X-Git-Url: https://git.libre-soc.org/?a=blobdiff_plain;f=src%2Fintel%2Fisl%2Fisl_surface_state.c;h=c47b8e0e9dad301262b435be0b2c73ea1aefd4da;hb=abc226cf41574454c79477c217e60e8ff1fddfad;hp=5845175de348efddecda90c5db188c5d1c201a54;hpb=7ddb21708c800dfbe3cea2df6aaaf7aed852761b;p=mesa.git diff --git a/src/intel/isl/isl_surface_state.c b/src/intel/isl/isl_surface_state.c index 5845175de34..c47b8e0e9da 100644 --- a/src/intel/isl/isl_surface_state.c +++ b/src/intel/isl/isl_surface_state.c @@ -26,8 +26,10 @@ #define __gen_address_type uint64_t #define __gen_user_data void -static inline uint64_t -__gen_combine_address(void *data, void *loc, uint64_t addr, uint32_t delta) +static uint64_t +__gen_combine_address(__attribute__((unused)) void *data, + __attribute__((unused)) void *loc, uint64_t addr, + uint32_t delta) { return addr + delta; } @@ -37,10 +39,6 @@ __gen_combine_address(void *data, void *loc, uint64_t addr, uint32_t delta) #include "isl_priv.h" -#define __PASTE2(x, y) x ## y -#define __PASTE(x, y) __PASTE2(x, y) -#define isl_genX(x) __PASTE(isl_, genX(x)) - #if GEN_GEN >= 8 static const uint8_t isl_to_gen_halign[] = { [4] = HALIGN4, @@ -113,12 +111,14 @@ get_surftype(enum isl_surf_dim dim, isl_surf_usage_flags_t usage) assert(!(usage & ISL_SURF_USAGE_CUBE_BIT)); return SURFTYPE_1D; case ISL_SURF_DIM_2D: - if (usage & ISL_SURF_USAGE_STORAGE_BIT) { - /* Storage images are always plain 2-D, not cube */ - return SURFTYPE_2D; - } else if (usage & ISL_SURF_USAGE_CUBE_BIT) { + if ((usage & ISL_SURF_USAGE_CUBE_BIT) && + (usage & ISL_SURF_USAGE_TEXTURE_BIT)) { + /* We need SURFTYPE_CUBE to make cube sampling work */ return SURFTYPE_CUBE; } else { + /* Everything else (render and storage) treat cubes as plain + * 2D array textures + */ return SURFTYPE_2D; } case ISL_SURF_DIM_3D: @@ -132,7 +132,7 @@ get_surftype(enum isl_surf_dim dim, isl_surf_usage_flags_t usage) * hardware. Note that this does NOT give you the actual hardware enum values * but an index into the isl_to_gen_[hv]align arrays above. */ -static inline struct isl_extent3d +UNUSED static struct isl_extent3d get_image_alignment(const struct isl_surf *surf) { if (GEN_GEN >= 9) { @@ -172,9 +172,21 @@ get_qpitch(const struct isl_surf *surf) default: unreachable("Bad isl_surf_dim"); case ISL_DIM_LAYOUT_GEN4_2D: - case ISL_DIM_LAYOUT_GEN4_3D: if (GEN_GEN >= 9) { - return isl_surf_get_array_pitch_el_rows(surf); + if (surf->dim == ISL_SURF_DIM_3D && surf->tiling == ISL_TILING_W) { + /* This is rather annoying and completely undocumented. It + * appears that the hardware has a bug (or undocumented feature) + * regarding stencil buffers most likely related to the way + * W-tiling is handled as modified Y-tiling. If you bind a 3-D + * stencil buffer normally, and use texelFetch on it, the z or + * array index will get implicitly multiplied by 2 for no obvious + * reason. The fix appears to be to divide qpitch by 2 for + * W-tiled surfaces. + */ + return isl_surf_get_array_pitch_el_rows(surf) / 2; + } else { + return isl_surf_get_array_pitch_el_rows(surf); + } } else { /* From the Broadwell PRM for RENDER_SURFACE_STATE.QPitch * @@ -199,6 +211,22 @@ get_qpitch(const struct isl_surf *surf) * slices. */ return isl_surf_get_array_pitch_el(surf); + case ISL_DIM_LAYOUT_GEN4_3D: + /* QPitch doesn't make sense for ISL_DIM_LAYOUT_GEN4_3D since it uses a + * different pitch at each LOD. Also, the QPitch field is ignored for + * these surfaces. From the Broadwell PRM documentation for QPitch: + * + * This field specifies the distance in rows between array slices. It + * is used only in the following cases: + * - Surface Array is enabled OR + * - Number of Mulitsamples is not NUMSAMPLES_1 and Multisampled + * Surface Storage Format set to MSFMT_MSS OR + * - Surface Type is SURFTYPE_CUBE + * + * None of the three conditions above can possibly apply to a 3D surface + * so it is safe to just set QPitch to 0. + */ + return 0; } } #endif /* GEN_GEN >= 8 */ @@ -215,12 +243,49 @@ isl_genX(surf_fill_state_s)(const struct isl_device *dev, void *state, assert(isl_format_supports_rendering(dev->info, info->view->format)); else if (info->view->usage & ISL_SURF_USAGE_TEXTURE_BIT) assert(isl_format_supports_sampling(dev->info, info->view->format)); + + /* From the Sky Lake PRM Vol. 2d, RENDER_SURFACE_STATE::SurfaceFormat + * + * This field cannot be a compressed (BC*, DXT*, FXT*, ETC*, EAC*) + * format if the Surface Type is SURFTYPE_1D + */ + if (info->surf->dim == ISL_SURF_DIM_1D) + assert(!isl_format_is_compressed(info->view->format)); + + if (isl_format_is_compressed(info->surf->format)) { + /* You're not allowed to make a view of a compressed format with any + * format other than the surface format. None of the userspace APIs + * allow for this directly and doing so would mess up a number of + * surface parameters such as Width, Height, and alignments. Ideally, + * we'd like to assert that the two formats match. However, we have an + * S3TC workaround that requires us to do reinterpretation. So assert + * that they're at least the same bpb and block size. + */ + ASSERTED const struct isl_format_layout *surf_fmtl = + isl_format_get_layout(info->surf->format); + ASSERTED const struct isl_format_layout *view_fmtl = + isl_format_get_layout(info->surf->format); + assert(surf_fmtl->bpb == view_fmtl->bpb); + assert(surf_fmtl->bw == view_fmtl->bw); + assert(surf_fmtl->bh == view_fmtl->bh); + } + s.SurfaceFormat = info->view->format; +#if GEN_GEN <= 5 + s.ColorBufferComponentWriteDisables = info->write_disables; +#else + assert(info->write_disables == 0); +#endif + #if GEN_IS_HASWELL - s.IntegerSurfaceFormat = isl_format_has_int_channel(s.SurfaceFormat); + s.IntegerSurfaceFormat = + isl_format_has_int_channel((enum isl_format) s.SurfaceFormat); #endif + assert(info->surf->logical_level0_px.width > 0 && + info->surf->logical_level0_px.height > 0); + s.Width = info->surf->logical_level0_px.width - 1; s.Height = info->surf->logical_level0_px.height - 1; @@ -365,7 +430,7 @@ isl_genX(surf_fill_state_s)(const struct isl_device *dev, void *state, /* For gen9 1-D textures, surface pitch is ignored */ s.SurfacePitch = 0; } else { - s.SurfacePitch = info->surf->row_pitch - 1; + s.SurfacePitch = info->surf->row_pitch_B - 1; } #if GEN_GEN >= 8 @@ -389,18 +454,12 @@ isl_genX(surf_fill_state_s)(const struct isl_device *dev, void *state, s.RenderCacheReadWriteMode = 0; #endif - if (info->view->usage & ISL_SURF_USAGE_CUBE_BIT) { -#if GEN_GEN >= 8 - s.CubeFaceEnablePositiveZ = 1; - s.CubeFaceEnableNegativeZ = 1; - s.CubeFaceEnablePositiveY = 1; - s.CubeFaceEnableNegativeY = 1; - s.CubeFaceEnablePositiveX = 1; - s.CubeFaceEnableNegativeX = 1; -#else - s.CubeFaceEnables = 0x3f; -#endif - } + s.CubeFaceEnablePositiveZ = 1; + s.CubeFaceEnableNegativeZ = 1; + s.CubeFaceEnablePositiveY = 1; + s.CubeFaceEnableNegativeY = 1; + s.CubeFaceEnablePositiveX = 1; + s.CubeFaceEnableNegativeX = 1; #if GEN_GEN >= 6 s.NumberofMultisamples = ffs(info->surf->samples) - 1; @@ -411,10 +470,15 @@ isl_genX(surf_fill_state_s)(const struct isl_device *dev, void *state, #endif #if (GEN_GEN >= 8 || GEN_IS_HASWELL) - s.ShaderChannelSelectRed = info->view->swizzle.r; - s.ShaderChannelSelectGreen = info->view->swizzle.g; - s.ShaderChannelSelectBlue = info->view->swizzle.b; - s.ShaderChannelSelectAlpha = info->view->swizzle.a; + if (info->view->usage & ISL_SURF_USAGE_RENDER_TARGET_BIT) + assert(isl_swizzle_supports_rendering(dev->info, info->view->swizzle)); + + s.ShaderChannelSelectRed = (enum GENX(ShaderChannelSelect)) info->view->swizzle.r; + s.ShaderChannelSelectGreen = (enum GENX(ShaderChannelSelect)) info->view->swizzle.g; + s.ShaderChannelSelectBlue = (enum GENX(ShaderChannelSelect)) info->view->swizzle.b; + s.ShaderChannelSelectAlpha = (enum GENX(ShaderChannelSelect)) info->view->swizzle.a; +#else + assert(isl_swizzle_is_identity(info->view->swizzle)); #endif s.SurfaceBaseAddress = info->address; @@ -463,27 +527,52 @@ isl_genX(surf_fill_state_s)(const struct isl_device *dev, void *state, #if GEN_GEN >= 7 if (info->aux_surf && info->aux_usage != ISL_AUX_USAGE_NONE) { + /* The docs don't appear to say anything whatsoever about compression + * and the data port. Testing seems to indicate that the data port + * completely ignores the AuxiliarySurfaceMode field. + */ + assert(!(info->view->usage & ISL_SURF_USAGE_STORAGE_BIT)); + struct isl_tile_info tile_info; - isl_surf_get_tile_info(dev, info->aux_surf, &tile_info); + isl_surf_get_tile_info(info->aux_surf, &tile_info); uint32_t pitch_in_tiles = - info->aux_surf->row_pitch / tile_info.phys_extent_B.width; + info->aux_surf->row_pitch_B / tile_info.phys_extent_B.width; + + s.AuxiliarySurfaceBaseAddress = info->aux_address; + s.AuxiliarySurfacePitch = pitch_in_tiles - 1; #if GEN_GEN >= 8 assert(GEN_GEN >= 9 || info->aux_usage != ISL_AUX_USAGE_CCS_E); - s.AuxiliarySurfacePitch = pitch_in_tiles - 1; /* Auxiliary surfaces in ISL have compressed formats but the hardware * doesn't expect our definition of the compression, it expects qpitch * in units of samples on the main surface. */ s.AuxiliarySurfaceQPitch = isl_surf_get_array_pitch_sa_rows(info->aux_surf) >> 2; - s.AuxiliarySurfaceBaseAddress = info->aux_address; + + if (info->aux_usage == ISL_AUX_USAGE_HIZ) { + /* The number of samples must be 1 */ + assert(info->surf->samples == 1); + + /* The dimension must not be 3D */ + assert(info->surf->dim != ISL_SURF_DIM_3D); + + /* The format must be one of the following: */ + switch (info->view->format) { + case ISL_FORMAT_R32_FLOAT: + case ISL_FORMAT_R24_UNORM_X8_TYPELESS: + case ISL_FORMAT_R16_UNORM: + break; + default: + assert(!"Incompatible HiZ Sampling format"); + break; + } + } + s.AuxiliarySurfaceMode = isl_to_gen_aux_mode[info->aux_usage]; #else assert(info->aux_usage == ISL_AUX_USAGE_MCS || info->aux_usage == ISL_AUX_USAGE_CCS_D); - s.MCSBaseAddress = info->aux_address, - s.MCSSurfacePitch = pitch_in_tiles - 1; s.MCSEnable = true; #endif } @@ -506,17 +595,57 @@ isl_genX(surf_fill_state_s)(const struct isl_device *dev, void *state, s.SamplerL2BypassModeDisable = true; break; default: + /* From the SKL PRM, Programming Note under Sampler Output Channel + * Mapping: + * + * If a surface has an associated HiZ Auxilliary surface, the + * Sampler L2 Bypass Mode Disable field in the RENDER_SURFACE_STATE + * must be set. + */ + if (GEN_GEN >= 9 && info->aux_usage == ISL_AUX_USAGE_HIZ) + s.SamplerL2BypassModeDisable = true; break; } } #endif if (info->aux_usage != ISL_AUX_USAGE_NONE) { + if (info->use_clear_address) { +#if GEN_GEN >= 10 + s.ClearValueAddressEnable = true; + s.ClearValueAddress = info->clear_address; +#else + unreachable("Gen9 and earlier do not support indirect clear colors"); +#endif + } + +#if GEN_GEN == 11 + /* + * From BXML > GT > Shared Functions > vol5c Shared Functions > + * [Structure] RENDER_SURFACE_STATE [BDW+] > ClearColorConversionEnable: + * + * Project: Gen11 + * + * "Enables Pixel backend hw to convert clear values into native format + * and write back to clear address, so that display and sampler can use + * the converted value for resolving fast cleared RTs." + * + * Summary: + * Clear color conversion must be enabled if the clear color is stored + * indirectly and fast color clears are enabled. + */ + if (info->use_clear_address) { + s.ClearColorConversionEnable = true; + } +#endif + #if GEN_GEN >= 9 - s.RedClearColor = info->clear_color.u32[0]; - s.GreenClearColor = info->clear_color.u32[1]; - s.BlueClearColor = info->clear_color.u32[2]; - s.AlphaClearColor = info->clear_color.u32[3]; + if (!info->use_clear_address) { + s.RedClearColor = info->clear_color.u32[0]; + s.GreenClearColor = info->clear_color.u32[1]; + s.BlueClearColor = info->clear_color.u32[2]; + s.AlphaClearColor = info->clear_color.u32[3]; + } #elif GEN_GEN >= 7 /* Prior to Sky Lake, we only have one bit for the clear color which * gives us 0 or 1 in whatever the surface's format happens to be. @@ -550,7 +679,27 @@ void isl_genX(buffer_fill_state_s)(void *state, const struct isl_buffer_fill_state_info *restrict info) { - uint32_t num_elements = info->size / info->stride; + uint64_t buffer_size = info->size_B; + + /* Uniform and Storage buffers need to have surface size not less that the + * aligned 32-bit size of the buffer. To calculate the array lenght on + * unsized arrays in StorageBuffer the last 2 bits store the padding size + * added to the surface, so we can calculate latter the original buffer + * size to know the number of elements. + * + * surface_size = isl_align(buffer_size, 4) + + * (isl_align(buffer_size) - buffer_size) + * + * buffer_size = (surface_size & ~3) - (surface_size & 3) + */ + if (info->format == ISL_FORMAT_RAW || + info->stride_B < isl_format_get_layout(info->format)->bpb / 8) { + assert(info->stride_B == 1); + uint64_t aligned_size = isl_align(buffer_size, 4); + buffer_size = aligned_size + (aligned_size - buffer_size); + } + + uint32_t num_elements = buffer_size / info->stride_B; if (GEN_GEN >= 7) { /* From the IVB PRM, SURFACE_STATE::Height, @@ -562,7 +711,7 @@ isl_genX(buffer_fill_state_s)(void *state, */ if (info->format == ISL_FORMAT_RAW) { assert(num_elements <= (1ull << 30)); - assert((num_elements & 3) == 0); + assert(num_elements > 0); } else { assert(num_elements <= (1ull << 27)); } @@ -593,7 +742,7 @@ isl_genX(buffer_fill_state_s)(void *state, s.Depth = ((num_elements - 1) >> 20) & 0x7f; #endif - s.SurfacePitch = info->stride - 1; + s.SurfacePitch = info->stride_B - 1; #if GEN_GEN >= 6 s.NumberofMultisamples = MULTISAMPLECOUNT_1; @@ -617,11 +766,37 @@ isl_genX(buffer_fill_state_s)(void *state, #endif #if (GEN_GEN >= 8 || GEN_IS_HASWELL) - s.ShaderChannelSelectRed = SCS_RED; - s.ShaderChannelSelectGreen = SCS_GREEN; - s.ShaderChannelSelectBlue = SCS_BLUE; - s.ShaderChannelSelectAlpha = SCS_ALPHA; + s.ShaderChannelSelectRed = (enum GENX(ShaderChannelSelect)) info->swizzle.r; + s.ShaderChannelSelectGreen = (enum GENX(ShaderChannelSelect)) info->swizzle.g; + s.ShaderChannelSelectBlue = (enum GENX(ShaderChannelSelect)) info->swizzle.b; + s.ShaderChannelSelectAlpha = (enum GENX(ShaderChannelSelect)) info->swizzle.a; #endif GENX(RENDER_SURFACE_STATE_pack)(NULL, state, &s); } + +void +isl_genX(null_fill_state)(void *state, struct isl_extent3d size) +{ + struct GENX(RENDER_SURFACE_STATE) s = { + .SurfaceType = SURFTYPE_NULL, + .SurfaceFormat = ISL_FORMAT_B8G8R8A8_UNORM, +#if GEN_GEN >= 7 + .SurfaceArray = size.depth > 0, +#endif +#if GEN_GEN >= 8 + .TileMode = YMAJOR, +#else + .TiledSurface = true, + .TileWalk = TILEWALK_YMAJOR, +#endif + .Width = size.width - 1, + .Height = size.height - 1, + .Depth = size.depth - 1, + .RenderTargetViewExtent = size.depth - 1, +#if GEN_GEN <= 5 + .ColorBufferComponentWriteDisables = 0xf, +#endif + }; + GENX(RENDER_SURFACE_STATE_pack)(NULL, state, &s); +}