFFFFFFC0
+#define S_0286D8_PARAM_GEN(x) (((x) & 0x1) << 6)
+#define G_0286D8_PARAM_GEN(x) (((x) >> 6) & 0x1)
+#define C_0286D8_PARAM_GEN 0xFFFFFFBF
+#define S_0286D8_FOG_ADDR(x) (((x) & 0x7F) << 7)
+#define G_0286D8_FOG_ADDR(x) (((x) >> 7) & 0x7F)
+#define C_0286D8_FOG_ADDR 0xFFFFC07F
+#define S_0286D8_BC_OPTIMIZE_DISABLE(x) (((x) & 0x1) << 14)
+#define G_0286D8_BC_OPTIMIZE_DISABLE(x) (((x) >> 14) & 0x1)
+#define C_0286D8_BC_OPTIMIZE_DISABLE 0xFFFFBFFF
+#define S_0286D8_PASS_FOG_THROUGH_PS(x) (((x) & 0x1) << 15)
+#define G_0286D8_PASS_FOG_THROUGH_PS(x) (((x) >> 15) & 0x1)
+#define C_0286D8_PASS_FOG_THROUGH_PS 0xFFFF7FFF
+#define R_0286E0_SPI_BARYC_CNTL 0x0286E0
+#define S_0286E0_PERSP_CENTER_CNTL(x) (((x) & 0x1) << 0)
+#define G_0286E0_PERSP_CENTER_CNTL(x) (((x) >> 0) & 0x1)
+#define C_0286E0_PERSP_CENTER_CNTL 0xFFFFFFFE
+#define S_0286E0_PERSP_CENTROID_CNTL(x) (((x) & 0x1) << 4)
+#define G_0286E0_PERSP_CENTROID_CNTL(x) (((x) >> 4) & 0x1)
+#define C_0286E0_PERSP_CENTROID_CNTL 0xFFFFFFEF
+#define S_0286E0_LINEAR_CENTER_CNTL(x) (((x) & 0x1) << 8)
+#define G_0286E0_LINEAR_CENTER_CNTL(x) (((x) >> 8) & 0x1)
+#define C_0286E0_LINEAR_CENTER_CNTL 0xFFFFFEFF
+#define S_0286E0_LINEAR_CENTROID_CNTL(x) (((x) & 0x1) << 12)
+#define G_0286E0_LINEAR_CENTROID_CNTL(x) (((x) >> 12) & 0x1)
+#define C_0286E0_LINEAR_CENTROID_CNTL 0xFFFFEFFF
+#define S_0286E0_POS_FLOAT_LOCATION(x) (((x) & 0x03) << 16)
+#define G_0286E0_POS_FLOAT_LOCATION(x) (((x) >> 16) & 0x03)
+#define C_0286E0_POS_FLOAT_LOCATION 0xFFFCFFFF
+#define V_0286E0_X_CALCULATE_PER_PIXEL_FLOATING_POINT_POSITION_AT 0x00
+#define S_0286E0_POS_FLOAT_ULC(x) (((x) & 0x1) << 20)
+#define G_0286E0_POS_FLOAT_ULC(x) (((x) >> 20) & 0x1)
+#define C_0286E0_POS_FLOAT_ULC 0xFFEFFFFF
+#define S_0286E0_FRONT_FACE_ALL_BITS(x) (((x) & 0x1) << 24)
+#define G_0286E0_FRONT_FACE_ALL_BITS(x) (((x) >> 24) & 0x1)
+#define C_0286E0_FRONT_FACE_ALL_BITS 0xFEFFFFFF
+#define R_0286E8_SPI_TMPRING_SIZE 0x0286E8
+#define S_0286E8_WAVES(x) (((x) & 0xFFF) << 0)
+#define G_0286E8_WAVES(x) (((x) >> 0) & 0xFFF)
+#define C_0286E8_WAVES 0xFFFFF000
+#define S_0286E8_WAVESIZE(x) (((x) & 0x1FFF) << 12)
+#define G_0286E8_WAVESIZE(x) (((x) >> 12) & 0x1FFF)
+#define C_0286E8_WAVESIZE 0xFE000FFF
+#define R_028704_SPI_WAVE_MGMT_1 0x028704
+#define S_028704_NUM_PS_WAVES(x) (((x) & 0x3F) << 0)
+#define G_028704_NUM_PS_WAVES(x) (((x) >> 0) & 0x3F)
+#define C_028704_NUM_PS_WAVES 0xFFFFFFC0
+#define S_028704_NUM_VS_WAVES(x) (((x) & 0x3F) << 6)
+#define G_028704_NUM_VS_WAVES(x) (((x) >> 6) & 0x3F)
+#define C_028704_NUM_VS_WAVES 0xFFFFF03F
+#define S_028704_NUM_GS_WAVES(x) (((x) & 0x3F) << 12)
+#define G_028704_NUM_GS_WAVES(x) (((x) >> 12) & 0x3F)
+#define C_028704_NUM_GS_WAVES 0xFFFC0FFF
+#define S_028704_NUM_ES_WAVES(x) (((x) & 0x3F) << 18)
+#define G_028704_NUM_ES_WAVES(x) (((x) >> 18) & 0x3F)
+#define C_028704_NUM_ES_WAVES 0xFF03FFFF
+#define S_028704_NUM_HS_WAVES(x) (((x) & 0x3F) << 24)
+#define G_028704_NUM_HS_WAVES(x) (((x) >> 24) & 0x3F)
+#define C_028704_NUM_HS_WAVES 0xC0FFFFFF
+#define R_028708_SPI_WAVE_MGMT_2 0x028708
+#define S_028708_NUM_LS_WAVES(x) (((x) & 0x3F) << 0)
+#define G_028708_NUM_LS_WAVES(x) (((x) >> 0) & 0x3F)
+#define C_028708_NUM_LS_WAVES 0xFFFFFFC0
+#define R_02870C_SPI_SHADER_POS_FORMAT 0x02870C
+#define S_02870C_POS0_EXPORT_FORMAT(x) (((x) & 0x0F) << 0)
+#define G_02870C_POS0_EXPORT_FORMAT(x) (((x) >> 0) & 0x0F)
+#define C_02870C_POS0_EXPORT_FORMAT 0xFFFFFFF0
+#define V_02870C_SPI_SHADER_NONE 0x00
+#define V_02870C_SPI_SHADER_1COMP 0x01
+#define V_02870C_SPI_SHADER_2COMP 0x02
+#define V_02870C_SPI_SHADER_4COMPRESS 0x03
+#define V_02870C_SPI_SHADER_4COMP 0x04
+#define S_02870C_POS1_EXPORT_FORMAT(x) (((x) & 0x0F) << 4)
+#define G_02870C_POS1_EXPORT_FORMAT(x) (((x) >> 4) & 0x0F)
+#define C_02870C_POS1_EXPORT_FORMAT 0xFFFFFF0F
+#define V_02870C_SPI_SHADER_NONE 0x00
+#define V_02870C_SPI_SHADER_1COMP 0x01
+#define V_02870C_SPI_SHADER_2COMP 0x02
+#define V_02870C_SPI_SHADER_4COMPRESS 0x03
+#define V_02870C_SPI_SHADER_4COMP 0x04
+#define S_02870C_POS2_EXPORT_FORMAT(x) (((x) & 0x0F) << 8)
+#define G_02870C_POS2_EXPORT_FORMAT(x) (((x) >> 8) & 0x0F)
+#define C_02870C_POS2_EXPORT_FORMAT 0xFFFFF0FF
+#define V_02870C_SPI_SHADER_NONE 0x00
+#define V_02870C_SPI_SHADER_1COMP 0x01
+#define V_02870C_SPI_SHADER_2COMP 0x02
+#define V_02870C_SPI_SHADER_4COMPRESS 0x03
+#define V_02870C_SPI_SHADER_4COMP 0x04
+#define S_02870C_POS3_EXPORT_FORMAT(x) (((x) & 0x0F) << 12)
+#define G_02870C_POS3_EXPORT_FORMAT(x) (((x) >> 12) & 0x0F)
+#define C_02870C_POS3_EXPORT_FORMAT 0xFFFF0FFF
+#define V_02870C_SPI_SHADER_NONE 0x00
+#define V_02870C_SPI_SHADER_1COMP 0x01
+#define V_02870C_SPI_SHADER_2COMP 0x02
+#define V_02870C_SPI_SHADER_4COMPRESS 0x03
+#define V_02870C_SPI_SHADER_4COMP 0x04
+#define R_028710_SPI_SHADER_Z_FORMAT 0x028710
+#define S_028710_Z_EXPORT_FORMAT(x) (((x) & 0x0F) << 0)
+#define G_028710_Z_EXPORT_FORMAT(x) (((x) >> 0) & 0x0F)
+#define C_028710_Z_EXPORT_FORMAT 0xFFFFFFF0
+#define V_028710_SPI_SHADER_ZERO 0x00
+#define V_028710_SPI_SHADER_32_R 0x01
+#define V_028710_SPI_SHADER_32_GR 0x02
+#define V_028710_SPI_SHADER_32_AR 0x03
+#define V_028710_SPI_SHADER_FP16_ABGR 0x04
+#define V_028710_SPI_SHADER_UNORM16_ABGR 0x05
+#define V_028710_SPI_SHADER_SNORM16_ABGR 0x06
+#define V_028710_SPI_SHADER_UINT16_ABGR 0x07
+#define V_028710_SPI_SHADER_SINT16_ABGR 0x08
+#define V_028710_SPI_SHADER_32_ABGR 0x09
+#define R_028714_SPI_SHADER_COL_FORMAT 0x028714
+#define S_028714_COL0_EXPORT_FORMAT(x) (((x) & 0x0F) << 0)
+#define G_028714_COL0_EXPORT_FORMAT(x) (((x) >> 0) & 0x0F)
+#define C_028714_COL0_EXPORT_FORMAT 0xFFFFFFF0
+#define V_028714_SPI_SHADER_ZERO 0x00
+#define V_028714_SPI_SHADER_32_R 0x01
+#define V_028714_SPI_SHADER_32_GR 0x02
+#define V_028714_SPI_SHADER_32_AR 0x03
+#define V_028714_SPI_SHADER_FP16_ABGR 0x04
+#define V_028714_SPI_SHADER_UNORM16_ABGR 0x05
+#define V_028714_SPI_SHADER_SNORM16_ABGR 0x06
+#define V_028714_SPI_SHADER_UINT16_ABGR 0x07
+#define V_028714_SPI_SHADER_SINT16_ABGR 0x08
+#define V_028714_SPI_SHADER_32_ABGR 0x09
+#define S_028714_COL1_EXPORT_FORMAT(x) (((x) & 0x0F) << 4)
+#define G_028714_COL1_EXPORT_FORMAT(x) (((x) >> 4) & 0x0F)
+#define C_028714_COL1_EXPORT_FORMAT 0xFFFFFF0F
+#define V_028714_SPI_SHADER_ZERO 0x00
+#define V_028714_SPI_SHADER_32_R 0x01
+#define V_028714_SPI_SHADER_32_GR 0x02
+#define V_028714_SPI_SHADER_32_AR 0x03
+#define V_028714_SPI_SHADER_FP16_ABGR 0x04
+#define V_028714_SPI_SHADER_UNORM16_ABGR 0x05
+#define V_028714_SPI_SHADER_SNORM16_ABGR 0x06
+#define V_028714_SPI_SHADER_UINT16_ABGR 0x07
+#define V_028714_SPI_SHADER_SINT16_ABGR 0x08
+#define V_028714_SPI_SHADER_32_ABGR 0x09
+#define S_028714_COL2_EXPORT_FORMAT(x) (((x) & 0x0F) << 8)
+#define G_028714_COL2_EXPORT_FORMAT(x) (((x) >> 8) & 0x0F)
+#define C_028714_COL2_EXPORT_FORMAT 0xFFFFF0FF
+#define V_028714_SPI_SHADER_ZERO 0x00
+#define V_028714_SPI_SHADER_32_R 0x01
+#define V_028714_SPI_SHADER_32_GR 0x02
+#define V_028714_SPI_SHADER_32_AR 0x03
+#define V_028714_SPI_SHADER_FP16_ABGR 0x04
+#define V_028714_SPI_SHADER_UNORM16_ABGR 0x05
+#define V_028714_SPI_SHADER_SNORM16_ABGR 0x06
+#define V_028714_SPI_SHADER_UINT16_ABGR 0x07
+#define V_028714_SPI_SHADER_SINT16_ABGR 0x08
+#define V_028714_SPI_SHADER_32_ABGR 0x09
+#define S_028714_COL3_EXPORT_FORMAT(x) (((x) & 0x0F) << 12)
+#define G_028714_COL3_EXPORT_FORMAT(x) (((x) >> 12) & 0x0F)
+#define C_028714_COL3_EXPORT_FORMAT 0xFFFF0FFF
+#define V_028714_SPI_SHADER_ZERO 0x00
+#define V_028714_SPI_SHADER_32_R 0x01
+#define V_028714_SPI_SHADER_32_GR 0x02
+#define V_028714_SPI_SHADER_32_AR 0x03
+#define V_028714_SPI_SHADER_FP16_ABGR 0x04
+#define V_028714_SPI_SHADER_UNORM16_ABGR 0x05
+#define V_028714_SPI_SHADER_SNORM16_ABGR 0x06
+#define V_028714_SPI_SHADER_UINT16_ABGR 0x07
+#define V_028714_SPI_SHADER_SINT16_ABGR 0x08
+#define V_028714_SPI_SHADER_32_ABGR 0x09
+#define S_028714_COL4_EXPORT_FORMAT(x) (((x) & 0x0F) << 16)
+#define G_028714_COL4_EXPORT_FORMAT(x) (((x) >> 16) & 0x0F)
+#define C_028714_COL4_EXPORT_FORMAT 0xFFF0FFFF
+#define V_028714_SPI_SHADER_ZERO 0x00
+#define V_028714_SPI_SHADER_32_R 0x01
+#define V_028714_SPI_SHADER_32_GR 0x02
+#define V_028714_SPI_SHADER_32_AR 0x03
+#define V_028714_SPI_SHADER_FP16_ABGR 0x04
+#define V_028714_SPI_SHADER_UNORM16_ABGR 0x05
+#define V_028714_SPI_SHADER_SNORM16_ABGR 0x06
+#define V_028714_SPI_SHADER_UINT16_ABGR 0x07
+#define V_028714_SPI_SHADER_SINT16_ABGR 0x08
+#define V_028714_SPI_SHADER_32_ABGR 0x09
+#define S_028714_COL5_EXPORT_FORMAT(x) (((x) & 0x0F) << 20)
+#define G_028714_COL5_EXPORT_FORMAT(x) (((x) >> 20) & 0x0F)
+#define C_028714_COL5_EXPORT_FORMAT 0xFF0FFFFF
+#define V_028714_SPI_SHADER_ZERO 0x00
+#define V_028714_SPI_SHADER_32_R 0x01
+#define V_028714_SPI_SHADER_32_GR 0x02
+#define V_028714_SPI_SHADER_32_AR 0x03
+#define V_028714_SPI_SHADER_FP16_ABGR 0x04
+#define V_028714_SPI_SHADER_UNORM16_ABGR 0x05
+#define V_028714_SPI_SHADER_SNORM16_ABGR 0x06
+#define V_028714_SPI_SHADER_UINT16_ABGR 0x07
+#define V_028714_SPI_SHADER_SINT16_ABGR 0x08
+#define V_028714_SPI_SHADER_32_ABGR 0x09
+#define S_028714_COL6_EXPORT_FORMAT(x) (((x) & 0x0F) << 24)
+#define G_028714_COL6_EXPORT_FORMAT(x) (((x) >> 24) & 0x0F)
+#define C_028714_COL6_EXPORT_FORMAT 0xF0FFFFFF
+#define V_028714_SPI_SHADER_ZERO 0x00
+#define V_028714_SPI_SHADER_32_R 0x01
+#define V_028714_SPI_SHADER_32_GR 0x02
+#define V_028714_SPI_SHADER_32_AR 0x03
+#define V_028714_SPI_SHADER_FP16_ABGR 0x04
+#define V_028714_SPI_SHADER_UNORM16_ABGR 0x05
+#define V_028714_SPI_SHADER_SNORM16_ABGR 0x06
+#define V_028714_SPI_SHADER_UINT16_ABGR 0x07
+#define V_028714_SPI_SHADER_SINT16_ABGR 0x08
+#define V_028714_SPI_SHADER_32_ABGR 0x09
+#define S_028714_COL7_EXPORT_FORMAT(x) (((x) & 0x0F) << 28)
+#define G_028714_COL7_EXPORT_FORMAT(x) (((x) >> 28) & 0x0F)
+#define C_028714_COL7_EXPORT_FORMAT 0x0FFFFFFF
+#define V_028714_SPI_SHADER_ZERO 0x00
+#define V_028714_SPI_SHADER_32_R 0x01
+#define V_028714_SPI_SHADER_32_GR 0x02
+#define V_028714_SPI_SHADER_32_AR 0x03
+#define V_028714_SPI_SHADER_FP16_ABGR 0x04
+#define V_028714_SPI_SHADER_UNORM16_ABGR 0x05
+#define V_028714_SPI_SHADER_SNORM16_ABGR 0x06
+#define V_028714_SPI_SHADER_UINT16_ABGR 0x07
+#define V_028714_SPI_SHADER_SINT16_ABGR 0x08
+#define V_028714_SPI_SHADER_32_ABGR 0x09
+#define R_028780_CB_BLEND0_CONTROL 0x028780
+#define S_028780_COLOR_SRCBLEND(x) (((x) & 0x1F) << 0)
+#define G_028780_COLOR_SRCBLEND(x) (((x) >> 0) & 0x1F)
+#define C_028780_COLOR_SRCBLEND 0xFFFFFFE0
+#define V_028780_BLEND_ZERO 0x00
+#define V_028780_BLEND_ONE 0x01
+#define V_028780_BLEND_SRC_COLOR 0x02
+#define V_028780_BLEND_ONE_MINUS_SRC_COLOR 0x03
+#define V_028780_BLEND_SRC_ALPHA 0x04
+#define V_028780_BLEND_ONE_MINUS_SRC_ALPHA 0x05
+#define V_028780_BLEND_DST_ALPHA 0x06
+#define V_028780_BLEND_ONE_MINUS_DST_ALPHA 0x07
+#define V_028780_BLEND_DST_COLOR 0x08
+#define V_028780_BLEND_ONE_MINUS_DST_COLOR 0x09
+#define V_028780_BLEND_SRC_ALPHA_SATURATE 0x0A
+#define V_028780_BLEND_CONSTANT_COLOR 0x0D
+#define V_028780_BLEND_ONE_MINUS_CONSTANT_COLOR 0x0E
+#define V_028780_BLEND_SRC1_COLOR 0x0F
+#define V_028780_BLEND_INV_SRC1_COLOR 0x10
+#define V_028780_BLEND_SRC1_ALPHA 0x11
+#define V_028780_BLEND_INV_SRC1_ALPHA 0x12
+#define V_028780_BLEND_CONSTANT_ALPHA 0x13
+#define V_028780_BLEND_ONE_MINUS_CONSTANT_ALPHA 0x14
+#define S_028780_COLOR_COMB_FCN(x) (((x) & 0x07) << 5)
+#define G_028780_COLOR_COMB_FCN(x) (((x) >> 5) & 0x07)
+#define C_028780_COLOR_COMB_FCN 0xFFFFFF1F
+#define V_028780_COMB_DST_PLUS_SRC 0x00
+#define V_028780_COMB_SRC_MINUS_DST 0x01
+#define V_028780_COMB_MIN_DST_SRC 0x02
+#define V_028780_COMB_MAX_DST_SRC 0x03
+#define V_028780_COMB_DST_MINUS_SRC 0x04
+#define S_028780_COLOR_DESTBLEND(x) (((x) & 0x1F) << 8)
+#define G_028780_COLOR_DESTBLEND(x) (((x) >> 8) & 0x1F)
+#define C_028780_COLOR_DESTBLEND 0xFFFFE0FF
+#define V_028780_BLEND_ZERO 0x00
+#define V_028780_BLEND_ONE 0x01
+#define V_028780_BLEND_SRC_COLOR 0x02
+#define V_028780_BLEND_ONE_MINUS_SRC_COLOR 0x03
+#define V_028780_BLEND_SRC_ALPHA 0x04
+#define V_028780_BLEND_ONE_MINUS_SRC_ALPHA 0x05
+#define V_028780_BLEND_DST_ALPHA 0x06
+#define V_028780_BLEND_ONE_MINUS_DST_ALPHA 0x07
+#define V_028780_BLEND_DST_COLOR 0x08
+#define V_028780_BLEND_ONE_MINUS_DST_COLOR 0x09
+#define V_028780_BLEND_SRC_ALPHA_SATURATE 0x0A
+#define V_028780_BLEND_CONSTANT_COLOR 0x0D
+#define V_028780_BLEND_ONE_MINUS_CONSTANT_COLOR 0x0E
+#define V_028780_BLEND_SRC1_COLOR 0x0F
+#define V_028780_BLEND_INV_SRC1_COLOR 0x10
+#define V_028780_BLEND_SRC1_ALPHA 0x11
+#define V_028780_BLEND_INV_SRC1_ALPHA 0x12
+#define V_028780_BLEND_CONSTANT_ALPHA 0x13
+#define V_028780_BLEND_ONE_MINUS_CONSTANT_ALPHA 0x14
+#define S_028780_ALPHA_SRCBLEND(x) (((x) & 0x1F) << 16)
+#define G_028780_ALPHA_SRCBLEND(x) (((x) >> 16) & 0x1F)
+#define C_028780_ALPHA_SRCBLEND 0xFFE0FFFF
+#define V_028780_BLEND_ZERO 0x00
+#define V_028780_BLEND_ONE 0x01
+#define V_028780_BLEND_SRC_COLOR 0x02
+#define V_028780_BLEND_ONE_MINUS_SRC_COLOR 0x03
+#define V_028780_BLEND_SRC_ALPHA 0x04
+#define V_028780_BLEND_ONE_MINUS_SRC_ALPHA 0x05
+#define V_028780_BLEND_DST_ALPHA 0x06
+#define V_028780_BLEND_ONE_MINUS_DST_ALPHA 0x07
+#define V_028780_BLEND_DST_COLOR 0x08
+#define V_028780_BLEND_ONE_MINUS_DST_COLOR 0x09
+#define V_028780_BLEND_SRC_ALPHA_SATURATE 0x0A
+#define V_028780_BLEND_CONSTANT_COLOR 0x0D
+#define V_028780_BLEND_ONE_MINUS_CONSTANT_COLOR 0x0E
+#define V_028780_BLEND_SRC1_COLOR 0x0F
+#define V_028780_BLEND_INV_SRC1_COLOR 0x10
+#define V_028780_BLEND_SRC1_ALPHA 0x11
+#define V_028780_BLEND_INV_SRC1_ALPHA 0x12
+#define V_028780_BLEND_CONSTANT_ALPHA 0x13
+#define V_028780_BLEND_ONE_MINUS_CONSTANT_ALPHA 0x14
+#define S_028780_ALPHA_COMB_FCN(x) (((x) & 0x07) << 21)
+#define G_028780_ALPHA_COMB_FCN(x) (((x) >> 21) & 0x07)
+#define C_028780_ALPHA_COMB_FCN 0xFF1FFFFF
+#define V_028780_COMB_DST_PLUS_SRC 0x00
+#define V_028780_COMB_SRC_MINUS_DST 0x01
+#define V_028780_COMB_MIN_DST_SRC 0x02
+#define V_028780_COMB_MAX_DST_SRC 0x03
+#define V_028780_COMB_DST_MINUS_SRC 0x04
+#define S_028780_ALPHA_DESTBLEND(x) (((x) & 0x1F) << 24)
+#define G_028780_ALPHA_DESTBLEND(x) (((x) >> 24) & 0x1F)
+#define C_028780_ALPHA_DESTBLEND 0xE0FFFFFF
+#define V_028780_BLEND_ZERO 0x00
+#define V_028780_BLEND_ONE 0x01
+#define V_028780_BLEND_SRC_COLOR 0x02
+#define V_028780_BLEND_ONE_MINUS_SRC_COLOR 0x03
+#define V_028780_BLEND_SRC_ALPHA 0x04
+#define V_028780_BLEND_ONE_MINUS_SRC_ALPHA 0x05
+#define V_028780_BLEND_DST_ALPHA 0x06
+#define V_028780_BLEND_ONE_MINUS_DST_ALPHA 0x07
+#define V_028780_BLEND_DST_COLOR 0x08
+#define V_028780_BLEND_ONE_MINUS_DST_COLOR 0x09
+#define V_028780_BLEND_SRC_ALPHA_SATURATE 0x0A
+#define V_028780_BLEND_CONSTANT_COLOR 0x0D
+#define V_028780_BLEND_ONE_MINUS_CONSTANT_COLOR 0x0E
+#define V_028780_BLEND_SRC1_COLOR 0x0F
+#define V_028780_BLEND_INV_SRC1_COLOR 0x10
+#define V_028780_BLEND_SRC1_ALPHA 0x11
+#define V_028780_BLEND_INV_SRC1_ALPHA 0x12
+#define V_028780_BLEND_CONSTANT_ALPHA 0x13
+#define V_028780_BLEND_ONE_MINUS_CONSTANT_ALPHA 0x14
+#define S_028780_SEPARATE_ALPHA_BLEND(x) (((x) & 0x1) << 29)
+#define G_028780_SEPARATE_ALPHA_BLEND(x) (((x) >> 29) & 0x1)
+#define C_028780_SEPARATE_ALPHA_BLEND 0xDFFFFFFF
+#define S_028780_ENABLE(x) (((x) & 0x1) << 30)
+#define G_028780_ENABLE(x) (((x) >> 30) & 0x1)
+#define C_028780_ENABLE 0xBFFFFFFF
+#define S_028780_DISABLE_ROP3(x) (((x) & 0x1) << 31)
+#define G_028780_DISABLE_ROP3(x) (((x) >> 31) & 0x1)
+#define C_028780_DISABLE_ROP3 0x7FFFFFFF
+#define R_028784_CB_BLEND1_CONTROL 0x028784
+#define R_028788_CB_BLEND2_CONTROL 0x028788
+#define R_02878C_CB_BLEND3_CONTROL 0x02878C
+#define R_028790_CB_BLEND4_CONTROL 0x028790
+#define R_028794_CB_BLEND5_CONTROL 0x028794
+#define R_028798_CB_BLEND6_CONTROL 0x028798
+#define R_02879C_CB_BLEND7_CONTROL 0x02879C
+#define R_0287D4_PA_CL_POINT_X_RAD 0x0287D4
+#define R_0287D8_PA_CL_POINT_Y_RAD 0x0287D8
+#define R_0287DC_PA_CL_POINT_SIZE 0x0287DC
+#define R_0287E0_PA_CL_POINT_CULL_RAD 0x0287E0
+#define R_0287E4_VGT_DMA_BASE_HI 0x0287E4
+#define S_0287E4_BASE_ADDR(x) (((x) & 0xFF) << 0)
+#define G_0287E4_BASE_ADDR(x) (((x) >> 0) & 0xFF)
+#define C_0287E4_BASE_ADDR 0xFFFFFF00
+#define R_0287E8_VGT_DMA_BASE 0x0287E8
+#define R_0287F0_VGT_DRAW_INITIATOR 0x0287F0
+#define S_0287F0_SOURCE_SELECT(x) (((x) & 0x03) << 0)
+#define G_0287F0_SOURCE_SELECT(x) (((x) >> 0) & 0x03)
+#define C_0287F0_SOURCE_SELECT 0xFFFFFFFC
+#define V_0287F0_DI_SRC_SEL_DMA 0x00
+#define V_0287F0_DI_SRC_SEL_IMMEDIATE 0x01
+#define V_0287F0_DI_SRC_SEL_AUTO_INDEX 0x02
+#define V_0287F0_DI_SRC_SEL_RESERVED 0x03
+#define S_0287F0_MAJOR_MODE(x) (((x) & 0x03) << 2)
+#define G_0287F0_MAJOR_MODE(x) (((x) >> 2) & 0x03)
+#define C_0287F0_MAJOR_MODE 0xFFFFFFF3
+#define V_0287F0_DI_MAJOR_MODE_0 0x00
+#define V_0287F0_DI_MAJOR_MODE_1 0x01
+#define S_0287F0_NOT_EOP(x) (((x) & 0x1) << 5)
+#define G_0287F0_NOT_EOP(x) (((x) >> 5) & 0x1)
+#define C_0287F0_NOT_EOP 0xFFFFFFDF
+#define S_0287F0_USE_OPAQUE(x) (((x) & 0x1) << 6)
+#define G_0287F0_USE_OPAQUE(x) (((x) >> 6) & 0x1)
+#define C_0287F0_USE_OPAQUE 0xFFFFFFBF
+#define R_0287F4_VGT_IMMED_DATA 0x0287F4
+#define R_028800_DB_DEPTH_CONTROL 0x028800
+#define S_028800_STENCIL_ENABLE(x) (((x) & 0x1) << 0)
+#define G_028800_STENCIL_ENABLE(x) (((x) >> 0) & 0x1)
+#define C_028800_STENCIL_ENABLE 0xFFFFFFFE
+#define S_028800_Z_ENABLE(x) (((x) & 0x1) << 1)
+#define G_028800_Z_ENABLE(x) (((x) >> 1) & 0x1)
+#define C_028800_Z_ENABLE 0xFFFFFFFD
+#define S_028800_Z_WRITE_ENABLE(x) (((x) & 0x1) << 2)
+#define G_028800_Z_WRITE_ENABLE(x) (((x) >> 2) & 0x1)
+#define C_028800_Z_WRITE_ENABLE 0xFFFFFFFB
+#define S_028800_DEPTH_BOUNDS_ENABLE(x) (((x) & 0x1) << 3)
+#define G_028800_DEPTH_BOUNDS_ENABLE(x) (((x) >> 3) & 0x1)
+#define C_028800_DEPTH_BOUNDS_ENABLE 0xFFFFFFF7
+#define S_028800_ZFUNC(x) (((x) & 0x07) << 4)
+#define G_028800_ZFUNC(x) (((x) >> 4) & 0x07)
+#define C_028800_ZFUNC 0xFFFFFF8F
+#define V_028800_FRAG_NEVER 0x00
+#define V_028800_FRAG_LESS 0x01
+#define V_028800_FRAG_EQUAL 0x02
+#define V_028800_FRAG_LEQUAL 0x03
+#define V_028800_FRAG_GREATER 0x04
+#define V_028800_FRAG_NOTEQUAL 0x05
+#define V_028800_FRAG_GEQUAL 0x06
+#define V_028800_FRAG_ALWAYS 0x07
+#define S_028800_BACKFACE_ENABLE(x) (((x) & 0x1) << 7)
+#define G_028800_BACKFACE_ENABLE(x) (((x) >> 7) & 0x1)
+#define C_028800_BACKFACE_ENABLE 0xFFFFFF7F
+#define S_028800_STENCILFUNC(x) (((x) & 0x07) << 8)
+#define G_028800_STENCILFUNC(x) (((x) >> 8) & 0x07)
+#define C_028800_STENCILFUNC 0xFFFFF8FF
+#define V_028800_REF_NEVER 0x00
+#define V_028800_REF_LESS 0x01
+#define V_028800_REF_EQUAL 0x02
+#define V_028800_REF_LEQUAL 0x03
+#define V_028800_REF_GREATER 0x04
+#define V_028800_REF_NOTEQUAL 0x05
+#define V_028800_REF_GEQUAL 0x06
+#define V_028800_REF_ALWAYS 0x07
+#define S_028800_STENCILFUNC_BF(x) (((x) & 0x07) << 20)
+#define G_028800_STENCILFUNC_BF(x) (((x) >> 20) & 0x07)
+#define C_028800_STENCILFUNC_BF 0xFF8FFFFF
+#define V_028800_REF_NEVER 0x00
+#define V_028800_REF_LESS 0x01
+#define V_028800_REF_EQUAL 0x02
+#define V_028800_REF_LEQUAL 0x03
+#define V_028800_REF_GREATER 0x04
+#define V_028800_REF_NOTEQUAL 0x05
+#define V_028800_REF_GEQUAL 0x06
+#define V_028800_REF_ALWAYS 0x07
+#define S_028800_ENABLE_COLOR_WRITES_ON_DEPTH_FAIL(x) (((x) & 0x1) << 30)
+#define G_028800_ENABLE_COLOR_WRITES_ON_DEPTH_FAIL(x) (((x) >> 30) & 0x1)
+#define C_028800_ENABLE_COLOR_WRITES_ON_DEPTH_FAIL 0xBFFFFFFF
+#define S_028800_DISABLE_COLOR_WRITES_ON_DEPTH_PASS(x) (((x) & 0x1) << 31)
+#define G_028800_DISABLE_COLOR_WRITES_ON_DEPTH_PASS(x) (((x) >> 31) & 0x1)
+#define C_028800_DISABLE_COLOR_WRITES_ON_DEPTH_PASS 0x7FFFFFFF
+#define R_028804_DB_EQAA 0x028804
+#define R_028808_CB_COLOR_CONTROL 0x028808
+#define S_028808_DEGAMMA_ENABLE(x) (((x) & 0x1) << 3)
+#define G_028808_DEGAMMA_ENABLE(x) (((x) >> 3) & 0x1)
+#define C_028808_DEGAMMA_ENABLE 0xFFFFFFF7
+#define S_028808_MODE(x) (((x) & 0x07) << 4)
+#define G_028808_MODE(x) (((x) >> 4) & 0x07)
+#define C_028808_MODE 0xFFFFFF8F
+#define V_028808_CB_DISABLE 0x00
+#define V_028808_CB_NORMAL 0x01
+#define V_028808_CB_ELIMINATE_FAST_CLEAR 0x02
+#define V_028808_CB_RESOLVE 0x03
+#define V_028808_CB_FMASK_DECOMPRESS 0x05
+#define S_028808_ROP3(x) (((x) & 0xFF) << 16)
+#define G_028808_ROP3(x) (((x) >> 16) & 0xFF)
+#define C_028808_ROP3 0xFF00FFFF
+#define V_028808_X_0X00 0x00
+#define V_028808_X_0X05 0x05
+#define V_028808_X_0X0A 0x0A
+#define V_028808_X_0X0F 0x0F
+#define V_028808_X_0X11 0x11
+#define V_028808_X_0X22 0x22
+#define V_028808_X_0X33 0x33
+#define V_028808_X_0X44 0x44
+#define V_028808_X_0X50 0x50
+#define V_028808_X_0X55 0x55
+#define V_028808_X_0X5A 0x5A
+#define V_028808_X_0X5F 0x5F
+#define V_028808_X_0X66 0x66
+#define V_028808_X_0X77 0x77
+#define V_028808_X_0X88 0x88
+#define V_028808_X_0X99 0x99
+#define V_028808_X_0XA0 0xA0
+#define V_028808_X_0XA5 0xA5
+#define V_028808_X_0XAA 0xAA
+#define V_028808_X_0XAF 0xAF
+#define V_028808_X_0XBB 0xBB
+#define V_028808_X_0XCC 0xCC
+#define V_028808_X_0XDD 0xDD
+#define V_028808_X_0XEE 0xEE
+#define V_028808_X_0XF0 0xF0
+#define V_028808_X_0XF5 0xF5
+#define V_028808_X_0XFA 0xFA
+#define V_028808_X_0XFF 0xFF
+#define R_02880C_DB_SHADER_CONTROL 0x02880C
+#define S_02880C_Z_EXPORT_ENABLE(x) (((x) & 0x1) << 0)
+#define G_02880C_Z_EXPORT_ENABLE(x) (((x) >> 0) & 0x1)
+#define C_02880C_Z_EXPORT_ENABLE 0xFFFFFFFE
+#define S_02880C_STENCIL_TEST_VAL_EXPORT_ENAB(x) (((x) & 0x1) << 1)
+#define G_02880C_STENCIL_TEST_VAL_EXPORT_ENAB(x) (((x) >> 1) & 0x1)
+#define C_02880C_STENCIL_TEST_VAL_EXPORT_ENAB 0xFFFFFFFD
+#define S_02880C_STENCIL_OP_VAL_EXPORT_ENABLE(x) (((x) & 0x1) << 2)
+#define G_02880C_STENCIL_OP_VAL_EXPORT_ENABLE(x) (((x) >> 2) & 0x1)
+#define C_02880C_STENCIL_OP_VAL_EXPORT_ENABLE 0xFFFFFFFB
+#define S_02880C_Z_ORDER(x) (((x) & 0x03) << 4)
+#define G_02880C_Z_ORDER(x) (((x) >> 4) & 0x03)
+#define C_02880C_Z_ORDER 0xFFFFFFCF
+#define V_02880C_LATE_Z 0x00
+#define V_02880C_EARLY_Z_THEN_LATE_Z 0x01
+#define V_02880C_RE_Z 0x02
+#define V_02880C_EARLY_Z_THEN_RE_Z 0x03
+#define S_02880C_KILL_ENABLE(x) (((x) & 0x1) << 6)
+#define G_02880C_KILL_ENABLE(x) (((x) >> 6) & 0x1)
+#define C_02880C_KILL_ENABLE 0xFFFFFFBF
+#define S_02880C_COVERAGE_TO_MASK_ENABLE(x) (((x) & 0x1) << 7)
+#define G_02880C_COVERAGE_TO_MASK_ENABLE(x) (((x) >> 7) & 0x1)
+#define C_02880C_COVERAGE_TO_MASK_ENABLE 0xFFFFFF7F
+#define S_02880C_MASK_EXPORT_ENABLE(x) (((x) & 0x1) << 8)
+#define G_02880C_MASK_EXPORT_ENABLE(x) (((x) >> 8) & 0x1)
+#define C_02880C_MASK_EXPORT_ENABLE 0xFFFFFEFF
+#define S_02880C_EXEC_ON_HIER_FAIL(x) (((x) & 0x1) << 9)
+#define G_02880C_EXEC_ON_HIER_FAIL(x) (((x) >> 9) & 0x1)
+#define C_02880C_EXEC_ON_HIER_FAIL 0xFFFFFDFF
+#define S_02880C_EXEC_ON_NOOP(x) (((x) & 0x1) << 10)
+#define G_02880C_EXEC_ON_NOOP(x) (((x) >> 10) & 0x1)
+#define C_02880C_EXEC_ON_NOOP 0xFFFFFBFF
+#define S_02880C_ALPHA_TO_MASK_DISABLE(x) (((x) & 0x1) << 11)
+#define G_02880C_ALPHA_TO_MASK_DISABLE(x) (((x) >> 11) & 0x1)
+#define C_02880C_ALPHA_TO_MASK_DISABLE 0xFFFFF7FF
+#define S_02880C_DEPTH_BEFORE_SHADER(x) (((x) & 0x1) << 12)
+#define G_02880C_DEPTH_BEFORE_SHADER(x) (((x) >> 12) & 0x1)
+#define C_02880C_DEPTH_BEFORE_SHADER 0xFFFFEFFF
+#define R_028810_PA_CL_CLIP_CNTL 0x028810
+#define S_028810_UCP_ENA_0(x) (((x) & 0x1) << 0)
+#define G_028810_UCP_ENA_0(x) (((x) >> 0) & 0x1)
+#define C_028810_UCP_ENA_0 0xFFFFFFFE
+#define S_028810_UCP_ENA_1(x) (((x) & 0x1) << 1)
+#define G_028810_UCP_ENA_1(x) (((x) >> 1) & 0x1)
+#define C_028810_UCP_ENA_1 0xFFFFFFFD
+#define S_028810_UCP_ENA_2(x) (((x) & 0x1) << 2)
+#define G_028810_UCP_ENA_2(x) (((x) >> 2) & 0x1)
+#define C_028810_UCP_ENA_2 0xFFFFFFFB
+#define S_028810_UCP_ENA_3(x) (((x) & 0x1) << 3)
+#define G_028810_UCP_ENA_3(x) (((x) >> 3) & 0x1)
+#define C_028810_UCP_ENA_3 0xFFFFFFF7
+#define S_028810_UCP_ENA_4(x) (((x) & 0x1) << 4)
+#define G_028810_UCP_ENA_4(x) (((x) >> 4) & 0x1)
+#define C_028810_UCP_ENA_4 0xFFFFFFEF
+#define S_028810_UCP_ENA_5(x) (((x) & 0x1) << 5)
+#define G_028810_UCP_ENA_5(x) (((x) >> 5) & 0x1)
+#define C_028810_UCP_ENA_5 0xFFFFFFDF
+#define S_028810_PS_UCP_Y_SCALE_NEG(x) (((x) & 0x1) << 13)
+#define G_028810_PS_UCP_Y_SCALE_NEG(x) (((x) >> 13) & 0x1)
+#define C_028810_PS_UCP_Y_SCALE_NEG 0xFFFFDFFF
+#define S_028810_PS_UCP_MODE(x) (((x) & 0x03) << 14)
+#define G_028810_PS_UCP_MODE(x) (((x) >> 14) & 0x03)
+#define C_028810_PS_UCP_MODE 0xFFFF3FFF
+#define S_028810_CLIP_DISABLE(x) (((x) & 0x1) << 16)
+#define G_028810_CLIP_DISABLE(x) (((x) >> 16) & 0x1)
+#define C_028810_CLIP_DISABLE 0xFFFEFFFF
+#define S_028810_UCP_CULL_ONLY_ENA(x) (((x) & 0x1) << 17)
+#define G_028810_UCP_CULL_ONLY_ENA(x) (((x) >> 17) & 0x1)
+#define C_028810_UCP_CULL_ONLY_ENA 0xFFFDFFFF
+#define S_028810_BOUNDARY_EDGE_FLAG_ENA(x) (((x) & 0x1) << 18)
+#define G_028810_BOUNDARY_EDGE_FLAG_ENA(x) (((x) >> 18) & 0x1)
+#define C_028810_BOUNDARY_EDGE_FLAG_ENA 0xFFFBFFFF
+#define S_028810_DX_CLIP_SPACE_DEF(x) (((x) & 0x1) << 19)
+#define G_028810_DX_CLIP_SPACE_DEF(x) (((x) >> 19) & 0x1)
+#define C_028810_DX_CLIP_SPACE_DEF 0xFFF7FFFF
+#define S_028810_DIS_CLIP_ERR_DETECT(x) (((x) & 0x1) << 20)
+#define G_028810_DIS_CLIP_ERR_DETECT(x) (((x) >> 20) & 0x1)
+#define C_028810_DIS_CLIP_ERR_DETECT 0xFFEFFFFF
+#define S_028810_VTX_KILL_OR(x) (((x) & 0x1) << 21)
+#define G_028810_VTX_KILL_OR(x) (((x) >> 21) & 0x1)
+#define C_028810_VTX_KILL_OR 0xFFDFFFFF
+#define S_028810_DX_RASTERIZATION_KILL(x) (((x) & 0x1) << 22)
+#define G_028810_DX_RASTERIZATION_KILL(x) (((x) >> 22) & 0x1)
+#define C_028810_DX_RASTERIZATION_KILL 0xFFBFFFFF
+#define S_028810_DX_LINEAR_ATTR_CLIP_ENA(x) (((x) & 0x1) << 24)
+#define G_028810_DX_LINEAR_ATTR_CLIP_ENA(x) (((x) >> 24) & 0x1)
+#define C_028810_DX_LINEAR_ATTR_CLIP_ENA 0xFEFFFFFF
+#define S_028810_VTE_VPORT_PROVOKE_DISABLE(x) (((x) & 0x1) << 25)
+#define G_028810_VTE_VPORT_PROVOKE_DISABLE(x) (((x) >> 25) & 0x1)
+#define C_028810_VTE_VPORT_PROVOKE_DISABLE 0xFDFFFFFF
+#define S_028810_ZCLIP_NEAR_DISABLE(x) (((x) & 0x1) << 26)
+#define G_028810_ZCLIP_NEAR_DISABLE(x) (((x) >> 26) & 0x1)
+#define C_028810_ZCLIP_NEAR_DISABLE 0xFBFFFFFF
+#define S_028810_ZCLIP_FAR_DISABLE(x) (((x) & 0x1) << 27)
+#define G_028810_ZCLIP_FAR_DISABLE(x) (((x) >> 27) & 0x1)
+#define C_028810_ZCLIP_FAR_DISABLE 0xF7FFFFFF
+#define R_028814_PA_SU_SC_MODE_CNTL 0x028814
+#define S_028814_CULL_FRONT(x) (((x) & 0x1) << 0)
+#define G_028814_CULL_FRONT(x) (((x) >> 0) & 0x1)
+#define C_028814_CULL_FRONT 0xFFFFFFFE
+#define S_028814_CULL_BACK(x) (((x) & 0x1) << 1)
+#define G_028814_CULL_BACK(x) (((x) >> 1) & 0x1)
+#define C_028814_CULL_BACK 0xFFFFFFFD
+#define S_028814_FACE(x) (((x) & 0x1) << 2)
+#define G_028814_FACE(x) (((x) >> 2) & 0x1)
+#define C_028814_FACE 0xFFFFFFFB
+#define S_028814_POLY_MODE(x) (((x) & 0x03) << 3)
+#define G_028814_POLY_MODE(x) (((x) >> 3) & 0x03)
+#define C_028814_POLY_MODE 0xFFFFFFE7
+#define V_028814_X_DISABLE_POLY_MODE 0x00
+#define V_028814_X_DUAL_MODE 0x01
+#define S_028814_POLYMODE_FRONT_PTYPE(x) (((x) & 0x07) << 5)
+#define G_028814_POLYMODE_FRONT_PTYPE(x) (((x) >> 5) & 0x07)
+#define C_028814_POLYMODE_FRONT_PTYPE 0xFFFFFF1F
+#define V_028814_X_DRAW_POINTS 0x00
+#define V_028814_X_DRAW_LINES 0x01
+#define V_028814_X_DRAW_TRIANGLES 0x02
+#define S_028814_POLYMODE_BACK_PTYPE(x) (((x) & 0x07) << 8)
+#define G_028814_POLYMODE_BACK_PTYPE(x) (((x) >> 8) & 0x07)
+#define C_028814_POLYMODE_BACK_PTYPE 0xFFFFF8FF
+#define V_028814_X_DRAW_POINTS 0x00
+#define V_028814_X_DRAW_LINES 0x01
+#define V_028814_X_DRAW_TRIANGLES 0x02
+#define S_028814_POLY_OFFSET_FRONT_ENABLE(x) (((x) & 0x1) << 11)
+#define G_028814_POLY_OFFSET_FRONT_ENABLE(x) (((x) >> 11) & 0x1)
+#define C_028814_POLY_OFFSET_FRONT_ENABLE 0xFFFFF7FF
+#define S_028814_POLY_OFFSET_BACK_ENABLE(x) (((x) & 0x1) << 12)
+#define G_028814_POLY_OFFSET_BACK_ENABLE(x) (((x) >> 12) & 0x1)
+#define C_028814_POLY_OFFSET_BACK_ENABLE 0xFFFFEFFF
+#define S_028814_POLY_OFFSET_PARA_ENABLE(x) (((x) & 0x1) << 13)
+#define G_028814_POLY_OFFSET_PARA_ENABLE(x) (((x) >> 13) & 0x1)
+#define C_028814_POLY_OFFSET_PARA_ENABLE 0xFFFFDFFF
+#define S_028814_VTX_WINDOW_OFFSET_ENABLE(x) (((x) & 0x1) << 16)
+#define G_028814_VTX_WINDOW_OFFSET_ENABLE(x) (((x) >> 16) & 0x1)
+#define C_028814_VTX_WINDOW_OFFSET_ENABLE 0xFFFEFFFF
+#define S_028814_PROVOKING_VTX_LAST(x) (((x) & 0x1) << 19)
+#define G_028814_PROVOKING_VTX_LAST(x) (((x) >> 19) & 0x1)
+#define C_028814_PROVOKING_VTX_LAST 0xFFF7FFFF
+#define S_028814_PERSP_CORR_DIS(x) (((x) & 0x1) << 20)
+#define G_028814_PERSP_CORR_DIS(x) (((x) >> 20) & 0x1)
+#define C_028814_PERSP_CORR_DIS 0xFFEFFFFF
+#define S_028814_MULTI_PRIM_IB_ENA(x) (((x) & 0x1) << 21)
+#define G_028814_MULTI_PRIM_IB_ENA(x) (((x) >> 21) & 0x1)
+#define C_028814_MULTI_PRIM_IB_ENA 0xFFDFFFFF
+#define R_028818_PA_CL_VTE_CNTL 0x028818
+#define S_028818_VPORT_X_SCALE_ENA(x) (((x) & 0x1) << 0)
+#define G_028818_VPORT_X_SCALE_ENA(x) (((x) >> 0) & 0x1)
+#define C_028818_VPORT_X_SCALE_ENA 0xFFFFFFFE
+#define S_028818_VPORT_X_OFFSET_ENA(x) (((x) & 0x1) << 1)
+#define G_028818_VPORT_X_OFFSET_ENA(x) (((x) >> 1) & 0x1)
+#define C_028818_VPORT_X_OFFSET_ENA 0xFFFFFFFD
+#define S_028818_VPORT_Y_SCALE_ENA(x) (((x) & 0x1) << 2)
+#define G_028818_VPORT_Y_SCALE_ENA(x) (((x) >> 2) & 0x1)
+#define C_028818_VPORT_Y_SCALE_ENA 0xFFFFFFFB
+#define S_028818_VPORT_Y_OFFSET_ENA(x) (((x) & 0x1) << 3)
+#define G_028818_VPORT_Y_OFFSET_ENA(x) (((x) >> 3) & 0x1)
+#define C_028818_VPORT_Y_OFFSET_ENA 0xFFFFFFF7
+#define S_028818_VPORT_Z_SCALE_ENA(x) (((x) & 0x1) << 4)
+#define G_028818_VPORT_Z_SCALE_ENA(x) (((x) >> 4) & 0x1)
+#define C_028818_VPORT_Z_SCALE_ENA 0xFFFFFFEF
+#define S_028818_VPORT_Z_OFFSET_ENA(x) (((x) & 0x1) << 5)
+#define G_028818_VPORT_Z_OFFSET_ENA(x) (((x) >> 5) & 0x1)
+#define C_028818_VPORT_Z_OFFSET_ENA 0xFFFFFFDF
+#define S_028818_VTX_XY_FMT(x) (((x) & 0x1) << 8)
+#define G_028818_VTX_XY_FMT(x) (((x) >> 8) & 0x1)
+#define C_028818_VTX_XY_FMT 0xFFFFFEFF
+#define S_028818_VTX_Z_FMT(x) (((x) & 0x1) << 9)
+#define G_028818_VTX_Z_FMT(x) (((x) >> 9) & 0x1)
+#define C_028818_VTX_Z_FMT 0xFFFFFDFF
+#define S_028818_VTX_W0_FMT(x) (((x) & 0x1) << 10)
+#define G_028818_VTX_W0_FMT(x) (((x) >> 10) & 0x1)
+#define C_028818_VTX_W0_FMT 0xFFFFFBFF
+#define R_02881C_PA_CL_VS_OUT_CNTL 0x02881C
+#define S_02881C_CLIP_DIST_ENA_0(x) (((x) & 0x1) << 0)
+#define G_02881C_CLIP_DIST_ENA_0(x) (((x) >> 0) & 0x1)
+#define C_02881C_CLIP_DIST_ENA_0 0xFFFFFFFE
+#define S_02881C_CLIP_DIST_ENA_1(x) (((x) & 0x1) << 1)
+#define G_02881C_CLIP_DIST_ENA_1(x) (((x) >> 1) & 0x1)
+#define C_02881C_CLIP_DIST_ENA_1 0xFFFFFFFD
+#define S_02881C_CLIP_DIST_ENA_2(x) (((x) & 0x1) << 2)
+#define G_02881C_CLIP_DIST_ENA_2(x) (((x) >> 2) & 0x1)
+#define C_02881C_CLIP_DIST_ENA_2 0xFFFFFFFB
+#define S_02881C_CLIP_DIST_ENA_3(x) (((x) & 0x1) << 3)
+#define G_02881C_CLIP_DIST_ENA_3(x) (((x) >> 3) & 0x1)
+#define C_02881C_CLIP_DIST_ENA_3 0xFFFFFFF7
+#define S_02881C_CLIP_DIST_ENA_4(x) (((x) & 0x1) << 4)
+#define G_02881C_CLIP_DIST_ENA_4(x) (((x) >> 4) & 0x1)
+#define C_02881C_CLIP_DIST_ENA_4 0xFFFFFFEF
+#define S_02881C_CLIP_DIST_ENA_5(x) (((x) & 0x1) << 5)
+#define G_02881C_CLIP_DIST_ENA_5(x) (((x) >> 5) & 0x1)
+#define C_02881C_CLIP_DIST_ENA_5 0xFFFFFFDF
+#define S_02881C_CLIP_DIST_ENA_6(x) (((x) & 0x1) << 6)
+#define G_02881C_CLIP_DIST_ENA_6(x) (((x) >> 6) & 0x1)
+#define C_02881C_CLIP_DIST_ENA_6 0xFFFFFFBF
+#define S_02881C_CLIP_DIST_ENA_7(x) (((x) & 0x1) << 7)
+#define G_02881C_CLIP_DIST_ENA_7(x) (((x) >> 7) & 0x1)
+#define C_02881C_CLIP_DIST_ENA_7 0xFFFFFF7F
+#define S_02881C_CULL_DIST_ENA_0(x) (((x) & 0x1) << 8)
+#define G_02881C_CULL_DIST_ENA_0(x) (((x) >> 8) & 0x1)
+#define C_02881C_CULL_DIST_ENA_0 0xFFFFFEFF
+#define S_02881C_CULL_DIST_ENA_1(x) (((x) & 0x1) << 9)
+#define G_02881C_CULL_DIST_ENA_1(x) (((x) >> 9) & 0x1)
+#define C_02881C_CULL_DIST_ENA_1 0xFFFFFDFF
+#define S_02881C_CULL_DIST_ENA_2(x) (((x) & 0x1) << 10)
+#define G_02881C_CULL_DIST_ENA_2(x) (((x) >> 10) & 0x1)
+#define C_02881C_CULL_DIST_ENA_2 0xFFFFFBFF
+#define S_02881C_CULL_DIST_ENA_3(x) (((x) & 0x1) << 11)
+#define G_02881C_CULL_DIST_ENA_3(x) (((x) >> 11) & 0x1)
+#define C_02881C_CULL_DIST_ENA_3 0xFFFFF7FF
+#define S_02881C_CULL_DIST_ENA_4(x) (((x) & 0x1) << 12)
+#define G_02881C_CULL_DIST_ENA_4(x) (((x) >> 12) & 0x1)
+#define C_02881C_CULL_DIST_ENA_4 0xFFFFEFFF
+#define S_02881C_CULL_DIST_ENA_5(x) (((x) & 0x1) << 13)
+#define G_02881C_CULL_DIST_ENA_5(x) (((x) >> 13) & 0x1)
+#define C_02881C_CULL_DIST_ENA_5 0xFFFFDFFF
+#define S_02881C_CULL_DIST_ENA_6(x) (((x) & 0x1) << 14)
+#define G_02881C_CULL_DIST_ENA_6(x) (((x) >> 14) & 0x1)
+#define C_02881C_CULL_DIST_ENA_6 0xFFFFBFFF
+#define S_02881C_CULL_DIST_ENA_7(x) (((x) & 0x1) << 15)
+#define G_02881C_CULL_DIST_ENA_7(x) (((x) >> 15) & 0x1)
+#define C_02881C_CULL_DIST_ENA_7 0xFFFF7FFF
+#define S_02881C_USE_VTX_POINT_SIZE(x) (((x) & 0x1) << 16)
+#define G_02881C_USE_VTX_POINT_SIZE(x) (((x) >> 16) & 0x1)
+#define C_02881C_USE_VTX_POINT_SIZE 0xFFFEFFFF
+#define S_02881C_USE_VTX_EDGE_FLAG(x) (((x) & 0x1) << 17)
+#define G_02881C_USE_VTX_EDGE_FLAG(x) (((x) >> 17) & 0x1)
+#define C_02881C_USE_VTX_EDGE_FLAG 0xFFFDFFFF
+#define S_02881C_USE_VTX_RENDER_TARGET_INDX(x) (((x) & 0x1) << 18)
+#define G_02881C_USE_VTX_RENDER_TARGET_INDX(x) (((x) >> 18) & 0x1)
+#define C_02881C_USE_VTX_RENDER_TARGET_INDX 0xFFFBFFFF
+#define S_02881C_USE_VTX_VIEWPORT_INDX(x) (((x) & 0x1) << 19)
+#define G_02881C_USE_VTX_VIEWPORT_INDX(x) (((x) >> 19) & 0x1)
+#define C_02881C_USE_VTX_VIEWPORT_INDX 0xFFF7FFFF
+#define S_02881C_USE_VTX_KILL_FLAG(x) (((x) & 0x1) << 20)
+#define G_02881C_USE_VTX_KILL_FLAG(x) (((x) >> 20) & 0x1)
+#define C_02881C_USE_VTX_KILL_FLAG 0xFFEFFFFF
+#define S_02881C_VS_OUT_MISC_VEC_ENA(x) (((x) & 0x1) << 21)
+#define G_02881C_VS_OUT_MISC_VEC_ENA(x) (((x) >> 21) & 0x1)
+#define C_02881C_VS_OUT_MISC_VEC_ENA 0xFFDFFFFF
+#define S_02881C_VS_OUT_CCDIST0_VEC_ENA(x) (((x) & 0x1) << 22)
+#define G_02881C_VS_OUT_CCDIST0_VEC_ENA(x) (((x) >> 22) & 0x1)
+#define C_02881C_VS_OUT_CCDIST0_VEC_ENA 0xFFBFFFFF
+#define S_02881C_VS_OUT_CCDIST1_VEC_ENA(x) (((x) & 0x1) << 23)
+#define G_02881C_VS_OUT_CCDIST1_VEC_ENA(x) (((x) >> 23) & 0x1)
+#define C_02881C_VS_OUT_CCDIST1_VEC_ENA 0xFF7FFFFF
+#define S_02881C_VS_OUT_MISC_SIDE_BUS_ENA(x) (((x) & 0x1) << 24)
+#define G_02881C_VS_OUT_MISC_SIDE_BUS_ENA(x) (((x) >> 24) & 0x1)
+#define C_02881C_VS_OUT_MISC_SIDE_BUS_ENA 0xFEFFFFFF
+#define S_02881C_USE_VTX_GS_CUT_FLAG(x) (((x) & 0x1) << 25)
+#define G_02881C_USE_VTX_GS_CUT_FLAG(x) (((x) >> 25) & 0x1)
+#define C_02881C_USE_VTX_GS_CUT_FLAG 0xFDFFFFFF
+#define R_028820_PA_CL_NANINF_CNTL 0x028820
+#define S_028820_VTE_XY_INF_DISCARD(x) (((x) & 0x1) << 0)
+#define G_028820_VTE_XY_INF_DISCARD(x) (((x) >> 0) & 0x1)
+#define C_028820_VTE_XY_INF_DISCARD 0xFFFFFFFE
+#define S_028820_VTE_Z_INF_DISCARD(x) (((x) & 0x1) << 1)
+#define G_028820_VTE_Z_INF_DISCARD(x) (((x) >> 1) & 0x1)
+#define C_028820_VTE_Z_INF_DISCARD 0xFFFFFFFD
+#define S_028820_VTE_W_INF_DISCARD(x) (((x) & 0x1) << 2)
+#define G_028820_VTE_W_INF_DISCARD(x) (((x) >> 2) & 0x1)
+#define C_028820_VTE_W_INF_DISCARD 0xFFFFFFFB
+#define S_028820_VTE_0XNANINF_IS_0(x) (((x) & 0x1) << 3)
+#define G_028820_VTE_0XNANINF_IS_0(x) (((x) >> 3) & 0x1)
+#define C_028820_VTE_0XNANINF_IS_0 0xFFFFFFF7
+#define S_028820_VTE_XY_NAN_RETAIN(x) (((x) & 0x1) << 4)
+#define G_028820_VTE_XY_NAN_RETAIN(x) (((x) >> 4) & 0x1)
+#define C_028820_VTE_XY_NAN_RETAIN 0xFFFFFFEF
+#define S_028820_VTE_Z_NAN_RETAIN(x) (((x) & 0x1) << 5)
+#define G_028820_VTE_Z_NAN_RETAIN(x) (((x) >> 5) & 0x1)
+#define C_028820_VTE_Z_NAN_RETAIN 0xFFFFFFDF
+#define S_028820_VTE_W_NAN_RETAIN(x) (((x) & 0x1) << 6)
+#define G_028820_VTE_W_NAN_RETAIN(x) (((x) >> 6) & 0x1)
+#define C_028820_VTE_W_NAN_RETAIN 0xFFFFFFBF
+#define S_028820_VTE_W_RECIP_NAN_IS_0(x) (((x) & 0x1) << 7)
+#define G_028820_VTE_W_RECIP_NAN_IS_0(x) (((x) >> 7) & 0x1)
+#define C_028820_VTE_W_RECIP_NAN_IS_0 0xFFFFFF7F
+#define S_028820_VS_XY_NAN_TO_INF(x) (((x) & 0x1) << 8)
+#define G_028820_VS_XY_NAN_TO_INF(x) (((x) >> 8) & 0x1)
+#define C_028820_VS_XY_NAN_TO_INF 0xFFFFFEFF
+#define S_028820_VS_XY_INF_RETAIN(x) (((x) & 0x1) << 9)
+#define G_028820_VS_XY_INF_RETAIN(x) (((x) >> 9) & 0x1)
+#define C_028820_VS_XY_INF_RETAIN 0xFFFFFDFF
+#define S_028820_VS_Z_NAN_TO_INF(x) (((x) & 0x1) << 10)
+#define G_028820_VS_Z_NAN_TO_INF(x) (((x) >> 10) & 0x1)
+#define C_028820_VS_Z_NAN_TO_INF 0xFFFFFBFF
+#define S_028820_VS_Z_INF_RETAIN(x) (((x) & 0x1) << 11)
+#define G_028820_VS_Z_INF_RETAIN(x) (((x) >> 11) & 0x1)
+#define C_028820_VS_Z_INF_RETAIN 0xFFFFF7FF
+#define S_028820_VS_W_NAN_TO_INF(x) (((x) & 0x1) << 12)
+#define G_028820_VS_W_NAN_TO_INF(x) (((x) >> 12) & 0x1)
+#define C_028820_VS_W_NAN_TO_INF 0xFFFFEFFF
+#define S_028820_VS_W_INF_RETAIN(x) (((x) & 0x1) << 13)
+#define G_028820_VS_W_INF_RETAIN(x) (((x) >> 13) & 0x1)
+#define C_028820_VS_W_INF_RETAIN 0xFFFFDFFF
+#define S_028820_VS_CLIP_DIST_INF_DISCARD(x) (((x) & 0x1) << 14)
+#define G_028820_VS_CLIP_DIST_INF_DISCARD(x) (((x) >> 14) & 0x1)
+#define C_028820_VS_CLIP_DIST_INF_DISCARD 0xFFFFBFFF
+#define S_028820_VTE_NO_OUTPUT_NEG_0(x) (((x) & 0x1) << 20)
+#define G_028820_VTE_NO_OUTPUT_NEG_0(x) (((x) >> 20) & 0x1)
+#define C_028820_VTE_NO_OUTPUT_NEG_0 0xFFEFFFFF
+#define R_028824_PA_SU_LINE_STIPPLE_CNTL 0x028824
+#define S_028824_LINE_STIPPLE_RESET(x) (((x) & 0x03) << 0)
+#define G_028824_LINE_STIPPLE_RESET(x) (((x) >> 0) & 0x03)
+#define C_028824_LINE_STIPPLE_RESET 0xFFFFFFFC
+#define S_028824_EXPAND_FULL_LENGTH(x) (((x) & 0x1) << 2)
+#define G_028824_EXPAND_FULL_LENGTH(x) (((x) >> 2) & 0x1)
+#define C_028824_EXPAND_FULL_LENGTH 0xFFFFFFFB
+#define S_028824_FRACTIONAL_ACCUM(x) (((x) & 0x1) << 3)
+#define G_028824_FRACTIONAL_ACCUM(x) (((x) >> 3) & 0x1)
+#define C_028824_FRACTIONAL_ACCUM 0xFFFFFFF7
+#define S_028824_DIAMOND_ADJUST(x) (((x) & 0x1) << 4)
+#define G_028824_DIAMOND_ADJUST(x) (((x) >> 4) & 0x1)
+#define C_028824_DIAMOND_ADJUST 0xFFFFFFEF
+#define R_028828_PA_SU_LINE_STIPPLE_SCALE 0x028828
+#define R_02882C_PA_SU_PRIM_FILTER_CNTL 0x02882C
+#define S_02882C_TRIANGLE_FILTER_DISABLE(x) (((x) & 0x1) << 0)
+#define G_02882C_TRIANGLE_FILTER_DISABLE(x) (((x) >> 0) & 0x1)
+#define C_02882C_TRIANGLE_FILTER_DISABLE 0xFFFFFFFE
+#define S_02882C_LINE_FILTER_DISABLE(x) (((x) & 0x1) << 1)
+#define G_02882C_LINE_FILTER_DISABLE(x) (((x) >> 1) & 0x1)
+#define C_02882C_LINE_FILTER_DISABLE 0xFFFFFFFD
+#define S_02882C_POINT_FILTER_DISABLE(x) (((x) & 0x1) << 2)
+#define G_02882C_POINT_FILTER_DISABLE(x) (((x) >> 2) & 0x1)
+#define C_02882C_POINT_FILTER_DISABLE 0xFFFFFFFB
+#define S_02882C_RECTANGLE_FILTER_DISABLE(x) (((x) & 0x1) << 3)
+#define G_02882C_RECTANGLE_FILTER_DISABLE(x) (((x) >> 3) & 0x1)
+#define C_02882C_RECTANGLE_FILTER_DISABLE 0xFFFFFFF7
+#define S_02882C_TRIANGLE_EXPAND_ENA(x) (((x) & 0x1) << 4)
+#define G_02882C_TRIANGLE_EXPAND_ENA(x) (((x) >> 4) & 0x1)
+#define C_02882C_TRIANGLE_EXPAND_ENA 0xFFFFFFEF
+#define S_02882C_LINE_EXPAND_ENA(x) (((x) & 0x1) << 5)
+#define G_02882C_LINE_EXPAND_ENA(x) (((x) >> 5) & 0x1)
+#define C_02882C_LINE_EXPAND_ENA 0xFFFFFFDF
+#define S_02882C_POINT_EXPAND_ENA(x) (((x) & 0x1) << 6)
+#define G_02882C_POINT_EXPAND_ENA(x) (((x) >> 6) & 0x1)
+#define C_02882C_POINT_EXPAND_ENA 0xFFFFFFBF
+#define S_02882C_RECTANGLE_EXPAND_ENA(x) (((x) & 0x1) << 7)
+#define G_02882C_RECTANGLE_EXPAND_ENA(x) (((x) >> 7) & 0x1)
+#define C_02882C_RECTANGLE_EXPAND_ENA 0xFFFFFF7F
+#define S_02882C_PRIM_EXPAND_CONSTANT(x) (((x) & 0xFF) << 8)
+#define G_02882C_PRIM_EXPAND_CONSTANT(x) (((x) >> 8) & 0xFF)
+#define C_02882C_PRIM_EXPAND_CONSTANT 0xFFFF00FF
+#define R_028A00_PA_SU_POINT_SIZE 0x028A00
+#define S_028A00_HEIGHT(x) (((x) & 0xFFFF) << 0)
+#define G_028A00_HEIGHT(x) (((x) >> 0) & 0xFFFF)
+#define C_028A00_HEIGHT 0xFFFF0000
+#define S_028A00_WIDTH(x) (((x) & 0xFFFF) << 16)
+#define G_028A00_WIDTH(x) (((x) >> 16) & 0xFFFF)
+#define C_028A00_WIDTH 0x0000FFFF
+#define R_028A04_PA_SU_POINT_MINMAX 0x028A04
+#define S_028A04_MIN_SIZE(x) (((x) & 0xFFFF) << 0)
+#define G_028A04_MIN_SIZE(x) (((x) >> 0) & 0xFFFF)
+#define C_028A04_MIN_SIZE 0xFFFF0000
+#define S_028A04_MAX_SIZE(x) (((x) & 0xFFFF) << 16)
+#define G_028A04_MAX_SIZE(x) (((x) >> 16) & 0xFFFF)
+#define C_028A04_MAX_SIZE 0x0000FFFF
+#define R_028A08_PA_SU_LINE_CNTL 0x028A08
+#define S_028A08_WIDTH(x) (((x) & 0xFFFF) << 0)
+#define G_028A08_WIDTH(x) (((x) >> 0) & 0xFFFF)
+#define C_028A08_WIDTH 0xFFFF0000
+#define R_028A0C_PA_SC_LINE_STIPPLE 0x028A0C
+#define S_028A0C_LINE_PATTERN(x) (((x) & 0xFFFF) << 0)
+#define G_028A0C_LINE_PATTERN(x) (((x) >> 0) & 0xFFFF)
+#define C_028A0C_LINE_PATTERN 0xFFFF0000
+#define S_028A0C_REPEAT_COUNT(x) (((x) & 0xFF) << 16)
+#define G_028A0C_REPEAT_COUNT(x) (((x) >> 16) & 0xFF)
+#define C_028A0C_REPEAT_COUNT 0xFF00FFFF
+#define S_028A0C_PATTERN_BIT_ORDER(x) (((x) & 0x1) << 28)
+#define G_028A0C_PATTERN_BIT_ORDER(x) (((x) >> 28) & 0x1)
+#define C_028A0C_PATTERN_BIT_ORDER 0xEFFFFFFF
+#define S_028A0C_AUTO_RESET_CNTL(x) (((x) & 0x03) << 29)
+#define G_028A0C_AUTO_RESET_CNTL(x) (((x) >> 29) & 0x03)
+#define C_028A0C_AUTO_RESET_CNTL 0x9FFFFFFF
+#define R_028A10_VGT_OUTPUT_PATH_CNTL 0x028A10
+#define S_028A10_PATH_SELECT(x) (((x) & 0x07) << 0)
+#define G_028A10_PATH_SELECT(x) (((x) >> 0) & 0x07)
+#define C_028A10_PATH_SELECT 0xFFFFFFF8
+#define V_028A10_VGT_OUTPATH_VTX_REUSE 0x00
+#define V_028A10_VGT_OUTPATH_TESS_EN 0x01
+#define V_028A10_VGT_OUTPATH_PASSTHRU 0x02
+#define V_028A10_VGT_OUTPATH_GS_BLOCK 0x03
+#define V_028A10_VGT_OUTPATH_HS_BLOCK 0x04
+#define R_028A14_VGT_HOS_CNTL 0x028A14
+#define S_028A14_TESS_MODE(x) (((x) & 0x03) << 0)
+#define G_028A14_TESS_MODE(x) (((x) >> 0) & 0x03)
+#define C_028A14_TESS_MODE 0xFFFFFFFC
+#define R_028A18_VGT_HOS_MAX_TESS_LEVEL 0x028A18
+#define R_028A1C_VGT_HOS_MIN_TESS_LEVEL 0x028A1C
+#define R_028A20_VGT_HOS_REUSE_DEPTH 0x028A20
+#define S_028A20_REUSE_DEPTH(x) (((x) & 0xFF) << 0)
+#define G_028A20_REUSE_DEPTH(x) (((x) >> 0) & 0xFF)
+#define C_028A20_REUSE_DEPTH 0xFFFFFF00
+#define R_028A24_VGT_GROUP_PRIM_TYPE 0x028A24
+#define S_028A24_PRIM_TYPE(x) (((x) & 0x1F) << 0)
+#define G_028A24_PRIM_TYPE(x) (((x) >> 0) & 0x1F)
+#define C_028A24_PRIM_TYPE 0xFFFFFFE0
+#define V_028A24_VGT_GRP_3D_POINT 0x00
+#define V_028A24_VGT_GRP_3D_LINE 0x01
+#define V_028A24_VGT_GRP_3D_TRI 0x02
+#define V_028A24_VGT_GRP_3D_RECT 0x03
+#define V_028A24_VGT_GRP_3D_QUAD 0x04
+#define V_028A24_VGT_GRP_2D_COPY_RECT_V0 0x05
+#define V_028A24_VGT_GRP_2D_COPY_RECT_V1 0x06
+#define V_028A24_VGT_GRP_2D_COPY_RECT_V2 0x07
+#define V_028A24_VGT_GRP_2D_COPY_RECT_V3 0x08
+#define V_028A24_VGT_GRP_2D_FILL_RECT 0x09
+#define V_028A24_VGT_GRP_2D_LINE 0x0A
+#define V_028A24_VGT_GRP_2D_TRI 0x0B
+#define V_028A24_VGT_GRP_PRIM_INDEX_LINE 0x0C
+#define V_028A24_VGT_GRP_PRIM_INDEX_TRI 0x0D
+#define V_028A24_VGT_GRP_PRIM_INDEX_QUAD 0x0E
+#define V_028A24_VGT_GRP_3D_LINE_ADJ 0x0F
+#define V_028A24_VGT_GRP_3D_TRI_ADJ 0x10
+#define V_028A24_VGT_GRP_3D_PATCH 0x11
+#define S_028A24_RETAIN_ORDER(x) (((x) & 0x1) << 14)
+#define G_028A24_RETAIN_ORDER(x) (((x) >> 14) & 0x1)
+#define C_028A24_RETAIN_ORDER 0xFFFFBFFF
+#define S_028A24_RETAIN_QUADS(x) (((x) & 0x1) << 15)
+#define G_028A24_RETAIN_QUADS(x) (((x) >> 15) & 0x1)
+#define C_028A24_RETAIN_QUADS 0xFFFF7FFF
+#define S_028A24_PRIM_ORDER(x) (((x) & 0x07) << 16)
+#define G_028A24_PRIM_ORDER(x) (((x) >> 16) & 0x07)
+#define C_028A24_PRIM_ORDER 0xFFF8FFFF
+#define V_028A24_VGT_GRP_LIST 0x00
+#define V_028A24_VGT_GRP_STRIP 0x01
+#define V_028A24_VGT_GRP_FAN 0x02
+#define V_028A24_VGT_GRP_LOOP 0x03
+#define V_028A24_VGT_GRP_POLYGON 0x04
+#define R_028A28_VGT_GROUP_FIRST_DECR 0x028A28
+#define S_028A28_FIRST_DECR(x) (((x) & 0x0F) << 0)
+#define G_028A28_FIRST_DECR(x) (((x) >> 0) & 0x0F)
+#define C_028A28_FIRST_DECR 0xFFFFFFF0
+#define R_028A2C_VGT_GROUP_DECR 0x028A2C
+#define S_028A2C_DECR(x) (((x) & 0x0F) << 0)
+#define G_028A2C_DECR(x) (((x) >> 0) & 0x0F)
+#define C_028A2C_DECR 0xFFFFFFF0
+#define R_028A30_VGT_GROUP_VECT_0_CNTL 0x028A30
+#define S_028A30_COMP_X_EN(x) (((x) & 0x1) << 0)
+#define G_028A30_COMP_X_EN(x) (((x) >> 0) & 0x1)
+#define C_028A30_COMP_X_EN 0xFFFFFFFE
+#define S_028A30_COMP_Y_EN(x) (((x) & 0x1) << 1)
+#define G_028A30_COMP_Y_EN(x) (((x) >> 1) & 0x1)
+#define C_028A30_COMP_Y_EN 0xFFFFFFFD
+#define S_028A30_COMP_Z_EN(x) (((x) & 0x1) << 2)
+#define G_028A30_COMP_Z_EN(x) (((x) >> 2) & 0x1)
+#define C_028A30_COMP_Z_EN 0xFFFFFFFB
+#define S_028A30_COMP_W_EN(x) (((x) & 0x1) << 3)
+#define G_028A30_COMP_W_EN(x) (((x) >> 3) & 0x1)
+#define C_028A30_COMP_W_EN 0xFFFFFFF7
+#define S_028A30_STRIDE(x) (((x) & 0xFF) << 8)
+#define G_028A30_STRIDE(x) (((x) >> 8) & 0xFF)
+#define C_028A30_STRIDE 0xFFFF00FF
+#define S_028A30_SHIFT(x) (((x) & 0xFF) << 16)
+#define G_028A30_SHIFT(x) (((x) >> 16) & 0xFF)
+#define C_028A30_SHIFT 0xFF00FFFF
+#define R_028A34_VGT_GROUP_VECT_1_CNTL 0x028A34
+#define S_028A34_COMP_X_EN(x) (((x) & 0x1) << 0)
+#define G_028A34_COMP_X_EN(x) (((x) >> 0) & 0x1)
+#define C_028A34_COMP_X_EN 0xFFFFFFFE
+#define S_028A34_COMP_Y_EN(x) (((x) & 0x1) << 1)
+#define G_028A34_COMP_Y_EN(x) (((x) >> 1) & 0x1)
+#define C_028A34_COMP_Y_EN 0xFFFFFFFD
+#define S_028A34_COMP_Z_EN(x) (((x) & 0x1) << 2)
+#define G_028A34_COMP_Z_EN(x) (((x) >> 2) & 0x1)
+#define C_028A34_COMP_Z_EN 0xFFFFFFFB
+#define S_028A34_COMP_W_EN(x) (((x) & 0x1) << 3)
+#define G_028A34_COMP_W_EN(x) (((x) >> 3) & 0x1)
+#define C_028A34_COMP_W_EN 0xFFFFFFF7
+#define S_028A34_STRIDE(x) (((x) & 0xFF) << 8)
+#define G_028A34_STRIDE(x) (((x) >> 8) & 0xFF)
+#define C_028A34_STRIDE 0xFFFF00FF
+#define S_028A34_SHIFT(x) (((x) & 0xFF) << 16)
+#define G_028A34_SHIFT(x) (((x) >> 16) & 0xFF)
+#define C_028A34_SHIFT 0xFF00FFFF
+#define R_028A38_VGT_GROUP_VECT_0_FMT_CNTL 0x028A38
+#define S_028A38_X_CONV(x) (((x) & 0x0F) << 0)
+#define G_028A38_X_CONV(x) (((x) >> 0) & 0x0F)
+#define C_028A38_X_CONV 0xFFFFFFF0
+#define V_028A38_VGT_GRP_INDEX_16 0x00
+#define V_028A38_VGT_GRP_INDEX_32 0x01
+#define V_028A38_VGT_GRP_UINT_16 0x02
+#define V_028A38_VGT_GRP_UINT_32 0x03
+#define V_028A38_VGT_GRP_SINT_16 0x04
+#define V_028A38_VGT_GRP_SINT_32 0x05
+#define V_028A38_VGT_GRP_FLOAT_32 0x06
+#define V_028A38_VGT_GRP_AUTO_PRIM 0x07
+#define V_028A38_VGT_GRP_FIX_1_23_TO_FLOAT 0x08
+#define S_028A38_X_OFFSET(x) (((x) & 0x0F) << 4)
+#define G_028A38_X_OFFSET(x) (((x) >> 4) & 0x0F)
+#define C_028A38_X_OFFSET 0xFFFFFF0F
+#define S_028A38_Y_CONV(x) (((x) & 0x0F) << 8)
+#define G_028A38_Y_CONV(x) (((x) >> 8) & 0x0F)
+#define C_028A38_Y_CONV 0xFFFFF0FF
+#define V_028A38_VGT_GRP_INDEX_16 0x00
+#define V_028A38_VGT_GRP_INDEX_32 0x01
+#define V_028A38_VGT_GRP_UINT_16 0x02
+#define V_028A38_VGT_GRP_UINT_32 0x03
+#define V_028A38_VGT_GRP_SINT_16 0x04
+#define V_028A38_VGT_GRP_SINT_32 0x05
+#define V_028A38_VGT_GRP_FLOAT_32 0x06
+#define V_028A38_VGT_GRP_AUTO_PRIM 0x07
+#define V_028A38_VGT_GRP_FIX_1_23_TO_FLOAT 0x08
+#define S_028A38_Y_OFFSET(x) (((x) & 0x0F) << 12)
+#define G_028A38_Y_OFFSET(x) (((x) >> 12) & 0x0F)
+#define C_028A38_Y_OFFSET 0xFFFF0FFF
+#define S_028A38_Z_CONV(x) (((x) & 0x0F) << 16)
+#define G_028A38_Z_CONV(x) (((x) >> 16) & 0x0F)
+#define C_028A38_Z_CONV 0xFFF0FFFF
+#define V_028A38_VGT_GRP_INDEX_16 0x00
+#define V_028A38_VGT_GRP_INDEX_32 0x01
+#define V_028A38_VGT_GRP_UINT_16 0x02
+#define V_028A38_VGT_GRP_UINT_32 0x03
+#define V_028A38_VGT_GRP_SINT_16 0x04
+#define V_028A38_VGT_GRP_SINT_32 0x05
+#define V_028A38_VGT_GRP_FLOAT_32 0x06
+#define V_028A38_VGT_GRP_AUTO_PRIM 0x07
+#define V_028A38_VGT_GRP_FIX_1_23_TO_FLOAT 0x08
+#define S_028A38_Z_OFFSET(x) (((x) & 0x0F) << 20)
+#define G_028A38_Z_OFFSET(x) (((x) >> 20) & 0x0F)
+#define C_028A38_Z_OFFSET 0xFF0FFFFF
+#define S_028A38_W_CONV(x) (((x) & 0x0F) << 24)
+#define G_028A38_W_CONV(x) (((x) >> 24) & 0x0F)
+#define C_028A38_W_CONV 0xF0FFFFFF
+#define V_028A38_VGT_GRP_INDEX_16 0x00
+#define V_028A38_VGT_GRP_INDEX_32 0x01
+#define V_028A38_VGT_GRP_UINT_16 0x02
+#define V_028A38_VGT_GRP_UINT_32 0x03
+#define V_028A38_VGT_GRP_SINT_16 0x04
+#define V_028A38_VGT_GRP_SINT_32 0x05
+#define V_028A38_VGT_GRP_FLOAT_32 0x06
+#define V_028A38_VGT_GRP_AUTO_PRIM 0x07
+#define V_028A38_VGT_GRP_FIX_1_23_TO_FLOAT 0x08
+#define S_028A38_W_OFFSET(x) (((x) & 0x0F) << 28)
+#define G_028A38_W_OFFSET(x) (((x) >> 28) & 0x0F)
+#define C_028A38_W_OFFSET 0x0FFFFFFF
+#define R_028A3C_VGT_GROUP_VECT_1_FMT_CNTL 0x028A3C
+#define S_028A3C_X_CONV(x) (((x) & 0x0F) << 0)
+#define G_028A3C_X_CONV(x) (((x) >> 0) & 0x0F)
+#define C_028A3C_X_CONV 0xFFFFFFF0
+#define V_028A3C_VGT_GRP_INDEX_16 0x00
+#define V_028A3C_VGT_GRP_INDEX_32 0x01
+#define V_028A3C_VGT_GRP_UINT_16 0x02
+#define V_028A3C_VGT_GRP_UINT_32 0x03
+#define V_028A3C_VGT_GRP_SINT_16 0x04
+#define V_028A3C_VGT_GRP_SINT_32 0x05
+#define V_028A3C_VGT_GRP_FLOAT_32 0x06
+#define V_028A3C_VGT_GRP_AUTO_PRIM 0x07
+#define V_028A3C_VGT_GRP_FIX_1_23_TO_FLOAT 0x08
+#define S_028A3C_X_OFFSET(x) (((x) & 0x0F) << 4)
+#define G_028A3C_X_OFFSET(x) (((x) >> 4) & 0x0F)
+#define C_028A3C_X_OFFSET 0xFFFFFF0F
+#define S_028A3C_Y_CONV(x) (((x) & 0x0F) << 8)
+#define G_028A3C_Y_CONV(x) (((x) >> 8) & 0x0F)
+#define C_028A3C_Y_CONV 0xFFFFF0FF
+#define V_028A3C_VGT_GRP_INDEX_16 0x00
+#define V_028A3C_VGT_GRP_INDEX_32 0x01
+#define V_028A3C_VGT_GRP_UINT_16 0x02
+#define V_028A3C_VGT_GRP_UINT_32 0x03
+#define V_028A3C_VGT_GRP_SINT_16 0x04
+#define V_028A3C_VGT_GRP_SINT_32 0x05
+#define V_028A3C_VGT_GRP_FLOAT_32 0x06
+#define V_028A3C_VGT_GRP_AUTO_PRIM 0x07
+#define V_028A3C_VGT_GRP_FIX_1_23_TO_FLOAT 0x08
+#define S_028A3C_Y_OFFSET(x) (((x) & 0x0F) << 12)
+#define G_028A3C_Y_OFFSET(x) (((x) >> 12) & 0x0F)
+#define C_028A3C_Y_OFFSET 0xFFFF0FFF
+#define S_028A3C_Z_CONV(x) (((x) & 0x0F) << 16)
+#define G_028A3C_Z_CONV(x) (((x) >> 16) & 0x0F)
+#define C_028A3C_Z_CONV 0xFFF0FFFF
+#define V_028A3C_VGT_GRP_INDEX_16 0x00
+#define V_028A3C_VGT_GRP_INDEX_32 0x01
+#define V_028A3C_VGT_GRP_UINT_16 0x02
+#define V_028A3C_VGT_GRP_UINT_32 0x03
+#define V_028A3C_VGT_GRP_SINT_16 0x04
+#define V_028A3C_VGT_GRP_SINT_32 0x05
+#define V_028A3C_VGT_GRP_FLOAT_32 0x06
+#define V_028A3C_VGT_GRP_AUTO_PRIM 0x07
+#define V_028A3C_VGT_GRP_FIX_1_23_TO_FLOAT 0x08
+#define S_028A3C_Z_OFFSET(x) (((x) & 0x0F) << 20)
+#define G_028A3C_Z_OFFSET(x) (((x) >> 20) & 0x0F)
+#define C_028A3C_Z_OFFSET 0xFF0FFFFF
+#define S_028A3C_W_CONV(x) (((x) & 0x0F) << 24)
+#define G_028A3C_W_CONV(x) (((x) >> 24) & 0x0F)
+#define C_028A3C_W_CONV 0xF0FFFFFF
+#define V_028A3C_VGT_GRP_INDEX_16 0x00
+#define V_028A3C_VGT_GRP_INDEX_32 0x01
+#define V_028A3C_VGT_GRP_UINT_16 0x02
+#define V_028A3C_VGT_GRP_UINT_32 0x03
+#define V_028A3C_VGT_GRP_SINT_16 0x04
+#define V_028A3C_VGT_GRP_SINT_32 0x05
+#define V_028A3C_VGT_GRP_FLOAT_32 0x06
+#define V_028A3C_VGT_GRP_AUTO_PRIM 0x07
+#define V_028A3C_VGT_GRP_FIX_1_23_TO_FLOAT 0x08
+#define S_028A3C_W_OFFSET(x) (((x) & 0x0F) << 28)
+#define G_028A3C_W_OFFSET(x) (((x) >> 28) & 0x0F)
+#define C_028A3C_W_OFFSET 0x0FFFFFFF
+#define R_028A40_VGT_GS_MODE 0x028A40
+#define S_028A40_MODE(x) (((x) & 0x07) << 0)
+#define G_028A40_MODE(x) (((x) >> 0) & 0x07)
+#define C_028A40_MODE 0xFFFFFFF8
+#define V_028A40_GS_OFF 0x00
+#define V_028A40_GS_SCENARIO_A 0x01
+#define V_028A40_GS_SCENARIO_B 0x02
+#define V_028A40_GS_SCENARIO_G 0x03
+#define V_028A40_GS_SCENARIO_C 0x04
+#define V_028A40_SPRITE_EN 0x05
+#define S_028A40_CUT_MODE(x) (((x) & 0x03) << 4)
+#define G_028A40_CUT_MODE(x) (((x) >> 4) & 0x03)
+#define C_028A40_CUT_MODE 0xFFFFFFCF
+#define V_028A40_GS_CUT_1024 0x00
+#define V_028A40_GS_CUT_512 0x01
+#define V_028A40_GS_CUT_256 0x02
+#define V_028A40_GS_CUT_128 0x03
+#define S_028A40_GS_C_PACK_EN(x) (((x) & 0x1) << 11)
+#define G_028A40_GS_C_PACK_EN(x) (((x) >> 11) & 0x1)
+#define C_028A40_GS_C_PACK_EN 0xFFFFF7FF
+#define S_028A40_ES_PASSTHRU(x) (((x) & 0x1) << 13)
+#define G_028A40_ES_PASSTHRU(x) (((x) >> 13) & 0x1)
+#define C_028A40_ES_PASSTHRU 0xFFFFDFFF
+#define S_028A40_COMPUTE_MODE(x) (((x) & 0x1) << 14)
+#define G_028A40_COMPUTE_MODE(x) (((x) >> 14) & 0x1)
+#define C_028A40_COMPUTE_MODE 0xFFFFBFFF
+#define S_028A40_FAST_COMPUTE_MODE(x) (((x) & 0x1) << 15)
+#define G_028A40_FAST_COMPUTE_MODE(x) (((x) >> 15) & 0x1)
+#define C_028A40_FAST_COMPUTE_MODE 0xFFFF7FFF
+#define S_028A40_ELEMENT_INFO_EN(x) (((x) & 0x1) << 16)
+#define G_028A40_ELEMENT_INFO_EN(x) (((x) >> 16) & 0x1)
+#define C_028A40_ELEMENT_INFO_EN 0xFFFEFFFF
+#define S_028A40_PARTIAL_THD_AT_EOI(x) (((x) & 0x1) << 17)
+#define G_028A40_PARTIAL_THD_AT_EOI(x) (((x) >> 17) & 0x1)
+#define C_028A40_PARTIAL_THD_AT_EOI 0xFFFDFFFF
+#define S_028A40_SUPPRESS_CUTS(x) (((x) & 0x1) << 18)
+#define G_028A40_SUPPRESS_CUTS(x) (((x) >> 18) & 0x1)
+#define C_028A40_SUPPRESS_CUTS 0xFFFBFFFF
+#define S_028A40_ES_WRITE_OPTIMIZE(x) (((x) & 0x1) << 19)
+#define G_028A40_ES_WRITE_OPTIMIZE(x) (((x) >> 19) & 0x1)
+#define C_028A40_ES_WRITE_OPTIMIZE 0xFFF7FFFF
+#define S_028A40_GS_WRITE_OPTIMIZE(x) (((x) & 0x1) << 20)
+#define G_028A40_GS_WRITE_OPTIMIZE(x) (((x) >> 20) & 0x1)
+#define C_028A40_GS_WRITE_OPTIMIZE 0xFFEFFFFF
+#define R_028A48_PA_SC_MODE_CNTL_0 0x028A48
+#define S_028A48_MSAA_ENABLE(x) (((x) & 0x1) << 0)
+#define G_028A48_MSAA_ENABLE(x) (((x) >> 0) & 0x1)
+#define C_028A48_MSAA_ENABLE 0xFFFFFFFE
+#define S_028A48_VPORT_SCISSOR_ENABLE(x) (((x) & 0x1) << 1)
+#define G_028A48_VPORT_SCISSOR_ENABLE(x) (((x) >> 1) & 0x1)
+#define C_028A48_VPORT_SCISSOR_ENABLE 0xFFFFFFFD
+#define S_028A48_LINE_STIPPLE_ENABLE(x) (((x) & 0x1) << 2)
+#define G_028A48_LINE_STIPPLE_ENABLE(x) (((x) >> 2) & 0x1)
+#define C_028A48_LINE_STIPPLE_ENABLE 0xFFFFFFFB
+#define S_028A48_SEND_UNLIT_STILES_TO_PKR(x) (((x) & 0x1) << 3)
+#define G_028A48_SEND_UNLIT_STILES_TO_PKR(x) (((x) >> 3) & 0x1)
+#define C_028A48_SEND_UNLIT_STILES_TO_PKR 0xFFFFFFF7
+#define R_028A4C_PA_SC_MODE_CNTL_1 0x028A4C
+#define S_028A4C_WALK_SIZE(x) (((x) & 0x1) << 0)
+#define G_028A4C_WALK_SIZE(x) (((x) >> 0) & 0x1)
+#define C_028A4C_WALK_SIZE 0xFFFFFFFE
+#define S_028A4C_WALK_ALIGNMENT(x) (((x) & 0x1) << 1)
+#define G_028A4C_WALK_ALIGNMENT(x) (((x) >> 1) & 0x1)
+#define C_028A4C_WALK_ALIGNMENT 0xFFFFFFFD
+#define S_028A4C_WALK_ALIGN8_PRIM_FITS_ST(x) (((x) & 0x1) << 2)
+#define G_028A4C_WALK_ALIGN8_PRIM_FITS_ST(x) (((x) >> 2) & 0x1)
+#define C_028A4C_WALK_ALIGN8_PRIM_FITS_ST 0xFFFFFFFB
+#define S_028A4C_WALK_FENCE_ENABLE(x) (((x) & 0x1) << 3)
+#define G_028A4C_WALK_FENCE_ENABLE(x) (((x) >> 3) & 0x1)
+#define C_028A4C_WALK_FENCE_ENABLE 0xFFFFFFF7
+#define S_028A4C_WALK_FENCE_SIZE(x) (((x) & 0x07) << 4)
+#define G_028A4C_WALK_FENCE_SIZE(x) (((x) >> 4) & 0x07)
+#define C_028A4C_WALK_FENCE_SIZE 0xFFFFFF8F
+#define S_028A4C_SUPERTILE_WALK_ORDER_ENABLE(x) (((x) & 0x1) << 7)
+#define G_028A4C_SUPERTILE_WALK_ORDER_ENABLE(x) (((x) >> 7) & 0x1)
+#define C_028A4C_SUPERTILE_WALK_ORDER_ENABLE 0xFFFFFF7F
+#define S_028A4C_TILE_WALK_ORDER_ENABLE(x) (((x) & 0x1) << 8)
+#define G_028A4C_TILE_WALK_ORDER_ENABLE(x) (((x) >> 8) & 0x1)
+#define C_028A4C_TILE_WALK_ORDER_ENABLE 0xFFFFFEFF
+#define S_028A4C_TILE_COVER_DISABLE(x) (((x) & 0x1) << 9)
+#define G_028A4C_TILE_COVER_DISABLE(x) (((x) >> 9) & 0x1)
+#define C_028A4C_TILE_COVER_DISABLE 0xFFFFFDFF
+#define S_028A4C_TILE_COVER_NO_SCISSOR(x) (((x) & 0x1) << 10)
+#define G_028A4C_TILE_COVER_NO_SCISSOR(x) (((x) >> 10) & 0x1)
+#define C_028A4C_TILE_COVER_NO_SCISSOR 0xFFFFFBFF
+#define S_028A4C_ZMM_LINE_EXTENT(x) (((x) & 0x1) << 11)
+#define G_028A4C_ZMM_LINE_EXTENT(x) (((x) >> 11) & 0x1)
+#define C_028A4C_ZMM_LINE_EXTENT 0xFFFFF7FF
+#define S_028A4C_ZMM_LINE_OFFSET(x) (((x) & 0x1) << 12)
+#define G_028A4C_ZMM_LINE_OFFSET(x) (((x) >> 12) & 0x1)
+#define C_028A4C_ZMM_LINE_OFFSET 0xFFFFEFFF
+#define S_028A4C_ZMM_RECT_EXTENT(x) (((x) & 0x1) << 13)
+#define G_028A4C_ZMM_RECT_EXTENT(x) (((x) >> 13) & 0x1)
+#define C_028A4C_ZMM_RECT_EXTENT 0xFFFFDFFF
+#define S_028A4C_KILL_PIX_POST_HI_Z(x) (((x) & 0x1) << 14)
+#define G_028A4C_KILL_PIX_POST_HI_Z(x) (((x) >> 14) & 0x1)
+#define C_028A4C_KILL_PIX_POST_HI_Z 0xFFFFBFFF
+#define S_028A4C_KILL_PIX_POST_DETAIL_MASK(x) (((x) & 0x1) << 15)
+#define G_028A4C_KILL_PIX_POST_DETAIL_MASK(x) (((x) >> 15) & 0x1)
+#define C_028A4C_KILL_PIX_POST_DETAIL_MASK 0xFFFF7FFF
+#define S_028A4C_PS_ITER_SAMPLE(x) (((x) & 0x1) << 16)
+#define G_028A4C_PS_ITER_SAMPLE(x) (((x) >> 16) & 0x1)
+#define C_028A4C_PS_ITER_SAMPLE 0xFFFEFFFF
+#define S_028A4C_MULTI_SHADER_ENGINE_PRIM_DISC(x) (((x) & 0x1) << 17)
+#define G_028A4C_MULTI_SHADER_ENGINE_PRIM_DISC(x) (((x) >> 17) & 0x1)
+#define C_028A4C_MULTI_SHADER_ENGINE_PRIM_DISC 0xFFFDFFFF
+#define S_028A4C_FORCE_EOV_CNTDWN_ENABLE(x) (((x) & 0x1) << 25)
+#define G_028A4C_FORCE_EOV_CNTDWN_ENABLE(x) (((x) >> 25) & 0x1)
+#define C_028A4C_FORCE_EOV_CNTDWN_ENABLE 0xFDFFFFFF
+#define S_028A4C_FORCE_EOV_REZ_ENABLE(x) (((x) & 0x1) << 26)
+#define G_028A4C_FORCE_EOV_REZ_ENABLE(x) (((x) >> 26) & 0x1)
+#define C_028A4C_FORCE_EOV_REZ_ENABLE 0xFBFFFFFF
+#define S_028A4C_OUT_OF_ORDER_PRIMITIVE_ENABLE(x) (((x) & 0x1) << 27)
+#define G_028A4C_OUT_OF_ORDER_PRIMITIVE_ENABLE(x) (((x) >> 27) & 0x1)
+#define C_028A4C_OUT_OF_ORDER_PRIMITIVE_ENABLE 0xF7FFFFFF
+#define S_028A4C_OUT_OF_ORDER_WATER_MARK(x) (((x) & 0x07) << 28)
+#define G_028A4C_OUT_OF_ORDER_WATER_MARK(x) (((x) >> 28) & 0x07)
+#define C_028A4C_OUT_OF_ORDER_WATER_MARK 0x8FFFFFFF
+#define R_028A50_VGT_ENHANCE 0x028A50
+#define R_028A54_VGT_GS_PER_ES 0x028A54
+#define S_028A54_GS_PER_ES(x) (((x) & 0x7FF) << 0)
+#define G_028A54_GS_PER_ES(x) (((x) >> 0) & 0x7FF)
+#define C_028A54_GS_PER_ES 0xFFFFF800
+#define R_028A58_VGT_ES_PER_GS 0x028A58
+#define S_028A58_ES_PER_GS(x) (((x) & 0x7FF) << 0)
+#define G_028A58_ES_PER_GS(x) (((x) >> 0) & 0x7FF)
+#define C_028A58_ES_PER_GS 0xFFFFF800
+#define R_028A5C_VGT_GS_PER_VS 0x028A5C
+#define S_028A5C_GS_PER_VS(x) (((x) & 0x0F) << 0)
+#define G_028A5C_GS_PER_VS(x) (((x) >> 0) & 0x0F)
+#define C_028A5C_GS_PER_VS 0xFFFFFFF0
+#define R_028A60_VGT_GSVS_RING_OFFSET_1 0x028A60
+#define S_028A60_OFFSET(x) (((x) & 0x7FFF) << 0)
+#define G_028A60_OFFSET(x) (((x) >> 0) & 0x7FFF)
+#define C_028A60_OFFSET 0xFFFF8000
+#define R_028A64_VGT_GSVS_RING_OFFSET_2 0x028A64
+#define S_028A64_OFFSET(x) (((x) & 0x7FFF) << 0)
+#define G_028A64_OFFSET(x) (((x) >> 0) & 0x7FFF)
+#define C_028A64_OFFSET 0xFFFF8000
+#define R_028A68_VGT_GSVS_RING_OFFSET_3 0x028A68
+#define S_028A68_OFFSET(x) (((x) & 0x7FFF) << 0)
+#define G_028A68_OFFSET(x) (((x) >> 0) & 0x7FFF)
+#define C_028A68_OFFSET 0xFFFF8000
+#define R_028A6C_VGT_GS_OUT_PRIM_TYPE 0x028A6C
+#define S_028A6C_OUTPRIM_TYPE(x) (((x) & 0x3F) << 0)
+#define G_028A6C_OUTPRIM_TYPE(x) (((x) >> 0) & 0x3F)
+#define C_028A6C_OUTPRIM_TYPE 0xFFFFFFC0
+#define S_028A6C_OUTPRIM_TYPE_1(x) (((x) & 0x3F) << 8)
+#define G_028A6C_OUTPRIM_TYPE_1(x) (((x) >> 8) & 0x3F)
+#define C_028A6C_OUTPRIM_TYPE_1 0xFFFFC0FF
+#define S_028A6C_OUTPRIM_TYPE_2(x) (((x) & 0x3F) << 16)
+#define G_028A6C_OUTPRIM_TYPE_2(x) (((x) >> 16) & 0x3F)
+#define C_028A6C_OUTPRIM_TYPE_2 0xFFC0FFFF
+#define S_028A6C_OUTPRIM_TYPE_3(x) (((x) & 0x3F) << 22)
+#define G_028A6C_OUTPRIM_TYPE_3(x) (((x) >> 22) & 0x3F)
+#define C_028A6C_OUTPRIM_TYPE_3 0xF03FFFFF
+#define S_028A6C_UNIQUE_TYPE_PER_STREAM(x) (((x) & 0x1) << 31)
+#define G_028A6C_UNIQUE_TYPE_PER_STREAM(x) (((x) >> 31) & 0x1)
+#define C_028A6C_UNIQUE_TYPE_PER_STREAM 0x7FFFFFFF
+#define R_028A70_IA_ENHANCE 0x028A70
+#define R_028A74_VGT_DMA_SIZE 0x028A74
+#define R_028A78_VGT_DMA_MAX_SIZE 0x028A78
+#define R_028A7C_VGT_DMA_INDEX_TYPE 0x028A7C
+#define S_028A7C_INDEX_TYPE(x) (((x) & 0x03) << 0)
+#define G_028A7C_INDEX_TYPE(x) (((x) >> 0) & 0x03)
+#define C_028A7C_INDEX_TYPE 0xFFFFFFFC
+#define V_028A7C_VGT_INDEX_16 0x00
+#define V_028A7C_VGT_INDEX_32 0x01
+#define S_028A7C_SWAP_MODE(x) (((x) & 0x03) << 2)
+#define G_028A7C_SWAP_MODE(x) (((x) >> 2) & 0x03)
+#define C_028A7C_SWAP_MODE 0xFFFFFFF3
+#define V_028A7C_VGT_DMA_SWAP_NONE 0x00
+#define V_028A7C_VGT_DMA_SWAP_16_BIT 0x01
+#define V_028A7C_VGT_DMA_SWAP_32_BIT 0x02
+#define V_028A7C_VGT_DMA_SWAP_WORD 0x03
+#define R_028A84_VGT_PRIMITIVEID_EN 0x028A84
+#define S_028A84_PRIMITIVEID_EN(x) (((x) & 0x1) << 0)
+#define G_028A84_PRIMITIVEID_EN(x) (((x) >> 0) & 0x1)
+#define C_028A84_PRIMITIVEID_EN 0xFFFFFFFE
+#define S_028A84_DISABLE_RESET_ON_EOI(x) (((x) & 0x1) << 1)
+#define G_028A84_DISABLE_RESET_ON_EOI(x) (((x) >> 1) & 0x1)
+#define C_028A84_DISABLE_RESET_ON_EOI 0xFFFFFFFD
+#define R_028A88_VGT_DMA_NUM_INSTANCES 0x028A88
+#define R_028A8C_VGT_PRIMITIVEID_RESET 0x028A8C
+#define R_028A90_VGT_EVENT_INITIATOR 0x028A90
+#define S_028A90_EVENT_TYPE(x) (((x) & 0x3F) << 0)
+#define G_028A90_EVENT_TYPE(x) (((x) >> 0) & 0x3F)
+#define C_028A90_EVENT_TYPE 0xFFFFFFC0
+#define V_028A90_SAMPLE_STREAMOUTSTATS1 0x01
+#define V_028A90_SAMPLE_STREAMOUTSTATS2 0x02
+#define V_028A90_SAMPLE_STREAMOUTSTATS3 0x03
+#define V_028A90_CACHE_FLUSH_TS 0x04
+#define V_028A90_CONTEXT_DONE 0x05
+#define V_028A90_CACHE_FLUSH 0x06
+#define V_028A90_CS_PARTIAL_FLUSH 0x07
+#define V_028A90_VGT_STREAMOUT_SYNC 0x08
+#define V_028A90_VGT_STREAMOUT_RESET 0x0A
+#define V_028A90_END_OF_PIPE_INCR_DE 0x0B
+#define V_028A90_END_OF_PIPE_IB_END 0x0C
+#define V_028A90_RST_PIX_CNT 0x0D
+#define V_028A90_VS_PARTIAL_FLUSH 0x0F
+#define V_028A90_PS_PARTIAL_FLUSH 0x10
+#define V_028A90_FLUSH_HS_OUTPUT 0x11
+#define V_028A90_FLUSH_LS_OUTPUT 0x12
+#define V_028A90_CACHE_FLUSH_AND_INV_TS_EVENT 0x14
+#define V_028A90_ZPASS_DONE 0x15
+#define V_028A90_CACHE_FLUSH_AND_INV_EVENT 0x16
+#define V_028A90_PERFCOUNTER_START 0x17
+#define V_028A90_PERFCOUNTER_STOP 0x18
+#define V_028A90_PIPELINESTAT_START 0x19
+#define V_028A90_PIPELINESTAT_STOP 0x1A
+#define V_028A90_PERFCOUNTER_SAMPLE 0x1B
+#define V_028A90_FLUSH_ES_OUTPUT 0x1C
+#define V_028A90_FLUSH_GS_OUTPUT 0x1D
+#define V_028A90_SAMPLE_PIPELINESTAT 0x1E
+#define V_028A90_SO_VGTSTREAMOUT_FLUSH 0x1F
+#define V_028A90_SAMPLE_STREAMOUTSTATS 0x20
+#define V_028A90_RESET_VTX_CNT 0x21
+#define V_028A90_BLOCK_CONTEXT_DONE 0x22
+#define V_028A90_CS_CONTEXT_DONE 0x23
+#define V_028A90_VGT_FLUSH 0x24
+#define V_028A90_SC_SEND_DB_VPZ 0x27
+#define V_028A90_BOTTOM_OF_PIPE_TS 0x28
+#define V_028A90_DB_CACHE_FLUSH_AND_INV 0x2A
+#define V_028A90_FLUSH_AND_INV_DB_DATA_TS 0x2B
+#define V_028A90_FLUSH_AND_INV_DB_META 0x2C
+#define V_028A90_FLUSH_AND_INV_CB_DATA_TS 0x2D
+#define V_028A90_FLUSH_AND_INV_CB_META 0x2E
+#define V_028A90_CS_DONE 0x2F
+#define V_028A90_PS_DONE 0x30
+#define V_028A90_FLUSH_AND_INV_CB_PIXEL_DATA 0x31
+#define V_028A90_THREAD_TRACE_START 0x33
+#define V_028A90_THREAD_TRACE_STOP 0x34
+#define V_028A90_THREAD_TRACE_MARKER 0x35
+#define V_028A90_THREAD_TRACE_FLUSH 0x36
+#define V_028A90_THREAD_TRACE_FINISH 0x37
+#define S_028A90_ADDRESS_HI(x) (((x) & 0x1FF) << 18)
+#define G_028A90_ADDRESS_HI(x) (((x) >> 18) & 0x1FF)
+#define C_028A90_ADDRESS_HI 0xF803FFFF
+#define S_028A90_EXTENDED_EVENT(x) (((x) & 0x1) << 27)
+#define G_028A90_EXTENDED_EVENT(x) (((x) >> 27) & 0x1)
+#define C_028A90_EXTENDED_EVENT 0xF7FFFFFF
+#define R_028A94_VGT_MULTI_PRIM_IB_RESET_EN 0x028A94
+#define S_028A94_RESET_EN(x) (((x) & 0x1) << 0)
+#define G_028A94_RESET_EN(x) (((x) >> 0) & 0x1)
+#define C_028A94_RESET_EN 0xFFFFFFFE
+#define R_028AA0_VGT_INSTANCE_STEP_RATE_0 0x028AA0
+#define R_028AA4_VGT_INSTANCE_STEP_RATE_1 0x028AA4
+#define R_028AA8_IA_MULTI_VGT_PARAM 0x028AA8
+#define S_028AA8_PRIMGROUP_SIZE(x) (((x) & 0xFFFF) << 0)
+#define G_028AA8_PRIMGROUP_SIZE(x) (((x) >> 0) & 0xFFFF)
+#define C_028AA8_PRIMGROUP_SIZE 0xFFFF0000
+#define S_028AA8_PARTIAL_VS_WAVE_ON(x) (((x) & 0x1) << 16)
+#define G_028AA8_PARTIAL_VS_WAVE_ON(x) (((x) >> 16) & 0x1)
+#define C_028AA8_PARTIAL_VS_WAVE_ON 0xFFFEFFFF
+#define S_028AA8_SWITCH_ON_EOP(x) (((x) & 0x1) << 17)
+#define G_028AA8_SWITCH_ON_EOP(x) (((x) >> 17) & 0x1)
+#define C_028AA8_SWITCH_ON_EOP 0xFFFDFFFF
+#define S_028AA8_PARTIAL_ES_WAVE_ON(x) (((x) & 0x1) << 18)
+#define G_028AA8_PARTIAL_ES_WAVE_ON(x) (((x) >> 18) & 0x1)
+#define C_028AA8_PARTIAL_ES_WAVE_ON 0xFFFBFFFF
+#define S_028AA8_SWITCH_ON_EOI(x) (((x) & 0x1) << 19)
+#define G_028AA8_SWITCH_ON_EOI(x) (((x) >> 19) & 0x1)
+#define C_028AA8_SWITCH_ON_EOI 0xFFF7FFFF
+#define R_028AAC_VGT_ESGS_RING_ITEMSIZE 0x028AAC
+#define S_028AAC_ITEMSIZE(x) (((x) & 0x7FFF) << 0)
+#define G_028AAC_ITEMSIZE(x) (((x) >> 0) & 0x7FFF)
+#define C_028AAC_ITEMSIZE 0xFFFF8000
+#define R_028AB0_VGT_GSVS_RING_ITEMSIZE 0x028AB0
+#define S_028AB0_ITEMSIZE(x) (((x) & 0x7FFF) << 0)
+#define G_028AB0_ITEMSIZE(x) (((x) >> 0) & 0x7FFF)
+#define C_028AB0_ITEMSIZE 0xFFFF8000
+#define R_028AB4_VGT_REUSE_OFF 0x028AB4
+#define S_028AB4_REUSE_OFF(x) (((x) & 0x1) << 0)
+#define G_028AB4_REUSE_OFF(x) (((x) >> 0) & 0x1)
+#define C_028AB4_REUSE_OFF 0xFFFFFFFE
+#define R_028AB8_VGT_VTX_CNT_EN 0x028AB8
+#define S_028AB8_VTX_CNT_EN(x) (((x) & 0x1) << 0)
+#define G_028AB8_VTX_CNT_EN(x) (((x) >> 0) & 0x1)
+#define C_028AB8_VTX_CNT_EN 0xFFFFFFFE
+#define R_028ABC_DB_HTILE_SURFACE 0x028ABC
+#define S_028ABC_LINEAR(x) (((x) & 0x1) << 0)
+#define G_028ABC_LINEAR(x) (((x) >> 0) & 0x1)
+#define C_028ABC_LINEAR 0xFFFFFFFE
+#define S_028ABC_FULL_CACHE(x) (((x) & 0x1) << 1)
+#define G_028ABC_FULL_CACHE(x) (((x) >> 1) & 0x1)
+#define C_028ABC_FULL_CACHE 0xFFFFFFFD
+#define S_028ABC_HTILE_USES_PRELOAD_WIN(x) (((x) & 0x1) << 2)
+#define G_028ABC_HTILE_USES_PRELOAD_WIN(x) (((x) >> 2) & 0x1)
+#define C_028ABC_HTILE_USES_PRELOAD_WIN 0xFFFFFFFB
+#define S_028ABC_PRELOAD(x) (((x) & 0x1) << 3)
+#define G_028ABC_PRELOAD(x) (((x) >> 3) & 0x1)
+#define C_028ABC_PRELOAD 0xFFFFFFF7
+#define S_028ABC_PREFETCH_WIDTH(x) (((x) & 0x3F) << 4)
+#define G_028ABC_PREFETCH_WIDTH(x) (((x) >> 4) & 0x3F)
+#define C_028ABC_PREFETCH_WIDTH 0xFFFFFC0F
+#define S_028ABC_PREFETCH_HEIGHT(x) (((x) & 0x3F) << 10)
+#define G_028ABC_PREFETCH_HEIGHT(x) (((x) >> 10) & 0x3F)
+#define C_028ABC_PREFETCH_HEIGHT 0xFFFF03FF
+#define S_028ABC_DST_OUTSIDE_ZERO_TO_ONE(x) (((x) & 0x1) << 16)
+#define G_028ABC_DST_OUTSIDE_ZERO_TO_ONE(x) (((x) >> 16) & 0x1)
+#define C_028ABC_DST_OUTSIDE_ZERO_TO_ONE 0xFFFEFFFF
+#define R_028AC0_DB_SRESULTS_COMPARE_STATE0 0x028AC0
+#define S_028AC0_COMPAREFUNC0(x) (((x) & 0x07) << 0)
+#define G_028AC0_COMPAREFUNC0(x) (((x) >> 0) & 0x07)
+#define C_028AC0_COMPAREFUNC0 0xFFFFFFF8
+#define V_028AC0_REF_NEVER 0x00
+#define V_028AC0_REF_LESS 0x01
+#define V_028AC0_REF_EQUAL 0x02
+#define V_028AC0_REF_LEQUAL 0x03
+#define V_028AC0_REF_GREATER 0x04
+#define V_028AC0_REF_NOTEQUAL 0x05
+#define V_028AC0_REF_GEQUAL 0x06
+#define V_028AC0_REF_ALWAYS 0x07
+#define S_028AC0_COMPAREVALUE0(x) (((x) & 0xFF) << 4)
+#define G_028AC0_COMPAREVALUE0(x) (((x) >> 4) & 0xFF)
+#define C_028AC0_COMPAREVALUE0 0xFFFFF00F
+#define S_028AC0_COMPAREMASK0(x) (((x) & 0xFF) << 12)
+#define G_028AC0_COMPAREMASK0(x) (((x) >> 12) & 0xFF)
+#define C_028AC0_COMPAREMASK0 0xFFF00FFF
+#define S_028AC0_ENABLE0(x) (((x) & 0x1) << 24)
+#define G_028AC0_ENABLE0(x) (((x) >> 24) & 0x1)
+#define C_028AC0_ENABLE0 0xFEFFFFFF
+#define R_028AC4_DB_SRESULTS_COMPARE_STATE1 0x028AC4
+#define S_028AC4_COMPAREFUNC1(x) (((x) & 0x07) << 0)
+#define G_028AC4_COMPAREFUNC1(x) (((x) >> 0) & 0x07)
+#define C_028AC4_COMPAREFUNC1 0xFFFFFFF8
+#define V_028AC4_REF_NEVER 0x00
+#define V_028AC4_REF_LESS 0x01
+#define V_028AC4_REF_EQUAL 0x02
+#define V_028AC4_REF_LEQUAL 0x03
+#define V_028AC4_REF_GREATER 0x04
+#define V_028AC4_REF_NOTEQUAL 0x05
+#define V_028AC4_REF_GEQUAL 0x06
+#define V_028AC4_REF_ALWAYS 0x07
+#define S_028AC4_COMPAREVALUE1(x) (((x) & 0xFF) << 4)
+#define G_028AC4_COMPAREVALUE1(x) (((x) >> 4) & 0xFF)
+#define C_028AC4_COMPAREVALUE1 0xFFFFF00F
+#define S_028AC4_COMPAREMASK1(x) (((x) & 0xFF) << 12)
+#define G_028AC4_COMPAREMASK1(x) (((x) >> 12) & 0xFF)
+#define C_028AC4_COMPAREMASK1 0xFFF00FFF
+#define S_028AC4_ENABLE1(x) (((x) & 0x1) << 24)
+#define G_028AC4_ENABLE1(x) (((x) >> 24) & 0x1)
+#define C_028AC4_ENABLE1 0xFEFFFFFF
+#define R_028AC8_DB_PRELOAD_CONTROL 0x028AC8
+#define S_028AC8_START_X(x) (((x) & 0xFF) << 0)
+#define G_028AC8_START_X(x) (((x) >> 0) & 0xFF)
+#define C_028AC8_START_X 0xFFFFFF00
+#define S_028AC8_START_Y(x) (((x) & 0xFF) << 8)
+#define G_028AC8_START_Y(x) (((x) >> 8) & 0xFF)
+#define C_028AC8_START_Y 0xFFFF00FF
+#define S_028AC8_MAX_X(x) (((x) & 0xFF) << 16)
+#define G_028AC8_MAX_X(x) (((x) >> 16) & 0xFF)
+#define C_028AC8_MAX_X 0xFF00FFFF
+#define S_028AC8_MAX_Y(x) (((x) & 0xFF) << 24)
+#define G_028AC8_MAX_Y(x) (((x) >> 24) & 0xFF)
+#define C_028AC8_MAX_Y 0x00FFFFFF
+#define R_028AD0_VGT_STRMOUT_BUFFER_SIZE_0 0x028AD0
+#define R_028AD4_VGT_STRMOUT_VTX_STRIDE_0 0x028AD4
+#define S_028AD4_STRIDE(x) (((x) & 0x3FF) << 0)
+#define G_028AD4_STRIDE(x) (((x) >> 0) & 0x3FF)
+#define C_028AD4_STRIDE 0xFFFFFC00
+#define R_028ADC_VGT_STRMOUT_BUFFER_OFFSET_0 0x028ADC
+#define R_028AE0_VGT_STRMOUT_BUFFER_SIZE_1 0x028AE0
+#define R_028AE4_VGT_STRMOUT_VTX_STRIDE_1 0x028AE4
+#define S_028AE4_STRIDE(x) (((x) & 0x3FF) << 0)
+#define G_028AE4_STRIDE(x) (((x) >> 0) & 0x3FF)
+#define C_028AE4_STRIDE 0xFFFFFC00
+#define R_028AEC_VGT_STRMOUT_BUFFER_OFFSET_1 0x028AEC
+#define R_028AF0_VGT_STRMOUT_BUFFER_SIZE_2 0x028AF0
+#define R_028AF4_VGT_STRMOUT_VTX_STRIDE_2 0x028AF4
+#define S_028AF4_STRIDE(x) (((x) & 0x3FF) << 0)
+#define G_028AF4_STRIDE(x) (((x) >> 0) & 0x3FF)
+#define C_028AF4_STRIDE 0xFFFFFC00
+#define R_028AFC_VGT_STRMOUT_BUFFER_OFFSET_2 0x028AFC
+#define R_028B00_VGT_STRMOUT_BUFFER_SIZE_3 0x028B00
+#define R_028B04_VGT_STRMOUT_VTX_STRIDE_3 0x028B04
+#define S_028B04_STRIDE(x) (((x) & 0x3FF) << 0)
+#define G_028B04_STRIDE(x) (((x) >> 0) & 0x3FF)
+#define C_028B04_STRIDE 0xFFFFFC00
+#define R_028B0C_VGT_STRMOUT_BUFFER_OFFSET_3 0x028B0C
+#define R_028B28_VGT_STRMOUT_DRAW_OPAQUE_OFFSET 0x028B28
+#define R_028B2C_VGT_STRMOUT_DRAW_OPAQUE_BUFFER_FILLED_SIZE 0x028B2C
+#define R_028B30_VGT_STRMOUT_DRAW_OPAQUE_VERTEX_STRIDE 0x028B30
+#define S_028B30_VERTEX_STRIDE(x) (((x) & 0x1FF) << 0)
+#define G_028B30_VERTEX_STRIDE(x) (((x) >> 0) & 0x1FF)
+#define C_028B30_VERTEX_STRIDE 0xFFFFFE00
+#define R_028B38_VGT_GS_MAX_VERT_OUT 0x028B38
+#define S_028B38_MAX_VERT_OUT(x) (((x) & 0x7FF) << 0)
+#define G_028B38_MAX_VERT_OUT(x) (((x) >> 0) & 0x7FF)
+#define C_028B38_MAX_VERT_OUT 0xFFFFF800
+#define R_028B54_VGT_SHADER_STAGES_EN 0x028B54
+#define S_028B54_LS_EN(x) (((x) & 0x03) << 0)
+#define G_028B54_LS_EN(x) (((x) >> 0) & 0x03)
+#define C_028B54_LS_EN 0xFFFFFFFC
+#define V_028B54_LS_STAGE_OFF 0x00
+#define V_028B54_LS_STAGE_ON 0x01
+#define V_028B54_CS_STAGE_ON 0x02
+#define S_028B54_HS_EN(x) (((x) & 0x1) << 2)
+#define G_028B54_HS_EN(x) (((x) >> 2) & 0x1)
+#define C_028B54_HS_EN 0xFFFFFFFB
+#define S_028B54_ES_EN(x) (((x) & 0x03) << 3)
+#define G_028B54_ES_EN(x) (((x) >> 3) & 0x03)
+#define C_028B54_ES_EN 0xFFFFFFE7
+#define V_028B54_ES_STAGE_OFF 0x00
+#define V_028B54_ES_STAGE_DS 0x01
+#define V_028B54_ES_STAGE_REAL 0x02
+#define S_028B54_GS_EN(x) (((x) & 0x1) << 5)
+#define G_028B54_GS_EN(x) (((x) >> 5) & 0x1)
+#define C_028B54_GS_EN 0xFFFFFFDF
+#define S_028B54_VS_EN(x) (((x) & 0x03) << 6)
+#define G_028B54_VS_EN(x) (((x) >> 6) & 0x03)
+#define C_028B54_VS_EN 0xFFFFFF3F
+#define V_028B54_VS_STAGE_REAL 0x00
+#define V_028B54_VS_STAGE_DS 0x01
+#define V_028B54_VS_STAGE_COPY_SHADER 0x02
+#define S_028B54_DYNAMIC_HS(x) (((x) & 0x1) << 8)
+#define G_028B54_DYNAMIC_HS(x) (((x) >> 8) & 0x1)
+#define C_028B54_DYNAMIC_HS 0xFFFFFEFF
+#define R_028B58_VGT_LS_HS_CONFIG 0x028B58
+#define S_028B58_NUM_PATCHES(x) (((x) & 0xFF) << 0)
+#define G_028B58_NUM_PATCHES(x) (((x) >> 0) & 0xFF)
+#define C_028B58_NUM_PATCHES 0xFFFFFF00
+#define S_028B58_HS_NUM_INPUT_CP(x) (((x) & 0x3F) << 8)
+#define G_028B58_HS_NUM_INPUT_CP(x) (((x) >> 8) & 0x3F)
+#define C_028B58_HS_NUM_INPUT_CP 0xFFFFC0FF
+#define S_028B58_HS_NUM_OUTPUT_CP(x) (((x) & 0x3F) << 14)
+#define G_028B58_HS_NUM_OUTPUT_CP(x) (((x) >> 14) & 0x3F)
+#define C_028B58_HS_NUM_OUTPUT_CP 0xFFF03FFF
+#define R_028B5C_VGT_GS_VERT_ITEMSIZE 0x028B5C
+#define S_028B5C_ITEMSIZE(x) (((x) & 0x7FFF) << 0)
+#define G_028B5C_ITEMSIZE(x) (((x) >> 0) & 0x7FFF)
+#define C_028B5C_ITEMSIZE 0xFFFF8000
+#define R_028B60_VGT_GS_VERT_ITEMSIZE_1 0x028B60
+#define S_028B60_ITEMSIZE(x) (((x) & 0x7FFF) << 0)
+#define G_028B60_ITEMSIZE(x) (((x) >> 0) & 0x7FFF)
+#define C_028B60_ITEMSIZE 0xFFFF8000
+#define R_028B64_VGT_GS_VERT_ITEMSIZE_2 0x028B64
+#define S_028B64_ITEMSIZE(x) (((x) & 0x7FFF) << 0)
+#define G_028B64_ITEMSIZE(x) (((x) >> 0) & 0x7FFF)
+#define C_028B64_ITEMSIZE 0xFFFF8000
+#define R_028B68_VGT_GS_VERT_ITEMSIZE_3 0x028B68
+#define S_028B68_ITEMSIZE(x) (((x) & 0x7FFF) << 0)
+#define G_028B68_ITEMSIZE(x) (((x) >> 0) & 0x7FFF)
+#define C_028B68_ITEMSIZE 0xFFFF8000
+#define R_028B6C_VGT_TF_PARAM 0x028B6C
+#define S_028B6C_TYPE(x) (((x) & 0x03) << 0)
+#define G_028B6C_TYPE(x) (((x) >> 0) & 0x03)
+#define C_028B6C_TYPE 0xFFFFFFFC
+#define V_028B6C_TESS_ISOLINE 0x00
+#define V_028B6C_TESS_TRIANGLE 0x01
+#define V_028B6C_TESS_QUAD 0x02
+#define S_028B6C_PARTITIONING(x) (((x) & 0x07) << 2)
+#define G_028B6C_PARTITIONING(x) (((x) >> 2) & 0x07)
+#define C_028B6C_PARTITIONING 0xFFFFFFE3
+#define V_028B6C_PART_INTEGER 0x00
+#define V_028B6C_PART_POW2 0x01
+#define V_028B6C_PART_FRAC_ODD 0x02
+#define V_028B6C_PART_FRAC_EVEN 0x03
+#define S_028B6C_TOPOLOGY(x) (((x) & 0x07) << 5)
+#define G_028B6C_TOPOLOGY(x) (((x) >> 5) & 0x07)
+#define C_028B6C_TOPOLOGY 0xFFFFFF1F
+#define V_028B6C_OUTPUT_POINT 0x00
+#define V_028B6C_OUTPUT_LINE 0x01
+#define V_028B6C_OUTPUT_TRIANGLE_CW 0x02
+#define V_028B6C_OUTPUT_TRIANGLE_CCW 0x03
+#define S_028B6C_RESERVED_REDUC_AXIS(x) (((x) & 0x1) << 8)
+#define G_028B6C_RESERVED_REDUC_AXIS(x) (((x) >> 8) & 0x1)
+#define C_028B6C_RESERVED_REDUC_AXIS 0xFFFFFEFF
+#define S_028B6C_NUM_DS_WAVES_PER_SIMD(x) (((x) & 0x0F) << 10)
+#define G_028B6C_NUM_DS_WAVES_PER_SIMD(x) (((x) >> 10) & 0x0F)
+#define C_028B6C_NUM_DS_WAVES_PER_SIMD 0xFFFFC3FF
+#define S_028B6C_DISABLE_DONUTS(x) (((x) & 0x1) << 14)
+#define G_028B6C_DISABLE_DONUTS(x) (((x) >> 14) & 0x1)
+#define C_028B6C_DISABLE_DONUTS 0xFFFFBFFF
+#define R_028B70_DB_ALPHA_TO_MASK 0x028B70
+#define S_028B70_ALPHA_TO_MASK_ENABLE(x) (((x) & 0x1) << 0)
+#define G_028B70_ALPHA_TO_MASK_ENABLE(x) (((x) >> 0) & 0x1)
+#define C_028B70_ALPHA_TO_MASK_ENABLE 0xFFFFFFFE
+#define S_028B70_ALPHA_TO_MASK_OFFSET0(x) (((x) & 0x03) << 8)
+#define G_028B70_ALPHA_TO_MASK_OFFSET0(x) (((x) >> 8) & 0x03)
+#define C_028B70_ALPHA_TO_MASK_OFFSET0 0xFFFFFCFF
+#define S_028B70_ALPHA_TO_MASK_OFFSET1(x) (((x) & 0x03) << 10)
+#define G_028B70_ALPHA_TO_MASK_OFFSET1(x) (((x) >> 10) & 0x03)
+#define C_028B70_ALPHA_TO_MASK_OFFSET1 0xFFFFF3FF
+#define S_028B70_ALPHA_TO_MASK_OFFSET2(x) (((x) & 0x03) << 12)
+#define G_028B70_ALPHA_TO_MASK_OFFSET2(x) (((x) >> 12) & 0x03)
+#define C_028B70_ALPHA_TO_MASK_OFFSET2 0xFFFFCFFF
+#define S_028B70_ALPHA_TO_MASK_OFFSET3(x) (((x) & 0x03) << 14)
+#define G_028B70_ALPHA_TO_MASK_OFFSET3(x) (((x) >> 14) & 0x03)
+#define C_028B70_ALPHA_TO_MASK_OFFSET3 0xFFFF3FFF
+#define S_028B70_OFFSET_ROUND(x) (((x) & 0x1) << 16)
+#define G_028B70_OFFSET_ROUND(x) (((x) >> 16) & 0x1)
+#define C_028B70_OFFSET_ROUND 0xFFFEFFFF
+#define R_028B78_PA_SU_POLY_OFFSET_DB_FMT_CNTL 0x028B78
+#define S_028B78_POLY_OFFSET_NEG_NUM_DB_BITS(x) (((x) & 0xFF) << 0)
+#define G_028B78_POLY_OFFSET_NEG_NUM_DB_BITS(x) (((x) >> 0) & 0xFF)
+#define C_028B78_POLY_OFFSET_NEG_NUM_DB_BITS 0xFFFFFF00
+#define S_028B78_POLY_OFFSET_DB_IS_FLOAT_FMT(x) (((x) & 0x1) << 8)
+#define G_028B78_POLY_OFFSET_DB_IS_FLOAT_FMT(x) (((x) >> 8) & 0x1)
+#define C_028B78_POLY_OFFSET_DB_IS_FLOAT_FMT 0xFFFFFEFF
+#define R_028B7C_PA_SU_POLY_OFFSET_CLAMP 0x028B7C
+#define R_028B80_PA_SU_POLY_OFFSET_FRONT_SCALE 0x028B80
+#define R_028B84_PA_SU_POLY_OFFSET_FRONT_OFFSET 0x028B84
+#define R_028B88_PA_SU_POLY_OFFSET_BACK_SCALE 0x028B88
+#define R_028B8C_PA_SU_POLY_OFFSET_BACK_OFFSET 0x028B8C
+#define R_028B90_VGT_GS_INSTANCE_CNT 0x028B90
+#define S_028B90_ENABLE(x) (((x) & 0x1) << 0)
+#define G_028B90_ENABLE(x) (((x) >> 0) & 0x1)
+#define C_028B90_ENABLE 0xFFFFFFFE
+#define S_028B90_CNT(x) (((x) & 0x7F) << 2)
+#define G_028B90_CNT(x) (((x) >> 2) & 0x7F)
+#define C_028B90_CNT 0xFFFFFE03
+#define R_028B94_VGT_STRMOUT_CONFIG 0x028B94
+#define S_028B94_STREAMOUT_0_EN(x) (((x) & 0x1) << 0)
+#define G_028B94_STREAMOUT_0_EN(x) (((x) >> 0) & 0x1)
+#define C_028B94_STREAMOUT_0_EN 0xFFFFFFFE
+#define S_028B94_STREAMOUT_1_EN(x) (((x) & 0x1) << 1)
+#define G_028B94_STREAMOUT_1_EN(x) (((x) >> 1) & 0x1)
+#define C_028B94_STREAMOUT_1_EN 0xFFFFFFFD
+#define S_028B94_STREAMOUT_2_EN(x) (((x) & 0x1) << 2)
+#define G_028B94_STREAMOUT_2_EN(x) (((x) >> 2) & 0x1)
+#define C_028B94_STREAMOUT_2_EN 0xFFFFFFFB
+#define S_028B94_STREAMOUT_3_EN(x) (((x) & 0x1) << 3)
+#define G_028B94_STREAMOUT_3_EN(x) (((x) >> 3) & 0x1)
+#define C_028B94_STREAMOUT_3_EN 0xFFFFFFF7
+#define S_028B94_RAST_STREAM(x) (((x) & 0x07) << 4)
+#define G_028B94_RAST_STREAM(x) (((x) >> 4) & 0x07)
+#define C_028B94_RAST_STREAM 0xFFFFFF8F
+#define S_028B94_RAST_STREAM_MASK(x) (((x) & 0x0F) << 8)
+#define G_028B94_RAST_STREAM_MASK(x) (((x) >> 8) & 0x0F)
+#define C_028B94_RAST_STREAM_MASK 0xFFFFF0FF
+#define S_028B94_USE_RAST_STREAM_MASK(x) (((x) & 0x1) << 31)
+#define G_028B94_USE_RAST_STREAM_MASK(x) (((x) >> 31) & 0x1)
+#define C_028B94_USE_RAST_STREAM_MASK 0x7FFFFFFF
+#define R_028B98_VGT_STRMOUT_BUFFER_CONFIG 0x028B98
+#define S_028B98_STREAM_0_BUFFER_EN(x) (((x) & 0x0F) << 0)
+#define G_028B98_STREAM_0_BUFFER_EN(x) (((x) >> 0) & 0x0F)
+#define C_028B98_STREAM_0_BUFFER_EN 0xFFFFFFF0
+#define S_028B98_STREAM_1_BUFFER_EN(x) (((x) & 0x0F) << 4)
+#define G_028B98_STREAM_1_BUFFER_EN(x) (((x) >> 4) & 0x0F)
+#define C_028B98_STREAM_1_BUFFER_EN 0xFFFFFF0F
+#define S_028B98_STREAM_2_BUFFER_EN(x) (((x) & 0x0F) << 8)
+#define G_028B98_STREAM_2_BUFFER_EN(x) (((x) >> 8) & 0x0F)
+#define C_028B98_STREAM_2_BUFFER_EN 0xFFFFF0FF
+#define S_028B98_STREAM_3_BUFFER_EN(x) (((x) & 0x0F) << 12)
+#define G_028B98_STREAM_3_BUFFER_EN(x) (((x) >> 12) & 0x0F)
+#define C_028B98_STREAM_3_BUFFER_EN 0xFFFF0FFF
+#define R_028BD4_PA_SC_CENTROID_PRIORITY_0 0x028BD4
+#define S_028BD4_DISTANCE_0(x) (((x) & 0x0F) << 0)
+#define G_028BD4_DISTANCE_0(x) (((x) >> 0) & 0x0F)
+#define C_028BD4_DISTANCE_0 0xFFFFFFF0
+#define S_028BD4_DISTANCE_1(x) (((x) & 0x0F) << 4)
+#define G_028BD4_DISTANCE_1(x) (((x) >> 4) & 0x0F)
+#define C_028BD4_DISTANCE_1 0xFFFFFF0F
+#define S_028BD4_DISTANCE_2(x) (((x) & 0x0F) << 8)
+#define G_028BD4_DISTANCE_2(x) (((x) >> 8) & 0x0F)
+#define C_028BD4_DISTANCE_2 0xFFFFF0FF
+#define S_028BD4_DISTANCE_3(x) (((x) & 0x0F) << 12)
+#define G_028BD4_DISTANCE_3(x) (((x) >> 12) & 0x0F)
+#define C_028BD4_DISTANCE_3 0xFFFF0FFF
+#define S_028BD4_DISTANCE_4(x) (((x) & 0x0F) << 16)
+#define G_028BD4_DISTANCE_4(x) (((x) >> 16) & 0x0F)
+#define C_028BD4_DISTANCE_4 0xFFF0FFFF
+#define S_028BD4_DISTANCE_5(x) (((x) & 0x0F) << 20)
+#define G_028BD4_DISTANCE_5(x) (((x) >> 20) & 0x0F)
+#define C_028BD4_DISTANCE_5 0xFF0FFFFF
+#define S_028BD4_DISTANCE_6(x) (((x) & 0x0F) << 24)
+#define G_028BD4_DISTANCE_6(x) (((x) >> 24) & 0x0F)
+#define C_028BD4_DISTANCE_6 0xF0FFFFFF
+#define S_028BD4_DISTANCE_7(x) (((x) & 0x0F) << 28)
+#define G_028BD4_DISTANCE_7(x) (((x) >> 28) & 0x0F)
+#define C_028BD4_DISTANCE_7 0x0FFFFFFF
+#define R_028BD8_PA_SC_CENTROID_PRIORITY_1 0x028BD8
+#define S_028BD8_DISTANCE_8(x) (((x) & 0x0F) << 0)
+#define G_028BD8_DISTANCE_8(x) (((x) >> 0) & 0x0F)
+#define C_028BD8_DISTANCE_8 0xFFFFFFF0
+#define S_028BD8_DISTANCE_9(x) (((x) & 0x0F) << 4)
+#define G_028BD8_DISTANCE_9(x) (((x) >> 4) & 0x0F)
+#define C_028BD8_DISTANCE_9 0xFFFFFF0F
+#define S_028BD8_DISTANCE_10(x) (((x) & 0x0F) << 8)
+#define G_028BD8_DISTANCE_10(x) (((x) >> 8) & 0x0F)
+#define C_028BD8_DISTANCE_10 0xFFFFF0FF
+#define S_028BD8_DISTANCE_11(x) (((x) & 0x0F) << 12)
+#define G_028BD8_DISTANCE_11(x) (((x) >> 12) & 0x0F)
+#define C_028BD8_DISTANCE_11 0xFFFF0FFF
+#define S_028BD8_DISTANCE_12(x) (((x) & 0x0F) << 16)
+#define G_028BD8_DISTANCE_12(x) (((x) >> 16) & 0x0F)
+#define C_028BD8_DISTANCE_12 0xFFF0FFFF
+#define S_028BD8_DISTANCE_13(x) (((x) & 0x0F) << 20)
+#define G_028BD8_DISTANCE_13(x) (((x) >> 20) & 0x0F)
+#define C_028BD8_DISTANCE_13 0xFF0FFFFF
+#define S_028BD8_DISTANCE_14(x) (((x) & 0x0F) << 24)
+#define G_028BD8_DISTANCE_14(x) (((x) >> 24) & 0x0F)
+#define C_028BD8_DISTANCE_14 0xF0FFFFFF
+#define S_028BD8_DISTANCE_15(x) (((x) & 0x0F) << 28)
+#define G_028BD8_DISTANCE_15(x) (((x) >> 28) & 0x0F)
+#define C_028BD8_DISTANCE_15 0x0FFFFFFF
+#define R_028BDC_PA_SC_LINE_CNTL 0x028BDC
+#define S_028BDC_EXPAND_LINE_WIDTH(x) (((x) & 0x1) << 9)
+#define G_028BDC_EXPAND_LINE_WIDTH(x) (((x) >> 9) & 0x1)
+#define C_028BDC_EXPAND_LINE_WIDTH 0xFFFFFDFF
+#define S_028BDC_LAST_PIXEL(x) (((x) & 0x1) << 10)
+#define G_028BDC_LAST_PIXEL(x) (((x) >> 10) & 0x1)
+#define C_028BDC_LAST_PIXEL 0xFFFFFBFF
+#define S_028BDC_PERPENDICULAR_ENDCAP_ENA(x) (((x) & 0x1) << 11)
+#define G_028BDC_PERPENDICULAR_ENDCAP_ENA(x) (((x) >> 11) & 0x1)
+#define C_028BDC_PERPENDICULAR_ENDCAP_ENA 0xFFFFF7FF
+#define S_028BDC_DX10_DIAMOND_TEST_ENA(x) (((x) & 0x1) << 12)
+#define G_028BDC_DX10_DIAMOND_TEST_ENA(x) (((x) >> 12) & 0x1)
+#define C_028BDC_DX10_DIAMOND_TEST_ENA 0xFFFFEFFF
+#define R_028BE0_PA_SC_AA_CONFIG 0x028BE0
+#define S_028BE0_MSAA_NUM_SAMPLES(x) (((x) & 0x07) << 0)
+#define G_028BE0_MSAA_NUM_SAMPLES(x) (((x) >> 0) & 0x07)
+#define C_028BE0_MSAA_NUM_SAMPLES 0xFFFFFFF8
+#define S_028BE0_AA_MASK_CENTROID_DTMN(x) (((x) & 0x1) << 4)
+#define G_028BE0_AA_MASK_CENTROID_DTMN(x) (((x) >> 4) & 0x1)
+#define C_028BE0_AA_MASK_CENTROID_DTMN 0xFFFFFFEF
+#define S_028BE0_MAX_SAMPLE_DIST(x) (((x) & 0x0F) << 13)
+#define G_028BE0_MAX_SAMPLE_DIST(x) (((x) >> 13) & 0x0F)
+#define C_028BE0_MAX_SAMPLE_DIST 0xFFFE1FFF
+#define S_028BE0_MSAA_EXPOSED_SAMPLES(x) (((x) & 0x07) << 20)
+#define G_028BE0_MSAA_EXPOSED_SAMPLES(x) (((x) >> 20) & 0x07)
+#define C_028BE0_MSAA_EXPOSED_SAMPLES 0xFF8FFFFF
+#define S_028BE0_DETAIL_TO_EXPOSED_MODE(x) (((x) & 0x03) << 24)
+#define G_028BE0_DETAIL_TO_EXPOSED_MODE(x) (((x) >> 24) & 0x03)
+#define C_028BE0_DETAIL_TO_EXPOSED_MODE 0xFCFFFFFF
+#define R_028BE4_PA_SU_VTX_CNTL 0x028BE4
+#define S_028BE4_PIX_CENTER(x) (((x) & 0x1) << 0)
+#define G_028BE4_PIX_CENTER(x) (((x) >> 0) & 0x1)
+#define C_028BE4_PIX_CENTER 0xFFFFFFFE
+#define S_028BE4_ROUND_MODE(x) (((x) & 0x03) << 1)
+#define G_028BE4_ROUND_MODE(x) (((x) >> 1) & 0x03)
+#define C_028BE4_ROUND_MODE 0xFFFFFFF9
+#define V_028BE4_X_TRUNCATE 0x00
+#define V_028BE4_X_ROUND 0x01
+#define V_028BE4_X_ROUND_TO_EVEN 0x02
+#define V_028BE4_X_ROUND_TO_ODD 0x03
+#define S_028BE4_QUANT_MODE(x) (((x) & 0x07) << 3)
+#define G_028BE4_QUANT_MODE(x) (((x) >> 3) & 0x07)
+#define C_028BE4_QUANT_MODE 0xFFFFFFC7
+#define V_028BE4_X_16_8_FIXED_POINT_1_16TH 0x00
+#define V_028BE4_X_16_8_FIXED_POINT_1_8TH 0x01
+#define V_028BE4_X_16_8_FIXED_POINT_1_4TH 0x02
+#define V_028BE4_X_16_8_FIXED_POINT_1_2 0x03
+#define V_028BE4_X_16_8_FIXED_POINT_1 0x04
+#define V_028BE4_X_16_8_FIXED_POINT_1_256TH 0x05
+#define V_028BE4_X_14_10_FIXED_POINT_1_1024TH 0x06
+#define V_028BE4_X_12_12_FIXED_POINT_1_4096TH 0x07
+#define R_028BE8_PA_CL_GB_VERT_CLIP_ADJ 0x028BE8
+#define R_028BEC_PA_CL_GB_VERT_DISC_ADJ 0x028BEC
+#define R_028BF0_PA_CL_GB_HORZ_CLIP_ADJ 0x028BF0
+#define R_028BF4_PA_CL_GB_HORZ_DISC_ADJ 0x028BF4
+#define R_028BF8_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_0 0x028BF8
+#define S_028BF8_S0_X(x) (((x) & 0x0F) << 0)
+#define G_028BF8_S0_X(x) (((x) >> 0) & 0x0F)
+#define C_028BF8_S0_X 0xFFFFFFF0
+#define S_028BF8_S0_Y(x) (((x) & 0x0F) << 4)
+#define G_028BF8_S0_Y(x) (((x) >> 4) & 0x0F)
+#define C_028BF8_S0_Y 0xFFFFFF0F
+#define S_028BF8_S1_X(x) (((x) & 0x0F) << 8)
+#define G_028BF8_S1_X(x) (((x) >> 8) & 0x0F)
+#define C_028BF8_S1_X 0xFFFFF0FF
+#define S_028BF8_S1_Y(x) (((x) & 0x0F) << 12)
+#define G_028BF8_S1_Y(x) (((x) >> 12) & 0x0F)
+#define C_028BF8_S1_Y 0xFFFF0FFF
+#define S_028BF8_S2_X(x) (((x) & 0x0F) << 16)
+#define G_028BF8_S2_X(x) (((x) >> 16) & 0x0F)
+#define C_028BF8_S2_X 0xFFF0FFFF
+#define S_028BF8_S2_Y(x) (((x) & 0x0F) << 20)
+#define G_028BF8_S2_Y(x) (((x) >> 20) & 0x0F)
+#define C_028BF8_S2_Y 0xFF0FFFFF
+#define S_028BF8_S3_X(x) (((x) & 0x0F) << 24)
+#define G_028BF8_S3_X(x) (((x) >> 24) & 0x0F)
+#define C_028BF8_S3_X 0xF0FFFFFF
+#define S_028BF8_S3_Y(x) (((x) & 0x0F) << 28)
+#define G_028BF8_S3_Y(x) (((x) >> 28) & 0x0F)
+#define C_028BF8_S3_Y 0x0FFFFFFF
+#define R_028BFC_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_1 0x028BFC
+#define S_028BFC_S4_X(x) (((x) & 0x0F) << 0)
+#define G_028BFC_S4_X(x) (((x) >> 0) & 0x0F)
+#define C_028BFC_S4_X 0xFFFFFFF0
+#define S_028BFC_S4_Y(x) (((x) & 0x0F) << 4)
+#define G_028BFC_S4_Y(x) (((x) >> 4) & 0x0F)
+#define C_028BFC_S4_Y 0xFFFFFF0F
+#define S_028BFC_S5_X(x) (((x) & 0x0F) << 8)
+#define G_028BFC_S5_X(x) (((x) >> 8) & 0x0F)
+#define C_028BFC_S5_X 0xFFFFF0FF
+#define S_028BFC_S5_Y(x) (((x) & 0x0F) << 12)
+#define G_028BFC_S5_Y(x) (((x) >> 12) & 0x0F)
+#define C_028BFC_S5_Y 0xFFFF0FFF
+#define S_028BFC_S6_X(x) (((x) & 0x0F) << 16)
+#define G_028BFC_S6_X(x) (((x) >> 16) & 0x0F)
+#define C_028BFC_S6_X 0xFFF0FFFF
+#define S_028BFC_S6_Y(x) (((x) & 0x0F) << 20)
+#define G_028BFC_S6_Y(x) (((x) >> 20) & 0x0F)
+#define C_028BFC_S6_Y 0xFF0FFFFF
+#define S_028BFC_S7_X(x) (((x) & 0x0F) << 24)
+#define G_028BFC_S7_X(x) (((x) >> 24) & 0x0F)
+#define C_028BFC_S7_X 0xF0FFFFFF
+#define S_028BFC_S7_Y(x) (((x) & 0x0F) << 28)
+#define G_028BFC_S7_Y(x) (((x) >> 28) & 0x0F)
+#define C_028BFC_S7_Y 0x0FFFFFFF
+#define R_028C00_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_2 0x028C00
+#define S_028C00_S8_X(x) (((x) & 0x0F) << 0)
+#define G_028C00_S8_X(x) (((x) >> 0) & 0x0F)
+#define C_028C00_S8_X 0xFFFFFFF0
+#define S_028C00_S8_Y(x) (((x) & 0x0F) << 4)
+#define G_028C00_S8_Y(x) (((x) >> 4) & 0x0F)
+#define C_028C00_S8_Y 0xFFFFFF0F
+#define S_028C00_S9_X(x) (((x) & 0x0F) << 8)
+#define G_028C00_S9_X(x) (((x) >> 8) & 0x0F)
+#define C_028C00_S9_X 0xFFFFF0FF
+#define S_028C00_S9_Y(x) (((x) & 0x0F) << 12)
+#define G_028C00_S9_Y(x) (((x) >> 12) & 0x0F)
+#define C_028C00_S9_Y 0xFFFF0FFF
+#define S_028C00_S10_X(x) (((x) & 0x0F) << 16)
+#define G_028C00_S10_X(x) (((x) >> 16) & 0x0F)
+#define C_028C00_S10_X 0xFFF0FFFF
+#define S_028C00_S10_Y(x) (((x) & 0x0F) << 20)
+#define G_028C00_S10_Y(x) (((x) >> 20) & 0x0F)
+#define C_028C00_S10_Y 0xFF0FFFFF
+#define S_028C00_S11_X(x) (((x) & 0x0F) << 24)
+#define G_028C00_S11_X(x) (((x) >> 24) & 0x0F)
+#define C_028C00_S11_X 0xF0FFFFFF
+#define S_028C00_S11_Y(x) (((x) & 0x0F) << 28)
+#define G_028C00_S11_Y(x) (((x) >> 28) & 0x0F)
+#define C_028C00_S11_Y 0x0FFFFFFF
+#define R_028C04_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_3 0x028C04
+#define S_028C04_S12_X(x) (((x) & 0x0F) << 0)
+#define G_028C04_S12_X(x) (((x) >> 0) & 0x0F)
+#define C_028C04_S12_X 0xFFFFFFF0
+#define S_028C04_S12_Y(x) (((x) & 0x0F) << 4)
+#define G_028C04_S12_Y(x) (((x) >> 4) & 0x0F)
+#define C_028C04_S12_Y 0xFFFFFF0F
+#define S_028C04_S13_X(x) (((x) & 0x0F) << 8)
+#define G_028C04_S13_X(x) (((x) >> 8) & 0x0F)
+#define C_028C04_S13_X 0xFFFFF0FF
+#define S_028C04_S13_Y(x) (((x) & 0x0F) << 12)
+#define G_028C04_S13_Y(x) (((x) >> 12) & 0x0F)
+#define C_028C04_S13_Y 0xFFFF0FFF
+#define S_028C04_S14_X(x) (((x) & 0x0F) << 16)
+#define G_028C04_S14_X(x) (((x) >> 16) & 0x0F)
+#define C_028C04_S14_X 0xFFF0FFFF
+#define S_028C04_S14_Y(x) (((x) & 0x0F) << 20)
+#define G_028C04_S14_Y(x) (((x) >> 20) & 0x0F)
+#define C_028C04_S14_Y 0xFF0FFFFF
+#define S_028C04_S15_X(x) (((x) & 0x0F) << 24)
+#define G_028C04_S15_X(x) (((x) >> 24) & 0x0F)
+#define C_028C04_S15_X 0xF0FFFFFF
+#define S_028C04_S15_Y(x) (((x) & 0x0F) << 28)
+#define G_028C04_S15_Y(x) (((x) >> 28) & 0x0F)
+#define C_028C04_S15_Y 0x0FFFFFFF
+#define R_028C08_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_0 0x028C08
+#define S_028C08_S0_X(x) (((x) & 0x0F) << 0)
+#define G_028C08_S0_X(x) (((x) >> 0) & 0x0F)
+#define C_028C08_S0_X 0xFFFFFFF0
+#define S_028C08_S0_Y(x) (((x) & 0x0F) << 4)
+#define G_028C08_S0_Y(x) (((x) >> 4) & 0x0F)
+#define C_028C08_S0_Y 0xFFFFFF0F
+#define S_028C08_S1_X(x) (((x) & 0x0F) << 8)
+#define G_028C08_S1_X(x) (((x) >> 8) & 0x0F)
+#define C_028C08_S1_X 0xFFFFF0FF
+#define S_028C08_S1_Y(x) (((x) & 0x0F) << 12)
+#define G_028C08_S1_Y(x) (((x) >> 12) & 0x0F)
+#define C_028C08_S1_Y 0xFFFF0FFF
+#define S_028C08_S2_X(x) (((x) & 0x0F) << 16)
+#define G_028C08_S2_X(x) (((x) >> 16) & 0x0F)
+#define C_028C08_S2_X 0xFFF0FFFF
+#define S_028C08_S2_Y(x) (((x) & 0x0F) << 20)
+#define G_028C08_S2_Y(x) (((x) >> 20) & 0x0F)
+#define C_028C08_S2_Y 0xFF0FFFFF
+#define S_028C08_S3_X(x) (((x) & 0x0F) << 24)
+#define G_028C08_S3_X(x) (((x) >> 24) & 0x0F)
+#define C_028C08_S3_X 0xF0FFFFFF
+#define S_028C08_S3_Y(x) (((x) & 0x0F) << 28)
+#define G_028C08_S3_Y(x) (((x) >> 28) & 0x0F)
+#define C_028C08_S3_Y 0x0FFFFFFF
+#define R_028C0C_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_1 0x028C0C
+#define S_028C0C_S4_X(x) (((x) & 0x0F) << 0)
+#define G_028C0C_S4_X(x) (((x) >> 0) & 0x0F)
+#define C_028C0C_S4_X 0xFFFFFFF0
+#define S_028C0C_S4_Y(x) (((x) & 0x0F) << 4)
+#define G_028C0C_S4_Y(x) (((x) >> 4) & 0x0F)
+#define C_028C0C_S4_Y 0xFFFFFF0F
+#define S_028C0C_S5_X(x) (((x) & 0x0F) << 8)
+#define G_028C0C_S5_X(x) (((x) >> 8) & 0x0F)
+#define C_028C0C_S5_X 0xFFFFF0FF
+#define S_028C0C_S5_Y(x) (((x) & 0x0F) << 12)
+#define G_028C0C_S5_Y(x) (((x) >> 12) & 0x0F)
+#define C_028C0C_S5_Y 0xFFFF0FFF
+#define S_028C0C_S6_X(x) (((x) & 0x0F) << 16)
+#define G_028C0C_S6_X(x) (((x) >> 16) & 0x0F)
+#define C_028C0C_S6_X 0xFFF0FFFF
+#define S_028C0C_S6_Y(x) (((x) & 0x0F) << 20)
+#define G_028C0C_S6_Y(x) (((x) >> 20) & 0x0F)
+#define C_028C0C_S6_Y 0xFF0FFFFF
+#define S_028C0C_S7_X(x) (((x) & 0x0F) << 24)
+#define G_028C0C_S7_X(x) (((x) >> 24) & 0x0F)
+#define C_028C0C_S7_X 0xF0FFFFFF
+#define S_028C0C_S7_Y(x) (((x) & 0x0F) << 28)
+#define G_028C0C_S7_Y(x) (((x) >> 28) & 0x0F)
+#define C_028C0C_S7_Y 0x0FFFFFFF
+#define R_028C10_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_2 0x028C10
+#define S_028C10_S8_X(x) (((x) & 0x0F) << 0)
+#define G_028C10_S8_X(x) (((x) >> 0) & 0x0F)
+#define C_028C10_S8_X 0xFFFFFFF0
+#define S_028C10_S8_Y(x) (((x) & 0x0F) << 4)
+#define G_028C10_S8_Y(x) (((x) >> 4) & 0x0F)
+#define C_028C10_S8_Y 0xFFFFFF0F
+#define S_028C10_S9_X(x) (((x) & 0x0F) << 8)
+#define G_028C10_S9_X(x) (((x) >> 8) & 0x0F)
+#define C_028C10_S9_X 0xFFFFF0FF
+#define S_028C10_S9_Y(x) (((x) & 0x0F) << 12)
+#define G_028C10_S9_Y(x) (((x) >> 12) & 0x0F)
+#define C_028C10_S9_Y 0xFFFF0FFF
+#define S_028C10_S10_X(x) (((x) & 0x0F) << 16)
+#define G_028C10_S10_X(x) (((x) >> 16) & 0x0F)
+#define C_028C10_S10_X 0xFFF0FFFF
+#define S_028C10_S10_Y(x) (((x) & 0x0F) << 20)
+#define G_028C10_S10_Y(x) (((x) >> 20) & 0x0F)
+#define C_028C10_S10_Y 0xFF0FFFFF
+#define S_028C10_S11_X(x) (((x) & 0x0F) << 24)
+#define G_028C10_S11_X(x) (((x) >> 24) & 0x0F)
+#define C_028C10_S11_X 0xF0FFFFFF
+#define S_028C10_S11_Y(x) (((x) & 0x0F) << 28)
+#define G_028C10_S11_Y(x) (((x) >> 28) & 0x0F)
+#define C_028C10_S11_Y 0x0FFFFFFF
+#define R_028C14_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_3 0x028C14
+#define S_028C14_S12_X(x) (((x) & 0x0F) << 0)
+#define G_028C14_S12_X(x) (((x) >> 0) & 0x0F)
+#define C_028C14_S12_X 0xFFFFFFF0
+#define S_028C14_S12_Y(x) (((x) & 0x0F) << 4)
+#define G_028C14_S12_Y(x) (((x) >> 4) & 0x0F)
+#define C_028C14_S12_Y 0xFFFFFF0F
+#define S_028C14_S13_X(x) (((x) & 0x0F) << 8)
+#define G_028C14_S13_X(x) (((x) >> 8) & 0x0F)
+#define C_028C14_S13_X 0xFFFFF0FF
+#define S_028C14_S13_Y(x) (((x) & 0x0F) << 12)
+#define G_028C14_S13_Y(x) (((x) >> 12) & 0x0F)
+#define C_028C14_S13_Y 0xFFFF0FFF
+#define S_028C14_S14_X(x) (((x) & 0x0F) << 16)
+#define G_028C14_S14_X(x) (((x) >> 16) & 0x0F)
+#define C_028C14_S14_X 0xFFF0FFFF
+#define S_028C14_S14_Y(x) (((x) & 0x0F) << 20)
+#define G_028C14_S14_Y(x) (((x) >> 20) & 0x0F)
+#define C_028C14_S14_Y 0xFF0FFFFF
+#define S_028C14_S15_X(x) (((x) & 0x0F) << 24)
+#define G_028C14_S15_X(x) (((x) >> 24) & 0x0F)
+#define C_028C14_S15_X 0xF0FFFFFF
+#define S_028C14_S15_Y(x) (((x) & 0x0F) << 28)
+#define G_028C14_S15_Y(x) (((x) >> 28) & 0x0F)
+#define C_028C14_S15_Y 0x0FFFFFFF
+#define R_028C18_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_0 0x028C18
+#define S_028C18_S0_X(x) (((x) & 0x0F) << 0)
+#define G_028C18_S0_X(x) (((x) >> 0) & 0x0F)
+#define C_028C18_S0_X 0xFFFFFFF0
+#define S_028C18_S0_Y(x) (((x) & 0x0F) << 4)
+#define G_028C18_S0_Y(x) (((x) >> 4) & 0x0F)
+#define C_028C18_S0_Y 0xFFFFFF0F
+#define S_028C18_S1_X(x) (((x) & 0x0F) << 8)
+#define G_028C18_S1_X(x) (((x) >> 8) & 0x0F)
+#define C_028C18_S1_X 0xFFFFF0FF
+#define S_028C18_S1_Y(x) (((x) & 0x0F) << 12)
+#define G_028C18_S1_Y(x) (((x) >> 12) & 0x0F)
+#define C_028C18_S1_Y 0xFFFF0FFF
+#define S_028C18_S2_X(x) (((x) & 0x0F) << 16)
+#define G_028C18_S2_X(x) (((x) >> 16) & 0x0F)
+#define C_028C18_S2_X 0xFFF0FFFF
+#define S_028C18_S2_Y(x) (((x) & 0x0F) << 20)
+#define G_028C18_S2_Y(x) (((x) >> 20) & 0x0F)
+#define C_028C18_S2_Y 0xFF0FFFFF
+#define S_028C18_S3_X(x) (((x) & 0x0F) << 24)
+#define G_028C18_S3_X(x) (((x) >> 24) & 0x0F)
+#define C_028C18_S3_X 0xF0FFFFFF
+#define S_028C18_S3_Y(x) (((x) & 0x0F) << 28)
+#define G_028C18_S3_Y(x) (((x) >> 28) & 0x0F)
+#define C_028C18_S3_Y 0x0FFFFFFF
+#define R_028C1C_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_1 0x028C1C
+#define S_028C1C_S4_X(x) (((x) & 0x0F) << 0)
+#define G_028C1C_S4_X(x) (((x) >> 0) & 0x0F)
+#define C_028C1C_S4_X 0xFFFFFFF0
+#define S_028C1C_S4_Y(x) (((x) & 0x0F) << 4)
+#define G_028C1C_S4_Y(x) (((x) >> 4) & 0x0F)
+#define C_028C1C_S4_Y 0xFFFFFF0F
+#define S_028C1C_S5_X(x) (((x) & 0x0F) << 8)
+#define G_028C1C_S5_X(x) (((x) >> 8) & 0x0F)
+#define C_028C1C_S5_X 0xFFFFF0FF
+#define S_028C1C_S5_Y(x) (((x) & 0x0F) << 12)
+#define G_028C1C_S5_Y(x) (((x) >> 12) & 0x0F)
+#define C_028C1C_S5_Y 0xFFFF0FFF
+#define S_028C1C_S6_X(x) (((x) & 0x0F) << 16)
+#define G_028C1C_S6_X(x) (((x) >> 16) & 0x0F)
+#define C_028C1C_S6_X 0xFFF0FFFF
+#define S_028C1C_S6_Y(x) (((x) & 0x0F) << 20)
+#define G_028C1C_S6_Y(x) (((x) >> 20) & 0x0F)
+#define C_028C1C_S6_Y 0xFF0FFFFF
+#define S_028C1C_S7_X(x) (((x) & 0x0F) << 24)
+#define G_028C1C_S7_X(x) (((x) >> 24) & 0x0F)
+#define C_028C1C_S7_X 0xF0FFFFFF
+#define S_028C1C_S7_Y(x) (((x) & 0x0F) << 28)
+#define G_028C1C_S7_Y(x) (((x) >> 28) & 0x0F)
+#define C_028C1C_S7_Y 0x0FFFFFFF
+#define R_028C20_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_2 0x028C20
+#define S_028C20_S8_X(x) (((x) & 0x0F) << 0)
+#define G_028C20_S8_X(x) (((x) >> 0) & 0x0F)
+#define C_028C20_S8_X 0xFFFFFFF0
+#define S_028C20_S8_Y(x) (((x) & 0x0F) << 4)
+#define G_028C20_S8_Y(x) (((x) >> 4) & 0x0F)
+#define C_028C20_S8_Y 0xFFFFFF0F
+#define S_028C20_S9_X(x) (((x) & 0x0F) << 8)
+#define G_028C20_S9_X(x) (((x) >> 8) & 0x0F)
+#define C_028C20_S9_X 0xFFFFF0FF
+#define S_028C20_S9_Y(x) (((x) & 0x0F) << 12)
+#define G_028C20_S9_Y(x) (((x) >> 12) & 0x0F)
+#define C_028C20_S9_Y 0xFFFF0FFF
+#define S_028C20_S10_X(x) (((x) & 0x0F) << 16)
+#define G_028C20_S10_X(x) (((x) >> 16) & 0x0F)
+#define C_028C20_S10_X 0xFFF0FFFF
+#define S_028C20_S10_Y(x) (((x) & 0x0F) << 20)
+#define G_028C20_S10_Y(x) (((x) >> 20) & 0x0F)
+#define C_028C20_S10_Y 0xFF0FFFFF
+#define S_028C20_S11_X(x) (((x) & 0x0F) << 24)
+#define G_028C20_S11_X(x) (((x) >> 24) & 0x0F)
+#define C_028C20_S11_X 0xF0FFFFFF
+#define S_028C20_S11_Y(x) (((x) & 0x0F) << 28)
+#define G_028C20_S11_Y(x) (((x) >> 28) & 0x0F)
+#define C_028C20_S11_Y 0x0FFFFFFF
+#define R_028C24_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_3 0x028C24
+#define S_028C24_S12_X(x) (((x) & 0x0F) << 0)
+#define G_028C24_S12_X(x) (((x) >> 0) & 0x0F)
+#define C_028C24_S12_X 0xFFFFFFF0
+#define S_028C24_S12_Y(x) (((x) & 0x0F) << 4)
+#define G_028C24_S12_Y(x) (((x) >> 4) & 0x0F)
+#define C_028C24_S12_Y 0xFFFFFF0F
+#define S_028C24_S13_X(x) (((x) & 0x0F) << 8)
+#define G_028C24_S13_X(x) (((x) >> 8) & 0x0F)
+#define C_028C24_S13_X 0xFFFFF0FF
+#define S_028C24_S13_Y(x) (((x) & 0x0F) << 12)
+#define G_028C24_S13_Y(x) (((x) >> 12) & 0x0F)
+#define C_028C24_S13_Y 0xFFFF0FFF
+#define S_028C24_S14_X(x) (((x) & 0x0F) << 16)
+#define G_028C24_S14_X(x) (((x) >> 16) & 0x0F)
+#define C_028C24_S14_X 0xFFF0FFFF
+#define S_028C24_S14_Y(x) (((x) & 0x0F) << 20)
+#define G_028C24_S14_Y(x) (((x) >> 20) & 0x0F)
+#define C_028C24_S14_Y 0xFF0FFFFF
+#define S_028C24_S15_X(x) (((x) & 0x0F) << 24)
+#define G_028C24_S15_X(x) (((x) >> 24) & 0x0F)
+#define C_028C24_S15_X 0xF0FFFFFF
+#define S_028C24_S15_Y(x) (((x) & 0x0F) << 28)
+#define G_028C24_S15_Y(x) (((x) >> 28) & 0x0F)
+#define C_028C24_S15_Y 0x0FFFFFFF
+#define R_028C28_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_0 0x028C28
+#define S_028C28_S0_X(x) (((x) & 0x0F) << 0)
+#define G_028C28_S0_X(x) (((x) >> 0) & 0x0F)
+#define C_028C28_S0_X 0xFFFFFFF0
+#define S_028C28_S0_Y(x) (((x) & 0x0F) << 4)
+#define G_028C28_S0_Y(x) (((x) >> 4) & 0x0F)
+#define C_028C28_S0_Y 0xFFFFFF0F
+#define S_028C28_S1_X(x) (((x) & 0x0F) << 8)
+#define G_028C28_S1_X(x) (((x) >> 8) & 0x0F)
+#define C_028C28_S1_X 0xFFFFF0FF
+#define S_028C28_S1_Y(x) (((x) & 0x0F) << 12)
+#define G_028C28_S1_Y(x) (((x) >> 12) & 0x0F)
+#define C_028C28_S1_Y 0xFFFF0FFF
+#define S_028C28_S2_X(x) (((x) & 0x0F) << 16)
+#define G_028C28_S2_X(x) (((x) >> 16) & 0x0F)
+#define C_028C28_S2_X 0xFFF0FFFF
+#define S_028C28_S2_Y(x) (((x) & 0x0F) << 20)
+#define G_028C28_S2_Y(x) (((x) >> 20) & 0x0F)
+#define C_028C28_S2_Y 0xFF0FFFFF
+#define S_028C28_S3_X(x) (((x) & 0x0F) << 24)
+#define G_028C28_S3_X(x) (((x) >> 24) & 0x0F)
+#define C_028C28_S3_X 0xF0FFFFFF
+#define S_028C28_S3_Y(x) (((x) & 0x0F) << 28)
+#define G_028C28_S3_Y(x) (((x) >> 28) & 0x0F)
+#define C_028C28_S3_Y 0x0FFFFFFF
+#define R_028C2C_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_1 0x028C2C
+#define S_028C2C_S4_X(x) (((x) & 0x0F) << 0)
+#define G_028C2C_S4_X(x) (((x) >> 0) & 0x0F)
+#define C_028C2C_S4_X 0xFFFFFFF0
+#define S_028C2C_S4_Y(x) (((x) & 0x0F) << 4)
+#define G_028C2C_S4_Y(x) (((x) >> 4) & 0x0F)
+#define C_028C2C_S4_Y 0xFFFFFF0F
+#define S_028C2C_S5_X(x) (((x) & 0x0F) << 8)
+#define G_028C2C_S5_X(x) (((x) >> 8) & 0x0F)
+#define C_028C2C_S5_X 0xFFFFF0FF
+#define S_028C2C_S5_Y(x) (((x) & 0x0F) << 12)
+#define G_028C2C_S5_Y(x) (((x) >> 12) & 0x0F)
+#define C_028C2C_S5_Y 0xFFFF0FFF
+#define S_028C2C_S6_X(x) (((x) & 0x0F) << 16)
+#define G_028C2C_S6_X(x) (((x) >> 16) & 0x0F)
+#define C_028C2C_S6_X 0xFFF0FFFF
+#define S_028C2C_S6_Y(x) (((x) & 0x0F) << 20)
+#define G_028C2C_S6_Y(x) (((x) >> 20) & 0x0F)
+#define C_028C2C_S6_Y 0xFF0FFFFF
+#define S_028C2C_S7_X(x) (((x) & 0x0F) << 24)
+#define G_028C2C_S7_X(x) (((x) >> 24) & 0x0F)
+#define C_028C2C_S7_X 0xF0FFFFFF
+#define S_028C2C_S7_Y(x) (((x) & 0x0F) << 28)
+#define G_028C2C_S7_Y(x) (((x) >> 28) & 0x0F)
+#define C_028C2C_S7_Y 0x0FFFFFFF
+#define R_028C30_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_2 0x028C30
+#define S_028C30_S8_X(x) (((x) & 0x0F) << 0)
+#define G_028C30_S8_X(x) (((x) >> 0) & 0x0F)
+#define C_028C30_S8_X 0xFFFFFFF0
+#define S_028C30_S8_Y(x) (((x) & 0x0F) << 4)
+#define G_028C30_S8_Y(x) (((x) >> 4) & 0x0F)
+#define C_028C30_S8_Y 0xFFFFFF0F
+#define S_028C30_S9_X(x) (((x) & 0x0F) << 8)
+#define G_028C30_S9_X(x) (((x) >> 8) & 0x0F)
+#define C_028C30_S9_X 0xFFFFF0FF
+#define S_028C30_S9_Y(x) (((x) & 0x0F) << 12)
+#define G_028C30_S9_Y(x) (((x) >> 12) & 0x0F)
+#define C_028C30_S9_Y 0xFFFF0FFF
+#define S_028C30_S10_X(x) (((x) & 0x0F) << 16)
+#define G_028C30_S10_X(x) (((x) >> 16) & 0x0F)
+#define C_028C30_S10_X 0xFFF0FFFF
+#define S_028C30_S10_Y(x) (((x) & 0x0F) << 20)
+#define G_028C30_S10_Y(x) (((x) >> 20) & 0x0F)
+#define C_028C30_S10_Y 0xFF0FFFFF
+#define S_028C30_S11_X(x) (((x) & 0x0F) << 24)
+#define G_028C30_S11_X(x) (((x) >> 24) & 0x0F)
+#define C_028C30_S11_X 0xF0FFFFFF
+#define S_028C30_S11_Y(x) (((x) & 0x0F) << 28)
+#define G_028C30_S11_Y(x) (((x) >> 28) & 0x0F)
+#define C_028C30_S11_Y 0x0FFFFFFF
+#define R_028C34_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_3 0x028C34
+#define S_028C34_S12_X(x) (((x) & 0x0F) << 0)
+#define G_028C34_S12_X(x) (((x) >> 0) & 0x0F)
+#define C_028C34_S12_X 0xFFFFFFF0
+#define S_028C34_S12_Y(x) (((x) & 0x0F) << 4)
+#define G_028C34_S12_Y(x) (((x) >> 4) & 0x0F)
+#define C_028C34_S12_Y 0xFFFFFF0F
+#define S_028C34_S13_X(x) (((x) & 0x0F) << 8)
+#define G_028C34_S13_X(x) (((x) >> 8) & 0x0F)
+#define C_028C34_S13_X 0xFFFFF0FF
+#define S_028C34_S13_Y(x) (((x) & 0x0F) << 12)
+#define G_028C34_S13_Y(x) (((x) >> 12) & 0x0F)
+#define C_028C34_S13_Y 0xFFFF0FFF
+#define S_028C34_S14_X(x) (((x) & 0x0F) << 16)
+#define G_028C34_S14_X(x) (((x) >> 16) & 0x0F)
+#define C_028C34_S14_X 0xFFF0FFFF
+#define S_028C34_S14_Y(x) (((x) & 0x0F) << 20)
+#define G_028C34_S14_Y(x) (((x) >> 20) & 0x0F)
+#define C_028C34_S14_Y 0xFF0FFFFF
+#define S_028C34_S15_X(x) (((x) & 0x0F) << 24)
+#define G_028C34_S15_X(x) (((x) >> 24) & 0x0F)
+#define C_028C34_S15_X 0xF0FFFFFF
+#define S_028C34_S15_Y(x) (((x) & 0x0F) << 28)
+#define G_028C34_S15_Y(x) (((x) >> 28) & 0x0F)
+#define C_028C34_S15_Y 0x0FFFFFFF
+#define R_028C38_PA_SC_AA_MASK_X0Y0_X1Y0 0x028C38
+#define S_028C38_AA_MASK_X0Y0(x) (((x) & 0xFFFF) << 0)
+#define G_028C38_AA_MASK_X0Y0(x) (((x) >> 0) & 0xFFFF)
+#define C_028C38_AA_MASK_X0Y0 0xFFFF0000
+#define S_028C38_AA_MASK_X1Y0(x) (((x) & 0xFFFF) << 16)
+#define G_028C38_AA_MASK_X1Y0(x) (((x) >> 16) & 0xFFFF)
+#define C_028C38_AA_MASK_X1Y0 0x0000FFFF
+#define R_028C3C_PA_SC_AA_MASK_X0Y1_X1Y1 0x028C3C
+#define S_028C3C_AA_MASK_X0Y1(x) (((x) & 0xFFFF) << 0)
+#define G_028C3C_AA_MASK_X0Y1(x) (((x) >> 0) & 0xFFFF)
+#define C_028C3C_AA_MASK_X0Y1 0xFFFF0000
+#define S_028C3C_AA_MASK_X1Y1(x) (((x) & 0xFFFF) << 16)
+#define G_028C3C_AA_MASK_X1Y1(x) (((x) >> 16) & 0xFFFF)
+#define C_028C3C_AA_MASK_X1Y1 0x0000FFFF
+#define R_028C58_VGT_VERTEX_REUSE_BLOCK_CNTL 0x028C58
+#define S_028C58_VTX_REUSE_DEPTH(x) (((x) & 0xFF) << 0)
+#define G_028C58_VTX_REUSE_DEPTH(x) (((x) >> 0) & 0xFF)
+#define C_028C58_VTX_REUSE_DEPTH 0xFFFFFF00
+#define R_028C5C_VGT_OUT_DEALLOC_CNTL 0x028C5C
+#define S_028C5C_DEALLOC_DIST(x) (((x) & 0x7F) << 0)
+#define G_028C5C_DEALLOC_DIST(x) (((x) >> 0) & 0x7F)
+#define C_028C5C_DEALLOC_DIST 0xFFFFFF80
+#define R_028C60_CB_COLOR0_BASE 0x028C60
+#define R_028C64_CB_COLOR0_PITCH 0x028C64
+#define S_028C64_TILE_MAX(x) (((x) & 0x7FF) << 0)
+#define G_028C64_TILE_MAX(x) (((x) >> 0) & 0x7FF)
+#define C_028C64_TILE_MAX 0xFFFFF800
+#define R_028C68_CB_COLOR0_SLICE 0x028C68
+#define S_028C68_TILE_MAX(x) (((x) & 0x3FFFFF) << 0)
+#define G_028C68_TILE_MAX(x) (((x) >> 0) & 0x3FFFFF)
+#define C_028C68_TILE_MAX 0xFFC00000
+#define R_028C6C_CB_COLOR0_VIEW 0x028C6C
+#define S_028C6C_SLICE_START(x) (((x) & 0x7FF) << 0)
+#define G_028C6C_SLICE_START(x) (((x) >> 0) & 0x7FF)
+#define C_028C6C_SLICE_START 0xFFFFF800
+#define S_028C6C_SLICE_MAX(x) (((x) & 0x7FF) << 13)
+#define G_028C6C_SLICE_MAX(x) (((x) >> 13) & 0x7FF)
+#define C_028C6C_SLICE_MAX 0xFF001FFF
+#define R_028C70_CB_COLOR0_INFO 0x028C70
+#define S_028C70_ENDIAN(x) (((x) & 0x03) << 0)
+#define G_028C70_ENDIAN(x) (((x) >> 0) & 0x03)
+#define C_028C70_ENDIAN 0xFFFFFFFC
+#define V_028C70_ENDIAN_NONE 0x00
+#define V_028C70_ENDIAN_8IN16 0x01
+#define V_028C70_ENDIAN_8IN32 0x02
+#define V_028C70_ENDIAN_8IN64 0x03
+#define S_028C70_FORMAT(x) (((x) & 0x1F) << 2)
+#define G_028C70_FORMAT(x) (((x) >> 2) & 0x1F)
+#define C_028C70_FORMAT 0xFFFFFF83
+#define V_028C70_COLOR_INVALID 0x00
+#define V_028C70_COLOR_8 0x01
+#define V_028C70_COLOR_16 0x02
+#define V_028C70_COLOR_8_8 0x03
+#define V_028C70_COLOR_32 0x04
+#define V_028C70_COLOR_16_16 0x05
+#define V_028C70_COLOR_10_11_11 0x06
+#define V_028C70_COLOR_11_11_10 0x07
+#define V_028C70_COLOR_10_10_10_2 0x08
+#define V_028C70_COLOR_2_10_10_10 0x09
+#define V_028C70_COLOR_8_8_8_8 0x0A
+#define V_028C70_COLOR_32_32 0x0B
+#define V_028C70_COLOR_16_16_16_16 0x0C
+#define V_028C70_COLOR_32_32_32_32 0x0E
+#define V_028C70_COLOR_5_6_5 0x10
+#define V_028C70_COLOR_1_5_5_5 0x11
+#define V_028C70_COLOR_5_5_5_1 0x12
+#define V_028C70_COLOR_4_4_4_4 0x13
+#define V_028C70_COLOR_8_24 0x14
+#define V_028C70_COLOR_24_8 0x15
+#define V_028C70_COLOR_X24_8_32_FLOAT 0x16
+#define S_028C70_LINEAR_GENERAL(x) (((x) & 0x1) << 7)
+#define G_028C70_LINEAR_GENERAL(x) (((x) >> 7) & 0x1)
+#define C_028C70_LINEAR_GENERAL 0xFFFFFF7F
+#define S_028C70_NUMBER_TYPE(x) (((x) & 0x07) << 8)
+#define G_028C70_NUMBER_TYPE(x) (((x) >> 8) & 0x07)
+#define C_028C70_NUMBER_TYPE 0xFFFFF8FF
+#define V_028C70_NUMBER_UNORM 0x00
+#define V_028C70_NUMBER_SNORM 0x01
+#define V_028C70_NUMBER_UINT 0x04
+#define V_028C70_NUMBER_SINT 0x05
+#define V_028C70_NUMBER_SRGB 0x06
+#define V_028C70_NUMBER_FLOAT 0x07
+#define S_028C70_COMP_SWAP(x) (((x) & 0x03) << 11)
+#define G_028C70_COMP_SWAP(x) (((x) >> 11) & 0x03)
+#define C_028C70_COMP_SWAP 0xFFFFE7FF
+#define V_028C70_SWAP_STD 0x00
+#define V_028C70_SWAP_ALT 0x01
+#define V_028C70_SWAP_STD_REV 0x02
+#define V_028C70_SWAP_ALT_REV 0x03
+#define S_028C70_FAST_CLEAR(x) (((x) & 0x1) << 13)
+#define G_028C70_FAST_CLEAR(x) (((x) >> 13) & 0x1)
+#define C_028C70_FAST_CLEAR 0xFFFFDFFF
+#define S_028C70_COMPRESSION(x) (((x) & 0x1) << 14)
+#define G_028C70_COMPRESSION(x) (((x) >> 14) & 0x1)
+#define C_028C70_COMPRESSION 0xFFFFBFFF
+#define S_028C70_BLEND_CLAMP(x) (((x) & 0x1) << 15)
+#define G_028C70_BLEND_CLAMP(x) (((x) >> 15) & 0x1)
+#define C_028C70_BLEND_CLAMP 0xFFFF7FFF
+#define S_028C70_BLEND_BYPASS(x) (((x) & 0x1) << 16)
+#define G_028C70_BLEND_BYPASS(x) (((x) >> 16) & 0x1)
+#define C_028C70_BLEND_BYPASS 0xFFFEFFFF
+#define S_028C70_SIMPLE_FLOAT(x) (((x) & 0x1) << 17)
+#define G_028C70_SIMPLE_FLOAT(x) (((x) >> 17) & 0x1)
+#define C_028C70_SIMPLE_FLOAT 0xFFFDFFFF
+#define S_028C70_ROUND_MODE(x) (((x) & 0x1) << 18)
+#define G_028C70_ROUND_MODE(x) (((x) >> 18) & 0x1)
+#define C_028C70_ROUND_MODE 0xFFFBFFFF
+#define S_028C70_CMASK_IS_LINEAR(x) (((x) & 0x1) << 19)
+#define G_028C70_CMASK_IS_LINEAR(x) (((x) >> 19) & 0x1)
+#define C_028C70_CMASK_IS_LINEAR 0xFFF7FFFF
+#define S_028C70_BLEND_OPT_DONT_RD_DST(x) (((x) & 0x07) << 20)
+#define G_028C70_BLEND_OPT_DONT_RD_DST(x) (((x) >> 20) & 0x07)
+#define C_028C70_BLEND_OPT_DONT_RD_DST 0xFF8FFFFF
+#define V_028C70_FORCE_OPT_AUTO 0x00
+#define V_028C70_FORCE_OPT_DISABLE 0x01
+#define V_028C70_FORCE_OPT_ENABLE_IF_SRC_A_0 0x02
+#define V_028C70_FORCE_OPT_ENABLE_IF_SRC_RGB_0 0x03
+#define V_028C70_FORCE_OPT_ENABLE_IF_SRC_ARGB_0 0x04
+#define V_028C70_FORCE_OPT_ENABLE_IF_SRC_A_1 0x05
+#define V_028C70_FORCE_OPT_ENABLE_IF_SRC_RGB_1 0x06
+#define V_028C70_FORCE_OPT_ENABLE_IF_SRC_ARGB_1 0x07
+#define S_028C70_BLEND_OPT_DISCARD_PIXEL(x) (((x) & 0x07) << 23)
+#define G_028C70_BLEND_OPT_DISCARD_PIXEL(x) (((x) >> 23) & 0x07)
+#define C_028C70_BLEND_OPT_DISCARD_PIXEL 0xFC7FFFFF
+#define V_028C70_FORCE_OPT_AUTO 0x00
+#define V_028C70_FORCE_OPT_DISABLE 0x01
+#define V_028C70_FORCE_OPT_ENABLE_IF_SRC_A_0 0x02
+#define V_028C70_FORCE_OPT_ENABLE_IF_SRC_RGB_0 0x03
+#define V_028C70_FORCE_OPT_ENABLE_IF_SRC_ARGB_0 0x04
+#define V_028C70_FORCE_OPT_ENABLE_IF_SRC_A_1 0x05
+#define V_028C70_FORCE_OPT_ENABLE_IF_SRC_RGB_1 0x06
+#define V_028C70_FORCE_OPT_ENABLE_IF_SRC_ARGB_1 0x07
+#define R_028C74_CB_COLOR0_ATTRIB 0x028C74
+#define S_028C74_TILE_MODE_INDEX(x) (((x) & 0x1F) << 0)
+#define G_028C74_TILE_MODE_INDEX(x) (((x) >> 0) & 0x1F)
+#define C_028C74_TILE_MODE_INDEX 0xFFFFFFE0
+#define S_028C74_FMASK_TILE_MODE_INDEX(x) (((x) & 0x1F) << 5)
+#define G_028C74_FMASK_TILE_MODE_INDEX(x) (((x) >> 5) & 0x1F)
+#define C_028C74_FMASK_TILE_MODE_INDEX 0xFFFFFC1F
+#define S_028C74_NUM_SAMPLES(x) (((x) & 0x07) << 12)
+#define G_028C74_NUM_SAMPLES(x) (((x) >> 12) & 0x07)
+#define C_028C74_NUM_SAMPLES 0xFFFF8FFF
+#define S_028C74_NUM_FRAGMENTS(x) (((x) & 0x03) << 15)
+#define G_028C74_NUM_FRAGMENTS(x) (((x) >> 15) & 0x03)
+#define C_028C74_NUM_FRAGMENTS 0xFFFE7FFF
+#define S_028C74_FORCE_DST_ALPHA_1(x) (((x) & 0x1) << 17)
+#define G_028C74_FORCE_DST_ALPHA_1(x) (((x) >> 17) & 0x1)
+#define C_028C74_FORCE_DST_ALPHA_1 0xFFFDFFFF
+#define R_028C7C_CB_COLOR0_CMASK 0x028C7C
+#define R_028C80_CB_COLOR0_CMASK_SLICE 0x028C80
+#define S_028C80_TILE_MAX(x) (((x) & 0x3FFF) << 0)
+#define G_028C80_TILE_MAX(x) (((x) >> 0) & 0x3FFF)
+#define C_028C80_TILE_MAX 0xFFFFC000
+#define R_028C84_CB_COLOR0_FMASK 0x028C84
+#define R_028C88_CB_COLOR0_FMASK_SLICE 0x028C88
+#define S_028C88_TILE_MAX(x) (((x) & 0x3FFFFF) << 0)
+#define G_028C88_TILE_MAX(x) (((x) >> 0) & 0x3FFFFF)
+#define C_028C88_TILE_MAX 0xFFC00000
+#define R_028C8C_CB_COLOR0_CLEAR_WORD0 0x028C8C
+#define R_028C90_CB_COLOR0_CLEAR_WORD1 0x028C90
+#define R_028C9C_CB_COLOR1_BASE 0x028C9C
+#define R_028CA0_CB_COLOR1_PITCH 0x028CA0
+#define R_028CA4_CB_COLOR1_SLICE 0x028CA4
+#define R_028CA8_CB_COLOR1_VIEW 0x028CA8
+#define R_028CAC_CB_COLOR1_INFO 0x028CAC
+#define R_028CB0_CB_COLOR1_ATTRIB 0x028CB0
+#define R_028CD4_CB_COLOR1_CMASK 0x028CB8
+#define R_028CBC_CB_COLOR1_CMASK_SLICE 0x028CBC
+#define R_028CC0_CB_COLOR1_FMASK 0x028CC0
+#define R_028CC4_CB_COLOR1_FMASK_SLICE 0x028CC4
+#define R_028CC8_CB_COLOR1_CLEAR_WORD0 0x028CC8
+#define R_028CCC_CB_COLOR1_CLEAR_WORD1 0x028CCC
+#define R_028CD8_CB_COLOR2_BASE 0x028CD8
+#define R_028CDC_CB_COLOR2_PITCH 0x028CDC
+#define R_028CE0_CB_COLOR2_SLICE 0x028CE0
+#define R_028CE4_CB_COLOR2_VIEW 0x028CE4
+#define R_028CE8_CB_COLOR2_INFO 0x028CE8
+#define R_028CEC_CB_COLOR2_ATTRIB 0x028CEC
+#define R_028CF4_CB_COLOR2_CMASK 0x028CF4
+#define R_028CF8_CB_COLOR2_CMASK_SLICE 0x028CF8
+#define R_028CFC_CB_COLOR2_FMASK 0x028CFC
+#define R_028D00_CB_COLOR2_FMASK_SLICE 0x028D00
+#define R_028D04_CB_COLOR2_CLEAR_WORD0 0x028D04
+#define R_028D08_CB_COLOR2_CLEAR_WORD1 0x028D08
+#define R_028D14_CB_COLOR3_BASE 0x028D14
+#define R_028D18_CB_COLOR3_PITCH 0x028D18
+#define R_028D1C_CB_COLOR3_SLICE 0x028D1C
+#define R_028D20_CB_COLOR3_VIEW 0x028D20
+#define R_028D24_CB_COLOR3_INFO 0x028D24
+#define R_028D28_CB_COLOR3_ATTRIB 0x028D28
+#define R_028D30_CB_COLOR3_CMASK 0x028D30
+#define R_028D34_CB_COLOR3_CMASK_SLICE 0x028D34
+#define R_028D38_CB_COLOR3_FMASK 0x028D38
+#define R_028D3C_CB_COLOR3_FMASK_SLICE 0x028D3C
+#define R_028D40_CB_COLOR3_CLEAR_WORD0 0x028D40
+#define R_028D44_CB_COLOR3_CLEAR_WORD1 0x028D44
+#define R_028D50_CB_COLOR4_BASE 0x028D50
+#define R_028D54_CB_COLOR4_PITCH 0x028D54
+#define R_028D58_CB_COLOR4_SLICE 0x028D58
+#define R_028D5C_CB_COLOR4_VIEW 0x028D5C
+#define R_028D60_CB_COLOR4_INFO 0x028D60
+#define R_028D64_CB_COLOR4_ATTRIB 0x028D64
+#define R_028D6C_CB_COLOR4_CMASK 0x028D6C
+#define R_028D70_CB_COLOR4_CMASK_SLICE 0x028D70
+#define R_028D74_CB_COLOR4_FMASK 0x028D74
+#define R_028D78_CB_COLOR4_FMASK_SLICE 0x028D78
+#define R_028D7C_CB_COLOR4_CLEAR_WORD0 0x028D7C
+#define R_028D80_CB_COLOR4_CLEAR_WORD1 0x028D80
+#define R_028D8C_CB_COLOR5_BASE 0x028D8C
+#define R_028D90_CB_COLOR5_PITCH 0x028D90
+#define R_028D94_CB_COLOR5_SLICE 0x028D94
+#define R_028D98_CB_COLOR5_VIEW 0x028D98
+#define R_028D9C_CB_COLOR5_INFO 0x028D9C
+#define R_028DA0_CB_COLOR5_ATTRIB 0x028DA0
+#define R_028DA8_CB_COLOR5_CMASK 0x028DA8
+#define R_028DAC_CB_COLOR5_CMASK_SLICE 0x028DAC
+#define R_028DB0_CB_COLOR5_FMASK 0x028DB0
+#define R_028DB4_CB_COLOR5_FMASK_SLICE 0x028DB4
+#define R_028DB8_CB_COLOR5_CLEAR_WORD0 0x028DB8
+#define R_028DBC_CB_COLOR5_CLEAR_WORD1 0x028DBC
+#define R_028DC8_CB_COLOR6_BASE 0x028DC8
+#define R_028DCC_CB_COLOR6_PITCH 0x028DCC
+#define R_028DD0_CB_COLOR6_SLICE 0x028DD0
+#define R_028DD4_CB_COLOR6_VIEW 0x028DD4
+#define R_028DD8_CB_COLOR6_INFO 0x028DD8
+#define R_028DDC_CB_COLOR6_ATTRIB 0x028DDC
+#define R_028DE4_CB_COLOR6_CMASK 0x028DE4
+#define R_028DE8_CB_COLOR6_CMASK_SLICE 0x028DE8
+#define R_028DEC_CB_COLOR6_FMASK 0x028DEC
+#define R_028DF0_CB_COLOR6_FMASK_SLICE 0x028DF0
+#define R_028DF4_CB_COLOR6_CLEAR_WORD0 0x028DF4
+#define R_028DF8_CB_COLOR6_CLEAR_WORD1 0x028DF8
+#define R_028E04_CB_COLOR7_BASE 0x028E04
+#define R_028E08_CB_COLOR7_PITCH 0x028E08
+#define R_028E0C_CB_COLOR7_SLICE 0x028E0C
+#define R_028E10_CB_COLOR7_VIEW 0x028E10
+#define R_028E14_CB_COLOR7_INFO 0x028E14
+#define R_028E18_CB_COLOR7_ATTRIB 0x028E18
+#define R_028E20_CB_COLOR7_CMASK 0x028E20
+#define R_028E24_CB_COLOR7_CMASK_SLICE 0x028E24
+#define R_028E28_CB_COLOR7_FMASK 0x028E28
+#define R_028E2C_CB_COLOR7_FMASK_SLICE 0x028E2C
+#define R_028E30_CB_COLOR7_CLEAR_WORD0 0x028E30
+#define R_028E34_CB_COLOR7_CLEAR_WORD1 0x028E34
+
+#endif /* _SID_H */
+