[Mesa-dev] [PATCH 4/6] mesa: add common format-independent memcpy-based ReadPixels path
Marek Olšák
maraeo at gmail.com
Thu Mar 14 11:45:37 PDT 2013
I'll need the _mesa_readpixels_needs_slow_path function for the blit-based
version, but it's also useful to have this memcpy-based path in one place
and not scattered across several functions.
---
src/mesa/main/readpix.c | 194 ++++++++++++++++++++++++++++++++++++++---------
src/mesa/main/readpix.h | 4 +
2 files changed, 164 insertions(+), 34 deletions(-)
diff --git a/src/mesa/main/readpix.c b/src/mesa/main/readpix.c
index e6d0ecd..790cf2a 100644
--- a/src/mesa/main/readpix.c
+++ b/src/mesa/main/readpix.c
@@ -108,11 +108,145 @@ get_readpixels_transfer_ops(struct gl_context *ctx, gl_format texFormat,
/**
- * Tries to implement glReadPixels() of GL_DEPTH_COMPONENT using memcpy of the
- * mapping.
+ * Return true if memcpy cannot be used for ReadPixels.
+ *
+ * If uses_blit is true, the function returns true if a simple 3D engine blit
+ * cannot be used for ReadPixels packing.
+ *
+ * NOTE: This doesn't take swizzling and format conversions between
+ * the readbuffer and the pixel pack buffer into account.
*/
+GLboolean
+_mesa_readpixels_needs_slow_path(struct gl_context *ctx, GLenum format,
+ GLenum type, GLboolean uses_blit)
+{
+ struct gl_renderbuffer *rb =
+ _mesa_get_read_renderbuffer_for_format(ctx, format);
+ GLenum srcType;
+
+ ASSERT(rb);
+
+ /* There are different rules depending on the base format. */
+ switch (format) {
+ case GL_DEPTH_STENCIL:
+ return !_mesa_has_depthstencil_combined(ctx->ReadBuffer) ||
+ ctx->Pixel.DepthScale != 1.0f || ctx->Pixel.DepthBias != 0.0f ||
+ ctx->Pixel.IndexShift || ctx->Pixel.IndexOffset ||
+ ctx->Pixel.MapStencilFlag;
+
+ case GL_DEPTH_COMPONENT:
+ return ctx->Pixel.DepthScale != 1.0f || ctx->Pixel.DepthBias != 0.0f;
+
+ case GL_STENCIL_INDEX:
+ return ctx->Pixel.IndexShift || ctx->Pixel.IndexOffset ||
+ ctx->Pixel.MapStencilFlag;
+
+ default:
+ /* Color formats. */
+ if (need_rgb_to_luminance_conversion(rb->Format, format)) {
+ return GL_TRUE;
+ }
+
+ /* Conversion between signed and unsigned integers needs masking
+ * (it isn't just memcpy). */
+ srcType = _mesa_get_format_datatype(rb->Format);
+
+ if ((srcType == GL_INT &&
+ (type == GL_UNSIGNED_INT ||
+ type == GL_UNSIGNED_SHORT ||
+ type == GL_UNSIGNED_BYTE)) ||
+ (srcType == GL_UNSIGNED_INT &&
+ (type == GL_INT ||
+ type == GL_SHORT ||
+ type == GL_BYTE))) {
+ return GL_TRUE;
+ }
+
+ /* And finally, see if there are any transfer ops. */
+ return get_readpixels_transfer_ops(ctx, rb->Format, format, type,
+ uses_blit) != 0;
+ }
+ return GL_FALSE;
+}
+
+
+static GLboolean
+readpixels_can_use_memcpy(struct gl_context *ctx, GLenum format, GLenum type,
+ const struct gl_pixelstore_attrib *packing)
+{
+ struct gl_renderbuffer *rb =
+ _mesa_get_read_renderbuffer_for_format(ctx, format);
+
+ ASSERT(rb);
+
+ if (_mesa_readpixels_needs_slow_path(ctx, format, type, GL_FALSE)) {
+ return GL_FALSE;
+ }
+
+ /* The base internal format and the base Mesa format must match. */
+ if (rb->_BaseFormat != _mesa_get_format_base_format(rb->Format)) {
+ return GL_FALSE;
+ }
+
+ /* The Mesa format must match the input format and type. */
+ if (!_mesa_format_matches_format_and_type(rb->Format, format, type,
+ packing->SwapBytes)) {
+ return GL_FALSE;
+ }
+
+ return GL_TRUE;
+}
+
+
static GLboolean
-fast_read_depth_pixels( struct gl_context *ctx,
+readpixels_memcpy(struct gl_context *ctx,
+ GLint x, GLint y,
+ GLsizei width, GLsizei height,
+ GLenum format, GLenum type,
+ GLvoid *pixels,
+ const struct gl_pixelstore_attrib *packing)
+{
+ struct gl_renderbuffer *rb =
+ _mesa_get_read_renderbuffer_for_format(ctx, format);
+ GLubyte *dst, *map;
+ int dstStride, stride, j, texelBytes;
+
+ /* Fail if memcpy cannot be used. */
+ if (!readpixels_can_use_memcpy(ctx, format, type, packing)) {
+ return GL_FALSE;
+ }
+
+ dstStride = _mesa_image_row_stride(packing, width, format, type);
+ dst = (GLubyte *) _mesa_image_address2d(packing, pixels, width, height,
+ format, type, 0, 0);
+
+ ctx->Driver.MapRenderbuffer(ctx, rb, x, y, width, height, GL_MAP_READ_BIT,
+ &map, &stride);
+ if (!map) {
+ _mesa_error(ctx, GL_OUT_OF_MEMORY, "glReadPixels");
+ return GL_TRUE; /* don't bother trying the slow path */
+ }
+
+ texelBytes = _mesa_get_format_bytes(rb->Format);
+
+ /* memcpy*/
+ for (j = 0; j < height; j++) {
+ memcpy(dst, map, width * texelBytes);
+ dst += dstStride;
+ map += stride;
+ }
+
+ ctx->Driver.UnmapRenderbuffer(ctx, rb);
+ return GL_TRUE;
+}
+
+
+/**
+ * Optimized path for conversion of depth values to GL_DEPTH_COMPONENT,
+ * GL_UNSIGNED_INT.
+ */
+static GLboolean
+read_uint_depth_pixels( struct gl_context *ctx,
GLint x, GLint y,
GLsizei width, GLsizei height,
GLenum type, GLvoid *pixels,
@@ -132,10 +266,6 @@ fast_read_depth_pixels( struct gl_context *ctx,
if (_mesa_get_format_datatype(rb->Format) != GL_UNSIGNED_NORMALIZED)
return GL_FALSE;
- if (!((type == GL_UNSIGNED_SHORT && rb->Format == MESA_FORMAT_Z16) ||
- type == GL_UNSIGNED_INT))
- return GL_FALSE;
-
ctx->Driver.MapRenderbuffer(ctx, rb, x, y, width, height, GL_MAP_READ_BIT,
&map, &stride);
@@ -149,12 +279,7 @@ fast_read_depth_pixels( struct gl_context *ctx,
GL_DEPTH_COMPONENT, type, 0, 0);
for (j = 0; j < height; j++) {
- if (type == GL_UNSIGNED_INT) {
- _mesa_unpack_uint_z_row(rb->Format, width, map, (GLuint *)dst);
- } else {
- ASSERT(type == GL_UNSIGNED_SHORT && rb->Format == MESA_FORMAT_Z16);
- memcpy(dst, map, width * 2);
- }
+ _mesa_unpack_uint_z_row(rb->Format, width, map, (GLuint *)dst);
map += stride;
dst += dstStride;
@@ -190,8 +315,10 @@ read_depth_pixels( struct gl_context *ctx,
ASSERT(x + width <= (GLint) rb->Width);
ASSERT(y + height <= (GLint) rb->Height);
- if (fast_read_depth_pixels(ctx, x, y, width, height, type, pixels, packing))
+ if (type == GL_UNSIGNED_INT &&
+ read_uint_depth_pixels(ctx, x, y, width, height, type, pixels, packing)) {
return;
+ }
dstStride = _mesa_image_row_stride(packing, width, GL_DEPTH_COMPONENT, type);
dst = (GLubyte *) _mesa_image_address2d(packing, pixels, width, height,
@@ -279,20 +406,20 @@ read_stencil_pixels( struct gl_context *ctx,
/**
- * Try to do glReadPixels of RGBA data using a simple memcpy or swizzle.
+ * Try to do glReadPixels of RGBA data using swizzle.
* \return GL_TRUE if successful, GL_FALSE otherwise (use the slow path)
*/
static GLboolean
-fast_read_rgba_pixels_memcpy( struct gl_context *ctx,
- GLint x, GLint y,
- GLsizei width, GLsizei height,
- GLenum format, GLenum type,
- GLvoid *pixels,
- const struct gl_pixelstore_attrib *packing)
+read_rgba_pixels_swizzle(struct gl_context *ctx,
+ GLint x, GLint y,
+ GLsizei width, GLsizei height,
+ GLenum format, GLenum type,
+ GLvoid *pixels,
+ const struct gl_pixelstore_attrib *packing)
{
struct gl_renderbuffer *rb = ctx->ReadBuffer->_ColorReadBuffer;
GLubyte *dst, *map;
- int dstStride, stride, j, texelBytes;
+ int dstStride, stride, j;
GLboolean swizzle_rb = GL_FALSE, copy_xrgb = GL_FALSE;
/* XXX we could check for other swizzle/special cases here as needed */
@@ -308,9 +435,9 @@ fast_read_rgba_pixels_memcpy( struct gl_context *ctx,
!ctx->Pack.SwapBytes) {
copy_xrgb = GL_TRUE;
}
- else if (!_mesa_format_matches_format_and_type(rb->Format, format, type,
- ctx->Pack.SwapBytes))
+ else {
return GL_FALSE;
+ }
dstStride = _mesa_image_row_stride(packing, width, format, type);
dst = (GLubyte *) _mesa_image_address2d(packing, pixels, width, height,
@@ -323,8 +450,6 @@ fast_read_rgba_pixels_memcpy( struct gl_context *ctx,
return GL_TRUE; /* don't bother trying the slow path */
}
- texelBytes = _mesa_get_format_bytes(rb->Format);
-
if (swizzle_rb) {
/* swap R/B */
for (j = 0; j < height; j++) {
@@ -350,13 +475,6 @@ fast_read_rgba_pixels_memcpy( struct gl_context *ctx,
dst += dstStride;
map += stride;
}
- } else {
- /* just memcpy */
- for (j = 0; j < height; j++) {
- memcpy(dst, map, width * texelBytes);
- dst += dstStride;
- map += stride;
- }
}
ctx->Driver.UnmapRenderbuffer(ctx, rb);
@@ -447,7 +565,7 @@ read_rgba_pixels( struct gl_context *ctx,
/* Try the optimized paths first. */
if (!transferOps &&
- fast_read_rgba_pixels_memcpy(ctx, x, y, width, height,
+ read_rgba_pixels_swizzle(ctx, x, y, width, height,
format, type, pixels, packing)) {
return;
}
@@ -703,6 +821,14 @@ _mesa_readpixels(struct gl_context *ctx,
pixels = _mesa_map_pbo_dest(ctx, &clippedPacking, pixels);
if (pixels) {
+ /* Try memcpy first. */
+ if (readpixels_memcpy(ctx, x, y, width, height, format, type,
+ pixels, packing)) {
+ _mesa_unmap_pbo_dest(ctx, &clippedPacking);
+ return;
+ }
+
+ /* Otherwise take the slow path. */
switch (format) {
case GL_STENCIL_INDEX:
read_stencil_pixels(ctx, x, y, width, height, type, pixels,
diff --git a/src/mesa/main/readpix.h b/src/mesa/main/readpix.h
index 5a5f73f..eb4e45c 100644
--- a/src/mesa/main/readpix.h
+++ b/src/mesa/main/readpix.h
@@ -33,6 +33,10 @@ struct gl_context;
struct gl_pixelstore_attrib;
+extern GLboolean
+_mesa_readpixels_needs_slow_path(struct gl_context *ctx, GLenum format,
+ GLenum type, GLboolean uses_blit);
+
extern void
_mesa_readpixels(struct gl_context *ctx,
GLint x, GLint y, GLsizei width, GLsizei height,
--
1.7.10.4
More information about the mesa-dev
mailing list