#include "intel_reg.h"
#include "intel_regions.h"
#include "intel_batchbuffer.h"
+#include "intel_mipmap_tree.h"
#define FILE_DEBUG_FLAG DEBUG_BLIT
}
}
+static uint32_t
+br13_for_cpp(int cpp)
+{
+ switch (cpp) {
+ case 4:
+ return BR13_8888;
+ break;
+ case 2:
+ return BR13_565;
+ break;
+ case 1:
+ return BR13_8;
+ break;
+ default:
+ assert(0);
+ return 0;
+ }
+}
/* Copy BitBlt
*/
-GLboolean
+bool
intelEmitCopyBlit(struct intel_context *intel,
GLuint cpp,
GLshort src_pitch,
- dri_bo *src_buffer,
+ drm_intel_bo *src_buffer,
GLuint src_offset,
uint32_t src_tiling,
GLshort dst_pitch,
- dri_bo *dst_buffer,
+ drm_intel_bo *dst_buffer,
GLuint dst_offset,
uint32_t dst_tiling,
GLshort src_x, GLshort src_y,
GLuint CMD, BR13, pass = 0;
int dst_y2 = dst_y + h;
int dst_x2 = dst_x + w;
- dri_bo *aper_array[3];
+ drm_intel_bo *aper_array[3];
BATCH_LOCALS;
- /* Blits are in a different ringbuffer so we don't use them. */
- if (intel->gen >= 6)
- return GL_FALSE;
-
if (dst_tiling != I915_TILING_NONE) {
if (dst_offset & 4095)
- return GL_FALSE;
+ return false;
if (dst_tiling == I915_TILING_Y)
- return GL_FALSE;
+ return false;
}
if (src_tiling != I915_TILING_NONE) {
if (src_offset & 4095)
- return GL_FALSE;
+ return false;
if (src_tiling == I915_TILING_Y)
- return GL_FALSE;
+ return false;
}
/* do space check before going any further */
do {
- aper_array[0] = intel->batch->buf;
+ aper_array[0] = intel->batch.bo;
aper_array[1] = dst_buffer;
aper_array[2] = src_buffer;
if (dri_bufmgr_check_aperture_space(aper_array, 3) != 0) {
- intel_batchbuffer_flush(intel->batch);
+ intel_batchbuffer_flush(intel);
pass++;
} else
break;
} while (pass < 2);
if (pass >= 2)
- return GL_FALSE;
+ return false;
- intel_batchbuffer_require_space(intel->batch, 8 * 4);
+ intel_batchbuffer_require_space(intel, 8 * 4, true);
DBG("%s src:buf(%p)/%d+%d %d,%d dst:buf(%p)/%d+%d %d,%d sz:%dx%d\n",
__FUNCTION__,
src_buffer, src_pitch, src_offset, src_x, src_y,
src_pitch *= cpp;
dst_pitch *= cpp;
- BR13 = translate_raster_op(logic_op) << 16;
+ /* Blit pitch must be dword-aligned. Otherwise, the hardware appears to drop
+ * the low bits.
+ */
+ assert(src_pitch % 4 == 0);
+ assert(dst_pitch % 4 == 0);
+
+ /* For big formats (such as floating point), do the copy using 32bpp and
+ * multiply the coordinates.
+ */
+ if (cpp > 4) {
+ assert(cpp % 4 == 0);
+ dst_x *= cpp / 4;
+ dst_x2 *= cpp / 4;
+ src_x *= cpp / 4;
+ cpp = 4;
+ }
+
+ BR13 = br13_for_cpp(cpp) | translate_raster_op(logic_op) << 16;
switch (cpp) {
case 1:
- CMD = XY_SRC_COPY_BLT_CMD;
- break;
case 2:
- BR13 |= BR13_565;
CMD = XY_SRC_COPY_BLT_CMD;
break;
case 4:
- BR13 |= BR13_8888;
CMD = XY_SRC_COPY_BLT_CMD | XY_BLT_WRITE_ALPHA | XY_BLT_WRITE_RGB;
break;
default:
- return GL_FALSE;
+ return false;
}
#ifndef I915
#endif
if (dst_y2 <= dst_y || dst_x2 <= dst_x) {
- return GL_TRUE;
+ return true;
}
assert(dst_x < dst_x2);
assert(dst_y < dst_y2);
- BEGIN_BATCH(8);
+ BEGIN_BATCH_BLT(8);
OUT_BATCH(CMD);
OUT_BATCH(BR13 | (uint16_t)dst_pitch);
OUT_BATCH((dst_y << 16) | dst_x);
src_offset);
ADVANCE_BATCH();
- intel_batchbuffer_emit_mi_flush(intel->batch);
+ intel_batchbuffer_emit_mi_flush(intel);
- return GL_TRUE;
+ return true;
}
* which we're clearing with triangles.
* \param mask bitmask of BUFFER_BIT_* values indicating buffers to clear
*/
-void
-intelClearWithBlit(GLcontext *ctx, GLbitfield mask)
+GLbitfield
+intelClearWithBlit(struct gl_context *ctx, GLbitfield mask)
{
struct intel_context *intel = intel_context(ctx);
struct gl_framebuffer *fb = ctx->DrawBuffer;
- GLuint clear_depth;
- GLboolean all;
+ GLuint clear_depth_value, clear_depth_mask;
GLint cx, cy, cw, ch;
+ GLbitfield fail_mask = 0;
BATCH_LOCALS;
- /* Blits are in a different ringbuffer so we don't use them. */
- assert(intel->gen < 6);
-
/*
* Compute values for clearing the buffers.
*/
- clear_depth = 0;
+ clear_depth_value = 0;
+ clear_depth_mask = 0;
if (mask & BUFFER_BIT_DEPTH) {
- clear_depth = (GLuint) (fb->_DepthMax * ctx->Depth.Clear);
+ clear_depth_value = (GLuint) (fb->_DepthMax * ctx->Depth.Clear);
+ clear_depth_mask = XY_BLT_WRITE_RGB;
}
if (mask & BUFFER_BIT_STENCIL) {
- clear_depth |= (ctx->Stencil.Clear & 0xff) << 24;
+ clear_depth_value |= (ctx->Stencil.Clear & 0xff) << 24;
+ clear_depth_mask |= XY_BLT_WRITE_ALPHA;
}
cx = fb->_Xmin;
ch = fb->_Ymax - fb->_Ymin;
if (cw == 0 || ch == 0)
- return;
-
- GLuint buf;
- all = (cw == fb->Width && ch == fb->Height);
-
- intel_prepare_render(intel);
+ return 0;
/* Loop over all renderbuffers */
- for (buf = 0; buf < BUFFER_COUNT && mask; buf++) {
- const GLbitfield bufBit = 1 << buf;
+ mask &= (1 << BUFFER_COUNT) - 1;
+ while (mask) {
+ GLuint buf = ffs(mask) - 1;
+ bool is_depth_stencil = buf == BUFFER_DEPTH || buf == BUFFER_STENCIL;
struct intel_renderbuffer *irb;
- drm_intel_bo *write_buffer;
int x1, y1, x2, y2;
uint32_t clear_val;
uint32_t BR13, CMD;
+ struct intel_region *region;
int pitch, cpp;
drm_intel_bo *aper_array[2];
- if (!(mask & bufBit))
- continue;
+ mask &= ~(1 << buf);
- /* OK, clear this renderbuffer */
irb = intel_get_renderbuffer(fb, buf);
- write_buffer = intel_region_buffer(intel, irb->region,
- all ? INTEL_WRITE_FULL :
- INTEL_WRITE_PART);
- x1 = cx + irb->region->draw_x;
- y1 = cy + irb->region->draw_y;
- x2 = cx + cw + irb->region->draw_x;
- y2 = cy + ch + irb->region->draw_y;
+ if (irb && irb->mt) {
+ region = irb->mt->region;
+ assert(region);
+ assert(region->bo);
+ } else {
+ fail_mask |= 1 << buf;
+ continue;
+ }
- pitch = irb->region->pitch;
- cpp = irb->region->cpp;
+ /* OK, clear this renderbuffer */
+ x1 = cx + irb->draw_x;
+ y1 = cy + irb->draw_y;
+ x2 = cx + cw + irb->draw_x;
+ y2 = cy + ch + irb->draw_y;
+
+ pitch = region->pitch;
+ cpp = region->cpp;
DBG("%s dst:buf(%p)/%d %d,%d sz:%dx%d\n",
__FUNCTION__,
- irb->region->buffer, (pitch * cpp),
+ region->bo, (pitch * cpp),
x1, y1, x2 - x1, y2 - y1);
BR13 = 0xf0 << 16;
/* Setup the blit command */
if (cpp == 4) {
- BR13 |= BR13_8888;
- if (buf == BUFFER_DEPTH || buf == BUFFER_STENCIL) {
- if (mask & BUFFER_BIT_DEPTH)
- CMD |= XY_BLT_WRITE_RGB;
- if (mask & BUFFER_BIT_STENCIL)
- CMD |= XY_BLT_WRITE_ALPHA;
+ if (is_depth_stencil) {
+ CMD |= clear_depth_mask;
} else {
/* clearing RGBA */
CMD |= XY_BLT_WRITE_ALPHA | XY_BLT_WRITE_RGB;
}
- } else {
- ASSERT(cpp == 2);
- BR13 |= BR13_565;
}
- assert(irb->region->tiling != I915_TILING_Y);
+ assert(region->tiling != I915_TILING_Y);
#ifndef I915
- if (irb->region->tiling != I915_TILING_NONE) {
+ if (region->tiling != I915_TILING_NONE) {
CMD |= XY_DST_TILED;
pitch /= 4;
}
#endif
BR13 |= (pitch * cpp);
- if (buf == BUFFER_DEPTH || buf == BUFFER_STENCIL) {
- clear_val = clear_depth;
+ if (is_depth_stencil) {
+ clear_val = clear_depth_value;
} else {
uint8_t clear[4];
- GLclampf *color = ctx->Color.ClearColor;
+ GLfloat *color = ctx->Color.ClearColor.f;
- CLAMPED_FLOAT_TO_UBYTE(clear[0], color[0]);
- CLAMPED_FLOAT_TO_UBYTE(clear[1], color[1]);
- CLAMPED_FLOAT_TO_UBYTE(clear[2], color[2]);
- CLAMPED_FLOAT_TO_UBYTE(clear[3], color[3]);
+ _mesa_unclamped_float_rgba_to_ubyte(clear, color);
switch (irb->Base.Format) {
case MESA_FORMAT_ARGB8888:
clear_val = PACK_COLOR_1555(clear[3], clear[0],
clear[1], clear[2]);
break;
+ case MESA_FORMAT_A8:
+ clear_val = PACK_COLOR_8888(clear[3], clear[3],
+ clear[3], clear[3]);
+ break;
default:
- _mesa_problem(ctx, "Unexpected renderbuffer format: %d\n",
- irb->Base.Format);
- clear_val = 0;
+ fail_mask |= 1 << buf;
+ continue;
}
}
+ BR13 |= br13_for_cpp(cpp);
+
assert(x1 < x2);
assert(y1 < y2);
/* do space check before going any further */
- aper_array[0] = intel->batch->buf;
- aper_array[1] = write_buffer;
+ aper_array[0] = intel->batch.bo;
+ aper_array[1] = region->bo;
if (drm_intel_bufmgr_check_aperture_space(aper_array,
ARRAY_SIZE(aper_array)) != 0) {
- intel_batchbuffer_flush(intel->batch);
+ intel_batchbuffer_flush(intel);
}
- BEGIN_BATCH(6);
+ BEGIN_BATCH_BLT(6);
OUT_BATCH(CMD);
OUT_BATCH(BR13);
OUT_BATCH((y1 << 16) | x1);
OUT_BATCH((y2 << 16) | x2);
- OUT_RELOC_FENCED(write_buffer,
+ OUT_RELOC_FENCED(region->bo,
I915_GEM_DOMAIN_RENDER, I915_GEM_DOMAIN_RENDER,
0);
OUT_BATCH(clear_val);
ADVANCE_BATCH();
+ if (intel->always_flush_cache)
+ intel_batchbuffer_emit_mi_flush(intel);
+
if (buf == BUFFER_DEPTH || buf == BUFFER_STENCIL)
mask &= ~(BUFFER_BIT_DEPTH | BUFFER_BIT_STENCIL);
- else
- mask &= ~bufBit; /* turn off bit, for faster loop exit */
}
+
+ return fail_mask;
}
-GLboolean
+bool
intelEmitImmediateColorExpandBlit(struct intel_context *intel,
GLuint cpp,
GLubyte *src_bits, GLuint src_size,
GLuint fg_color,
GLshort dst_pitch,
- dri_bo *dst_buffer,
+ drm_intel_bo *dst_buffer,
GLuint dst_offset,
uint32_t dst_tiling,
GLshort x, GLshort y,
int dwords = ALIGN(src_size, 8) / 4;
uint32_t opcode, br13, blit_cmd;
- /* Blits are in a different ringbuffer so we don't use them. */
- if (intel->gen >= 6)
- return GL_FALSE;
-
if (dst_tiling != I915_TILING_NONE) {
if (dst_offset & 4095)
- return GL_FALSE;
+ return false;
if (dst_tiling == I915_TILING_Y)
- return GL_FALSE;
+ return false;
}
assert( logic_op - GL_CLEAR >= 0 );
assert(dst_pitch > 0);
if (w < 0 || h < 0)
- return GL_TRUE;
+ return true;
dst_pitch *= cpp;
__FUNCTION__,
dst_buffer, dst_pitch, dst_offset, x, y, w, h, src_size, dwords);
- intel_batchbuffer_require_space( intel->batch,
- (8 * 4) +
- (3 * 4) +
- dwords * 4 );
+ intel_batchbuffer_require_space(intel,
+ (8 * 4) +
+ (3 * 4) +
+ dwords * 4, true);
opcode = XY_SETUP_BLT_CMD;
if (cpp == 4)
#endif
br13 = dst_pitch | (translate_raster_op(logic_op) << 16) | (1 << 29);
- if (cpp == 2)
- br13 |= BR13_565;
- else
- br13 |= BR13_8888;
+ br13 |= br13_for_cpp(cpp);
blit_cmd = XY_TEXT_IMMEDIATE_BLIT_CMD | XY_TEXT_BYTE_PACKED; /* packing? */
if (dst_tiling != I915_TILING_NONE)
blit_cmd |= XY_DST_TILED;
- BEGIN_BATCH(8 + 3);
+ BEGIN_BATCH_BLT(8 + 3);
OUT_BATCH(opcode);
OUT_BATCH(br13);
OUT_BATCH((0 << 16) | 0); /* clip x1, y1 */
OUT_BATCH(((y + h) << 16) | (x + w));
ADVANCE_BATCH();
- intel_batchbuffer_data( intel->batch,
- src_bits,
- dwords * 4 );
+ intel_batchbuffer_data(intel, src_bits, dwords * 4, true);
- intel_batchbuffer_emit_mi_flush(intel->batch);
+ intel_batchbuffer_emit_mi_flush(intel);
- return GL_TRUE;
+ return true;
}
/* We don't have a memmove-type blit like some other hardware, so we'll do a
unsigned int size)
{
GLuint pitch, height;
- GLboolean ok;
-
- /* Blits are in a different ringbuffer so we don't use them. */
- assert(intel->gen < 6);
+ bool ok;
- /* The pitch is a signed value. */
- pitch = MIN2(size, (1 << 15) - 1);
+ /* The pitch given to the GPU must be DWORD aligned, and
+ * we want width to match pitch. Max width is (1 << 15 - 1),
+ * rounding that down to the nearest DWORD is 1 << 15 - 4
+ */
+ pitch = ROUND_DOWN_TO(MIN2(size, (1 << 15) - 1), 4);
height = size / pitch;
ok = intelEmitCopyBlit(intel, 1,
pitch, src_bo, src_offset, I915_TILING_NONE,
dst_offset += pitch * height;
size -= pitch * height;
assert (size < (1 << 15));
+ pitch = ALIGN(size, 4);
if (size != 0) {
ok = intelEmitCopyBlit(intel, 1,
- size, src_bo, src_offset, I915_TILING_NONE,
- size, dst_bo, dst_offset, I915_TILING_NONE,
+ pitch, src_bo, src_offset, I915_TILING_NONE,
+ pitch, dst_bo, dst_offset, I915_TILING_NONE,
0, 0, /* src x/y */
0, 0, /* dst x/y */
size, 1, /* w, h */
assert(ok);
}
}
+
+/**
+ * Used to initialize the alpha value of an ARGB8888 teximage after
+ * loading it from an XRGB8888 source.
+ *
+ * This is very common with glCopyTexImage2D().
+ */
+void
+intel_set_teximage_alpha_to_one(struct gl_context *ctx,
+ struct intel_texture_image *intel_image)
+{
+ struct intel_context *intel = intel_context(ctx);
+ unsigned int image_x, image_y;
+ uint32_t x1, y1, x2, y2;
+ uint32_t BR13, CMD;
+ int pitch, cpp;
+ drm_intel_bo *aper_array[2];
+ struct intel_region *region = intel_image->mt->region;
+ int width, height, depth;
+ BATCH_LOCALS;
+
+ intel_miptree_get_dimensions_for_image(&intel_image->base.Base,
+ &width, &height, &depth);
+ assert(depth == 1);
+
+ assert(intel_image->base.Base.TexFormat == MESA_FORMAT_ARGB8888);
+
+ /* get dest x/y in destination texture */
+ intel_miptree_get_image_offset(intel_image->mt,
+ intel_image->base.Base.Level,
+ intel_image->base.Base.Face,
+ 0,
+ &image_x, &image_y);
+
+ x1 = image_x;
+ y1 = image_y;
+ x2 = image_x + width;
+ y2 = image_y + height;
+
+ pitch = region->pitch;
+ cpp = region->cpp;
+
+ DBG("%s dst:buf(%p)/%d %d,%d sz:%dx%d\n",
+ __FUNCTION__,
+ intel_image->mt->region->bo, (pitch * cpp),
+ x1, y1, x2 - x1, y2 - y1);
+
+ BR13 = br13_for_cpp(cpp) | 0xf0 << 16;
+ CMD = XY_COLOR_BLT_CMD;
+ CMD |= XY_BLT_WRITE_ALPHA;
+
+ assert(region->tiling != I915_TILING_Y);
+
+#ifndef I915
+ if (region->tiling != I915_TILING_NONE) {
+ CMD |= XY_DST_TILED;
+ pitch /= 4;
+ }
+#endif
+ BR13 |= (pitch * cpp);
+
+ /* do space check before going any further */
+ aper_array[0] = intel->batch.bo;
+ aper_array[1] = region->bo;
+
+ if (drm_intel_bufmgr_check_aperture_space(aper_array,
+ ARRAY_SIZE(aper_array)) != 0) {
+ intel_batchbuffer_flush(intel);
+ }
+
+ BEGIN_BATCH_BLT(6);
+ OUT_BATCH(CMD);
+ OUT_BATCH(BR13);
+ OUT_BATCH((y1 << 16) | x1);
+ OUT_BATCH((y2 << 16) | x2);
+ OUT_RELOC_FENCED(region->bo,
+ I915_GEM_DOMAIN_RENDER, I915_GEM_DOMAIN_RENDER,
+ 0);
+ OUT_BATCH(0xffffffff); /* white, but only alpha gets written */
+ ADVANCE_BATCH();
+
+ intel_batchbuffer_emit_mi_flush(intel);
+}