* [igt-dev] [PATCH v3] kms_rotation_crc:Add HW rotation test case for amdgpu
@ 2021-01-21 23:03 Sung Joon Kim
2021-01-21 23:08 ` [igt-dev] ✗ Fi.CI.BUILD: failure for kms_rotation_crc:Add HW rotation test case for amdgpu (rev3) Patchwork
2021-01-22 7:26 ` [igt-dev] [PATCH v3] kms_rotation_crc:Add HW rotation test case for amdgpu Petri Latvala
0 siblings, 2 replies; 3+ messages in thread
From: Sung Joon Kim @ 2021-01-21 23:03 UTC (permalink / raw
To: igt-dev
Added Hw rotation case specifically for amdgpu. Currently, kms_rotation_crc tests intel gpus. Added conditions to bypass all the requirements needed for intel when testing amdgpu.
v3: code cleanup. Freed unused fb. Merged all files into one patch due to dependency. Added tiling methods for amdgpu. Updated drm header for amdgpu tiling modifiers.
Signed-off-by: Sung Joon Kim <sungkim@amd.com>
---
include/drm-uapi/drm_fourcc.h | 134 +++++++++++++++++
lib/igt_amd.c | 103 ++++++++++---
lib/igt_amd.h | 35 +----
lib/igt_fb.c | 23 ++-
tests/kms_rotation_crc.c | 270 +++++++++++++++-------------------
5 files changed, 357 insertions(+), 208 deletions(-)
diff --git a/include/drm-uapi/drm_fourcc.h b/include/drm-uapi/drm_fourcc.h
index 82f32780..a3286346 100644
--- a/include/drm-uapi/drm_fourcc.h
+++ b/include/drm-uapi/drm_fourcc.h
@@ -1056,6 +1056,140 @@ drm_fourcc_canonicalize_nvidia_format_mod(__u64 modifier)
*/
#define AMLOGIC_FBC_OPTION_MEM_SAVING (1ULL << 0)
+/*
+ * AMD modifiers
+ *
+ * Memory layout:
+ *
+ * without DCC:
+ * - main surface
+ *
+ * with DCC & without DCC_RETILE:
+ * - main surface in plane 0
+ * - DCC surface in plane 1 (RB-aligned, pipe-aligned if DCC_PIPE_ALIGN is set)
+ *
+ * with DCC & DCC_RETILE:
+ * - main surface in plane 0
+ * - displayable DCC surface in plane 1 (not RB-aligned & not pipe-aligned)
+ * - pipe-aligned DCC surface in plane 2 (RB-aligned & pipe-aligned)
+ *
+ * For multi-plane formats the above surfaces get merged into one plane for
+ * each format plane, based on the required alignment only.
+ *
+ * Bits Parameter Notes
+ * ----- ------------------------ ---------------------------------------------
+ *
+ * 7:0 TILE_VERSION Values are AMD_FMT_MOD_TILE_VER_*
+ * 12:8 TILE Values are AMD_FMT_MOD_TILE_<version>_*
+ * 13 DCC
+ * 14 DCC_RETILE
+ * 15 DCC_PIPE_ALIGN
+ * 16 DCC_INDEPENDENT_64B
+ * 17 DCC_INDEPENDENT_128B
+ * 19:18 DCC_MAX_COMPRESSED_BLOCK Values are AMD_FMT_MOD_DCC_BLOCK_*
+ * 20 DCC_CONSTANT_ENCODE
+ * 23:21 PIPE_XOR_BITS Only for some chips
+ * 26:24 BANK_XOR_BITS Only for some chips
+ * 29:27 PACKERS Only for some chips
+ * 32:30 RB Only for some chips
+ * 35:33 PIPE Only for some chips
+ * 55:36 - Reserved for future use, must be zero
+ */
+#define AMD_FMT_MOD fourcc_mod_code(AMD, 0)
+
+#define IS_AMD_FMT_MOD(val) (((val) >> 56) == DRM_FORMAT_MOD_VENDOR_AMD)
+
+/* Reserve 0 for GFX8 and older */
+#define AMD_FMT_MOD_TILE_VER_GFX9 1
+#define AMD_FMT_MOD_TILE_VER_GFX10 2
+#define AMD_FMT_MOD_TILE_VER_GFX10_RBPLUS 3
+
+/*
+ * 64K_S is the same for GFX9/GFX10/GFX10_RBPLUS and hence has GFX9 as canonical
+ * version.
+ */
+#define AMD_FMT_MOD_TILE_GFX9_64K_S 9
+
+/*
+ * 64K_D for non-32 bpp is the same for GFX9/GFX10/GFX10_RBPLUS and hence has
+ * GFX9 as canonical version.
+ */
+#define AMD_FMT_MOD_TILE_GFX9_64K_D 10
+#define AMD_FMT_MOD_TILE_GFX9_64K_S_X 25
+#define AMD_FMT_MOD_TILE_GFX9_64K_D_X 26
+#define AMD_FMT_MOD_TILE_GFX9_64K_R_X 27
+
+#define AMD_FMT_MOD_DCC_BLOCK_64B 0
+#define AMD_FMT_MOD_DCC_BLOCK_128B 1
+#define AMD_FMT_MOD_DCC_BLOCK_256B 2
+
+#define AMD_FMT_MOD_TILE_VERSION_SHIFT 0
+#define AMD_FMT_MOD_TILE_VERSION_MASK 0xFF
+#define AMD_FMT_MOD_TILE_SHIFT 8
+#define AMD_FMT_MOD_TILE_MASK 0x1F
+
+/* Whether DCC compression is enabled. */
+#define AMD_FMT_MOD_DCC_SHIFT 13
+#define AMD_FMT_MOD_DCC_MASK 0x1
+
+/*
+ * Whether to include two DCC surfaces, one which is rb & pipe aligned, and
+ * one which is not-aligned.
+ */
+#define AMD_FMT_MOD_DCC_RETILE_SHIFT 14
+#define AMD_FMT_MOD_DCC_RETILE_MASK 0x1
+
+/* Only set if DCC_RETILE = false */
+#define AMD_FMT_MOD_DCC_PIPE_ALIGN_SHIFT 15
+#define AMD_FMT_MOD_DCC_PIPE_ALIGN_MASK 0x1
+
+#define AMD_FMT_MOD_DCC_INDEPENDENT_64B_SHIFT 16
+#define AMD_FMT_MOD_DCC_INDEPENDENT_64B_MASK 0x1
+#define AMD_FMT_MOD_DCC_INDEPENDENT_128B_SHIFT 17
+#define AMD_FMT_MOD_DCC_INDEPENDENT_128B_MASK 0x1
+#define AMD_FMT_MOD_DCC_MAX_COMPRESSED_BLOCK_SHIFT 18
+#define AMD_FMT_MOD_DCC_MAX_COMPRESSED_BLOCK_MASK 0x3
+
+/*
+ * DCC supports embedding some clear colors directly in the DCC surface.
+ * However, on older GPUs the rendering HW ignores the embedded clear color
+ * and prefers the driver provided color. This necessitates doing a fastclear
+ * eliminate operation before a process transfers control.
+ *
+ * If this bit is set that means the fastclear eliminate is not needed for these
+ * embeddable colors.
+ */
+#define AMD_FMT_MOD_DCC_CONSTANT_ENCODE_SHIFT 20
+#define AMD_FMT_MOD_DCC_CONSTANT_ENCODE_MASK 0x1
+
+/*
+ * The below fields are for accounting for per GPU differences. These are only
+ * relevant for GFX9 and later and if the tile field is *_X/_T.
+ *
+ * PIPE_XOR_BITS = always needed
+ * BANK_XOR_BITS = only for TILE_VER_GFX9
+ * PACKERS = only for TILE_VER_GFX10_RBPLUS
+ * RB = only for TILE_VER_GFX9 & DCC
+ * PIPE = only for TILE_VER_GFX9 & DCC & (DCC_RETILE | DCC_PIPE_ALIGN)
+ */
+#define AMD_FMT_MOD_PIPE_XOR_BITS_SHIFT 21
+#define AMD_FMT_MOD_PIPE_XOR_BITS_MASK 0x7
+#define AMD_FMT_MOD_BANK_XOR_BITS_SHIFT 24
+#define AMD_FMT_MOD_BANK_XOR_BITS_MASK 0x7
+#define AMD_FMT_MOD_PACKERS_SHIFT 27
+#define AMD_FMT_MOD_PACKERS_MASK 0x7
+#define AMD_FMT_MOD_RB_SHIFT 30
+#define AMD_FMT_MOD_RB_MASK 0x7
+#define AMD_FMT_MOD_PIPE_SHIFT 33
+#define AMD_FMT_MOD_PIPE_MASK 0x7
+
+#define AMD_FMT_MOD_SET(field, value) \
+ ((uint64_t)(value) << AMD_FMT_MOD_##field##_SHIFT)
+#define AMD_FMT_MOD_GET(field, value) \
+ (((value) >> AMD_FMT_MOD_##field##_SHIFT) & AMD_FMT_MOD_##field##_MASK)
+#define AMD_FMT_MOD_CLEAR(field) \
+ (~((uint64_t)AMD_FMT_MOD_##field##_MASK << AMD_FMT_MOD_##field##_SHIFT))
+
#if defined(__cplusplus)
}
#endif
diff --git a/lib/igt_amd.c b/lib/igt_amd.c
index 737e5670..54082ba9 100644
--- a/lib/igt_amd.c
+++ b/lib/igt_amd.c
@@ -24,6 +24,30 @@
#include "igt.h"
#include <amdgpu_drm.h>
+#define X0 1
+#define X1 2
+#define X2 4
+#define X3 8
+#define X4 16
+#define X5 32
+#define X6 64
+#define X7 128
+#define Y0 1
+#define Y1 2
+#define Y2 4
+#define Y3 8
+#define Y4 16
+#define Y5 32
+#define Y6 64
+#define Y7 128
+
+struct dim2d
+{
+ int w;
+ int h;
+};
+
+
uint32_t igt_amd_create_bo(int fd, uint64_t size)
{
union drm_amdgpu_gem_create create;
@@ -55,10 +79,13 @@ void *igt_amd_mmap_bo(int fd, uint32_t handle, uint64_t size, int prot)
return ptr == MAP_FAILED ? NULL : ptr;
}
-unsigned int igt_amd_compute_offset(unsigned int* swizzle_pattern, unsigned int x, unsigned int y)
+unsigned int igt_amd_compute_offset(unsigned int* swizzle_pattern,
+ unsigned int x, unsigned int y)
{
- unsigned int offset = 0, index = 0, blk_size_table_index = 0, interleave = 0;
- unsigned int channel[16] = {0, 0, 1, 1, 2, 2, 2, 1, 2, 1, 2, 1, 2, 1, 2, 1};
+ unsigned int offset = 0, index = 0;
+ unsigned int blk_size_table_index = 0, interleave = 0;
+ unsigned int channel[16] =
+ {0, 0, 1, 1, 2, 2, 2, 1, 2, 1, 2, 1, 2, 1, 2, 1};
unsigned int i, v;
for (i = 0; i < 16; i++)
@@ -97,39 +124,63 @@ unsigned int igt_amd_compute_offset(unsigned int* swizzle_pattern, unsigned int
return offset;
}
-uint32_t igt_amd_fb_tiled_offset(unsigned int bpp, unsigned int x_input, unsigned int y_input,
- unsigned int width_input)
+unsigned int igt_amd_fb_get_blk_size_table_idx(unsigned int bpp)
{
- unsigned int blk_size_table_index, blk_size_log2, blk_size_log2_256B;
- unsigned int element_bytes, width_amp, height_amp, width, height, pitch;
- unsigned int pb, yb, xb, blk_idx, blk_offset, addr;
- unsigned int* swizzle_pattern;
- unsigned int block256_2d[][2] = {{16, 16}, {16, 8}, {8, 8}, {8, 4}, {4, 4}};
- unsigned int sw_64k_s[][16]=
- {
- {X0, X1, X2, X3, Y0, Y1, Y2, Y3, Y4, X4, Y5, X5, Y6, X6, Y7, X7},
- {0, X0, X1, X2, Y0, Y1, Y2, X3, Y3, X4, Y4, X5, Y5, X6, Y6, X7},
- {0, 0, X0, X1, Y0, Y1, Y2, X2, Y3, X3, Y4, X4, Y5, X5, Y6, X6},
- {0, 0, 0, X0, Y0, Y1, X1, X2, Y2, X3, Y3, X4, Y4, X5, Y5, X6},
- {0, 0, 0, 0, Y0, Y1, X0, X1, Y2, X2, Y3, X3, Y4, X4, Y5, X5},
- };
+ unsigned int element_bytes;
+ unsigned int blk_size_table_index = 0;
element_bytes = bpp >> 3;
- blk_size_table_index = 0;
while (element_bytes > 1) {
blk_size_table_index++;
element_bytes = (element_bytes + 1) >> 1;
}
+ return blk_size_table_index;
+}
+
+void igt_amd_fb_calculate_tile_dimension(unsigned int bpp,
+ unsigned int *width, unsigned int *height)
+{
+ unsigned int blk_size_table_index;
+ unsigned int blk_size_log2, blk_size_log2_256B;
+ unsigned int width_amp, height_amp;
+
+ // swizzle 64kb tile block
+ unsigned int block256_2d[][2] = {{16, 16}, {16, 8}, {8, 8}, {8, 4}, {4, 4}};
blk_size_log2 = 16;
+
+ blk_size_table_index = igt_amd_fb_get_blk_size_table_idx(bpp);
+
blk_size_log2_256B = blk_size_log2 - 8;
width_amp = blk_size_log2_256B / 2;
height_amp = blk_size_log2_256B - width_amp;
- width = (block256_2d[blk_size_table_index][0] << width_amp);
- height = (block256_2d[blk_size_table_index][1] << height_amp);
+ *width = (block256_2d[blk_size_table_index][0] << width_amp);
+ *height = (block256_2d[blk_size_table_index][1] << height_amp);
+}
+
+uint32_t igt_amd_fb_tiled_offset(unsigned int bpp, unsigned int x_input,
+ unsigned int y_input, unsigned int width_input)
+{
+ unsigned int width, height, pitch;
+ unsigned int pb, yb, xb, blk_idx, blk_offset, addr;
+ unsigned int blk_size_table_index, blk_size_log2;
+ unsigned int* swizzle_pattern;
+
+ // swizzle 64kb tile block
+ unsigned int sw_64k_s[][16]=
+ {
+ {X0, X1, X2, X3, Y0, Y1, Y2, Y3, Y4, X4, Y5, X5, Y6, X6, Y7, X7},
+ {0, X0, X1, X2, Y0, Y1, Y2, X3, Y3, X4, Y4, X5, Y5, X6, Y6, X7},
+ {0, 0, X0, X1, Y0, Y1, Y2, X2, Y3, X3, Y4, X4, Y5, X5, Y6, X6},
+ {0, 0, 0, X0, Y0, Y1, X1, X2, Y2, X3, Y3, X4, Y4, X5, Y5, X6},
+ {0, 0, 0, 0, Y0, Y1, X0, X1, Y2, X2, Y3, X3, Y4, X4, Y5, X5},
+ };
+ igt_amd_fb_calculate_tile_dimension(bpp, &width, &height);
+ blk_size_table_index = igt_amd_fb_get_blk_size_table_idx(bpp);
+ blk_size_log2 = 16;
pitch = (width_input + (width - 1)) & (~(width - 1));
@@ -183,6 +234,16 @@ void igt_amd_fb_convert_plane_to_tiled(struct igt_fb *dst, void *dst_buf,
unsigned int plane;
for (plane = 0; plane < src->num_planes; plane++) {
+ igt_require(AMD_FMT_MOD_GET(TILE, dst->modifier) ==
+ AMD_FMT_MOD_TILE_GFX9_64K_S);
igt_amd_fb_to_tiled(dst, dst_buf, src, src_buf, plane);
}
}
+
+bool igt_amd_is_tiled(uint64_t modifier)
+{
+ if (IS_AMD_FMT_MOD(modifier) && AMD_FMT_MOD_GET(TILE, modifier))
+ return true;
+ else
+ return false;
+}
diff --git a/lib/igt_amd.h b/lib/igt_amd.h
index d5b421b4..6656d901 100644
--- a/lib/igt_amd.h
+++ b/lib/igt_amd.h
@@ -28,35 +28,16 @@
uint32_t igt_amd_create_bo(int fd, uint64_t size);
void *igt_amd_mmap_bo(int fd, uint32_t handle, uint64_t size, int prot);
-unsigned int igt_amd_compute_offset(unsigned int* SwizzlePattern, unsigned int x, unsigned int y);
-uint32_t igt_amd_fb_tiled_offset(unsigned int bpp, unsigned int x_input, unsigned int y_input,
- unsigned int width_input);
+unsigned int igt_amd_compute_offset(unsigned int* swizzle_pattern,
+ unsigned int x, unsigned int y);
+unsigned int igt_amd_fb_get_blk_size_table_idx(unsigned int bpp);
+void igt_amd_fb_calculate_tile_dimension(unsigned int bpp,
+ unsigned int *width, unsigned int *height);
+uint32_t igt_amd_fb_tiled_offset(unsigned int bpp, unsigned int x_input,
+ unsigned int y_input, unsigned int width_input);
void igt_amd_fb_to_tiled(struct igt_fb *dst, void *dst_buf, struct igt_fb *src,
void *src_buf, unsigned int plane);
void igt_amd_fb_convert_plane_to_tiled(struct igt_fb *dst, void *dst_buf,
struct igt_fb *src, void *src_buf);
-
-#define X0 1
-#define X1 2
-#define X2 4
-#define X3 8
-#define X4 16
-#define X5 32
-#define X6 64
-#define X7 128
-#define Y0 1
-#define Y1 2
-#define Y2 4
-#define Y3 8
-#define Y4 16
-#define Y5 32
-#define Y6 64
-#define Y7 128
-
-struct Dim2d
-{
- int w;
- int h;
-};
-
+bool igt_amd_is_tiled(uint64_t modifier);
#endif /* IGT_AMD_H */
diff --git a/lib/igt_fb.c b/lib/igt_fb.c
index 6eebe048..e41fbcff 100644
--- a/lib/igt_fb.c
+++ b/lib/igt_fb.c
@@ -675,7 +675,13 @@ static uint32_t calc_plane_stride(struct igt_fb *fb, int plane)
/*
* For amdgpu device with tiling mode
*/
- return ALIGN(min_stride, 512);
+ unsigned int tile_width, tile_height;
+
+ igt_amd_fb_calculate_tile_dimension(fb->plane_bpp[plane],
+ &tile_width, &tile_height);
+ tile_width *= (fb->plane_bpp[plane] / 8);
+
+ return ALIGN(min_stride, tile_width);
} else if (is_gen12_ccs_cc_plane(fb, plane)) {
/* clear color always fixed to 64 bytes */
return 64;
@@ -720,8 +726,14 @@ static uint64_t calc_plane_size(struct igt_fb *fb, int plane)
/*
* For amdgpu device with tiling mode
*/
+ unsigned int tile_width, tile_height;
+
+ igt_amd_fb_calculate_tile_dimension(fb->plane_bpp[plane],
+ &tile_width, &tile_height);
+ tile_height *= (fb->plane_bpp[plane] / 8);
+
return (uint64_t) fb->strides[plane] *
- ALIGN(fb->plane_height[plane], 512);
+ ALIGN(fb->plane_height[plane], tile_height);
} else if (is_gen12_ccs_plane(fb, plane)) {
/* The AUX CCS surface must be page aligned */
return (uint64_t)fb->strides[plane] *
@@ -2364,7 +2376,7 @@ static void free_linear_mapping(struct fb_blit_upload *blit)
vc4_fb_convert_plane_to_tiled(fb, map, &linear->fb, &linear->map);
munmap(map, fb->size);
- } else if (is_amdgpu_device(fd) && fb->modifier != 0) {
+ } else if (igt_amd_is_tiled(fb->modifier)) {
void *map = igt_amd_mmap_bo(fd, fb->gem_handle, fb->size, PROT_WRITE);
igt_amd_fb_convert_plane_to_tiled(fb, map, &linear->fb, linear->map);
@@ -2437,7 +2449,7 @@ static void setup_linear_mapping(struct fb_blit_upload *blit)
vc4_fb_convert_plane_from_tiled(&linear->fb, &linear->map, fb, map);
munmap(map, fb->size);
- } else if (is_amdgpu_device(fd) && fb->modifier != 0) {
+ } else if (igt_amd_is_tiled(fb->modifier)) {
linear->map = igt_amd_mmap_bo(fd, linear->fb.gem_handle,
linear->fb.size,
PROT_READ | PROT_WRITE);
@@ -3647,7 +3659,8 @@ cairo_surface_t *igt_get_cairo_surface(int fd, struct igt_fb *fb)
if (use_convert(fb))
create_cairo_surface__convert(fd, fb);
else if (use_blitter(fb) || use_enginecopy(fb) ||
- igt_vc4_is_tiled(fb->modifier) || (is_amdgpu_device(fd) && fb->modifier != 0))
+ igt_vc4_is_tiled(fb->modifier) ||
+ igt_amd_is_tiled(fb->modifier))
create_cairo_surface__gpu(fd, fb);
else
create_cairo_surface__gtt(fd, fb);
diff --git a/tests/kms_rotation_crc.c b/tests/kms_rotation_crc.c
index 31c7499c..5657235c 100644
--- a/tests/kms_rotation_crc.c
+++ b/tests/kms_rotation_crc.c
@@ -63,6 +63,7 @@ typedef struct {
struct igt_fb fb;
struct igt_fb fb_reference;
struct igt_fb fb_flip;
+ struct igt_fb fb_crtc;
igt_crc_t ref_crc;
igt_crc_t flip_crc;
igt_pipe_crc_t *pipe_crc;
@@ -185,7 +186,7 @@ static void cleanup_crtc(data_t *data)
}
static void prepare_crtc(data_t *data, igt_output_t *output, enum pipe pipe,
- igt_plane_t *plane, bool start_crc)
+ igt_plane_t *plane)
{
igt_display_t *display = &data->display;
@@ -198,22 +199,18 @@ static void prepare_crtc(data_t *data, igt_output_t *output, enum pipe pipe,
igt_pipe_crc_free(data->pipe_crc);
if (is_amdgpu_device(data->gfx_fd)) {
- igt_fb_t fb_temp;
drmModeModeInfo *mode = igt_output_get_mode(output);
igt_create_fb(data->gfx_fd, mode->hdisplay, mode->vdisplay,
- DRM_FORMAT_XRGB8888, 0, &fb_temp);
- igt_plane_set_fb(plane, &fb_temp);
- paint_squares(data, IGT_ROTATION_0, &fb_temp, 1.0);
+ data->override_fmt, LOCAL_DRM_FORMAT_MOD_NONE, &data->fb_crtc);
+ igt_plane_set_fb(plane, &data->fb_crtc);
+ paint_squares(data, IGT_ROTATION_0, &data->fb_crtc, 1.0);
if (plane->type != DRM_PLANE_TYPE_CURSOR)
igt_plane_set_position(plane, data->pos_x, data->pos_y);
}
igt_display_commit2(display, COMMIT_ATOMIC);
data->pipe_crc = igt_pipe_crc_new(data->gfx_fd, pipe, INTEL_PIPE_CRC_SOURCE_AUTO);
-
- if (start_crc)
- igt_pipe_crc_start(data->pipe_crc);
}
static void prepare_fbs(data_t *data, igt_output_t *output,
@@ -225,8 +222,6 @@ static void prepare_fbs(data_t *data, igt_output_t *output,
uint64_t tiling = data->override_tiling ?: LOCAL_DRM_FORMAT_MOD_NONE;
uint32_t pixel_format = data->override_fmt ?: DRM_FORMAT_XRGB8888;
const float flip_opacity = 0.75;
- bool amd_gpu = is_amdgpu_device(data->gfx_fd);
- bool intel_gpu = is_i915_device(data->gfx_fd);
remove_fbs(data);
@@ -234,21 +229,16 @@ static void prepare_fbs(data_t *data, igt_output_t *output,
mode = igt_output_get_mode(output);
if (plane->type != DRM_PLANE_TYPE_CURSOR) {
- if (amd_gpu) {
+ if (data->use_native_resolution) {
w = mode->hdisplay;
h = mode->vdisplay;
- } else if (intel_gpu) {
- if (data->use_native_resolution) {
- w = mode->hdisplay;
- h = mode->vdisplay;
- } else {
- w = min(TEST_MAX_WIDTH, mode->hdisplay);
- h = min(TEST_MAX_HEIGHT, mode->vdisplay);
- }
-
- min_w = 256;
- min_h = 256;
+ } else {
+ w = min(TEST_MAX_WIDTH, mode->hdisplay);
+ h = min(TEST_MAX_HEIGHT, mode->vdisplay);
}
+
+ min_w = 256;
+ min_h = 256;
} else {
pixel_format = data->override_fmt ?: DRM_FORMAT_ARGB8888;
@@ -280,8 +270,7 @@ static void prepare_fbs(data_t *data, igt_output_t *output,
* frame can fit in
*/
if (data->rotation & (IGT_ROTATION_90 | IGT_ROTATION_270)) {
- if (intel_gpu)
- tiling = data->override_tiling ?: LOCAL_I915_FORMAT_MOD_Y_TILED;
+ tiling = data->override_tiling ?: LOCAL_I915_FORMAT_MOD_Y_TILED;
igt_swap(w, h);
}
@@ -292,63 +281,48 @@ static void prepare_fbs(data_t *data, igt_output_t *output,
*/
igt_require(igt_display_has_format_mod(display, pixel_format, tiling));
- if (intel_gpu) {
- if (!data->crc_rect[rect].valid) {
- /*
- * Create a reference software rotated flip framebuffer.
- */
- igt_create_fb(data->gfx_fd, ref_w, ref_h, pixel_format, tiling,
- &data->fb_flip);
- paint_squares(data, data->rotation, &data->fb_flip,
- flip_opacity);
- igt_plane_set_fb(plane, &data->fb_flip);
- if (plane->type != DRM_PLANE_TYPE_CURSOR)
- igt_plane_set_position(plane, data->pos_x, data->pos_y);
- igt_display_commit2(display, COMMIT_ATOMIC);
-
- igt_pipe_crc_get_current(display->drm_fd, data->pipe_crc, &data->crc_rect[rect].flip_crc);
- igt_remove_fb(data->gfx_fd, &data->fb_flip);
-
- /*
- * Create a reference CRC for a software-rotated fb.
- */
- igt_create_fb(data->gfx_fd, ref_w, ref_h, pixel_format,
- data->override_tiling ?: LOCAL_DRM_FORMAT_MOD_NONE, &data->fb_reference);
- paint_squares(data, data->rotation, &data->fb_reference, 1.0);
-
- igt_plane_set_fb(plane, &data->fb_reference);
- if (plane->type != DRM_PLANE_TYPE_CURSOR)
- igt_plane_set_position(plane, data->pos_x, data->pos_y);
- igt_display_commit2(display, COMMIT_ATOMIC);
-
- igt_pipe_crc_get_current(display->drm_fd, data->pipe_crc, &data->crc_rect[rect].ref_crc);
- data->crc_rect[rect].valid = true;
- }
+ if (!data->crc_rect[rect].valid) {
+ /*
+ * Create a reference software rotated flip framebuffer.
+ */
+ igt_create_fb(data->gfx_fd, ref_w, ref_h, pixel_format, LOCAL_DRM_FORMAT_MOD_NONE,
+ &data->fb_flip);
+ paint_squares(data, data->rotation, &data->fb_flip,
+ flip_opacity);
+ igt_plane_set_fb(plane, &data->fb_flip);
+ if (plane->type != DRM_PLANE_TYPE_CURSOR)
+ igt_plane_set_position(plane, data->pos_x, data->pos_y);
+ igt_display_commit2(display, COMMIT_ATOMIC);
+ igt_pipe_crc_collect_crc(data->pipe_crc, &data->crc_rect[rect].flip_crc);
+
+ if (is_amdgpu_device(data->gfx_fd))
+ igt_remove_fb(data->gfx_fd, &data->fb_crtc);
/*
- * Prepare the non-rotated flip fb.
- */
- igt_create_fb(data->gfx_fd, w, h, pixel_format, tiling,
- &data->fb_flip);
- paint_squares(data, IGT_ROTATION_0, &data->fb_flip,
- flip_opacity);
- } else if (amd_gpu) {
- tiling = 0x900;
+ * Create a reference CRC for a software-rotated fb.
+ */
igt_create_fb(data->gfx_fd, ref_w, ref_h, pixel_format,
- LOCAL_DRM_FORMAT_MOD_NONE, &data->fb_reference);
+ LOCAL_DRM_FORMAT_MOD_NONE, &data->fb_reference);
paint_squares(data, data->rotation, &data->fb_reference, 1.0);
igt_plane_set_fb(plane, &data->fb_reference);
if (plane->type != DRM_PLANE_TYPE_CURSOR)
igt_plane_set_position(plane, data->pos_x, data->pos_y);
-
- if (data->rotation & (IGT_ROTATION_90 | IGT_ROTATION_270))
- igt_plane_set_size(plane, ref_w, ref_h);
-
igt_display_commit2(display, COMMIT_ATOMIC);
- igt_pipe_crc_collect_crc(data->pipe_crc, &data->ref_crc);
+ igt_pipe_crc_collect_crc(data->pipe_crc, &data->crc_rect[rect].ref_crc);
+ data->crc_rect[rect].valid = true;
+
+ igt_remove_fb(data->gfx_fd, &data->fb_flip);
}
+ /*
+ * Prepare the non-rotated flip fb.
+ */
+ igt_create_fb(data->gfx_fd, w, h, pixel_format, tiling,
+ &data->fb_flip);
+ paint_squares(data, IGT_ROTATION_0, &data->fb_flip,
+ flip_opacity);
+
/*
* Prepare the plane with an non-rotated fb let the hw rotate it.
*/
@@ -386,37 +360,32 @@ static void test_single_case(data_t *data, enum pipe pipe,
/* Verify commit was ok. */
igt_assert_eq(ret, 0);
- if (is_i915_device(data->gfx_fd)) {
- /* Check CRC */
- igt_pipe_crc_get_current(display->drm_fd, data->pipe_crc, &crc_output);
- igt_assert_crc_equal(&data->crc_rect[rect].ref_crc, &crc_output);
+ /* Check CRC */
+ igt_pipe_crc_collect_crc(data->pipe_crc, &crc_output);
+ igt_assert_crc_equal(&data->crc_rect[rect].ref_crc, &crc_output);
- /*
- * If flips are requested flip to a different fb and
- * check CRC against that one as well.
- */
- if (data->fb_flip.fb_id) {
- igt_plane_set_fb(plane, &data->fb_flip);
- if (data->rotation == IGT_ROTATION_90 || data->rotation == IGT_ROTATION_270)
- igt_plane_set_size(plane, data->fb.height, data->fb.width);
-
- if (plane->type != DRM_PLANE_TYPE_PRIMARY) {
- igt_display_commit_atomic(display, DRM_MODE_PAGE_FLIP_EVENT | DRM_MODE_ATOMIC_NONBLOCK, NULL);
- } else {
- ret = drmModePageFlip(data->gfx_fd,
- output->config.crtc->crtc_id,
- data->fb_flip.fb_id,
- DRM_MODE_PAGE_FLIP_EVENT,
- NULL);
- igt_assert_eq(ret, 0);
- }
- kmstest_wait_for_pageflip(data->gfx_fd);
- igt_pipe_crc_get_current(display->drm_fd, data->pipe_crc, &crc_output);
- igt_assert_crc_equal(&data->crc_rect[rect].flip_crc, &crc_output);
+ /*
+ * If flips are requested flip to a different fb and
+ * check CRC against that one as well.
+ */
+ if (data->fb_flip.fb_id) {
+ igt_plane_set_fb(plane, &data->fb_flip);
+ if (data->rotation == IGT_ROTATION_90 || data->rotation == IGT_ROTATION_270)
+ igt_plane_set_size(plane, data->fb.height, data->fb.width);
+
+ if (plane->type != DRM_PLANE_TYPE_PRIMARY) {
+ igt_display_commit_atomic(display, DRM_MODE_PAGE_FLIP_EVENT | DRM_MODE_ATOMIC_NONBLOCK, NULL);
+ } else {
+ ret = drmModePageFlip(data->gfx_fd,
+ output->config.crtc->crtc_id,
+ data->fb_flip.fb_id,
+ DRM_MODE_PAGE_FLIP_EVENT,
+ NULL);
+ igt_assert_eq(ret, 0);
}
- } else if (is_amdgpu_device(data->gfx_fd)) {
+ kmstest_wait_for_pageflip(data->gfx_fd);
igt_pipe_crc_collect_crc(data->pipe_crc, &crc_output);
- igt_assert_crc_equal(&data->ref_crc, &crc_output);
+ igt_assert_crc_equal(&data->crc_rect[rect].flip_crc, &crc_output);
}
}
@@ -448,6 +417,9 @@ static void test_plane_rotation(data_t *data, int plane_type, bool test_bad_form
igt_output_t *output;
enum pipe pipe;
+ if (is_amdgpu_device(data->gfx_fd))
+ igt_require(plane_type != DRM_PLANE_TYPE_OVERLAY);
+
if (plane_type == DRM_PLANE_TYPE_CURSOR)
igt_require(display->has_cursor_plane);
@@ -457,78 +429,58 @@ static void test_plane_rotation(data_t *data, int plane_type, bool test_bad_form
igt_plane_t *plane;
int i, j, c;
- if (is_i915_device(data->gfx_fd)) {
- for (c = 0; c < num_rectangle_types; c++)
- data->crc_rect[c].valid = false;
+ for (c = 0; c < num_rectangle_types; c++)
+ data->crc_rect[c].valid = false;
- if (IS_CHERRYVIEW(data->devid) && pipe != PIPE_B)
- continue;
+ if (is_i915_device(data->gfx_fd) && IS_CHERRYVIEW(data->devid) && pipe != PIPE_B)
+ continue;
- igt_output_set_pipe(output, pipe);
+ igt_output_set_pipe(output, pipe);
- plane = igt_output_get_plane_type(output, plane_type);
- igt_require(igt_plane_has_prop(plane, IGT_PLANE_ROTATION));
+ plane = igt_output_get_plane_type(output, plane_type);
+ igt_require(igt_plane_has_prop(plane, IGT_PLANE_ROTATION));
- prepare_crtc(data, output, pipe, plane, true);
+ prepare_crtc(data, output, pipe, plane);
- for (i = 0; i < num_rectangle_types; i++) {
- /* Unsupported on i915 */
- if (plane_type == DRM_PLANE_TYPE_CURSOR &&
- i != square)
- continue;
+ for (i = 0; i < num_rectangle_types; i++) {
+ /* Unsupported on i915 */
+ if (plane_type == DRM_PLANE_TYPE_CURSOR &&
+ i != square)
+ continue;
- /* Only support partial covering primary plane on gen9+ */
- if (plane_type == DRM_PLANE_TYPE_PRIMARY &&
- intel_gen(intel_get_drm_devid(data->gfx_fd)) < 9) {
- if (i != rectangle)
- continue;
- else
- data->use_native_resolution = true;
- } else {
- data->use_native_resolution = false;
- }
+ /* Only support partial covering primary plane on gen9+ */
+ if (is_amdgpu_device(data->gfx_fd) ||
+ (plane_type == DRM_PLANE_TYPE_PRIMARY &&
+ intel_gen(intel_get_drm_devid(data->gfx_fd)) < 9)) {
+ if (i != rectangle)
+ continue;
+ else
+ data->use_native_resolution = true;
+ } else {
+ data->use_native_resolution = false;
+ }
- if (!data->override_fmt) {
- struct igt_vec tested_formats;
+ if (!data->override_fmt) {
+ struct igt_vec tested_formats;
- igt_vec_init(&tested_formats, sizeof(uint32_t));
+ igt_vec_init(&tested_formats, sizeof(uint32_t));
- for (j = 0; j < plane->drm_plane->count_formats; j++) {
- uint32_t format = plane->drm_plane->formats[j];
+ for (j = 0; j < plane->drm_plane->count_formats; j++) {
+ uint32_t format = plane->drm_plane->formats[j];
- if (!test_format(data, &tested_formats, format))
- continue;
-
- test_single_case(data, pipe, output, plane, i,
- format, test_bad_format);
- }
+ if (!test_format(data, &tested_formats, format))
+ continue;
- igt_vec_fini(&tested_formats);
- } else {
test_single_case(data, pipe, output, plane, i,
- data->override_fmt, test_bad_format);
+ format, test_bad_format);
}
- }
- } else if (is_amdgpu_device(data->gfx_fd)) {
- uint32_t format = DRM_FORMAT_XRGB8888;
-
- igt_output_set_pipe(output, pipe);
-
- plane = igt_output_get_plane_type(output, DRM_PLANE_TYPE_PRIMARY);
- igt_require(igt_plane_has_prop(plane, IGT_PLANE_ROTATION));
- prepare_crtc(data, output, pipe, plane, false);
-
- if (plane_type != DRM_PLANE_TYPE_PRIMARY) {
- plane = igt_output_get_plane_type(output, plane_type);
- igt_require(igt_plane_has_prop(plane, IGT_PLANE_ROTATION));
+ igt_vec_fini(&tested_formats);
+ } else {
+ test_single_case(data, pipe, output, plane, i,
+ data->override_fmt, test_bad_format);
}
-
- test_single_case(data, pipe, output, plane,
- rectangle, format, test_bad_format);
}
-
- igt_pipe_crc_stop(data->pipe_crc);
}
}
@@ -771,7 +723,7 @@ static void test_plane_rotation_exhaust_fences(data_t *data,
igt_require(igt_plane_has_prop(plane, IGT_PLANE_ROTATION));
- prepare_crtc(data, output, pipe, plane, false);
+ prepare_crtc(data, output, pipe, plane);
mode = igt_output_get_mode(output);
w = mode->hdisplay;
@@ -928,6 +880,14 @@ igt_main_args("", long_opts, help_str, opt_handler, &data)
igt_require(!(subtest->rot &
(IGT_ROTATION_90 | IGT_ROTATION_270)) ||
gen >= 9);
+ } else if (is_amdgpu_device(data.gfx_fd)) {
+ data.override_fmt = DRM_FORMAT_XRGB8888;
+ if (subtest->rot & (IGT_ROTATION_90 | IGT_ROTATION_270))
+ data.override_tiling = 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);
+ else
+ data.override_tiling = LOCAL_DRM_FORMAT_MOD_NONE;
}
data.rotation = subtest->rot;
test_plane_rotation(&data, subtest->plane, false);
--
2.25.1
_______________________________________________
igt-dev mailing list
igt-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/igt-dev
^ permalink raw reply related [flat|nested] 3+ messages in thread
* [igt-dev] ✗ Fi.CI.BUILD: failure for kms_rotation_crc:Add HW rotation test case for amdgpu (rev3)
2021-01-21 23:03 [igt-dev] [PATCH v3] kms_rotation_crc:Add HW rotation test case for amdgpu Sung Joon Kim
@ 2021-01-21 23:08 ` Patchwork
2021-01-22 7:26 ` [igt-dev] [PATCH v3] kms_rotation_crc:Add HW rotation test case for amdgpu Petri Latvala
1 sibling, 0 replies; 3+ messages in thread
From: Patchwork @ 2021-01-21 23:08 UTC (permalink / raw
To: Kim, Sung joon; +Cc: igt-dev
== Series Details ==
Series: kms_rotation_crc:Add HW rotation test case for amdgpu (rev3)
URL : https://patchwork.freedesktop.org/series/85018/
State : failure
== Summary ==
Applying: kms_rotation_crc:Add HW rotation test case for amdgpu
Using index info to reconstruct a base tree...
M lib/igt_amd.c
M lib/igt_amd.h
M lib/igt_fb.c
M tests/kms_rotation_crc.c
Falling back to patching base and 3-way merge...
Auto-merging tests/kms_rotation_crc.c
CONFLICT (content): Merge conflict in tests/kms_rotation_crc.c
Auto-merging lib/igt_fb.c
CONFLICT (content): Merge conflict in lib/igt_fb.c
Auto-merging lib/igt_amd.h
CONFLICT (content): Merge conflict in lib/igt_amd.h
Auto-merging lib/igt_amd.c
CONFLICT (content): Merge conflict in lib/igt_amd.c
Patch failed at 0001 kms_rotation_crc:Add HW rotation test case for amdgpu
When you have resolved this problem, run "git am --continue".
If you prefer to skip this patch, run "git am --skip" instead.
To restore the original branch and stop patching, run "git am --abort".
_______________________________________________
igt-dev mailing list
igt-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/igt-dev
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [igt-dev] [PATCH v3] kms_rotation_crc:Add HW rotation test case for amdgpu
2021-01-21 23:03 [igt-dev] [PATCH v3] kms_rotation_crc:Add HW rotation test case for amdgpu Sung Joon Kim
2021-01-21 23:08 ` [igt-dev] ✗ Fi.CI.BUILD: failure for kms_rotation_crc:Add HW rotation test case for amdgpu (rev3) Patchwork
@ 2021-01-22 7:26 ` Petri Latvala
1 sibling, 0 replies; 3+ messages in thread
From: Petri Latvala @ 2021-01-22 7:26 UTC (permalink / raw
To: Sung Joon Kim; +Cc: igt-dev
On Thu, Jan 21, 2021 at 06:03:04PM -0500, Sung Joon Kim wrote:
> Added Hw rotation case specifically for amdgpu. Currently, kms_rotation_crc tests intel gpus. Added conditions to bypass all the requirements needed for intel when testing amdgpu.
>
> v3: code cleanup. Freed unused fb. Merged all files into one patch due to dependency. Added tiling methods for amdgpu. Updated drm header for amdgpu tiling modifiers.
> Signed-off-by: Sung Joon Kim <sungkim@amd.com>
> ---
> include/drm-uapi/drm_fourcc.h | 134 +++++++++++++++++
> lib/igt_amd.c | 103 ++++++++++---
> lib/igt_amd.h | 35 +----
> lib/igt_fb.c | 23 ++-
> tests/kms_rotation_crc.c | 270 +++++++++++++++-------------------
> 5 files changed, 357 insertions(+), 208 deletions(-)
>
> diff --git a/include/drm-uapi/drm_fourcc.h b/include/drm-uapi/drm_fourcc.h
> index 82f32780..a3286346 100644
> --- a/include/drm-uapi/drm_fourcc.h
> +++ b/include/drm-uapi/drm_fourcc.h
Please update drm_fourcc.h with a direct copy from the kernel header
and state in the commit message which kernel commit it's from.
(I didn't actually check if it is a copy here, I'm only guessing based
on the kernel commit missing from this commit message.)
--
Petri Latvala
> @@ -1056,6 +1056,140 @@ drm_fourcc_canonicalize_nvidia_format_mod(__u64 modifier)
> */
> #define AMLOGIC_FBC_OPTION_MEM_SAVING (1ULL << 0)
>
> +/*
> + * AMD modifiers
> + *
> + * Memory layout:
> + *
> + * without DCC:
> + * - main surface
> + *
> + * with DCC & without DCC_RETILE:
> + * - main surface in plane 0
> + * - DCC surface in plane 1 (RB-aligned, pipe-aligned if DCC_PIPE_ALIGN is set)
> + *
> + * with DCC & DCC_RETILE:
> + * - main surface in plane 0
> + * - displayable DCC surface in plane 1 (not RB-aligned & not pipe-aligned)
> + * - pipe-aligned DCC surface in plane 2 (RB-aligned & pipe-aligned)
> + *
> + * For multi-plane formats the above surfaces get merged into one plane for
> + * each format plane, based on the required alignment only.
> + *
> + * Bits Parameter Notes
> + * ----- ------------------------ ---------------------------------------------
> + *
> + * 7:0 TILE_VERSION Values are AMD_FMT_MOD_TILE_VER_*
> + * 12:8 TILE Values are AMD_FMT_MOD_TILE_<version>_*
> + * 13 DCC
> + * 14 DCC_RETILE
> + * 15 DCC_PIPE_ALIGN
> + * 16 DCC_INDEPENDENT_64B
> + * 17 DCC_INDEPENDENT_128B
> + * 19:18 DCC_MAX_COMPRESSED_BLOCK Values are AMD_FMT_MOD_DCC_BLOCK_*
> + * 20 DCC_CONSTANT_ENCODE
> + * 23:21 PIPE_XOR_BITS Only for some chips
> + * 26:24 BANK_XOR_BITS Only for some chips
> + * 29:27 PACKERS Only for some chips
> + * 32:30 RB Only for some chips
> + * 35:33 PIPE Only for some chips
> + * 55:36 - Reserved for future use, must be zero
> + */
> +#define AMD_FMT_MOD fourcc_mod_code(AMD, 0)
> +
> +#define IS_AMD_FMT_MOD(val) (((val) >> 56) == DRM_FORMAT_MOD_VENDOR_AMD)
> +
> +/* Reserve 0 for GFX8 and older */
> +#define AMD_FMT_MOD_TILE_VER_GFX9 1
> +#define AMD_FMT_MOD_TILE_VER_GFX10 2
> +#define AMD_FMT_MOD_TILE_VER_GFX10_RBPLUS 3
> +
> +/*
> + * 64K_S is the same for GFX9/GFX10/GFX10_RBPLUS and hence has GFX9 as canonical
> + * version.
> + */
> +#define AMD_FMT_MOD_TILE_GFX9_64K_S 9
> +
> +/*
> + * 64K_D for non-32 bpp is the same for GFX9/GFX10/GFX10_RBPLUS and hence has
> + * GFX9 as canonical version.
> + */
> +#define AMD_FMT_MOD_TILE_GFX9_64K_D 10
> +#define AMD_FMT_MOD_TILE_GFX9_64K_S_X 25
> +#define AMD_FMT_MOD_TILE_GFX9_64K_D_X 26
> +#define AMD_FMT_MOD_TILE_GFX9_64K_R_X 27
> +
> +#define AMD_FMT_MOD_DCC_BLOCK_64B 0
> +#define AMD_FMT_MOD_DCC_BLOCK_128B 1
> +#define AMD_FMT_MOD_DCC_BLOCK_256B 2
> +
> +#define AMD_FMT_MOD_TILE_VERSION_SHIFT 0
> +#define AMD_FMT_MOD_TILE_VERSION_MASK 0xFF
> +#define AMD_FMT_MOD_TILE_SHIFT 8
> +#define AMD_FMT_MOD_TILE_MASK 0x1F
> +
> +/* Whether DCC compression is enabled. */
> +#define AMD_FMT_MOD_DCC_SHIFT 13
> +#define AMD_FMT_MOD_DCC_MASK 0x1
> +
> +/*
> + * Whether to include two DCC surfaces, one which is rb & pipe aligned, and
> + * one which is not-aligned.
> + */
> +#define AMD_FMT_MOD_DCC_RETILE_SHIFT 14
> +#define AMD_FMT_MOD_DCC_RETILE_MASK 0x1
> +
> +/* Only set if DCC_RETILE = false */
> +#define AMD_FMT_MOD_DCC_PIPE_ALIGN_SHIFT 15
> +#define AMD_FMT_MOD_DCC_PIPE_ALIGN_MASK 0x1
> +
> +#define AMD_FMT_MOD_DCC_INDEPENDENT_64B_SHIFT 16
> +#define AMD_FMT_MOD_DCC_INDEPENDENT_64B_MASK 0x1
> +#define AMD_FMT_MOD_DCC_INDEPENDENT_128B_SHIFT 17
> +#define AMD_FMT_MOD_DCC_INDEPENDENT_128B_MASK 0x1
> +#define AMD_FMT_MOD_DCC_MAX_COMPRESSED_BLOCK_SHIFT 18
> +#define AMD_FMT_MOD_DCC_MAX_COMPRESSED_BLOCK_MASK 0x3
> +
> +/*
> + * DCC supports embedding some clear colors directly in the DCC surface.
> + * However, on older GPUs the rendering HW ignores the embedded clear color
> + * and prefers the driver provided color. This necessitates doing a fastclear
> + * eliminate operation before a process transfers control.
> + *
> + * If this bit is set that means the fastclear eliminate is not needed for these
> + * embeddable colors.
> + */
> +#define AMD_FMT_MOD_DCC_CONSTANT_ENCODE_SHIFT 20
> +#define AMD_FMT_MOD_DCC_CONSTANT_ENCODE_MASK 0x1
> +
> +/*
> + * The below fields are for accounting for per GPU differences. These are only
> + * relevant for GFX9 and later and if the tile field is *_X/_T.
> + *
> + * PIPE_XOR_BITS = always needed
> + * BANK_XOR_BITS = only for TILE_VER_GFX9
> + * PACKERS = only for TILE_VER_GFX10_RBPLUS
> + * RB = only for TILE_VER_GFX9 & DCC
> + * PIPE = only for TILE_VER_GFX9 & DCC & (DCC_RETILE | DCC_PIPE_ALIGN)
> + */
> +#define AMD_FMT_MOD_PIPE_XOR_BITS_SHIFT 21
> +#define AMD_FMT_MOD_PIPE_XOR_BITS_MASK 0x7
> +#define AMD_FMT_MOD_BANK_XOR_BITS_SHIFT 24
> +#define AMD_FMT_MOD_BANK_XOR_BITS_MASK 0x7
> +#define AMD_FMT_MOD_PACKERS_SHIFT 27
> +#define AMD_FMT_MOD_PACKERS_MASK 0x7
> +#define AMD_FMT_MOD_RB_SHIFT 30
> +#define AMD_FMT_MOD_RB_MASK 0x7
> +#define AMD_FMT_MOD_PIPE_SHIFT 33
> +#define AMD_FMT_MOD_PIPE_MASK 0x7
> +
> +#define AMD_FMT_MOD_SET(field, value) \
> + ((uint64_t)(value) << AMD_FMT_MOD_##field##_SHIFT)
> +#define AMD_FMT_MOD_GET(field, value) \
> + (((value) >> AMD_FMT_MOD_##field##_SHIFT) & AMD_FMT_MOD_##field##_MASK)
> +#define AMD_FMT_MOD_CLEAR(field) \
> + (~((uint64_t)AMD_FMT_MOD_##field##_MASK << AMD_FMT_MOD_##field##_SHIFT))
> +
> #if defined(__cplusplus)
> }
> #endif
> diff --git a/lib/igt_amd.c b/lib/igt_amd.c
> index 737e5670..54082ba9 100644
> --- a/lib/igt_amd.c
> +++ b/lib/igt_amd.c
> @@ -24,6 +24,30 @@
> #include "igt.h"
> #include <amdgpu_drm.h>
>
> +#define X0 1
> +#define X1 2
> +#define X2 4
> +#define X3 8
> +#define X4 16
> +#define X5 32
> +#define X6 64
> +#define X7 128
> +#define Y0 1
> +#define Y1 2
> +#define Y2 4
> +#define Y3 8
> +#define Y4 16
> +#define Y5 32
> +#define Y6 64
> +#define Y7 128
> +
> +struct dim2d
> +{
> + int w;
> + int h;
> +};
> +
> +
> uint32_t igt_amd_create_bo(int fd, uint64_t size)
> {
> union drm_amdgpu_gem_create create;
> @@ -55,10 +79,13 @@ void *igt_amd_mmap_bo(int fd, uint32_t handle, uint64_t size, int prot)
> return ptr == MAP_FAILED ? NULL : ptr;
> }
>
> -unsigned int igt_amd_compute_offset(unsigned int* swizzle_pattern, unsigned int x, unsigned int y)
> +unsigned int igt_amd_compute_offset(unsigned int* swizzle_pattern,
> + unsigned int x, unsigned int y)
> {
> - unsigned int offset = 0, index = 0, blk_size_table_index = 0, interleave = 0;
> - unsigned int channel[16] = {0, 0, 1, 1, 2, 2, 2, 1, 2, 1, 2, 1, 2, 1, 2, 1};
> + unsigned int offset = 0, index = 0;
> + unsigned int blk_size_table_index = 0, interleave = 0;
> + unsigned int channel[16] =
> + {0, 0, 1, 1, 2, 2, 2, 1, 2, 1, 2, 1, 2, 1, 2, 1};
> unsigned int i, v;
>
> for (i = 0; i < 16; i++)
> @@ -97,39 +124,63 @@ unsigned int igt_amd_compute_offset(unsigned int* swizzle_pattern, unsigned int
> return offset;
> }
>
> -uint32_t igt_amd_fb_tiled_offset(unsigned int bpp, unsigned int x_input, unsigned int y_input,
> - unsigned int width_input)
> +unsigned int igt_amd_fb_get_blk_size_table_idx(unsigned int bpp)
> {
> - unsigned int blk_size_table_index, blk_size_log2, blk_size_log2_256B;
> - unsigned int element_bytes, width_amp, height_amp, width, height, pitch;
> - unsigned int pb, yb, xb, blk_idx, blk_offset, addr;
> - unsigned int* swizzle_pattern;
> - unsigned int block256_2d[][2] = {{16, 16}, {16, 8}, {8, 8}, {8, 4}, {4, 4}};
> - unsigned int sw_64k_s[][16]=
> - {
> - {X0, X1, X2, X3, Y0, Y1, Y2, Y3, Y4, X4, Y5, X5, Y6, X6, Y7, X7},
> - {0, X0, X1, X2, Y0, Y1, Y2, X3, Y3, X4, Y4, X5, Y5, X6, Y6, X7},
> - {0, 0, X0, X1, Y0, Y1, Y2, X2, Y3, X3, Y4, X4, Y5, X5, Y6, X6},
> - {0, 0, 0, X0, Y0, Y1, X1, X2, Y2, X3, Y3, X4, Y4, X5, Y5, X6},
> - {0, 0, 0, 0, Y0, Y1, X0, X1, Y2, X2, Y3, X3, Y4, X4, Y5, X5},
> - };
> + unsigned int element_bytes;
> + unsigned int blk_size_table_index = 0;
>
> element_bytes = bpp >> 3;
> - blk_size_table_index = 0;
>
> while (element_bytes > 1) {
> blk_size_table_index++;
> element_bytes = (element_bytes + 1) >> 1;
> }
>
> + return blk_size_table_index;
> +}
> +
> +void igt_amd_fb_calculate_tile_dimension(unsigned int bpp,
> + unsigned int *width, unsigned int *height)
> +{
> + unsigned int blk_size_table_index;
> + unsigned int blk_size_log2, blk_size_log2_256B;
> + unsigned int width_amp, height_amp;
> +
> + // swizzle 64kb tile block
> + unsigned int block256_2d[][2] = {{16, 16}, {16, 8}, {8, 8}, {8, 4}, {4, 4}};
> blk_size_log2 = 16;
> +
> + blk_size_table_index = igt_amd_fb_get_blk_size_table_idx(bpp);
> +
> blk_size_log2_256B = blk_size_log2 - 8;
>
> width_amp = blk_size_log2_256B / 2;
> height_amp = blk_size_log2_256B - width_amp;
>
> - width = (block256_2d[blk_size_table_index][0] << width_amp);
> - height = (block256_2d[blk_size_table_index][1] << height_amp);
> + *width = (block256_2d[blk_size_table_index][0] << width_amp);
> + *height = (block256_2d[blk_size_table_index][1] << height_amp);
> +}
> +
> +uint32_t igt_amd_fb_tiled_offset(unsigned int bpp, unsigned int x_input,
> + unsigned int y_input, unsigned int width_input)
> +{
> + unsigned int width, height, pitch;
> + unsigned int pb, yb, xb, blk_idx, blk_offset, addr;
> + unsigned int blk_size_table_index, blk_size_log2;
> + unsigned int* swizzle_pattern;
> +
> + // swizzle 64kb tile block
> + unsigned int sw_64k_s[][16]=
> + {
> + {X0, X1, X2, X3, Y0, Y1, Y2, Y3, Y4, X4, Y5, X5, Y6, X6, Y7, X7},
> + {0, X0, X1, X2, Y0, Y1, Y2, X3, Y3, X4, Y4, X5, Y5, X6, Y6, X7},
> + {0, 0, X0, X1, Y0, Y1, Y2, X2, Y3, X3, Y4, X4, Y5, X5, Y6, X6},
> + {0, 0, 0, X0, Y0, Y1, X1, X2, Y2, X3, Y3, X4, Y4, X5, Y5, X6},
> + {0, 0, 0, 0, Y0, Y1, X0, X1, Y2, X2, Y3, X3, Y4, X4, Y5, X5},
> + };
> + igt_amd_fb_calculate_tile_dimension(bpp, &width, &height);
> + blk_size_table_index = igt_amd_fb_get_blk_size_table_idx(bpp);
> + blk_size_log2 = 16;
>
> pitch = (width_input + (width - 1)) & (~(width - 1));
>
> @@ -183,6 +234,16 @@ void igt_amd_fb_convert_plane_to_tiled(struct igt_fb *dst, void *dst_buf,
> unsigned int plane;
>
> for (plane = 0; plane < src->num_planes; plane++) {
> + igt_require(AMD_FMT_MOD_GET(TILE, dst->modifier) ==
> + AMD_FMT_MOD_TILE_GFX9_64K_S);
> igt_amd_fb_to_tiled(dst, dst_buf, src, src_buf, plane);
> }
> }
> +
> +bool igt_amd_is_tiled(uint64_t modifier)
> +{
> + if (IS_AMD_FMT_MOD(modifier) && AMD_FMT_MOD_GET(TILE, modifier))
> + return true;
> + else
> + return false;
> +}
> diff --git a/lib/igt_amd.h b/lib/igt_amd.h
> index d5b421b4..6656d901 100644
> --- a/lib/igt_amd.h
> +++ b/lib/igt_amd.h
> @@ -28,35 +28,16 @@
>
> uint32_t igt_amd_create_bo(int fd, uint64_t size);
> void *igt_amd_mmap_bo(int fd, uint32_t handle, uint64_t size, int prot);
> -unsigned int igt_amd_compute_offset(unsigned int* SwizzlePattern, unsigned int x, unsigned int y);
> -uint32_t igt_amd_fb_tiled_offset(unsigned int bpp, unsigned int x_input, unsigned int y_input,
> - unsigned int width_input);
> +unsigned int igt_amd_compute_offset(unsigned int* swizzle_pattern,
> + unsigned int x, unsigned int y);
> +unsigned int igt_amd_fb_get_blk_size_table_idx(unsigned int bpp);
> +void igt_amd_fb_calculate_tile_dimension(unsigned int bpp,
> + unsigned int *width, unsigned int *height);
> +uint32_t igt_amd_fb_tiled_offset(unsigned int bpp, unsigned int x_input,
> + unsigned int y_input, unsigned int width_input);
> void igt_amd_fb_to_tiled(struct igt_fb *dst, void *dst_buf, struct igt_fb *src,
> void *src_buf, unsigned int plane);
> void igt_amd_fb_convert_plane_to_tiled(struct igt_fb *dst, void *dst_buf,
> struct igt_fb *src, void *src_buf);
> -
> -#define X0 1
> -#define X1 2
> -#define X2 4
> -#define X3 8
> -#define X4 16
> -#define X5 32
> -#define X6 64
> -#define X7 128
> -#define Y0 1
> -#define Y1 2
> -#define Y2 4
> -#define Y3 8
> -#define Y4 16
> -#define Y5 32
> -#define Y6 64
> -#define Y7 128
> -
> -struct Dim2d
> -{
> - int w;
> - int h;
> -};
> -
> +bool igt_amd_is_tiled(uint64_t modifier);
> #endif /* IGT_AMD_H */
> diff --git a/lib/igt_fb.c b/lib/igt_fb.c
> index 6eebe048..e41fbcff 100644
> --- a/lib/igt_fb.c
> +++ b/lib/igt_fb.c
> @@ -675,7 +675,13 @@ static uint32_t calc_plane_stride(struct igt_fb *fb, int plane)
> /*
> * For amdgpu device with tiling mode
> */
> - return ALIGN(min_stride, 512);
> + unsigned int tile_width, tile_height;
> +
> + igt_amd_fb_calculate_tile_dimension(fb->plane_bpp[plane],
> + &tile_width, &tile_height);
> + tile_width *= (fb->plane_bpp[plane] / 8);
> +
> + return ALIGN(min_stride, tile_width);
> } else if (is_gen12_ccs_cc_plane(fb, plane)) {
> /* clear color always fixed to 64 bytes */
> return 64;
> @@ -720,8 +726,14 @@ static uint64_t calc_plane_size(struct igt_fb *fb, int plane)
> /*
> * For amdgpu device with tiling mode
> */
> + unsigned int tile_width, tile_height;
> +
> + igt_amd_fb_calculate_tile_dimension(fb->plane_bpp[plane],
> + &tile_width, &tile_height);
> + tile_height *= (fb->plane_bpp[plane] / 8);
> +
> return (uint64_t) fb->strides[plane] *
> - ALIGN(fb->plane_height[plane], 512);
> + ALIGN(fb->plane_height[plane], tile_height);
> } else if (is_gen12_ccs_plane(fb, plane)) {
> /* The AUX CCS surface must be page aligned */
> return (uint64_t)fb->strides[plane] *
> @@ -2364,7 +2376,7 @@ static void free_linear_mapping(struct fb_blit_upload *blit)
> vc4_fb_convert_plane_to_tiled(fb, map, &linear->fb, &linear->map);
>
> munmap(map, fb->size);
> - } else if (is_amdgpu_device(fd) && fb->modifier != 0) {
> + } else if (igt_amd_is_tiled(fb->modifier)) {
> void *map = igt_amd_mmap_bo(fd, fb->gem_handle, fb->size, PROT_WRITE);
>
> igt_amd_fb_convert_plane_to_tiled(fb, map, &linear->fb, linear->map);
> @@ -2437,7 +2449,7 @@ static void setup_linear_mapping(struct fb_blit_upload *blit)
> vc4_fb_convert_plane_from_tiled(&linear->fb, &linear->map, fb, map);
>
> munmap(map, fb->size);
> - } else if (is_amdgpu_device(fd) && fb->modifier != 0) {
> + } else if (igt_amd_is_tiled(fb->modifier)) {
> linear->map = igt_amd_mmap_bo(fd, linear->fb.gem_handle,
> linear->fb.size,
> PROT_READ | PROT_WRITE);
> @@ -3647,7 +3659,8 @@ cairo_surface_t *igt_get_cairo_surface(int fd, struct igt_fb *fb)
> if (use_convert(fb))
> create_cairo_surface__convert(fd, fb);
> else if (use_blitter(fb) || use_enginecopy(fb) ||
> - igt_vc4_is_tiled(fb->modifier) || (is_amdgpu_device(fd) && fb->modifier != 0))
> + igt_vc4_is_tiled(fb->modifier) ||
> + igt_amd_is_tiled(fb->modifier))
> create_cairo_surface__gpu(fd, fb);
> else
> create_cairo_surface__gtt(fd, fb);
> diff --git a/tests/kms_rotation_crc.c b/tests/kms_rotation_crc.c
> index 31c7499c..5657235c 100644
> --- a/tests/kms_rotation_crc.c
> +++ b/tests/kms_rotation_crc.c
> @@ -63,6 +63,7 @@ typedef struct {
> struct igt_fb fb;
> struct igt_fb fb_reference;
> struct igt_fb fb_flip;
> + struct igt_fb fb_crtc;
> igt_crc_t ref_crc;
> igt_crc_t flip_crc;
> igt_pipe_crc_t *pipe_crc;
> @@ -185,7 +186,7 @@ static void cleanup_crtc(data_t *data)
> }
>
> static void prepare_crtc(data_t *data, igt_output_t *output, enum pipe pipe,
> - igt_plane_t *plane, bool start_crc)
> + igt_plane_t *plane)
> {
> igt_display_t *display = &data->display;
>
> @@ -198,22 +199,18 @@ static void prepare_crtc(data_t *data, igt_output_t *output, enum pipe pipe,
> igt_pipe_crc_free(data->pipe_crc);
>
> if (is_amdgpu_device(data->gfx_fd)) {
> - igt_fb_t fb_temp;
> drmModeModeInfo *mode = igt_output_get_mode(output);
>
> igt_create_fb(data->gfx_fd, mode->hdisplay, mode->vdisplay,
> - DRM_FORMAT_XRGB8888, 0, &fb_temp);
> - igt_plane_set_fb(plane, &fb_temp);
> - paint_squares(data, IGT_ROTATION_0, &fb_temp, 1.0);
> + data->override_fmt, LOCAL_DRM_FORMAT_MOD_NONE, &data->fb_crtc);
> + igt_plane_set_fb(plane, &data->fb_crtc);
> + paint_squares(data, IGT_ROTATION_0, &data->fb_crtc, 1.0);
>
> if (plane->type != DRM_PLANE_TYPE_CURSOR)
> igt_plane_set_position(plane, data->pos_x, data->pos_y);
> }
> igt_display_commit2(display, COMMIT_ATOMIC);
> data->pipe_crc = igt_pipe_crc_new(data->gfx_fd, pipe, INTEL_PIPE_CRC_SOURCE_AUTO);
> -
> - if (start_crc)
> - igt_pipe_crc_start(data->pipe_crc);
> }
>
> static void prepare_fbs(data_t *data, igt_output_t *output,
> @@ -225,8 +222,6 @@ static void prepare_fbs(data_t *data, igt_output_t *output,
> uint64_t tiling = data->override_tiling ?: LOCAL_DRM_FORMAT_MOD_NONE;
> uint32_t pixel_format = data->override_fmt ?: DRM_FORMAT_XRGB8888;
> const float flip_opacity = 0.75;
> - bool amd_gpu = is_amdgpu_device(data->gfx_fd);
> - bool intel_gpu = is_i915_device(data->gfx_fd);
>
> remove_fbs(data);
>
> @@ -234,21 +229,16 @@ static void prepare_fbs(data_t *data, igt_output_t *output,
>
> mode = igt_output_get_mode(output);
> if (plane->type != DRM_PLANE_TYPE_CURSOR) {
> - if (amd_gpu) {
> + if (data->use_native_resolution) {
> w = mode->hdisplay;
> h = mode->vdisplay;
> - } else if (intel_gpu) {
> - if (data->use_native_resolution) {
> - w = mode->hdisplay;
> - h = mode->vdisplay;
> - } else {
> - w = min(TEST_MAX_WIDTH, mode->hdisplay);
> - h = min(TEST_MAX_HEIGHT, mode->vdisplay);
> - }
> -
> - min_w = 256;
> - min_h = 256;
> + } else {
> + w = min(TEST_MAX_WIDTH, mode->hdisplay);
> + h = min(TEST_MAX_HEIGHT, mode->vdisplay);
> }
> +
> + min_w = 256;
> + min_h = 256;
> } else {
> pixel_format = data->override_fmt ?: DRM_FORMAT_ARGB8888;
>
> @@ -280,8 +270,7 @@ static void prepare_fbs(data_t *data, igt_output_t *output,
> * frame can fit in
> */
> if (data->rotation & (IGT_ROTATION_90 | IGT_ROTATION_270)) {
> - if (intel_gpu)
> - tiling = data->override_tiling ?: LOCAL_I915_FORMAT_MOD_Y_TILED;
> + tiling = data->override_tiling ?: LOCAL_I915_FORMAT_MOD_Y_TILED;
>
> igt_swap(w, h);
> }
> @@ -292,63 +281,48 @@ static void prepare_fbs(data_t *data, igt_output_t *output,
> */
> igt_require(igt_display_has_format_mod(display, pixel_format, tiling));
>
> - if (intel_gpu) {
> - if (!data->crc_rect[rect].valid) {
> - /*
> - * Create a reference software rotated flip framebuffer.
> - */
> - igt_create_fb(data->gfx_fd, ref_w, ref_h, pixel_format, tiling,
> - &data->fb_flip);
> - paint_squares(data, data->rotation, &data->fb_flip,
> - flip_opacity);
> - igt_plane_set_fb(plane, &data->fb_flip);
> - if (plane->type != DRM_PLANE_TYPE_CURSOR)
> - igt_plane_set_position(plane, data->pos_x, data->pos_y);
> - igt_display_commit2(display, COMMIT_ATOMIC);
> -
> - igt_pipe_crc_get_current(display->drm_fd, data->pipe_crc, &data->crc_rect[rect].flip_crc);
> - igt_remove_fb(data->gfx_fd, &data->fb_flip);
> -
> - /*
> - * Create a reference CRC for a software-rotated fb.
> - */
> - igt_create_fb(data->gfx_fd, ref_w, ref_h, pixel_format,
> - data->override_tiling ?: LOCAL_DRM_FORMAT_MOD_NONE, &data->fb_reference);
> - paint_squares(data, data->rotation, &data->fb_reference, 1.0);
> -
> - igt_plane_set_fb(plane, &data->fb_reference);
> - if (plane->type != DRM_PLANE_TYPE_CURSOR)
> - igt_plane_set_position(plane, data->pos_x, data->pos_y);
> - igt_display_commit2(display, COMMIT_ATOMIC);
> -
> - igt_pipe_crc_get_current(display->drm_fd, data->pipe_crc, &data->crc_rect[rect].ref_crc);
> - data->crc_rect[rect].valid = true;
> - }
> + if (!data->crc_rect[rect].valid) {
> + /*
> + * Create a reference software rotated flip framebuffer.
> + */
> + igt_create_fb(data->gfx_fd, ref_w, ref_h, pixel_format, LOCAL_DRM_FORMAT_MOD_NONE,
> + &data->fb_flip);
> + paint_squares(data, data->rotation, &data->fb_flip,
> + flip_opacity);
> + igt_plane_set_fb(plane, &data->fb_flip);
> + if (plane->type != DRM_PLANE_TYPE_CURSOR)
> + igt_plane_set_position(plane, data->pos_x, data->pos_y);
> + igt_display_commit2(display, COMMIT_ATOMIC);
> + igt_pipe_crc_collect_crc(data->pipe_crc, &data->crc_rect[rect].flip_crc);
> +
> + if (is_amdgpu_device(data->gfx_fd))
> + igt_remove_fb(data->gfx_fd, &data->fb_crtc);
>
> /*
> - * Prepare the non-rotated flip fb.
> - */
> - igt_create_fb(data->gfx_fd, w, h, pixel_format, tiling,
> - &data->fb_flip);
> - paint_squares(data, IGT_ROTATION_0, &data->fb_flip,
> - flip_opacity);
> - } else if (amd_gpu) {
> - tiling = 0x900;
> + * Create a reference CRC for a software-rotated fb.
> + */
> igt_create_fb(data->gfx_fd, ref_w, ref_h, pixel_format,
> - LOCAL_DRM_FORMAT_MOD_NONE, &data->fb_reference);
> + LOCAL_DRM_FORMAT_MOD_NONE, &data->fb_reference);
> paint_squares(data, data->rotation, &data->fb_reference, 1.0);
>
> igt_plane_set_fb(plane, &data->fb_reference);
> if (plane->type != DRM_PLANE_TYPE_CURSOR)
> igt_plane_set_position(plane, data->pos_x, data->pos_y);
> -
> - if (data->rotation & (IGT_ROTATION_90 | IGT_ROTATION_270))
> - igt_plane_set_size(plane, ref_w, ref_h);
> -
> igt_display_commit2(display, COMMIT_ATOMIC);
> - igt_pipe_crc_collect_crc(data->pipe_crc, &data->ref_crc);
> + igt_pipe_crc_collect_crc(data->pipe_crc, &data->crc_rect[rect].ref_crc);
> + data->crc_rect[rect].valid = true;
> +
> + igt_remove_fb(data->gfx_fd, &data->fb_flip);
> }
>
> + /*
> + * Prepare the non-rotated flip fb.
> + */
> + igt_create_fb(data->gfx_fd, w, h, pixel_format, tiling,
> + &data->fb_flip);
> + paint_squares(data, IGT_ROTATION_0, &data->fb_flip,
> + flip_opacity);
> +
> /*
> * Prepare the plane with an non-rotated fb let the hw rotate it.
> */
> @@ -386,37 +360,32 @@ static void test_single_case(data_t *data, enum pipe pipe,
> /* Verify commit was ok. */
> igt_assert_eq(ret, 0);
>
> - if (is_i915_device(data->gfx_fd)) {
> - /* Check CRC */
> - igt_pipe_crc_get_current(display->drm_fd, data->pipe_crc, &crc_output);
> - igt_assert_crc_equal(&data->crc_rect[rect].ref_crc, &crc_output);
> + /* Check CRC */
> + igt_pipe_crc_collect_crc(data->pipe_crc, &crc_output);
> + igt_assert_crc_equal(&data->crc_rect[rect].ref_crc, &crc_output);
>
> - /*
> - * If flips are requested flip to a different fb and
> - * check CRC against that one as well.
> - */
> - if (data->fb_flip.fb_id) {
> - igt_plane_set_fb(plane, &data->fb_flip);
> - if (data->rotation == IGT_ROTATION_90 || data->rotation == IGT_ROTATION_270)
> - igt_plane_set_size(plane, data->fb.height, data->fb.width);
> -
> - if (plane->type != DRM_PLANE_TYPE_PRIMARY) {
> - igt_display_commit_atomic(display, DRM_MODE_PAGE_FLIP_EVENT | DRM_MODE_ATOMIC_NONBLOCK, NULL);
> - } else {
> - ret = drmModePageFlip(data->gfx_fd,
> - output->config.crtc->crtc_id,
> - data->fb_flip.fb_id,
> - DRM_MODE_PAGE_FLIP_EVENT,
> - NULL);
> - igt_assert_eq(ret, 0);
> - }
> - kmstest_wait_for_pageflip(data->gfx_fd);
> - igt_pipe_crc_get_current(display->drm_fd, data->pipe_crc, &crc_output);
> - igt_assert_crc_equal(&data->crc_rect[rect].flip_crc, &crc_output);
> + /*
> + * If flips are requested flip to a different fb and
> + * check CRC against that one as well.
> + */
> + if (data->fb_flip.fb_id) {
> + igt_plane_set_fb(plane, &data->fb_flip);
> + if (data->rotation == IGT_ROTATION_90 || data->rotation == IGT_ROTATION_270)
> + igt_plane_set_size(plane, data->fb.height, data->fb.width);
> +
> + if (plane->type != DRM_PLANE_TYPE_PRIMARY) {
> + igt_display_commit_atomic(display, DRM_MODE_PAGE_FLIP_EVENT | DRM_MODE_ATOMIC_NONBLOCK, NULL);
> + } else {
> + ret = drmModePageFlip(data->gfx_fd,
> + output->config.crtc->crtc_id,
> + data->fb_flip.fb_id,
> + DRM_MODE_PAGE_FLIP_EVENT,
> + NULL);
> + igt_assert_eq(ret, 0);
> }
> - } else if (is_amdgpu_device(data->gfx_fd)) {
> + kmstest_wait_for_pageflip(data->gfx_fd);
> igt_pipe_crc_collect_crc(data->pipe_crc, &crc_output);
> - igt_assert_crc_equal(&data->ref_crc, &crc_output);
> + igt_assert_crc_equal(&data->crc_rect[rect].flip_crc, &crc_output);
> }
> }
>
> @@ -448,6 +417,9 @@ static void test_plane_rotation(data_t *data, int plane_type, bool test_bad_form
> igt_output_t *output;
> enum pipe pipe;
>
> + if (is_amdgpu_device(data->gfx_fd))
> + igt_require(plane_type != DRM_PLANE_TYPE_OVERLAY);
> +
> if (plane_type == DRM_PLANE_TYPE_CURSOR)
> igt_require(display->has_cursor_plane);
>
> @@ -457,78 +429,58 @@ static void test_plane_rotation(data_t *data, int plane_type, bool test_bad_form
> igt_plane_t *plane;
> int i, j, c;
>
> - if (is_i915_device(data->gfx_fd)) {
> - for (c = 0; c < num_rectangle_types; c++)
> - data->crc_rect[c].valid = false;
> + for (c = 0; c < num_rectangle_types; c++)
> + data->crc_rect[c].valid = false;
>
> - if (IS_CHERRYVIEW(data->devid) && pipe != PIPE_B)
> - continue;
> + if (is_i915_device(data->gfx_fd) && IS_CHERRYVIEW(data->devid) && pipe != PIPE_B)
> + continue;
>
> - igt_output_set_pipe(output, pipe);
> + igt_output_set_pipe(output, pipe);
>
> - plane = igt_output_get_plane_type(output, plane_type);
> - igt_require(igt_plane_has_prop(plane, IGT_PLANE_ROTATION));
> + plane = igt_output_get_plane_type(output, plane_type);
> + igt_require(igt_plane_has_prop(plane, IGT_PLANE_ROTATION));
>
> - prepare_crtc(data, output, pipe, plane, true);
> + prepare_crtc(data, output, pipe, plane);
>
> - for (i = 0; i < num_rectangle_types; i++) {
> - /* Unsupported on i915 */
> - if (plane_type == DRM_PLANE_TYPE_CURSOR &&
> - i != square)
> - continue;
> + for (i = 0; i < num_rectangle_types; i++) {
> + /* Unsupported on i915 */
> + if (plane_type == DRM_PLANE_TYPE_CURSOR &&
> + i != square)
> + continue;
>
> - /* Only support partial covering primary plane on gen9+ */
> - if (plane_type == DRM_PLANE_TYPE_PRIMARY &&
> - intel_gen(intel_get_drm_devid(data->gfx_fd)) < 9) {
> - if (i != rectangle)
> - continue;
> - else
> - data->use_native_resolution = true;
> - } else {
> - data->use_native_resolution = false;
> - }
> + /* Only support partial covering primary plane on gen9+ */
> + if (is_amdgpu_device(data->gfx_fd) ||
> + (plane_type == DRM_PLANE_TYPE_PRIMARY &&
> + intel_gen(intel_get_drm_devid(data->gfx_fd)) < 9)) {
> + if (i != rectangle)
> + continue;
> + else
> + data->use_native_resolution = true;
> + } else {
> + data->use_native_resolution = false;
> + }
>
> - if (!data->override_fmt) {
> - struct igt_vec tested_formats;
> + if (!data->override_fmt) {
> + struct igt_vec tested_formats;
>
> - igt_vec_init(&tested_formats, sizeof(uint32_t));
> + igt_vec_init(&tested_formats, sizeof(uint32_t));
>
> - for (j = 0; j < plane->drm_plane->count_formats; j++) {
> - uint32_t format = plane->drm_plane->formats[j];
> + for (j = 0; j < plane->drm_plane->count_formats; j++) {
> + uint32_t format = plane->drm_plane->formats[j];
>
> - if (!test_format(data, &tested_formats, format))
> - continue;
> -
> - test_single_case(data, pipe, output, plane, i,
> - format, test_bad_format);
> - }
> + if (!test_format(data, &tested_formats, format))
> + continue;
>
> - igt_vec_fini(&tested_formats);
> - } else {
> test_single_case(data, pipe, output, plane, i,
> - data->override_fmt, test_bad_format);
> + format, test_bad_format);
> }
> - }
> - } else if (is_amdgpu_device(data->gfx_fd)) {
> - uint32_t format = DRM_FORMAT_XRGB8888;
> -
> - igt_output_set_pipe(output, pipe);
> -
> - plane = igt_output_get_plane_type(output, DRM_PLANE_TYPE_PRIMARY);
> - igt_require(igt_plane_has_prop(plane, IGT_PLANE_ROTATION));
>
> - prepare_crtc(data, output, pipe, plane, false);
> -
> - if (plane_type != DRM_PLANE_TYPE_PRIMARY) {
> - plane = igt_output_get_plane_type(output, plane_type);
> - igt_require(igt_plane_has_prop(plane, IGT_PLANE_ROTATION));
> + igt_vec_fini(&tested_formats);
> + } else {
> + test_single_case(data, pipe, output, plane, i,
> + data->override_fmt, test_bad_format);
> }
> -
> - test_single_case(data, pipe, output, plane,
> - rectangle, format, test_bad_format);
> }
> -
> - igt_pipe_crc_stop(data->pipe_crc);
> }
> }
>
> @@ -771,7 +723,7 @@ static void test_plane_rotation_exhaust_fences(data_t *data,
>
> igt_require(igt_plane_has_prop(plane, IGT_PLANE_ROTATION));
>
> - prepare_crtc(data, output, pipe, plane, false);
> + prepare_crtc(data, output, pipe, plane);
>
> mode = igt_output_get_mode(output);
> w = mode->hdisplay;
> @@ -928,6 +880,14 @@ igt_main_args("", long_opts, help_str, opt_handler, &data)
> igt_require(!(subtest->rot &
> (IGT_ROTATION_90 | IGT_ROTATION_270)) ||
> gen >= 9);
> + } else if (is_amdgpu_device(data.gfx_fd)) {
> + data.override_fmt = DRM_FORMAT_XRGB8888;
> + if (subtest->rot & (IGT_ROTATION_90 | IGT_ROTATION_270))
> + data.override_tiling = 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);
> + else
> + data.override_tiling = LOCAL_DRM_FORMAT_MOD_NONE;
> }
> data.rotation = subtest->rot;
> test_plane_rotation(&data, subtest->plane, false);
> --
> 2.25.1
>
> _______________________________________________
> igt-dev mailing list
> igt-dev@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/igt-dev
_______________________________________________
igt-dev mailing list
igt-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/igt-dev
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2021-01-22 7:26 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2021-01-21 23:03 [igt-dev] [PATCH v3] kms_rotation_crc:Add HW rotation test case for amdgpu Sung Joon Kim
2021-01-21 23:08 ` [igt-dev] ✗ Fi.CI.BUILD: failure for kms_rotation_crc:Add HW rotation test case for amdgpu (rev3) Patchwork
2021-01-22 7:26 ` [igt-dev] [PATCH v3] kms_rotation_crc:Add HW rotation test case for amdgpu Petri Latvala
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.