]> git.baikalelectronics.ru Git - kernel.git/commitdiff
Merge drm/drm-next into drm-misc-next
authorThomas Zimmermann <tzimmermann@suse.de>
Mon, 1 Aug 2022 13:54:52 +0000 (15:54 +0200)
committerThomas Zimmermann <tzimmermann@suse.de>
Mon, 1 Aug 2022 14:04:00 +0000 (16:04 +0200)
Backmerging to pick up fixes from amdgpu.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
17 files changed:
1  2 
drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.c
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_plane.c
drivers/gpu/drm/i915/gem/i915_gem_ttm.c
drivers/gpu/drm/mgag200/mgag200_drv.c
drivers/gpu/drm/nouveau/dispnv50/base507c.c
drivers/gpu/drm/nouveau/dispnv50/curs507a.c
drivers/gpu/drm/nouveau/dispnv50/disp.c
drivers/gpu/drm/nouveau/dispnv50/ovly507e.c
drivers/gpu/drm/nouveau/dispnv50/wndwc37e.c
drivers/gpu/drm/nouveau/dispnv50/wndwc57e.c
drivers/gpu/drm/rcar-du/rcar_du_crtc.c
drivers/gpu/drm/rcar-du/rcar_du_plane.c
drivers/gpu/drm/rcar-du/rcar_du_vsp.c

index 0000000000000000000000000000000000000000,8cd25b2ea0dca7f01e25171347dd5582a1a365b8..5eb5d31e591dede7a73eb854973024da490c84f3
mode 000000,100644..100644
--- /dev/null
@@@ -1,0 -1,1646 +1,1646 @@@
 -      .destroy        = drm_primary_helper_destroy,
+ // SPDX-License-Identifier: MIT
+ /*
+  * Copyright 2022 Advanced Micro Devices, Inc.
+  *
+  * Permission is hereby granted, free of charge, to any person obtaining a
+  * copy of this software and associated documentation files (the "Software"),
+  * to deal in the Software without restriction, including without limitation
+  * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+  * and/or sell copies of the Software, and to permit persons to whom the
+  * Software is furnished to do so, subject to the following conditions:
+  *
+  * The above copyright notice and this permission notice shall be included in
+  * all copies or substantial portions of the Software.
+  *
+  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+  * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+  * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+  * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+  * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+  * OTHER DEALINGS IN THE SOFTWARE.
+  *
+  * Authors: AMD
+  *
+  */
+ #include <drm/drm_atomic_helper.h>
+ #include <drm/drm_blend.h>
+ #include <drm/drm_gem_atomic_helper.h>
+ #include <drm/drm_plane_helper.h>
+ #include <drm/drm_fourcc.h>
+ #include "amdgpu.h"
+ #include "dal_asic_id.h"
+ #include "amdgpu_display.h"
+ #include "amdgpu_dm_trace.h"
+ #include "gc/gc_11_0_0_offset.h"
+ #include "gc/gc_11_0_0_sh_mask.h"
+ /*
+  * TODO: these are currently initialized to rgb formats only.
+  * For future use cases we should either initialize them dynamically based on
+  * plane capabilities, or initialize this array to all formats, so internal drm
+  * check will succeed, and let DC implement proper check
+  */
+ static const uint32_t rgb_formats[] = {
+       DRM_FORMAT_XRGB8888,
+       DRM_FORMAT_ARGB8888,
+       DRM_FORMAT_RGBA8888,
+       DRM_FORMAT_XRGB2101010,
+       DRM_FORMAT_XBGR2101010,
+       DRM_FORMAT_ARGB2101010,
+       DRM_FORMAT_ABGR2101010,
+       DRM_FORMAT_XRGB16161616,
+       DRM_FORMAT_XBGR16161616,
+       DRM_FORMAT_ARGB16161616,
+       DRM_FORMAT_ABGR16161616,
+       DRM_FORMAT_XBGR8888,
+       DRM_FORMAT_ABGR8888,
+       DRM_FORMAT_RGB565,
+ };
+ static const uint32_t overlay_formats[] = {
+       DRM_FORMAT_XRGB8888,
+       DRM_FORMAT_ARGB8888,
+       DRM_FORMAT_RGBA8888,
+       DRM_FORMAT_XBGR8888,
+       DRM_FORMAT_ABGR8888,
+       DRM_FORMAT_RGB565
+ };
+ static const u32 cursor_formats[] = {
+       DRM_FORMAT_ARGB8888
+ };
+ enum dm_micro_swizzle {
+       MICRO_SWIZZLE_Z = 0,
+       MICRO_SWIZZLE_S = 1,
+       MICRO_SWIZZLE_D = 2,
+       MICRO_SWIZZLE_R = 3
+ };
+ const struct drm_format_info *amd_get_format_info(const struct drm_mode_fb_cmd2 *cmd)
+ {
+       return amdgpu_lookup_format_info(cmd->pixel_format, cmd->modifier[0]);
+ }
+ void fill_blending_from_plane_state(const struct drm_plane_state *plane_state,
+                              bool *per_pixel_alpha, bool *pre_multiplied_alpha,
+                              bool *global_alpha, int *global_alpha_value)
+ {
+       *per_pixel_alpha = false;
+       *pre_multiplied_alpha = true;
+       *global_alpha = false;
+       *global_alpha_value = 0xff;
+       if (plane_state->plane->type != DRM_PLANE_TYPE_OVERLAY)
+               return;
+       if (plane_state->pixel_blend_mode == DRM_MODE_BLEND_PREMULTI ||
+               plane_state->pixel_blend_mode == DRM_MODE_BLEND_COVERAGE) {
+               static const uint32_t alpha_formats[] = {
+                       DRM_FORMAT_ARGB8888,
+                       DRM_FORMAT_RGBA8888,
+                       DRM_FORMAT_ABGR8888,
+               };
+               uint32_t format = plane_state->fb->format->format;
+               unsigned int i;
+               for (i = 0; i < ARRAY_SIZE(alpha_formats); ++i) {
+                       if (format == alpha_formats[i]) {
+                               *per_pixel_alpha = true;
+                               break;
+                       }
+               }
+               if (*per_pixel_alpha && plane_state->pixel_blend_mode == DRM_MODE_BLEND_COVERAGE)
+                       *pre_multiplied_alpha = false;
+       }
+       if (plane_state->alpha < 0xffff) {
+               *global_alpha = true;
+               *global_alpha_value = plane_state->alpha >> 8;
+       }
+ }
+ static void add_modifier(uint64_t **mods, uint64_t *size, uint64_t *cap, uint64_t mod)
+ {
+       if (!*mods)
+               return;
+       if (*cap - *size < 1) {
+               uint64_t new_cap = *cap * 2;
+               uint64_t *new_mods = kmalloc(new_cap * sizeof(uint64_t), GFP_KERNEL);
+               if (!new_mods) {
+                       kfree(*mods);
+                       *mods = NULL;
+                       return;
+               }
+               memcpy(new_mods, *mods, sizeof(uint64_t) * *size);
+               kfree(*mods);
+               *mods = new_mods;
+               *cap = new_cap;
+       }
+       (*mods)[*size] = mod;
+       *size += 1;
+ }
+ bool modifier_has_dcc(uint64_t modifier)
+ {
+       return IS_AMD_FMT_MOD(modifier) && AMD_FMT_MOD_GET(DCC, modifier);
+ }
+ unsigned modifier_gfx9_swizzle_mode(uint64_t modifier)
+ {
+       if (modifier == DRM_FORMAT_MOD_LINEAR)
+               return 0;
+       return AMD_FMT_MOD_GET(TILE, modifier);
+ }
+ static void fill_gfx8_tiling_info_from_flags(union dc_tiling_info *tiling_info,
+                                uint64_t tiling_flags)
+ {
+       /* Fill GFX8 params */
+       if (AMDGPU_TILING_GET(tiling_flags, ARRAY_MODE) == DC_ARRAY_2D_TILED_THIN1) {
+               unsigned int bankw, bankh, mtaspect, tile_split, num_banks;
+               bankw = AMDGPU_TILING_GET(tiling_flags, BANK_WIDTH);
+               bankh = AMDGPU_TILING_GET(tiling_flags, BANK_HEIGHT);
+               mtaspect = AMDGPU_TILING_GET(tiling_flags, MACRO_TILE_ASPECT);
+               tile_split = AMDGPU_TILING_GET(tiling_flags, TILE_SPLIT);
+               num_banks = AMDGPU_TILING_GET(tiling_flags, NUM_BANKS);
+               /* XXX fix me for VI */
+               tiling_info->gfx8.num_banks = num_banks;
+               tiling_info->gfx8.array_mode =
+                               DC_ARRAY_2D_TILED_THIN1;
+               tiling_info->gfx8.tile_split = tile_split;
+               tiling_info->gfx8.bank_width = bankw;
+               tiling_info->gfx8.bank_height = bankh;
+               tiling_info->gfx8.tile_aspect = mtaspect;
+               tiling_info->gfx8.tile_mode =
+                               DC_ADDR_SURF_MICRO_TILING_DISPLAY;
+       } else if (AMDGPU_TILING_GET(tiling_flags, ARRAY_MODE)
+                       == DC_ARRAY_1D_TILED_THIN1) {
+               tiling_info->gfx8.array_mode = DC_ARRAY_1D_TILED_THIN1;
+       }
+       tiling_info->gfx8.pipe_config =
+                       AMDGPU_TILING_GET(tiling_flags, PIPE_CONFIG);
+ }
+ static void fill_gfx9_tiling_info_from_device(const struct amdgpu_device *adev,
+                                 union dc_tiling_info *tiling_info)
+ {
+       /* Fill GFX9 params */
+       tiling_info->gfx9.num_pipes =
+               adev->gfx.config.gb_addr_config_fields.num_pipes;
+       tiling_info->gfx9.num_banks =
+               adev->gfx.config.gb_addr_config_fields.num_banks;
+       tiling_info->gfx9.pipe_interleave =
+               adev->gfx.config.gb_addr_config_fields.pipe_interleave_size;
+       tiling_info->gfx9.num_shader_engines =
+               adev->gfx.config.gb_addr_config_fields.num_se;
+       tiling_info->gfx9.max_compressed_frags =
+               adev->gfx.config.gb_addr_config_fields.max_compress_frags;
+       tiling_info->gfx9.num_rb_per_se =
+               adev->gfx.config.gb_addr_config_fields.num_rb_per_se;
+       tiling_info->gfx9.shaderEnable = 1;
+       if (adev->ip_versions[GC_HWIP][0] >= IP_VERSION(10, 3, 0))
+               tiling_info->gfx9.num_pkrs = adev->gfx.config.gb_addr_config_fields.num_pkrs;
+ }
+ static void fill_gfx9_tiling_info_from_modifier(const struct amdgpu_device *adev,
+                                   union dc_tiling_info *tiling_info,
+                                   uint64_t modifier)
+ {
+       unsigned int mod_bank_xor_bits = AMD_FMT_MOD_GET(BANK_XOR_BITS, modifier);
+       unsigned int mod_pipe_xor_bits = AMD_FMT_MOD_GET(PIPE_XOR_BITS, modifier);
+       unsigned int pkrs_log2 = AMD_FMT_MOD_GET(PACKERS, modifier);
+       unsigned int pipes_log2;
+       pipes_log2 = min(5u, mod_pipe_xor_bits);
+       fill_gfx9_tiling_info_from_device(adev, tiling_info);
+       if (!IS_AMD_FMT_MOD(modifier))
+               return;
+       tiling_info->gfx9.num_pipes = 1u << pipes_log2;
+       tiling_info->gfx9.num_shader_engines = 1u << (mod_pipe_xor_bits - pipes_log2);
+       if (adev->family >= AMDGPU_FAMILY_NV) {
+               tiling_info->gfx9.num_pkrs = 1u << pkrs_log2;
+       } else {
+               tiling_info->gfx9.num_banks = 1u << mod_bank_xor_bits;
+               /* for DCC we know it isn't rb aligned, so rb_per_se doesn't matter. */
+       }
+ }
+ static int validate_dcc(struct amdgpu_device *adev,
+            const enum surface_pixel_format format,
+            const enum dc_rotation_angle rotation,
+            const union dc_tiling_info *tiling_info,
+            const struct dc_plane_dcc_param *dcc,
+            const struct dc_plane_address *address,
+            const struct plane_size *plane_size)
+ {
+       struct dc *dc = adev->dm.dc;
+       struct dc_dcc_surface_param input;
+       struct dc_surface_dcc_cap output;
+       memset(&input, 0, sizeof(input));
+       memset(&output, 0, sizeof(output));
+       if (!dcc->enable)
+               return 0;
+       if (format >= SURFACE_PIXEL_FORMAT_VIDEO_BEGIN ||
+           !dc->cap_funcs.get_dcc_compression_cap)
+               return -EINVAL;
+       input.format = format;
+       input.surface_size.width = plane_size->surface_size.width;
+       input.surface_size.height = plane_size->surface_size.height;
+       input.swizzle_mode = tiling_info->gfx9.swizzle;
+       if (rotation == ROTATION_ANGLE_0 || rotation == ROTATION_ANGLE_180)
+               input.scan = SCAN_DIRECTION_HORIZONTAL;
+       else if (rotation == ROTATION_ANGLE_90 || rotation == ROTATION_ANGLE_270)
+               input.scan = SCAN_DIRECTION_VERTICAL;
+       if (!dc->cap_funcs.get_dcc_compression_cap(dc, &input, &output))
+               return -EINVAL;
+       if (!output.capable)
+               return -EINVAL;
+       if (dcc->independent_64b_blks == 0 &&
+           output.grph.rgb.independent_64b_blks != 0)
+               return -EINVAL;
+       return 0;
+ }
+ static int fill_gfx9_plane_attributes_from_modifiers(struct amdgpu_device *adev,
+                                         const struct amdgpu_framebuffer *afb,
+                                         const enum surface_pixel_format format,
+                                         const enum dc_rotation_angle rotation,
+                                         const struct plane_size *plane_size,
+                                         union dc_tiling_info *tiling_info,
+                                         struct dc_plane_dcc_param *dcc,
+                                         struct dc_plane_address *address,
+                                         const bool force_disable_dcc)
+ {
+       const uint64_t modifier = afb->base.modifier;
+       int ret = 0;
+       fill_gfx9_tiling_info_from_modifier(adev, tiling_info, modifier);
+       tiling_info->gfx9.swizzle = modifier_gfx9_swizzle_mode(modifier);
+       if (modifier_has_dcc(modifier) && !force_disable_dcc) {
+               uint64_t dcc_address = afb->address + afb->base.offsets[1];
+               bool independent_64b_blks = AMD_FMT_MOD_GET(DCC_INDEPENDENT_64B, modifier);
+               bool independent_128b_blks = AMD_FMT_MOD_GET(DCC_INDEPENDENT_128B, modifier);
+               dcc->enable = 1;
+               dcc->meta_pitch = afb->base.pitches[1];
+               dcc->independent_64b_blks = independent_64b_blks;
+               if (AMD_FMT_MOD_GET(TILE_VERSION, modifier) >= AMD_FMT_MOD_TILE_VER_GFX10_RBPLUS) {
+                       if (independent_64b_blks && independent_128b_blks)
+                               dcc->dcc_ind_blk = hubp_ind_block_64b_no_128bcl;
+                       else if (independent_128b_blks)
+                               dcc->dcc_ind_blk = hubp_ind_block_128b;
+                       else if (independent_64b_blks && !independent_128b_blks)
+                               dcc->dcc_ind_blk = hubp_ind_block_64b;
+                       else
+                               dcc->dcc_ind_blk = hubp_ind_block_unconstrained;
+               } else {
+                       if (independent_64b_blks)
+                               dcc->dcc_ind_blk = hubp_ind_block_64b;
+                       else
+                               dcc->dcc_ind_blk = hubp_ind_block_unconstrained;
+               }
+               address->grph.meta_addr.low_part = lower_32_bits(dcc_address);
+               address->grph.meta_addr.high_part = upper_32_bits(dcc_address);
+       }
+       ret = validate_dcc(adev, format, rotation, tiling_info, dcc, address, plane_size);
+       if (ret)
+               drm_dbg_kms(adev_to_drm(adev), "validate_dcc: returned error: %d\n", ret);
+       return ret;
+ }
+ static void add_gfx10_1_modifiers(const struct amdgpu_device *adev,
+                     uint64_t **mods, uint64_t *size, uint64_t *capacity)
+ {
+       int pipe_xor_bits = ilog2(adev->gfx.config.gb_addr_config_fields.num_pipes);
+       add_modifier(mods, size, capacity, AMD_FMT_MOD |
+                   AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_R_X) |
+                   AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX10) |
+                   AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
+                   AMD_FMT_MOD_SET(DCC, 1) |
+                   AMD_FMT_MOD_SET(DCC_CONSTANT_ENCODE, 1) |
+                   AMD_FMT_MOD_SET(DCC_INDEPENDENT_64B, 1) |
+                   AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_64B));
+       add_modifier(mods, size, capacity, AMD_FMT_MOD |
+                   AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_R_X) |
+                   AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX10) |
+                   AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
+                   AMD_FMT_MOD_SET(DCC, 1) |
+                   AMD_FMT_MOD_SET(DCC_RETILE, 1) |
+                   AMD_FMT_MOD_SET(DCC_CONSTANT_ENCODE, 1) |
+                   AMD_FMT_MOD_SET(DCC_INDEPENDENT_64B, 1) |
+                   AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_64B));
+       add_modifier(mods, size, capacity, AMD_FMT_MOD |
+                   AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_R_X) |
+                   AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX10) |
+                   AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits));
+       add_modifier(mods, size, capacity, AMD_FMT_MOD |
+                   AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_S_X) |
+                   AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX10) |
+                   AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits));
+       /* Only supported for 64bpp, will be filtered in dm_plane_format_mod_supported */
+       add_modifier(mods, size, capacity, AMD_FMT_MOD |
+                   AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_D) |
+                   AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9));
+       add_modifier(mods, size, capacity, AMD_FMT_MOD |
+                   AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_S) |
+                   AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9));
+ }
+ static void add_gfx9_modifiers(const struct amdgpu_device *adev,
+                  uint64_t **mods, uint64_t *size, uint64_t *capacity)
+ {
+       int pipes = ilog2(adev->gfx.config.gb_addr_config_fields.num_pipes);
+       int pipe_xor_bits = min(8, pipes +
+                               ilog2(adev->gfx.config.gb_addr_config_fields.num_se));
+       int bank_xor_bits = min(8 - pipe_xor_bits,
+                               ilog2(adev->gfx.config.gb_addr_config_fields.num_banks));
+       int rb = ilog2(adev->gfx.config.gb_addr_config_fields.num_se) +
+                ilog2(adev->gfx.config.gb_addr_config_fields.num_rb_per_se);
+       if (adev->family == AMDGPU_FAMILY_RV) {
+               /* Raven2 and later */
+               bool has_constant_encode = adev->asic_type > CHIP_RAVEN || adev->external_rev_id >= 0x81;
+               /*
+                * No _D DCC swizzles yet because we only allow 32bpp, which
+                * doesn't support _D on DCN
+                */
+               if (has_constant_encode) {
+                       add_modifier(mods, size, capacity, AMD_FMT_MOD |
+                                   AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_S_X) |
+                                   AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9) |
+                                   AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
+                                   AMD_FMT_MOD_SET(BANK_XOR_BITS, bank_xor_bits) |
+                                   AMD_FMT_MOD_SET(DCC, 1) |
+                                   AMD_FMT_MOD_SET(DCC_INDEPENDENT_64B, 1) |
+                                   AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_64B) |
+                                   AMD_FMT_MOD_SET(DCC_CONSTANT_ENCODE, 1));
+               }
+               add_modifier(mods, size, capacity, AMD_FMT_MOD |
+                           AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_S_X) |
+                           AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9) |
+                           AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
+                           AMD_FMT_MOD_SET(BANK_XOR_BITS, bank_xor_bits) |
+                           AMD_FMT_MOD_SET(DCC, 1) |
+                           AMD_FMT_MOD_SET(DCC_INDEPENDENT_64B, 1) |
+                           AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_64B) |
+                           AMD_FMT_MOD_SET(DCC_CONSTANT_ENCODE, 0));
+               if (has_constant_encode) {
+                       add_modifier(mods, size, capacity, AMD_FMT_MOD |
+                                   AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_S_X) |
+                                   AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9) |
+                                   AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
+                                   AMD_FMT_MOD_SET(BANK_XOR_BITS, bank_xor_bits) |
+                                   AMD_FMT_MOD_SET(DCC, 1) |
+                                   AMD_FMT_MOD_SET(DCC_RETILE, 1) |
+                                   AMD_FMT_MOD_SET(DCC_INDEPENDENT_64B, 1) |
+                                   AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_64B) |
+                                   AMD_FMT_MOD_SET(DCC_CONSTANT_ENCODE, 1) |
+                                   AMD_FMT_MOD_SET(RB, rb) |
+                                   AMD_FMT_MOD_SET(PIPE, pipes));
+               }
+               add_modifier(mods, size, capacity, AMD_FMT_MOD |
+                           AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_S_X) |
+                           AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9) |
+                           AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
+                           AMD_FMT_MOD_SET(BANK_XOR_BITS, bank_xor_bits) |
+                           AMD_FMT_MOD_SET(DCC, 1) |
+                           AMD_FMT_MOD_SET(DCC_RETILE, 1) |
+                           AMD_FMT_MOD_SET(DCC_INDEPENDENT_64B, 1) |
+                           AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_64B) |
+                           AMD_FMT_MOD_SET(DCC_CONSTANT_ENCODE, 0) |
+                           AMD_FMT_MOD_SET(RB, rb) |
+                           AMD_FMT_MOD_SET(PIPE, pipes));
+       }
+       /*
+        * Only supported for 64bpp on Raven, will be filtered on format in
+        * dm_plane_format_mod_supported.
+        */
+       add_modifier(mods, size, capacity, AMD_FMT_MOD |
+                   AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_D_X) |
+                   AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9) |
+                   AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
+                   AMD_FMT_MOD_SET(BANK_XOR_BITS, bank_xor_bits));
+       if (adev->family == AMDGPU_FAMILY_RV) {
+               add_modifier(mods, size, capacity, AMD_FMT_MOD |
+                           AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_S_X) |
+                           AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9) |
+                           AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
+                           AMD_FMT_MOD_SET(BANK_XOR_BITS, bank_xor_bits));
+       }
+       /*
+        * Only supported for 64bpp on Raven, will be filtered on format in
+        * dm_plane_format_mod_supported.
+        */
+       add_modifier(mods, size, capacity, AMD_FMT_MOD |
+                   AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_D) |
+                   AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9));
+       if (adev->family == AMDGPU_FAMILY_RV) {
+               add_modifier(mods, size, capacity, AMD_FMT_MOD |
+                           AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_S) |
+                           AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9));
+       }
+ }
+ static void add_gfx10_3_modifiers(const struct amdgpu_device *adev,
+                     uint64_t **mods, uint64_t *size, uint64_t *capacity)
+ {
+       int pipe_xor_bits = ilog2(adev->gfx.config.gb_addr_config_fields.num_pipes);
+       int pkrs = ilog2(adev->gfx.config.gb_addr_config_fields.num_pkrs);
+       add_modifier(mods, size, capacity, AMD_FMT_MOD |
+                   AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_R_X) |
+                   AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX10_RBPLUS) |
+                   AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
+                   AMD_FMT_MOD_SET(PACKERS, pkrs) |
+                   AMD_FMT_MOD_SET(DCC, 1) |
+                   AMD_FMT_MOD_SET(DCC_CONSTANT_ENCODE, 1) |
+                   AMD_FMT_MOD_SET(DCC_INDEPENDENT_64B, 1) |
+                   AMD_FMT_MOD_SET(DCC_INDEPENDENT_128B, 1) |
+                   AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_64B));
+       add_modifier(mods, size, capacity, AMD_FMT_MOD |
+                   AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_R_X) |
+                   AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX10_RBPLUS) |
+                   AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
+                   AMD_FMT_MOD_SET(PACKERS, pkrs) |
+                   AMD_FMT_MOD_SET(DCC, 1) |
+                   AMD_FMT_MOD_SET(DCC_CONSTANT_ENCODE, 1) |
+                   AMD_FMT_MOD_SET(DCC_INDEPENDENT_128B, 1) |
+                   AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_128B));
+       add_modifier(mods, size, capacity, AMD_FMT_MOD |
+                   AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_R_X) |
+                   AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX10_RBPLUS) |
+                   AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
+                   AMD_FMT_MOD_SET(PACKERS, pkrs) |
+                   AMD_FMT_MOD_SET(DCC, 1) |
+                   AMD_FMT_MOD_SET(DCC_RETILE, 1) |
+                   AMD_FMT_MOD_SET(DCC_CONSTANT_ENCODE, 1) |
+                   AMD_FMT_MOD_SET(DCC_INDEPENDENT_64B, 1) |
+                   AMD_FMT_MOD_SET(DCC_INDEPENDENT_128B, 1) |
+                   AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_64B));
+       add_modifier(mods, size, capacity, AMD_FMT_MOD |
+                   AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_R_X) |
+                   AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX10_RBPLUS) |
+                   AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
+                   AMD_FMT_MOD_SET(PACKERS, pkrs) |
+                   AMD_FMT_MOD_SET(DCC, 1) |
+                   AMD_FMT_MOD_SET(DCC_RETILE, 1) |
+                   AMD_FMT_MOD_SET(DCC_CONSTANT_ENCODE, 1) |
+                   AMD_FMT_MOD_SET(DCC_INDEPENDENT_128B, 1) |
+                   AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_128B));
+       add_modifier(mods, size, capacity, AMD_FMT_MOD |
+                   AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_R_X) |
+                   AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX10_RBPLUS) |
+                   AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
+                   AMD_FMT_MOD_SET(PACKERS, pkrs));
+       add_modifier(mods, size, capacity, AMD_FMT_MOD |
+                   AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_S_X) |
+                   AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX10_RBPLUS) |
+                   AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
+                   AMD_FMT_MOD_SET(PACKERS, pkrs));
+       /* Only supported for 64bpp, will be filtered in dm_plane_format_mod_supported */
+       add_modifier(mods, size, capacity, AMD_FMT_MOD |
+                   AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_D) |
+                   AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9));
+       add_modifier(mods, size, capacity, AMD_FMT_MOD |
+                   AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_S) |
+                   AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9));
+ }
+ static void add_gfx11_modifiers(struct amdgpu_device *adev,
+                     uint64_t **mods, uint64_t *size, uint64_t *capacity)
+ {
+       int num_pipes = 0;
+       int pipe_xor_bits = 0;
+       int num_pkrs = 0;
+       int pkrs = 0;
+       u32 gb_addr_config;
+       u8 i = 0;
+       unsigned swizzle_r_x;
+       uint64_t modifier_r_x;
+       uint64_t modifier_dcc_best;
+       uint64_t modifier_dcc_4k;
+       /* TODO: GFX11 IP HW init hasnt finish and we get zero if we read from
+        * adev->gfx.config.gb_addr_config_fields.num_{pkrs,pipes}
+        */
+       gb_addr_config = RREG32_SOC15(GC, 0, regGB_ADDR_CONFIG);
+       ASSERT(gb_addr_config != 0);
+       num_pkrs = 1 << REG_GET_FIELD(gb_addr_config, GB_ADDR_CONFIG, NUM_PKRS);
+       pkrs = ilog2(num_pkrs);
+       num_pipes = 1 << REG_GET_FIELD(gb_addr_config, GB_ADDR_CONFIG, NUM_PIPES);
+       pipe_xor_bits = ilog2(num_pipes);
+       for (i = 0; i < 2; i++) {
+               /* Insert the best one first. */
+               /* R_X swizzle modes are the best for rendering and DCC requires them. */
+               if (num_pipes > 16)
+                       swizzle_r_x = !i ? AMD_FMT_MOD_TILE_GFX11_256K_R_X : AMD_FMT_MOD_TILE_GFX9_64K_R_X;
+               else
+                       swizzle_r_x = !i ? AMD_FMT_MOD_TILE_GFX9_64K_R_X : AMD_FMT_MOD_TILE_GFX11_256K_R_X;
+               modifier_r_x = AMD_FMT_MOD |
+                              AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX11) |
+                              AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
+                              AMD_FMT_MOD_SET(TILE, swizzle_r_x) |
+                              AMD_FMT_MOD_SET(PACKERS, pkrs);
+               /* DCC_CONSTANT_ENCODE is not set because it can't vary with gfx11 (it's implied to be 1). */
+               modifier_dcc_best = modifier_r_x | AMD_FMT_MOD_SET(DCC, 1) |
+                                   AMD_FMT_MOD_SET(DCC_INDEPENDENT_64B, 0) |
+                                   AMD_FMT_MOD_SET(DCC_INDEPENDENT_128B, 1) |
+                                   AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_128B);
+               /* DCC settings for 4K and greater resolutions. (required by display hw) */
+               modifier_dcc_4k = modifier_r_x | AMD_FMT_MOD_SET(DCC, 1) |
+                                 AMD_FMT_MOD_SET(DCC_INDEPENDENT_64B, 1) |
+                                 AMD_FMT_MOD_SET(DCC_INDEPENDENT_128B, 1) |
+                                 AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_64B);
+               add_modifier(mods, size, capacity, modifier_dcc_best);
+               add_modifier(mods, size, capacity, modifier_dcc_4k);
+               add_modifier(mods, size, capacity, modifier_dcc_best | AMD_FMT_MOD_SET(DCC_RETILE, 1));
+               add_modifier(mods, size, capacity, modifier_dcc_4k | AMD_FMT_MOD_SET(DCC_RETILE, 1));
+               add_modifier(mods, size, capacity, modifier_r_x);
+       }
+       add_modifier(mods, size, capacity, AMD_FMT_MOD |
+                       AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX11) |
+                       AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_D));
+ }
+ static int get_plane_modifiers(struct amdgpu_device *adev, unsigned int plane_type, uint64_t **mods)
+ {
+       uint64_t size = 0, capacity = 128;
+       *mods = NULL;
+       /* We have not hooked up any pre-GFX9 modifiers. */
+       if (adev->family < AMDGPU_FAMILY_AI)
+               return 0;
+       *mods = kmalloc(capacity * sizeof(uint64_t), GFP_KERNEL);
+       if (plane_type == DRM_PLANE_TYPE_CURSOR) {
+               add_modifier(mods, &size, &capacity, DRM_FORMAT_MOD_LINEAR);
+               add_modifier(mods, &size, &capacity, DRM_FORMAT_MOD_INVALID);
+               return *mods ? 0 : -ENOMEM;
+       }
+       switch (adev->family) {
+       case AMDGPU_FAMILY_AI:
+       case AMDGPU_FAMILY_RV:
+               add_gfx9_modifiers(adev, mods, &size, &capacity);
+               break;
+       case AMDGPU_FAMILY_NV:
+       case AMDGPU_FAMILY_VGH:
+       case AMDGPU_FAMILY_YC:
+       case AMDGPU_FAMILY_GC_10_3_6:
+       case AMDGPU_FAMILY_GC_10_3_7:
+               if (adev->ip_versions[GC_HWIP][0] >= IP_VERSION(10, 3, 0))
+                       add_gfx10_3_modifiers(adev, mods, &size, &capacity);
+               else
+                       add_gfx10_1_modifiers(adev, mods, &size, &capacity);
+               break;
+       case AMDGPU_FAMILY_GC_11_0_0:
+       case AMDGPU_FAMILY_GC_11_0_2:
+               add_gfx11_modifiers(adev, mods, &size, &capacity);
+               break;
+       }
+       add_modifier(mods, &size, &capacity, DRM_FORMAT_MOD_LINEAR);
+       /* INVALID marks the end of the list. */
+       add_modifier(mods, &size, &capacity, DRM_FORMAT_MOD_INVALID);
+       if (!*mods)
+               return -ENOMEM;
+       return 0;
+ }
+ static int get_plane_formats(const struct drm_plane *plane,
+                            const struct dc_plane_cap *plane_cap,
+                            uint32_t *formats, int max_formats)
+ {
+       int i, num_formats = 0;
+       /*
+        * TODO: Query support for each group of formats directly from
+        * DC plane caps. This will require adding more formats to the
+        * caps list.
+        */
+       switch (plane->type) {
+       case DRM_PLANE_TYPE_PRIMARY:
+               for (i = 0; i < ARRAY_SIZE(rgb_formats); ++i) {
+                       if (num_formats >= max_formats)
+                               break;
+                       formats[num_formats++] = rgb_formats[i];
+               }
+               if (plane_cap && plane_cap->pixel_format_support.nv12)
+                       formats[num_formats++] = DRM_FORMAT_NV12;
+               if (plane_cap && plane_cap->pixel_format_support.p010)
+                       formats[num_formats++] = DRM_FORMAT_P010;
+               if (plane_cap && plane_cap->pixel_format_support.fp16) {
+                       formats[num_formats++] = DRM_FORMAT_XRGB16161616F;
+                       formats[num_formats++] = DRM_FORMAT_ARGB16161616F;
+                       formats[num_formats++] = DRM_FORMAT_XBGR16161616F;
+                       formats[num_formats++] = DRM_FORMAT_ABGR16161616F;
+               }
+               break;
+       case DRM_PLANE_TYPE_OVERLAY:
+               for (i = 0; i < ARRAY_SIZE(overlay_formats); ++i) {
+                       if (num_formats >= max_formats)
+                               break;
+                       formats[num_formats++] = overlay_formats[i];
+               }
+               break;
+       case DRM_PLANE_TYPE_CURSOR:
+               for (i = 0; i < ARRAY_SIZE(cursor_formats); ++i) {
+                       if (num_formats >= max_formats)
+                               break;
+                       formats[num_formats++] = cursor_formats[i];
+               }
+               break;
+       }
+       return num_formats;
+ }
+ #ifdef CONFIG_DRM_AMD_DC_HDR
+ static int attach_color_mgmt_properties(struct amdgpu_display_manager *dm, struct drm_plane *plane)
+ {
+       drm_object_attach_property(&plane->base,
+                                  dm->degamma_lut_property,
+                                  0);
+       drm_object_attach_property(&plane->base,
+                                  dm->degamma_lut_size_property,
+                                  MAX_COLOR_LUT_ENTRIES);
+       drm_object_attach_property(&plane->base, dm->ctm_property,
+                                  0);
+       drm_object_attach_property(&plane->base, dm->sdr_boost_property,
+                                  DEFAULT_SDR_BOOST);
+       return 0;
+ }
+ #endif
+ int fill_plane_buffer_attributes(struct amdgpu_device *adev,
+                            const struct amdgpu_framebuffer *afb,
+                            const enum surface_pixel_format format,
+                            const enum dc_rotation_angle rotation,
+                            const uint64_t tiling_flags,
+                            union dc_tiling_info *tiling_info,
+                            struct plane_size *plane_size,
+                            struct dc_plane_dcc_param *dcc,
+                            struct dc_plane_address *address,
+                            bool tmz_surface,
+                            bool force_disable_dcc)
+ {
+       const struct drm_framebuffer *fb = &afb->base;
+       int ret;
+       memset(tiling_info, 0, sizeof(*tiling_info));
+       memset(plane_size, 0, sizeof(*plane_size));
+       memset(dcc, 0, sizeof(*dcc));
+       memset(address, 0, sizeof(*address));
+       address->tmz_surface = tmz_surface;
+       if (format < SURFACE_PIXEL_FORMAT_VIDEO_BEGIN) {
+               uint64_t addr = afb->address + fb->offsets[0];
+               plane_size->surface_size.x = 0;
+               plane_size->surface_size.y = 0;
+               plane_size->surface_size.width = fb->width;
+               plane_size->surface_size.height = fb->height;
+               plane_size->surface_pitch =
+                       fb->pitches[0] / fb->format->cpp[0];
+               address->type = PLN_ADDR_TYPE_GRAPHICS;
+               address->grph.addr.low_part = lower_32_bits(addr);
+               address->grph.addr.high_part = upper_32_bits(addr);
+       } else if (format < SURFACE_PIXEL_FORMAT_INVALID) {
+               uint64_t luma_addr = afb->address + fb->offsets[0];
+               uint64_t chroma_addr = afb->address + fb->offsets[1];
+               plane_size->surface_size.x = 0;
+               plane_size->surface_size.y = 0;
+               plane_size->surface_size.width = fb->width;
+               plane_size->surface_size.height = fb->height;
+               plane_size->surface_pitch =
+                       fb->pitches[0] / fb->format->cpp[0];
+               plane_size->chroma_size.x = 0;
+               plane_size->chroma_size.y = 0;
+               /* TODO: set these based on surface format */
+               plane_size->chroma_size.width = fb->width / 2;
+               plane_size->chroma_size.height = fb->height / 2;
+               plane_size->chroma_pitch =
+                       fb->pitches[1] / fb->format->cpp[1];
+               address->type = PLN_ADDR_TYPE_VIDEO_PROGRESSIVE;
+               address->video_progressive.luma_addr.low_part =
+                       lower_32_bits(luma_addr);
+               address->video_progressive.luma_addr.high_part =
+                       upper_32_bits(luma_addr);
+               address->video_progressive.chroma_addr.low_part =
+                       lower_32_bits(chroma_addr);
+               address->video_progressive.chroma_addr.high_part =
+                       upper_32_bits(chroma_addr);
+       }
+       if (adev->family >= AMDGPU_FAMILY_AI) {
+               ret = fill_gfx9_plane_attributes_from_modifiers(adev, afb, format,
+                                                               rotation, plane_size,
+                                                               tiling_info, dcc,
+                                                               address,
+                                                               force_disable_dcc);
+               if (ret)
+                       return ret;
+       } else {
+               fill_gfx8_tiling_info_from_flags(tiling_info, tiling_flags);
+       }
+       return 0;
+ }
+ static int dm_plane_helper_prepare_fb(struct drm_plane *plane,
+                                     struct drm_plane_state *new_state)
+ {
+       struct amdgpu_framebuffer *afb;
+       struct drm_gem_object *obj;
+       struct amdgpu_device *adev;
+       struct amdgpu_bo *rbo;
+       struct dm_plane_state *dm_plane_state_new, *dm_plane_state_old;
+       uint32_t domain;
+       int r;
+       if (!new_state->fb) {
+               DRM_DEBUG_KMS("No FB bound\n");
+               return 0;
+       }
+       afb = to_amdgpu_framebuffer(new_state->fb);
+       obj = new_state->fb->obj[0];
+       rbo = gem_to_amdgpu_bo(obj);
+       adev = amdgpu_ttm_adev(rbo->tbo.bdev);
+       r = amdgpu_bo_reserve(rbo, true);
+       if (r) {
+               dev_err(adev->dev, "fail to reserve bo (%d)\n", r);
+               return r;
+       }
+       r = dma_resv_reserve_fences(rbo->tbo.base.resv, 1);
+       if (r) {
+               dev_err(adev->dev, "reserving fence slot failed (%d)\n", r);
+               goto error_unlock;
+       }
+       if (plane->type != DRM_PLANE_TYPE_CURSOR)
+               domain = amdgpu_display_supported_domains(adev, rbo->flags);
+       else
+               domain = AMDGPU_GEM_DOMAIN_VRAM;
+       r = amdgpu_bo_pin(rbo, domain);
+       if (unlikely(r != 0)) {
+               if (r != -ERESTARTSYS)
+                       DRM_ERROR("Failed to pin framebuffer with error %d\n", r);
+               goto error_unlock;
+       }
+       r = amdgpu_ttm_alloc_gart(&rbo->tbo);
+       if (unlikely(r != 0)) {
+               DRM_ERROR("%p bind failed\n", rbo);
+               goto error_unpin;
+       }
+       r = drm_gem_plane_helper_prepare_fb(plane, new_state);
+       if (unlikely(r != 0))
+               goto error_unpin;
+       amdgpu_bo_unreserve(rbo);
+       afb->address = amdgpu_bo_gpu_offset(rbo);
+       amdgpu_bo_ref(rbo);
+       /**
+        * We don't do surface updates on planes that have been newly created,
+        * but we also don't have the afb->address during atomic check.
+        *
+        * Fill in buffer attributes depending on the address here, but only on
+        * newly created planes since they're not being used by DC yet and this
+        * won't modify global state.
+        */
+       dm_plane_state_old = to_dm_plane_state(plane->state);
+       dm_plane_state_new = to_dm_plane_state(new_state);
+       if (dm_plane_state_new->dc_state &&
+           dm_plane_state_old->dc_state != dm_plane_state_new->dc_state) {
+               struct dc_plane_state *plane_state =
+                       dm_plane_state_new->dc_state;
+               bool force_disable_dcc = !plane_state->dcc.enable;
+               fill_plane_buffer_attributes(
+                       adev, afb, plane_state->format, plane_state->rotation,
+                       afb->tiling_flags,
+                       &plane_state->tiling_info, &plane_state->plane_size,
+                       &plane_state->dcc, &plane_state->address,
+                       afb->tmz_surface, force_disable_dcc);
+       }
+       return 0;
+ error_unpin:
+       amdgpu_bo_unpin(rbo);
+ error_unlock:
+       amdgpu_bo_unreserve(rbo);
+       return r;
+ }
+ static void dm_plane_helper_cleanup_fb(struct drm_plane *plane,
+                                      struct drm_plane_state *old_state)
+ {
+       struct amdgpu_bo *rbo;
+       int r;
+       if (!old_state->fb)
+               return;
+       rbo = gem_to_amdgpu_bo(old_state->fb->obj[0]);
+       r = amdgpu_bo_reserve(rbo, false);
+       if (unlikely(r)) {
+               DRM_ERROR("failed to reserve rbo before unpin\n");
+               return;
+       }
+       amdgpu_bo_unpin(rbo);
+       amdgpu_bo_unreserve(rbo);
+       amdgpu_bo_unref(&rbo);
+ }
+ static void get_min_max_dc_plane_scaling(struct drm_device *dev,
+                                        struct drm_framebuffer *fb,
+                                        int *min_downscale, int *max_upscale)
+ {
+       struct amdgpu_device *adev = drm_to_adev(dev);
+       struct dc *dc = adev->dm.dc;
+       /* Caps for all supported planes are the same on DCE and DCN 1 - 3 */
+       struct dc_plane_cap *plane_cap = &dc->caps.planes[0];
+       switch (fb->format->format) {
+       case DRM_FORMAT_P010:
+       case DRM_FORMAT_NV12:
+       case DRM_FORMAT_NV21:
+               *max_upscale = plane_cap->max_upscale_factor.nv12;
+               *min_downscale = plane_cap->max_downscale_factor.nv12;
+               break;
+       case DRM_FORMAT_XRGB16161616F:
+       case DRM_FORMAT_ARGB16161616F:
+       case DRM_FORMAT_XBGR16161616F:
+       case DRM_FORMAT_ABGR16161616F:
+               *max_upscale = plane_cap->max_upscale_factor.fp16;
+               *min_downscale = plane_cap->max_downscale_factor.fp16;
+               break;
+       default:
+               *max_upscale = plane_cap->max_upscale_factor.argb8888;
+               *min_downscale = plane_cap->max_downscale_factor.argb8888;
+               break;
+       }
+       /*
+        * A factor of 1 in the plane_cap means to not allow scaling, ie. use a
+        * scaling factor of 1.0 == 1000 units.
+        */
+       if (*max_upscale == 1)
+               *max_upscale = 1000;
+       if (*min_downscale == 1)
+               *min_downscale = 1000;
+ }
+ int dm_plane_helper_check_state(struct drm_plane_state *state,
+                                      struct drm_crtc_state *new_crtc_state)
+ {
+       struct drm_framebuffer *fb = state->fb;
+       int min_downscale, max_upscale;
+       int min_scale = 0;
+       int max_scale = INT_MAX;
+       /* Plane enabled? Validate viewport and get scaling factors from plane caps. */
+       if (fb && state->crtc) {
+               /* Validate viewport to cover the case when only the position changes */
+               if (state->plane->type != DRM_PLANE_TYPE_CURSOR) {
+                       int viewport_width = state->crtc_w;
+                       int viewport_height = state->crtc_h;
+                       if (state->crtc_x < 0)
+                               viewport_width += state->crtc_x;
+                       else if (state->crtc_x + state->crtc_w > new_crtc_state->mode.crtc_hdisplay)
+                               viewport_width = new_crtc_state->mode.crtc_hdisplay - state->crtc_x;
+                       if (state->crtc_y < 0)
+                               viewport_height += state->crtc_y;
+                       else if (state->crtc_y + state->crtc_h > new_crtc_state->mode.crtc_vdisplay)
+                               viewport_height = new_crtc_state->mode.crtc_vdisplay - state->crtc_y;
+                       if (viewport_width < 0 || viewport_height < 0) {
+                               DRM_DEBUG_ATOMIC("Plane completely outside of screen\n");
+                               return -EINVAL;
+                       } else if (viewport_width < MIN_VIEWPORT_SIZE*2) { /* x2 for width is because of pipe-split. */
+                               DRM_DEBUG_ATOMIC("Viewport width %d smaller than %d\n", viewport_width, MIN_VIEWPORT_SIZE*2);
+                               return -EINVAL;
+                       } else if (viewport_height < MIN_VIEWPORT_SIZE) {
+                               DRM_DEBUG_ATOMIC("Viewport height %d smaller than %d\n", viewport_height, MIN_VIEWPORT_SIZE);
+                               return -EINVAL;
+                       }
+               }
+               /* Get min/max allowed scaling factors from plane caps. */
+               get_min_max_dc_plane_scaling(state->crtc->dev, fb,
+                                            &min_downscale, &max_upscale);
+               /*
+                * Convert to drm convention: 16.16 fixed point, instead of dc's
+                * 1.0 == 1000. Also drm scaling is src/dst instead of dc's
+                * dst/src, so min_scale = 1.0 / max_upscale, etc.
+                */
+               min_scale = (1000 << 16) / max_upscale;
+               max_scale = (1000 << 16) / min_downscale;
+       }
+       return drm_atomic_helper_check_plane_state(
+               state, new_crtc_state, min_scale, max_scale, true, true);
+ }
+ int fill_dc_scaling_info(struct amdgpu_device *adev,
+                               const struct drm_plane_state *state,
+                               struct dc_scaling_info *scaling_info)
+ {
+       int scale_w, scale_h, min_downscale, max_upscale;
+       memset(scaling_info, 0, sizeof(*scaling_info));
+       /* Source is fixed 16.16 but we ignore mantissa for now... */
+       scaling_info->src_rect.x = state->src_x >> 16;
+       scaling_info->src_rect.y = state->src_y >> 16;
+       /*
+        * For reasons we don't (yet) fully understand a non-zero
+        * src_y coordinate into an NV12 buffer can cause a
+        * system hang on DCN1x.
+        * To avoid hangs (and maybe be overly cautious)
+        * let's reject both non-zero src_x and src_y.
+        *
+        * We currently know of only one use-case to reproduce a
+        * scenario with non-zero src_x and src_y for NV12, which
+        * is to gesture the YouTube Android app into full screen
+        * on ChromeOS.
+        */
+       if (((adev->ip_versions[DCE_HWIP][0] == IP_VERSION(1, 0, 0)) ||
+           (adev->ip_versions[DCE_HWIP][0] == IP_VERSION(1, 0, 1))) &&
+           (state->fb && state->fb->format->format == DRM_FORMAT_NV12 &&
+           (scaling_info->src_rect.x != 0 || scaling_info->src_rect.y != 0)))
+               return -EINVAL;
+       scaling_info->src_rect.width = state->src_w >> 16;
+       if (scaling_info->src_rect.width == 0)
+               return -EINVAL;
+       scaling_info->src_rect.height = state->src_h >> 16;
+       if (scaling_info->src_rect.height == 0)
+               return -EINVAL;
+       scaling_info->dst_rect.x = state->crtc_x;
+       scaling_info->dst_rect.y = state->crtc_y;
+       if (state->crtc_w == 0)
+               return -EINVAL;
+       scaling_info->dst_rect.width = state->crtc_w;
+       if (state->crtc_h == 0)
+               return -EINVAL;
+       scaling_info->dst_rect.height = state->crtc_h;
+       /* DRM doesn't specify clipping on destination output. */
+       scaling_info->clip_rect = scaling_info->dst_rect;
+       /* Validate scaling per-format with DC plane caps */
+       if (state->plane && state->plane->dev && state->fb) {
+               get_min_max_dc_plane_scaling(state->plane->dev, state->fb,
+                                            &min_downscale, &max_upscale);
+       } else {
+               min_downscale = 250;
+               max_upscale = 16000;
+       }
+       scale_w = scaling_info->dst_rect.width * 1000 /
+                 scaling_info->src_rect.width;
+       if (scale_w < min_downscale || scale_w > max_upscale)
+               return -EINVAL;
+       scale_h = scaling_info->dst_rect.height * 1000 /
+                 scaling_info->src_rect.height;
+       if (scale_h < min_downscale || scale_h > max_upscale)
+               return -EINVAL;
+       /*
+        * The "scaling_quality" can be ignored for now, quality = 0 has DC
+        * assume reasonable defaults based on the format.
+        */
+       return 0;
+ }
+ static int dm_plane_atomic_check(struct drm_plane *plane,
+                                struct drm_atomic_state *state)
+ {
+       struct drm_plane_state *new_plane_state = drm_atomic_get_new_plane_state(state,
+                                                                                plane);
+       struct amdgpu_device *adev = drm_to_adev(plane->dev);
+       struct dc *dc = adev->dm.dc;
+       struct dm_plane_state *dm_plane_state;
+       struct dc_scaling_info scaling_info;
+       struct drm_crtc_state *new_crtc_state;
+       int ret;
+       trace_amdgpu_dm_plane_atomic_check(new_plane_state);
+       dm_plane_state = to_dm_plane_state(new_plane_state);
+       if (!dm_plane_state->dc_state)
+               return 0;
+       new_crtc_state =
+               drm_atomic_get_new_crtc_state(state,
+                                             new_plane_state->crtc);
+       if (!new_crtc_state)
+               return -EINVAL;
+       ret = dm_plane_helper_check_state(new_plane_state, new_crtc_state);
+       if (ret)
+               return ret;
+       ret = fill_dc_scaling_info(adev, new_plane_state, &scaling_info);
+       if (ret)
+               return ret;
+       if (dc_validate_plane(dc, dm_plane_state->dc_state) == DC_OK)
+               return 0;
+       return -EINVAL;
+ }
+ static int dm_plane_atomic_async_check(struct drm_plane *plane,
+                                      struct drm_atomic_state *state)
+ {
+       /* Only support async updates on cursor planes. */
+       if (plane->type != DRM_PLANE_TYPE_CURSOR)
+               return -EINVAL;
+       return 0;
+ }
+ static int get_cursor_position(struct drm_plane *plane, struct drm_crtc *crtc,
+                              struct dc_cursor_position *position)
+ {
+       struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+       int x, y;
+       int xorigin = 0, yorigin = 0;
+       if (!crtc || !plane->state->fb)
+               return 0;
+       if ((plane->state->crtc_w > amdgpu_crtc->max_cursor_width) ||
+           (plane->state->crtc_h > amdgpu_crtc->max_cursor_height)) {
+               DRM_ERROR("%s: bad cursor width or height %d x %d\n",
+                         __func__,
+                         plane->state->crtc_w,
+                         plane->state->crtc_h);
+               return -EINVAL;
+       }
+       x = plane->state->crtc_x;
+       y = plane->state->crtc_y;
+       if (x <= -amdgpu_crtc->max_cursor_width ||
+           y <= -amdgpu_crtc->max_cursor_height)
+               return 0;
+       if (x < 0) {
+               xorigin = min(-x, amdgpu_crtc->max_cursor_width - 1);
+               x = 0;
+       }
+       if (y < 0) {
+               yorigin = min(-y, amdgpu_crtc->max_cursor_height - 1);
+               y = 0;
+       }
+       position->enable = true;
+       position->translate_by_source = true;
+       position->x = x;
+       position->y = y;
+       position->x_hotspot = xorigin;
+       position->y_hotspot = yorigin;
+       return 0;
+ }
+ void handle_cursor_update(struct drm_plane *plane,
+                                struct drm_plane_state *old_plane_state)
+ {
+       struct amdgpu_device *adev = drm_to_adev(plane->dev);
+       struct amdgpu_framebuffer *afb = to_amdgpu_framebuffer(plane->state->fb);
+       struct drm_crtc *crtc = afb ? plane->state->crtc : old_plane_state->crtc;
+       struct dm_crtc_state *crtc_state = crtc ? to_dm_crtc_state(crtc->state) : NULL;
+       struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+       uint64_t address = afb ? afb->address : 0;
+       struct dc_cursor_position position = {0};
+       struct dc_cursor_attributes attributes;
+       int ret;
+       if (!plane->state->fb && !old_plane_state->fb)
+               return;
+       DC_LOG_CURSOR("%s: crtc_id=%d with size %d to %d\n",
+                     __func__,
+                     amdgpu_crtc->crtc_id,
+                     plane->state->crtc_w,
+                     plane->state->crtc_h);
+       ret = get_cursor_position(plane, crtc, &position);
+       if (ret)
+               return;
+       if (!position.enable) {
+               /* turn off cursor */
+               if (crtc_state && crtc_state->stream) {
+                       mutex_lock(&adev->dm.dc_lock);
+                       dc_stream_set_cursor_position(crtc_state->stream,
+                                                     &position);
+                       mutex_unlock(&adev->dm.dc_lock);
+               }
+               return;
+       }
+       amdgpu_crtc->cursor_width = plane->state->crtc_w;
+       amdgpu_crtc->cursor_height = plane->state->crtc_h;
+       memset(&attributes, 0, sizeof(attributes));
+       attributes.address.high_part = upper_32_bits(address);
+       attributes.address.low_part  = lower_32_bits(address);
+       attributes.width             = plane->state->crtc_w;
+       attributes.height            = plane->state->crtc_h;
+       attributes.color_format      = CURSOR_MODE_COLOR_PRE_MULTIPLIED_ALPHA;
+       attributes.rotation_angle    = 0;
+       attributes.attribute_flags.value = 0;
+       attributes.pitch = afb->base.pitches[0] / afb->base.format->cpp[0];
+       if (crtc_state->stream) {
+               mutex_lock(&adev->dm.dc_lock);
+               if (!dc_stream_set_cursor_attributes(crtc_state->stream,
+                                                        &attributes))
+                       DRM_ERROR("DC failed to set cursor attributes\n");
+               if (!dc_stream_set_cursor_position(crtc_state->stream,
+                                                  &position))
+                       DRM_ERROR("DC failed to set cursor position\n");
+               mutex_unlock(&adev->dm.dc_lock);
+       }
+ }
+ static void dm_plane_atomic_async_update(struct drm_plane *plane,
+                                        struct drm_atomic_state *state)
+ {
+       struct drm_plane_state *new_state = drm_atomic_get_new_plane_state(state,
+                                                                          plane);
+       struct drm_plane_state *old_state =
+               drm_atomic_get_old_plane_state(state, plane);
+       trace_amdgpu_dm_atomic_update_cursor(new_state);
+       swap(plane->state->fb, new_state->fb);
+       plane->state->src_x = new_state->src_x;
+       plane->state->src_y = new_state->src_y;
+       plane->state->src_w = new_state->src_w;
+       plane->state->src_h = new_state->src_h;
+       plane->state->crtc_x = new_state->crtc_x;
+       plane->state->crtc_y = new_state->crtc_y;
+       plane->state->crtc_w = new_state->crtc_w;
+       plane->state->crtc_h = new_state->crtc_h;
+       handle_cursor_update(plane, old_state);
+ }
+ static const struct drm_plane_helper_funcs dm_plane_helper_funcs = {
+       .prepare_fb = dm_plane_helper_prepare_fb,
+       .cleanup_fb = dm_plane_helper_cleanup_fb,
+       .atomic_check = dm_plane_atomic_check,
+       .atomic_async_check = dm_plane_atomic_async_check,
+       .atomic_async_update = dm_plane_atomic_async_update
+ };
+ static void dm_drm_plane_reset(struct drm_plane *plane)
+ {
+       struct dm_plane_state *amdgpu_state = NULL;
+       if (plane->state)
+               plane->funcs->atomic_destroy_state(plane, plane->state);
+       amdgpu_state = kzalloc(sizeof(*amdgpu_state), GFP_KERNEL);
+       WARN_ON(amdgpu_state == NULL);
+       if (amdgpu_state)
+               __drm_atomic_helper_plane_reset(plane, &amdgpu_state->base);
+ #ifdef CONFIG_DRM_AMD_DC_HDR
+       if (amdgpu_state)
+               amdgpu_state->sdr_boost = DEFAULT_SDR_BOOST;
+ #endif
+ }
+ static struct drm_plane_state *
+ dm_drm_plane_duplicate_state(struct drm_plane *plane)
+ {
+       struct dm_plane_state *dm_plane_state, *old_dm_plane_state;
+       old_dm_plane_state = to_dm_plane_state(plane->state);
+       dm_plane_state = kzalloc(sizeof(*dm_plane_state), GFP_KERNEL);
+       if (!dm_plane_state)
+               return NULL;
+       __drm_atomic_helper_plane_duplicate_state(plane, &dm_plane_state->base);
+       if (old_dm_plane_state->dc_state) {
+               dm_plane_state->dc_state = old_dm_plane_state->dc_state;
+               dc_plane_state_retain(dm_plane_state->dc_state);
+       }
+ #ifdef CONFIG_DRM_AMD_DC_HDR
+       if (dm_plane_state->degamma_lut)
+               drm_property_blob_get(dm_plane_state->degamma_lut);
+       if (dm_plane_state->ctm)
+               drm_property_blob_get(dm_plane_state->ctm);
+       dm_plane_state->sdr_boost = old_dm_plane_state->sdr_boost;
+ #endif
+       return &dm_plane_state->base;
+ }
+ static bool dm_plane_format_mod_supported(struct drm_plane *plane,
+                                         uint32_t format,
+                                         uint64_t modifier)
+ {
+       struct amdgpu_device *adev = drm_to_adev(plane->dev);
+       const struct drm_format_info *info = drm_format_info(format);
+       struct hw_asic_id asic_id = adev->dm.dc->ctx->asic_id;
+       enum dm_micro_swizzle microtile = modifier_gfx9_swizzle_mode(modifier) & 3;
+       if (!info)
+               return false;
+       /*
+        * We always have to allow these modifiers:
+        * 1. Core DRM checks for LINEAR support if userspace does not provide modifiers.
+        * 2. Not passing any modifiers is the same as explicitly passing INVALID.
+        */
+       if (modifier == DRM_FORMAT_MOD_LINEAR ||
+           modifier == DRM_FORMAT_MOD_INVALID) {
+               return true;
+       }
+       /* check if swizzle mode is supported by this version of DCN */
+       switch (asic_id.chip_family) {
+       case FAMILY_SI:
+       case FAMILY_CI:
+       case FAMILY_KV:
+       case FAMILY_CZ:
+       case FAMILY_VI:
+               /* asics before AI does not have modifier support */
+               return false;
+       case FAMILY_AI:
+       case FAMILY_RV:
+       case FAMILY_NV:
+       case FAMILY_VGH:
+       case FAMILY_YELLOW_CARP:
+       case AMDGPU_FAMILY_GC_10_3_6:
+       case AMDGPU_FAMILY_GC_10_3_7:
+               switch (AMD_FMT_MOD_GET(TILE, modifier)) {
+               case AMD_FMT_MOD_TILE_GFX9_64K_R_X:
+               case AMD_FMT_MOD_TILE_GFX9_64K_D_X:
+               case AMD_FMT_MOD_TILE_GFX9_64K_S_X:
+               case AMD_FMT_MOD_TILE_GFX9_64K_D:
+                       return true;
+               default:
+                       return false;
+               }
+               break;
+       case AMDGPU_FAMILY_GC_11_0_0:
+       case AMDGPU_FAMILY_GC_11_0_2:
+               switch (AMD_FMT_MOD_GET(TILE, modifier)) {
+               case AMD_FMT_MOD_TILE_GFX11_256K_R_X:
+               case AMD_FMT_MOD_TILE_GFX9_64K_R_X:
+               case AMD_FMT_MOD_TILE_GFX9_64K_D_X:
+               case AMD_FMT_MOD_TILE_GFX9_64K_S_X:
+               case AMD_FMT_MOD_TILE_GFX9_64K_D:
+                       return true;
+               default:
+                       return false;
+               }
+               break;
+       default:
+               ASSERT(0); /* Unknown asic */
+               break;
+       }
+       /*
+        * For D swizzle the canonical modifier depends on the bpp, so check
+        * it here.
+        */
+       if (AMD_FMT_MOD_GET(TILE_VERSION, modifier) == AMD_FMT_MOD_TILE_VER_GFX9 &&
+           adev->family >= AMDGPU_FAMILY_NV) {
+               if (microtile == MICRO_SWIZZLE_D && info->cpp[0] == 4)
+                       return false;
+       }
+       if (adev->family >= AMDGPU_FAMILY_RV && microtile == MICRO_SWIZZLE_D &&
+           info->cpp[0] < 8)
+               return false;
+       if (modifier_has_dcc(modifier)) {
+               /* Per radeonsi comments 16/64 bpp are more complicated. */
+               if (info->cpp[0] != 4)
+                       return false;
+               /* We support multi-planar formats, but not when combined with
+                * additional DCC metadata planes.
+                */
+               if (info->num_planes > 1)
+                       return false;
+       }
+       return true;
+ }
+ static void dm_drm_plane_destroy_state(struct drm_plane *plane,
+                               struct drm_plane_state *state)
+ {
+       struct dm_plane_state *dm_plane_state = to_dm_plane_state(state);
+ #ifdef CONFIG_DRM_AMD_DC_HDR
+       drm_property_blob_put(dm_plane_state->degamma_lut);
+       drm_property_blob_put(dm_plane_state->ctm);
+ #endif
+       if (dm_plane_state->dc_state)
+               dc_plane_state_release(dm_plane_state->dc_state);
+       drm_atomic_helper_plane_destroy_state(plane, state);
+ }
+ #ifdef CONFIG_DRM_AMD_DC_HDR
+ /* copied from drm_atomic_uapi.c */
+ static int atomic_replace_property_blob_from_id(struct drm_device *dev,
+                                        struct drm_property_blob **blob,
+                                        uint64_t blob_id,
+                                        ssize_t expected_size,
+                                        ssize_t expected_elem_size,
+                                        bool *replaced)
+ {
+       struct drm_property_blob *new_blob = NULL;
+       if (blob_id != 0) {
+               new_blob = drm_property_lookup_blob(dev, blob_id);
+               if (new_blob == NULL)
+                       return -EINVAL;
+               if (expected_size > 0 &&
+                   new_blob->length != expected_size) {
+                       drm_property_blob_put(new_blob);
+                       return -EINVAL;
+               }
+               if (expected_elem_size > 0 &&
+                   new_blob->length % expected_elem_size != 0) {
+                       drm_property_blob_put(new_blob);
+                       return -EINVAL;
+               }
+       }
+       *replaced |= drm_property_replace_blob(blob, new_blob);
+       drm_property_blob_put(new_blob);
+       return 0;
+ }
+ int dm_drm_plane_set_property(struct drm_plane *plane,
+                             struct drm_plane_state *state,
+                             struct drm_property *property,
+                             uint64_t val)
+ {
+       struct amdgpu_device *adev = drm_to_adev(plane->dev);
+       struct dm_plane_state *dm_plane_state = to_dm_plane_state(state);
+       int ret = 0;
+       bool replaced;
+       if (property == adev->dm.degamma_lut_property) {
+               ret = atomic_replace_property_blob_from_id(adev_to_drm(adev),
+                               &dm_plane_state->degamma_lut,
+                               val, -1, sizeof(struct drm_color_lut),
+                               &replaced);
+       } else if (property == adev->dm.ctm_property) {
+               ret = atomic_replace_property_blob_from_id(adev_to_drm(adev),
+                               &dm_plane_state->ctm,
+                               val,
+                               sizeof(struct drm_color_ctm), -1,
+                               &replaced);
+       } else if (property == adev->dm.sdr_boost_property) {
+               dm_plane_state->sdr_boost = val;
+       } else {
+               return -EINVAL;
+       }
+       return ret;
+ }
+ int dm_drm_plane_get_property(struct drm_plane *plane,
+                             const struct drm_plane_state *state,
+                             struct drm_property *property,
+                             uint64_t *val)
+ {
+       struct dm_plane_state *dm_plane_state = to_dm_plane_state(state);
+       struct amdgpu_device *adev = drm_to_adev(plane->dev);
+       if (property == adev->dm.degamma_lut_property) {
+               *val = (dm_plane_state->degamma_lut) ?
+                       dm_plane_state->degamma_lut->base.id : 0;
+       } else if (property == adev->dm.ctm_property) {
+               *val = (dm_plane_state->ctm) ? dm_plane_state->ctm->base.id : 0;
+       } else if (property == adev->dm.sdr_boost_property) {
+               *val = dm_plane_state->sdr_boost;
+       } else {
+               return -EINVAL;
+       }
+       return 0;
+ }
+ #endif
+ static const struct drm_plane_funcs dm_plane_funcs = {
+       .update_plane   = drm_atomic_helper_update_plane,
+       .disable_plane  = drm_atomic_helper_disable_plane,
++      .destroy        = drm_plane_helper_destroy,
+       .reset = dm_drm_plane_reset,
+       .atomic_duplicate_state = dm_drm_plane_duplicate_state,
+       .atomic_destroy_state = dm_drm_plane_destroy_state,
+       .format_mod_supported = dm_plane_format_mod_supported,
+ #ifdef CONFIG_DRM_AMD_DC_HDR
+       .atomic_set_property = dm_drm_plane_set_property,
+       .atomic_get_property = dm_drm_plane_get_property,
+ #endif
+ };
+ int amdgpu_dm_plane_init(struct amdgpu_display_manager *dm,
+                               struct drm_plane *plane,
+                               unsigned long possible_crtcs,
+                               const struct dc_plane_cap *plane_cap)
+ {
+       uint32_t formats[32];
+       int num_formats;
+       int res = -EPERM;
+       unsigned int supported_rotations;
+       uint64_t *modifiers = NULL;
+       num_formats = get_plane_formats(plane, plane_cap, formats,
+                                       ARRAY_SIZE(formats));
+       res = get_plane_modifiers(dm->adev, plane->type, &modifiers);
+       if (res)
+               return res;
+       res = drm_universal_plane_init(adev_to_drm(dm->adev), plane, possible_crtcs,
+                                      &dm_plane_funcs, formats, num_formats,
+                                      modifiers, plane->type, NULL);
+       kfree(modifiers);
+       if (res)
+               return res;
+       if (plane->type == DRM_PLANE_TYPE_OVERLAY &&
+           plane_cap && plane_cap->per_pixel_alpha) {
+               unsigned int blend_caps = BIT(DRM_MODE_BLEND_PIXEL_NONE) |
+                                         BIT(DRM_MODE_BLEND_PREMULTI) |
+                                         BIT(DRM_MODE_BLEND_COVERAGE);
+               drm_plane_create_alpha_property(plane);
+               drm_plane_create_blend_mode_property(plane, blend_caps);
+       }
+       if (plane->type == DRM_PLANE_TYPE_PRIMARY &&
+           plane_cap &&
+           (plane_cap->pixel_format_support.nv12 ||
+            plane_cap->pixel_format_support.p010)) {
+               /* This only affects YUV formats. */
+               drm_plane_create_color_properties(
+                       plane,
+                       BIT(DRM_COLOR_YCBCR_BT601) |
+                       BIT(DRM_COLOR_YCBCR_BT709) |
+                       BIT(DRM_COLOR_YCBCR_BT2020),
+                       BIT(DRM_COLOR_YCBCR_LIMITED_RANGE) |
+                       BIT(DRM_COLOR_YCBCR_FULL_RANGE),
+                       DRM_COLOR_YCBCR_BT709, DRM_COLOR_YCBCR_LIMITED_RANGE);
+       }
+       supported_rotations =
+               DRM_MODE_ROTATE_0 | DRM_MODE_ROTATE_90 |
+               DRM_MODE_ROTATE_180 | DRM_MODE_ROTATE_270;
+       if (dm->adev->asic_type >= CHIP_BONAIRE &&
+           plane->type != DRM_PLANE_TYPE_CURSOR)
+               drm_plane_create_rotation_property(plane, DRM_MODE_ROTATE_0,
+                                                  supported_rotations);
+       drm_plane_helper_add(plane, &dm_plane_helper_funcs);
+ #ifdef CONFIG_DRM_AMD_DC_HDR
+       attach_color_mgmt_properties(dm, plane);
+ #endif
+       /* Create (reset) the plane state */
+       if (plane->funcs->reset)
+               plane->funcs->reset(plane);
+       return 0;
+ }
Simple merge
index 8c2da56defe0dd85d0f0b5c97faed9a05ec21fbd,d4af69e903ad7cacf7ec8850f328bd7a9b76c1f2..797c1e4e0eaa35f01a474ec3d8a65b18fe6eddcd
@@@ -24,9 -24,9 +24,8 @@@
  
  #include <drm/drm_atomic_helper.h>
  #include <drm/drm_fourcc.h>
 -#include <drm/drm_plane_helper.h>
  
- #include <nvif/cl507e.h>
- #include <nvif/event.h>
+ #include <nvif/if0014.h>
  #include <nvif/push507c.h>
  
  #include <nvhw/class/cl507e.h>
index 9cf599d14ce8b76d26a126ead974b83c2ccf3c80,082a66d595068bc148a6dca5948779b9ffe39c47..b3deea5aca58d79f6803b9a66f5a78b057bf480f
  #include "atom.h"
  
  #include <drm/drm_atomic_helper.h>
 -#include <drm/drm_plane_helper.h>
  #include <nouveau_bo.h>
  
- #include <nvif/clc37e.h>
+ #include <nvif/if0014.h>
  #include <nvif/pushc37b.h>
  
  #include <nvhw/class/clc37e.h>
index 1a9a07c77da1c613ae7c085322722e3c8cb2c263,31167c3987085323c4fcb27d61751e5431e05d50..1d214a4b960ab9021a1be543d6c36b36e3db22c9
@@@ -23,9 -23,9 +23,8 @@@
  #include "atom.h"
  
  #include <drm/drm_atomic_helper.h>
 -#include <drm/drm_plane_helper.h>
  #include <nouveau_bo.h>
  
- #include <nvif/clc37e.h>
  #include <nvif/pushc37b.h>
  
  #include <nvhw/class/clc57e.h>
Simple merge