#include "evergreen_compute.h"
#include "util/u_math.h"
-static INLINE unsigned evergreen_array_mode(unsigned mode)
+static inline unsigned evergreen_array_mode(unsigned mode)
{
switch (mode) {
case RADEON_SURF_MODE_LINEAR_ALIGNED: return V_028C70_ARRAY_LINEAR_ALIGNED;
}
}
-static uint32_t r600_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_SNORM:
- case PIPE_FORMAT_A8_UINT:
- case PIPE_FORMAT_A8_SINT:
- case PIPE_FORMAT_A16_UNORM:
- case PIPE_FORMAT_A16_SNORM:
- case PIPE_FORMAT_A16_UINT:
- case PIPE_FORMAT_A16_SINT:
- case PIPE_FORMAT_A16_FLOAT:
- case PIPE_FORMAT_A32_UINT:
- case PIPE_FORMAT_A32_SINT:
- case PIPE_FORMAT_A32_FLOAT:
- case PIPE_FORMAT_R4A4_UNORM:
- return V_028C70_SWAP_ALT_REV;
- case PIPE_FORMAT_I8_UNORM:
- case PIPE_FORMAT_I8_SNORM:
- case PIPE_FORMAT_I8_UINT:
- case PIPE_FORMAT_I8_SINT:
- case PIPE_FORMAT_I16_UNORM:
- case PIPE_FORMAT_I16_SNORM:
- case PIPE_FORMAT_I16_UINT:
- case PIPE_FORMAT_I16_SINT:
- case PIPE_FORMAT_I16_FLOAT:
- case PIPE_FORMAT_I32_UINT:
- case PIPE_FORMAT_I32_SINT:
- case PIPE_FORMAT_I32_FLOAT:
- case PIPE_FORMAT_L8_UNORM:
- case PIPE_FORMAT_L8_SNORM:
- case PIPE_FORMAT_L8_UINT:
- case PIPE_FORMAT_L8_SINT:
- case PIPE_FORMAT_L8_SRGB:
- case PIPE_FORMAT_L16_UNORM:
- case PIPE_FORMAT_L16_SNORM:
- case PIPE_FORMAT_L16_UINT:
- case PIPE_FORMAT_L16_SINT:
- case PIPE_FORMAT_L16_FLOAT:
- case PIPE_FORMAT_L32_UINT:
- case PIPE_FORMAT_L32_SINT:
- case PIPE_FORMAT_L32_FLOAT:
- 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_SNORM:
- case PIPE_FORMAT_L8A8_UINT:
- case PIPE_FORMAT_L8A8_SINT:
- case PIPE_FORMAT_L8A8_SRGB:
- case PIPE_FORMAT_L16A16_UNORM:
- case PIPE_FORMAT_L16A16_SNORM:
- case PIPE_FORMAT_L16A16_UINT:
- case PIPE_FORMAT_L16A16_SINT:
- case PIPE_FORMAT_L16A16_FLOAT:
- case PIPE_FORMAT_L32A32_UINT:
- case PIPE_FORMAT_L32A32_SINT:
- case PIPE_FORMAT_L32A32_FLOAT:
- case PIPE_FORMAT_R8A8_UNORM:
- case PIPE_FORMAT_R8A8_SNORM:
- case PIPE_FORMAT_R8A8_UINT:
- case PIPE_FORMAT_R8A8_SINT:
- case PIPE_FORMAT_R16A16_UNORM:
- case PIPE_FORMAT_R16A16_SNORM:
- case PIPE_FORMAT_R16A16_UINT:
- case PIPE_FORMAT_R16A16_SINT:
- case PIPE_FORMAT_R16A16_FLOAT:
- case PIPE_FORMAT_R32A32_UINT:
- case PIPE_FORMAT_R32A32_SINT:
- case PIPE_FORMAT_R32A32_FLOAT:
- return V_028C70_SWAP_ALT;
- case PIPE_FORMAT_R8G8_UNORM:
- case PIPE_FORMAT_R8G8_SNORM:
- case PIPE_FORMAT_R8G8_UINT:
- case PIPE_FORMAT_R8G8_SINT:
- return V_028C70_SWAP_STD;
-
- case PIPE_FORMAT_R16_UNORM:
- case PIPE_FORMAT_R16_SNORM:
- 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_SINT:
- case PIPE_FORMAT_R8G8B8A8_UINT:
- case PIPE_FORMAT_R8G8B8X8_UNORM:
- case PIPE_FORMAT_R8G8B8X8_SNORM:
- case PIPE_FORMAT_R8G8B8X8_SRGB:
- case PIPE_FORMAT_R8G8B8X8_UINT:
- case PIPE_FORMAT_R8G8B8X8_SINT:
- 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_REV;
-
- 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:
- case PIPE_FORMAT_B10G10R10X2_UNORM:
- 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_SNORM:
- 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_UINT:
- case PIPE_FORMAT_R16G16B16A16_SINT:
- case PIPE_FORMAT_R16G16B16A16_FLOAT:
- case PIPE_FORMAT_R16G16B16X16_UNORM:
- case PIPE_FORMAT_R16G16B16X16_SNORM:
- case PIPE_FORMAT_R16G16B16X16_FLOAT:
- case PIPE_FORMAT_R16G16B16X16_UINT:
- case PIPE_FORMAT_R16G16B16X16_SINT:
- 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_SINT:
- case PIPE_FORMAT_R32G32B32A32_UINT:
- case PIPE_FORMAT_R32G32B32X32_FLOAT:
- case PIPE_FORMAT_R32G32B32X32_UINT:
- case PIPE_FORMAT_R32G32B32X32_SINT:
- return V_028C70_SWAP_STD;
- default:
- R600_ERR("unsupported colorswap format %d\n", format);
- return ~0U;
- }
- return ~0U;
-}
-
-static uint32_t r600_translate_colorformat(enum pipe_format format)
-{
- switch (format) {
- /* 8-bit buffers. */
- case PIPE_FORMAT_A8_UNORM:
- case PIPE_FORMAT_A8_SNORM:
- case PIPE_FORMAT_A8_UINT:
- case PIPE_FORMAT_A8_SINT:
- case PIPE_FORMAT_I8_UNORM:
- case PIPE_FORMAT_I8_SNORM:
- case PIPE_FORMAT_I8_UINT:
- case PIPE_FORMAT_I8_SINT:
- case PIPE_FORMAT_L8_UNORM:
- case PIPE_FORMAT_L8_SNORM:
- 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_Z16_UNORM:
- return V_028C70_COLOR_16;
-
- case PIPE_FORMAT_L8A8_UNORM:
- case PIPE_FORMAT_L8A8_SNORM:
- case PIPE_FORMAT_L8A8_UINT:
- case PIPE_FORMAT_L8A8_SINT:
- case PIPE_FORMAT_L8A8_SRGB:
- case PIPE_FORMAT_R8G8_UNORM:
- case PIPE_FORMAT_R8G8_SNORM:
- case PIPE_FORMAT_R8G8_UINT:
- case PIPE_FORMAT_R8G8_SINT:
- case PIPE_FORMAT_R8A8_UNORM:
- case PIPE_FORMAT_R8A8_SNORM:
- case PIPE_FORMAT_R8A8_UINT:
- case PIPE_FORMAT_R8A8_SINT:
- return V_028C70_COLOR_8_8;
-
- case PIPE_FORMAT_R16_UNORM:
- case PIPE_FORMAT_R16_SNORM:
- case PIPE_FORMAT_R16_UINT:
- case PIPE_FORMAT_R16_SINT:
- case PIPE_FORMAT_A16_UNORM:
- case PIPE_FORMAT_A16_SNORM:
- case PIPE_FORMAT_A16_UINT:
- case PIPE_FORMAT_A16_SINT:
- case PIPE_FORMAT_L16_UNORM:
- case PIPE_FORMAT_L16_SNORM:
- case PIPE_FORMAT_L16_UINT:
- case PIPE_FORMAT_L16_SINT:
- case PIPE_FORMAT_I16_UNORM:
- case PIPE_FORMAT_I16_SNORM:
- case PIPE_FORMAT_I16_UINT:
- case PIPE_FORMAT_I16_SINT:
- return V_028C70_COLOR_16;
-
- case PIPE_FORMAT_R16_FLOAT:
- case PIPE_FORMAT_A16_FLOAT:
- case PIPE_FORMAT_L16_FLOAT:
- case PIPE_FORMAT_I16_FLOAT:
- return V_028C70_COLOR_16_FLOAT;
-
- /* 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_R8G8B8X8_SNORM:
- case PIPE_FORMAT_R8G8B8X8_SRGB:
- case PIPE_FORMAT_R8G8B8X8_UINT:
- case PIPE_FORMAT_R8G8B8X8_SINT:
- case PIPE_FORMAT_R8SG8SB8UX8U_NORM:
- case PIPE_FORMAT_X8B8G8R8_UNORM:
- case PIPE_FORMAT_X8R8G8B8_UNORM:
- case PIPE_FORMAT_R8G8B8_UNORM:
- 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_B10G10R10X2_UNORM:
- 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_UINT:
- case PIPE_FORMAT_R32_SINT:
- case PIPE_FORMAT_A32_UINT:
- case PIPE_FORMAT_A32_SINT:
- case PIPE_FORMAT_L32_UINT:
- case PIPE_FORMAT_L32_SINT:
- case PIPE_FORMAT_I32_UINT:
- case PIPE_FORMAT_I32_SINT:
- return V_028C70_COLOR_32;
-
- case PIPE_FORMAT_R32_FLOAT:
- case PIPE_FORMAT_A32_FLOAT:
- case PIPE_FORMAT_L32_FLOAT:
- case PIPE_FORMAT_I32_FLOAT:
- case PIPE_FORMAT_Z32_FLOAT:
- return V_028C70_COLOR_32_FLOAT;
-
- case PIPE_FORMAT_R16G16_FLOAT:
- case PIPE_FORMAT_L16A16_FLOAT:
- case PIPE_FORMAT_R16A16_FLOAT:
- return V_028C70_COLOR_16_16_FLOAT;
-
- case PIPE_FORMAT_R16G16_UNORM:
- case PIPE_FORMAT_R16G16_SNORM:
- case PIPE_FORMAT_R16G16_UINT:
- case PIPE_FORMAT_R16G16_SINT:
- case PIPE_FORMAT_L16A16_UNORM:
- case PIPE_FORMAT_L16A16_SNORM:
- case PIPE_FORMAT_L16A16_UINT:
- case PIPE_FORMAT_L16A16_SINT:
- case PIPE_FORMAT_R16A16_UNORM:
- case PIPE_FORMAT_R16A16_SNORM:
- case PIPE_FORMAT_R16A16_UINT:
- case PIPE_FORMAT_R16A16_SINT:
- return V_028C70_COLOR_16_16;
-
- case PIPE_FORMAT_R11G11B10_FLOAT:
- return V_028C70_COLOR_10_11_11_FLOAT;
-
- /* 64-bit buffers. */
- case PIPE_FORMAT_R16G16B16A16_UINT:
- case PIPE_FORMAT_R16G16B16A16_SINT:
- case PIPE_FORMAT_R16G16B16A16_UNORM:
- case PIPE_FORMAT_R16G16B16A16_SNORM:
- case PIPE_FORMAT_R16G16B16X16_UNORM:
- case PIPE_FORMAT_R16G16B16X16_SNORM:
- case PIPE_FORMAT_R16G16B16X16_UINT:
- case PIPE_FORMAT_R16G16B16X16_SINT:
- return V_028C70_COLOR_16_16_16_16;
-
- case PIPE_FORMAT_R16G16B16A16_FLOAT:
- case PIPE_FORMAT_R16G16B16X16_FLOAT:
- return V_028C70_COLOR_16_16_16_16_FLOAT;
-
- case PIPE_FORMAT_R32G32_FLOAT:
- case PIPE_FORMAT_L32A32_FLOAT:
- case PIPE_FORMAT_R32A32_FLOAT:
- return V_028C70_COLOR_32_32_FLOAT;
-
- case PIPE_FORMAT_R32G32_SINT:
- case PIPE_FORMAT_R32G32_UINT:
- case PIPE_FORMAT_L32A32_UINT:
- case PIPE_FORMAT_L32A32_SINT:
- return V_028C70_COLOR_32_32;
-
- /* 128-bit buffers. */
- case PIPE_FORMAT_R32G32B32A32_SNORM:
- case PIPE_FORMAT_R32G32B32A32_UNORM:
- case PIPE_FORMAT_R32G32B32A32_SINT:
- case PIPE_FORMAT_R32G32B32A32_UINT:
- case PIPE_FORMAT_R32G32B32X32_UINT:
- case PIPE_FORMAT_R32G32B32X32_SINT:
- return V_028C70_COLOR_32_32_32_32;
- case PIPE_FORMAT_R32G32B32A32_FLOAT:
- case PIPE_FORMAT_R32G32B32X32_FLOAT:
- return V_028C70_COLOR_32_32_32_32_FLOAT;
-
- /* YUV buffers. */
- case PIPE_FORMAT_UYVY:
- case PIPE_FORMAT_YUYV:
- default:
- return ~0U; /* Unsupported. */
- }
-}
-
-static uint32_t r600_colorformat_endian_swap(uint32_t colorformat)
-{
- if (R600_BIG_ENDIAN) {
- switch(colorformat) {
-
- /* 8-bit buffers. */
- case V_028C70_COLOR_8:
- return 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 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_32_FLOAT:
- case V_028C70_COLOR_16_16_FLOAT:
- case V_028C70_COLOR_16_16:
- return ENDIAN_8IN32;
-
- /* 64-bit buffers. */
- case V_028C70_COLOR_16_16_16_16:
- case V_028C70_COLOR_16_16_16_16_FLOAT:
- return ENDIAN_8IN16;
-
- case V_028C70_COLOR_32_32_FLOAT:
- case V_028C70_COLOR_32_32:
- case V_028C70_COLOR_X24_8_32_FLOAT:
- return ENDIAN_8IN32;
-
- /* 96-bit buffers. */
- case V_028C70_COLOR_32_32_32_FLOAT:
- /* 128-bit buffers. */
- case V_028C70_COLOR_32_32_32_32_FLOAT:
- case V_028C70_COLOR_32_32_32_32:
- return ENDIAN_8IN32;
- default:
- return ENDIAN_NONE; /* Unsupported. */
- }
- } else {
- return ENDIAN_NONE;
- }
-}
-
static bool r600_is_sampler_format_supported(struct pipe_screen *screen, enum pipe_format format)
{
return r600_translate_texformat(screen, format, NULL, NULL, NULL) != ~0U;
}
-static bool r600_is_colorbuffer_format_supported(enum pipe_format format)
+static bool r600_is_colorbuffer_format_supported(enum chip_class chip, enum pipe_format format)
{
- return r600_translate_colorformat(format) != ~0U &&
+ return r600_translate_colorformat(chip, format) != ~0U &&
r600_translate_colorswap(format) != ~0U;
}
}
}
- if ((usage & PIPE_BIND_SAMPLER_VIEW) &&
- r600_is_sampler_format_supported(screen, format)) {
- retval |= PIPE_BIND_SAMPLER_VIEW;
+ if (usage & PIPE_BIND_SAMPLER_VIEW) {
+ if (target == PIPE_BUFFER) {
+ if (r600_is_vertex_format_supported(format))
+ retval |= PIPE_BIND_SAMPLER_VIEW;
+ } else {
+ if (r600_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)) {
+ PIPE_BIND_SHARED |
+ PIPE_BIND_BLENDABLE)) &&
+ r600_is_colorbuffer_format_supported(rscreen->b.chip_class, format)) {
retval |= usage &
(PIPE_BIND_RENDER_TARGET |
PIPE_BIND_DISPLAY_TARGET |
PIPE_BIND_SCANOUT |
PIPE_BIND_SHARED);
+ if (!util_format_is_pure_integer(format) &&
+ !util_format_is_depth_or_stencil(format))
+ retval |= usage & PIPE_BIND_BLENDABLE;
}
if ((usage & PIPE_BIND_DEPTH_STENCIL) &&
S_028A0C_REPEAT_COUNT(state->line_stipple_factor) : 0;
rs->pa_cl_clip_cntl =
S_028810_PS_UCP_MODE(3) |
+ S_028810_DX_CLIP_SPACE_DEF(state->clip_halfz) |
S_028810_ZCLIP_NEAR_DISABLE(!state->depth_clip) |
S_028810_ZCLIP_FAR_DISABLE(!state->depth_clip) |
- S_028810_DX_LINEAR_ATTR_CLIP_ENA(1);
+ S_028810_DX_LINEAR_ATTR_CLIP_ENA(1) |
+ S_028810_DX_RASTERIZATION_KILL(state->rasterizer_discard);
rs->multisample_enable = state->multisample;
/* offset */
rs->offset_units = state->offset_units;
- rs->offset_scale = state->offset_scale * 12.0f;
+ rs->offset_scale = state->offset_scale * 16.0f;
rs->offset_enable = state->offset_point || state->offset_line || state->offset_tri;
if (state->point_size_per_vertex) {
S_028A48_VPORT_SCISSOR_ENABLE(state->scissor) |
S_028A48_LINE_STIPPLE_ENABLE(state->line_stipple_enable));
- if (rctx->chip_class == CAYMAN) {
+ if (rctx->b.chip_class == CAYMAN) {
r600_store_context_reg(&rs->buffer, CM_R_028BE4_PA_SU_VTX_CNTL,
- S_028C08_PIX_CENTER_HALF(state->gl_rasterization_rules) |
+ S_028C08_PIX_CENTER_HALF(state->half_pixel_center) |
S_028C08_QUANT_MODE(V_028C08_X_1_256TH));
} else {
r600_store_context_reg(&rs->buffer, R_028C08_PA_SU_VTX_CNTL,
- S_028C08_PIX_CENTER_HALF(state->gl_rasterization_rules) |
+ S_028C08_PIX_CENTER_HALF(state->half_pixel_center) |
S_028C08_QUANT_MODE(V_028C08_X_1_256TH));
}
S_028814_CULL_FRONT((state->cull_face & PIPE_FACE_FRONT) ? 1 : 0) |
S_028814_CULL_BACK((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_OFFSET_FRONT_ENABLE(util_get_offset(state, state->fill_front)) |
+ S_028814_POLY_OFFSET_BACK_ENABLE(util_get_offset(state, state->fill_back)) |
+ S_028814_POLY_OFFSET_PARA_ENABLE(state->offset_point || state->offset_line) |
S_028814_POLY_MODE(state->fill_front != PIPE_POLYGON_MODE_FILL ||
state->fill_back != PIPE_POLYGON_MODE_FILL) |
S_028814_POLYMODE_FRONT_PTYPE(r600_translate_fill(state->fill_front)) |
S_028814_POLYMODE_BACK_PTYPE(r600_translate_fill(state->fill_back)));
- r600_store_context_reg(&rs->buffer, R_028350_SX_MISC, S_028350_MULTIPASS(state->rasterizer_discard));
return rs;
}
}
static struct pipe_sampler_view *
-texture_buffer_sampler_view(struct r600_pipe_sampler_view *view,
+texture_buffer_sampler_view(struct r600_context *rctx,
+ struct r600_pipe_sampler_view *view,
unsigned width0, unsigned height0)
{
- struct pipe_context *ctx = view->base.context;
struct r600_texture *tmp = (struct r600_texture*)view->base.texture;
uint64_t va;
int stride = util_format_get_blocksize(view->base.format);
unsigned swizzle_res;
unsigned char swizzle[4];
const struct util_format_description *desc;
+ unsigned offset = view->base.u.buf.first_element * stride;
+ unsigned size = (view->base.u.buf.last_element - view->base.u.buf.first_element + 1) * stride;
swizzle[0] = view->base.swizzle_r;
swizzle[1] = view->base.swizzle_g;
swizzle_res = r600_get_swizzle_combined(desc->swizzle, swizzle, TRUE);
- va = r600_resource_va(ctx->screen, view->base.texture);
+ va = tmp->resource.gpu_address + offset;
view->tex_resource = &tmp->resource;
view->skip_mip_address_reloc = true;
view->tex_resource_words[0] = va;
- view->tex_resource_words[1] = width0 - 1;
+ view->tex_resource_words[1] = size - 1;
view->tex_resource_words[2] = S_030008_BASE_ADDRESS_HI(va >> 32UL) |
S_030008_STRIDE(stride) |
S_030008_DATA_FORMAT(format) |
S_030008_NUM_FORMAT_ALL(num_format) |
S_030008_FORMAT_COMP_ALL(format_comp) |
- S_030008_SRF_MODE_ALL(1) |
S_030008_ENDIAN_SWAP(endian);
view->tex_resource_words[3] = swizzle_res;
/*
view->tex_resource_words[4] = 0;
view->tex_resource_words[5] = view->tex_resource_words[6] = 0;
view->tex_resource_words[7] = S_03001C_TYPE(V_03001C_SQ_TEX_VTX_VALID_BUFFER);
+
+ if (tmp->resource.gpu_address)
+ LIST_ADDTAIL(&view->list, &rctx->b.texture_buffers);
return &view->base;
}
evergreen_create_sampler_view_custom(struct pipe_context *ctx,
struct pipe_resource *texture,
const struct pipe_sampler_view *state,
- unsigned width0, unsigned height0)
+ unsigned width0, unsigned height0,
+ unsigned force_level)
{
+ struct r600_context *rctx = (struct r600_context*)ctx;
struct r600_screen *rscreen = (struct r600_screen*)ctx->screen;
struct r600_pipe_sampler_view *view = CALLOC_STRUCT(r600_pipe_sampler_view);
struct r600_texture *tmp = (struct r600_texture*)texture;
uint32_t word4 = 0, yuv_format = 0, pitch = 0;
unsigned char swizzle[4], array_mode = 0, non_disp_tiling = 0;
unsigned height, depth, width;
- unsigned macro_aspect, tile_split, bankh, bankw, nbanks;
+ unsigned macro_aspect, tile_split, bankh, bankw, nbanks, fmask_bankh;
enum pipe_format pipe_format = state->format;
- struct radeon_surface_level *surflevel;
+ struct radeon_surf_level *surflevel;
+ unsigned base_level, first_level, last_level;
+ uint64_t va;
if (view == NULL)
return NULL;
view->base.context = ctx;
if (texture->target == PIPE_BUFFER)
- return texture_buffer_sampler_view(view, width0, height0);
+ return texture_buffer_sampler_view(rctx, view, width0, height0);
swizzle[0] = state->swizzle_r;
swizzle[1] = state->swizzle_g;
endian = r600_colorformat_endian_swap(format);
+ base_level = 0;
+ first_level = state->u.tex.first_level;
+ last_level = state->u.tex.last_level;
width = width0;
height = height0;
depth = texture->depth0;
- pitch = surflevel[0].nblk_x * util_format_get_blockwidth(pipe_format);
+
+ if (force_level) {
+ base_level = force_level;
+ first_level = 0;
+ last_level = 0;
+ width = u_minify(width, force_level);
+ height = u_minify(height, force_level);
+ depth = u_minify(depth, force_level);
+ }
+
+ pitch = surflevel[base_level].nblk_x * util_format_get_blockwidth(pipe_format);
non_disp_tiling = tmp->non_disp_tiling;
- switch (surflevel[0].mode) {
+ switch (surflevel[base_level].mode) {
case RADEON_SURF_MODE_LINEAR_ALIGNED:
array_mode = V_028C70_ARRAY_LINEAR_ALIGNED;
break;
macro_aspect = eg_macro_tile_aspect(macro_aspect);
bankw = eg_bank_wh(bankw);
bankh = eg_bank_wh(bankh);
+ fmask_bankh = eg_bank_wh(tmp->fmask.bank_height);
/* 128 bit formats require tile type = 1 */
- if (rscreen->chip_class == CAYMAN) {
+ if (rscreen->b.chip_class == CAYMAN) {
if (util_format_get_blocksize(pipe_format) >= 16)
non_disp_tiling = 1;
}
- nbanks = eg_num_banks(rscreen->tiling_info.num_banks);
+ nbanks = eg_num_banks(rscreen->b.tiling_info.num_banks);
if (texture->target == PIPE_TEXTURE_1D_ARRAY) {
height = 1;
} else if (texture->target == PIPE_TEXTURE_CUBE_ARRAY)
depth = texture->array_size / 6;
+ va = tmp->resource.gpu_address;
+
view->tex_resource = &tmp->resource;
view->tex_resource_words[0] = (S_030000_DIM(r600_tex_dim(texture->target, texture->nr_samples)) |
S_030000_PITCH((pitch / 8) - 1) |
S_030000_TEX_WIDTH(width - 1));
- if (rscreen->chip_class == CAYMAN)
+ if (rscreen->b.chip_class == CAYMAN)
view->tex_resource_words[0] |= CM_S_030000_NON_DISP_TILING_ORDER(non_disp_tiling);
else
view->tex_resource_words[0] |= S_030000_NON_DISP_TILING_ORDER(non_disp_tiling);
view->tex_resource_words[1] = (S_030004_TEX_HEIGHT(height - 1) |
S_030004_TEX_DEPTH(depth - 1) |
S_030004_ARRAY_MODE(array_mode));
- view->tex_resource_words[2] = (surflevel[0].offset + r600_resource_va(ctx->screen, texture)) >> 8;
+ view->tex_resource_words[2] = (surflevel[base_level].offset + va) >> 8;
/* TEX_RESOURCE_WORD3.MIP_ADDRESS */
- if (texture->nr_samples > 1 && rscreen->msaa_texture_support == MSAA_TEXTURE_COMPRESSED) {
- /* XXX the 2x and 4x cases are broken. */
- if (tmp->is_depth || tmp->resource.b.b.nr_samples != 8) {
+ if (texture->nr_samples > 1 && rscreen->has_compressed_msaa_texturing) {
+ if (tmp->is_depth) {
/* disable FMASK (0 = disabled) */
view->tex_resource_words[3] = 0;
view->skip_mip_address_reloc = true;
} else {
/* FMASK should be in MIP_ADDRESS for multisample textures */
- view->tex_resource_words[3] = (tmp->fmask_offset + r600_resource_va(ctx->screen, texture)) >> 8;
+ view->tex_resource_words[3] = (tmp->fmask.offset + va) >> 8;
}
- } else if (state->u.tex.last_level && texture->nr_samples <= 1) {
- view->tex_resource_words[3] = (surflevel[1].offset + r600_resource_va(ctx->screen, texture)) >> 8;
+ } else if (last_level && texture->nr_samples <= 1) {
+ view->tex_resource_words[3] = (surflevel[1].offset + va) >> 8;
} else {
- view->tex_resource_words[3] = (surflevel[0].offset + r600_resource_va(ctx->screen, texture)) >> 8;
+ view->tex_resource_words[3] = (surflevel[base_level].offset + va) >> 8;
}
view->tex_resource_words[4] = (word4 |
- S_030010_SRF_MODE_ALL(V_030010_SRF_MODE_ZERO_CLAMP_MINUS_ONE) |
S_030010_ENDIAN_SWAP(endian));
view->tex_resource_words[5] = S_030014_BASE_ARRAY(state->u.tex.first_layer) |
S_030014_LAST_ARRAY(state->u.tex.last_layer);
+ view->tex_resource_words[6] = S_030018_TILE_SPLIT(tile_split);
+
if (texture->nr_samples > 1) {
unsigned log_samples = util_logbase2(texture->nr_samples);
- if (rscreen->chip_class == CAYMAN) {
+ if (rscreen->b.chip_class == CAYMAN) {
view->tex_resource_words[4] |= S_030010_LOG2_NUM_FRAGMENTS(log_samples);
}
/* LAST_LEVEL holds log2(nr_samples) for multisample textures */
view->tex_resource_words[5] |= S_030014_LAST_LEVEL(log_samples);
+ view->tex_resource_words[6] |= S_030018_FMASK_BANK_HEIGHT(fmask_bankh);
} else {
- view->tex_resource_words[4] |= S_030010_BASE_LEVEL(state->u.tex.first_level);
- view->tex_resource_words[5] |= S_030014_LAST_LEVEL(state->u.tex.last_level);
+ view->tex_resource_words[4] |= S_030010_BASE_LEVEL(first_level);
+ view->tex_resource_words[5] |= S_030014_LAST_LEVEL(last_level);
+ /* aniso max 16 samples */
+ view->tex_resource_words[6] |= S_030018_MAX_ANISO(4);
}
- /* aniso max 16 samples */
- view->tex_resource_words[6] = (S_030018_MAX_ANISO(4)) |
- (S_030018_TILE_SPLIT(tile_split));
+
view->tex_resource_words[7] = S_03001C_DATA_FORMAT(format) |
S_03001C_TYPE(V_03001C_SQ_TEX_VTX_VALID_TEXTURE) |
S_03001C_BANK_WIDTH(bankw) |
const struct pipe_sampler_view *state)
{
return evergreen_create_sampler_view_custom(ctx, tex, state,
- tex->width0, tex->height0);
+ tex->width0, tex->height0, 0);
}
static void evergreen_emit_clip_state(struct r600_context *rctx, struct r600_atom *atom)
{
- struct radeon_winsys_cs *cs = rctx->rings.gfx.cs;
+ struct radeon_winsys_cs *cs = rctx->b.rings.gfx.cs;
struct pipe_clip_state *state = &rctx->clip_state.state;
- r600_write_context_reg_seq(cs, R_0285BC_PA_CL_UCP0_X, 6*4);
- r600_write_array(cs, 6*4, (unsigned*)state);
+ radeon_set_context_reg_seq(cs, R_0285BC_PA_CL_UCP0_X, 6*4);
+ radeon_emit_array(cs, (unsigned*)state, 6*4);
}
static void evergreen_set_polygon_stipple(struct pipe_context *ctx,
tl_y = 1;
/* cayman hw workaround */
- if (rctx->chip_class == CAYMAN) {
+ if (rctx->b.chip_class == CAYMAN) {
if (br_x == 1 && br_y == 1)
br_x = 2;
}
*br = S_028244_BR_X(br_x) | S_028244_BR_Y(br_y);
}
-static void evergreen_set_scissor_state(struct pipe_context *ctx,
+static void evergreen_set_scissor_states(struct pipe_context *ctx,
+ unsigned start_slot,
+ unsigned num_scissors,
const struct pipe_scissor_state *state)
{
struct r600_context *rctx = (struct r600_context *)ctx;
+ int i;
- rctx->scissor.scissor = *state;
- rctx->scissor.atom.dirty = true;
+ for (i = start_slot; i < start_slot + num_scissors; i++) {
+ rctx->scissor[i].scissor = state[i - start_slot];
+ r600_mark_atom_dirty(rctx, &rctx->scissor[i].atom);
+ }
}
static void evergreen_emit_scissor_state(struct r600_context *rctx, struct r600_atom *atom)
{
- struct radeon_winsys_cs *cs = rctx->rings.gfx.cs;
- struct pipe_scissor_state *state = &rctx->scissor.scissor;
+ struct radeon_winsys_cs *cs = rctx->b.rings.gfx.cs;
+ struct r600_scissor_state *rstate = (struct r600_scissor_state *)atom;
+ struct pipe_scissor_state *state = &rstate->scissor;
+ unsigned offset = rstate->idx * 4 * 2;
uint32_t tl, br;
evergreen_get_scissor_rect(rctx, state->minx, state->miny, state->maxx, state->maxy, &tl, &br);
- r600_write_context_reg_seq(cs, R_028250_PA_SC_VPORT_SCISSOR_0_TL, 2);
- r600_write_value(cs, tl);
- r600_write_value(cs, br);
+ radeon_set_context_reg_seq(cs, R_028250_PA_SC_VPORT_SCISSOR_0_TL + offset, 2);
+ radeon_emit(cs, tl);
+ radeon_emit(cs, br);
}
/**
* This function intializes the CB* register values for RATs. It is meant
* to be used for 1D aligned buffers that do not have an associated
- * radeon_surface.
+ * radeon_surf.
*/
void evergreen_init_color_surface_rat(struct r600_context *rctx,
struct r600_surface *surf)
{
struct pipe_resource *pipe_buffer = surf->base.texture;
- unsigned format = r600_translate_colorformat(surf->base.format);
+ unsigned format = r600_translate_colorformat(rctx->b.chip_class,
+ surf->base.format);
unsigned endian = r600_colorformat_endian_swap(format);
unsigned swap = r600_translate_colorswap(surf->base.format);
unsigned block_size =
align(util_format_get_blocksize(pipe_buffer->format), 4);
unsigned pitch_alignment =
- MAX2(64, rctx->screen->tiling_info.group_bytes / block_size);
+ MAX2(64, rctx->screen->b.tiling_info.group_bytes / block_size);
unsigned pitch = align(pipe_buffer->width0, pitch_alignment);
/* XXX: This is copied from evergreen_init_color_surface(). I don't
endian = ENDIAN_NONE;
}
- surf->cb_color_base =
- r600_resource_va(rctx->context.screen, pipe_buffer) >> 8;
+ surf->cb_color_base = r600_resource(pipe_buffer)->gpu_address >> 8;
surf->cb_color_pitch = (pitch / 8) - 1;
* elements. */
surf->cb_color_dim = pipe_buffer->width0;
- surf->cb_color_cmask = surf->cb_color_base;
- surf->cb_color_cmask_slice = 0;
+ /* Set the buffer range the GPU will have access to: */
+ util_range_add(&r600_resource(pipe_buffer)->valid_buffer_range,
+ 0, pipe_buffer->width0);
+
surf->cb_color_fmask = surf->cb_color_base;
surf->cb_color_fmask_slice = 0;
}
{
struct r600_screen *rscreen = rctx->screen;
struct r600_texture *rtex = (struct r600_texture*)surf->base.texture;
- struct pipe_resource *pipe_tex = surf->base.texture;
unsigned level = surf->base.u.tex.level;
unsigned pitch, slice;
- unsigned color_info, color_attrib, color_dim = 0;
+ unsigned color_info, color_attrib, color_dim = 0, color_view;
unsigned format, swap, ntype, endian;
uint64_t offset, base_offset;
unsigned non_disp_tiling, macro_aspect, tile_split, bankh, bankw, fmask_bankh, nbanks;
bool blend_clamp = 0, blend_bypass = 0;
offset = rtex->surface.level[level].offset;
- if (rtex->surface.level[level].mode < RADEON_SURF_MODE_1D) {
+ if (rtex->surface.level[level].mode == RADEON_SURF_MODE_LINEAR) {
+ assert(surf->base.u.tex.first_layer == surf->base.u.tex.last_layer);
offset += rtex->surface.level[level].slice_size *
surf->base.u.tex.first_layer;
- }
+ color_view = 0;
+ } else
+ color_view = S_028C6C_SLICE_START(surf->base.u.tex.first_layer) |
+ S_028C6C_SLICE_MAX(surf->base.u.tex.last_layer);
+
pitch = (rtex->surface.level[level].nblk_x) / 8 - 1;
slice = (rtex->surface.level[level].nblk_x * rtex->surface.level[level].nblk_y) / 64;
if (slice) {
macro_aspect = rtex->surface.mtilea;
bankw = rtex->surface.bankw;
bankh = rtex->surface.bankh;
- fmask_bankh = rtex->fmask_bank_height;
+ if (rtex->fmask.size)
+ fmask_bankh = rtex->fmask.bank_height;
+ else
+ fmask_bankh = rtex->surface.bankh;
tile_split = eg_tile_split(tile_split);
macro_aspect = eg_macro_tile_aspect(macro_aspect);
bankw = eg_bank_wh(bankw);
fmask_bankh = eg_bank_wh(fmask_bankh);
/* 128 bit formats require tile type = 1 */
- if (rscreen->chip_class == CAYMAN) {
+ if (rscreen->b.chip_class == CAYMAN) {
if (util_format_get_blocksize(surf->base.format) >= 16)
non_disp_tiling = 1;
}
- nbanks = eg_num_banks(rscreen->tiling_info.num_banks);
+ nbanks = eg_num_banks(rscreen->b.tiling_info.num_banks);
desc = util_format_description(surf->base.format);
for (i = 0; i < 4; i++) {
if (desc->channel[i].type != UTIL_FORMAT_TYPE_VOID) {
S_028C74_NON_DISP_TILING_ORDER(non_disp_tiling) |
S_028C74_FMASK_BANK_HEIGHT(fmask_bankh);
- if (rctx->chip_class == CAYMAN) {
+ if (rctx->b.chip_class == CAYMAN) {
color_attrib |= S_028C74_FORCE_DST_ALPHA_1(desc->swizzle[3] ==
UTIL_FORMAT_SWIZZLE_1);
ntype = V_028C70_NUMBER_UINT;
}
- format = r600_translate_colorformat(surf->base.format);
+ format = r600_translate_colorformat(rctx->b.chip_class, surf->base.format);
assert(format != ~0);
swap = r600_translate_colorswap(surf->base.format);
S_028C70_NUMBER_TYPE(ntype) |
S_028C70_ENDIAN(endian);
- if (rtex->is_rat) {
- color_info |= S_028C70_RAT(1);
- color_dim = S_028C78_WIDTH_MAX(pipe_tex->width0 & 0xffff)
- | S_028C78_HEIGHT_MAX((pipe_tex->width0 >> 16) & 0xffff);
- }
-
/* EXPORT_NORM is an optimzation that can be enabled for better
* performance in certain cases.
* EXPORT_NORM can be enabled if:
surf->export_16bpc = true;
}
- if (rtex->fmask_size && rtex->cmask_size) {
- color_info |= S_028C70_COMPRESSION(1) | S_028C70_FAST_CLEAR(1);
+ if (rtex->fmask.size) {
+ color_info |= S_028C70_COMPRESSION(1);
}
- base_offset = r600_resource_va(rctx->context.screen, pipe_tex);
+ base_offset = rtex->resource.gpu_address;
/* XXX handle enabling of CB beyond BASE8 which has different offset */
surf->cb_color_base = (base_offset + offset) >> 8;
surf->cb_color_info = color_info;
surf->cb_color_pitch = S_028C64_PITCH_TILE_MAX(pitch);
surf->cb_color_slice = S_028C68_SLICE_TILE_MAX(slice);
- if (rtex->surface.level[level].mode < RADEON_SURF_MODE_1D) {
- surf->cb_color_view = 0;
- } else {
- surf->cb_color_view = S_028C6C_SLICE_START(surf->base.u.tex.first_layer) |
- S_028C6C_SLICE_MAX(surf->base.u.tex.last_layer);
- }
+ surf->cb_color_view = color_view;
surf->cb_color_attrib = color_attrib;
- if (rtex->fmask_size && rtex->cmask_size) {
- surf->cb_color_fmask = (base_offset + rtex->fmask_offset) >> 8;
- surf->cb_color_cmask = (base_offset + rtex->cmask_offset) >> 8;
+ if (rtex->fmask.size) {
+ surf->cb_color_fmask = (base_offset + rtex->fmask.offset) >> 8;
+ surf->cb_color_fmask_slice = S_028C88_TILE_MAX(rtex->fmask.slice_tile_max);
} else {
surf->cb_color_fmask = surf->cb_color_base;
- surf->cb_color_cmask = surf->cb_color_base;
+ surf->cb_color_fmask_slice = S_028C88_TILE_MAX(slice);
}
- surf->cb_color_fmask_slice = S_028C88_TILE_MAX(slice);
- surf->cb_color_cmask_slice = S_028C80_TILE_MAX(rtex->cmask_slice_tile_max);
surf->color_initialized = true;
}
struct r600_surface *surf)
{
struct r600_screen *rscreen = rctx->screen;
- struct pipe_screen *screen = &rscreen->screen;
struct r600_texture *rtex = (struct r600_texture*)surf->base.texture;
+ unsigned level = surf->base.u.tex.level;
+ struct radeon_surf_level *levelinfo = &rtex->surface.level[level];
uint64_t offset;
- unsigned level, pitch, slice, format, array_mode;
+ unsigned format, array_mode;
unsigned macro_aspect, tile_split, bankh, bankw, nbanks;
- level = surf->base.u.tex.level;
+
format = r600_translate_dbformat(surf->base.format);
assert(format != ~0);
- offset = r600_resource_va(screen, surf->base.texture);
+ offset = rtex->resource.gpu_address;
offset += rtex->surface.level[level].offset;
- pitch = (rtex->surface.level[level].nblk_x / 8) - 1;
- slice = (rtex->surface.level[level].nblk_x * rtex->surface.level[level].nblk_y) / 64;
- if (slice) {
- slice = slice - 1;
- }
+
switch (rtex->surface.level[level].mode) {
case RADEON_SURF_MODE_2D:
array_mode = V_028C70_ARRAY_2D_TILED_THIN1;
macro_aspect = eg_macro_tile_aspect(macro_aspect);
bankw = eg_bank_wh(bankw);
bankh = eg_bank_wh(bankh);
- nbanks = eg_num_banks(rscreen->tiling_info.num_banks);
+ nbanks = eg_num_banks(rscreen->b.tiling_info.num_banks);
offset >>= 8;
- surf->db_depth_info = S_028040_ARRAY_MODE(array_mode) |
- S_028040_FORMAT(format) |
- S_028040_TILE_SPLIT(tile_split)|
- S_028040_NUM_BANKS(nbanks) |
- S_028040_BANK_WIDTH(bankw) |
- S_028040_BANK_HEIGHT(bankh) |
- S_028040_MACRO_TILE_ASPECT(macro_aspect);
- if (rscreen->chip_class == CAYMAN && rtex->resource.b.b.nr_samples > 1) {
- surf->db_depth_info |= S_028040_NUM_SAMPLES(util_logbase2(rtex->resource.b.b.nr_samples));
+ surf->db_z_info = S_028040_ARRAY_MODE(array_mode) |
+ S_028040_FORMAT(format) |
+ S_028040_TILE_SPLIT(tile_split)|
+ S_028040_NUM_BANKS(nbanks) |
+ S_028040_BANK_WIDTH(bankw) |
+ S_028040_BANK_HEIGHT(bankh) |
+ S_028040_MACRO_TILE_ASPECT(macro_aspect);
+ if (rscreen->b.chip_class == CAYMAN && rtex->resource.b.b.nr_samples > 1) {
+ surf->db_z_info |= S_028040_NUM_SAMPLES(util_logbase2(rtex->resource.b.b.nr_samples));
}
+
+ assert(levelinfo->nblk_x % 8 == 0 && levelinfo->nblk_y % 8 == 0);
+
surf->db_depth_base = offset;
surf->db_depth_view = S_028008_SLICE_START(surf->base.u.tex.first_layer) |
S_028008_SLICE_MAX(surf->base.u.tex.last_layer);
- surf->db_depth_size = S_028058_PITCH_TILE_MAX(pitch);
- surf->db_depth_slice = S_02805C_SLICE_TILE_MAX(slice);
+ surf->db_depth_size = S_028058_PITCH_TILE_MAX(levelinfo->nblk_x / 8 - 1) |
+ S_028058_HEIGHT_TILE_MAX(levelinfo->nblk_y / 8 - 1);
+ surf->db_depth_slice = S_02805C_SLICE_TILE_MAX(levelinfo->nblk_x *
+ levelinfo->nblk_y / 64 - 1);
switch (surf->base.format) {
case PIPE_FORMAT_Z24X8_UNORM:
stile_split = eg_tile_split(stile_split);
stencil_offset = rtex->surface.stencil_level[level].offset;
- stencil_offset += r600_resource_va(screen, surf->base.texture);
+ stencil_offset += rtex->resource.gpu_address;
surf->db_stencil_base = stencil_offset >> 8;
surf->db_stencil_info = S_028044_FORMAT(V_028044_STENCIL_8) |
surf->db_stencil_base = offset;
/* DRM 2.6.18 allows the INVALID format to disable stencil.
* Older kernels are out of luck. */
- surf->db_stencil_info = rctx->screen->info.drm_minor >= 18 ?
+ surf->db_stencil_info = rctx->screen->b.info.drm_minor >= 18 ?
S_028044_FORMAT(V_028044_STENCIL_INVALID) :
S_028044_FORMAT(V_028044_STENCIL_8);
}
- surf->htile_enabled = 0;
/* use htile only for first level */
- if (rtex->htile && !level) {
- uint64_t va = r600_resource_va(&rctx->screen->screen, &rtex->htile->b.b);
- surf->htile_enabled = 1;
+ if (rtex->htile_buffer && !level) {
+ uint64_t va = rtex->htile_buffer->gpu_address;
surf->db_htile_data_base = va >> 8;
surf->db_htile_surface = S_028ABC_HTILE_WIDTH(1) |
- S_028ABC_HTILE_HEIGHT(1) |
- S_028ABC_LINEAR(1);
- surf->db_depth_info |= S_028040_TILE_SURFACE_ENABLE(1);
+ S_028ABC_HTILE_HEIGHT(1) |
+ S_028ABC_FULL_CACHE(1);
+ surf->db_z_info |= S_028040_TILE_SURFACE_ENABLE(1);
surf->db_preload_control = 0;
}
uint32_t i, log_samples;
if (rctx->framebuffer.state.nr_cbufs) {
- rctx->flags |= R600_CONTEXT_WAIT_3D_IDLE | R600_CONTEXT_FLUSH_AND_INV;
-
- if (rctx->framebuffer.state.cbufs[0]->texture->nr_samples > 1) {
- rctx->flags |= R600_CONTEXT_FLUSH_AND_INV_CB_META;
- }
+ rctx->b.flags |= R600_CONTEXT_WAIT_3D_IDLE | R600_CONTEXT_FLUSH_AND_INV;
+ rctx->b.flags |= R600_CONTEXT_FLUSH_AND_INV_CB |
+ R600_CONTEXT_FLUSH_AND_INV_CB_META;
}
if (rctx->framebuffer.state.zsbuf) {
- rctx->flags |= R600_CONTEXT_WAIT_3D_IDLE | R600_CONTEXT_FLUSH_AND_INV;
+ rctx->b.flags |= R600_CONTEXT_WAIT_3D_IDLE | R600_CONTEXT_FLUSH_AND_INV;
+ rctx->b.flags |= R600_CONTEXT_FLUSH_AND_INV_DB;
rtex = (struct r600_texture*)rctx->framebuffer.state.zsbuf->texture;
- if (rtex->htile) {
- rctx->flags |= R600_CONTEXT_FLUSH_AND_INV_DB_META;
+ if (rtex->htile_buffer) {
+ rctx->b.flags |= R600_CONTEXT_FLUSH_AND_INV_DB_META;
}
}
/* Colorbuffers. */
rctx->framebuffer.export_16bpc = state->nr_cbufs != 0;
- rctx->framebuffer.cb0_is_integer = state->nr_cbufs &&
+ rctx->framebuffer.cb0_is_integer = state->nr_cbufs && state->cbufs[0] &&
util_format_is_pure_integer(state->cbufs[0]->format);
rctx->framebuffer.compressed_cb_mask = 0;
-
- if (state->nr_cbufs)
- rctx->framebuffer.nr_samples = state->cbufs[0]->texture->nr_samples;
- else if (state->zsbuf)
- rctx->framebuffer.nr_samples = state->zsbuf->texture->nr_samples;
- else
- rctx->framebuffer.nr_samples = 0;
+ rctx->framebuffer.nr_samples = util_framebuffer_get_num_samples(state);
for (i = 0; i < state->nr_cbufs; i++) {
surf = (struct r600_surface*)state->cbufs[i];
+ if (!surf)
+ continue;
+
rtex = (struct r600_texture*)surf->base.texture;
r600_context_add_resource_size(ctx, state->cbufs[i]->texture);
rctx->framebuffer.export_16bpc = false;
}
- if (rtex->fmask_size && rtex->cmask_size) {
+ if (rtex->fmask.size && rtex->cmask.size) {
rctx->framebuffer.compressed_cb_mask |= 1 << i;
}
}
/* Update alpha-test state dependencies.
* Alpha-test is done on the first colorbuffer only. */
if (state->nr_cbufs) {
+ bool alphatest_bypass = false;
+ bool export_16bpc = true;
+
surf = (struct r600_surface*)state->cbufs[0];
- if (rctx->alphatest_state.bypass != surf->alphatest_bypass) {
- rctx->alphatest_state.bypass = surf->alphatest_bypass;
- rctx->alphatest_state.atom.dirty = true;
+ if (surf) {
+ alphatest_bypass = surf->alphatest_bypass;
+ export_16bpc = surf->export_16bpc;
+ }
+
+ if (rctx->alphatest_state.bypass != alphatest_bypass) {
+ rctx->alphatest_state.bypass = alphatest_bypass;
+ r600_mark_atom_dirty(rctx, &rctx->alphatest_state.atom);
}
- if (rctx->alphatest_state.cb0_export_16bpc != surf->export_16bpc) {
- rctx->alphatest_state.cb0_export_16bpc = surf->export_16bpc;
- rctx->alphatest_state.atom.dirty = true;
+ if (rctx->alphatest_state.cb0_export_16bpc != export_16bpc) {
+ rctx->alphatest_state.cb0_export_16bpc = export_16bpc;
+ r600_mark_atom_dirty(rctx, &rctx->alphatest_state.atom);
}
}
if (state->zsbuf->format != rctx->poly_offset_state.zs_format) {
rctx->poly_offset_state.zs_format = state->zsbuf->format;
- rctx->poly_offset_state.atom.dirty = true;
+ r600_mark_atom_dirty(rctx, &rctx->poly_offset_state.atom);
}
if (rctx->db_state.rsurf != surf) {
rctx->db_state.rsurf = surf;
- rctx->db_state.atom.dirty = true;
- rctx->db_misc_state.atom.dirty = true;
+ r600_mark_atom_dirty(rctx, &rctx->db_state.atom);
+ r600_mark_atom_dirty(rctx, &rctx->db_misc_state.atom);
}
} else if (rctx->db_state.rsurf) {
rctx->db_state.rsurf = NULL;
- rctx->db_state.atom.dirty = true;
- rctx->db_misc_state.atom.dirty = true;
+ r600_mark_atom_dirty(rctx, &rctx->db_state.atom);
+ r600_mark_atom_dirty(rctx, &rctx->db_misc_state.atom);
}
if (rctx->cb_misc_state.nr_cbufs != state->nr_cbufs) {
rctx->cb_misc_state.nr_cbufs = state->nr_cbufs;
- rctx->cb_misc_state.atom.dirty = true;
+ r600_mark_atom_dirty(rctx, &rctx->cb_misc_state.atom);
}
if (state->nr_cbufs == 0 && rctx->alphatest_state.bypass) {
rctx->alphatest_state.bypass = false;
- rctx->alphatest_state.atom.dirty = true;
+ r600_mark_atom_dirty(rctx, &rctx->alphatest_state.atom);
}
log_samples = util_logbase2(rctx->framebuffer.nr_samples);
- if (rctx->chip_class == CAYMAN && rctx->db_misc_state.log_samples != log_samples) {
+ /* This is for Cayman to program SAMPLE_RATE, and for RV770 to fix a hw bug. */
+ if ((rctx->b.chip_class == CAYMAN ||
+ rctx->b.family == CHIP_RV770) &&
+ rctx->db_misc_state.log_samples != log_samples) {
rctx->db_misc_state.log_samples = log_samples;
- rctx->db_misc_state.atom.dirty = true;
+ r600_mark_atom_dirty(rctx, &rctx->db_misc_state.atom);
}
- evergreen_update_db_shader_control(rctx);
/* Calculate the CS size. */
rctx->framebuffer.atom.num_dw = 4; /* SCISSOR */
/* MSAA. */
- if (rctx->chip_class == EVERGREEN) {
- switch (rctx->framebuffer.nr_samples) {
- case 2:
- case 4:
- rctx->framebuffer.atom.num_dw += 6;
- break;
- case 8:
- rctx->framebuffer.atom.num_dw += 10;
- break;
- }
- rctx->framebuffer.atom.num_dw += 4;
- } else {
- switch (rctx->framebuffer.nr_samples) {
- case 2:
- case 4:
- rctx->framebuffer.atom.num_dw += 12;
- break;
- case 8:
- rctx->framebuffer.atom.num_dw += 16;
- break;
- case 16:
- rctx->framebuffer.atom.num_dw += 18;
- break;
- }
- rctx->framebuffer.atom.num_dw += 7;
- }
+ if (rctx->b.chip_class == EVERGREEN)
+ rctx->framebuffer.atom.num_dw += 17; /* Evergreen */
+ else
+ rctx->framebuffer.atom.num_dw += 28; /* Cayman */
/* Colorbuffers. */
- rctx->framebuffer.atom.num_dw += state->nr_cbufs * 21;
+ rctx->framebuffer.atom.num_dw += state->nr_cbufs * 23;
if (rctx->keep_tiling_flags)
rctx->framebuffer.atom.num_dw += state->nr_cbufs * 2;
rctx->framebuffer.atom.num_dw += (12 - state->nr_cbufs) * 3;
rctx->framebuffer.atom.num_dw += 24;
if (rctx->keep_tiling_flags)
rctx->framebuffer.atom.num_dw += 2;
- } else if (rctx->screen->info.drm_minor >= 18) {
+ } else if (rctx->screen->b.info.drm_minor >= 18) {
rctx->framebuffer.atom.num_dw += 4;
}
- rctx->framebuffer.atom.dirty = true;
-}
+ r600_mark_atom_dirty(rctx, &rctx->framebuffer.atom);
-#define FILL_SREG(s0x, s0y, s1x, s1y, s2x, s2y, s3x, s3y) \
- (((s0x) & 0xf) | (((s0y) & 0xf) << 4) | \
- (((s1x) & 0xf) << 8) | (((s1y) & 0xf) << 12) | \
- (((s2x) & 0xf) << 16) | (((s2y) & 0xf) << 20) | \
- (((s3x) & 0xf) << 24) | (((s3y) & 0xf) << 28))
+ r600_set_sample_locations_constant_buffer(rctx);
+}
-static void evergreen_emit_msaa_state(struct r600_context *rctx, int nr_samples)
+static void evergreen_set_min_samples(struct pipe_context *ctx, unsigned min_samples)
{
- /* 2xMSAA
- * There are two locations (-4, 4), (4, -4). */
- static uint32_t sample_locs_2x[] = {
- FILL_SREG(-4, 4, 4, -4, -4, 4, 4, -4),
- FILL_SREG(-4, 4, 4, -4, -4, 4, 4, -4),
- FILL_SREG(-4, 4, 4, -4, -4, 4, 4, -4),
- FILL_SREG(-4, 4, 4, -4, -4, 4, 4, -4),
- };
- static unsigned max_dist_2x = 4;
- /* 4xMSAA
- * There are 4 locations: (-2, -2), (2, 2), (-6, 6), (6, -6). */
- static uint32_t sample_locs_4x[] = {
- FILL_SREG(-2, -2, 2, 2, -6, 6, 6, -6),
- FILL_SREG(-2, -2, 2, 2, -6, 6, 6, -6),
- FILL_SREG(-2, -2, 2, 2, -6, 6, 6, -6),
- FILL_SREG(-2, -2, 2, 2, -6, 6, 6, -6),
- };
- static unsigned max_dist_4x = 6;
- /* 8xMSAA */
- static uint32_t sample_locs_8x[] = {
- FILL_SREG(-1, 1, 1, 5, 3, -5, 5, 3),
- FILL_SREG(-7, -1, -3, -7, 7, -3, -5, 7),
- FILL_SREG(-1, 1, 1, 5, 3, -5, 5, 3),
- FILL_SREG(-7, -1, -3, -7, 7, -3, -5, 7),
- FILL_SREG(-1, 1, 1, 5, 3, -5, 5, 3),
- FILL_SREG(-7, -1, -3, -7, 7, -3, -5, 7),
- FILL_SREG(-1, 1, 1, 5, 3, -5, 5, 3),
- FILL_SREG(-7, -1, -3, -7, 7, -3, -5, 7),
- };
- static unsigned max_dist_8x = 7;
+ struct r600_context *rctx = (struct r600_context *)ctx;
- struct radeon_winsys_cs *cs = rctx->rings.gfx.cs;
- unsigned max_dist = 0;
+ if (rctx->ps_iter_samples == min_samples)
+ return;
- switch (nr_samples) {
+ rctx->ps_iter_samples = min_samples;
+ if (rctx->framebuffer.nr_samples > 1) {
+ r600_mark_atom_dirty(rctx, &rctx->framebuffer.atom);
+ }
+}
+
+/* 8xMSAA */
+static uint32_t sample_locs_8x[] = {
+ FILL_SREG(-1, 1, 1, 5, 3, -5, 5, 3),
+ FILL_SREG(-7, -1, -3, -7, 7, -3, -5, 7),
+ FILL_SREG(-1, 1, 1, 5, 3, -5, 5, 3),
+ FILL_SREG(-7, -1, -3, -7, 7, -3, -5, 7),
+ FILL_SREG(-1, 1, 1, 5, 3, -5, 5, 3),
+ FILL_SREG(-7, -1, -3, -7, 7, -3, -5, 7),
+ FILL_SREG(-1, 1, 1, 5, 3, -5, 5, 3),
+ FILL_SREG(-7, -1, -3, -7, 7, -3, -5, 7),
+};
+static unsigned max_dist_8x = 7;
+
+static void evergreen_get_sample_position(struct pipe_context *ctx,
+ unsigned sample_count,
+ unsigned sample_index,
+ float *out_value)
+{
+ int offset, index;
+ struct {
+ int idx:4;
+ } val;
+ switch (sample_count) {
+ case 1:
default:
- nr_samples = 0;
+ out_value[0] = out_value[1] = 0.5;
break;
case 2:
- r600_write_context_reg_seq(cs, R_028C1C_PA_SC_AA_SAMPLE_LOCS_0, Elements(sample_locs_2x));
- r600_write_array(cs, Elements(sample_locs_2x), sample_locs_2x);
- max_dist = max_dist_2x;
+ offset = 4 * (sample_index * 2);
+ val.idx = (eg_sample_locs_2x[0] >> offset) & 0xf;
+ out_value[0] = (float)(val.idx + 8) / 16.0f;
+ val.idx = (eg_sample_locs_2x[0] >> (offset + 4)) & 0xf;
+ out_value[1] = (float)(val.idx + 8) / 16.0f;
break;
case 4:
- r600_write_context_reg_seq(cs, R_028C1C_PA_SC_AA_SAMPLE_LOCS_0, Elements(sample_locs_4x));
- r600_write_array(cs, Elements(sample_locs_4x), sample_locs_4x);
- max_dist = max_dist_4x;
+ offset = 4 * (sample_index * 2);
+ val.idx = (eg_sample_locs_4x[0] >> offset) & 0xf;
+ out_value[0] = (float)(val.idx + 8) / 16.0f;
+ val.idx = (eg_sample_locs_4x[0] >> (offset + 4)) & 0xf;
+ out_value[1] = (float)(val.idx + 8) / 16.0f;
break;
case 8:
- r600_write_context_reg_seq(cs, R_028C1C_PA_SC_AA_SAMPLE_LOCS_0, Elements(sample_locs_8x));
- r600_write_array(cs, Elements(sample_locs_8x), sample_locs_8x);
- max_dist = max_dist_8x;
+ offset = 4 * (sample_index % 4 * 2);
+ index = (sample_index / 4);
+ val.idx = (sample_locs_8x[index] >> offset) & 0xf;
+ out_value[0] = (float)(val.idx + 8) / 16.0f;
+ val.idx = (sample_locs_8x[index] >> (offset + 4)) & 0xf;
+ out_value[1] = (float)(val.idx + 8) / 16.0f;
break;
}
-
- if (nr_samples > 1) {
- r600_write_context_reg_seq(cs, R_028C00_PA_SC_LINE_CNTL, 2);
- r600_write_value(cs, S_028C00_LAST_PIXEL(1) |
- S_028C00_EXPAND_LINE_WIDTH(1)); /* R_028C00_PA_SC_LINE_CNTL */
- r600_write_value(cs, S_028C04_MSAA_NUM_SAMPLES(util_logbase2(nr_samples)) |
- S_028C04_MAX_SAMPLE_DIST(max_dist)); /* R_028C04_PA_SC_AA_CONFIG */
- } else {
- r600_write_context_reg_seq(cs, R_028C00_PA_SC_LINE_CNTL, 2);
- r600_write_value(cs, S_028C00_LAST_PIXEL(1)); /* R_028C00_PA_SC_LINE_CNTL */
- r600_write_value(cs, 0); /* R_028C04_PA_SC_AA_CONFIG */
- }
}
-static void cayman_emit_msaa_state(struct r600_context *rctx, int nr_samples)
+static void evergreen_emit_msaa_state(struct r600_context *rctx, int nr_samples, int ps_iter_samples)
{
- /* 2xMSAA
- * There are two locations (-4, 4), (4, -4). */
- static uint32_t sample_locs_2x[] = {
- FILL_SREG(-4, 4, 4, -4, -4, 4, 4, -4),
- FILL_SREG(-4, 4, 4, -4, -4, 4, 4, -4),
- FILL_SREG(-4, 4, 4, -4, -4, 4, 4, -4),
- FILL_SREG(-4, 4, 4, -4, -4, 4, 4, -4),
- };
- static unsigned max_dist_2x = 4;
- /* 4xMSAA
- * There are 4 locations: (-2, -2), (2, 2), (-6, 6), (6, -6). */
- static uint32_t sample_locs_4x[] = {
- FILL_SREG(-2, -2, 2, 2, -6, 6, 6, -6),
- FILL_SREG(-2, -2, 2, 2, -6, 6, 6, -6),
- FILL_SREG(-2, -2, 2, 2, -6, 6, 6, -6),
- FILL_SREG(-2, -2, 2, 2, -6, 6, 6, -6),
- };
- static unsigned max_dist_4x = 6;
- /* 8xMSAA */
- static uint32_t sample_locs_8x[] = {
- FILL_SREG(-2, -5, 3, -4, -1, 5, -6, -2),
- FILL_SREG(-2, -5, 3, -4, -1, 5, -6, -2),
- FILL_SREG(-2, -5, 3, -4, -1, 5, -6, -2),
- FILL_SREG(-2, -5, 3, -4, -1, 5, -6, -2),
- FILL_SREG( 6, 0, 0, 0, -5, 3, 4, 4),
- FILL_SREG( 6, 0, 0, 0, -5, 3, 4, 4),
- FILL_SREG( 6, 0, 0, 0, -5, 3, 4, 4),
- FILL_SREG( 6, 0, 0, 0, -5, 3, 4, 4),
- };
- static unsigned max_dist_8x = 8;
- /* 16xMSAA */
- static uint32_t sample_locs_16x[] = {
- FILL_SREG(-7, -3, 7, 3, 1, -5, -5, 5),
- FILL_SREG(-7, -3, 7, 3, 1, -5, -5, 5),
- FILL_SREG(-7, -3, 7, 3, 1, -5, -5, 5),
- FILL_SREG(-7, -3, 7, 3, 1, -5, -5, 5),
- FILL_SREG(-3, -7, 3, 7, 5, -1, -1, 1),
- FILL_SREG(-3, -7, 3, 7, 5, -1, -1, 1),
- FILL_SREG(-3, -7, 3, 7, 5, -1, -1, 1),
- FILL_SREG(-3, -7, 3, 7, 5, -1, -1, 1),
- FILL_SREG(-8, -6, 4, 2, 2, -8, -2, 6),
- FILL_SREG(-8, -6, 4, 2, 2, -8, -2, 6),
- FILL_SREG(-8, -6, 4, 2, 2, -8, -2, 6),
- FILL_SREG(-8, -6, 4, 2, 2, -8, -2, 6),
- FILL_SREG(-4, -2, 0, 4, 6, -4, -6, 0),
- FILL_SREG(-4, -2, 0, 4, 6, -4, -6, 0),
- FILL_SREG(-4, -2, 0, 4, 6, -4, -6, 0),
- FILL_SREG(-4, -2, 0, 4, 6, -4, -6, 0),
- };
- static unsigned max_dist_16x = 8;
- struct radeon_winsys_cs *cs = rctx->rings.gfx.cs;
+ struct radeon_winsys_cs *cs = rctx->b.rings.gfx.cs;
unsigned max_dist = 0;
switch (nr_samples) {
nr_samples = 0;
break;
case 2:
- r600_write_context_reg(cs, CM_R_028BF8_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_0, sample_locs_2x[0]);
- r600_write_context_reg(cs, CM_R_028C08_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_0, sample_locs_2x[1]);
- r600_write_context_reg(cs, CM_R_028C18_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_0, sample_locs_2x[2]);
- r600_write_context_reg(cs, CM_R_028C28_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_0, sample_locs_2x[3]);
- max_dist = max_dist_2x;
+ radeon_set_context_reg_seq(cs, R_028C1C_PA_SC_AA_SAMPLE_LOCS_0, Elements(eg_sample_locs_2x));
+ radeon_emit_array(cs, eg_sample_locs_2x, Elements(eg_sample_locs_2x));
+ max_dist = eg_max_dist_2x;
break;
case 4:
- r600_write_context_reg(cs, CM_R_028BF8_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_0, sample_locs_4x[0]);
- r600_write_context_reg(cs, CM_R_028C08_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_0, sample_locs_4x[1]);
- r600_write_context_reg(cs, CM_R_028C18_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_0, sample_locs_4x[2]);
- r600_write_context_reg(cs, CM_R_028C28_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_0, sample_locs_4x[3]);
- max_dist = max_dist_4x;
+ radeon_set_context_reg_seq(cs, R_028C1C_PA_SC_AA_SAMPLE_LOCS_0, Elements(eg_sample_locs_4x));
+ radeon_emit_array(cs, eg_sample_locs_4x, Elements(eg_sample_locs_4x));
+ max_dist = eg_max_dist_4x;
break;
case 8:
- r600_write_context_reg_seq(cs, CM_R_028BF8_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_0, 14);
- r600_write_value(cs, sample_locs_8x[0]);
- r600_write_value(cs, sample_locs_8x[4]);
- r600_write_value(cs, 0);
- r600_write_value(cs, 0);
- r600_write_value(cs, sample_locs_8x[1]);
- r600_write_value(cs, sample_locs_8x[5]);
- r600_write_value(cs, 0);
- r600_write_value(cs, 0);
- r600_write_value(cs, sample_locs_8x[2]);
- r600_write_value(cs, sample_locs_8x[6]);
- r600_write_value(cs, 0);
- r600_write_value(cs, 0);
- r600_write_value(cs, sample_locs_8x[3]);
- r600_write_value(cs, sample_locs_8x[7]);
+ radeon_set_context_reg_seq(cs, R_028C1C_PA_SC_AA_SAMPLE_LOCS_0, Elements(sample_locs_8x));
+ radeon_emit_array(cs, sample_locs_8x, Elements(sample_locs_8x));
max_dist = max_dist_8x;
break;
- case 16:
- r600_write_context_reg_seq(cs, CM_R_028BF8_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_0, 16);
- r600_write_value(cs, sample_locs_16x[0]);
- r600_write_value(cs, sample_locs_16x[4]);
- r600_write_value(cs, sample_locs_16x[8]);
- r600_write_value(cs, sample_locs_16x[12]);
- r600_write_value(cs, sample_locs_16x[1]);
- r600_write_value(cs, sample_locs_16x[5]);
- r600_write_value(cs, sample_locs_16x[9]);
- r600_write_value(cs, sample_locs_16x[13]);
- r600_write_value(cs, sample_locs_16x[2]);
- r600_write_value(cs, sample_locs_16x[6]);
- r600_write_value(cs, sample_locs_16x[10]);
- r600_write_value(cs, sample_locs_16x[14]);
- r600_write_value(cs, sample_locs_16x[3]);
- r600_write_value(cs, sample_locs_16x[7]);
- r600_write_value(cs, sample_locs_16x[11]);
- r600_write_value(cs, sample_locs_16x[15]);
- max_dist = max_dist_16x;
- break;
}
if (nr_samples > 1) {
- unsigned log_samples = util_logbase2(nr_samples);
-
- r600_write_context_reg_seq(cs, CM_R_028BDC_PA_SC_LINE_CNTL, 2);
- r600_write_value(cs, S_028C00_LAST_PIXEL(1) |
- S_028C00_EXPAND_LINE_WIDTH(1)); /* CM_R_028BDC_PA_SC_LINE_CNTL */
- r600_write_value(cs, S_028BE0_MSAA_NUM_SAMPLES(log_samples) |
- S_028BE0_MAX_SAMPLE_DIST(max_dist) |
- S_028BE0_MSAA_EXPOSED_SAMPLES(log_samples)); /* CM_R_028BE0_PA_SC_AA_CONFIG */
-
- r600_write_context_reg(cs, CM_R_028804_DB_EQAA,
- S_028804_MAX_ANCHOR_SAMPLES(log_samples) |
- S_028804_PS_ITER_SAMPLES(log_samples) |
- S_028804_MASK_EXPORT_NUM_SAMPLES(log_samples) |
- S_028804_ALPHA_TO_MASK_NUM_SAMPLES(log_samples) |
- S_028804_HIGH_QUALITY_INTERSECTIONS(1) |
- S_028804_STATIC_ANCHOR_ASSOCIATIONS(1));
+ radeon_set_context_reg_seq(cs, R_028C00_PA_SC_LINE_CNTL, 2);
+ radeon_emit(cs, S_028C00_LAST_PIXEL(1) |
+ S_028C00_EXPAND_LINE_WIDTH(1)); /* R_028C00_PA_SC_LINE_CNTL */
+ radeon_emit(cs, S_028C04_MSAA_NUM_SAMPLES(util_logbase2(nr_samples)) |
+ S_028C04_MAX_SAMPLE_DIST(max_dist)); /* R_028C04_PA_SC_AA_CONFIG */
+ radeon_set_context_reg(cs, EG_R_028A4C_PA_SC_MODE_CNTL_1, EG_S_028A4C_PS_ITER_SAMPLE(ps_iter_samples > 1));
} else {
- r600_write_context_reg_seq(cs, CM_R_028BDC_PA_SC_LINE_CNTL, 2);
- r600_write_value(cs, S_028C00_LAST_PIXEL(1)); /* CM_R_028BDC_PA_SC_LINE_CNTL */
- r600_write_value(cs, 0); /* CM_R_028BE0_PA_SC_AA_CONFIG */
-
- r600_write_context_reg(cs, CM_R_028804_DB_EQAA,
- S_028804_HIGH_QUALITY_INTERSECTIONS(1) |
- S_028804_STATIC_ANCHOR_ASSOCIATIONS(1));
+ radeon_set_context_reg_seq(cs, R_028C00_PA_SC_LINE_CNTL, 2);
+ radeon_emit(cs, S_028C00_LAST_PIXEL(1)); /* R_028C00_PA_SC_LINE_CNTL */
+ radeon_emit(cs, 0); /* R_028C04_PA_SC_AA_CONFIG */
+ radeon_set_context_reg(cs, EG_R_028A4C_PA_SC_MODE_CNTL_1, 0);
}
}
static void evergreen_emit_framebuffer_state(struct r600_context *rctx, struct r600_atom *atom)
{
- struct radeon_winsys_cs *cs = rctx->rings.gfx.cs;
+ struct radeon_winsys_cs *cs = rctx->b.rings.gfx.cs;
struct pipe_framebuffer_state *state = &rctx->framebuffer.state;
unsigned nr_cbufs = state->nr_cbufs;
unsigned i, tl, br;
+ struct r600_texture *tex = NULL;
+ struct r600_surface *cb = NULL;
/* XXX support more colorbuffers once we need them */
assert(nr_cbufs <= 8);
/* Colorbuffers. */
for (i = 0; i < nr_cbufs; i++) {
- struct r600_surface *cb = (struct r600_surface*)state->cbufs[i];
- unsigned reloc = r600_context_bo_reloc(rctx,
- &rctx->rings.gfx,
- (struct r600_resource*)cb->base.texture,
- RADEON_USAGE_READWRITE);
-
- r600_write_context_reg_seq(cs, R_028C60_CB_COLOR0_BASE + i * 0x3C, 11);
- r600_write_value(cs, cb->cb_color_base); /* R_028C60_CB_COLOR0_BASE */
- r600_write_value(cs, cb->cb_color_pitch); /* R_028C64_CB_COLOR0_PITCH */
- r600_write_value(cs, cb->cb_color_slice); /* R_028C68_CB_COLOR0_SLICE */
- r600_write_value(cs, cb->cb_color_view); /* R_028C6C_CB_COLOR0_VIEW */
- r600_write_value(cs, cb->cb_color_info); /* R_028C70_CB_COLOR0_INFO */
- r600_write_value(cs, cb->cb_color_attrib); /* R_028C74_CB_COLOR0_ATTRIB */
- r600_write_value(cs, cb->cb_color_dim); /* R_028C78_CB_COLOR0_DIM */
- r600_write_value(cs, cb->cb_color_cmask); /* R_028C7C_CB_COLOR0_CMASK */
- r600_write_value(cs, cb->cb_color_cmask_slice); /* R_028C80_CB_COLOR0_CMASK_SLICE */
- r600_write_value(cs, cb->cb_color_fmask); /* R_028C84_CB_COLOR0_FMASK */
- r600_write_value(cs, cb->cb_color_fmask_slice); /* R_028C88_CB_COLOR0_FMASK_SLICE */
-
- r600_write_value(cs, PKT3(PKT3_NOP, 0, 0)); /* R_028C60_CB_COLOR0_BASE */
- r600_write_value(cs, reloc);
+ unsigned reloc, cmask_reloc;
+
+ cb = (struct r600_surface*)state->cbufs[i];
+ if (!cb) {
+ radeon_set_context_reg(cs, R_028C70_CB_COLOR0_INFO + i * 0x3C,
+ S_028C70_FORMAT(V_028C70_COLOR_INVALID));
+ continue;
+ }
+
+ tex = (struct r600_texture *)cb->base.texture;
+ reloc = radeon_add_to_buffer_list(&rctx->b,
+ &rctx->b.rings.gfx,
+ (struct r600_resource*)cb->base.texture,
+ RADEON_USAGE_READWRITE,
+ tex->surface.nsamples > 1 ?
+ RADEON_PRIO_COLOR_BUFFER_MSAA :
+ RADEON_PRIO_COLOR_BUFFER);
+
+ if (tex->cmask_buffer && tex->cmask_buffer != &tex->resource) {
+ cmask_reloc = radeon_add_to_buffer_list(&rctx->b, &rctx->b.rings.gfx,
+ tex->cmask_buffer, RADEON_USAGE_READWRITE,
+ RADEON_PRIO_COLOR_META);
+ } else {
+ cmask_reloc = reloc;
+ }
+
+ radeon_set_context_reg_seq(cs, R_028C60_CB_COLOR0_BASE + i * 0x3C, 13);
+ radeon_emit(cs, cb->cb_color_base); /* R_028C60_CB_COLOR0_BASE */
+ radeon_emit(cs, cb->cb_color_pitch); /* R_028C64_CB_COLOR0_PITCH */
+ radeon_emit(cs, cb->cb_color_slice); /* R_028C68_CB_COLOR0_SLICE */
+ radeon_emit(cs, cb->cb_color_view); /* R_028C6C_CB_COLOR0_VIEW */
+ radeon_emit(cs, cb->cb_color_info | tex->cb_color_info); /* R_028C70_CB_COLOR0_INFO */
+ radeon_emit(cs, cb->cb_color_attrib); /* R_028C74_CB_COLOR0_ATTRIB */
+ radeon_emit(cs, cb->cb_color_dim); /* R_028C78_CB_COLOR0_DIM */
+ radeon_emit(cs, tex->cmask.base_address_reg); /* R_028C7C_CB_COLOR0_CMASK */
+ radeon_emit(cs, tex->cmask.slice_tile_max); /* R_028C80_CB_COLOR0_CMASK_SLICE */
+ radeon_emit(cs, cb->cb_color_fmask); /* R_028C84_CB_COLOR0_FMASK */
+ radeon_emit(cs, cb->cb_color_fmask_slice); /* R_028C88_CB_COLOR0_FMASK_SLICE */
+ radeon_emit(cs, tex->color_clear_value[0]); /* R_028C8C_CB_COLOR0_CLEAR_WORD0 */
+ radeon_emit(cs, tex->color_clear_value[1]); /* R_028C90_CB_COLOR0_CLEAR_WORD1 */
+
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0)); /* R_028C60_CB_COLOR0_BASE */
+ radeon_emit(cs, reloc);
if (!rctx->keep_tiling_flags) {
- r600_write_value(cs, PKT3(PKT3_NOP, 0, 0)); /* R_028C70_CB_COLOR0_INFO */
- r600_write_value(cs, reloc);
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0)); /* R_028C70_CB_COLOR0_INFO */
+ radeon_emit(cs, reloc);
}
- r600_write_value(cs, PKT3(PKT3_NOP, 0, 0)); /* R_028C74_CB_COLOR0_ATTRIB */
- r600_write_value(cs, reloc);
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0)); /* R_028C74_CB_COLOR0_ATTRIB */
+ radeon_emit(cs, reloc);
- r600_write_value(cs, PKT3(PKT3_NOP, 0, 0)); /* R_028C7C_CB_COLOR0_CMASK */
- r600_write_value(cs, reloc);
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0)); /* R_028C7C_CB_COLOR0_CMASK */
+ radeon_emit(cs, cmask_reloc);
- r600_write_value(cs, PKT3(PKT3_NOP, 0, 0)); /* R_028C84_CB_COLOR0_FMASK */
- r600_write_value(cs, reloc);
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0)); /* R_028C84_CB_COLOR0_FMASK */
+ radeon_emit(cs, reloc);
}
/* set CB_COLOR1_INFO for possible dual-src blending */
- if (i == 1 && !((struct r600_texture*)state->cbufs[0]->texture)->is_rat) {
- r600_write_context_reg(cs, R_028C70_CB_COLOR0_INFO + 1 * 0x3C,
- ((struct r600_surface*)state->cbufs[0])->cb_color_info);
+ if (i == 1 && state->cbufs[0]) {
+ radeon_set_context_reg(cs, R_028C70_CB_COLOR0_INFO + 1 * 0x3C,
+ cb->cb_color_info | tex->cb_color_info);
if (!rctx->keep_tiling_flags) {
- unsigned reloc = r600_context_bo_reloc(rctx,
- &rctx->rings.gfx,
+ unsigned reloc = radeon_add_to_buffer_list(&rctx->b,
+ &rctx->b.rings.gfx,
(struct r600_resource*)state->cbufs[0]->texture,
- RADEON_USAGE_READWRITE);
+ RADEON_USAGE_READWRITE,
+ RADEON_PRIO_COLOR_BUFFER);
- r600_write_value(cs, PKT3(PKT3_NOP, 0, 0)); /* R_028C70_CB_COLOR0_INFO */
- r600_write_value(cs, reloc);
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0)); /* R_028C70_CB_COLOR0_INFO */
+ radeon_emit(cs, reloc);
}
i++;
}
if (rctx->keep_tiling_flags) {
for (; i < 8 ; i++) {
- r600_write_context_reg(cs, R_028C70_CB_COLOR0_INFO + i * 0x3C, 0);
+ radeon_set_context_reg(cs, R_028C70_CB_COLOR0_INFO + i * 0x3C, 0);
}
for (; i < 12; i++) {
- r600_write_context_reg(cs, R_028E50_CB_COLOR8_INFO + (i - 8) * 0x1C, 0);
+ radeon_set_context_reg(cs, R_028E50_CB_COLOR8_INFO + (i - 8) * 0x1C, 0);
}
}
/* ZS buffer. */
if (state->zsbuf) {
struct r600_surface *zb = (struct r600_surface*)state->zsbuf;
- unsigned reloc = r600_context_bo_reloc(rctx,
- &rctx->rings.gfx,
+ unsigned reloc = radeon_add_to_buffer_list(&rctx->b,
+ &rctx->b.rings.gfx,
(struct r600_resource*)state->zsbuf->texture,
- RADEON_USAGE_READWRITE);
+ RADEON_USAGE_READWRITE,
+ zb->base.texture->nr_samples > 1 ?
+ RADEON_PRIO_DEPTH_BUFFER_MSAA :
+ RADEON_PRIO_DEPTH_BUFFER);
- r600_write_context_reg(cs, R_028B78_PA_SU_POLY_OFFSET_DB_FMT_CNTL,
+ radeon_set_context_reg(cs, R_028B78_PA_SU_POLY_OFFSET_DB_FMT_CNTL,
zb->pa_su_poly_offset_db_fmt_cntl);
- r600_write_context_reg(cs, R_028008_DB_DEPTH_VIEW, zb->db_depth_view);
-
- r600_write_context_reg_seq(cs, R_028040_DB_Z_INFO, 8);
- r600_write_value(cs, zb->db_depth_info); /* R_028040_DB_Z_INFO */
- r600_write_value(cs, zb->db_stencil_info); /* R_028044_DB_STENCIL_INFO */
- r600_write_value(cs, zb->db_depth_base); /* R_028048_DB_Z_READ_BASE */
- r600_write_value(cs, zb->db_stencil_base); /* R_02804C_DB_STENCIL_READ_BASE */
- r600_write_value(cs, zb->db_depth_base); /* R_028050_DB_Z_WRITE_BASE */
- r600_write_value(cs, zb->db_stencil_base); /* R_028054_DB_STENCIL_WRITE_BASE */
- r600_write_value(cs, zb->db_depth_size); /* R_028058_DB_DEPTH_SIZE */
- r600_write_value(cs, zb->db_depth_slice); /* R_02805C_DB_DEPTH_SLICE */
+ radeon_set_context_reg(cs, R_028008_DB_DEPTH_VIEW, zb->db_depth_view);
+
+ radeon_set_context_reg_seq(cs, R_028040_DB_Z_INFO, 8);
+ radeon_emit(cs, zb->db_z_info); /* R_028040_DB_Z_INFO */
+ radeon_emit(cs, zb->db_stencil_info); /* R_028044_DB_STENCIL_INFO */
+ radeon_emit(cs, zb->db_depth_base); /* R_028048_DB_Z_READ_BASE */
+ radeon_emit(cs, zb->db_stencil_base); /* R_02804C_DB_STENCIL_READ_BASE */
+ radeon_emit(cs, zb->db_depth_base); /* R_028050_DB_Z_WRITE_BASE */
+ radeon_emit(cs, zb->db_stencil_base); /* R_028054_DB_STENCIL_WRITE_BASE */
+ radeon_emit(cs, zb->db_depth_size); /* R_028058_DB_DEPTH_SIZE */
+ radeon_emit(cs, zb->db_depth_slice); /* R_02805C_DB_DEPTH_SLICE */
if (!rctx->keep_tiling_flags) {
- r600_write_value(cs, PKT3(PKT3_NOP, 0, 0)); /* R_028040_DB_Z_INFO */
- r600_write_value(cs, reloc);
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0)); /* R_028040_DB_Z_INFO */
+ radeon_emit(cs, reloc);
}
- r600_write_value(cs, PKT3(PKT3_NOP, 0, 0)); /* R_028048_DB_Z_READ_BASE */
- r600_write_value(cs, reloc);
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0)); /* R_028048_DB_Z_READ_BASE */
+ radeon_emit(cs, reloc);
- r600_write_value(cs, PKT3(PKT3_NOP, 0, 0)); /* R_02804C_DB_STENCIL_READ_BASE */
- r600_write_value(cs, reloc);
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0)); /* R_02804C_DB_STENCIL_READ_BASE */
+ radeon_emit(cs, reloc);
- r600_write_value(cs, PKT3(PKT3_NOP, 0, 0)); /* R_028050_DB_Z_WRITE_BASE */
- r600_write_value(cs, reloc);
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0)); /* R_028050_DB_Z_WRITE_BASE */
+ radeon_emit(cs, reloc);
- r600_write_value(cs, PKT3(PKT3_NOP, 0, 0)); /* R_028054_DB_STENCIL_WRITE_BASE */
- r600_write_value(cs, reloc);
- } else if (rctx->screen->info.drm_minor >= 18) {
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0)); /* R_028054_DB_STENCIL_WRITE_BASE */
+ radeon_emit(cs, reloc);
+ } else if (rctx->screen->b.info.drm_minor >= 18) {
/* DRM 2.6.18 allows the INVALID format to disable depth/stencil.
* Older kernels are out of luck. */
- r600_write_context_reg_seq(cs, R_028040_DB_Z_INFO, 2);
- r600_write_value(cs, S_028040_FORMAT(V_028040_Z_INVALID)); /* R_028040_DB_Z_INFO */
- r600_write_value(cs, S_028044_FORMAT(V_028044_STENCIL_INVALID)); /* R_028044_DB_STENCIL_INFO */
+ radeon_set_context_reg_seq(cs, R_028040_DB_Z_INFO, 2);
+ radeon_emit(cs, S_028040_FORMAT(V_028040_Z_INVALID)); /* R_028040_DB_Z_INFO */
+ radeon_emit(cs, S_028044_FORMAT(V_028044_STENCIL_INVALID)); /* R_028044_DB_STENCIL_INFO */
}
/* Framebuffer dimensions. */
evergreen_get_scissor_rect(rctx, 0, 0, state->width, state->height, &tl, &br);
- r600_write_context_reg_seq(cs, R_028204_PA_SC_WINDOW_SCISSOR_TL, 2);
- r600_write_value(cs, tl); /* R_028204_PA_SC_WINDOW_SCISSOR_TL */
- r600_write_value(cs, br); /* R_028208_PA_SC_WINDOW_SCISSOR_BR */
+ radeon_set_context_reg_seq(cs, R_028204_PA_SC_WINDOW_SCISSOR_TL, 2);
+ radeon_emit(cs, tl); /* R_028204_PA_SC_WINDOW_SCISSOR_TL */
+ radeon_emit(cs, br); /* R_028208_PA_SC_WINDOW_SCISSOR_BR */
- if (rctx->chip_class == EVERGREEN) {
- evergreen_emit_msaa_state(rctx, rctx->framebuffer.nr_samples);
+ if (rctx->b.chip_class == EVERGREEN) {
+ evergreen_emit_msaa_state(rctx, rctx->framebuffer.nr_samples, rctx->ps_iter_samples);
} else {
- cayman_emit_msaa_state(rctx, rctx->framebuffer.nr_samples);
+ cayman_emit_msaa_sample_locs(cs, rctx->framebuffer.nr_samples);
+ cayman_emit_msaa_config(cs, rctx->framebuffer.nr_samples, rctx->ps_iter_samples, 0);
}
}
static void evergreen_emit_polygon_offset(struct r600_context *rctx, struct r600_atom *a)
{
- struct radeon_winsys_cs *cs = rctx->rings.gfx.cs;
+ struct radeon_winsys_cs *cs = rctx->b.rings.gfx.cs;
struct r600_poly_offset_state *state = (struct r600_poly_offset_state*)a;
float offset_units = state->offset_units;
float offset_scale = state->offset_scale;
default:;
}
- r600_write_context_reg_seq(cs, R_028B80_PA_SU_POLY_OFFSET_FRONT_SCALE, 4);
- r600_write_value(cs, fui(offset_scale));
- r600_write_value(cs, fui(offset_units));
- r600_write_value(cs, fui(offset_scale));
- r600_write_value(cs, fui(offset_units));
+ radeon_set_context_reg_seq(cs, R_028B80_PA_SU_POLY_OFFSET_FRONT_SCALE, 4);
+ radeon_emit(cs, fui(offset_scale));
+ radeon_emit(cs, fui(offset_units));
+ radeon_emit(cs, fui(offset_scale));
+ radeon_emit(cs, fui(offset_units));
}
static void evergreen_emit_cb_misc_state(struct r600_context *rctx, struct r600_atom *atom)
{
- struct radeon_winsys_cs *cs = rctx->rings.gfx.cs;
+ struct radeon_winsys_cs *cs = rctx->b.rings.gfx.cs;
struct r600_cb_misc_state *a = (struct r600_cb_misc_state*)atom;
unsigned fb_colormask = (1ULL << ((unsigned)a->nr_cbufs * 4)) - 1;
unsigned ps_colormask = (1ULL << ((unsigned)a->nr_ps_color_outputs * 4)) - 1;
- r600_write_context_reg_seq(cs, R_028238_CB_TARGET_MASK, 2);
- r600_write_value(cs, a->blend_colormask & fb_colormask); /* R_028238_CB_TARGET_MASK */
- /* Always enable the first colorbuffer in CB_SHADER_MASK. This
- * will assure that the alpha-test will work even if there is
- * no colorbuffer bound. */
- r600_write_value(cs, 0xf | (a->dual_src_blend ? ps_colormask : 0) | fb_colormask); /* R_02823C_CB_SHADER_MASK */
+ radeon_set_context_reg_seq(cs, R_028238_CB_TARGET_MASK, 2);
+ radeon_emit(cs, a->blend_colormask & fb_colormask); /* R_028238_CB_TARGET_MASK */
+ /* This must match the used export instructions exactly.
+ * Other values may lead to undefined behavior and hangs.
+ */
+ radeon_emit(cs, ps_colormask); /* R_02823C_CB_SHADER_MASK */
}
static void evergreen_emit_db_state(struct r600_context *rctx, struct r600_atom *atom)
{
- struct radeon_winsys_cs *cs = rctx->rings.gfx.cs;
+ struct radeon_winsys_cs *cs = rctx->b.rings.gfx.cs;
struct r600_db_state *a = (struct r600_db_state*)atom;
- if (a->rsurf && a->rsurf->htile_enabled) {
+ if (a->rsurf && a->rsurf->db_htile_surface) {
struct r600_texture *rtex = (struct r600_texture *)a->rsurf->base.texture;
unsigned reloc_idx;
- r600_write_context_reg(cs, R_02802C_DB_DEPTH_CLEAR, fui(rtex->depth_clear));
- r600_write_context_reg(cs, R_028ABC_DB_HTILE_SURFACE, a->rsurf->db_htile_surface);
- r600_write_context_reg(cs, R_028AC8_DB_PRELOAD_CONTROL, a->rsurf->db_preload_control);
- r600_write_context_reg(cs, R_028014_DB_HTILE_DATA_BASE, a->rsurf->db_htile_data_base);
- reloc_idx = r600_context_bo_reloc(rctx, &rctx->rings.gfx, rtex->htile, RADEON_USAGE_READWRITE);
+ radeon_set_context_reg(cs, R_02802C_DB_DEPTH_CLEAR, fui(rtex->depth_clear_value));
+ radeon_set_context_reg(cs, R_028ABC_DB_HTILE_SURFACE, a->rsurf->db_htile_surface);
+ radeon_set_context_reg(cs, R_028AC8_DB_PRELOAD_CONTROL, a->rsurf->db_preload_control);
+ radeon_set_context_reg(cs, R_028014_DB_HTILE_DATA_BASE, a->rsurf->db_htile_data_base);
+ reloc_idx = radeon_add_to_buffer_list(&rctx->b, &rctx->b.rings.gfx, rtex->htile_buffer,
+ RADEON_USAGE_READWRITE, RADEON_PRIO_DEPTH_META);
cs->buf[cs->cdw++] = PKT3(PKT3_NOP, 0, 0);
cs->buf[cs->cdw++] = reloc_idx;
} else {
- r600_write_context_reg(cs, R_028ABC_DB_HTILE_SURFACE, 0);
- r600_write_context_reg(cs, R_028AC8_DB_PRELOAD_CONTROL, 0);
+ radeon_set_context_reg(cs, R_028ABC_DB_HTILE_SURFACE, 0);
+ radeon_set_context_reg(cs, R_028AC8_DB_PRELOAD_CONTROL, 0);
}
}
static void evergreen_emit_db_misc_state(struct r600_context *rctx, struct r600_atom *atom)
{
- struct radeon_winsys_cs *cs = rctx->rings.gfx.cs;
+ struct radeon_winsys_cs *cs = rctx->b.rings.gfx.cs;
struct r600_db_misc_state *a = (struct r600_db_misc_state*)atom;
unsigned db_render_control = 0;
unsigned db_count_control = 0;
unsigned db_render_override =
S_02800C_FORCE_HIS_ENABLE0(V_02800C_FORCE_DISABLE) |
- S_02800C_FORCE_HIS_ENABLE1(V_02800C_FORCE_DISABLE);
+ S_02800C_FORCE_HIS_ENABLE1(V_02800C_FORCE_DISABLE) |
+ /* There is a hang with HTILE if stencil is used and
+ * fast stencil is enabled. */
+ S_02800C_FAST_STENCIL_DISABLE(1);
if (a->occlusion_query_enabled) {
db_count_control |= S_028004_PERFECT_ZPASS_COUNTS(1);
- if (rctx->chip_class == CAYMAN) {
+ if (rctx->b.chip_class == CAYMAN) {
db_count_control |= S_028004_SAMPLE_RATE(a->log_samples);
}
db_render_override |= S_02800C_NOOP_CULL_DISABLE(1);
*
* Disable hyperz for now if not writing to zbuffer.
*/
- if (rctx->db_state.rsurf && rctx->db_state.rsurf->htile_enabled && rctx->zwritemask) {
+ if (rctx->db_state.rsurf && rctx->db_state.rsurf->db_htile_surface && rctx->zwritemask) {
/* FORCE_OFF means HiZ/HiS are determined by DB_SHADER_CONTROL */
db_render_override |= S_02800C_FORCE_HIZ_ENABLE(V_02800C_FORCE_OFF);
/* This is to fix a lockup when hyperz and alpha test are enabled at
db_render_control |= S_028000_DEPTH_CLEAR_ENABLE(1);
}
- r600_write_context_reg_seq(cs, R_028000_DB_RENDER_CONTROL, 2);
- r600_write_value(cs, db_render_control); /* R_028000_DB_RENDER_CONTROL */
- r600_write_value(cs, db_count_control); /* R_028004_DB_COUNT_CONTROL */
- r600_write_context_reg(cs, R_02800C_DB_RENDER_OVERRIDE, db_render_override);
- r600_write_context_reg(cs, R_02880C_DB_SHADER_CONTROL, a->db_shader_control);
+ radeon_set_context_reg_seq(cs, R_028000_DB_RENDER_CONTROL, 2);
+ radeon_emit(cs, db_render_control); /* R_028000_DB_RENDER_CONTROL */
+ radeon_emit(cs, db_count_control); /* R_028004_DB_COUNT_CONTROL */
+ radeon_set_context_reg(cs, R_02800C_DB_RENDER_OVERRIDE, db_render_override);
+ radeon_set_context_reg(cs, R_02880C_DB_SHADER_CONTROL, a->db_shader_control);
}
static void evergreen_emit_vertex_buffers(struct r600_context *rctx,
unsigned resource_offset,
unsigned pkt_flags)
{
- struct radeon_winsys_cs *cs = rctx->rings.gfx.cs;
+ struct radeon_winsys_cs *cs = rctx->b.rings.gfx.cs;
uint32_t dirty_mask = state->dirty_mask;
while (dirty_mask) {
rbuffer = (struct r600_resource*)vb->buffer;
assert(rbuffer);
- va = r600_resource_va(&rctx->screen->screen, &rbuffer->b.b);
- va += vb->buffer_offset;
+ va = rbuffer->gpu_address + vb->buffer_offset;
/* fetch resources start at index 992 */
- r600_write_value(cs, PKT3(PKT3_SET_RESOURCE, 8, 0) | pkt_flags);
- r600_write_value(cs, (resource_offset + buffer_index) * 8);
- r600_write_value(cs, va); /* RESOURCEi_WORD0 */
- r600_write_value(cs, rbuffer->buf->size - vb->buffer_offset - 1); /* RESOURCEi_WORD1 */
- r600_write_value(cs, /* RESOURCEi_WORD2 */
+ radeon_emit(cs, PKT3(PKT3_SET_RESOURCE, 8, 0) | pkt_flags);
+ radeon_emit(cs, (resource_offset + buffer_index) * 8);
+ radeon_emit(cs, va); /* RESOURCEi_WORD0 */
+ radeon_emit(cs, rbuffer->buf->size - vb->buffer_offset - 1); /* RESOURCEi_WORD1 */
+ radeon_emit(cs, /* RESOURCEi_WORD2 */
S_030008_ENDIAN_SWAP(r600_endian_swap(32)) |
S_030008_STRIDE(vb->stride) |
S_030008_BASE_ADDRESS_HI(va >> 32UL));
- r600_write_value(cs, /* RESOURCEi_WORD3 */
+ radeon_emit(cs, /* RESOURCEi_WORD3 */
S_03000C_DST_SEL_X(V_03000C_SQ_SEL_X) |
S_03000C_DST_SEL_Y(V_03000C_SQ_SEL_Y) |
S_03000C_DST_SEL_Z(V_03000C_SQ_SEL_Z) |
S_03000C_DST_SEL_W(V_03000C_SQ_SEL_W));
- r600_write_value(cs, 0); /* RESOURCEi_WORD4 */
- r600_write_value(cs, 0); /* RESOURCEi_WORD5 */
- r600_write_value(cs, 0); /* RESOURCEi_WORD6 */
- r600_write_value(cs, 0xc0000000); /* RESOURCEi_WORD7 */
+ radeon_emit(cs, 0); /* RESOURCEi_WORD4 */
+ radeon_emit(cs, 0); /* RESOURCEi_WORD5 */
+ radeon_emit(cs, 0); /* RESOURCEi_WORD6 */
+ radeon_emit(cs, 0xc0000000); /* RESOURCEi_WORD7 */
- r600_write_value(cs, PKT3(PKT3_NOP, 0, 0) | pkt_flags);
- r600_write_value(cs, r600_context_bo_reloc(rctx, &rctx->rings.gfx, rbuffer, RADEON_USAGE_READ));
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0) | pkt_flags);
+ radeon_emit(cs, radeon_add_to_buffer_list(&rctx->b, &rctx->b.rings.gfx, rbuffer,
+ RADEON_USAGE_READ, RADEON_PRIO_SHADER_BUFFER_RO));
}
state->dirty_mask = 0;
}
struct r600_constbuf_state *state,
unsigned buffer_id_base,
unsigned reg_alu_constbuf_size,
- unsigned reg_alu_const_cache)
+ unsigned reg_alu_const_cache,
+ unsigned pkt_flags)
{
- struct radeon_winsys_cs *cs = rctx->rings.gfx.cs;
+ struct radeon_winsys_cs *cs = rctx->b.rings.gfx.cs;
uint32_t dirty_mask = state->dirty_mask;
while (dirty_mask) {
struct r600_resource *rbuffer;
uint64_t va;
unsigned buffer_index = ffs(dirty_mask) - 1;
+ unsigned gs_ring_buffer = (buffer_index == R600_GS_RING_CONST_BUFFER);
cb = &state->cb[buffer_index];
rbuffer = (struct r600_resource*)cb->buffer;
assert(rbuffer);
- va = r600_resource_va(&rctx->screen->screen, &rbuffer->b.b);
- va += cb->buffer_offset;
+ va = rbuffer->gpu_address + cb->buffer_offset;
- r600_write_context_reg(cs, reg_alu_constbuf_size + buffer_index * 4,
- ALIGN_DIVUP(cb->buffer_size >> 4, 16));
- r600_write_context_reg(cs, reg_alu_const_cache + buffer_index * 4, va >> 8);
-
- r600_write_value(cs, PKT3(PKT3_NOP, 0, 0));
- r600_write_value(cs, r600_context_bo_reloc(rctx, &rctx->rings.gfx, rbuffer, RADEON_USAGE_READ));
+ if (!gs_ring_buffer) {
+ radeon_set_context_reg_flag(cs, reg_alu_constbuf_size + buffer_index * 4,
+ ALIGN_DIVUP(cb->buffer_size >> 4, 16), pkt_flags);
+ radeon_set_context_reg_flag(cs, reg_alu_const_cache + buffer_index * 4, va >> 8,
+ pkt_flags);
+ }
- r600_write_value(cs, PKT3(PKT3_SET_RESOURCE, 8, 0));
- r600_write_value(cs, (buffer_id_base + buffer_index) * 8);
- r600_write_value(cs, va); /* RESOURCEi_WORD0 */
- r600_write_value(cs, rbuffer->buf->size - cb->buffer_offset - 1); /* RESOURCEi_WORD1 */
- r600_write_value(cs, /* RESOURCEi_WORD2 */
- S_030008_ENDIAN_SWAP(r600_endian_swap(32)) |
- S_030008_STRIDE(16) |
- S_030008_BASE_ADDRESS_HI(va >> 32UL));
- r600_write_value(cs, /* RESOURCEi_WORD3 */
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0) | pkt_flags);
+ radeon_emit(cs, radeon_add_to_buffer_list(&rctx->b, &rctx->b.rings.gfx, rbuffer,
+ RADEON_USAGE_READ, RADEON_PRIO_SHADER_BUFFER_RO));
+
+ radeon_emit(cs, PKT3(PKT3_SET_RESOURCE, 8, 0) | pkt_flags);
+ radeon_emit(cs, (buffer_id_base + buffer_index) * 8);
+ radeon_emit(cs, va); /* RESOURCEi_WORD0 */
+ radeon_emit(cs, rbuffer->buf->size - cb->buffer_offset - 1); /* RESOURCEi_WORD1 */
+ radeon_emit(cs, /* RESOURCEi_WORD2 */
+ S_030008_ENDIAN_SWAP(gs_ring_buffer ? ENDIAN_NONE : r600_endian_swap(32)) |
+ S_030008_STRIDE(gs_ring_buffer ? 4 : 16) |
+ S_030008_BASE_ADDRESS_HI(va >> 32UL) |
+ S_030008_DATA_FORMAT(FMT_32_32_32_32_FLOAT));
+ radeon_emit(cs, /* RESOURCEi_WORD3 */
+ S_03000C_UNCACHED(gs_ring_buffer ? 1 : 0) |
S_03000C_DST_SEL_X(V_03000C_SQ_SEL_X) |
S_03000C_DST_SEL_Y(V_03000C_SQ_SEL_Y) |
S_03000C_DST_SEL_Z(V_03000C_SQ_SEL_Z) |
S_03000C_DST_SEL_W(V_03000C_SQ_SEL_W));
- r600_write_value(cs, 0); /* RESOURCEi_WORD4 */
- r600_write_value(cs, 0); /* RESOURCEi_WORD5 */
- r600_write_value(cs, 0); /* RESOURCEi_WORD6 */
- r600_write_value(cs, 0xc0000000); /* RESOURCEi_WORD7 */
+ radeon_emit(cs, 0); /* RESOURCEi_WORD4 */
+ radeon_emit(cs, 0); /* RESOURCEi_WORD5 */
+ radeon_emit(cs, 0); /* RESOURCEi_WORD6 */
+ radeon_emit(cs, /* RESOURCEi_WORD7 */
+ S_03001C_TYPE(V_03001C_SQ_TEX_VTX_VALID_BUFFER));
- r600_write_value(cs, PKT3(PKT3_NOP, 0, 0));
- r600_write_value(cs, r600_context_bo_reloc(rctx, &rctx->rings.gfx, rbuffer, RADEON_USAGE_READ));
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0) | pkt_flags);
+ radeon_emit(cs, radeon_add_to_buffer_list(&rctx->b, &rctx->b.rings.gfx, rbuffer,
+ RADEON_USAGE_READ, RADEON_PRIO_SHADER_BUFFER_RO));
dirty_mask &= ~(1 << buffer_index);
}
{
evergreen_emit_constant_buffers(rctx, &rctx->constbuf_state[PIPE_SHADER_VERTEX], 176,
R_028180_ALU_CONST_BUFFER_SIZE_VS_0,
- R_028980_ALU_CONST_CACHE_VS_0);
+ R_028980_ALU_CONST_CACHE_VS_0,
+ 0 /* PKT3 flags */);
}
static void evergreen_emit_gs_constant_buffers(struct r600_context *rctx, struct r600_atom *atom)
{
evergreen_emit_constant_buffers(rctx, &rctx->constbuf_state[PIPE_SHADER_GEOMETRY], 336,
R_0281C0_ALU_CONST_BUFFER_SIZE_GS_0,
- R_0289C0_ALU_CONST_CACHE_GS_0);
+ R_0289C0_ALU_CONST_CACHE_GS_0,
+ 0 /* PKT3 flags */);
}
static void evergreen_emit_ps_constant_buffers(struct r600_context *rctx, struct r600_atom *atom)
{
evergreen_emit_constant_buffers(rctx, &rctx->constbuf_state[PIPE_SHADER_FRAGMENT], 0,
R_028140_ALU_CONST_BUFFER_SIZE_PS_0,
- R_028940_ALU_CONST_CACHE_PS_0);
+ R_028940_ALU_CONST_CACHE_PS_0,
+ 0 /* PKT3 flags */);
+}
+
+static void evergreen_emit_cs_constant_buffers(struct r600_context *rctx, struct r600_atom *atom)
+{
+ evergreen_emit_constant_buffers(rctx, &rctx->constbuf_state[PIPE_SHADER_COMPUTE], 816,
+ R_028FC0_ALU_CONST_BUFFER_SIZE_LS_0,
+ R_028F40_ALU_CONST_CACHE_LS_0,
+ RADEON_CP_PACKET3_COMPUTE_MODE);
}
static void evergreen_emit_sampler_views(struct r600_context *rctx,
struct r600_samplerview_state *state,
- unsigned resource_id_base)
+ unsigned resource_id_base, unsigned pkt_flags)
{
- struct radeon_winsys_cs *cs = rctx->rings.gfx.cs;
+ struct radeon_winsys_cs *cs = rctx->b.rings.gfx.cs;
uint32_t dirty_mask = state->dirty_mask;
while (dirty_mask) {
rview = state->views[resource_index];
assert(rview);
- r600_write_value(cs, PKT3(PKT3_SET_RESOURCE, 8, 0));
- r600_write_value(cs, (resource_id_base + resource_index) * 8);
- r600_write_array(cs, 8, rview->tex_resource_words);
+ radeon_emit(cs, PKT3(PKT3_SET_RESOURCE, 8, 0) | pkt_flags);
+ radeon_emit(cs, (resource_id_base + resource_index) * 8);
+ radeon_emit_array(cs, rview->tex_resource_words, 8);
- reloc = r600_context_bo_reloc(rctx, &rctx->rings.gfx, rview->tex_resource,
- RADEON_USAGE_READ);
- r600_write_value(cs, PKT3(PKT3_NOP, 0, 0));
- r600_write_value(cs, reloc);
+ reloc = radeon_add_to_buffer_list(&rctx->b, &rctx->b.rings.gfx, rview->tex_resource,
+ RADEON_USAGE_READ,
+ rview->tex_resource->b.b.nr_samples > 1 ?
+ RADEON_PRIO_SHADER_TEXTURE_MSAA :
+ RADEON_PRIO_SHADER_TEXTURE_RO);
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0) | pkt_flags);
+ radeon_emit(cs, reloc);
if (!rview->skip_mip_address_reloc) {
- r600_write_value(cs, PKT3(PKT3_NOP, 0, 0));
- r600_write_value(cs, reloc);
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0) | pkt_flags);
+ radeon_emit(cs, reloc);
}
}
state->dirty_mask = 0;
static void evergreen_emit_vs_sampler_views(struct r600_context *rctx, struct r600_atom *atom)
{
- evergreen_emit_sampler_views(rctx, &rctx->samplers[PIPE_SHADER_VERTEX].views, 176 + R600_MAX_CONST_BUFFERS);
+ evergreen_emit_sampler_views(rctx, &rctx->samplers[PIPE_SHADER_VERTEX].views,
+ 176 + R600_MAX_CONST_BUFFERS, 0);
}
static void evergreen_emit_gs_sampler_views(struct r600_context *rctx, struct r600_atom *atom)
{
- evergreen_emit_sampler_views(rctx, &rctx->samplers[PIPE_SHADER_GEOMETRY].views, 336 + R600_MAX_CONST_BUFFERS);
+ evergreen_emit_sampler_views(rctx, &rctx->samplers[PIPE_SHADER_GEOMETRY].views,
+ 336 + R600_MAX_CONST_BUFFERS, 0);
}
static void evergreen_emit_ps_sampler_views(struct r600_context *rctx, struct r600_atom *atom)
{
- evergreen_emit_sampler_views(rctx, &rctx->samplers[PIPE_SHADER_FRAGMENT].views, R600_MAX_CONST_BUFFERS);
+ evergreen_emit_sampler_views(rctx, &rctx->samplers[PIPE_SHADER_FRAGMENT].views,
+ R600_MAX_CONST_BUFFERS, 0);
+}
+
+static void evergreen_emit_cs_sampler_views(struct r600_context *rctx, struct r600_atom *atom)
+{
+ evergreen_emit_sampler_views(rctx, &rctx->samplers[PIPE_SHADER_COMPUTE].views,
+ 816 + 2, RADEON_CP_PACKET3_COMPUTE_MODE);
}
static void evergreen_emit_sampler_states(struct r600_context *rctx,
struct r600_textures_info *texinfo,
unsigned resource_id_base,
- unsigned border_index_reg)
+ unsigned border_index_reg,
+ unsigned pkt_flags)
{
- struct radeon_winsys_cs *cs = rctx->rings.gfx.cs;
+ struct radeon_winsys_cs *cs = rctx->b.rings.gfx.cs;
uint32_t dirty_mask = texinfo->states.dirty_mask;
while (dirty_mask) {
rstate = texinfo->states.states[i];
assert(rstate);
- r600_write_value(cs, PKT3(PKT3_SET_SAMPLER, 3, 0));
- r600_write_value(cs, (resource_id_base + i) * 3);
- r600_write_array(cs, 3, rstate->tex_sampler_words);
+ radeon_emit(cs, PKT3(PKT3_SET_SAMPLER, 3, 0) | pkt_flags);
+ radeon_emit(cs, (resource_id_base + i) * 3);
+ radeon_emit_array(cs, rstate->tex_sampler_words, 3);
if (rstate->border_color_use) {
- r600_write_config_reg_seq(cs, border_index_reg, 5);
- r600_write_value(cs, i);
- r600_write_array(cs, 4, rstate->border_color.ui);
+ radeon_set_config_reg_seq(cs, border_index_reg, 5);
+ radeon_emit(cs, i);
+ radeon_emit_array(cs, rstate->border_color.ui, 4);
}
}
texinfo->states.dirty_mask = 0;
static void evergreen_emit_vs_sampler_states(struct r600_context *rctx, struct r600_atom *atom)
{
- evergreen_emit_sampler_states(rctx, &rctx->samplers[PIPE_SHADER_VERTEX], 18, R_00A414_TD_VS_SAMPLER0_BORDER_INDEX);
+ evergreen_emit_sampler_states(rctx, &rctx->samplers[PIPE_SHADER_VERTEX], 18,
+ R_00A414_TD_VS_SAMPLER0_BORDER_INDEX, 0);
}
static void evergreen_emit_gs_sampler_states(struct r600_context *rctx, struct r600_atom *atom)
{
- evergreen_emit_sampler_states(rctx, &rctx->samplers[PIPE_SHADER_GEOMETRY], 36, R_00A428_TD_GS_SAMPLER0_BORDER_INDEX);
+ evergreen_emit_sampler_states(rctx, &rctx->samplers[PIPE_SHADER_GEOMETRY], 36,
+ R_00A428_TD_GS_SAMPLER0_BORDER_INDEX, 0);
}
static void evergreen_emit_ps_sampler_states(struct r600_context *rctx, struct r600_atom *atom)
{
- evergreen_emit_sampler_states(rctx, &rctx->samplers[PIPE_SHADER_FRAGMENT], 0, R_00A400_TD_PS_SAMPLER0_BORDER_INDEX);
+ evergreen_emit_sampler_states(rctx, &rctx->samplers[PIPE_SHADER_FRAGMENT], 0,
+ R_00A400_TD_PS_SAMPLER0_BORDER_INDEX, 0);
+}
+
+static void evergreen_emit_cs_sampler_states(struct r600_context *rctx, struct r600_atom *atom)
+{
+ evergreen_emit_sampler_states(rctx, &rctx->samplers[PIPE_SHADER_COMPUTE], 90,
+ R_00A464_TD_CS_SAMPLER0_BORDER_INDEX,
+ RADEON_CP_PACKET3_COMPUTE_MODE);
}
static void evergreen_emit_sample_mask(struct r600_context *rctx, struct r600_atom *a)
struct r600_sample_mask *s = (struct r600_sample_mask*)a;
uint8_t mask = s->sample_mask;
- r600_write_context_reg(rctx->rings.gfx.cs, R_028C3C_PA_SC_AA_MASK,
+ radeon_set_context_reg(rctx->b.rings.gfx.cs, R_028C3C_PA_SC_AA_MASK,
mask | (mask << 8) | (mask << 16) | (mask << 24));
}
static void cayman_emit_sample_mask(struct r600_context *rctx, struct r600_atom *a)
{
struct r600_sample_mask *s = (struct r600_sample_mask*)a;
- struct radeon_winsys_cs *cs = rctx->rings.gfx.cs;
+ struct radeon_winsys_cs *cs = rctx->b.rings.gfx.cs;
uint16_t mask = s->sample_mask;
- r600_write_context_reg_seq(cs, CM_R_028C38_PA_SC_AA_MASK_X0Y0_X1Y0, 2);
- r600_write_value(cs, mask | (mask << 16)); /* X0Y0_X1Y0 */
- r600_write_value(cs, mask | (mask << 16)); /* X0Y1_X1Y1 */
+ radeon_set_context_reg_seq(cs, CM_R_028C38_PA_SC_AA_MASK_X0Y0_X1Y0, 2);
+ radeon_emit(cs, mask | (mask << 16)); /* X0Y0_X1Y0 */
+ radeon_emit(cs, mask | (mask << 16)); /* X0Y1_X1Y1 */
}
static void evergreen_emit_vertex_fetch_shader(struct r600_context *rctx, struct r600_atom *a)
{
- struct radeon_winsys_cs *cs = rctx->rings.gfx.cs;
+ struct radeon_winsys_cs *cs = rctx->b.rings.gfx.cs;
struct r600_cso_state *state = (struct r600_cso_state*)a;
struct r600_fetch_shader *shader = (struct r600_fetch_shader*)state->cso;
- r600_write_context_reg(cs, R_0288A4_SQ_PGM_START_FS,
- (r600_resource_va(rctx->context.screen, &shader->buffer->b.b) + shader->offset) >> 8);
- r600_write_value(cs, PKT3(PKT3_NOP, 0, 0));
- r600_write_value(cs, r600_context_bo_reloc(rctx, &rctx->rings.gfx, shader->buffer, RADEON_USAGE_READ));
+ radeon_set_context_reg(cs, R_0288A4_SQ_PGM_START_FS,
+ (shader->buffer->gpu_address + shader->offset) >> 8);
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0));
+ radeon_emit(cs, radeon_add_to_buffer_list(&rctx->b, &rctx->b.rings.gfx, shader->buffer,
+ RADEON_USAGE_READ, RADEON_PRIO_SHADER_DATA));
}
-void evergreen_init_state_functions(struct r600_context *rctx)
+static void evergreen_emit_shader_stages(struct r600_context *rctx, struct r600_atom *a)
{
- unsigned id = 4;
+ struct radeon_winsys_cs *cs = rctx->b.rings.gfx.cs;
+ struct r600_shader_stages_state *state = (struct r600_shader_stages_state*)a;
- /* !!!
- * To avoid GPU lockup registers must be emited in a specific order
- * (no kidding ...). The order below is important and have been
- * partialy infered from analyzing fglrx command stream.
- *
- * Don't reorder atom without carefully checking the effect (GPU lockup
- * or piglit regression).
- * !!!
- */
+ uint32_t v = 0, v2 = 0, primid = 0;
- r600_init_atom(rctx, &rctx->framebuffer.atom, id++, evergreen_emit_framebuffer_state, 0);
- /* shader const */
- r600_init_atom(rctx, &rctx->constbuf_state[PIPE_SHADER_VERTEX].atom, id++, evergreen_emit_vs_constant_buffers, 0);
- r600_init_atom(rctx, &rctx->constbuf_state[PIPE_SHADER_GEOMETRY].atom, id++, evergreen_emit_gs_constant_buffers, 0);
- r600_init_atom(rctx, &rctx->constbuf_state[PIPE_SHADER_FRAGMENT].atom, id++, evergreen_emit_ps_constant_buffers, 0);
- /* shader program */
- r600_init_atom(rctx, &rctx->cs_shader_state.atom, id++, evergreen_emit_cs_shader, 0);
- /* sampler */
- r600_init_atom(rctx, &rctx->samplers[PIPE_SHADER_VERTEX].states.atom, id++, evergreen_emit_vs_sampler_states, 0);
- r600_init_atom(rctx, &rctx->samplers[PIPE_SHADER_GEOMETRY].states.atom, id++, evergreen_emit_gs_sampler_states, 0);
- r600_init_atom(rctx, &rctx->samplers[PIPE_SHADER_FRAGMENT].states.atom, id++, evergreen_emit_ps_sampler_states, 0);
- /* resources */
- r600_init_atom(rctx, &rctx->vertex_buffer_state.atom, id++, evergreen_fs_emit_vertex_buffers, 0);
- r600_init_atom(rctx, &rctx->cs_vertex_buffer_state.atom, id++, evergreen_cs_emit_vertex_buffers, 0);
- r600_init_atom(rctx, &rctx->samplers[PIPE_SHADER_VERTEX].views.atom, id++, evergreen_emit_vs_sampler_views, 0);
- r600_init_atom(rctx, &rctx->samplers[PIPE_SHADER_GEOMETRY].views.atom, id++, evergreen_emit_gs_sampler_views, 0);
- r600_init_atom(rctx, &rctx->samplers[PIPE_SHADER_FRAGMENT].views.atom, id++, evergreen_emit_ps_sampler_views, 0);
+ if (rctx->vs_shader->current->shader.vs_as_gs_a) {
+ v2 = S_028A40_MODE(V_028A40_GS_SCENARIO_A);
+ primid = 1;
+ }
- r600_init_atom(rctx, &rctx->vgt_state.atom, id++, r600_emit_vgt_state, 7);
+ if (state->geom_enable) {
+ uint32_t cut_val;
- if (rctx->chip_class == EVERGREEN) {
- r600_init_atom(rctx, &rctx->sample_mask.atom, id++, evergreen_emit_sample_mask, 3);
+ if (rctx->gs_shader->gs_max_out_vertices <= 128)
+ cut_val = V_028A40_GS_CUT_128;
+ else if (rctx->gs_shader->gs_max_out_vertices <= 256)
+ cut_val = V_028A40_GS_CUT_256;
+ else if (rctx->gs_shader->gs_max_out_vertices <= 512)
+ cut_val = V_028A40_GS_CUT_512;
+ else
+ cut_val = V_028A40_GS_CUT_1024;
+ v = S_028B54_ES_EN(V_028B54_ES_STAGE_REAL) |
+ S_028B54_GS_EN(1) |
+ S_028B54_VS_EN(V_028B54_VS_STAGE_COPY_SHADER);
+
+ v2 = S_028A40_MODE(V_028A40_GS_SCENARIO_G) |
+ S_028A40_CUT_MODE(cut_val);
+
+ if (rctx->gs_shader->current->shader.gs_prim_id_input)
+ primid = 1;
+ }
+
+ radeon_set_context_reg(cs, R_028B54_VGT_SHADER_STAGES_EN, v);
+ radeon_set_context_reg(cs, R_028A40_VGT_GS_MODE, v2);
+ radeon_set_context_reg(cs, R_028A84_VGT_PRIMITIVEID_EN, primid);
+}
+
+static void evergreen_emit_gs_rings(struct r600_context *rctx, struct r600_atom *a)
+{
+ struct radeon_winsys_cs *cs = rctx->b.rings.gfx.cs;
+ struct r600_gs_rings_state *state = (struct r600_gs_rings_state*)a;
+ struct r600_resource *rbuffer;
+
+ radeon_set_config_reg(cs, R_008040_WAIT_UNTIL, S_008040_WAIT_3D_IDLE(1));
+ radeon_emit(cs, PKT3(PKT3_EVENT_WRITE, 0, 0));
+ radeon_emit(cs, EVENT_TYPE(EVENT_TYPE_VGT_FLUSH));
+
+ if (state->enable) {
+ rbuffer =(struct r600_resource*)state->esgs_ring.buffer;
+ radeon_set_config_reg(cs, R_008C40_SQ_ESGS_RING_BASE,
+ rbuffer->gpu_address >> 8);
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0));
+ radeon_emit(cs, radeon_add_to_buffer_list(&rctx->b, &rctx->b.rings.gfx, rbuffer,
+ RADEON_USAGE_READWRITE,
+ RADEON_PRIO_SHADER_RESOURCE_RW));
+ radeon_set_config_reg(cs, R_008C44_SQ_ESGS_RING_SIZE,
+ state->esgs_ring.buffer_size >> 8);
+
+ rbuffer =(struct r600_resource*)state->gsvs_ring.buffer;
+ radeon_set_config_reg(cs, R_008C48_SQ_GSVS_RING_BASE,
+ rbuffer->gpu_address >> 8);
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0));
+ radeon_emit(cs, radeon_add_to_buffer_list(&rctx->b, &rctx->b.rings.gfx, rbuffer,
+ RADEON_USAGE_READWRITE,
+ RADEON_PRIO_SHADER_RESOURCE_RW));
+ radeon_set_config_reg(cs, R_008C4C_SQ_GSVS_RING_SIZE,
+ state->gsvs_ring.buffer_size >> 8);
} else {
- r600_init_atom(rctx, &rctx->sample_mask.atom, id++, cayman_emit_sample_mask, 4);
+ radeon_set_config_reg(cs, R_008C44_SQ_ESGS_RING_SIZE, 0);
+ radeon_set_config_reg(cs, R_008C4C_SQ_GSVS_RING_SIZE, 0);
}
- rctx->sample_mask.sample_mask = ~0;
- r600_init_atom(rctx, &rctx->alphatest_state.atom, id++, r600_emit_alphatest_state, 6);
- r600_init_atom(rctx, &rctx->blend_color.atom, id++, r600_emit_blend_color, 6);
- r600_init_atom(rctx, &rctx->blend_state.atom, id++, r600_emit_cso_state, 0);
- r600_init_atom(rctx, &rctx->cb_misc_state.atom, id++, evergreen_emit_cb_misc_state, 4);
- r600_init_atom(rctx, &rctx->clip_misc_state.atom, id++, r600_emit_clip_misc_state, 6);
- r600_init_atom(rctx, &rctx->clip_state.atom, id++, evergreen_emit_clip_state, 26);
- r600_init_atom(rctx, &rctx->db_misc_state.atom, id++, evergreen_emit_db_misc_state, 10);
- r600_init_atom(rctx, &rctx->db_state.atom, id++, evergreen_emit_db_state, 14);
- r600_init_atom(rctx, &rctx->dsa_state.atom, id++, r600_emit_cso_state, 0);
- r600_init_atom(rctx, &rctx->poly_offset_state.atom, id++, evergreen_emit_polygon_offset, 6);
- r600_init_atom(rctx, &rctx->rasterizer_state.atom, id++, r600_emit_cso_state, 0);
- r600_init_atom(rctx, &rctx->scissor.atom, id++, evergreen_emit_scissor_state, 4);
- r600_init_atom(rctx, &rctx->stencil_ref.atom, id++, r600_emit_stencil_ref, 4);
- r600_init_atom(rctx, &rctx->viewport.atom, id++, r600_emit_viewport_state, 8);
- r600_init_atom(rctx, &rctx->vertex_fetch_shader.atom, id++, evergreen_emit_vertex_fetch_shader, 5);
- r600_init_atom(rctx, &rctx->streamout.begin_atom, id++, r600_emit_streamout_begin, 0);
-
- rctx->context.create_blend_state = evergreen_create_blend_state;
- rctx->context.create_depth_stencil_alpha_state = evergreen_create_dsa_state;
- rctx->context.create_rasterizer_state = evergreen_create_rs_state;
- rctx->context.create_sampler_state = evergreen_create_sampler_state;
- rctx->context.create_sampler_view = evergreen_create_sampler_view;
- rctx->context.set_framebuffer_state = evergreen_set_framebuffer_state;
- rctx->context.set_polygon_stipple = evergreen_set_polygon_stipple;
- rctx->context.set_scissor_state = evergreen_set_scissor_state;
- evergreen_init_compute_state_functions(rctx);
+ radeon_set_config_reg(cs, R_008040_WAIT_UNTIL, S_008040_WAIT_3D_IDLE(1));
+ radeon_emit(cs, PKT3(PKT3_EVENT_WRITE, 0, 0));
+ radeon_emit(cs, EVENT_TYPE(EVENT_TYPE_VGT_FLUSH));
}
void cayman_init_common_regs(struct r600_command_buffer *cb,
r600_store_config_reg(cb, R_008D8C_SQ_DYN_GPR_CNTL_PS_FLUSH_REQ, (1 << 8));
- r600_store_context_reg(cb, R_028A4C_PA_SC_MODE_CNTL_1, 0);
-
- r600_store_context_reg(cb, R_028354_SX_SURFACE_SYNC, S_028354_SURFACE_SYNC_MASK(0xf));
+ r600_store_context_reg_seq(cb, R_028350_SX_MISC, 2);
+ r600_store_value(cb, 0);
+ r600_store_value(cb, S_028354_SURFACE_SYNC_MASK(0xf));
r600_store_context_reg(cb, R_028800_DB_DEPTH_CONTROL, 0);
}
static void cayman_init_atom_start_cs(struct r600_context *rctx)
{
struct r600_command_buffer *cb = &rctx->start_cs_cmd;
+ int tmp, i;
- r600_init_command_buffer(cb, 256);
+ r600_init_command_buffer(cb, 320);
/* This must be first. */
r600_store_value(cb, PKT3(PKT3_CONTEXT_CONTROL, 1, 0));
r600_store_value(cb, PKT3(PKT3_EVENT_WRITE, 0, 0));
r600_store_value(cb, EVENT_TYPE(EVENT_TYPE_PS_PARTIAL_FLUSH) | EVENT_INDEX(4));
- cayman_init_common_regs(cb, rctx->chip_class,
- rctx->family, rctx->screen->info.drm_minor);
+ cayman_init_common_regs(cb, rctx->b.chip_class,
+ rctx->b.family, rctx->screen->b.info.drm_minor);
r600_store_config_reg(cb, R_009100_SPI_CONFIG_CNTL, 0);
r600_store_config_reg(cb, R_00913C_SPI_CONFIG_CNTL_1, S_00913C_VTX_DONE_DELAY(4));
r600_store_value(cb, 0); /* R_028A3C_VGT_GROUP_VECT_1_FMT_CNTL */
r600_store_value(cb, 0); /* R_028A40_VGT_GS_MODE */
- r600_store_context_reg_seq(cb, R_028B94_VGT_STRMOUT_CONFIG, 2);
- r600_store_value(cb, 0); /* R_028B94_VGT_STRMOUT_CONFIG */
- r600_store_value(cb, 0); /* R_028B98_VGT_STRMOUT_BUFFER_CONFIG */
+ r600_store_context_reg(cb, R_028B98_VGT_STRMOUT_BUFFER_CONFIG, 0);
r600_store_context_reg_seq(cb, R_028AB4_VGT_REUSE_OFF, 2);
r600_store_value(cb, 0); /* R_028AB4_VGT_REUSE_OFF */
r600_store_config_reg(cb, R_008A14_PA_CL_ENHANCE, (3 << 1) | 1);
- r600_store_context_reg(cb, CM_R_028AA8_IA_MULTI_VGT_PARAM, S_028AA8_SWITCH_ON_EOP(1) | S_028AA8_PARTIAL_VS_WAVE_ON(1) | S_028AA8_PRIMGROUP_SIZE(63));
-
r600_store_context_reg_seq(cb, CM_R_028BD4_PA_SC_CENTROID_PRIORITY_0, 2);
r600_store_value(cb, 0x76543210); /* CM_R_028BD4_PA_SC_CENTROID_PRIORITY_0 */
r600_store_value(cb, 0xfedcba98); /* CM_R_028BD8_PA_SC_CENTROID_PRIORITY_1 */
r600_store_context_reg(cb, R_028200_PA_SC_WINDOW_OFFSET, 0);
r600_store_context_reg(cb, R_02820C_PA_SC_CLIPRECT_RULE, 0xFFFF);
- r600_store_context_reg_seq(cb, R_0282D0_PA_SC_VPORT_ZMIN_0, 2);
- r600_store_value(cb, 0); /* R_0282D0_PA_SC_VPORT_ZMIN_0 */
- r600_store_value(cb, 0x3F800000); /* R_0282D4_PA_SC_VPORT_ZMAX_0 */
+ r600_store_context_reg_seq(cb, R_0282D0_PA_SC_VPORT_ZMIN_0, 2 * R600_MAX_VIEWPORTS);
+ for (tmp = 0; tmp < R600_MAX_VIEWPORTS; tmp++) {
+ r600_store_value(cb, 0); /* R_0282D0_PA_SC_VPORT_ZMIN_0 */
+ r600_store_value(cb, fui(1.0)); /* R_0282D4_PA_SC_VPORT_ZMAX_0 */
+ }
r600_store_context_reg(cb, R_028230_PA_SC_EDGERULE, 0xAAAAAAAA);
- r600_store_context_reg(cb, R_028818_PA_CL_VTE_CNTL, 0x0000043F);
r600_store_context_reg(cb, R_028820_PA_CL_NANINF_CNTL, 0);
r600_store_context_reg_seq(cb, CM_R_028BE8_PA_CL_GB_VERT_CLIP_ADJ, 4);
- r600_store_value(cb, 0x3F800000); /* CM_R_028BE8_PA_CL_GB_VERT_CLIP_ADJ */
- r600_store_value(cb, 0x3F800000); /* CM_R_028BEC_PA_CL_GB_VERT_DISC_ADJ */
- r600_store_value(cb, 0x3F800000); /* CM_R_028BF0_PA_CL_GB_HORZ_CLIP_ADJ */
- r600_store_value(cb, 0x3F800000); /* CM_R_028BF4_PA_CL_GB_HORZ_DISC_ADJ */
+ r600_store_value(cb, fui(1.0)); /* CM_R_028BE8_PA_CL_GB_VERT_CLIP_ADJ */
+ r600_store_value(cb, fui(1.0)); /* CM_R_028BEC_PA_CL_GB_VERT_DISC_ADJ */
+ r600_store_value(cb, fui(1.0)); /* CM_R_028BF0_PA_CL_GB_HORZ_CLIP_ADJ */
+ r600_store_value(cb, fui(1.0)); /* CM_R_028BF4_PA_CL_GB_HORZ_DISC_ADJ */
r600_store_context_reg_seq(cb, R_028240_PA_SC_GENERIC_SCISSOR_TL, 2);
r600_store_value(cb, 0); /* R_028240_PA_SC_GENERIC_SCISSOR_TL */
/* to avoid GPU doing any preloading of constant from random address */
r600_store_context_reg_seq(cb, R_028140_ALU_CONST_BUFFER_SIZE_PS_0, 16);
- r600_store_value(cb, 0); /* R_028140_ALU_CONST_BUFFER_SIZE_PS_0 */
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
+ for (i = 0; i < 16; i++)
+ r600_store_value(cb, 0);
r600_store_context_reg_seq(cb, R_028180_ALU_CONST_BUFFER_SIZE_VS_0, 16);
- r600_store_value(cb, 0); /* R_028180_ALU_CONST_BUFFER_SIZE_VS_0 */
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
+ for (i = 0; i < 16; i++)
+ r600_store_value(cb, 0);
+
+ r600_store_context_reg_seq(cb, R_0281C0_ALU_CONST_BUFFER_SIZE_GS_0, 16);
+ for (i = 0; i < 16; i++)
+ r600_store_value(cb, 0);
+
+ r600_store_context_reg_seq(cb, R_028FC0_ALU_CONST_BUFFER_SIZE_LS_0, 16);
+ for (i = 0; i < 16; i++)
+ r600_store_value(cb, 0);
- if (rctx->screen->has_streamout) {
+ r600_store_context_reg_seq(cb, R_028F80_ALU_CONST_BUFFER_SIZE_HS_0, 16);
+ for (i = 0; i < 16; i++)
+ r600_store_value(cb, 0);
+
+ if (rctx->screen->b.has_streamout) {
r600_store_context_reg(cb, R_028B28_VGT_STRMOUT_DRAW_OPAQUE_OFFSET, 0);
}
+ r600_store_context_reg(cb, R_028010_DB_RENDER_OVERRIDE2, 0);
+ r600_store_context_reg(cb, R_028234_PA_SU_HARDWARE_SCREEN_OFFSET, 0);
+ r600_store_context_reg(cb, R_0286C8_SPI_THREAD_GROUPING, 0);
+ r600_store_context_reg_seq(cb, R_0286E4_SPI_PS_IN_CONTROL_2, 2);
+ r600_store_value(cb, 0); /* R_0286E4_SPI_PS_IN_CONTROL_2 */
+ r600_store_value(cb, 0); /* R_0286E8_SPI_COMPUTE_INPUT_CNTL */
+ r600_store_context_reg(cb, R_028B54_VGT_SHADER_STAGES_EN, 0);
+
eg_store_loop_const(cb, R_03A200_SQ_LOOP_CONST_0, 0x01000FFF);
eg_store_loop_const(cb, R_03A200_SQ_LOOP_CONST_0 + (32 * 4), 0x01000FFF);
+ eg_store_loop_const(cb, R_03A200_SQ_LOOP_CONST_0 + (64 * 4), 0x01000FFF);
}
void evergreen_init_common_regs(struct r600_command_buffer *cb,
r600_store_value(cb, tmp); /* R_008C0C_SQ_GPR_RESOURCE_MGMT_3 */
}
- r600_store_config_reg(cb, R_008E2C_SQ_LDS_RESOURCE_MGMT,
- S_008E2C_NUM_PS_LDS(0x1000) | S_008E2C_NUM_LS_LDS(0x1000));
-
- r600_store_context_reg(cb, R_028A4C_PA_SC_MODE_CNTL_1, 0);
-
/* The cs checker requires this register to be set. */
r600_store_context_reg(cb, R_028800_DB_DEPTH_CONTROL, 0);
- r600_store_context_reg(cb, R_028354_SX_SURFACE_SYNC, S_028354_SURFACE_SYNC_MASK(0xf));
+ r600_store_context_reg_seq(cb, R_028350_SX_MISC, 2);
+ r600_store_value(cb, 0);
+ r600_store_value(cb, S_028354_SURFACE_SYNC_MASK(0xf));
return;
}
int num_hs_stack_entries;
int num_ls_stack_entries;
enum radeon_family family;
- unsigned tmp;
+ unsigned tmp, i;
- if (rctx->chip_class == CAYMAN) {
+ if (rctx->b.chip_class == CAYMAN) {
cayman_init_atom_start_cs(rctx);
return;
}
- r600_init_command_buffer(cb, 256);
+ r600_init_command_buffer(cb, 320);
/* This must be first. */
r600_store_value(cb, PKT3(PKT3_CONTEXT_CONTROL, 1, 0));
r600_store_value(cb, PKT3(PKT3_EVENT_WRITE, 0, 0));
r600_store_value(cb, EVENT_TYPE(EVENT_TYPE_PS_PARTIAL_FLUSH) | EVENT_INDEX(4));
- evergreen_init_common_regs(cb, rctx->chip_class,
- rctx->family, rctx->screen->info.drm_minor);
+ evergreen_init_common_regs(cb, rctx->b.chip_class,
+ rctx->b.family, rctx->screen->b.info.drm_minor);
- family = rctx->family;
+ family = rctx->b.family;
switch (family) {
case CHIP_CEDAR:
default:
tmp |= S_008C28_NUM_LS_STACK_ENTRIES(num_ls_stack_entries);
r600_store_value(cb, tmp); /* R_008C28_SQ_STACK_RESOURCE_MGMT_3 */
+ r600_store_config_reg(cb, R_008E2C_SQ_LDS_RESOURCE_MGMT,
+ S_008E2C_NUM_PS_LDS(0x1000) | S_008E2C_NUM_LS_LDS(0x1000));
+
r600_store_config_reg(cb, R_009100_SPI_CONFIG_CNTL, 0);
r600_store_config_reg(cb, R_00913C_SPI_CONFIG_CNTL_1, S_00913C_VTX_DONE_DELAY(4));
r600_store_context_reg(cb, R_02820C_PA_SC_CLIPRECT_RULE, 0xFFFF);
r600_store_context_reg(cb, R_028230_PA_SC_EDGERULE, 0xAAAAAAAA);
- r600_store_context_reg_seq(cb, R_0282D0_PA_SC_VPORT_ZMIN_0, 2);
- r600_store_value(cb, 0); /* R_0282D0_PA_SC_VPORT_ZMIN_0 */
- r600_store_value(cb, 0x3F800000); /* R_0282D4_PA_SC_VPORT_ZMAX_0 */
+ r600_store_context_reg_seq(cb, R_0282D0_PA_SC_VPORT_ZMIN_0, 2 * R600_MAX_VIEWPORTS);
+ for (tmp = 0; tmp < R600_MAX_VIEWPORTS; tmp++) {
+ r600_store_value(cb, 0); /* R_0282D0_PA_SC_VPORT_ZMIN_0 */
+ r600_store_value(cb, fui(1.0)); /* R_0282D4_PA_SC_VPORT_ZMAX_0 */
+ }
r600_store_context_reg(cb, R_0286DC_SPI_FOG_CNTL, 0);
- r600_store_context_reg(cb, R_028818_PA_CL_VTE_CNTL, 0x0000043F);
r600_store_context_reg(cb, R_028820_PA_CL_NANINF_CNTL, 0);
r600_store_context_reg_seq(cb, R_028AC0_DB_SRESULTS_COMPARE_STATE0, 3);
r600_store_value(cb, 0); /* R_028AC8_DB_PRELOAD_CONTROL */
r600_store_context_reg_seq(cb, R_028C0C_PA_CL_GB_VERT_CLIP_ADJ, 4);
- r600_store_value(cb, 0x3F800000); /* R_028C0C_PA_CL_GB_VERT_CLIP_ADJ */
- r600_store_value(cb, 0x3F800000); /* R_028C10_PA_CL_GB_VERT_DISC_ADJ */
- r600_store_value(cb, 0x3F800000); /* R_028C14_PA_CL_GB_HORZ_CLIP_ADJ */
- r600_store_value(cb, 0x3F800000); /* R_028C18_PA_CL_GB_HORZ_DISC_ADJ */
+ r600_store_value(cb, fui(1.0)); /* R_028C0C_PA_CL_GB_VERT_CLIP_ADJ */
+ r600_store_value(cb, fui(1.0)); /* R_028C10_PA_CL_GB_VERT_DISC_ADJ */
+ r600_store_value(cb, fui(1.0)); /* R_028C14_PA_CL_GB_HORZ_CLIP_ADJ */
+ r600_store_value(cb, fui(1.0)); /* R_028C18_PA_CL_GB_HORZ_DISC_ADJ */
r600_store_context_reg_seq(cb, R_028240_PA_SC_GENERIC_SCISSOR_TL, 2);
r600_store_value(cb, 0); /* R_028240_PA_SC_GENERIC_SCISSOR_TL */
/* to avoid GPU doing any preloading of constant from random address */
r600_store_context_reg_seq(cb, R_028140_ALU_CONST_BUFFER_SIZE_PS_0, 16);
- r600_store_value(cb, 0); /* R_028140_ALU_CONST_BUFFER_SIZE_PS_0 */
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
+ for (i = 0; i < 16; i++)
+ r600_store_value(cb, 0);
r600_store_context_reg_seq(cb, R_028180_ALU_CONST_BUFFER_SIZE_VS_0, 16);
- r600_store_value(cb, 0); /* R_028180_ALU_CONST_BUFFER_SIZE_VS_0 */
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
+ for (i = 0; i < 16; i++)
+ r600_store_value(cb, 0);
+
+ r600_store_context_reg_seq(cb, R_0281C0_ALU_CONST_BUFFER_SIZE_GS_0, 16);
+ for (i = 0; i < 16; i++)
+ r600_store_value(cb, 0);
- r600_store_context_reg_seq(cb, R_028B94_VGT_STRMOUT_CONFIG, 2);
- r600_store_value(cb, 0); /* R_028B94_VGT_STRMOUT_CONFIG */
- r600_store_value(cb, 0); /* R_028B98_VGT_STRMOUT_BUFFER_CONFIG */
+ r600_store_context_reg_seq(cb, R_028FC0_ALU_CONST_BUFFER_SIZE_LS_0, 16);
+ for (i = 0; i < 16; i++)
+ r600_store_value(cb, 0);
- if (rctx->screen->has_streamout) {
+ r600_store_context_reg_seq(cb, R_028F80_ALU_CONST_BUFFER_SIZE_HS_0, 16);
+ for (i = 0; i < 16; i++)
+ r600_store_value(cb, 0);
+
+ r600_store_context_reg(cb, R_028B98_VGT_STRMOUT_BUFFER_CONFIG, 0);
+
+ if (rctx->screen->b.has_streamout) {
r600_store_context_reg(cb, R_028B28_VGT_STRMOUT_DRAW_OPAQUE_OFFSET, 0);
}
+ r600_store_context_reg(cb, R_028010_DB_RENDER_OVERRIDE2, 0);
+ r600_store_context_reg(cb, R_028234_PA_SU_HARDWARE_SCREEN_OFFSET, 0);
+ r600_store_context_reg(cb, R_0286C8_SPI_THREAD_GROUPING, 0);
+ r600_store_context_reg_seq(cb, R_0286E4_SPI_PS_IN_CONTROL_2, 2);
+ r600_store_value(cb, 0); /* R_0286E4_SPI_PS_IN_CONTROL_2 */
+ r600_store_value(cb, 0); /* R_0286E8_SPI_COMPUTE_INPUT_CNTL */
+ r600_store_context_reg(cb, R_0288EC_SQ_LDS_ALLOC_PS, 0);
+ r600_store_context_reg(cb, R_028B54_VGT_SHADER_STAGES_EN, 0);
+
eg_store_loop_const(cb, R_03A200_SQ_LOOP_CONST_0, 0x01000FFF);
eg_store_loop_const(cb, R_03A200_SQ_LOOP_CONST_0 + (32 * 4), 0x01000FFF);
+ eg_store_loop_const(cb, R_03A200_SQ_LOOP_CONST_0 + (64 * 4), 0x01000FFF);
}
-void evergreen_pipe_shader_ps(struct pipe_context *ctx, struct r600_pipe_shader *shader)
+void evergreen_update_ps_state(struct pipe_context *ctx, struct r600_pipe_shader *shader)
{
struct r600_context *rctx = (struct r600_context *)ctx;
- struct r600_pipe_state *rstate = &shader->rstate;
+ struct r600_command_buffer *cb = &shader->command_buffer;
struct r600_shader *rshader = &shader->shader;
unsigned i, exports_ps, num_cout, spi_ps_in_control_0, spi_input_z, spi_ps_in_control_1, db_shader_control = 0;
- int pos_index = -1, face_index = -1;
+ int pos_index = -1, face_index = -1, fixed_pt_position_index = -1;
int ninterp = 0;
- boolean have_linear = FALSE, have_centroid = FALSE, have_perspective = FALSE;
- unsigned spi_baryc_cntl, sid, tmp, idx = 0;
- unsigned z_export = 0, stencil_export = 0;
+ boolean have_perspective = FALSE, have_linear = FALSE;
+ static const unsigned spi_baryc_enable_bit[6] = {
+ S_0286E0_PERSP_SAMPLE_ENA(1),
+ S_0286E0_PERSP_CENTER_ENA(1),
+ S_0286E0_PERSP_CENTROID_ENA(1),
+ S_0286E0_LINEAR_SAMPLE_ENA(1),
+ S_0286E0_LINEAR_CENTER_ENA(1),
+ S_0286E0_LINEAR_CENTROID_ENA(1)
+ };
+ unsigned spi_baryc_cntl = 0, sid, tmp, num = 0;
+ unsigned z_export = 0, stencil_export = 0, mask_export = 0;
unsigned sprite_coord_enable = rctx->rasterizer ? rctx->rasterizer->sprite_coord_enable : 0;
+ uint32_t spi_ps_input_cntl[32];
- rstate->nregs = 0;
+ if (!cb->buf) {
+ r600_init_command_buffer(cb, 64);
+ } else {
+ cb->num_dw = 0;
+ }
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 if (rshader->input[i].name == TGSI_SEMANTIC_FACE) {
+ if (face_index == -1)
+ face_index = i;
+ }
+ else if (rshader->input[i].name == TGSI_SEMANTIC_SAMPLEMASK) {
+ if (face_index == -1)
+ face_index = i; /* lives in same register, same enable bit */
+ }
+ else if (rshader->input[i].name == TGSI_SEMANTIC_SAMPLEID) {
+ fixed_pt_position_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;
+ int k = eg_get_interpolator_index(
+ rshader->input[i].interpolate,
+ rshader->input[i].interpolate_location);
+ if (k >= 0) {
+ spi_baryc_cntl |= spi_baryc_enable_bit[k];
+ have_perspective |= k < 3;
+ have_linear |= !(k < 3);
+ }
}
sid = rshader->input[i].spi_sid;
if (sid) {
-
tmp = S_028644_SEMANTIC(sid);
if (rshader->input[i].name == TGSI_SEMANTIC_POSITION ||
tmp |= S_028644_PT_SPRITE_TEX(1);
}
- r600_pipe_state_add_reg(rstate, R_028644_SPI_PS_INPUT_CNTL_0 + idx * 4,
- tmp);
-
- idx++;
+ spi_ps_input_cntl[num++] = tmp;
}
}
+ r600_store_context_reg_seq(cb, R_028644_SPI_PS_INPUT_CNTL_0, num);
+ r600_store_array(cb, num, spi_ps_input_cntl);
+
for (i = 0; i < rshader->noutput; i++) {
if (rshader->output[i].name == TGSI_SEMANTIC_POSITION)
z_export = 1;
if (rshader->output[i].name == TGSI_SEMANTIC_STENCIL)
stencil_export = 1;
+ if (rshader->output[i].name == TGSI_SEMANTIC_SAMPLEMASK &&
+ rctx->framebuffer.nr_samples > 1 && rctx->ps_iter_samples > 0)
+ mask_export = 1;
}
if (rshader->uses_kill)
db_shader_control |= S_02880C_KILL_ENABLE(1);
db_shader_control |= S_02880C_Z_EXPORT_ENABLE(z_export);
db_shader_control |= S_02880C_STENCIL_EXPORT_ENABLE(stencil_export);
+ db_shader_control |= S_02880C_MASK_EXPORT_ENABLE(mask_export);
exports_ps = 0;
for (i = 0; i < rshader->noutput; i++) {
if (rshader->output[i].name == TGSI_SEMANTIC_POSITION ||
- rshader->output[i].name == TGSI_SEMANTIC_STENCIL)
+ rshader->output[i].name == TGSI_SEMANTIC_STENCIL ||
+ rshader->output[i].name == TGSI_SEMANTIC_SAMPLEMASK)
exports_ps |= 1;
}
ninterp = 1;
have_perspective = TRUE;
}
+ if (!spi_baryc_cntl)
+ spi_baryc_cntl |= spi_baryc_enable_bit[0];
if (!have_perspective && !have_linear)
have_perspective = TRUE;
spi_input_z = 0;
if (pos_index != -1) {
spi_ps_in_control_0 |= S_0286CC_POSITION_ENA(1) |
- S_0286CC_POSITION_CENTROID(rshader->input[pos_index].centroid) |
+ S_0286CC_POSITION_CENTROID(rshader->input[pos_index].interpolate_location == TGSI_INTERPOLATE_LOC_CENTROID) |
S_0286CC_POSITION_ADDR(rshader->input[pos_index].gpr);
- spi_input_z |= 1;
+ spi_input_z |= S_0286D8_PROVIDE_Z_TO_SPI(1);
}
spi_ps_in_control_1 = 0;
spi_ps_in_control_1 |= S_0286D0_FRONT_FACE_ENA(1) |
S_0286D0_FRONT_FACE_ADDR(rshader->input[face_index].gpr);
}
+ if (fixed_pt_position_index != -1) {
+ spi_ps_in_control_1 |= S_0286D0_FIXED_PT_POSITION_ENA(1) |
+ S_0286D0_FIXED_PT_POSITION_ADDR(rshader->input[fixed_pt_position_index].gpr);
+ }
+
+ r600_store_context_reg_seq(cb, R_0286CC_SPI_PS_IN_CONTROL_0, 2);
+ r600_store_value(cb, spi_ps_in_control_0); /* R_0286CC_SPI_PS_IN_CONTROL_0 */
+ r600_store_value(cb, spi_ps_in_control_1); /* R_0286D0_SPI_PS_IN_CONTROL_1 */
+
+ r600_store_context_reg(cb, R_0286E0_SPI_BARYC_CNTL, spi_baryc_cntl);
+ r600_store_context_reg(cb, R_0286D8_SPI_INPUT_Z, spi_input_z);
+ r600_store_context_reg(cb, R_02884C_SQ_PGM_EXPORTS_PS, exports_ps);
- spi_baryc_cntl = 0;
- if (have_perspective)
- spi_baryc_cntl |= S_0286E0_PERSP_CENTER_ENA(1) |
- S_0286E0_PERSP_CENTROID_ENA(have_centroid);
- if (have_linear)
- spi_baryc_cntl |= S_0286E0_LINEAR_CENTER_ENA(1) |
- S_0286E0_LINEAR_CENTROID_ENA(have_centroid);
-
- r600_pipe_state_add_reg(rstate, R_0286CC_SPI_PS_IN_CONTROL_0,
- spi_ps_in_control_0);
- r600_pipe_state_add_reg(rstate, R_0286D0_SPI_PS_IN_CONTROL_1,
- spi_ps_in_control_1);
- r600_pipe_state_add_reg(rstate, R_0286E4_SPI_PS_IN_CONTROL_2,
- 0);
- r600_pipe_state_add_reg(rstate, R_0286D8_SPI_INPUT_Z, spi_input_z);
- r600_pipe_state_add_reg(rstate,
- R_0286E0_SPI_BARYC_CNTL,
- spi_baryc_cntl);
-
- r600_pipe_state_add_reg_bo(rstate,
- R_028840_SQ_PGM_START_PS,
- r600_resource_va(ctx->screen, (void *)shader->bo) >> 8,
- shader->bo, RADEON_USAGE_READ);
- r600_pipe_state_add_reg(rstate,
- R_028844_SQ_PGM_RESOURCES_PS,
- S_028844_NUM_GPRS(rshader->bc.ngpr) |
- S_028844_PRIME_CACHE_ON_DRAW(1) |
- S_028844_STACK_SIZE(rshader->bc.nstack));
- r600_pipe_state_add_reg(rstate,
- R_02884C_SQ_PGM_EXPORTS_PS,
- exports_ps);
+ r600_store_context_reg_seq(cb, R_028840_SQ_PGM_START_PS, 2);
+ r600_store_value(cb, shader->bo->gpu_address >> 8);
+ r600_store_value(cb, /* R_028844_SQ_PGM_RESOURCES_PS */
+ S_028844_NUM_GPRS(rshader->bc.ngpr) |
+ S_028844_PRIME_CACHE_ON_DRAW(1) |
+ S_028844_STACK_SIZE(rshader->bc.nstack));
+ /* After that, the NOP relocation packet must be emitted (shader->bo, RADEON_USAGE_READ). */
shader->db_shader_control = db_shader_control;
- shader->ps_depth_export = z_export | stencil_export;
+ shader->ps_depth_export = z_export | stencil_export | mask_export;
shader->sprite_coord_enable = sprite_coord_enable;
if (rctx->rasterizer)
shader->flatshade = rctx->rasterizer->flatshade;
}
-void evergreen_pipe_shader_vs(struct pipe_context *ctx, struct r600_pipe_shader *shader)
+void evergreen_update_es_state(struct pipe_context *ctx, struct r600_pipe_shader *shader)
+{
+ struct r600_command_buffer *cb = &shader->command_buffer;
+ struct r600_shader *rshader = &shader->shader;
+
+ r600_init_command_buffer(cb, 32);
+
+ r600_store_context_reg(cb, R_028890_SQ_PGM_RESOURCES_ES,
+ S_028890_NUM_GPRS(rshader->bc.ngpr) |
+ S_028890_STACK_SIZE(rshader->bc.nstack));
+ r600_store_context_reg(cb, R_02888C_SQ_PGM_START_ES,
+ shader->bo->gpu_address >> 8);
+ /* After that, the NOP relocation packet must be emitted (shader->bo, RADEON_USAGE_READ). */
+}
+
+void evergreen_update_gs_state(struct pipe_context *ctx, struct r600_pipe_shader *shader)
{
struct r600_context *rctx = (struct r600_context *)ctx;
- struct r600_pipe_state *rstate = &shader->rstate;
+ struct r600_command_buffer *cb = &shader->command_buffer;
+ struct r600_shader *rshader = &shader->shader;
+ struct r600_shader *cp_shader = &shader->gs_copy_shader->shader;
+ unsigned gsvs_itemsizes[4] = {
+ (cp_shader->ring_item_sizes[0] * shader->selector->gs_max_out_vertices) >> 2,
+ (cp_shader->ring_item_sizes[1] * shader->selector->gs_max_out_vertices) >> 2,
+ (cp_shader->ring_item_sizes[2] * shader->selector->gs_max_out_vertices) >> 2,
+ (cp_shader->ring_item_sizes[3] * shader->selector->gs_max_out_vertices) >> 2
+ };
+
+ r600_init_command_buffer(cb, 64);
+
+ /* VGT_GS_MODE is written by evergreen_emit_shader_stages */
+
+ r600_store_context_reg(cb, R_028AB8_VGT_VTX_CNT_EN, 1);
+
+ r600_store_context_reg(cb, R_028B38_VGT_GS_MAX_VERT_OUT,
+ S_028B38_MAX_VERT_OUT(shader->selector->gs_max_out_vertices));
+ r600_store_context_reg(cb, R_028A6C_VGT_GS_OUT_PRIM_TYPE,
+ r600_conv_prim_to_gs_out(shader->selector->gs_output_prim));
+
+ if (rctx->screen->b.info.drm_minor >= 35) {
+ r600_store_context_reg(cb, R_028B90_VGT_GS_INSTANCE_CNT,
+ S_028B90_CNT(MIN2(shader->selector->gs_num_invocations, 127)) |
+ S_028B90_ENABLE(shader->selector->gs_num_invocations > 0));
+ }
+ r600_store_context_reg_seq(cb, R_02891C_SQ_GS_VERT_ITEMSIZE, 4);
+ r600_store_value(cb, cp_shader->ring_item_sizes[0] >> 2);
+ r600_store_value(cb, cp_shader->ring_item_sizes[1] >> 2);
+ r600_store_value(cb, cp_shader->ring_item_sizes[2] >> 2);
+ r600_store_value(cb, cp_shader->ring_item_sizes[3] >> 2);
+
+ r600_store_context_reg(cb, R_028900_SQ_ESGS_RING_ITEMSIZE,
+ (rshader->ring_item_sizes[0]) >> 2);
+
+ r600_store_context_reg(cb, R_028904_SQ_GSVS_RING_ITEMSIZE,
+ gsvs_itemsizes[0] +
+ gsvs_itemsizes[1] +
+ gsvs_itemsizes[2] +
+ gsvs_itemsizes[3]);
+
+ r600_store_context_reg_seq(cb, R_02892C_SQ_GSVS_RING_OFFSET_1, 3);
+ r600_store_value(cb, gsvs_itemsizes[0]);
+ r600_store_value(cb, gsvs_itemsizes[0] + gsvs_itemsizes[1]);
+ r600_store_value(cb, gsvs_itemsizes[0] + gsvs_itemsizes[1] + gsvs_itemsizes[2]);
+
+ /* FIXME calculate these values somehow ??? */
+ r600_store_context_reg_seq(cb, R_028A54_GS_PER_ES, 3);
+ r600_store_value(cb, 0x80); /* GS_PER_ES */
+ r600_store_value(cb, 0x100); /* ES_PER_GS */
+ r600_store_value(cb, 0x2); /* GS_PER_VS */
+
+ r600_store_context_reg(cb, R_028878_SQ_PGM_RESOURCES_GS,
+ S_028878_NUM_GPRS(rshader->bc.ngpr) |
+ S_028878_STACK_SIZE(rshader->bc.nstack));
+ r600_store_context_reg(cb, R_028874_SQ_PGM_START_GS,
+ shader->bo->gpu_address >> 8);
+ /* After that, the NOP relocation packet must be emitted (shader->bo, RADEON_USAGE_READ). */
+}
+
+
+void evergreen_update_vs_state(struct pipe_context *ctx, struct r600_pipe_shader *shader)
+{
+ struct r600_command_buffer *cb = &shader->command_buffer;
struct r600_shader *rshader = &shader->shader;
unsigned spi_vs_out_id[10] = {};
unsigned i, tmp, nparams = 0;
- /* clear previous register */
- rstate->nregs = 0;
-
for (i = 0; i < rshader->noutput; i++) {
if (rshader->output[i].spi_sid) {
tmp = rshader->output[i].spi_sid << ((nparams & 3) * 8);
}
}
+ r600_init_command_buffer(cb, 32);
+
+ r600_store_context_reg_seq(cb, R_02861C_SPI_VS_OUT_ID_0, 10);
for (i = 0; i < 10; i++) {
- r600_pipe_state_add_reg(rstate,
- R_02861C_SPI_VS_OUT_ID_0 + i * 4,
- spi_vs_out_id[i]);
+ r600_store_value(cb, spi_vs_out_id[i]);
}
/* Certain attributes (position, psize, etc.) don't count as params.
if (nparams < 1)
nparams = 1;
- r600_pipe_state_add_reg(rstate,
- R_0286C4_SPI_VS_OUT_CONFIG,
- S_0286C4_VS_EXPORT_COUNT(nparams - 1));
- r600_pipe_state_add_reg(rstate,
- R_028860_SQ_PGM_RESOURCES_VS,
- S_028860_NUM_GPRS(rshader->bc.ngpr) |
- S_028860_STACK_SIZE(rshader->bc.nstack));
- r600_pipe_state_add_reg_bo(rstate,
- R_02885C_SQ_PGM_START_VS,
- r600_resource_va(ctx->screen, (void *)shader->bo) >> 8,
- shader->bo, RADEON_USAGE_READ);
+ r600_store_context_reg(cb, R_0286C4_SPI_VS_OUT_CONFIG,
+ S_0286C4_VS_EXPORT_COUNT(nparams - 1));
+ r600_store_context_reg(cb, R_028860_SQ_PGM_RESOURCES_VS,
+ S_028860_NUM_GPRS(rshader->bc.ngpr) |
+ S_028860_STACK_SIZE(rshader->bc.nstack));
+ if (rshader->vs_position_window_space) {
+ r600_store_context_reg(cb, R_028818_PA_CL_VTE_CNTL,
+ S_028818_VTX_XY_FMT(1) | S_028818_VTX_Z_FMT(1));
+ } else {
+ r600_store_context_reg(cb, R_028818_PA_CL_VTE_CNTL,
+ S_028818_VTX_W0_FMT(1) |
+ S_028818_VPORT_X_SCALE_ENA(1) | S_028818_VPORT_X_OFFSET_ENA(1) |
+ S_028818_VPORT_Y_SCALE_ENA(1) | S_028818_VPORT_Y_OFFSET_ENA(1) |
+ S_028818_VPORT_Z_SCALE_ENA(1) | S_028818_VPORT_Z_OFFSET_ENA(1));
+
+ }
+ r600_store_context_reg(cb, R_02885C_SQ_PGM_START_VS,
+ shader->bo->gpu_address >> 8);
+ /* After that, the NOP relocation packet must be emitted (shader->bo, RADEON_USAGE_READ). */
shader->pa_cl_vs_out_cntl =
S_02881C_VS_OUT_CCDIST0_VEC_ENA((rshader->clip_dist_write & 0x0F) != 0) |
S_02881C_VS_OUT_CCDIST1_VEC_ENA((rshader->clip_dist_write & 0xF0) != 0) |
S_02881C_VS_OUT_MISC_VEC_ENA(rshader->vs_out_misc_write) |
- S_02881C_USE_VTX_POINT_SIZE(rshader->vs_out_point_size);
+ S_02881C_USE_VTX_POINT_SIZE(rshader->vs_out_point_size) |
+ S_02881C_USE_VTX_EDGE_FLAG(rshader->vs_out_edgeflag) |
+ S_02881C_USE_VTX_VIEWPORT_INDX(rshader->vs_out_viewport) |
+ S_02881C_USE_VTX_RENDER_TARGET_INDX(rshader->vs_out_layer);
}
void *evergreen_create_resolve_blend(struct r600_context *rctx)
memset(&blend, 0, sizeof(blend));
blend.independent_blend_enable = true;
blend.rt[0].colormask = 0xf;
- return evergreen_create_blend_state_mode(&rctx->context, &blend, V_028808_CB_RESOLVE);
+ return evergreen_create_blend_state_mode(&rctx->b.b, &blend, V_028808_CB_RESOLVE);
}
void *evergreen_create_decompress_blend(struct r600_context *rctx)
{
struct pipe_blend_state blend;
+ unsigned mode = rctx->screen->has_compressed_msaa_texturing ?
+ V_028808_CB_FMASK_DECOMPRESS : V_028808_CB_DECOMPRESS;
memset(&blend, 0, sizeof(blend));
blend.independent_blend_enable = true;
blend.rt[0].colormask = 0xf;
- return evergreen_create_blend_state_mode(&rctx->context, &blend, V_028808_CB_DECOMPRESS);
+ return evergreen_create_blend_state_mode(&rctx->b.b, &blend, mode);
}
-void *evergreen_create_fmask_decompress_blend(struct r600_context *rctx)
+void *evergreen_create_fastclear_blend(struct r600_context *rctx)
{
struct pipe_blend_state blend;
+ unsigned mode = V_028808_CB_ELIMINATE_FAST_CLEAR;
memset(&blend, 0, sizeof(blend));
blend.independent_blend_enable = true;
blend.rt[0].colormask = 0xf;
- return evergreen_create_blend_state_mode(&rctx->context, &blend, V_028808_CB_FMASK_DECOMPRESS);
+ return evergreen_create_blend_state_mode(&rctx->b.b, &blend, mode);
}
void *evergreen_create_db_flush_dsa(struct r600_context *rctx)
{
struct pipe_depth_stencil_alpha_state dsa = {{0}};
- return rctx->context.create_depth_stencil_alpha_state(&rctx->context, &dsa);
+ return rctx->b.b.create_depth_stencil_alpha_state(&rctx->b.b, &dsa);
}
void evergreen_update_db_shader_control(struct r600_context * rctx)
{
- bool dual_export = rctx->framebuffer.export_16bpc &&
- !rctx->ps_shader->current->ps_depth_export;
+ bool dual_export;
+ unsigned db_shader_control;
+
+ if (!rctx->ps_shader) {
+ return;
+ }
+
+ dual_export = rctx->framebuffer.export_16bpc &&
+ !rctx->ps_shader->current->ps_depth_export;
- unsigned db_shader_control = rctx->ps_shader->current->db_shader_control |
- S_02880C_DUAL_EXPORT_ENABLE(dual_export) |
- S_02880C_DB_SOURCE_FORMAT(dual_export ? V_02880C_EXPORT_DB_TWO :
- V_02880C_EXPORT_DB_FULL) |
- S_02880C_ALPHA_TO_MASK_DISABLE(rctx->framebuffer.cb0_is_integer);
+ db_shader_control = rctx->ps_shader->current->db_shader_control |
+ S_02880C_DUAL_EXPORT_ENABLE(dual_export) |
+ S_02880C_DB_SOURCE_FORMAT(dual_export ? V_02880C_EXPORT_DB_TWO :
+ V_02880C_EXPORT_DB_FULL) |
+ S_02880C_ALPHA_TO_MASK_DISABLE(rctx->framebuffer.cb0_is_integer);
/* When alpha test is enabled we can't trust the hw to make the proper
* decision on the order in which ztest should be run related to fragment
if (db_shader_control != rctx->db_misc_state.db_shader_control) {
rctx->db_misc_state.db_shader_control = db_shader_control;
- rctx->db_misc_state.atom.dirty = true;
+ r600_mark_atom_dirty(rctx, &rctx->db_misc_state.atom);
}
}
unsigned pitch,
unsigned bpp)
{
- struct radeon_winsys_cs *cs = rctx->rings.dma.cs;
+ struct radeon_winsys_cs *cs = rctx->b.rings.dma.cs;
struct r600_texture *rsrc = (struct r600_texture*)src;
struct r600_texture *rdst = (struct r600_texture*)dst;
unsigned array_mode, lbpp, pitch_tile_max, slice_tile_max, size;
unsigned ncopy, height, cheight, detile, i, x, y, z, src_mode, dst_mode;
- unsigned sub_cmd, bank_h, bank_w, mt_aspect, nbanks, tile_split;
+ unsigned sub_cmd, bank_h, bank_w, mt_aspect, nbanks, tile_split, non_disp_tiling = 0;
uint64_t base, addr;
- /* make sure that the dma ring is only one active */
- rctx->rings.gfx.flush(rctx, RADEON_FLUSH_ASYNC);
-
dst_mode = rdst->surface.level[dst_level].mode;
src_mode = rsrc->surface.level[src_level].mode;
/* downcast linear aligned to linear to simplify test */
dst_mode = dst_mode == RADEON_SURF_MODE_LINEAR_ALIGNED ? RADEON_SURF_MODE_LINEAR : dst_mode;
assert(dst_mode != src_mode);
+ /* non_disp_tiling bit needs to be set for depth, stencil, and fmask surfaces */
+ if (util_format_has_depth(util_format_description(src->format)))
+ non_disp_tiling = 1;
+
y = 0;
- sub_cmd = 0x8;
+ sub_cmd = EG_DMA_COPY_TILED;
lbpp = util_logbase2(bpp);
- pitch_tile_max = ((pitch / bpp) >> 3) - 1;
- nbanks = eg_num_banks(rctx->screen->tiling_info.num_banks);
+ pitch_tile_max = ((pitch / bpp) / 8) - 1;
+ nbanks = eg_num_banks(rctx->screen->b.tiling_info.num_banks);
if (dst_mode == RADEON_SURF_MODE_LINEAR) {
/* T2L */
array_mode = evergreen_array_mode(src_mode);
- slice_tile_max = (rsrc->surface.level[src_level].nblk_x * rsrc->surface.level[src_level].nblk_y) >> 6;
+ slice_tile_max = (rsrc->surface.level[src_level].nblk_x * rsrc->surface.level[src_level].nblk_y) / (8*8);
slice_tile_max = slice_tile_max ? slice_tile_max - 1 : 0;
/* linear height must be the same as the slice tile max height, it's ok even
* if the linear destination/source have smaller heigh as the size of the
bank_w = eg_bank_wh(rsrc->surface.bankw);
mt_aspect = eg_macro_tile_aspect(rsrc->surface.mtilea);
tile_split = eg_tile_split(rsrc->surface.tile_split);
- base += r600_resource_va(&rctx->screen->screen, src);
- addr += r600_resource_va(&rctx->screen->screen, dst);
+ base += rsrc->resource.gpu_address;
+ addr += rdst->resource.gpu_address;
} else {
/* L2T */
array_mode = evergreen_array_mode(dst_mode);
- slice_tile_max = (rdst->surface.level[dst_level].nblk_x * rdst->surface.level[dst_level].nblk_y) >> 6;
+ slice_tile_max = (rdst->surface.level[dst_level].nblk_x * rdst->surface.level[dst_level].nblk_y) / (8*8);
slice_tile_max = slice_tile_max ? slice_tile_max - 1 : 0;
/* linear height must be the same as the slice tile max height, it's ok even
* if the linear destination/source have smaller heigh as the size of the
bank_w = eg_bank_wh(rdst->surface.bankw);
mt_aspect = eg_macro_tile_aspect(rdst->surface.mtilea);
tile_split = eg_tile_split(rdst->surface.tile_split);
- base += r600_resource_va(&rctx->screen->screen, dst);
- addr += r600_resource_va(&rctx->screen->screen, src);
+ base += rdst->resource.gpu_address;
+ addr += rsrc->resource.gpu_address;
}
- size = (copy_height * pitch) >> 2;
- ncopy = (size / 0x000fffff) + !!(size % 0x000fffff);
- r600_need_dma_space(rctx, ncopy * 9);
+ size = (copy_height * pitch) / 4;
+ ncopy = (size / EG_DMA_COPY_MAX_SIZE) + !!(size % EG_DMA_COPY_MAX_SIZE);
+ r600_need_dma_space(&rctx->b, ncopy * 9);
for (i = 0; i < ncopy; i++) {
cheight = copy_height;
- if (((cheight * pitch) >> 2) > 0x000fffff) {
- cheight = (0x000fffff << 2) / pitch;
+ if (((cheight * pitch) / 4) > EG_DMA_COPY_MAX_SIZE) {
+ cheight = (EG_DMA_COPY_MAX_SIZE * 4) / pitch;
}
- size = (cheight * pitch) >> 2;
- /* emit reloc before writting cs so that cs is always in consistent state */
- r600_context_bo_reloc(rctx, &rctx->rings.dma, &rsrc->resource, RADEON_USAGE_READ);
- r600_context_bo_reloc(rctx, &rctx->rings.dma, &rdst->resource, RADEON_USAGE_WRITE);
+ size = (cheight * pitch) / 4;
+ /* emit reloc before writing cs so that cs is always in consistent state */
+ radeon_add_to_buffer_list(&rctx->b, &rctx->b.rings.dma, &rsrc->resource,
+ RADEON_USAGE_READ, RADEON_PRIO_MIN);
+ radeon_add_to_buffer_list(&rctx->b, &rctx->b.rings.dma, &rdst->resource,
+ RADEON_USAGE_WRITE, RADEON_PRIO_MIN);
cs->buf[cs->cdw++] = DMA_PACKET(DMA_PACKET_COPY, sub_cmd, size);
cs->buf[cs->cdw++] = base >> 8;
cs->buf[cs->cdw++] = (detile << 31) | (array_mode << 27) |
cs->buf[cs->cdw++] = (pitch_tile_max << 0) | ((height - 1) << 16);
cs->buf[cs->cdw++] = (slice_tile_max << 0);
cs->buf[cs->cdw++] = (x << 0) | (z << 18);
- cs->buf[cs->cdw++] = (y << 0) | (tile_split << 21) | (nbanks << 25);
+ cs->buf[cs->cdw++] = (y << 0) | (tile_split << 21) | (nbanks << 25) | (non_disp_tiling << 28);
cs->buf[cs->cdw++] = addr & 0xfffffffc;
cs->buf[cs->cdw++] = (addr >> 32UL) & 0xff;
copy_height -= cheight;
}
}
-boolean evergreen_dma_blit(struct pipe_context *ctx,
- struct pipe_resource *dst,
- unsigned dst_level,
- unsigned dst_x, unsigned dst_y, unsigned dst_z,
- struct pipe_resource *src,
- unsigned src_level,
- const struct pipe_box *src_box)
+static void evergreen_dma_copy(struct pipe_context *ctx,
+ struct pipe_resource *dst,
+ unsigned dst_level,
+ unsigned dstx, unsigned dsty, unsigned dstz,
+ struct pipe_resource *src,
+ unsigned src_level,
+ const struct pipe_box *src_box)
{
struct r600_context *rctx = (struct r600_context *)ctx;
struct r600_texture *rsrc = (struct r600_texture*)src;
struct r600_texture *rdst = (struct r600_texture*)dst;
unsigned dst_pitch, src_pitch, bpp, dst_mode, src_mode, copy_height;
unsigned src_w, dst_w;
+ unsigned src_x, src_y;
+ unsigned dst_x = dstx, dst_y = dsty, dst_z = dstz;
- if (rctx->rings.dma.cs == NULL) {
- return FALSE;
+ if (rctx->b.rings.dma.cs == NULL) {
+ goto fallback;
}
- if (src->format != dst->format) {
- return FALSE;
+
+ if (dst->target == PIPE_BUFFER && src->target == PIPE_BUFFER) {
+ evergreen_dma_copy_buffer(rctx, dst, src, dst_x, src_box->x, src_box->width);
+ return;
}
+ if (src->format != dst->format || src_box->depth > 1 ||
+ rdst->dirty_level_mask != 0) {
+ goto fallback;
+ }
+
+ if (rsrc->dirty_level_mask) {
+ ctx->flush_resource(ctx, src);
+ }
+
+ src_x = util_format_get_nblocksx(src->format, src_box->x);
+ dst_x = util_format_get_nblocksx(src->format, dst_x);
+ src_y = util_format_get_nblocksy(src->format, src_box->y);
+ dst_y = util_format_get_nblocksy(src->format, dst_y);
+
bpp = rdst->surface.bpe;
dst_pitch = rdst->surface.level[dst_level].pitch_bytes;
src_pitch = rsrc->surface.level[src_level].pitch_bytes;
if (src_pitch != dst_pitch || src_box->x || dst_x || src_w != dst_w) {
/* FIXME evergreen can do partial blit */
- return FALSE;
+ goto fallback;
}
/* the x test here are currently useless (because we don't support partial blit)
* but keep them around so we don't forget about those
*/
- if ((src_pitch & 0x7) || (src_box->x & 0x7) || (dst_x & 0x7) || (src_box->y & 0x7) || (dst_y & 0x7)) {
- return FALSE;
+ if (src_pitch % 8 || src_box->x % 8 || dst_x % 8 || src_box->y % 8 || dst_y % 8) {
+ goto fallback;
+ }
+
+ /* 128 bpp surfaces require non_disp_tiling for both
+ * tiled and linear buffers on cayman. However, async
+ * DMA only supports it on the tiled side. As such
+ * the tile order is backwards after a L2T/T2L packet.
+ */
+ if ((rctx->b.chip_class == CAYMAN) &&
+ (src_mode != dst_mode) &&
+ (util_format_get_blocksize(src->format) >= 16)) {
+ goto fallback;
}
if (src_mode == dst_mode) {
*/
src_offset= rsrc->surface.level[src_level].offset;
src_offset += rsrc->surface.level[src_level].slice_size * src_box->z;
- src_offset += src_box->y * src_pitch + src_box->x * bpp;
+ src_offset += src_y * src_pitch + src_x * bpp;
dst_offset = rdst->surface.level[dst_level].offset;
dst_offset += rdst->surface.level[dst_level].slice_size * dst_z;
dst_offset += dst_y * dst_pitch + dst_x * bpp;
- evergreen_dma_copy(rctx, dst, src, dst_offset, src_offset,
+ evergreen_dma_copy_buffer(rctx, dst, src, dst_offset, src_offset,
src_box->height * src_pitch);
} else {
evergreen_dma_copy_tile(rctx, dst, dst_level, dst_x, dst_y, dst_z,
- src, src_level, src_box->x, src_box->y, src_box->z,
+ src, src_level, src_x, src_y, src_box->z,
copy_height, dst_pitch, bpp);
}
- return TRUE;
+ return;
+
+fallback:
+ r600_resource_copy_region(ctx, dst, dst_level, dstx, dsty, dstz,
+ src, src_level, src_box);
+}
+
+void evergreen_init_state_functions(struct r600_context *rctx)
+{
+ unsigned id = 4;
+ int i;
+ /* !!!
+ * To avoid GPU lockup registers must be emited in a specific order
+ * (no kidding ...). The order below is important and have been
+ * partialy infered from analyzing fglrx command stream.
+ *
+ * Don't reorder atom without carefully checking the effect (GPU lockup
+ * or piglit regression).
+ * !!!
+ */
+
+ r600_init_atom(rctx, &rctx->framebuffer.atom, id++, evergreen_emit_framebuffer_state, 0);
+ /* shader const */
+ r600_init_atom(rctx, &rctx->constbuf_state[PIPE_SHADER_VERTEX].atom, id++, evergreen_emit_vs_constant_buffers, 0);
+ r600_init_atom(rctx, &rctx->constbuf_state[PIPE_SHADER_GEOMETRY].atom, id++, evergreen_emit_gs_constant_buffers, 0);
+ r600_init_atom(rctx, &rctx->constbuf_state[PIPE_SHADER_FRAGMENT].atom, id++, evergreen_emit_ps_constant_buffers, 0);
+ r600_init_atom(rctx, &rctx->constbuf_state[PIPE_SHADER_COMPUTE].atom, id++, evergreen_emit_cs_constant_buffers, 0);
+ /* shader program */
+ r600_init_atom(rctx, &rctx->cs_shader_state.atom, id++, evergreen_emit_cs_shader, 0);
+ /* sampler */
+ r600_init_atom(rctx, &rctx->samplers[PIPE_SHADER_VERTEX].states.atom, id++, evergreen_emit_vs_sampler_states, 0);
+ r600_init_atom(rctx, &rctx->samplers[PIPE_SHADER_GEOMETRY].states.atom, id++, evergreen_emit_gs_sampler_states, 0);
+ r600_init_atom(rctx, &rctx->samplers[PIPE_SHADER_FRAGMENT].states.atom, id++, evergreen_emit_ps_sampler_states, 0);
+ r600_init_atom(rctx, &rctx->samplers[PIPE_SHADER_COMPUTE].states.atom, id++, evergreen_emit_cs_sampler_states, 0);
+ /* resources */
+ r600_init_atom(rctx, &rctx->vertex_buffer_state.atom, id++, evergreen_fs_emit_vertex_buffers, 0);
+ r600_init_atom(rctx, &rctx->cs_vertex_buffer_state.atom, id++, evergreen_cs_emit_vertex_buffers, 0);
+ r600_init_atom(rctx, &rctx->samplers[PIPE_SHADER_VERTEX].views.atom, id++, evergreen_emit_vs_sampler_views, 0);
+ r600_init_atom(rctx, &rctx->samplers[PIPE_SHADER_GEOMETRY].views.atom, id++, evergreen_emit_gs_sampler_views, 0);
+ r600_init_atom(rctx, &rctx->samplers[PIPE_SHADER_FRAGMENT].views.atom, id++, evergreen_emit_ps_sampler_views, 0);
+ r600_init_atom(rctx, &rctx->samplers[PIPE_SHADER_COMPUTE].views.atom, id++, evergreen_emit_cs_sampler_views, 0);
+
+ r600_init_atom(rctx, &rctx->vgt_state.atom, id++, r600_emit_vgt_state, 10);
+
+ if (rctx->b.chip_class == EVERGREEN) {
+ r600_init_atom(rctx, &rctx->sample_mask.atom, id++, evergreen_emit_sample_mask, 3);
+ } else {
+ r600_init_atom(rctx, &rctx->sample_mask.atom, id++, cayman_emit_sample_mask, 4);
+ }
+ rctx->sample_mask.sample_mask = ~0;
+
+ r600_init_atom(rctx, &rctx->alphatest_state.atom, id++, r600_emit_alphatest_state, 6);
+ r600_init_atom(rctx, &rctx->blend_color.atom, id++, r600_emit_blend_color, 6);
+ r600_init_atom(rctx, &rctx->blend_state.atom, id++, r600_emit_cso_state, 0);
+ r600_init_atom(rctx, &rctx->cb_misc_state.atom, id++, evergreen_emit_cb_misc_state, 4);
+ r600_init_atom(rctx, &rctx->clip_misc_state.atom, id++, r600_emit_clip_misc_state, 6);
+ r600_init_atom(rctx, &rctx->clip_state.atom, id++, evergreen_emit_clip_state, 26);
+ r600_init_atom(rctx, &rctx->db_misc_state.atom, id++, evergreen_emit_db_misc_state, 10);
+ r600_init_atom(rctx, &rctx->db_state.atom, id++, evergreen_emit_db_state, 14);
+ r600_init_atom(rctx, &rctx->dsa_state.atom, id++, r600_emit_cso_state, 0);
+ r600_init_atom(rctx, &rctx->poly_offset_state.atom, id++, evergreen_emit_polygon_offset, 6);
+ r600_init_atom(rctx, &rctx->rasterizer_state.atom, id++, r600_emit_cso_state, 0);
+ for (i = 0; i < R600_MAX_VIEWPORTS; i++) {
+ r600_init_atom(rctx, &rctx->viewport[i].atom, id++, r600_emit_viewport_state, 8);
+ r600_init_atom(rctx, &rctx->scissor[i].atom, id++, evergreen_emit_scissor_state, 4);
+ rctx->viewport[i].idx = i;
+ rctx->scissor[i].idx = i;
+ }
+ r600_init_atom(rctx, &rctx->stencil_ref.atom, id++, r600_emit_stencil_ref, 4);
+ r600_init_atom(rctx, &rctx->vertex_fetch_shader.atom, id++, evergreen_emit_vertex_fetch_shader, 5);
+ r600_add_atom(rctx, &rctx->b.streamout.begin_atom, id++);
+ r600_add_atom(rctx, &rctx->b.streamout.enable_atom, id++);
+ r600_init_atom(rctx, &rctx->vertex_shader.atom, id++, r600_emit_shader, 23);
+ r600_init_atom(rctx, &rctx->pixel_shader.atom, id++, r600_emit_shader, 0);
+ r600_init_atom(rctx, &rctx->geometry_shader.atom, id++, r600_emit_shader, 0);
+ r600_init_atom(rctx, &rctx->export_shader.atom, id++, r600_emit_shader, 0);
+ r600_init_atom(rctx, &rctx->shader_stages.atom, id++, evergreen_emit_shader_stages, 6);
+ r600_init_atom(rctx, &rctx->gs_rings.atom, id++, evergreen_emit_gs_rings, 26);
+
+ rctx->b.b.create_blend_state = evergreen_create_blend_state;
+ rctx->b.b.create_depth_stencil_alpha_state = evergreen_create_dsa_state;
+ rctx->b.b.create_rasterizer_state = evergreen_create_rs_state;
+ rctx->b.b.create_sampler_state = evergreen_create_sampler_state;
+ rctx->b.b.create_sampler_view = evergreen_create_sampler_view;
+ rctx->b.b.set_framebuffer_state = evergreen_set_framebuffer_state;
+ rctx->b.b.set_polygon_stipple = evergreen_set_polygon_stipple;
+ rctx->b.b.set_min_samples = evergreen_set_min_samples;
+ rctx->b.b.set_scissor_states = evergreen_set_scissor_states;
+
+ if (rctx->b.chip_class == EVERGREEN)
+ rctx->b.b.get_sample_position = evergreen_get_sample_position;
+ else
+ rctx->b.b.get_sample_position = cayman_get_sample_position;
+ rctx->b.dma_copy = evergreen_dma_copy;
+
+ evergreen_init_compute_state_functions(rctx);
}
projects / mesa.git / blobdiff